mirrors
/
SDL
mirror of https://github.com/libsdl-org/SDL
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity
SDL/test/testautomation_video.c

2532 lines
105 KiB
C
Raw Normal View History

Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
/**
* Video test suite
*/
Switch header convention from `#include "SDL.h"` to `#include <SDL3/SDLh>` I ran this script in the include directory: ```sh sed -i '' -e 's,#include "\(SDL.*\)",#include <SDL3/\1>,' *.h ``` I ran this script in the src directory: ```sh for i in ../include/SDL3/SDL*.h do hdr=$(basename $i) if [ x"$(echo $hdr | egrep 'SDL_main|SDL_name|SDL_test|SDL_syswm|SDL_opengl|SDL_egl|SDL_vulkan')" != x ]; then find . -type f -exec sed -i '' -e 's,#include "\('$hdr'\)",#include <SDL3/\1>,' {} \; else find . -type f -exec sed -i '' -e '/#include "'$hdr'"/d' {} \; fi done ``` Fixes https://github.com/libsdl-org/SDL/issues/6575
2022-11-27 07:43:38 +03:00
#include <SDL3/SDL.h>
#include <SDL3/SDL_test.h>
More fix warnings about static function and prototype
2023-03-08 18:12:45 +03:00
#include "testautomation_suites.h"
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
/* Private helpers */
test: don't use wiki urls for documentation comments Also make consistent use of \ as documentation escape character.
2023-02-02 02:21:53 +03:00
/**
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
* Create a test window
*/
Remove more reserved identifiers (#6925)
2022-12-30 00:58:16 +03:00
static SDL_Window *createVideoSuiteTestWindow(const char *title)
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
{
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
SDL_Window *window;
Removed temporary memory from the API It was intended to make the API easier to use, but various automatic garbage collection all had flaws, and making the application periodically clean up temporary memory added cognitive load to using the API, and in many cases was it was difficult to restructure threaded code to handle this. So, we're largely going back to the original system, where the API returns allocated results and you free them. In addition, to solve the problems we originally wanted temporary memory for: * Short strings with a finite count, like device names, get stored in a per-thread string pool. * Events continue to use temporary memory internally, which is cleaned up on the next event processing cycle.
2024-07-27 04:57:18 +03:00
SDL_Window **windows;
video: Implement asynchronous windowing SDL window size, state, and position functions have been considered immediate, with their effects assuming to have taken effect upon successful return of the function. However, several windowing systems handle these requests asynchronously, resulting in the functions blocking until the changes have taken effect, potentially for long periods of time. Additionally, some windowing systems treat these as requests, and can potentially deny or fulfill the request in a manner differently than the application expects, such as not allowing a window to be positioned or sized beyond desktop borders, prohibiting fullscreen, and so on. With these changes, applications can make requests of the window manager that do not block, with the understanding that an associated event will be sent if the request is fulfilled. Currently, size, position, maximize, minimize, and fullscreen calls are handled as asynchronous requests, with events being returned if the request is honored. If the application requires that the change take effect immediately, it can call the new SDL_SyncWindow function, which will attempt to block until the request is fulfilled, or some arbitrary timeout period elapses, the duration of which depends not only on the windowing system, but on the operation requested as well (e.g. a 100ms timeout is fine for most X11 events, but maximizing a window can take considerably longer for some reason). There is also a new hint 'SDL_VIDEO_SYNC_ALL_WINDOW_OPS' that will mimic the old behavior by synchronizing after every window operation with, again, the understanding that using this may result in the associated calls blocking for a relatively long period. The deferred model also results in the window size and position getters not reporting false coordinates anymore, as they only forward what the window manager reports vs allowing applications to set arbitrary values, and fullscreen enter/leave events that were initiated via the window manager update the window state appropriately, where they didn't before. Care was taken to ensure that order of operations is maintained, and that requests are not ignored or dropped. This does require some implicit internal synchronization in the various backends if many requests are made in a short period, as some state and behavior depends on other bits of state that need to be known at that particular point in time, but this isn't something that typical applications will hit, unless they are sending a lot of window state in a short time as the tests do. The automated tests developed to test the previous behavior also resulted in previously undefined behavior being defined and normalized across platforms, particularly when it comes to the sizing and positioning of windows when they are in a fixed-size state, such as maximized or fullscreen. Size and position requests made when the window is not in a movable or resizable state will be deferred until it can be applied, so no requests are lost. These changes fix another long-standing issue with renderers recreating maximized windows, where the original non-maximized size was lost, resulting in the window being restored to the wrong size. All automated video tests pass across all platforms. Overall, the "make a request/get an event" model better reflects how most windowing systems work, and some backends avoid spending significant time blocking while waiting for operations to complete.
2023-10-25 22:09:55 +03:00
SDL_Event event;
SDL_CreateWindow() has been simplified and no longer takes a window position.
2023-03-06 01:44:38 +03:00
int w, h;
Added SDL_GetWindows()
2024-06-24 22:22:12 +03:00
int count;
Update some internal files to use `SDL_WindowFlags`
2024-03-03 22:30:57 +03:00
SDL_WindowFlags flags;
Removed SDL_bool in favor of plain bool We require stdbool.h in the build environment, so we might as well use the plain bool type. If your environment doesn't have stdbool.h, this simple replacement will suffice: typedef signed char bool;
2024-09-18 17:52:28 +03:00
bool needs_renderer = false;
bool needs_events_pumped = false;
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
/* Standard window */
w = SDLTest_RandomIntegerInRange(320, 1024);
h = SDLTest_RandomIntegerInRange(320, 768);
Remove SDL_WINDOW_SHOW flag, as redundant with SDL_WINDOW_HIDDEN
2022-12-28 22:17:55 +03:00
flags = SDL_WINDOW_RESIZABLE | SDL_WINDOW_BORDERLESS;
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
SDL_CreateWindow() has been simplified and no longer takes a window position.
2023-03-06 01:44:38 +03:00
window = SDL_CreateWindow(title, w, h, flags);
SDL_WindowFlags is now 64-bit Fixes https://github.com/libsdl-org/SDL/issues/7321
2024-05-09 20:11:44 +03:00
SDLTest_AssertPass("Call to SDL_CreateWindow('Title',%d,%d,%" SDL_PRIu64 ")", w, h, flags);
Clean up any renderer in SDL_DestroyWindowSurface() Also added an automated test to verify window surface functionality. Fixes https://github.com/libsdl-org/SDL/issues/10133
2024-06-29 12:34:17 +03:00
SDLTest_AssertCheck(window != NULL, "Validate that returned window is not NULL");
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
Added SDL_GetWindows()
2024-06-24 22:22:12 +03:00
/* Check the window is available in the window list */
windows = SDL_GetWindows(&count);
SDLTest_AssertCheck(windows != NULL, "Validate that returned window list is not NULL");
SDLTest_AssertCheck(windows[0] == window, "Validate that the window is first in the window list");
Removed temporary memory from the API It was intended to make the API easier to use, but various automatic garbage collection all had flaws, and making the application periodically clean up temporary memory added cognitive load to using the API, and in many cases was it was difficult to restructure threaded code to handle this. So, we're largely going back to the original system, where the API returns allocated results and you free them. In addition, to solve the problems we originally wanted temporary memory for: * Short strings with a finite count, like device names, get stored in a per-thread string pool. * Events continue to use temporary memory internally, which is cleaned up on the next event processing cycle.
2024-07-27 04:57:18 +03:00
SDL_free(windows);
Added SDL_GetWindows()
2024-06-24 22:22:12 +03:00
tests: Fix automated window grab and positioning tests under Wayland/XWayland When using the Wayland video driver or X11 under XWayland, create a renderer and present a frame as part of window creation, as Wayland requires that a frame be presented for the window to be fully mapped and displayed onscreen. This fixes the grab and expected window size tests. This also disables the window positioning tests when running under the Wayland driver, as Wayland does not allow application windows to position themselves in the desktop space, which renders the tests unreliable and subject to spurious failure.
2023-07-23 18:25:15 +03:00
/* Wayland and XWayland windows require that a frame be presented before they are fully mapped and visible onscreen.
* This is required for the mouse/keyboard grab tests to pass.
*/
if (SDL_strcmp(SDL_GetCurrentVideoDriver(), "wayland") == 0) {
Removed SDL_bool in favor of plain bool We require stdbool.h in the build environment, so we might as well use the plain bool type. If your environment doesn't have stdbool.h, this simple replacement will suffice: typedef signed char bool;
2024-09-18 17:52:28 +03:00
needs_renderer = true;
tests: Fix automated window grab and positioning tests under Wayland/XWayland When using the Wayland video driver or X11 under XWayland, create a renderer and present a frame as part of window creation, as Wayland requires that a frame be presented for the window to be fully mapped and displayed onscreen. This fixes the grab and expected window size tests. This also disables the window positioning tests when running under the Wayland driver, as Wayland does not allow application windows to position themselves in the desktop space, which renders the tests unreliable and subject to spurious failure.
2023-07-23 18:25:15 +03:00
} else if (SDL_strcmp(SDL_GetCurrentVideoDriver(), "x11") == 0) {
/* Try to detect if the x11 driver is running under XWayland */
Added thread-safe environment functions Also marked the existing functions as unsafe, as they can cause crashes if used in multi-threaded applications. As a bonus, since the new functions are hashtable based, hint environment lookups are much faster.
2024-09-14 03:00:15 +03:00
const char *session_type = SDL_GetEnvironmentVariable(SDL_GetEnvironment(), "XDG_SESSION_TYPE");
tests: Fix automated window grab and positioning tests under Wayland/XWayland When using the Wayland video driver or X11 under XWayland, create a renderer and present a frame as part of window creation, as Wayland requires that a frame be presented for the window to be fully mapped and displayed onscreen. This fixes the grab and expected window size tests. This also disables the window positioning tests when running under the Wayland driver, as Wayland does not allow application windows to position themselves in the desktop space, which renders the tests unreliable and subject to spurious failure.
2023-07-23 18:25:15 +03:00
if (session_type && SDL_strcasecmp(session_type, "wayland") == 0) {
Removed SDL_bool in favor of plain bool We require stdbool.h in the build environment, so we might as well use the plain bool type. If your environment doesn't have stdbool.h, this simple replacement will suffice: typedef signed char bool;
2024-09-18 17:52:28 +03:00
needs_renderer = true;
tests: Fix automated window grab and positioning tests under Wayland/XWayland When using the Wayland video driver or X11 under XWayland, create a renderer and present a frame as part of window creation, as Wayland requires that a frame be presented for the window to be fully mapped and displayed onscreen. This fixes the grab and expected window size tests. This also disables the window positioning tests when running under the Wayland driver, as Wayland does not allow application windows to position themselves in the desktop space, which renders the tests unreliable and subject to spurious failure.
2023-07-23 18:25:15 +03:00
}
video: Implement asynchronous windowing SDL window size, state, and position functions have been considered immediate, with their effects assuming to have taken effect upon successful return of the function. However, several windowing systems handle these requests asynchronously, resulting in the functions blocking until the changes have taken effect, potentially for long periods of time. Additionally, some windowing systems treat these as requests, and can potentially deny or fulfill the request in a manner differently than the application expects, such as not allowing a window to be positioned or sized beyond desktop borders, prohibiting fullscreen, and so on. With these changes, applications can make requests of the window manager that do not block, with the understanding that an associated event will be sent if the request is fulfilled. Currently, size, position, maximize, minimize, and fullscreen calls are handled as asynchronous requests, with events being returned if the request is honored. If the application requires that the change take effect immediately, it can call the new SDL_SyncWindow function, which will attempt to block until the request is fulfilled, or some arbitrary timeout period elapses, the duration of which depends not only on the windowing system, but on the operation requested as well (e.g. a 100ms timeout is fine for most X11 events, but maximizing a window can take considerably longer for some reason). There is also a new hint 'SDL_VIDEO_SYNC_ALL_WINDOW_OPS' that will mimic the old behavior by synchronizing after every window operation with, again, the understanding that using this may result in the associated calls blocking for a relatively long period. The deferred model also results in the window size and position getters not reporting false coordinates anymore, as they only forward what the window manager reports vs allowing applications to set arbitrary values, and fullscreen enter/leave events that were initiated via the window manager update the window state appropriately, where they didn't before. Care was taken to ensure that order of operations is maintained, and that requests are not ignored or dropped. This does require some implicit internal synchronization in the various backends if many requests are made in a short period, as some state and behavior depends on other bits of state that need to be known at that particular point in time, but this isn't something that typical applications will hit, unless they are sending a lot of window state in a short time as the tests do. The automated tests developed to test the previous behavior also resulted in previously undefined behavior being defined and normalized across platforms, particularly when it comes to the sizing and positioning of windows when they are in a fixed-size state, such as maximized or fullscreen. Size and position requests made when the window is not in a movable or resizable state will be deferred until it can be applied, so no requests are lost. These changes fix another long-standing issue with renderers recreating maximized windows, where the original non-maximized size was lost, resulting in the window being restored to the wrong size. All automated video tests pass across all platforms. Overall, the "make a request/get an event" model better reflects how most windowing systems work, and some backends avoid spending significant time blocking while waiting for operations to complete.
2023-10-25 22:09:55 +03:00
/* X11 needs the initial events pumped, or it can erroneously deliver old configuration events at a later time. */
Removed SDL_bool in favor of plain bool We require stdbool.h in the build environment, so we might as well use the plain bool type. If your environment doesn't have stdbool.h, this simple replacement will suffice: typedef signed char bool;
2024-09-18 17:52:28 +03:00
needs_events_pumped = true;
tests: Fix automated window grab and positioning tests under Wayland/XWayland When using the Wayland video driver or X11 under XWayland, create a renderer and present a frame as part of window creation, as Wayland requires that a frame be presented for the window to be fully mapped and displayed onscreen. This fixes the grab and expected window size tests. This also disables the window positioning tests when running under the Wayland driver, as Wayland does not allow application windows to position themselves in the desktop space, which renders the tests unreliable and subject to spurious failure.
2023-07-23 18:25:15 +03:00
}
if (needs_renderer) {
Removed SDL_RendererFlags The flags parameter has been removed from SDL_CreateRenderer() and SDL_RENDERER_PRESENTVSYNC has been replaced with SDL_PROP_RENDERER_CREATE_PRESENT_VSYNC_NUMBER during window creation and SDL_PROP_RENDERER_VSYNC_NUMBER after renderer creation. SDL_SetRenderVSync() now takes additional values besides 0 and 1. The maximum texture size has been removed from SDL_RendererInfo, replaced with SDL_PROP_RENDERER_MAX_TEXTURE_SIZE_NUMBER.
2024-05-13 20:31:55 +03:00
SDL_Renderer *renderer = SDL_CreateRenderer(window, NULL);
tests: Fix automated window grab and positioning tests under Wayland/XWayland When using the Wayland video driver or X11 under XWayland, create a renderer and present a frame as part of window creation, as Wayland requires that a frame be presented for the window to be fully mapped and displayed onscreen. This fixes the grab and expected window size tests. This also disables the window positioning tests when running under the Wayland driver, as Wayland does not allow application windows to position themselves in the desktop space, which renders the tests unreliable and subject to spurious failure.
2023-07-23 18:25:15 +03:00
if (renderer) {
SDL_SetRenderDrawColor(renderer, 0x00, 0x00, 0x00, 0xFF);
SDL_RenderClear(renderer);
SDL_RenderPresent(renderer);
/* Some desktops don't display the window immediately after presentation,
* so delay to give the window time to actually appear on the desktop.
*/
SDL_Delay(100);
} else {
SDLTest_Log("Unable to create a renderer, some tests may fail on Wayland/XWayland");
}
}
video: Implement asynchronous windowing SDL window size, state, and position functions have been considered immediate, with their effects assuming to have taken effect upon successful return of the function. However, several windowing systems handle these requests asynchronously, resulting in the functions blocking until the changes have taken effect, potentially for long periods of time. Additionally, some windowing systems treat these as requests, and can potentially deny or fulfill the request in a manner differently than the application expects, such as not allowing a window to be positioned or sized beyond desktop borders, prohibiting fullscreen, and so on. With these changes, applications can make requests of the window manager that do not block, with the understanding that an associated event will be sent if the request is fulfilled. Currently, size, position, maximize, minimize, and fullscreen calls are handled as asynchronous requests, with events being returned if the request is honored. If the application requires that the change take effect immediately, it can call the new SDL_SyncWindow function, which will attempt to block until the request is fulfilled, or some arbitrary timeout period elapses, the duration of which depends not only on the windowing system, but on the operation requested as well (e.g. a 100ms timeout is fine for most X11 events, but maximizing a window can take considerably longer for some reason). There is also a new hint 'SDL_VIDEO_SYNC_ALL_WINDOW_OPS' that will mimic the old behavior by synchronizing after every window operation with, again, the understanding that using this may result in the associated calls blocking for a relatively long period. The deferred model also results in the window size and position getters not reporting false coordinates anymore, as they only forward what the window manager reports vs allowing applications to set arbitrary values, and fullscreen enter/leave events that were initiated via the window manager update the window state appropriately, where they didn't before. Care was taken to ensure that order of operations is maintained, and that requests are not ignored or dropped. This does require some implicit internal synchronization in the various backends if many requests are made in a short period, as some state and behavior depends on other bits of state that need to be known at that particular point in time, but this isn't something that typical applications will hit, unless they are sending a lot of window state in a short time as the tests do. The automated tests developed to test the previous behavior also resulted in previously undefined behavior being defined and normalized across platforms, particularly when it comes to the sizing and positioning of windows when they are in a fixed-size state, such as maximized or fullscreen. Size and position requests made when the window is not in a movable or resizable state will be deferred until it can be applied, so no requests are lost. These changes fix another long-standing issue with renderers recreating maximized windows, where the original non-maximized size was lost, resulting in the window being restored to the wrong size. All automated video tests pass across all platforms. Overall, the "make a request/get an event" model better reflects how most windowing systems work, and some backends avoid spending significant time blocking while waiting for operations to complete.
2023-10-25 22:09:55 +03:00
if (needs_events_pumped) {
/* Pump out the event queue */
while (SDL_PollEvent(&event)) {
}
}
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
return window;
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
}
test: don't use wiki urls for documentation comments Also make consistent use of \ as documentation escape character.
2023-02-02 02:21:53 +03:00
/**
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
* Destroy test window
*/
Remove more reserved identifiers (#6925)
2022-12-30 00:58:16 +03:00
static void destroyVideoSuiteTestWindow(SDL_Window *window)
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
{
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
if (window != NULL) {
tests: Clean up the renderer in the video test suite since it isn't automatic anymore.
2024-04-19 19:19:10 +03:00
SDL_Renderer *renderer = SDL_GetRenderer(window);
if (renderer) {
SDL_DestroyRenderer(renderer);
}
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
SDL_DestroyWindow(window);
window = NULL;
SDLTest_AssertPass("Call to SDL_DestroyWindow()");
}
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
}
/* Test case functions */
/**
docs: Remove Doxygen `\brief` tags. Doxygen and the wiki bridge don't need them; they'll both just use the first line/sentence instead. Fixes #8446.
2023-11-06 18:26:06 +03:00
* Enable and disable screensaver while checking state
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
*/
SDL_test: use SDLCALL calling convention This is needed when using a pre-built static SDL3_test library.
2024-09-06 04:21:13 +03:00
static int SDLCALL video_enableDisableScreensaver(void *arg)
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
{
Removed SDL_bool in favor of plain bool We require stdbool.h in the build environment, so we might as well use the plain bool type. If your environment doesn't have stdbool.h, this simple replacement will suffice: typedef signed char bool;
2024-09-18 17:52:28 +03:00
bool initialResult;
bool result;
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
/* Get current state and proceed according to current state */
SDL API renaming: *Is* functions Feedback from @icculus: "IsTablet" uses "is" as a form of "to be" ...like, the actual question is of its nature. The rest is just a superfluous word in the question and it flows as better English with if (RectEmpty) than if (IsRectEmpty) Fixes https://github.com/libsdl-org/SDL/issues/6932
2022-12-29 06:34:01 +03:00
initialResult = SDL_ScreenSaverEnabled();
SDLTest_AssertPass("Call to SDL_ScreenSaverEnabled()");
Removed SDL_bool in favor of plain bool We require stdbool.h in the build environment, so we might as well use the plain bool type. If your environment doesn't have stdbool.h, this simple replacement will suffice: typedef signed char bool;
2024-09-18 17:52:28 +03:00
if (initialResult == true) {
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
/* Currently enabled: disable first, then enable again */
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
/* Disable screensaver and check */
SDL_DisableScreenSaver();
SDLTest_AssertPass("Call to SDL_DisableScreenSaver()");
SDL API renaming: *Is* functions Feedback from @icculus: "IsTablet" uses "is" as a form of "to be" ...like, the actual question is of its nature. The rest is just a superfluous word in the question and it flows as better English with if (RectEmpty) than if (IsRectEmpty) Fixes https://github.com/libsdl-org/SDL/issues/6932
2022-12-29 06:34:01 +03:00
result = SDL_ScreenSaverEnabled();
SDLTest_AssertPass("Call to SDL_ScreenSaverEnabled()");
Removed SDL_bool in favor of plain bool We require stdbool.h in the build environment, so we might as well use the plain bool type. If your environment doesn't have stdbool.h, this simple replacement will suffice: typedef signed char bool;
2024-09-18 17:52:28 +03:00
SDLTest_AssertCheck(result == false, "Verify result from SDL_ScreenSaverEnabled, expected: %i, got: %i", false, result);
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
/* Enable screensaver and check */
SDL_EnableScreenSaver();
SDLTest_AssertPass("Call to SDL_EnableScreenSaver()");
SDL API renaming: *Is* functions Feedback from @icculus: "IsTablet" uses "is" as a form of "to be" ...like, the actual question is of its nature. The rest is just a superfluous word in the question and it flows as better English with if (RectEmpty) than if (IsRectEmpty) Fixes https://github.com/libsdl-org/SDL/issues/6932
2022-12-29 06:34:01 +03:00
result = SDL_ScreenSaverEnabled();
SDLTest_AssertPass("Call to SDL_ScreenSaverEnabled()");
Removed SDL_bool in favor of plain bool We require stdbool.h in the build environment, so we might as well use the plain bool type. If your environment doesn't have stdbool.h, this simple replacement will suffice: typedef signed char bool;
2024-09-18 17:52:28 +03:00
SDLTest_AssertCheck(result == true, "Verify result from SDL_ScreenSaverEnabled, expected: %i, got: %i", true, result);
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
} else {
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
/* Currently disabled: enable first, then disable again */
/* Enable screensaver and check */
SDL_EnableScreenSaver();
SDLTest_AssertPass("Call to SDL_EnableScreenSaver()");
SDL API renaming: *Is* functions Feedback from @icculus: "IsTablet" uses "is" as a form of "to be" ...like, the actual question is of its nature. The rest is just a superfluous word in the question and it flows as better English with if (RectEmpty) than if (IsRectEmpty) Fixes https://github.com/libsdl-org/SDL/issues/6932
2022-12-29 06:34:01 +03:00
result = SDL_ScreenSaverEnabled();
SDLTest_AssertPass("Call to SDL_ScreenSaverEnabled()");
Removed SDL_bool in favor of plain bool We require stdbool.h in the build environment, so we might as well use the plain bool type. If your environment doesn't have stdbool.h, this simple replacement will suffice: typedef signed char bool;
2024-09-18 17:52:28 +03:00
SDLTest_AssertCheck(result == true, "Verify result from SDL_ScreenSaverEnabled, expected: %i, got: %i", true, result);
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
/* Disable screensaver and check */
SDL_DisableScreenSaver();
SDLTest_AssertPass("Call to SDL_DisableScreenSaver()");
SDL API renaming: *Is* functions Feedback from @icculus: "IsTablet" uses "is" as a form of "to be" ...like, the actual question is of its nature. The rest is just a superfluous word in the question and it flows as better English with if (RectEmpty) than if (IsRectEmpty) Fixes https://github.com/libsdl-org/SDL/issues/6932
2022-12-29 06:34:01 +03:00
result = SDL_ScreenSaverEnabled();
SDLTest_AssertPass("Call to SDL_ScreenSaverEnabled()");
Removed SDL_bool in favor of plain bool We require stdbool.h in the build environment, so we might as well use the plain bool type. If your environment doesn't have stdbool.h, this simple replacement will suffice: typedef signed char bool;
2024-09-18 17:52:28 +03:00
SDLTest_AssertCheck(result == false, "Verify result from SDL_ScreenSaverEnabled, expected: %i, got: %i", false, result);
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
}
return TEST_COMPLETED;
}
/**
docs: Remove Doxygen `\brief` tags. Doxygen and the wiki bridge don't need them; they'll both just use the first line/sentence instead. Fixes #8446.
2023-11-06 18:26:06 +03:00
* Tests the functionality of the SDL_CreateWindow function using different sizes
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
*/
SDL_test: use SDLCALL calling convention This is needed when using a pre-built static SDL3_test library.
2024-09-06 04:21:13 +03:00
static int SDLCALL video_createWindowVariousSizes(void *arg)
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
{
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
SDL_Window *window;
const char *title = "video_createWindowVariousSizes Test Window";
More fix warnings about static function and prototype
2023-03-08 18:12:45 +03:00
int w = 0, h = 0;
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
int wVariation, hVariation;
for (wVariation = 0; wVariation < 3; wVariation++) {
for (hVariation = 0; hVariation < 3; hVariation++) {
switch (wVariation) {
case 0:
/* Width of 1 */
w = 1;
break;
case 1:
/* Random "normal" width */
w = SDLTest_RandomIntegerInRange(320, 1920);
break;
case 2:
/* Random "large" width */
w = SDLTest_RandomIntegerInRange(2048, 4095);
break;
}
switch (hVariation) {
case 0:
/* Height of 1 */
h = 1;
break;
case 1:
/* Random "normal" height */
h = SDLTest_RandomIntegerInRange(320, 1080);
break;
case 2:
/* Random "large" height */
h = SDLTest_RandomIntegerInRange(2048, 4095);
break;
}
SDL_CreateWindow() has been simplified and no longer takes a window position.
2023-03-06 01:44:38 +03:00
window = SDL_CreateWindow(title, w, h, 0);
SDLTest_AssertPass("Call to SDL_CreateWindow('Title',%d,%d,SHOWN)", w, h);
Clean up any renderer in SDL_DestroyWindowSurface() Also added an automated test to verify window surface functionality. Fixes https://github.com/libsdl-org/SDL/issues/10133
2024-06-29 12:34:17 +03:00
SDLTest_AssertCheck(window != NULL, "Validate that returned window is not NULL");
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
/* Clean up */
Remove more reserved identifiers (#6925)
2022-12-30 00:58:16 +03:00
destroyVideoSuiteTestWindow(window);
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
}
}
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
return TEST_COMPLETED;
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
}
/**
docs: Remove Doxygen `\brief` tags. Doxygen and the wiki bridge don't need them; they'll both just use the first line/sentence instead. Fixes #8446.
2023-11-06 18:26:06 +03:00
* Tests the functionality of the SDL_CreateWindow function using different flags
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
*/
SDL_test: use SDLCALL calling convention This is needed when using a pre-built static SDL3_test library.
2024-09-06 04:21:13 +03:00
static int SDLCALL video_createWindowVariousFlags(void *arg)
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
{
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
SDL_Window *window;
const char *title = "video_createWindowVariousFlags Test Window";
SDL_CreateWindow() has been simplified and no longer takes a window position.
2023-03-06 01:44:38 +03:00
int w, h;
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
int fVariation;
Update some internal files to use `SDL_WindowFlags`
2024-03-03 22:30:57 +03:00
SDL_WindowFlags flags;
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
/* Standard window */
w = SDLTest_RandomIntegerInRange(320, 1024);
h = SDLTest_RandomIntegerInRange(320, 768);
Folded SDL_WINDOW_FULLSCREEN_EXCLUSIVE and SDL_WINDOW_FULLSCREEN_DESKTOP into a single SDL_WINDOW_FULLSCREEN flag The fullscreen video mode used by the window can be used to determine whether it's in exclusive fullscreen or fullscreen desktop mode.
2023-02-01 22:30:28 +03:00
for (fVariation = 1; fVariation < 14; fVariation++) {
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
switch (fVariation) {
default:
case 1:
Folded SDL_WINDOW_FULLSCREEN_EXCLUSIVE and SDL_WINDOW_FULLSCREEN_DESKTOP into a single SDL_WINDOW_FULLSCREEN flag The fullscreen video mode used by the window can be used to determine whether it's in exclusive fullscreen or fullscreen desktop mode.
2023-02-01 22:30:28 +03:00
flags = SDL_WINDOW_FULLSCREEN;
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
/* Skip - blanks screen; comment out next line to run test */
continue;
break;
case 2:
flags = SDL_WINDOW_OPENGL;
Fixed automation tests using the dummy video driver
2023-07-25 21:48:10 +03:00
/* Skip - not every video driver supports OpenGL; comment out next line to run test */
continue;
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
break;
case 3:
Remove SDL_WINDOW_SHOW flag, as redundant with SDL_WINDOW_HIDDEN
2022-12-28 22:17:55 +03:00
flags = 0;
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
break;
case 4:
flags = SDL_WINDOW_HIDDEN;
break;
case 5:
flags = SDL_WINDOW_BORDERLESS;
break;
case 6:
flags = SDL_WINDOW_RESIZABLE;
break;
case 7:
flags = SDL_WINDOW_MINIMIZED;
break;
case 8:
flags = SDL_WINDOW_MAXIMIZED;
break;
case 9:
flags = SDL_WINDOW_MOUSE_GRABBED;
break;
case 10:
flags = SDL_WINDOW_INPUT_FOCUS;
break;
case 11:
flags = SDL_WINDOW_MOUSE_FOCUS;
break;
case 12:
Renamed SDL_WINDOW_FOREIGN to SDL_WINDOW_EXTERNAL
2023-11-13 23:21:23 +03:00
flags = SDL_WINDOW_EXTERNAL;
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
break;
case 13:
flags = SDL_WINDOW_KEYBOARD_GRABBED;
break;
}
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
SDL_CreateWindow() has been simplified and no longer takes a window position.
2023-03-06 01:44:38 +03:00
window = SDL_CreateWindow(title, w, h, flags);
SDL_WindowFlags is now 64-bit Fixes https://github.com/libsdl-org/SDL/issues/7321
2024-05-09 20:11:44 +03:00
SDLTest_AssertPass("Call to SDL_CreateWindow('Title',%d,%d,%" SDL_PRIu64 ")", w, h, flags);
Clean up any renderer in SDL_DestroyWindowSurface() Also added an automated test to verify window surface functionality. Fixes https://github.com/libsdl-org/SDL/issues/10133
2024-06-29 12:34:17 +03:00
SDLTest_AssertCheck(window != NULL, "Validate that returned window is not NULL");
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
/* Clean up */
Remove more reserved identifiers (#6925)
2022-12-30 00:58:16 +03:00
destroyVideoSuiteTestWindow(window);
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
}
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
return TEST_COMPLETED;
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
}
/**
docs: Remove Doxygen `\brief` tags. Doxygen and the wiki bridge don't need them; they'll both just use the first line/sentence instead. Fixes #8446.
2023-11-06 18:26:06 +03:00
* Tests the functionality of the SDL_GetWindowFlags function
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
*/
SDL_test: use SDLCALL calling convention This is needed when using a pre-built static SDL3_test library.
2024-09-06 04:21:13 +03:00
static int SDLCALL video_getWindowFlags(void *arg)
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
{
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
SDL_Window *window;
const char *title = "video_getWindowFlags Test Window";
Update some internal files to use `SDL_WindowFlags`
2024-03-03 22:30:57 +03:00
SDL_WindowFlags flags;
SDL_WindowFlags is now 64-bit Fixes https://github.com/libsdl-org/SDL/issues/7321
2024-05-09 20:11:44 +03:00
SDL_WindowFlags actualFlags;
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
/* Reliable flag set always set in test window */
Remove SDL_WINDOW_SHOW flag, as redundant with SDL_WINDOW_HIDDEN
2022-12-28 22:17:55 +03:00
flags = 0;
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
/* Call against new test window */
Remove more reserved identifiers (#6925)
2022-12-30 00:58:16 +03:00
window = createVideoSuiteTestWindow(title);
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
if (window != NULL) {
actualFlags = SDL_GetWindowFlags(window);
SDLTest_AssertPass("Call to SDL_GetWindowFlags()");
SDL_WindowFlags is now 64-bit Fixes https://github.com/libsdl-org/SDL/issues/7321
2024-05-09 20:11:44 +03:00
SDLTest_AssertCheck((flags & actualFlags) == flags, "Verify returned value has flags %" SDL_PRIu64 " set, got: %" SDL_PRIu64, flags, actualFlags);
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
}
/* Clean up */
Remove more reserved identifiers (#6925)
2022-12-30 00:58:16 +03:00
destroyVideoSuiteTestWindow(window);
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
return TEST_COMPLETED;
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
}
/**
docs: Remove Doxygen `\brief` tags. Doxygen and the wiki bridge don't need them; they'll both just use the first line/sentence instead. Fixes #8446.
2023-11-06 18:26:06 +03:00
* Tests the functionality of the SDL_GetFullscreenDisplayModes function
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
*/
SDL_test: use SDLCALL calling convention This is needed when using a pre-built static SDL3_test library.
2024-09-06 04:21:13 +03:00
static int SDLCALL video_getFullscreenDisplayModes(void *arg)
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
{
Removed temporary memory from the API It was intended to make the API easier to use, but various automatic garbage collection all had flaws, and making the application periodically clean up temporary memory added cognitive load to using the API, and in many cases was it was difficult to restructure threaded code to handle this. So, we're largely going back to the original system, where the API returns allocated results and you free them. In addition, to solve the problems we originally wanted temporary memory for: * Short strings with a finite count, like device names, get stored in a per-thread string pool. * Events continue to use temporary memory internally, which is cleaned up on the next event processing cycle.
2024-07-27 04:57:18 +03:00
SDL_DisplayID *displays;
SDL_DisplayMode **modes;
Windows default to fullscreen desktop mode if they don't pick an explicit video mode Rather than iterating over display modes using an index, there is a new function SDL_GetFullscreenDisplayModes() to get the list of available fullscreen modes on a display. { SDL_DisplayID display = SDL_GetPrimaryDisplay(); int num_modes = 0; SDL_DisplayMode **modes = SDL_GetFullscreenDisplayModes(display, &num_modes); if (modes) { for (i = 0; i < num_modes; ++i) { SDL_DisplayMode *mode = modes[i]; SDL_Log("Display %" SDL_PRIu32 " mode %d: %dx%d@%gHz, %d%% scale\n", display, i, mode->pixel_w, mode->pixel_h, mode->refresh_rate, (int)(mode->display_scale * 100.0f)); } SDL_free(modes); } } SDL_GetDesktopDisplayMode() and SDL_GetCurrentDisplayMode() return pointers to display modes rather than filling in application memory. Windows now have an explicit fullscreen mode that is set, using SDL_SetWindowFullscreenMode(). The fullscreen mode for a window can be queried with SDL_GetWindowFullscreenMode(), which returns a pointer to the mode, or NULL if the window will be fullscreen desktop. SDL_SetWindowFullscreen() just takes a boolean value, setting the correct fullscreen state based on the selected mode.
2023-02-01 08:23:14 +03:00
int count;
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
int i;
/* Get number of displays */
Displays are now referenced by instance ID instead of index
2023-01-30 00:30:55 +03:00
displays = SDL_GetDisplays(NULL);
if (displays) {
SDLTest_AssertPass("Call to SDL_GetDisplays()");
/* Make call for each display */
for (i = 0; displays[i]; ++i) {
Windows default to fullscreen desktop mode if they don't pick an explicit video mode Rather than iterating over display modes using an index, there is a new function SDL_GetFullscreenDisplayModes() to get the list of available fullscreen modes on a display. { SDL_DisplayID display = SDL_GetPrimaryDisplay(); int num_modes = 0; SDL_DisplayMode **modes = SDL_GetFullscreenDisplayModes(display, &num_modes); if (modes) { for (i = 0; i < num_modes; ++i) { SDL_DisplayMode *mode = modes[i]; SDL_Log("Display %" SDL_PRIu32 " mode %d: %dx%d@%gHz, %d%% scale\n", display, i, mode->pixel_w, mode->pixel_h, mode->refresh_rate, (int)(mode->display_scale * 100.0f)); } SDL_free(modes); } } SDL_GetDesktopDisplayMode() and SDL_GetCurrentDisplayMode() return pointers to display modes rather than filling in application memory. Windows now have an explicit fullscreen mode that is set, using SDL_SetWindowFullscreenMode(). The fullscreen mode for a window can be queried with SDL_GetWindowFullscreenMode(), which returns a pointer to the mode, or NULL if the window will be fullscreen desktop. SDL_SetWindowFullscreen() just takes a boolean value, setting the correct fullscreen state based on the selected mode.
2023-02-01 08:23:14 +03:00
modes = SDL_GetFullscreenDisplayModes(displays[i], &count);
SDLTest_AssertPass("Call to SDL_GetFullscreenDisplayModes(%" SDL_PRIu32 ")", displays[i]);
SDLTest_AssertCheck(modes != NULL, "Validate returned value from function; expected != NULL; got: %p", modes);
SDLTest_AssertCheck(count >= 0, "Validate number of modes; expected: >= 0; got: %d", count);
Removed temporary memory from the API It was intended to make the API easier to use, but various automatic garbage collection all had flaws, and making the application periodically clean up temporary memory added cognitive load to using the API, and in many cases was it was difficult to restructure threaded code to handle this. So, we're largely going back to the original system, where the API returns allocated results and you free them. In addition, to solve the problems we originally wanted temporary memory for: * Short strings with a finite count, like device names, get stored in a per-thread string pool. * Events continue to use temporary memory internally, which is cleaned up on the next event processing cycle.
2024-07-27 04:57:18 +03:00
SDL_free(modes);
Displays are now referenced by instance ID instead of index
2023-01-30 00:30:55 +03:00
}
Removed temporary memory from the API It was intended to make the API easier to use, but various automatic garbage collection all had flaws, and making the application periodically clean up temporary memory added cognitive load to using the API, and in many cases was it was difficult to restructure threaded code to handle this. So, we're largely going back to the original system, where the API returns allocated results and you free them. In addition, to solve the problems we originally wanted temporary memory for: * Short strings with a finite count, like device names, get stored in a per-thread string pool. * Events continue to use temporary memory internally, which is cleaned up on the next event processing cycle.
2024-07-27 04:57:18 +03:00
SDL_free(displays);
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
}
return TEST_COMPLETED;
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
}
/**
docs: Remove Doxygen `\brief` tags. Doxygen and the wiki bridge don't need them; they'll both just use the first line/sentence instead. Fixes #8446.
2023-11-06 18:26:06 +03:00
* Tests the functionality of the SDL_GetClosestFullscreenDisplayMode function against current resolution
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
*/
SDL_test: use SDLCALL calling convention This is needed when using a pre-built static SDL3_test library.
2024-09-06 04:21:13 +03:00
static int SDLCALL video_getClosestDisplayModeCurrentResolution(void *arg)
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
{
Removed temporary memory from the API It was intended to make the API easier to use, but various automatic garbage collection all had flaws, and making the application periodically clean up temporary memory added cognitive load to using the API, and in many cases was it was difficult to restructure threaded code to handle this. So, we're largely going back to the original system, where the API returns allocated results and you free them. In addition, to solve the problems we originally wanted temporary memory for: * Short strings with a finite count, like device names, get stored in a per-thread string pool. * Events continue to use temporary memory internally, which is cleaned up on the next event processing cycle.
2024-07-27 04:57:18 +03:00
SDL_DisplayID *displays;
SDL_DisplayMode **modes;
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
SDL_DisplayMode current;
Removed temporary memory from the API It was intended to make the API easier to use, but various automatic garbage collection all had flaws, and making the application periodically clean up temporary memory added cognitive load to using the API, and in many cases was it was difficult to restructure threaded code to handle this. So, we're largely going back to the original system, where the API returns allocated results and you free them. In addition, to solve the problems we originally wanted temporary memory for: * Short strings with a finite count, like device names, get stored in a per-thread string pool. * Events continue to use temporary memory internally, which is cleaned up on the next event processing cycle.
2024-07-27 04:57:18 +03:00
SDL_DisplayMode closest;
int i, result, num_modes;
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
/* Get number of displays */
Displays are now referenced by instance ID instead of index
2023-01-30 00:30:55 +03:00
displays = SDL_GetDisplays(NULL);
if (displays) {
SDLTest_AssertPass("Call to SDL_GetDisplays()");
/* Make calls for each display */
for (i = 0; displays[i]; ++i) {
remove several instances of unnecessary emptry str "" after SDL_PRI???
2024-01-18 17:25:50 +03:00
SDLTest_Log("Testing against display: %" SDL_PRIu32, displays[i]);
Displays are now referenced by instance ID instead of index
2023-01-30 00:30:55 +03:00
/* Get first display mode to get a sane resolution; this should always work */
Windows default to fullscreen desktop mode if they don't pick an explicit video mode Rather than iterating over display modes using an index, there is a new function SDL_GetFullscreenDisplayModes() to get the list of available fullscreen modes on a display. { SDL_DisplayID display = SDL_GetPrimaryDisplay(); int num_modes = 0; SDL_DisplayMode **modes = SDL_GetFullscreenDisplayModes(display, &num_modes); if (modes) { for (i = 0; i < num_modes; ++i) { SDL_DisplayMode *mode = modes[i]; SDL_Log("Display %" SDL_PRIu32 " mode %d: %dx%d@%gHz, %d%% scale\n", display, i, mode->pixel_w, mode->pixel_h, mode->refresh_rate, (int)(mode->display_scale * 100.0f)); } SDL_free(modes); } } SDL_GetDesktopDisplayMode() and SDL_GetCurrentDisplayMode() return pointers to display modes rather than filling in application memory. Windows now have an explicit fullscreen mode that is set, using SDL_SetWindowFullscreenMode(). The fullscreen mode for a window can be queried with SDL_GetWindowFullscreenMode(), which returns a pointer to the mode, or NULL if the window will be fullscreen desktop. SDL_SetWindowFullscreen() just takes a boolean value, setting the correct fullscreen state based on the selected mode.
2023-02-01 08:23:14 +03:00
modes = SDL_GetFullscreenDisplayModes(displays[i], &num_modes);
SDLTest_AssertPass("Call to SDL_GetDisplayModes()");
SDLTest_Assert(modes != NULL, "Verify returned value is not NULL");
if (num_modes > 0) {
SDL_memcpy(&current, modes[0], sizeof(current));
Displays are now referenced by instance ID instead of index
2023-01-30 00:30:55 +03:00
/* Make call */
Removed SDL_bool in favor of plain bool We require stdbool.h in the build environment, so we might as well use the plain bool type. If your environment doesn't have stdbool.h, this simple replacement will suffice: typedef signed char bool;
2024-09-18 17:52:28 +03:00
result = SDL_GetClosestFullscreenDisplayMode(displays[i], current.w, current.h, current.refresh_rate, false, &closest);
Windows default to fullscreen desktop mode if they don't pick an explicit video mode Rather than iterating over display modes using an index, there is a new function SDL_GetFullscreenDisplayModes() to get the list of available fullscreen modes on a display. { SDL_DisplayID display = SDL_GetPrimaryDisplay(); int num_modes = 0; SDL_DisplayMode **modes = SDL_GetFullscreenDisplayModes(display, &num_modes); if (modes) { for (i = 0; i < num_modes; ++i) { SDL_DisplayMode *mode = modes[i]; SDL_Log("Display %" SDL_PRIu32 " mode %d: %dx%d@%gHz, %d%% scale\n", display, i, mode->pixel_w, mode->pixel_h, mode->refresh_rate, (int)(mode->display_scale * 100.0f)); } SDL_free(modes); } } SDL_GetDesktopDisplayMode() and SDL_GetCurrentDisplayMode() return pointers to display modes rather than filling in application memory. Windows now have an explicit fullscreen mode that is set, using SDL_SetWindowFullscreenMode(). The fullscreen mode for a window can be queried with SDL_GetWindowFullscreenMode(), which returns a pointer to the mode, or NULL if the window will be fullscreen desktop. SDL_SetWindowFullscreen() just takes a boolean value, setting the correct fullscreen state based on the selected mode.
2023-02-01 08:23:14 +03:00
SDLTest_AssertPass("Call to SDL_GetClosestFullscreenDisplayMode(target=current)");
Removed SDL_bool in favor of plain bool We require stdbool.h in the build environment, so we might as well use the plain bool type. If your environment doesn't have stdbool.h, this simple replacement will suffice: typedef signed char bool;
2024-09-18 17:52:28 +03:00
SDLTest_AssertCheck(result == true, "Verify return value; expected: true, got: %d", result);
Displays are now referenced by instance ID instead of index
2023-01-30 00:30:55 +03:00
/* Check that one gets the current resolution back again */
Removed temporary memory from the API It was intended to make the API easier to use, but various automatic garbage collection all had flaws, and making the application periodically clean up temporary memory added cognitive load to using the API, and in many cases was it was difficult to restructure threaded code to handle this. So, we're largely going back to the original system, where the API returns allocated results and you free them. In addition, to solve the problems we originally wanted temporary memory for: * Short strings with a finite count, like device names, get stored in a per-thread string pool. * Events continue to use temporary memory internally, which is cleaned up on the next event processing cycle.
2024-07-27 04:57:18 +03:00
if (result == 0) {
SDLTest_AssertCheck(closest.w == current.w,
Updated SDL high DPI support We have gotten feedback that abstracting the coordinate system based on the display scale is unexpected and it is difficult to adapt existing applications to the proposed API. The new approach is to provide the coordinate systems that people expect, but provide additional information that will help applications properly handle high DPI situations. The concepts needed for high DPI support are documented in README-highdpi.md. An example of automatically adapting the content to display scale changes can be found in SDL_test_common.c, where auto_scale_content is checked. Also, the SDL_WINDOW_ALLOW_HIGHDPI window flag has been replaced by the SDL_HINT_VIDEO_ENABLE_HIGH_PIXEL_DENSITY hint. Fixes https://github.com/libsdl-org/SDL/issues/7709
2023-05-17 02:29:52 +03:00
"Verify returned width matches current width; expected: %d, got: %d",
Removed temporary memory from the API It was intended to make the API easier to use, but various automatic garbage collection all had flaws, and making the application periodically clean up temporary memory added cognitive load to using the API, and in many cases was it was difficult to restructure threaded code to handle this. So, we're largely going back to the original system, where the API returns allocated results and you free them. In addition, to solve the problems we originally wanted temporary memory for: * Short strings with a finite count, like device names, get stored in a per-thread string pool. * Events continue to use temporary memory internally, which is cleaned up on the next event processing cycle.
2024-07-27 04:57:18 +03:00
current.w, closest.w);
SDLTest_AssertCheck(closest.h == current.h,
Updated SDL high DPI support We have gotten feedback that abstracting the coordinate system based on the display scale is unexpected and it is difficult to adapt existing applications to the proposed API. The new approach is to provide the coordinate systems that people expect, but provide additional information that will help applications properly handle high DPI situations. The concepts needed for high DPI support are documented in README-highdpi.md. An example of automatically adapting the content to display scale changes can be found in SDL_test_common.c, where auto_scale_content is checked. Also, the SDL_WINDOW_ALLOW_HIGHDPI window flag has been replaced by the SDL_HINT_VIDEO_ENABLE_HIGH_PIXEL_DENSITY hint. Fixes https://github.com/libsdl-org/SDL/issues/7709
2023-05-17 02:29:52 +03:00
"Verify returned height matches current height; expected: %d, got: %d",
Removed temporary memory from the API It was intended to make the API easier to use, but various automatic garbage collection all had flaws, and making the application periodically clean up temporary memory added cognitive load to using the API, and in many cases was it was difficult to restructure threaded code to handle this. So, we're largely going back to the original system, where the API returns allocated results and you free them. In addition, to solve the problems we originally wanted temporary memory for: * Short strings with a finite count, like device names, get stored in a per-thread string pool. * Events continue to use temporary memory internally, which is cleaned up on the next event processing cycle.
2024-07-27 04:57:18 +03:00
current.h, closest.h);
Windows default to fullscreen desktop mode if they don't pick an explicit video mode Rather than iterating over display modes using an index, there is a new function SDL_GetFullscreenDisplayModes() to get the list of available fullscreen modes on a display. { SDL_DisplayID display = SDL_GetPrimaryDisplay(); int num_modes = 0; SDL_DisplayMode **modes = SDL_GetFullscreenDisplayModes(display, &num_modes); if (modes) { for (i = 0; i < num_modes; ++i) { SDL_DisplayMode *mode = modes[i]; SDL_Log("Display %" SDL_PRIu32 " mode %d: %dx%d@%gHz, %d%% scale\n", display, i, mode->pixel_w, mode->pixel_h, mode->refresh_rate, (int)(mode->display_scale * 100.0f)); } SDL_free(modes); } } SDL_GetDesktopDisplayMode() and SDL_GetCurrentDisplayMode() return pointers to display modes rather than filling in application memory. Windows now have an explicit fullscreen mode that is set, using SDL_SetWindowFullscreenMode(). The fullscreen mode for a window can be queried with SDL_GetWindowFullscreenMode(), which returns a pointer to the mode, or NULL if the window will be fullscreen desktop. SDL_SetWindowFullscreen() just takes a boolean value, setting the correct fullscreen state based on the selected mode.
2023-02-01 08:23:14 +03:00
}
Displays are now referenced by instance ID instead of index
2023-01-30 00:30:55 +03:00
}
Removed temporary memory from the API It was intended to make the API easier to use, but various automatic garbage collection all had flaws, and making the application periodically clean up temporary memory added cognitive load to using the API, and in many cases was it was difficult to restructure threaded code to handle this. So, we're largely going back to the original system, where the API returns allocated results and you free them. In addition, to solve the problems we originally wanted temporary memory for: * Short strings with a finite count, like device names, get stored in a per-thread string pool. * Events continue to use temporary memory internally, which is cleaned up on the next event processing cycle.
2024-07-27 04:57:18 +03:00
SDL_free(modes);
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
}
Removed temporary memory from the API It was intended to make the API easier to use, but various automatic garbage collection all had flaws, and making the application periodically clean up temporary memory added cognitive load to using the API, and in many cases was it was difficult to restructure threaded code to handle this. So, we're largely going back to the original system, where the API returns allocated results and you free them. In addition, to solve the problems we originally wanted temporary memory for: * Short strings with a finite count, like device names, get stored in a per-thread string pool. * Events continue to use temporary memory internally, which is cleaned up on the next event processing cycle.
2024-07-27 04:57:18 +03:00
SDL_free(displays);
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
}
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
return TEST_COMPLETED;
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
}
/**
docs: Remove Doxygen `\brief` tags. Doxygen and the wiki bridge don't need them; they'll both just use the first line/sentence instead. Fixes #8446.
2023-11-06 18:26:06 +03:00
* Tests the functionality of the SDL_GetClosestFullscreenDisplayMode function against random resolution
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
*/
SDL_test: use SDLCALL calling convention This is needed when using a pre-built static SDL3_test library.
2024-09-06 04:21:13 +03:00
static int SDLCALL video_getClosestDisplayModeRandomResolution(void *arg)
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
{
Removed temporary memory from the API It was intended to make the API easier to use, but various automatic garbage collection all had flaws, and making the application periodically clean up temporary memory added cognitive load to using the API, and in many cases was it was difficult to restructure threaded code to handle this. So, we're largely going back to the original system, where the API returns allocated results and you free them. In addition, to solve the problems we originally wanted temporary memory for: * Short strings with a finite count, like device names, get stored in a per-thread string pool. * Events continue to use temporary memory internally, which is cleaned up on the next event processing cycle.
2024-07-27 04:57:18 +03:00
SDL_DisplayID *displays;
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
SDL_DisplayMode target;
Removed temporary memory from the API It was intended to make the API easier to use, but various automatic garbage collection all had flaws, and making the application periodically clean up temporary memory added cognitive load to using the API, and in many cases was it was difficult to restructure threaded code to handle this. So, we're largely going back to the original system, where the API returns allocated results and you free them. In addition, to solve the problems we originally wanted temporary memory for: * Short strings with a finite count, like device names, get stored in a per-thread string pool. * Events continue to use temporary memory internally, which is cleaned up on the next event processing cycle.
2024-07-27 04:57:18 +03:00
SDL_DisplayMode closest;
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
int i;
int variation;
/* Get number of displays */
Displays are now referenced by instance ID instead of index
2023-01-30 00:30:55 +03:00
displays = SDL_GetDisplays(NULL);
if (displays) {
SDLTest_AssertPass("Call to SDL_GetDisplays()");
/* Make calls for each display */
for (i = 0; displays[i]; ++i) {
remove several instances of unnecessary emptry str "" after SDL_PRI???
2024-01-18 17:25:50 +03:00
SDLTest_Log("Testing against display: %" SDL_PRIu32, displays[i]);
Displays are now referenced by instance ID instead of index
2023-01-30 00:30:55 +03:00
for (variation = 0; variation < 16; variation++) {
/* Set random constraints */
SDL_zero(target);
Updated SDL high DPI support We have gotten feedback that abstracting the coordinate system based on the display scale is unexpected and it is difficult to adapt existing applications to the proposed API. The new approach is to provide the coordinate systems that people expect, but provide additional information that will help applications properly handle high DPI situations. The concepts needed for high DPI support are documented in README-highdpi.md. An example of automatically adapting the content to display scale changes can be found in SDL_test_common.c, where auto_scale_content is checked. Also, the SDL_WINDOW_ALLOW_HIGHDPI window flag has been replaced by the SDL_HINT_VIDEO_ENABLE_HIGH_PIXEL_DENSITY hint. Fixes https://github.com/libsdl-org/SDL/issues/7709
2023-05-17 02:29:52 +03:00
target.w = (variation & 1) ? SDLTest_RandomIntegerInRange(1, 4096) : 0;
target.h = (variation & 2) ? SDLTest_RandomIntegerInRange(1, 4096) : 0;
Displays are now referenced by instance ID instead of index
2023-01-30 00:30:55 +03:00
target.refresh_rate = (variation & 8) ? (float)SDLTest_RandomIntegerInRange(25, 120) : 0.0f;
/* Make call; may or may not find anything, so don't validate any further */
Removed SDL_bool in favor of plain bool We require stdbool.h in the build environment, so we might as well use the plain bool type. If your environment doesn't have stdbool.h, this simple replacement will suffice: typedef signed char bool;
2024-09-18 17:52:28 +03:00
SDL_GetClosestFullscreenDisplayMode(displays[i], target.w, target.h, target.refresh_rate, false, &closest);
Windows default to fullscreen desktop mode if they don't pick an explicit video mode Rather than iterating over display modes using an index, there is a new function SDL_GetFullscreenDisplayModes() to get the list of available fullscreen modes on a display. { SDL_DisplayID display = SDL_GetPrimaryDisplay(); int num_modes = 0; SDL_DisplayMode **modes = SDL_GetFullscreenDisplayModes(display, &num_modes); if (modes) { for (i = 0; i < num_modes; ++i) { SDL_DisplayMode *mode = modes[i]; SDL_Log("Display %" SDL_PRIu32 " mode %d: %dx%d@%gHz, %d%% scale\n", display, i, mode->pixel_w, mode->pixel_h, mode->refresh_rate, (int)(mode->display_scale * 100.0f)); } SDL_free(modes); } } SDL_GetDesktopDisplayMode() and SDL_GetCurrentDisplayMode() return pointers to display modes rather than filling in application memory. Windows now have an explicit fullscreen mode that is set, using SDL_SetWindowFullscreenMode(). The fullscreen mode for a window can be queried with SDL_GetWindowFullscreenMode(), which returns a pointer to the mode, or NULL if the window will be fullscreen desktop. SDL_SetWindowFullscreen() just takes a boolean value, setting the correct fullscreen state based on the selected mode.
2023-02-01 08:23:14 +03:00
SDLTest_AssertPass("Call to SDL_GetClosestFullscreenDisplayMode(target=random/variation%d)", variation);
Displays are now referenced by instance ID instead of index
2023-01-30 00:30:55 +03:00
}
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
}
Removed temporary memory from the API It was intended to make the API easier to use, but various automatic garbage collection all had flaws, and making the application periodically clean up temporary memory added cognitive load to using the API, and in many cases was it was difficult to restructure threaded code to handle this. So, we're largely going back to the original system, where the API returns allocated results and you free them. In addition, to solve the problems we originally wanted temporary memory for: * Short strings with a finite count, like device names, get stored in a per-thread string pool. * Events continue to use temporary memory internally, which is cleaned up on the next event processing cycle.
2024-07-27 04:57:18 +03:00
SDL_free(displays);
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
}
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
return TEST_COMPLETED;
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
}
/**
docs: Remove Doxygen `\brief` tags. Doxygen and the wiki bridge don't need them; they'll both just use the first line/sentence instead. Fixes #8446.
2023-11-06 18:26:06 +03:00
* Tests call to SDL_GetWindowFullscreenMode
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
*
test: don't use wiki urls for documentation comments Also make consistent use of \ as documentation escape character.
2023-02-02 02:21:53 +03:00
* \sa SDL_GetWindowFullscreenMode
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
*/
SDL_test: use SDLCALL calling convention This is needed when using a pre-built static SDL3_test library.
2024-09-06 04:21:13 +03:00
static int SDLCALL video_getWindowDisplayMode(void *arg)
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
{
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
SDL_Window *window;
const char *title = "video_getWindowDisplayMode Test Window";
Windows default to fullscreen desktop mode if they don't pick an explicit video mode Rather than iterating over display modes using an index, there is a new function SDL_GetFullscreenDisplayModes() to get the list of available fullscreen modes on a display. { SDL_DisplayID display = SDL_GetPrimaryDisplay(); int num_modes = 0; SDL_DisplayMode **modes = SDL_GetFullscreenDisplayModes(display, &num_modes); if (modes) { for (i = 0; i < num_modes; ++i) { SDL_DisplayMode *mode = modes[i]; SDL_Log("Display %" SDL_PRIu32 " mode %d: %dx%d@%gHz, %d%% scale\n", display, i, mode->pixel_w, mode->pixel_h, mode->refresh_rate, (int)(mode->display_scale * 100.0f)); } SDL_free(modes); } } SDL_GetDesktopDisplayMode() and SDL_GetCurrentDisplayMode() return pointers to display modes rather than filling in application memory. Windows now have an explicit fullscreen mode that is set, using SDL_SetWindowFullscreenMode(). The fullscreen mode for a window can be queried with SDL_GetWindowFullscreenMode(), which returns a pointer to the mode, or NULL if the window will be fullscreen desktop. SDL_SetWindowFullscreen() just takes a boolean value, setting the correct fullscreen state based on the selected mode.
2023-02-01 08:23:14 +03:00
const SDL_DisplayMode *mode;
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
/* Call against new test window */
Remove more reserved identifiers (#6925)
2022-12-30 00:58:16 +03:00
window = createVideoSuiteTestWindow(title);
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
if (window != NULL) {
Windows default to fullscreen desktop mode if they don't pick an explicit video mode Rather than iterating over display modes using an index, there is a new function SDL_GetFullscreenDisplayModes() to get the list of available fullscreen modes on a display. { SDL_DisplayID display = SDL_GetPrimaryDisplay(); int num_modes = 0; SDL_DisplayMode **modes = SDL_GetFullscreenDisplayModes(display, &num_modes); if (modes) { for (i = 0; i < num_modes; ++i) { SDL_DisplayMode *mode = modes[i]; SDL_Log("Display %" SDL_PRIu32 " mode %d: %dx%d@%gHz, %d%% scale\n", display, i, mode->pixel_w, mode->pixel_h, mode->refresh_rate, (int)(mode->display_scale * 100.0f)); } SDL_free(modes); } } SDL_GetDesktopDisplayMode() and SDL_GetCurrentDisplayMode() return pointers to display modes rather than filling in application memory. Windows now have an explicit fullscreen mode that is set, using SDL_SetWindowFullscreenMode(). The fullscreen mode for a window can be queried with SDL_GetWindowFullscreenMode(), which returns a pointer to the mode, or NULL if the window will be fullscreen desktop. SDL_SetWindowFullscreen() just takes a boolean value, setting the correct fullscreen state based on the selected mode.
2023-02-01 08:23:14 +03:00
mode = SDL_GetWindowFullscreenMode(window);
SDLTest_AssertPass("Call to SDL_GetWindowFullscreenMode()");
SDLTest_AssertCheck(mode == NULL, "Validate result value; expected: NULL, got: %p", mode);
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
}
/* Clean up */
Remove more reserved identifiers (#6925)
2022-12-30 00:58:16 +03:00
destroyVideoSuiteTestWindow(window);
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
return TEST_COMPLETED;
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
}
/* Helper function that checks for an 'Invalid window' error */
More fix warnings about static function and prototype
2023-03-08 18:12:45 +03:00
static void checkInvalidWindowError(void)
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
{
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
const char *invalidWindowError = "Invalid window";
Fix warning drop const qualifier
2023-03-08 23:18:02 +03:00
const char *lastError;
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
Fix warning drop const qualifier
2023-03-08 23:18:02 +03:00
lastError = SDL_GetError();
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
SDLTest_AssertPass("SDL_GetError()");
SDLTest_AssertCheck(lastError != NULL, "Verify error message is not NULL");
if (lastError != NULL) {
SDLTest_AssertCheck(SDL_strcmp(lastError, invalidWindowError) == 0,
"SDL_GetError(): expected message '%s', was message: '%s'",
invalidWindowError,
lastError);
SDL_ClearError();
SDLTest_AssertPass("Call to SDL_ClearError()");
}
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
}
/**
docs: Remove Doxygen `\brief` tags. Doxygen and the wiki bridge don't need them; they'll both just use the first line/sentence instead. Fixes #8446.
2023-11-06 18:26:06 +03:00
* Tests call to SDL_GetWindowFullscreenMode with invalid input
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
*
test: don't use wiki urls for documentation comments Also make consistent use of \ as documentation escape character.
2023-02-02 02:21:53 +03:00
* \sa SDL_GetWindowFullscreenMode
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
*/
SDL_test: use SDLCALL calling convention This is needed when using a pre-built static SDL3_test library.
2024-09-06 04:21:13 +03:00
static int SDLCALL video_getWindowDisplayModeNegative(void *arg)
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
{
Windows default to fullscreen desktop mode if they don't pick an explicit video mode Rather than iterating over display modes using an index, there is a new function SDL_GetFullscreenDisplayModes() to get the list of available fullscreen modes on a display. { SDL_DisplayID display = SDL_GetPrimaryDisplay(); int num_modes = 0; SDL_DisplayMode **modes = SDL_GetFullscreenDisplayModes(display, &num_modes); if (modes) { for (i = 0; i < num_modes; ++i) { SDL_DisplayMode *mode = modes[i]; SDL_Log("Display %" SDL_PRIu32 " mode %d: %dx%d@%gHz, %d%% scale\n", display, i, mode->pixel_w, mode->pixel_h, mode->refresh_rate, (int)(mode->display_scale * 100.0f)); } SDL_free(modes); } } SDL_GetDesktopDisplayMode() and SDL_GetCurrentDisplayMode() return pointers to display modes rather than filling in application memory. Windows now have an explicit fullscreen mode that is set, using SDL_SetWindowFullscreenMode(). The fullscreen mode for a window can be queried with SDL_GetWindowFullscreenMode(), which returns a pointer to the mode, or NULL if the window will be fullscreen desktop. SDL_SetWindowFullscreen() just takes a boolean value, setting the correct fullscreen state based on the selected mode.
2023-02-01 08:23:14 +03:00
const SDL_DisplayMode *mode;
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
/* Call against invalid window */
Windows default to fullscreen desktop mode if they don't pick an explicit video mode Rather than iterating over display modes using an index, there is a new function SDL_GetFullscreenDisplayModes() to get the list of available fullscreen modes on a display. { SDL_DisplayID display = SDL_GetPrimaryDisplay(); int num_modes = 0; SDL_DisplayMode **modes = SDL_GetFullscreenDisplayModes(display, &num_modes); if (modes) { for (i = 0; i < num_modes; ++i) { SDL_DisplayMode *mode = modes[i]; SDL_Log("Display %" SDL_PRIu32 " mode %d: %dx%d@%gHz, %d%% scale\n", display, i, mode->pixel_w, mode->pixel_h, mode->refresh_rate, (int)(mode->display_scale * 100.0f)); } SDL_free(modes); } } SDL_GetDesktopDisplayMode() and SDL_GetCurrentDisplayMode() return pointers to display modes rather than filling in application memory. Windows now have an explicit fullscreen mode that is set, using SDL_SetWindowFullscreenMode(). The fullscreen mode for a window can be queried with SDL_GetWindowFullscreenMode(), which returns a pointer to the mode, or NULL if the window will be fullscreen desktop. SDL_SetWindowFullscreen() just takes a boolean value, setting the correct fullscreen state based on the selected mode.
2023-02-01 08:23:14 +03:00
mode = SDL_GetWindowFullscreenMode(NULL);
SDLTest_AssertPass("Call to SDL_GetWindowFullscreenMode(window=NULL)");
SDLTest_AssertCheck(mode == NULL, "Validate result value; expected: NULL, got: %p", mode);
Remove more reserved identifiers (#6925)
2022-12-30 00:58:16 +03:00
checkInvalidWindowError();
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
return TEST_COMPLETED;
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
}
Expose separate keyboard and mouse grab support This adds SDL_SetWindowKeyboardGrab(), SDL_GetWindowKeyboardGrab(), SDL_SetWindowMouseGrab(), SDL_GetWindowMouseGrab(), and new SDL_WINDOW_KEYBOARD_GRABBED flag. It also updates the test harness to exercise this functionality and makes a minor fix to X11 that I missed in https://hg.libsdl.org/SDL/rev/02a2d609369b To fit in with this new support, SDL_WINDOW_INPUT_CAPTURE has been renamed to SDL_WINDOW_MOUSE_CAPTURE with the old name remaining as an alias for backwards compatibility with older code.
2021-01-27 04:16:17 +03:00
/* Helper for setting and checking the window mouse grab state */
Removed SDL_bool in favor of plain bool We require stdbool.h in the build environment, so we might as well use the plain bool type. If your environment doesn't have stdbool.h, this simple replacement will suffice: typedef signed char bool;
2024-09-18 17:52:28 +03:00
static void setAndCheckWindowMouseGrabState(SDL_Window *window, bool desiredState)
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
{
Removed SDL_bool in favor of plain bool We require stdbool.h in the build environment, so we might as well use the plain bool type. If your environment doesn't have stdbool.h, this simple replacement will suffice: typedef signed char bool;
2024-09-18 17:52:28 +03:00
bool currentState;
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
/* Set state */
SDL_SetWindowMouseGrab(window, desiredState);
Removed SDL_bool in favor of plain bool We require stdbool.h in the build environment, so we might as well use the plain bool type. If your environment doesn't have stdbool.h, this simple replacement will suffice: typedef signed char bool;
2024-09-18 17:52:28 +03:00
SDLTest_AssertPass("Call to SDL_SetWindowMouseGrab(%s)", (desiredState == false) ? "false" : "true");
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
/* Get and check state */
currentState = SDL_GetWindowMouseGrab(window);
SDLTest_AssertPass("Call to SDL_GetWindowMouseGrab()");
Expose separate keyboard and mouse grab support This adds SDL_SetWindowKeyboardGrab(), SDL_GetWindowKeyboardGrab(), SDL_SetWindowMouseGrab(), SDL_GetWindowMouseGrab(), and new SDL_WINDOW_KEYBOARD_GRABBED flag. It also updates the test harness to exercise this functionality and makes a minor fix to X11 that I missed in https://hg.libsdl.org/SDL/rev/02a2d609369b To fit in with this new support, SDL_WINDOW_INPUT_CAPTURE has been renamed to SDL_WINDOW_MOUSE_CAPTURE with the old name remaining as an alias for backwards compatibility with older code.
2021-01-27 04:16:17 +03:00
SDLTest_AssertCheck(
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
currentState == desiredState,
"Validate returned state; expected: %s, got: %s",
Removed SDL_bool in favor of plain bool We require stdbool.h in the build environment, so we might as well use the plain bool type. If your environment doesn't have stdbool.h, this simple replacement will suffice: typedef signed char bool;
2024-09-18 17:52:28 +03:00
(desiredState == false) ? "false" : "true",
(currentState == false) ? "false" : "true");
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
if (desiredState) {
SDLTest_AssertCheck(
SDL_GetGrabbedWindow() == window,
"Grabbed window should be to our window");
SDLTest_AssertCheck(
SDL_GetWindowFlags(window) & SDL_WINDOW_MOUSE_GRABBED,
"SDL_WINDOW_MOUSE_GRABBED should be set");
} else {
SDLTest_AssertCheck(
!(SDL_GetWindowFlags(window) & SDL_WINDOW_MOUSE_GRABBED),
"SDL_WINDOW_MOUSE_GRABBED should be unset");
}
Expose separate keyboard and mouse grab support This adds SDL_SetWindowKeyboardGrab(), SDL_GetWindowKeyboardGrab(), SDL_SetWindowMouseGrab(), SDL_GetWindowMouseGrab(), and new SDL_WINDOW_KEYBOARD_GRABBED flag. It also updates the test harness to exercise this functionality and makes a minor fix to X11 that I missed in https://hg.libsdl.org/SDL/rev/02a2d609369b To fit in with this new support, SDL_WINDOW_INPUT_CAPTURE has been renamed to SDL_WINDOW_MOUSE_CAPTURE with the old name remaining as an alias for backwards compatibility with older code.
2021-01-27 04:16:17 +03:00
}
/* Helper for setting and checking the window keyboard grab state */
Removed SDL_bool in favor of plain bool We require stdbool.h in the build environment, so we might as well use the plain bool type. If your environment doesn't have stdbool.h, this simple replacement will suffice: typedef signed char bool;
2024-09-18 17:52:28 +03:00
static void setAndCheckWindowKeyboardGrabState(SDL_Window *window, bool desiredState)
Expose separate keyboard and mouse grab support This adds SDL_SetWindowKeyboardGrab(), SDL_GetWindowKeyboardGrab(), SDL_SetWindowMouseGrab(), SDL_GetWindowMouseGrab(), and new SDL_WINDOW_KEYBOARD_GRABBED flag. It also updates the test harness to exercise this functionality and makes a minor fix to X11 that I missed in https://hg.libsdl.org/SDL/rev/02a2d609369b To fit in with this new support, SDL_WINDOW_INPUT_CAPTURE has been renamed to SDL_WINDOW_MOUSE_CAPTURE with the old name remaining as an alias for backwards compatibility with older code.
2021-01-27 04:16:17 +03:00
{
Removed SDL_bool in favor of plain bool We require stdbool.h in the build environment, so we might as well use the plain bool type. If your environment doesn't have stdbool.h, this simple replacement will suffice: typedef signed char bool;
2024-09-18 17:52:28 +03:00
bool currentState;
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
/* Set state */
SDL_SetWindowKeyboardGrab(window, desiredState);
Removed SDL_bool in favor of plain bool We require stdbool.h in the build environment, so we might as well use the plain bool type. If your environment doesn't have stdbool.h, this simple replacement will suffice: typedef signed char bool;
2024-09-18 17:52:28 +03:00
SDLTest_AssertPass("Call to SDL_SetWindowKeyboardGrab(%s)", (desiredState == false) ? "false" : "true");
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
/* Get and check state */
currentState = SDL_GetWindowKeyboardGrab(window);
SDLTest_AssertPass("Call to SDL_GetWindowKeyboardGrab()");
Expose separate keyboard and mouse grab support This adds SDL_SetWindowKeyboardGrab(), SDL_GetWindowKeyboardGrab(), SDL_SetWindowMouseGrab(), SDL_GetWindowMouseGrab(), and new SDL_WINDOW_KEYBOARD_GRABBED flag. It also updates the test harness to exercise this functionality and makes a minor fix to X11 that I missed in https://hg.libsdl.org/SDL/rev/02a2d609369b To fit in with this new support, SDL_WINDOW_INPUT_CAPTURE has been renamed to SDL_WINDOW_MOUSE_CAPTURE with the old name remaining as an alias for backwards compatibility with older code.
2021-01-27 04:16:17 +03:00
SDLTest_AssertCheck(
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
currentState == desiredState,
"Validate returned state; expected: %s, got: %s",
Removed SDL_bool in favor of plain bool We require stdbool.h in the build environment, so we might as well use the plain bool type. If your environment doesn't have stdbool.h, this simple replacement will suffice: typedef signed char bool;
2024-09-18 17:52:28 +03:00
(desiredState == false) ? "false" : "true",
(currentState == false) ? "false" : "true");
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
if (desiredState) {
SDLTest_AssertCheck(
SDL_GetGrabbedWindow() == window,
"Grabbed window should be set to our window");
SDLTest_AssertCheck(
SDL_GetWindowFlags(window) & SDL_WINDOW_KEYBOARD_GRABBED,
"SDL_WINDOW_KEYBOARD_GRABBED should be set");
} else {
SDLTest_AssertCheck(
!(SDL_GetWindowFlags(window) & SDL_WINDOW_KEYBOARD_GRABBED),
"SDL_WINDOW_KEYBOARD_GRABBED should be unset");
}
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
}
/**
docs: Remove Doxygen `\brief` tags. Doxygen and the wiki bridge don't need them; they'll both just use the first line/sentence instead. Fixes #8446.
2023-11-06 18:26:06 +03:00
* Tests keyboard and mouse grab support
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
*
Remove legacy SDL2 input grab API
2024-03-25 02:35:40 +03:00
* \sa SDL_GetWindowMouseGrab
* \sa SDL_GetWindowKeyboardGrab
* \sa SDL_SetWindowMouseGrab
* \sa SDL_SetWindowKeyboardGrab
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
*/
SDL_test: use SDLCALL calling convention This is needed when using a pre-built static SDL3_test library.
2024-09-06 04:21:13 +03:00
static int SDLCALL video_getSetWindowGrab(void *arg)
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
{
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
const char *title = "video_getSetWindowGrab Test Window";
SDL_Window *window;
Removed SDL_bool in favor of plain bool We require stdbool.h in the build environment, so we might as well use the plain bool type. If your environment doesn't have stdbool.h, this simple replacement will suffice: typedef signed char bool;
2024-09-18 17:52:28 +03:00
bool originalMouseState, originalKeyboardState;
bool hasFocusGained = false;
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
/* Call against new test window */
Remove more reserved identifiers (#6925)
2022-12-30 00:58:16 +03:00
window = createVideoSuiteTestWindow(title);
Pointer as bool (libsdl-org#7214)
2023-11-10 00:29:15 +03:00
if (!window) {
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
return TEST_ABORTED;
}
Fix video_getSetWindowGrab(): need to raise the window, and wait for focus gained
2023-03-20 14:34:36 +03:00
/* Need to raise the window to have and SDL_EVENT_WINDOW_FOCUS_GAINED,
* so that the window gets the flags SDL_WINDOW_INPUT_FOCUS,
* so that it can be "grabbed" */
SDL_RaiseWindow(window);
video: Implement asynchronous windowing SDL window size, state, and position functions have been considered immediate, with their effects assuming to have taken effect upon successful return of the function. However, several windowing systems handle these requests asynchronously, resulting in the functions blocking until the changes have taken effect, potentially for long periods of time. Additionally, some windowing systems treat these as requests, and can potentially deny or fulfill the request in a manner differently than the application expects, such as not allowing a window to be positioned or sized beyond desktop borders, prohibiting fullscreen, and so on. With these changes, applications can make requests of the window manager that do not block, with the understanding that an associated event will be sent if the request is fulfilled. Currently, size, position, maximize, minimize, and fullscreen calls are handled as asynchronous requests, with events being returned if the request is honored. If the application requires that the change take effect immediately, it can call the new SDL_SyncWindow function, which will attempt to block until the request is fulfilled, or some arbitrary timeout period elapses, the duration of which depends not only on the windowing system, but on the operation requested as well (e.g. a 100ms timeout is fine for most X11 events, but maximizing a window can take considerably longer for some reason). There is also a new hint 'SDL_VIDEO_SYNC_ALL_WINDOW_OPS' that will mimic the old behavior by synchronizing after every window operation with, again, the understanding that using this may result in the associated calls blocking for a relatively long period. The deferred model also results in the window size and position getters not reporting false coordinates anymore, as they only forward what the window manager reports vs allowing applications to set arbitrary values, and fullscreen enter/leave events that were initiated via the window manager update the window state appropriately, where they didn't before. Care was taken to ensure that order of operations is maintained, and that requests are not ignored or dropped. This does require some implicit internal synchronization in the various backends if many requests are made in a short period, as some state and behavior depends on other bits of state that need to be known at that particular point in time, but this isn't something that typical applications will hit, unless they are sending a lot of window state in a short time as the tests do. The automated tests developed to test the previous behavior also resulted in previously undefined behavior being defined and normalized across platforms, particularly when it comes to the sizing and positioning of windows when they are in a fixed-size state, such as maximized or fullscreen. Size and position requests made when the window is not in a movable or resizable state will be deferred until it can be applied, so no requests are lost. These changes fix another long-standing issue with renderers recreating maximized windows, where the original non-maximized size was lost, resulting in the window being restored to the wrong size. All automated video tests pass across all platforms. Overall, the "make a request/get an event" model better reflects how most windowing systems work, and some backends avoid spending significant time blocking while waiting for operations to complete.
2023-10-25 22:09:55 +03:00
if (!(SDL_GetWindowFlags(window) & SDL_WINDOW_INPUT_FOCUS)) {
int count = 0;
Fix video_getSetWindowGrab(): need to raise the window, and wait for focus gained
2023-03-20 14:34:36 +03:00
SDL_Event evt;
SDL_zero(evt);
video: Implement asynchronous windowing SDL window size, state, and position functions have been considered immediate, with their effects assuming to have taken effect upon successful return of the function. However, several windowing systems handle these requests asynchronously, resulting in the functions blocking until the changes have taken effect, potentially for long periods of time. Additionally, some windowing systems treat these as requests, and can potentially deny or fulfill the request in a manner differently than the application expects, such as not allowing a window to be positioned or sized beyond desktop borders, prohibiting fullscreen, and so on. With these changes, applications can make requests of the window manager that do not block, with the understanding that an associated event will be sent if the request is fulfilled. Currently, size, position, maximize, minimize, and fullscreen calls are handled as asynchronous requests, with events being returned if the request is honored. If the application requires that the change take effect immediately, it can call the new SDL_SyncWindow function, which will attempt to block until the request is fulfilled, or some arbitrary timeout period elapses, the duration of which depends not only on the windowing system, but on the operation requested as well (e.g. a 100ms timeout is fine for most X11 events, but maximizing a window can take considerably longer for some reason). There is also a new hint 'SDL_VIDEO_SYNC_ALL_WINDOW_OPS' that will mimic the old behavior by synchronizing after every window operation with, again, the understanding that using this may result in the associated calls blocking for a relatively long period. The deferred model also results in the window size and position getters not reporting false coordinates anymore, as they only forward what the window manager reports vs allowing applications to set arbitrary values, and fullscreen enter/leave events that were initiated via the window manager update the window state appropriately, where they didn't before. Care was taken to ensure that order of operations is maintained, and that requests are not ignored or dropped. This does require some implicit internal synchronization in the various backends if many requests are made in a short period, as some state and behavior depends on other bits of state that need to be known at that particular point in time, but this isn't something that typical applications will hit, unless they are sending a lot of window state in a short time as the tests do. The automated tests developed to test the previous behavior also resulted in previously undefined behavior being defined and normalized across platforms, particularly when it comes to the sizing and positioning of windows when they are in a fixed-size state, such as maximized or fullscreen. Size and position requests made when the window is not in a movable or resizable state will be deferred until it can be applied, so no requests are lost. These changes fix another long-standing issue with renderers recreating maximized windows, where the original non-maximized size was lost, resulting in the window being restored to the wrong size. All automated video tests pass across all platforms. Overall, the "make a request/get an event" model better reflects how most windowing systems work, and some backends avoid spending significant time blocking while waiting for operations to complete.
2023-10-25 22:09:55 +03:00
while (!hasFocusGained && count++ < 3) {
while (SDL_PollEvent(&evt)) {
if (evt.type == SDL_EVENT_WINDOW_FOCUS_GAINED) {
Removed SDL_bool in favor of plain bool We require stdbool.h in the build environment, so we might as well use the plain bool type. If your environment doesn't have stdbool.h, this simple replacement will suffice: typedef signed char bool;
2024-09-18 17:52:28 +03:00
hasFocusGained = true;
video: Implement asynchronous windowing SDL window size, state, and position functions have been considered immediate, with their effects assuming to have taken effect upon successful return of the function. However, several windowing systems handle these requests asynchronously, resulting in the functions blocking until the changes have taken effect, potentially for long periods of time. Additionally, some windowing systems treat these as requests, and can potentially deny or fulfill the request in a manner differently than the application expects, such as not allowing a window to be positioned or sized beyond desktop borders, prohibiting fullscreen, and so on. With these changes, applications can make requests of the window manager that do not block, with the understanding that an associated event will be sent if the request is fulfilled. Currently, size, position, maximize, minimize, and fullscreen calls are handled as asynchronous requests, with events being returned if the request is honored. If the application requires that the change take effect immediately, it can call the new SDL_SyncWindow function, which will attempt to block until the request is fulfilled, or some arbitrary timeout period elapses, the duration of which depends not only on the windowing system, but on the operation requested as well (e.g. a 100ms timeout is fine for most X11 events, but maximizing a window can take considerably longer for some reason). There is also a new hint 'SDL_VIDEO_SYNC_ALL_WINDOW_OPS' that will mimic the old behavior by synchronizing after every window operation with, again, the understanding that using this may result in the associated calls blocking for a relatively long period. The deferred model also results in the window size and position getters not reporting false coordinates anymore, as they only forward what the window manager reports vs allowing applications to set arbitrary values, and fullscreen enter/leave events that were initiated via the window manager update the window state appropriately, where they didn't before. Care was taken to ensure that order of operations is maintained, and that requests are not ignored or dropped. This does require some implicit internal synchronization in the various backends if many requests are made in a short period, as some state and behavior depends on other bits of state that need to be known at that particular point in time, but this isn't something that typical applications will hit, unless they are sending a lot of window state in a short time as the tests do. The automated tests developed to test the previous behavior also resulted in previously undefined behavior being defined and normalized across platforms, particularly when it comes to the sizing and positioning of windows when they are in a fixed-size state, such as maximized or fullscreen. Size and position requests made when the window is not in a movable or resizable state will be deferred until it can be applied, so no requests are lost. These changes fix another long-standing issue with renderers recreating maximized windows, where the original non-maximized size was lost, resulting in the window being restored to the wrong size. All automated video tests pass across all platforms. Overall, the "make a request/get an event" model better reflects how most windowing systems work, and some backends avoid spending significant time blocking while waiting for operations to complete.
2023-10-25 22:09:55 +03:00
}
Fix video_getSetWindowGrab(): need to raise the window, and wait for focus gained
2023-03-20 14:34:36 +03:00
}
}
video: Implement asynchronous windowing SDL window size, state, and position functions have been considered immediate, with their effects assuming to have taken effect upon successful return of the function. However, several windowing systems handle these requests asynchronously, resulting in the functions blocking until the changes have taken effect, potentially for long periods of time. Additionally, some windowing systems treat these as requests, and can potentially deny or fulfill the request in a manner differently than the application expects, such as not allowing a window to be positioned or sized beyond desktop borders, prohibiting fullscreen, and so on. With these changes, applications can make requests of the window manager that do not block, with the understanding that an associated event will be sent if the request is fulfilled. Currently, size, position, maximize, minimize, and fullscreen calls are handled as asynchronous requests, with events being returned if the request is honored. If the application requires that the change take effect immediately, it can call the new SDL_SyncWindow function, which will attempt to block until the request is fulfilled, or some arbitrary timeout period elapses, the duration of which depends not only on the windowing system, but on the operation requested as well (e.g. a 100ms timeout is fine for most X11 events, but maximizing a window can take considerably longer for some reason). There is also a new hint 'SDL_VIDEO_SYNC_ALL_WINDOW_OPS' that will mimic the old behavior by synchronizing after every window operation with, again, the understanding that using this may result in the associated calls blocking for a relatively long period. The deferred model also results in the window size and position getters not reporting false coordinates anymore, as they only forward what the window manager reports vs allowing applications to set arbitrary values, and fullscreen enter/leave events that were initiated via the window manager update the window state appropriately, where they didn't before. Care was taken to ensure that order of operations is maintained, and that requests are not ignored or dropped. This does require some implicit internal synchronization in the various backends if many requests are made in a short period, as some state and behavior depends on other bits of state that need to be known at that particular point in time, but this isn't something that typical applications will hit, unless they are sending a lot of window state in a short time as the tests do. The automated tests developed to test the previous behavior also resulted in previously undefined behavior being defined and normalized across platforms, particularly when it comes to the sizing and positioning of windows when they are in a fixed-size state, such as maximized or fullscreen. Size and position requests made when the window is not in a movable or resizable state will be deferred until it can be applied, so no requests are lost. These changes fix another long-standing issue with renderers recreating maximized windows, where the original non-maximized size was lost, resulting in the window being restored to the wrong size. All automated video tests pass across all platforms. Overall, the "make a request/get an event" model better reflects how most windowing systems work, and some backends avoid spending significant time blocking while waiting for operations to complete.
2023-10-25 22:09:55 +03:00
} else {
Removed SDL_bool in favor of plain bool We require stdbool.h in the build environment, so we might as well use the plain bool type. If your environment doesn't have stdbool.h, this simple replacement will suffice: typedef signed char bool;
2024-09-18 17:52:28 +03:00
hasFocusGained = true;
Fix video_getSetWindowGrab(): need to raise the window, and wait for focus gained
2023-03-20 14:34:36 +03:00
}
Removed SDL_bool in favor of plain bool We require stdbool.h in the build environment, so we might as well use the plain bool type. If your environment doesn't have stdbool.h, this simple replacement will suffice: typedef signed char bool;
2024-09-18 17:52:28 +03:00
SDLTest_AssertCheck(hasFocusGained == true, "Expectded window with focus");
Fix video_getSetWindowGrab(): need to raise the window, and wait for focus gained
2023-03-20 14:34:36 +03:00
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
/* Get state */
originalMouseState = SDL_GetWindowMouseGrab(window);
SDLTest_AssertPass("Call to SDL_GetWindowMouseGrab()");
originalKeyboardState = SDL_GetWindowKeyboardGrab(window);
SDLTest_AssertPass("Call to SDL_GetWindowKeyboardGrab()");
/* F */
Removed SDL_bool in favor of plain bool We require stdbool.h in the build environment, so we might as well use the plain bool type. If your environment doesn't have stdbool.h, this simple replacement will suffice: typedef signed char bool;
2024-09-18 17:52:28 +03:00
setAndCheckWindowKeyboardGrabState(window, false);
setAndCheckWindowMouseGrabState(window, false);
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
SDLTest_AssertCheck(SDL_GetGrabbedWindow() == NULL,
"Expected NULL grabbed window");
/* F --> F */
Removed SDL_bool in favor of plain bool We require stdbool.h in the build environment, so we might as well use the plain bool type. If your environment doesn't have stdbool.h, this simple replacement will suffice: typedef signed char bool;
2024-09-18 17:52:28 +03:00
setAndCheckWindowMouseGrabState(window, false);
setAndCheckWindowKeyboardGrabState(window, false);
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
SDLTest_AssertCheck(SDL_GetGrabbedWindow() == NULL,
"Expected NULL grabbed window");
/* F --> T */
Removed SDL_bool in favor of plain bool We require stdbool.h in the build environment, so we might as well use the plain bool type. If your environment doesn't have stdbool.h, this simple replacement will suffice: typedef signed char bool;
2024-09-18 17:52:28 +03:00
setAndCheckWindowMouseGrabState(window, true);
setAndCheckWindowKeyboardGrabState(window, true);
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
/* T --> T */
Removed SDL_bool in favor of plain bool We require stdbool.h in the build environment, so we might as well use the plain bool type. If your environment doesn't have stdbool.h, this simple replacement will suffice: typedef signed char bool;
2024-09-18 17:52:28 +03:00
setAndCheckWindowKeyboardGrabState(window, true);
setAndCheckWindowMouseGrabState(window, true);
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
/* M: T --> F */
/* K: T --> T */
Removed SDL_bool in favor of plain bool We require stdbool.h in the build environment, so we might as well use the plain bool type. If your environment doesn't have stdbool.h, this simple replacement will suffice: typedef signed char bool;
2024-09-18 17:52:28 +03:00
setAndCheckWindowKeyboardGrabState(window, true);
setAndCheckWindowMouseGrabState(window, false);
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
/* M: F --> T */
/* K: T --> F */
Removed SDL_bool in favor of plain bool We require stdbool.h in the build environment, so we might as well use the plain bool type. If your environment doesn't have stdbool.h, this simple replacement will suffice: typedef signed char bool;
2024-09-18 17:52:28 +03:00
setAndCheckWindowMouseGrabState(window, true);
setAndCheckWindowKeyboardGrabState(window, false);
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
/* M: T --> F */
/* K: F --> F */
Removed SDL_bool in favor of plain bool We require stdbool.h in the build environment, so we might as well use the plain bool type. If your environment doesn't have stdbool.h, this simple replacement will suffice: typedef signed char bool;
2024-09-18 17:52:28 +03:00
setAndCheckWindowMouseGrabState(window, false);
setAndCheckWindowKeyboardGrabState(window, false);
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
SDLTest_AssertCheck(SDL_GetGrabbedWindow() == NULL,
"Expected NULL grabbed window");
/* Negative tests */
Remove legacy SDL2 input grab API
2024-03-25 02:35:40 +03:00
SDL_GetWindowMouseGrab(NULL);
SDLTest_AssertPass("Call to SDL_GetWindowMouseGrab(window=NULL)");
Remove more reserved identifiers (#6925)
2022-12-30 00:58:16 +03:00
checkInvalidWindowError();
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
SDL_GetWindowKeyboardGrab(NULL);
SDLTest_AssertPass("Call to SDL_GetWindowKeyboardGrab(window=NULL)");
Remove more reserved identifiers (#6925)
2022-12-30 00:58:16 +03:00
checkInvalidWindowError();
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
Removed SDL_bool in favor of plain bool We require stdbool.h in the build environment, so we might as well use the plain bool type. If your environment doesn't have stdbool.h, this simple replacement will suffice: typedef signed char bool;
2024-09-18 17:52:28 +03:00
SDL_SetWindowMouseGrab(NULL, false);
SDLTest_AssertPass("Call to SDL_SetWindowMouseGrab(window=NULL,false)");
Remove more reserved identifiers (#6925)
2022-12-30 00:58:16 +03:00
checkInvalidWindowError();
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
Removed SDL_bool in favor of plain bool We require stdbool.h in the build environment, so we might as well use the plain bool type. If your environment doesn't have stdbool.h, this simple replacement will suffice: typedef signed char bool;
2024-09-18 17:52:28 +03:00
SDL_SetWindowKeyboardGrab(NULL, false);
SDLTest_AssertPass("Call to SDL_SetWindowKeyboardGrab(window=NULL,false)");
Remove more reserved identifiers (#6925)
2022-12-30 00:58:16 +03:00
checkInvalidWindowError();
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
Removed SDL_bool in favor of plain bool We require stdbool.h in the build environment, so we might as well use the plain bool type. If your environment doesn't have stdbool.h, this simple replacement will suffice: typedef signed char bool;
2024-09-18 17:52:28 +03:00
SDL_SetWindowMouseGrab(NULL, true);
SDLTest_AssertPass("Call to SDL_SetWindowMouseGrab(window=NULL,true)");
Remove more reserved identifiers (#6925)
2022-12-30 00:58:16 +03:00
checkInvalidWindowError();
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
Removed SDL_bool in favor of plain bool We require stdbool.h in the build environment, so we might as well use the plain bool type. If your environment doesn't have stdbool.h, this simple replacement will suffice: typedef signed char bool;
2024-09-18 17:52:28 +03:00
SDL_SetWindowKeyboardGrab(NULL, true);
SDLTest_AssertPass("Call to SDL_SetWindowKeyboardGrab(window=NULL,true)");
Remove more reserved identifiers (#6925)
2022-12-30 00:58:16 +03:00
checkInvalidWindowError();
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
/* Restore state */
Remove more reserved identifiers (#6925)
2022-12-30 00:58:16 +03:00
setAndCheckWindowMouseGrabState(window, originalMouseState);
setAndCheckWindowKeyboardGrabState(window, originalKeyboardState);
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
/* Clean up */
Remove more reserved identifiers (#6925)
2022-12-30 00:58:16 +03:00
destroyVideoSuiteTestWindow(window);
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
return TEST_COMPLETED;
}
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
/**
docs: Remove Doxygen `\brief` tags. Doxygen and the wiki bridge don't need them; they'll both just use the first line/sentence instead. Fixes #8446.
2023-11-06 18:26:06 +03:00
* Tests call to SDL_GetWindowID and SDL_GetWindowFromID
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
*
test: don't use wiki urls for documentation comments Also make consistent use of \ as documentation escape character.
2023-02-02 02:21:53 +03:00
* \sa SDL_GetWindowID
* \sa SDL_SetWindowFromID
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
*/
SDL_test: use SDLCALL calling convention This is needed when using a pre-built static SDL3_test library.
2024-09-06 04:21:13 +03:00
static int SDLCALL video_getWindowId(void *arg)
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
{
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
const char *title = "video_getWindowId Test Window";
SDL_Window *window;
SDL_Window *result;
Uint32 id, randomId;
/* Call against new test window */
Remove more reserved identifiers (#6925)
2022-12-30 00:58:16 +03:00
window = createVideoSuiteTestWindow(title);
Pointer as bool (libsdl-org#7214)
2023-11-10 00:29:15 +03:00
if (!window) {
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
return TEST_ABORTED;
}
/* Get ID */
id = SDL_GetWindowID(window);
SDLTest_AssertPass("Call to SDL_GetWindowID()");
/* Get window from ID */
result = SDL_GetWindowFromID(id);
SDLTest_AssertPass("Call to SDL_GetWindowID(%" SDL_PRIu32 ")", id);
SDLTest_AssertCheck(result == window, "Verify result matches window pointer");
/* Get window from random large ID, no result check */
randomId = SDLTest_RandomIntegerInRange(UINT8_MAX, UINT16_MAX);
result = SDL_GetWindowFromID(randomId);
SDLTest_AssertPass("Call to SDL_GetWindowID(%" SDL_PRIu32 "/random_large)", randomId);
/* Get window from 0 and Uint32 max ID, no result check */
result = SDL_GetWindowFromID(0);
SDLTest_AssertPass("Call to SDL_GetWindowID(0)");
result = SDL_GetWindowFromID(UINT32_MAX);
SDLTest_AssertPass("Call to SDL_GetWindowID(UINT32_MAX)");
/* Clean up */
Remove more reserved identifiers (#6925)
2022-12-30 00:58:16 +03:00
destroyVideoSuiteTestWindow(window);
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
/* Get window from ID for closed window */
result = SDL_GetWindowFromID(id);
SDLTest_AssertPass("Call to SDL_GetWindowID(%" SDL_PRIu32 "/closed_window)", id);
SDLTest_AssertCheck(result == NULL, "Verify result is NULL");
/* Negative test */
SDL_ClearError();
SDLTest_AssertPass("Call to SDL_ClearError()");
id = SDL_GetWindowID(NULL);
SDLTest_AssertPass("Call to SDL_GetWindowID(window=NULL)");
Remove more reserved identifiers (#6925)
2022-12-30 00:58:16 +03:00
checkInvalidWindowError();
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
return TEST_COMPLETED;
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
}
/**
docs: Remove Doxygen `\brief` tags. Doxygen and the wiki bridge don't need them; they'll both just use the first line/sentence instead. Fixes #8446.
2023-11-06 18:26:06 +03:00
* Tests call to SDL_GetWindowPixelFormat
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
*
test: don't use wiki urls for documentation comments Also make consistent use of \ as documentation escape character.
2023-02-02 02:21:53 +03:00
* \sa SDL_GetWindowPixelFormat
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
*/
SDL_test: use SDLCALL calling convention This is needed when using a pre-built static SDL3_test library.
2024-09-06 04:21:13 +03:00
static int SDLCALL video_getWindowPixelFormat(void *arg)
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
{
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
const char *title = "video_getWindowPixelFormat Test Window";
SDL_Window *window;
Changed SDL_GetWindowPixelFormat() to return SDL_PixelFormat Fixes https://github.com/libsdl-org/SDL/issues/10257 Closes https://github.com/libsdl-org/SDL/pull/10258
2024-07-14 17:19:20 +03:00
SDL_PixelFormat format;
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
/* Call against new test window */
Remove more reserved identifiers (#6925)
2022-12-30 00:58:16 +03:00
window = createVideoSuiteTestWindow(title);
Pointer as bool (libsdl-org#7214)
2023-11-10 00:29:15 +03:00
if (!window) {
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
return TEST_ABORTED;
}
/* Get format */
format = SDL_GetWindowPixelFormat(window);
SDLTest_AssertPass("Call to SDL_GetWindowPixelFormat()");
Fixed error: format '%lu' expects argument of type 'long unsigned int', but argument 4 has type 'unsigned int'
2024-07-14 18:51:43 +03:00
SDLTest_AssertCheck(format != SDL_PIXELFORMAT_UNKNOWN, "Verify that returned format is valid; expected: != SDL_PIXELFORMAT_UNKNOWN, got: SDL_PIXELFORMAT_UNKNOWN");
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
/* Clean up */
Remove more reserved identifiers (#6925)
2022-12-30 00:58:16 +03:00
destroyVideoSuiteTestWindow(window);
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
/* Negative test */
SDL_ClearError();
SDLTest_AssertPass("Call to SDL_ClearError()");
format = SDL_GetWindowPixelFormat(NULL);
SDLTest_AssertPass("Call to SDL_GetWindowPixelFormat(window=NULL)");
Remove more reserved identifiers (#6925)
2022-12-30 00:58:16 +03:00
checkInvalidWindowError();
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
return TEST_COMPLETED;
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
}
testautomation_video: if SDL_SetWindowSize/Position isn't honored, we should check there is an event x11: send the events if various occasions
2023-03-20 13:25:10 +03:00
Removed SDL_bool in favor of plain bool We require stdbool.h in the build environment, so we might as well use the plain bool type. If your environment doesn't have stdbool.h, this simple replacement will suffice: typedef signed char bool;
2024-09-18 17:52:28 +03:00
static bool getPositionFromEvent(int *x, int *y)
testautomation_video: if SDL_SetWindowSize/Position isn't honored, we should check there is an event x11: send the events if various occasions
2023-03-20 13:25:10 +03:00
{
Removed SDL_bool in favor of plain bool We require stdbool.h in the build environment, so we might as well use the plain bool type. If your environment doesn't have stdbool.h, this simple replacement will suffice: typedef signed char bool;
2024-09-18 17:52:28 +03:00
bool ret = false;
testautomation_video: if SDL_SetWindowSize/Position isn't honored, we should check there is an event x11: send the events if various occasions
2023-03-20 13:25:10 +03:00
SDL_Event evt;
SDL_zero(evt);
while (SDL_PollEvent(&evt)) {
if (evt.type == SDL_EVENT_WINDOW_MOVED) {
*x = evt.window.data1;
*y = evt.window.data2;
Removed SDL_bool in favor of plain bool We require stdbool.h in the build environment, so we might as well use the plain bool type. If your environment doesn't have stdbool.h, this simple replacement will suffice: typedef signed char bool;
2024-09-18 17:52:28 +03:00
ret = true;
testautomation_video: if SDL_SetWindowSize/Position isn't honored, we should check there is an event x11: send the events if various occasions
2023-03-20 13:25:10 +03:00
}
}
return ret;
}
Removed SDL_bool in favor of plain bool We require stdbool.h in the build environment, so we might as well use the plain bool type. If your environment doesn't have stdbool.h, this simple replacement will suffice: typedef signed char bool;
2024-09-18 17:52:28 +03:00
static bool getSizeFromEvent(int *w, int *h)
testautomation_video: if SDL_SetWindowSize/Position isn't honored, we should check there is an event x11: send the events if various occasions
2023-03-20 13:25:10 +03:00
{
Removed SDL_bool in favor of plain bool We require stdbool.h in the build environment, so we might as well use the plain bool type. If your environment doesn't have stdbool.h, this simple replacement will suffice: typedef signed char bool;
2024-09-18 17:52:28 +03:00
bool ret = false;
testautomation_video: if SDL_SetWindowSize/Position isn't honored, we should check there is an event x11: send the events if various occasions
2023-03-20 13:25:10 +03:00
SDL_Event evt;
SDL_zero(evt);
while (SDL_PollEvent(&evt)) {
if (evt.type == SDL_EVENT_WINDOW_RESIZED) {
*w = evt.window.data1;
*h = evt.window.data2;
Removed SDL_bool in favor of plain bool We require stdbool.h in the build environment, so we might as well use the plain bool type. If your environment doesn't have stdbool.h, this simple replacement will suffice: typedef signed char bool;
2024-09-18 17:52:28 +03:00
ret = true;
testautomation_video: if SDL_SetWindowSize/Position isn't honored, we should check there is an event x11: send the events if various occasions
2023-03-20 13:25:10 +03:00
}
}
return ret;
}
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
/**
docs: Remove Doxygen `\brief` tags. Doxygen and the wiki bridge don't need them; they'll both just use the first line/sentence instead. Fixes #8446.
2023-11-06 18:26:06 +03:00
* Tests call to SDL_GetWindowPosition and SDL_SetWindowPosition
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
*
test: don't use wiki urls for documentation comments Also make consistent use of \ as documentation escape character.
2023-02-02 02:21:53 +03:00
* \sa SDL_GetWindowPosition
* \sa SDL_SetWindowPosition
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
*/
SDL_test: use SDLCALL calling convention This is needed when using a pre-built static SDL3_test library.
2024-09-06 04:21:13 +03:00
static int SDLCALL video_getSetWindowPosition(void *arg)
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
{
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
const char *title = "video_getSetWindowPosition Test Window";
SDL_Window *window;
video: Implement asynchronous windowing SDL window size, state, and position functions have been considered immediate, with their effects assuming to have taken effect upon successful return of the function. However, several windowing systems handle these requests asynchronously, resulting in the functions blocking until the changes have taken effect, potentially for long periods of time. Additionally, some windowing systems treat these as requests, and can potentially deny or fulfill the request in a manner differently than the application expects, such as not allowing a window to be positioned or sized beyond desktop borders, prohibiting fullscreen, and so on. With these changes, applications can make requests of the window manager that do not block, with the understanding that an associated event will be sent if the request is fulfilled. Currently, size, position, maximize, minimize, and fullscreen calls are handled as asynchronous requests, with events being returned if the request is honored. If the application requires that the change take effect immediately, it can call the new SDL_SyncWindow function, which will attempt to block until the request is fulfilled, or some arbitrary timeout period elapses, the duration of which depends not only on the windowing system, but on the operation requested as well (e.g. a 100ms timeout is fine for most X11 events, but maximizing a window can take considerably longer for some reason). There is also a new hint 'SDL_VIDEO_SYNC_ALL_WINDOW_OPS' that will mimic the old behavior by synchronizing after every window operation with, again, the understanding that using this may result in the associated calls blocking for a relatively long period. The deferred model also results in the window size and position getters not reporting false coordinates anymore, as they only forward what the window manager reports vs allowing applications to set arbitrary values, and fullscreen enter/leave events that were initiated via the window manager update the window state appropriately, where they didn't before. Care was taken to ensure that order of operations is maintained, and that requests are not ignored or dropped. This does require some implicit internal synchronization in the various backends if many requests are made in a short period, as some state and behavior depends on other bits of state that need to be known at that particular point in time, but this isn't something that typical applications will hit, unless they are sending a lot of window state in a short time as the tests do. The automated tests developed to test the previous behavior also resulted in previously undefined behavior being defined and normalized across platforms, particularly when it comes to the sizing and positioning of windows when they are in a fixed-size state, such as maximized or fullscreen. Size and position requests made when the window is not in a movable or resizable state will be deferred until it can be applied, so no requests are lost. These changes fix another long-standing issue with renderers recreating maximized windows, where the original non-maximized size was lost, resulting in the window being restored to the wrong size. All automated video tests pass across all platforms. Overall, the "make a request/get an event" model better reflects how most windowing systems work, and some backends avoid spending significant time blocking while waiting for operations to complete.
2023-10-25 22:09:55 +03:00
int result;
int maxxVariation, maxyVariation;
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
int xVariation, yVariation;
int referenceX, referenceY;
int currentX, currentY;
int desiredX, desiredY;
video: Implement asynchronous windowing SDL window size, state, and position functions have been considered immediate, with their effects assuming to have taken effect upon successful return of the function. However, several windowing systems handle these requests asynchronously, resulting in the functions blocking until the changes have taken effect, potentially for long periods of time. Additionally, some windowing systems treat these as requests, and can potentially deny or fulfill the request in a manner differently than the application expects, such as not allowing a window to be positioned or sized beyond desktop borders, prohibiting fullscreen, and so on. With these changes, applications can make requests of the window manager that do not block, with the understanding that an associated event will be sent if the request is fulfilled. Currently, size, position, maximize, minimize, and fullscreen calls are handled as asynchronous requests, with events being returned if the request is honored. If the application requires that the change take effect immediately, it can call the new SDL_SyncWindow function, which will attempt to block until the request is fulfilled, or some arbitrary timeout period elapses, the duration of which depends not only on the windowing system, but on the operation requested as well (e.g. a 100ms timeout is fine for most X11 events, but maximizing a window can take considerably longer for some reason). There is also a new hint 'SDL_VIDEO_SYNC_ALL_WINDOW_OPS' that will mimic the old behavior by synchronizing after every window operation with, again, the understanding that using this may result in the associated calls blocking for a relatively long period. The deferred model also results in the window size and position getters not reporting false coordinates anymore, as they only forward what the window manager reports vs allowing applications to set arbitrary values, and fullscreen enter/leave events that were initiated via the window manager update the window state appropriately, where they didn't before. Care was taken to ensure that order of operations is maintained, and that requests are not ignored or dropped. This does require some implicit internal synchronization in the various backends if many requests are made in a short period, as some state and behavior depends on other bits of state that need to be known at that particular point in time, but this isn't something that typical applications will hit, unless they are sending a lot of window state in a short time as the tests do. The automated tests developed to test the previous behavior also resulted in previously undefined behavior being defined and normalized across platforms, particularly when it comes to the sizing and positioning of windows when they are in a fixed-size state, such as maximized or fullscreen. Size and position requests made when the window is not in a movable or resizable state will be deferred until it can be applied, so no requests are lost. These changes fix another long-standing issue with renderers recreating maximized windows, where the original non-maximized size was lost, resulting in the window being restored to the wrong size. All automated video tests pass across all platforms. Overall, the "make a request/get an event" model better reflects how most windowing systems work, and some backends avoid spending significant time blocking while waiting for operations to complete.
2023-10-25 22:09:55 +03:00
SDL_Rect display_bounds;
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
/* Call against new test window */
Remove more reserved identifiers (#6925)
2022-12-30 00:58:16 +03:00
window = createVideoSuiteTestWindow(title);
Pointer as bool (libsdl-org#7214)
2023-11-10 00:29:15 +03:00
if (!window) {
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
return TEST_ABORTED;
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
}
video: Implement asynchronous windowing SDL window size, state, and position functions have been considered immediate, with their effects assuming to have taken effect upon successful return of the function. However, several windowing systems handle these requests asynchronously, resulting in the functions blocking until the changes have taken effect, potentially for long periods of time. Additionally, some windowing systems treat these as requests, and can potentially deny or fulfill the request in a manner differently than the application expects, such as not allowing a window to be positioned or sized beyond desktop borders, prohibiting fullscreen, and so on. With these changes, applications can make requests of the window manager that do not block, with the understanding that an associated event will be sent if the request is fulfilled. Currently, size, position, maximize, minimize, and fullscreen calls are handled as asynchronous requests, with events being returned if the request is honored. If the application requires that the change take effect immediately, it can call the new SDL_SyncWindow function, which will attempt to block until the request is fulfilled, or some arbitrary timeout period elapses, the duration of which depends not only on the windowing system, but on the operation requested as well (e.g. a 100ms timeout is fine for most X11 events, but maximizing a window can take considerably longer for some reason). There is also a new hint 'SDL_VIDEO_SYNC_ALL_WINDOW_OPS' that will mimic the old behavior by synchronizing after every window operation with, again, the understanding that using this may result in the associated calls blocking for a relatively long period. The deferred model also results in the window size and position getters not reporting false coordinates anymore, as they only forward what the window manager reports vs allowing applications to set arbitrary values, and fullscreen enter/leave events that were initiated via the window manager update the window state appropriately, where they didn't before. Care was taken to ensure that order of operations is maintained, and that requests are not ignored or dropped. This does require some implicit internal synchronization in the various backends if many requests are made in a short period, as some state and behavior depends on other bits of state that need to be known at that particular point in time, but this isn't something that typical applications will hit, unless they are sending a lot of window state in a short time as the tests do. The automated tests developed to test the previous behavior also resulted in previously undefined behavior being defined and normalized across platforms, particularly when it comes to the sizing and positioning of windows when they are in a fixed-size state, such as maximized or fullscreen. Size and position requests made when the window is not in a movable or resizable state will be deferred until it can be applied, so no requests are lost. These changes fix another long-standing issue with renderers recreating maximized windows, where the original non-maximized size was lost, resulting in the window being restored to the wrong size. All automated video tests pass across all platforms. Overall, the "make a request/get an event" model better reflects how most windowing systems work, and some backends avoid spending significant time blocking while waiting for operations to complete.
2023-10-25 22:09:55 +03:00
/* Sanity check */
SDL_GetWindowPosition(window, &currentX, &currentY);
Use SDL_bool instead an int return code in the SDL API Most SDL functions used to indicate success or failure using an int return code. These functions have been changed to return SDL_bool. Here is a coccinelle patch to change code that previously compared the return value to 0 and changes it to a boolean test: @ bool_return_type @ identifier func =~ "^(SDL_AddEventWatch|SDL_AddHintCallback|SDL_AddSurfaceAlternateImage|SDL_AddVulkanRenderSemaphores|SDL_BindAudioStream|SDL_BindAudioStreams|SDL_BlitSurface|SDL_BlitSurface9Grid|SDL_BlitSurfaceScaled|SDL_BlitSurfaceTiled|SDL_BlitSurfaceTiledWithScale|SDL_BlitSurfaceUnchecked|SDL_BlitSurfaceUncheckedScaled|SDL_CaptureMouse|SDL_ClearAudioStream|SDL_ClearClipboardData|SDL_ClearComposition|SDL_ClearError|SDL_ClearProperty|SDL_ClearSurface|SDL_CloseIO|SDL_CloseStorage|SDL_ConvertAudioSamples|SDL_ConvertEventToRenderCoordinates|SDL_ConvertPixels|SDL_ConvertPixelsAndColorspace|SDL_CopyFile|SDL_CopyProperties|SDL_CopyStorageFile|SDL_CreateDirectory|SDL_CreateStorageDirectory|SDL_CreateWindowAndRenderer|SDL_DateTimeToTime|SDL_DestroyWindowSurface|SDL_DetachVirtualJoystick|SDL_DisableScreenSaver|SDL_EnableScreenSaver|SDL_EnumerateDirectory|SDL_EnumerateProperties|SDL_EnumerateStorageDirectory|SDL_FillSurfaceRect|SDL_FillSurfaceRects|SDL_FlashWindow|SDL_FlipSurface|SDL_FlushAudioStream|SDL_FlushRenderer|SDL_GL_DestroyContext|SDL_GL_GetAttribute|SDL_GL_GetSwapInterval|SDL_GL_LoadLibrary|SDL_GL_MakeCurrent|SDL_GL_SetAttribute|SDL_GL_SetSwapInterval|SDL_GL_SwapWindow|SDL_GetAudioDeviceFormat|SDL_GetAudioStreamFormat|SDL_GetCameraFormat|SDL_GetClosestFullscreenDisplayMode|SDL_GetCurrentRenderOutputSize|SDL_GetCurrentTime|SDL_GetDXGIOutputInfo|SDL_GetDateTimeLocalePreferences|SDL_GetDisplayBounds|SDL_GetDisplayUsableBounds|SDL_GetGDKDefaultUser|SDL_GetGDKTaskQueue|SDL_GetGamepadSensorData|SDL_GetGamepadTouchpadFinger|SDL_GetHapticEffectStatus|SDL_GetJoystickBall|SDL_GetMasksForPixelFormat|SDL_GetPathInfo|SDL_GetRectUnion|SDL_GetRectUnionFloat|SDL_GetRenderClipRect|SDL_GetRenderColorScale|SDL_GetRenderDrawBlendMode|SDL_GetRenderDrawColor|SDL_GetRenderDrawColorFloat|SDL_GetRenderLogicalPresentation|SDL_GetRenderLogicalPresentationRect|SDL_GetRenderOutputSize|SDL_GetRenderSafeArea|SDL_GetRenderScale|SDL_GetRenderVSync|SDL_GetRenderViewport|SDL_GetSensorData|SDL_GetStorageFileSize|SDL_GetStoragePathInfo|SDL_GetSurfaceAlphaMod|SDL_GetSurfaceBlendMode|SDL_GetSurfaceClipRect|SDL_GetSurfaceColorKey|SDL_GetSurfaceColorMod|SDL_GetTextInputArea|SDL_GetTextureAlphaMod|SDL_GetTextureAlphaModFloat|SDL_GetTextureBlendMode|SDL_GetTextureColorMod|SDL_GetTextureColorModFloat|SDL_GetTextureScaleMode|SDL_GetTextureSize|SDL_GetWindowAspectRatio|SDL_GetWindowBordersSize|SDL_GetWindowMaximumSize|SDL_GetWindowMinimumSize|SDL_GetWindowPosition|SDL_GetWindowRelativeMouseMode|SDL_GetWindowSafeArea|SDL_GetWindowSize|SDL_GetWindowSizeInPixels|SDL_GetWindowSurfaceVSync|SDL_HideCursor|SDL_HideWindow|SDL_Init|SDL_InitHapticRumble|SDL_InitSubSystem|SDL_LoadWAV|SDL_LoadWAV_IO|SDL_LockAudioStream|SDL_LockProperties|SDL_LockSurface|SDL_LockTexture|SDL_LockTextureToSurface|SDL_MaximizeWindow|SDL_MinimizeWindow|SDL_MixAudio|SDL_OpenURL|SDL_OutOfMemory|SDL_PauseAudioDevice|SDL_PauseAudioStreamDevice|SDL_PauseHaptic|SDL_PlayHapticRumble|SDL_PremultiplyAlpha|SDL_PremultiplySurfaceAlpha|SDL_PushEvent|SDL_PutAudioStreamData|SDL_RaiseWindow|SDL_ReadStorageFile|SDL_ReadSurfacePixel|SDL_ReadSurfacePixelFloat|SDL_RegisterApp|SDL_ReloadGamepadMappings|SDL_RemovePath|SDL_RemoveStoragePath|SDL_RemoveTimer|SDL_RenamePath|SDL_RenameStoragePath|SDL_RenderClear|SDL_RenderCoordinatesFromWindow|SDL_RenderCoordinatesToWindow|SDL_RenderFillRect|SDL_RenderFillRects|SDL_RenderGeometry|SDL_RenderGeometryRaw|SDL_RenderLine|SDL_RenderLines|SDL_RenderPoint|SDL_RenderPoints|SDL_RenderPresent|SDL_RenderRect|SDL_RenderRects|SDL_RenderTexture|SDL_RenderTexture9Grid|SDL_RenderTextureRotated|SDL_RenderTextureTiled|SDL_RequestAndroidPermission|SDL_RestoreWindow|SDL_ResumeAudioDevice|SDL_ResumeAudioStreamDevice|SDL_ResumeHaptic|SDL_RumbleGamepad|SDL_RumbleGamepadTriggers|SDL_RumbleJoystick|SDL_RumbleJoystickTriggers|SDL_RunHapticEffect|SDL_SaveBMP|SDL_SaveBMP_IO|SDL_SendAndroidMessage|SDL_SendGamepadEffect|SDL_SendJoystickEffect|SDL_SendJoystickVirtualSensorData|SDL_SetAppMetadata|SDL_SetAppMetadataProperty|SDL_SetAudioDeviceGain|SDL_SetAudioPostmixCallback|SDL_SetAudioStreamFormat|SDL_SetAudioStreamFrequencyRatio|SDL_SetAudioStreamGain|SDL_SetAudioStreamGetCallback|SDL_SetAudioStreamInputChannelMap|SDL_SetAudioStreamOutputChannelMap|SDL_SetAudioStreamPutCallback|SDL_SetBooleanProperty|SDL_SetClipboardData|SDL_SetClipboardText|SDL_SetCursor|SDL_SetFloatProperty|SDL_SetGamepadLED|SDL_SetGamepadMapping|SDL_SetGamepadPlayerIndex|SDL_SetGamepadSensorEnabled|SDL_SetHapticAutocenter|SDL_SetHapticGain|SDL_SetJoystickLED|SDL_SetJoystickPlayerIndex|SDL_SetJoystickVirtualAxis|SDL_SetJoystickVirtualBall|SDL_SetJoystickVirtualButton|SDL_SetJoystickVirtualHat|SDL_SetJoystickVirtualTouchpad|SDL_SetLinuxThreadPriority|SDL_SetLinuxThreadPriorityAndPolicy|SDL_SetLogPriorityPrefix|SDL_SetMemoryFunctions|SDL_SetNumberProperty|SDL_SetPaletteColors|SDL_SetPointerProperty|SDL_SetPointerPropertyWithCleanup|SDL_SetPrimarySelectionText|SDL_SetRenderClipRect|SDL_SetRenderColorScale|SDL_SetRenderDrawBlendMode|SDL_SetRenderDrawColor|SDL_SetRenderDrawColorFloat|SDL_SetRenderLogicalPresentation|SDL_SetRenderScale|SDL_SetRenderTarget|SDL_SetRenderVSync|SDL_SetRenderViewport|SDL_SetScancodeName|SDL_SetStringProperty|SDL_SetSurfaceAlphaMod|SDL_SetSurfaceBlendMode|SDL_SetSurfaceColorKey|SDL_SetSurfaceColorMod|SDL_SetSurfaceColorspace|SDL_SetSurfacePalette|SDL_SetSurfaceRLE|SDL_SetTLS|SDL_SetTextInputArea|SDL_SetTextureAlphaMod|SDL_SetTextureAlphaModFloat|SDL_SetTextureBlendMode|SDL_SetTextureColorMod|SDL_SetTextureColorModFloat|SDL_SetTextureScaleMode|SDL_SetThreadPriority|SDL_SetWindowAlwaysOnTop|SDL_SetWindowAspectRatio|SDL_SetWindowBordered|SDL_SetWindowFocusable|SDL_SetWindowFullscreen|SDL_SetWindowFullscreenMode|SDL_SetWindowHitTest|SDL_SetWindowIcon|SDL_SetWindowKeyboardGrab|SDL_SetWindowMaximumSize|SDL_SetWindowMinimumSize|SDL_SetWindowModalFor|SDL_SetWindowMouseGrab|SDL_SetWindowMouseRect|SDL_SetWindowOpacity|SDL_SetWindowPosition|SDL_SetWindowRelativeMouseMode|SDL_SetWindowResizable|SDL_SetWindowShape|SDL_SetWindowSize|SDL_SetWindowSurfaceVSync|SDL_SetWindowTitle|SDL_SetiOSAnimationCallback|SDL_ShowAndroidToast|SDL_ShowCursor|SDL_ShowMessageBox|SDL_ShowSimpleMessageBox|SDL_ShowWindow|SDL_ShowWindowSystemMenu|SDL_StartTextInput|SDL_StartTextInputWithProperties|SDL_StopHapticEffect|SDL_StopHapticEffects|SDL_StopHapticRumble|SDL_StopTextInput|SDL_SyncWindow|SDL_TimeToDateTime|SDL_TryLockMutex|SDL_TryLockRWLockForReading|SDL_TryLockRWLockForWriting|SDL_TryWaitSemaphore|SDL_UnlockAudioStream|SDL_UpdateHapticEffect|SDL_UpdateNVTexture|SDL_UpdateTexture|SDL_UpdateWindowSurface|SDL_UpdateWindowSurfaceRects|SDL_UpdateYUVTexture|SDL_Vulkan_CreateSurface|SDL_Vulkan_LoadLibrary|SDL_WaitConditionTimeout|SDL_WaitSemaphoreTimeout|SDL_WarpMouseGlobal|SDL_WriteStorageFile|SDL_WriteSurfacePixel|SDL_WriteSurfacePixelFloat)$"; @@ ( func( ... ) - == 0 | - func( + !func( ... ) - < 0 | - func( + !func( ... ) - != 0 | - func( + !func( ... ) - == -1 )
2024-08-23 03:33:49 +03:00
if (!SDL_SetWindowPosition(window, currentX, currentY)) {
video: Implement asynchronous windowing SDL window size, state, and position functions have been considered immediate, with their effects assuming to have taken effect upon successful return of the function. However, several windowing systems handle these requests asynchronously, resulting in the functions blocking until the changes have taken effect, potentially for long periods of time. Additionally, some windowing systems treat these as requests, and can potentially deny or fulfill the request in a manner differently than the application expects, such as not allowing a window to be positioned or sized beyond desktop borders, prohibiting fullscreen, and so on. With these changes, applications can make requests of the window manager that do not block, with the understanding that an associated event will be sent if the request is fulfilled. Currently, size, position, maximize, minimize, and fullscreen calls are handled as asynchronous requests, with events being returned if the request is honored. If the application requires that the change take effect immediately, it can call the new SDL_SyncWindow function, which will attempt to block until the request is fulfilled, or some arbitrary timeout period elapses, the duration of which depends not only on the windowing system, but on the operation requested as well (e.g. a 100ms timeout is fine for most X11 events, but maximizing a window can take considerably longer for some reason). There is also a new hint 'SDL_VIDEO_SYNC_ALL_WINDOW_OPS' that will mimic the old behavior by synchronizing after every window operation with, again, the understanding that using this may result in the associated calls blocking for a relatively long period. The deferred model also results in the window size and position getters not reporting false coordinates anymore, as they only forward what the window manager reports vs allowing applications to set arbitrary values, and fullscreen enter/leave events that were initiated via the window manager update the window state appropriately, where they didn't before. Care was taken to ensure that order of operations is maintained, and that requests are not ignored or dropped. This does require some implicit internal synchronization in the various backends if many requests are made in a short period, as some state and behavior depends on other bits of state that need to be known at that particular point in time, but this isn't something that typical applications will hit, unless they are sending a lot of window state in a short time as the tests do. The automated tests developed to test the previous behavior also resulted in previously undefined behavior being defined and normalized across platforms, particularly when it comes to the sizing and positioning of windows when they are in a fixed-size state, such as maximized or fullscreen. Size and position requests made when the window is not in a movable or resizable state will be deferred until it can be applied, so no requests are lost. These changes fix another long-standing issue with renderers recreating maximized windows, where the original non-maximized size was lost, resulting in the window being restored to the wrong size. All automated video tests pass across all platforms. Overall, the "make a request/get an event" model better reflects how most windowing systems work, and some backends avoid spending significant time blocking while waiting for operations to complete.
2023-10-25 22:09:55 +03:00
SDLTest_Log("Skipping window positioning tests: %s reports window positioning as unsupported", SDL_GetCurrentVideoDriver());
goto null_tests;
}
if (SDL_strcmp(SDL_GetCurrentVideoDriver(), "x11") == 0) {
/* The X11 server allows arbitrary window placement, but compositing
* window managers such as GNOME and KDE force windows to be within
* desktop bounds.
*/
maxxVariation = 2;
maxyVariation = 2;
SDL_GetDisplayUsableBounds(SDL_GetPrimaryDisplay(), &display_bounds);
} else if (SDL_strcmp(SDL_GetCurrentVideoDriver(), "cocoa") == 0) {
/* Platform doesn't allow windows with negative Y desktop bounds */
maxxVariation = 4;
maxyVariation = 3;
SDL_GetDisplayUsableBounds(SDL_GetPrimaryDisplay(), &display_bounds);
} else {
/* Platform allows windows to be placed out of bounds */
maxxVariation = 4;
maxyVariation = 4;
SDL_GetDisplayBounds(SDL_GetPrimaryDisplay(), &display_bounds);
}
for (xVariation = 0; xVariation < maxxVariation; xVariation++) {
for (yVariation = 0; yVariation < maxyVariation; yVariation++) {
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
switch (xVariation) {
default:
case 0:
/* Zero X Position */
video: Implement asynchronous windowing SDL window size, state, and position functions have been considered immediate, with their effects assuming to have taken effect upon successful return of the function. However, several windowing systems handle these requests asynchronously, resulting in the functions blocking until the changes have taken effect, potentially for long periods of time. Additionally, some windowing systems treat these as requests, and can potentially deny or fulfill the request in a manner differently than the application expects, such as not allowing a window to be positioned or sized beyond desktop borders, prohibiting fullscreen, and so on. With these changes, applications can make requests of the window manager that do not block, with the understanding that an associated event will be sent if the request is fulfilled. Currently, size, position, maximize, minimize, and fullscreen calls are handled as asynchronous requests, with events being returned if the request is honored. If the application requires that the change take effect immediately, it can call the new SDL_SyncWindow function, which will attempt to block until the request is fulfilled, or some arbitrary timeout period elapses, the duration of which depends not only on the windowing system, but on the operation requested as well (e.g. a 100ms timeout is fine for most X11 events, but maximizing a window can take considerably longer for some reason). There is also a new hint 'SDL_VIDEO_SYNC_ALL_WINDOW_OPS' that will mimic the old behavior by synchronizing after every window operation with, again, the understanding that using this may result in the associated calls blocking for a relatively long period. The deferred model also results in the window size and position getters not reporting false coordinates anymore, as they only forward what the window manager reports vs allowing applications to set arbitrary values, and fullscreen enter/leave events that were initiated via the window manager update the window state appropriately, where they didn't before. Care was taken to ensure that order of operations is maintained, and that requests are not ignored or dropped. This does require some implicit internal synchronization in the various backends if many requests are made in a short period, as some state and behavior depends on other bits of state that need to be known at that particular point in time, but this isn't something that typical applications will hit, unless they are sending a lot of window state in a short time as the tests do. The automated tests developed to test the previous behavior also resulted in previously undefined behavior being defined and normalized across platforms, particularly when it comes to the sizing and positioning of windows when they are in a fixed-size state, such as maximized or fullscreen. Size and position requests made when the window is not in a movable or resizable state will be deferred until it can be applied, so no requests are lost. These changes fix another long-standing issue with renderers recreating maximized windows, where the original non-maximized size was lost, resulting in the window being restored to the wrong size. All automated video tests pass across all platforms. Overall, the "make a request/get an event" model better reflects how most windowing systems work, and some backends avoid spending significant time blocking while waiting for operations to complete.
2023-10-25 22:09:55 +03:00
desiredX = display_bounds.x > 0 ? display_bounds.x : 0;
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
break;
case 1:
/* Random X position inside screen */
video: Implement asynchronous windowing SDL window size, state, and position functions have been considered immediate, with their effects assuming to have taken effect upon successful return of the function. However, several windowing systems handle these requests asynchronously, resulting in the functions blocking until the changes have taken effect, potentially for long periods of time. Additionally, some windowing systems treat these as requests, and can potentially deny or fulfill the request in a manner differently than the application expects, such as not allowing a window to be positioned or sized beyond desktop borders, prohibiting fullscreen, and so on. With these changes, applications can make requests of the window manager that do not block, with the understanding that an associated event will be sent if the request is fulfilled. Currently, size, position, maximize, minimize, and fullscreen calls are handled as asynchronous requests, with events being returned if the request is honored. If the application requires that the change take effect immediately, it can call the new SDL_SyncWindow function, which will attempt to block until the request is fulfilled, or some arbitrary timeout period elapses, the duration of which depends not only on the windowing system, but on the operation requested as well (e.g. a 100ms timeout is fine for most X11 events, but maximizing a window can take considerably longer for some reason). There is also a new hint 'SDL_VIDEO_SYNC_ALL_WINDOW_OPS' that will mimic the old behavior by synchronizing after every window operation with, again, the understanding that using this may result in the associated calls blocking for a relatively long period. The deferred model also results in the window size and position getters not reporting false coordinates anymore, as they only forward what the window manager reports vs allowing applications to set arbitrary values, and fullscreen enter/leave events that were initiated via the window manager update the window state appropriately, where they didn't before. Care was taken to ensure that order of operations is maintained, and that requests are not ignored or dropped. This does require some implicit internal synchronization in the various backends if many requests are made in a short period, as some state and behavior depends on other bits of state that need to be known at that particular point in time, but this isn't something that typical applications will hit, unless they are sending a lot of window state in a short time as the tests do. The automated tests developed to test the previous behavior also resulted in previously undefined behavior being defined and normalized across platforms, particularly when it comes to the sizing and positioning of windows when they are in a fixed-size state, such as maximized or fullscreen. Size and position requests made when the window is not in a movable or resizable state will be deferred until it can be applied, so no requests are lost. These changes fix another long-standing issue with renderers recreating maximized windows, where the original non-maximized size was lost, resulting in the window being restored to the wrong size. All automated video tests pass across all platforms. Overall, the "make a request/get an event" model better reflects how most windowing systems work, and some backends avoid spending significant time blocking while waiting for operations to complete.
2023-10-25 22:09:55 +03:00
desiredX = SDLTest_RandomIntegerInRange(display_bounds.x + 1, display_bounds.x + 100);
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
break;
case 2:
/* Random X position outside screen (positive) */
desiredX = SDLTest_RandomIntegerInRange(10000, 11000);
break;
case 3:
/* Random X position outside screen (negative) */
desiredX = SDLTest_RandomIntegerInRange(-1000, -100);
break;
}
switch (yVariation) {
default:
case 0:
video: Implement asynchronous windowing SDL window size, state, and position functions have been considered immediate, with their effects assuming to have taken effect upon successful return of the function. However, several windowing systems handle these requests asynchronously, resulting in the functions blocking until the changes have taken effect, potentially for long periods of time. Additionally, some windowing systems treat these as requests, and can potentially deny or fulfill the request in a manner differently than the application expects, such as not allowing a window to be positioned or sized beyond desktop borders, prohibiting fullscreen, and so on. With these changes, applications can make requests of the window manager that do not block, with the understanding that an associated event will be sent if the request is fulfilled. Currently, size, position, maximize, minimize, and fullscreen calls are handled as asynchronous requests, with events being returned if the request is honored. If the application requires that the change take effect immediately, it can call the new SDL_SyncWindow function, which will attempt to block until the request is fulfilled, or some arbitrary timeout period elapses, the duration of which depends not only on the windowing system, but on the operation requested as well (e.g. a 100ms timeout is fine for most X11 events, but maximizing a window can take considerably longer for some reason). There is also a new hint 'SDL_VIDEO_SYNC_ALL_WINDOW_OPS' that will mimic the old behavior by synchronizing after every window operation with, again, the understanding that using this may result in the associated calls blocking for a relatively long period. The deferred model also results in the window size and position getters not reporting false coordinates anymore, as they only forward what the window manager reports vs allowing applications to set arbitrary values, and fullscreen enter/leave events that were initiated via the window manager update the window state appropriately, where they didn't before. Care was taken to ensure that order of operations is maintained, and that requests are not ignored or dropped. This does require some implicit internal synchronization in the various backends if many requests are made in a short period, as some state and behavior depends on other bits of state that need to be known at that particular point in time, but this isn't something that typical applications will hit, unless they are sending a lot of window state in a short time as the tests do. The automated tests developed to test the previous behavior also resulted in previously undefined behavior being defined and normalized across platforms, particularly when it comes to the sizing and positioning of windows when they are in a fixed-size state, such as maximized or fullscreen. Size and position requests made when the window is not in a movable or resizable state will be deferred until it can be applied, so no requests are lost. These changes fix another long-standing issue with renderers recreating maximized windows, where the original non-maximized size was lost, resulting in the window being restored to the wrong size. All automated video tests pass across all platforms. Overall, the "make a request/get an event" model better reflects how most windowing systems work, and some backends avoid spending significant time blocking while waiting for operations to complete.
2023-10-25 22:09:55 +03:00
/* Zero Y Position */
desiredY = display_bounds.y > 0 ? display_bounds.y : 0;
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
break;
case 1:
video: Implement asynchronous windowing SDL window size, state, and position functions have been considered immediate, with their effects assuming to have taken effect upon successful return of the function. However, several windowing systems handle these requests asynchronously, resulting in the functions blocking until the changes have taken effect, potentially for long periods of time. Additionally, some windowing systems treat these as requests, and can potentially deny or fulfill the request in a manner differently than the application expects, such as not allowing a window to be positioned or sized beyond desktop borders, prohibiting fullscreen, and so on. With these changes, applications can make requests of the window manager that do not block, with the understanding that an associated event will be sent if the request is fulfilled. Currently, size, position, maximize, minimize, and fullscreen calls are handled as asynchronous requests, with events being returned if the request is honored. If the application requires that the change take effect immediately, it can call the new SDL_SyncWindow function, which will attempt to block until the request is fulfilled, or some arbitrary timeout period elapses, the duration of which depends not only on the windowing system, but on the operation requested as well (e.g. a 100ms timeout is fine for most X11 events, but maximizing a window can take considerably longer for some reason). There is also a new hint 'SDL_VIDEO_SYNC_ALL_WINDOW_OPS' that will mimic the old behavior by synchronizing after every window operation with, again, the understanding that using this may result in the associated calls blocking for a relatively long period. The deferred model also results in the window size and position getters not reporting false coordinates anymore, as they only forward what the window manager reports vs allowing applications to set arbitrary values, and fullscreen enter/leave events that were initiated via the window manager update the window state appropriately, where they didn't before. Care was taken to ensure that order of operations is maintained, and that requests are not ignored or dropped. This does require some implicit internal synchronization in the various backends if many requests are made in a short period, as some state and behavior depends on other bits of state that need to be known at that particular point in time, but this isn't something that typical applications will hit, unless they are sending a lot of window state in a short time as the tests do. The automated tests developed to test the previous behavior also resulted in previously undefined behavior being defined and normalized across platforms, particularly when it comes to the sizing and positioning of windows when they are in a fixed-size state, such as maximized or fullscreen. Size and position requests made when the window is not in a movable or resizable state will be deferred until it can be applied, so no requests are lost. These changes fix another long-standing issue with renderers recreating maximized windows, where the original non-maximized size was lost, resulting in the window being restored to the wrong size. All automated video tests pass across all platforms. Overall, the "make a request/get an event" model better reflects how most windowing systems work, and some backends avoid spending significant time blocking while waiting for operations to complete.
2023-10-25 22:09:55 +03:00
/* Random Y position inside screen */
desiredY = SDLTest_RandomIntegerInRange(display_bounds.y + 1, display_bounds.y + 100);
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
break;
case 2:
video: Implement asynchronous windowing SDL window size, state, and position functions have been considered immediate, with their effects assuming to have taken effect upon successful return of the function. However, several windowing systems handle these requests asynchronously, resulting in the functions blocking until the changes have taken effect, potentially for long periods of time. Additionally, some windowing systems treat these as requests, and can potentially deny or fulfill the request in a manner differently than the application expects, such as not allowing a window to be positioned or sized beyond desktop borders, prohibiting fullscreen, and so on. With these changes, applications can make requests of the window manager that do not block, with the understanding that an associated event will be sent if the request is fulfilled. Currently, size, position, maximize, minimize, and fullscreen calls are handled as asynchronous requests, with events being returned if the request is honored. If the application requires that the change take effect immediately, it can call the new SDL_SyncWindow function, which will attempt to block until the request is fulfilled, or some arbitrary timeout period elapses, the duration of which depends not only on the windowing system, but on the operation requested as well (e.g. a 100ms timeout is fine for most X11 events, but maximizing a window can take considerably longer for some reason). There is also a new hint 'SDL_VIDEO_SYNC_ALL_WINDOW_OPS' that will mimic the old behavior by synchronizing after every window operation with, again, the understanding that using this may result in the associated calls blocking for a relatively long period. The deferred model also results in the window size and position getters not reporting false coordinates anymore, as they only forward what the window manager reports vs allowing applications to set arbitrary values, and fullscreen enter/leave events that were initiated via the window manager update the window state appropriately, where they didn't before. Care was taken to ensure that order of operations is maintained, and that requests are not ignored or dropped. This does require some implicit internal synchronization in the various backends if many requests are made in a short period, as some state and behavior depends on other bits of state that need to be known at that particular point in time, but this isn't something that typical applications will hit, unless they are sending a lot of window state in a short time as the tests do. The automated tests developed to test the previous behavior also resulted in previously undefined behavior being defined and normalized across platforms, particularly when it comes to the sizing and positioning of windows when they are in a fixed-size state, such as maximized or fullscreen. Size and position requests made when the window is not in a movable or resizable state will be deferred until it can be applied, so no requests are lost. These changes fix another long-standing issue with renderers recreating maximized windows, where the original non-maximized size was lost, resulting in the window being restored to the wrong size. All automated video tests pass across all platforms. Overall, the "make a request/get an event" model better reflects how most windowing systems work, and some backends avoid spending significant time blocking while waiting for operations to complete.
2023-10-25 22:09:55 +03:00
/* Random Y position outside screen (positive) */
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
desiredY = SDLTest_RandomIntegerInRange(10000, 11000);
break;
case 3:
/* Random Y position outside screen (negative) */
desiredY = SDLTest_RandomIntegerInRange(-1000, -100);
break;
}
/* Set position */
SDL_SetWindowPosition(window, desiredX, desiredY);
SDLTest_AssertPass("Call to SDL_SetWindowPosition(...,%d,%d)", desiredX, desiredY);
video: Implement asynchronous windowing SDL window size, state, and position functions have been considered immediate, with their effects assuming to have taken effect upon successful return of the function. However, several windowing systems handle these requests asynchronously, resulting in the functions blocking until the changes have taken effect, potentially for long periods of time. Additionally, some windowing systems treat these as requests, and can potentially deny or fulfill the request in a manner differently than the application expects, such as not allowing a window to be positioned or sized beyond desktop borders, prohibiting fullscreen, and so on. With these changes, applications can make requests of the window manager that do not block, with the understanding that an associated event will be sent if the request is fulfilled. Currently, size, position, maximize, minimize, and fullscreen calls are handled as asynchronous requests, with events being returned if the request is honored. If the application requires that the change take effect immediately, it can call the new SDL_SyncWindow function, which will attempt to block until the request is fulfilled, or some arbitrary timeout period elapses, the duration of which depends not only on the windowing system, but on the operation requested as well (e.g. a 100ms timeout is fine for most X11 events, but maximizing a window can take considerably longer for some reason). There is also a new hint 'SDL_VIDEO_SYNC_ALL_WINDOW_OPS' that will mimic the old behavior by synchronizing after every window operation with, again, the understanding that using this may result in the associated calls blocking for a relatively long period. The deferred model also results in the window size and position getters not reporting false coordinates anymore, as they only forward what the window manager reports vs allowing applications to set arbitrary values, and fullscreen enter/leave events that were initiated via the window manager update the window state appropriately, where they didn't before. Care was taken to ensure that order of operations is maintained, and that requests are not ignored or dropped. This does require some implicit internal synchronization in the various backends if many requests are made in a short period, as some state and behavior depends on other bits of state that need to be known at that particular point in time, but this isn't something that typical applications will hit, unless they are sending a lot of window state in a short time as the tests do. The automated tests developed to test the previous behavior also resulted in previously undefined behavior being defined and normalized across platforms, particularly when it comes to the sizing and positioning of windows when they are in a fixed-size state, such as maximized or fullscreen. Size and position requests made when the window is not in a movable or resizable state will be deferred until it can be applied, so no requests are lost. These changes fix another long-standing issue with renderers recreating maximized windows, where the original non-maximized size was lost, resulting in the window being restored to the wrong size. All automated video tests pass across all platforms. Overall, the "make a request/get an event" model better reflects how most windowing systems work, and some backends avoid spending significant time blocking while waiting for operations to complete.
2023-10-25 22:09:55 +03:00
result = SDL_SyncWindow(window);
SDLTest_AssertPass("SDL_SyncWindow()");
Removed SDL_bool in favor of plain bool We require stdbool.h in the build environment, so we might as well use the plain bool type. If your environment doesn't have stdbool.h, this simple replacement will suffice: typedef signed char bool;
2024-09-18 17:52:28 +03:00
SDLTest_AssertCheck(result == true, "Verify return value; expected: true, got: %d", result);
video: Implement asynchronous windowing SDL window size, state, and position functions have been considered immediate, with their effects assuming to have taken effect upon successful return of the function. However, several windowing systems handle these requests asynchronously, resulting in the functions blocking until the changes have taken effect, potentially for long periods of time. Additionally, some windowing systems treat these as requests, and can potentially deny or fulfill the request in a manner differently than the application expects, such as not allowing a window to be positioned or sized beyond desktop borders, prohibiting fullscreen, and so on. With these changes, applications can make requests of the window manager that do not block, with the understanding that an associated event will be sent if the request is fulfilled. Currently, size, position, maximize, minimize, and fullscreen calls are handled as asynchronous requests, with events being returned if the request is honored. If the application requires that the change take effect immediately, it can call the new SDL_SyncWindow function, which will attempt to block until the request is fulfilled, or some arbitrary timeout period elapses, the duration of which depends not only on the windowing system, but on the operation requested as well (e.g. a 100ms timeout is fine for most X11 events, but maximizing a window can take considerably longer for some reason). There is also a new hint 'SDL_VIDEO_SYNC_ALL_WINDOW_OPS' that will mimic the old behavior by synchronizing after every window operation with, again, the understanding that using this may result in the associated calls blocking for a relatively long period. The deferred model also results in the window size and position getters not reporting false coordinates anymore, as they only forward what the window manager reports vs allowing applications to set arbitrary values, and fullscreen enter/leave events that were initiated via the window manager update the window state appropriately, where they didn't before. Care was taken to ensure that order of operations is maintained, and that requests are not ignored or dropped. This does require some implicit internal synchronization in the various backends if many requests are made in a short period, as some state and behavior depends on other bits of state that need to be known at that particular point in time, but this isn't something that typical applications will hit, unless they are sending a lot of window state in a short time as the tests do. The automated tests developed to test the previous behavior also resulted in previously undefined behavior being defined and normalized across platforms, particularly when it comes to the sizing and positioning of windows when they are in a fixed-size state, such as maximized or fullscreen. Size and position requests made when the window is not in a movable or resizable state will be deferred until it can be applied, so no requests are lost. These changes fix another long-standing issue with renderers recreating maximized windows, where the original non-maximized size was lost, resulting in the window being restored to the wrong size. All automated video tests pass across all platforms. Overall, the "make a request/get an event" model better reflects how most windowing systems work, and some backends avoid spending significant time blocking while waiting for operations to complete.
2023-10-25 22:09:55 +03:00
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
/* Get position */
currentX = desiredX + 1;
currentY = desiredY + 1;
SDL_GetWindowPosition(window, &currentX, &currentY);
SDLTest_AssertPass("Call to SDL_GetWindowPosition()");
testautomation_video: if SDL_SetWindowSize/Position isn't honored, we should check there is an event x11: send the events if various occasions
2023-03-20 13:25:10 +03:00
if (desiredX == currentX && desiredY == currentY) {
SDLTest_AssertCheck(desiredX == currentX, "Verify returned X position; expected: %d, got: %d", desiredX, currentX);
SDLTest_AssertCheck(desiredY == currentY, "Verify returned Y position; expected: %d, got: %d", desiredY, currentY);
} else {
Removed SDL_bool in favor of plain bool We require stdbool.h in the build environment, so we might as well use the plain bool type. If your environment doesn't have stdbool.h, this simple replacement will suffice: typedef signed char bool;
2024-09-18 17:52:28 +03:00
bool hasEvent;
testautomation_video: if SDL_SetWindowSize/Position isn't honored, we should check there is an event x11: send the events if various occasions
2023-03-20 13:25:10 +03:00
/* SDL_SetWindowPosition() and SDL_SetWindowSize() will make requests of the window manager and set the internal position and size,
* and then we get events signaling what actually happened, and they get passed on to the application if they're not what we expect. */
video: Implement asynchronous windowing SDL window size, state, and position functions have been considered immediate, with their effects assuming to have taken effect upon successful return of the function. However, several windowing systems handle these requests asynchronously, resulting in the functions blocking until the changes have taken effect, potentially for long periods of time. Additionally, some windowing systems treat these as requests, and can potentially deny or fulfill the request in a manner differently than the application expects, such as not allowing a window to be positioned or sized beyond desktop borders, prohibiting fullscreen, and so on. With these changes, applications can make requests of the window manager that do not block, with the understanding that an associated event will be sent if the request is fulfilled. Currently, size, position, maximize, minimize, and fullscreen calls are handled as asynchronous requests, with events being returned if the request is honored. If the application requires that the change take effect immediately, it can call the new SDL_SyncWindow function, which will attempt to block until the request is fulfilled, or some arbitrary timeout period elapses, the duration of which depends not only on the windowing system, but on the operation requested as well (e.g. a 100ms timeout is fine for most X11 events, but maximizing a window can take considerably longer for some reason). There is also a new hint 'SDL_VIDEO_SYNC_ALL_WINDOW_OPS' that will mimic the old behavior by synchronizing after every window operation with, again, the understanding that using this may result in the associated calls blocking for a relatively long period. The deferred model also results in the window size and position getters not reporting false coordinates anymore, as they only forward what the window manager reports vs allowing applications to set arbitrary values, and fullscreen enter/leave events that were initiated via the window manager update the window state appropriately, where they didn't before. Care was taken to ensure that order of operations is maintained, and that requests are not ignored or dropped. This does require some implicit internal synchronization in the various backends if many requests are made in a short period, as some state and behavior depends on other bits of state that need to be known at that particular point in time, but this isn't something that typical applications will hit, unless they are sending a lot of window state in a short time as the tests do. The automated tests developed to test the previous behavior also resulted in previously undefined behavior being defined and normalized across platforms, particularly when it comes to the sizing and positioning of windows when they are in a fixed-size state, such as maximized or fullscreen. Size and position requests made when the window is not in a movable or resizable state will be deferred until it can be applied, so no requests are lost. These changes fix another long-standing issue with renderers recreating maximized windows, where the original non-maximized size was lost, resulting in the window being restored to the wrong size. All automated video tests pass across all platforms. Overall, the "make a request/get an event" model better reflects how most windowing systems work, and some backends avoid spending significant time blocking while waiting for operations to complete.
2023-10-25 22:09:55 +03:00
currentX = desiredX + 1;
currentY = desiredY + 1;
hasEvent = getPositionFromEvent(&currentX, &currentY);
Removed SDL_bool in favor of plain bool We require stdbool.h in the build environment, so we might as well use the plain bool type. If your environment doesn't have stdbool.h, this simple replacement will suffice: typedef signed char bool;
2024-09-18 17:52:28 +03:00
SDLTest_AssertCheck(hasEvent == true, "Changing position was not honored by WM, checking present of SDL_EVENT_WINDOW_MOVED");
testautomation_video: if SDL_SetWindowSize/Position isn't honored, we should check there is an event x11: send the events if various occasions
2023-03-20 13:25:10 +03:00
if (hasEvent) {
SDLTest_AssertCheck(desiredX == currentX, "Verify returned X position is the position from SDL event; expected: %d, got: %d", desiredX, currentX);
SDLTest_AssertCheck(desiredY == currentY, "Verify returned Y position is the position from SDL event; expected: %d, got: %d", desiredY, currentY);
}
}
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
/* Get position X */
currentX = desiredX + 1;
SDL_GetWindowPosition(window, &currentX, NULL);
SDLTest_AssertPass("Call to SDL_GetWindowPosition(&y=NULL)");
SDLTest_AssertCheck(desiredX == currentX, "Verify returned X position; expected: %d, got: %d", desiredX, currentX);
/* Get position Y */
currentY = desiredY + 1;
SDL_GetWindowPosition(window, NULL, &currentY);
SDLTest_AssertPass("Call to SDL_GetWindowPosition(&x=NULL)");
SDLTest_AssertCheck(desiredY == currentY, "Verify returned Y position; expected: %d, got: %d", desiredY, currentY);
}
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
}
video: Implement asynchronous windowing SDL window size, state, and position functions have been considered immediate, with their effects assuming to have taken effect upon successful return of the function. However, several windowing systems handle these requests asynchronously, resulting in the functions blocking until the changes have taken effect, potentially for long periods of time. Additionally, some windowing systems treat these as requests, and can potentially deny or fulfill the request in a manner differently than the application expects, such as not allowing a window to be positioned or sized beyond desktop borders, prohibiting fullscreen, and so on. With these changes, applications can make requests of the window manager that do not block, with the understanding that an associated event will be sent if the request is fulfilled. Currently, size, position, maximize, minimize, and fullscreen calls are handled as asynchronous requests, with events being returned if the request is honored. If the application requires that the change take effect immediately, it can call the new SDL_SyncWindow function, which will attempt to block until the request is fulfilled, or some arbitrary timeout period elapses, the duration of which depends not only on the windowing system, but on the operation requested as well (e.g. a 100ms timeout is fine for most X11 events, but maximizing a window can take considerably longer for some reason). There is also a new hint 'SDL_VIDEO_SYNC_ALL_WINDOW_OPS' that will mimic the old behavior by synchronizing after every window operation with, again, the understanding that using this may result in the associated calls blocking for a relatively long period. The deferred model also results in the window size and position getters not reporting false coordinates anymore, as they only forward what the window manager reports vs allowing applications to set arbitrary values, and fullscreen enter/leave events that were initiated via the window manager update the window state appropriately, where they didn't before. Care was taken to ensure that order of operations is maintained, and that requests are not ignored or dropped. This does require some implicit internal synchronization in the various backends if many requests are made in a short period, as some state and behavior depends on other bits of state that need to be known at that particular point in time, but this isn't something that typical applications will hit, unless they are sending a lot of window state in a short time as the tests do. The automated tests developed to test the previous behavior also resulted in previously undefined behavior being defined and normalized across platforms, particularly when it comes to the sizing and positioning of windows when they are in a fixed-size state, such as maximized or fullscreen. Size and position requests made when the window is not in a movable or resizable state will be deferred until it can be applied, so no requests are lost. These changes fix another long-standing issue with renderers recreating maximized windows, where the original non-maximized size was lost, resulting in the window being restored to the wrong size. All automated video tests pass across all platforms. Overall, the "make a request/get an event" model better reflects how most windowing systems work, and some backends avoid spending significant time blocking while waiting for operations to complete.
2023-10-25 22:09:55 +03:00
null_tests:
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
/* Dummy call with both pointers NULL */
SDL_GetWindowPosition(window, NULL, NULL);
SDLTest_AssertPass("Call to SDL_GetWindowPosition(&x=NULL,&y=NULL)");
/* Clean up */
Remove more reserved identifiers (#6925)
2022-12-30 00:58:16 +03:00
destroyVideoSuiteTestWindow(window);
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
/* Set some 'magic' value for later check that nothing was changed */
referenceX = SDLTest_RandomSint32();
referenceY = SDLTest_RandomSint32();
currentX = referenceX;
currentY = referenceY;
desiredX = SDLTest_RandomSint32();
desiredY = SDLTest_RandomSint32();
/* Negative tests */
SDL_ClearError();
SDLTest_AssertPass("Call to SDL_ClearError()");
SDL_GetWindowPosition(NULL, &currentX, &currentY);
SDLTest_AssertPass("Call to SDL_GetWindowPosition(window=NULL)");
SDLTest_AssertCheck(
currentX == referenceX && currentY == referenceY,
"Verify that content of X and Y pointers has not been modified; expected: %d,%d; got: %d,%d",
referenceX, referenceY,
currentX, currentY);
Remove more reserved identifiers (#6925)
2022-12-30 00:58:16 +03:00
checkInvalidWindowError();
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
SDL_GetWindowPosition(NULL, NULL, NULL);
SDLTest_AssertPass("Call to SDL_GetWindowPosition(NULL, NULL, NULL)");
Remove more reserved identifiers (#6925)
2022-12-30 00:58:16 +03:00
checkInvalidWindowError();
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
SDL_SetWindowPosition(NULL, desiredX, desiredY);
SDLTest_AssertPass("Call to SDL_SetWindowPosition(window=NULL)");
Remove more reserved identifiers (#6925)
2022-12-30 00:58:16 +03:00
checkInvalidWindowError();
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
return TEST_COMPLETED;
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
}
/* Helper function that checks for an 'Invalid parameter' error */
More fix warnings about static function and prototype
2023-03-08 18:12:45 +03:00
static void checkInvalidParameterError(void)
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
{
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
const char *invalidParameterError = "Parameter";
Fix warning drop const qualifier
2023-03-08 23:18:02 +03:00
const char *lastError;
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
Fix warning drop const qualifier
2023-03-08 23:18:02 +03:00
lastError = SDL_GetError();
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
SDLTest_AssertPass("SDL_GetError()");
SDLTest_AssertCheck(lastError != NULL, "Verify error message is not NULL");
if (lastError != NULL) {
SDLTest_AssertCheck(SDL_strncmp(lastError, invalidParameterError, SDL_strlen(invalidParameterError)) == 0,
"SDL_GetError(): expected message starts with '%s', was message: '%s'",
invalidParameterError,
lastError);
SDL_ClearError();
SDLTest_AssertPass("Call to SDL_ClearError()");
}
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
}
/**
docs: Remove Doxygen `\brief` tags. Doxygen and the wiki bridge don't need them; they'll both just use the first line/sentence instead. Fixes #8446.
2023-11-06 18:26:06 +03:00
* Tests call to SDL_GetWindowSize and SDL_SetWindowSize
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
*
test: don't use wiki urls for documentation comments Also make consistent use of \ as documentation escape character.
2023-02-02 02:21:53 +03:00
* \sa SDL_GetWindowSize
* \sa SDL_SetWindowSize
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
*/
SDL_test: use SDLCALL calling convention This is needed when using a pre-built static SDL3_test library.
2024-09-06 04:21:13 +03:00
static int SDLCALL video_getSetWindowSize(void *arg)
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
{
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
const char *title = "video_getSetWindowSize Test Window";
SDL_Window *window;
int result;
SDL_Rect display;
int maxwVariation, maxhVariation;
int wVariation, hVariation;
int referenceW, referenceH;
int currentW, currentH;
int desiredW, desiredH;
Removed SDL_bool in favor of plain bool We require stdbool.h in the build environment, so we might as well use the plain bool type. If your environment doesn't have stdbool.h, this simple replacement will suffice: typedef signed char bool;
2024-09-18 17:52:28 +03:00
const bool restoreHint = SDL_GetHintBoolean("SDL_BORDERLESS_RESIZABLE_STYLE", true);
video: Implement asynchronous windowing SDL window size, state, and position functions have been considered immediate, with their effects assuming to have taken effect upon successful return of the function. However, several windowing systems handle these requests asynchronously, resulting in the functions blocking until the changes have taken effect, potentially for long periods of time. Additionally, some windowing systems treat these as requests, and can potentially deny or fulfill the request in a manner differently than the application expects, such as not allowing a window to be positioned or sized beyond desktop borders, prohibiting fullscreen, and so on. With these changes, applications can make requests of the window manager that do not block, with the understanding that an associated event will be sent if the request is fulfilled. Currently, size, position, maximize, minimize, and fullscreen calls are handled as asynchronous requests, with events being returned if the request is honored. If the application requires that the change take effect immediately, it can call the new SDL_SyncWindow function, which will attempt to block until the request is fulfilled, or some arbitrary timeout period elapses, the duration of which depends not only on the windowing system, but on the operation requested as well (e.g. a 100ms timeout is fine for most X11 events, but maximizing a window can take considerably longer for some reason). There is also a new hint 'SDL_VIDEO_SYNC_ALL_WINDOW_OPS' that will mimic the old behavior by synchronizing after every window operation with, again, the understanding that using this may result in the associated calls blocking for a relatively long period. The deferred model also results in the window size and position getters not reporting false coordinates anymore, as they only forward what the window manager reports vs allowing applications to set arbitrary values, and fullscreen enter/leave events that were initiated via the window manager update the window state appropriately, where they didn't before. Care was taken to ensure that order of operations is maintained, and that requests are not ignored or dropped. This does require some implicit internal synchronization in the various backends if many requests are made in a short period, as some state and behavior depends on other bits of state that need to be known at that particular point in time, but this isn't something that typical applications will hit, unless they are sending a lot of window state in a short time as the tests do. The automated tests developed to test the previous behavior also resulted in previously undefined behavior being defined and normalized across platforms, particularly when it comes to the sizing and positioning of windows when they are in a fixed-size state, such as maximized or fullscreen. Size and position requests made when the window is not in a movable or resizable state will be deferred until it can be applied, so no requests are lost. These changes fix another long-standing issue with renderers recreating maximized windows, where the original non-maximized size was lost, resulting in the window being restored to the wrong size. All automated video tests pass across all platforms. Overall, the "make a request/get an event" model better reflects how most windowing systems work, and some backends avoid spending significant time blocking while waiting for operations to complete.
2023-10-25 22:09:55 +03:00
/* Win32 borderless windows are not resizable by default and need this undocumented hint */
SDL_SetHint("SDL_BORDERLESS_RESIZABLE_STYLE", "1");
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
/* Get display bounds for size range */
video: Implement asynchronous windowing SDL window size, state, and position functions have been considered immediate, with their effects assuming to have taken effect upon successful return of the function. However, several windowing systems handle these requests asynchronously, resulting in the functions blocking until the changes have taken effect, potentially for long periods of time. Additionally, some windowing systems treat these as requests, and can potentially deny or fulfill the request in a manner differently than the application expects, such as not allowing a window to be positioned or sized beyond desktop borders, prohibiting fullscreen, and so on. With these changes, applications can make requests of the window manager that do not block, with the understanding that an associated event will be sent if the request is fulfilled. Currently, size, position, maximize, minimize, and fullscreen calls are handled as asynchronous requests, with events being returned if the request is honored. If the application requires that the change take effect immediately, it can call the new SDL_SyncWindow function, which will attempt to block until the request is fulfilled, or some arbitrary timeout period elapses, the duration of which depends not only on the windowing system, but on the operation requested as well (e.g. a 100ms timeout is fine for most X11 events, but maximizing a window can take considerably longer for some reason). There is also a new hint 'SDL_VIDEO_SYNC_ALL_WINDOW_OPS' that will mimic the old behavior by synchronizing after every window operation with, again, the understanding that using this may result in the associated calls blocking for a relatively long period. The deferred model also results in the window size and position getters not reporting false coordinates anymore, as they only forward what the window manager reports vs allowing applications to set arbitrary values, and fullscreen enter/leave events that were initiated via the window manager update the window state appropriately, where they didn't before. Care was taken to ensure that order of operations is maintained, and that requests are not ignored or dropped. This does require some implicit internal synchronization in the various backends if many requests are made in a short period, as some state and behavior depends on other bits of state that need to be known at that particular point in time, but this isn't something that typical applications will hit, unless they are sending a lot of window state in a short time as the tests do. The automated tests developed to test the previous behavior also resulted in previously undefined behavior being defined and normalized across platforms, particularly when it comes to the sizing and positioning of windows when they are in a fixed-size state, such as maximized or fullscreen. Size and position requests made when the window is not in a movable or resizable state will be deferred until it can be applied, so no requests are lost. These changes fix another long-standing issue with renderers recreating maximized windows, where the original non-maximized size was lost, resulting in the window being restored to the wrong size. All automated video tests pass across all platforms. Overall, the "make a request/get an event" model better reflects how most windowing systems work, and some backends avoid spending significant time blocking while waiting for operations to complete.
2023-10-25 22:09:55 +03:00
result = SDL_GetDisplayUsableBounds(SDL_GetPrimaryDisplay(), &display);
Use SDL_bool instead an int return code in the SDL API Most SDL functions used to indicate success or failure using an int return code. These functions have been changed to return SDL_bool. Here is a coccinelle patch to change code that previously compared the return value to 0 and changes it to a boolean test: @ bool_return_type @ identifier func =~ "^(SDL_AddEventWatch|SDL_AddHintCallback|SDL_AddSurfaceAlternateImage|SDL_AddVulkanRenderSemaphores|SDL_BindAudioStream|SDL_BindAudioStreams|SDL_BlitSurface|SDL_BlitSurface9Grid|SDL_BlitSurfaceScaled|SDL_BlitSurfaceTiled|SDL_BlitSurfaceTiledWithScale|SDL_BlitSurfaceUnchecked|SDL_BlitSurfaceUncheckedScaled|SDL_CaptureMouse|SDL_ClearAudioStream|SDL_ClearClipboardData|SDL_ClearComposition|SDL_ClearError|SDL_ClearProperty|SDL_ClearSurface|SDL_CloseIO|SDL_CloseStorage|SDL_ConvertAudioSamples|SDL_ConvertEventToRenderCoordinates|SDL_ConvertPixels|SDL_ConvertPixelsAndColorspace|SDL_CopyFile|SDL_CopyProperties|SDL_CopyStorageFile|SDL_CreateDirectory|SDL_CreateStorageDirectory|SDL_CreateWindowAndRenderer|SDL_DateTimeToTime|SDL_DestroyWindowSurface|SDL_DetachVirtualJoystick|SDL_DisableScreenSaver|SDL_EnableScreenSaver|SDL_EnumerateDirectory|SDL_EnumerateProperties|SDL_EnumerateStorageDirectory|SDL_FillSurfaceRect|SDL_FillSurfaceRects|SDL_FlashWindow|SDL_FlipSurface|SDL_FlushAudioStream|SDL_FlushRenderer|SDL_GL_DestroyContext|SDL_GL_GetAttribute|SDL_GL_GetSwapInterval|SDL_GL_LoadLibrary|SDL_GL_MakeCurrent|SDL_GL_SetAttribute|SDL_GL_SetSwapInterval|SDL_GL_SwapWindow|SDL_GetAudioDeviceFormat|SDL_GetAudioStreamFormat|SDL_GetCameraFormat|SDL_GetClosestFullscreenDisplayMode|SDL_GetCurrentRenderOutputSize|SDL_GetCurrentTime|SDL_GetDXGIOutputInfo|SDL_GetDateTimeLocalePreferences|SDL_GetDisplayBounds|SDL_GetDisplayUsableBounds|SDL_GetGDKDefaultUser|SDL_GetGDKTaskQueue|SDL_GetGamepadSensorData|SDL_GetGamepadTouchpadFinger|SDL_GetHapticEffectStatus|SDL_GetJoystickBall|SDL_GetMasksForPixelFormat|SDL_GetPathInfo|SDL_GetRectUnion|SDL_GetRectUnionFloat|SDL_GetRenderClipRect|SDL_GetRenderColorScale|SDL_GetRenderDrawBlendMode|SDL_GetRenderDrawColor|SDL_GetRenderDrawColorFloat|SDL_GetRenderLogicalPresentation|SDL_GetRenderLogicalPresentationRect|SDL_GetRenderOutputSize|SDL_GetRenderSafeArea|SDL_GetRenderScale|SDL_GetRenderVSync|SDL_GetRenderViewport|SDL_GetSensorData|SDL_GetStorageFileSize|SDL_GetStoragePathInfo|SDL_GetSurfaceAlphaMod|SDL_GetSurfaceBlendMode|SDL_GetSurfaceClipRect|SDL_GetSurfaceColorKey|SDL_GetSurfaceColorMod|SDL_GetTextInputArea|SDL_GetTextureAlphaMod|SDL_GetTextureAlphaModFloat|SDL_GetTextureBlendMode|SDL_GetTextureColorMod|SDL_GetTextureColorModFloat|SDL_GetTextureScaleMode|SDL_GetTextureSize|SDL_GetWindowAspectRatio|SDL_GetWindowBordersSize|SDL_GetWindowMaximumSize|SDL_GetWindowMinimumSize|SDL_GetWindowPosition|SDL_GetWindowRelativeMouseMode|SDL_GetWindowSafeArea|SDL_GetWindowSize|SDL_GetWindowSizeInPixels|SDL_GetWindowSurfaceVSync|SDL_HideCursor|SDL_HideWindow|SDL_Init|SDL_InitHapticRumble|SDL_InitSubSystem|SDL_LoadWAV|SDL_LoadWAV_IO|SDL_LockAudioStream|SDL_LockProperties|SDL_LockSurface|SDL_LockTexture|SDL_LockTextureToSurface|SDL_MaximizeWindow|SDL_MinimizeWindow|SDL_MixAudio|SDL_OpenURL|SDL_OutOfMemory|SDL_PauseAudioDevice|SDL_PauseAudioStreamDevice|SDL_PauseHaptic|SDL_PlayHapticRumble|SDL_PremultiplyAlpha|SDL_PremultiplySurfaceAlpha|SDL_PushEvent|SDL_PutAudioStreamData|SDL_RaiseWindow|SDL_ReadStorageFile|SDL_ReadSurfacePixel|SDL_ReadSurfacePixelFloat|SDL_RegisterApp|SDL_ReloadGamepadMappings|SDL_RemovePath|SDL_RemoveStoragePath|SDL_RemoveTimer|SDL_RenamePath|SDL_RenameStoragePath|SDL_RenderClear|SDL_RenderCoordinatesFromWindow|SDL_RenderCoordinatesToWindow|SDL_RenderFillRect|SDL_RenderFillRects|SDL_RenderGeometry|SDL_RenderGeometryRaw|SDL_RenderLine|SDL_RenderLines|SDL_RenderPoint|SDL_RenderPoints|SDL_RenderPresent|SDL_RenderRect|SDL_RenderRects|SDL_RenderTexture|SDL_RenderTexture9Grid|SDL_RenderTextureRotated|SDL_RenderTextureTiled|SDL_RequestAndroidPermission|SDL_RestoreWindow|SDL_ResumeAudioDevice|SDL_ResumeAudioStreamDevice|SDL_ResumeHaptic|SDL_RumbleGamepad|SDL_RumbleGamepadTriggers|SDL_RumbleJoystick|SDL_RumbleJoystickTriggers|SDL_RunHapticEffect|SDL_SaveBMP|SDL_SaveBMP_IO|SDL_SendAndroidMessage|SDL_SendGamepadEffect|SDL_SendJoystickEffect|SDL_SendJoystickVirtualSensorData|SDL_SetAppMetadata|SDL_SetAppMetadataProperty|SDL_SetAudioDeviceGain|SDL_SetAudioPostmixCallback|SDL_SetAudioStreamFormat|SDL_SetAudioStreamFrequencyRatio|SDL_SetAudioStreamGain|SDL_SetAudioStreamGetCallback|SDL_SetAudioStreamInputChannelMap|SDL_SetAudioStreamOutputChannelMap|SDL_SetAudioStreamPutCallback|SDL_SetBooleanProperty|SDL_SetClipboardData|SDL_SetClipboardText|SDL_SetCursor|SDL_SetFloatProperty|SDL_SetGamepadLED|SDL_SetGamepadMapping|SDL_SetGamepadPlayerIndex|SDL_SetGamepadSensorEnabled|SDL_SetHapticAutocenter|SDL_SetHapticGain|SDL_SetJoystickLED|SDL_SetJoystickPlayerIndex|SDL_SetJoystickVirtualAxis|SDL_SetJoystickVirtualBall|SDL_SetJoystickVirtualButton|SDL_SetJoystickVirtualHat|SDL_SetJoystickVirtualTouchpad|SDL_SetLinuxThreadPriority|SDL_SetLinuxThreadPriorityAndPolicy|SDL_SetLogPriorityPrefix|SDL_SetMemoryFunctions|SDL_SetNumberProperty|SDL_SetPaletteColors|SDL_SetPointerProperty|SDL_SetPointerPropertyWithCleanup|SDL_SetPrimarySelectionText|SDL_SetRenderClipRect|SDL_SetRenderColorScale|SDL_SetRenderDrawBlendMode|SDL_SetRenderDrawColor|SDL_SetRenderDrawColorFloat|SDL_SetRenderLogicalPresentation|SDL_SetRenderScale|SDL_SetRenderTarget|SDL_SetRenderVSync|SDL_SetRenderViewport|SDL_SetScancodeName|SDL_SetStringProperty|SDL_SetSurfaceAlphaMod|SDL_SetSurfaceBlendMode|SDL_SetSurfaceColorKey|SDL_SetSurfaceColorMod|SDL_SetSurfaceColorspace|SDL_SetSurfacePalette|SDL_SetSurfaceRLE|SDL_SetTLS|SDL_SetTextInputArea|SDL_SetTextureAlphaMod|SDL_SetTextureAlphaModFloat|SDL_SetTextureBlendMode|SDL_SetTextureColorMod|SDL_SetTextureColorModFloat|SDL_SetTextureScaleMode|SDL_SetThreadPriority|SDL_SetWindowAlwaysOnTop|SDL_SetWindowAspectRatio|SDL_SetWindowBordered|SDL_SetWindowFocusable|SDL_SetWindowFullscreen|SDL_SetWindowFullscreenMode|SDL_SetWindowHitTest|SDL_SetWindowIcon|SDL_SetWindowKeyboardGrab|SDL_SetWindowMaximumSize|SDL_SetWindowMinimumSize|SDL_SetWindowModalFor|SDL_SetWindowMouseGrab|SDL_SetWindowMouseRect|SDL_SetWindowOpacity|SDL_SetWindowPosition|SDL_SetWindowRelativeMouseMode|SDL_SetWindowResizable|SDL_SetWindowShape|SDL_SetWindowSize|SDL_SetWindowSurfaceVSync|SDL_SetWindowTitle|SDL_SetiOSAnimationCallback|SDL_ShowAndroidToast|SDL_ShowCursor|SDL_ShowMessageBox|SDL_ShowSimpleMessageBox|SDL_ShowWindow|SDL_ShowWindowSystemMenu|SDL_StartTextInput|SDL_StartTextInputWithProperties|SDL_StopHapticEffect|SDL_StopHapticEffects|SDL_StopHapticRumble|SDL_StopTextInput|SDL_SyncWindow|SDL_TimeToDateTime|SDL_TryLockMutex|SDL_TryLockRWLockForReading|SDL_TryLockRWLockForWriting|SDL_TryWaitSemaphore|SDL_UnlockAudioStream|SDL_UpdateHapticEffect|SDL_UpdateNVTexture|SDL_UpdateTexture|SDL_UpdateWindowSurface|SDL_UpdateWindowSurfaceRects|SDL_UpdateYUVTexture|SDL_Vulkan_CreateSurface|SDL_Vulkan_LoadLibrary|SDL_WaitConditionTimeout|SDL_WaitSemaphoreTimeout|SDL_WarpMouseGlobal|SDL_WriteStorageFile|SDL_WriteSurfacePixel|SDL_WriteSurfacePixelFloat)$"; @@ ( func( ... ) - == 0 | - func( + !func( ... ) - < 0 | - func( + !func( ... ) - != 0 | - func( + !func( ... ) - == -1 )
2024-08-23 03:33:49 +03:00
SDLTest_AssertPass("SDL_GetDisplayUsableBounds()");
Removed SDL_bool in favor of plain bool We require stdbool.h in the build environment, so we might as well use the plain bool type. If your environment doesn't have stdbool.h, this simple replacement will suffice: typedef signed char bool;
2024-09-18 17:52:28 +03:00
SDLTest_AssertCheck(result == true, "Verify return value; expected: true, got: %d", result);
Use SDL_bool instead an int return code in the SDL API Most SDL functions used to indicate success or failure using an int return code. These functions have been changed to return SDL_bool. Here is a coccinelle patch to change code that previously compared the return value to 0 and changes it to a boolean test: @ bool_return_type @ identifier func =~ "^(SDL_AddEventWatch|SDL_AddHintCallback|SDL_AddSurfaceAlternateImage|SDL_AddVulkanRenderSemaphores|SDL_BindAudioStream|SDL_BindAudioStreams|SDL_BlitSurface|SDL_BlitSurface9Grid|SDL_BlitSurfaceScaled|SDL_BlitSurfaceTiled|SDL_BlitSurfaceTiledWithScale|SDL_BlitSurfaceUnchecked|SDL_BlitSurfaceUncheckedScaled|SDL_CaptureMouse|SDL_ClearAudioStream|SDL_ClearClipboardData|SDL_ClearComposition|SDL_ClearError|SDL_ClearProperty|SDL_ClearSurface|SDL_CloseIO|SDL_CloseStorage|SDL_ConvertAudioSamples|SDL_ConvertEventToRenderCoordinates|SDL_ConvertPixels|SDL_ConvertPixelsAndColorspace|SDL_CopyFile|SDL_CopyProperties|SDL_CopyStorageFile|SDL_CreateDirectory|SDL_CreateStorageDirectory|SDL_CreateWindowAndRenderer|SDL_DateTimeToTime|SDL_DestroyWindowSurface|SDL_DetachVirtualJoystick|SDL_DisableScreenSaver|SDL_EnableScreenSaver|SDL_EnumerateDirectory|SDL_EnumerateProperties|SDL_EnumerateStorageDirectory|SDL_FillSurfaceRect|SDL_FillSurfaceRects|SDL_FlashWindow|SDL_FlipSurface|SDL_FlushAudioStream|SDL_FlushRenderer|SDL_GL_DestroyContext|SDL_GL_GetAttribute|SDL_GL_GetSwapInterval|SDL_GL_LoadLibrary|SDL_GL_MakeCurrent|SDL_GL_SetAttribute|SDL_GL_SetSwapInterval|SDL_GL_SwapWindow|SDL_GetAudioDeviceFormat|SDL_GetAudioStreamFormat|SDL_GetCameraFormat|SDL_GetClosestFullscreenDisplayMode|SDL_GetCurrentRenderOutputSize|SDL_GetCurrentTime|SDL_GetDXGIOutputInfo|SDL_GetDateTimeLocalePreferences|SDL_GetDisplayBounds|SDL_GetDisplayUsableBounds|SDL_GetGDKDefaultUser|SDL_GetGDKTaskQueue|SDL_GetGamepadSensorData|SDL_GetGamepadTouchpadFinger|SDL_GetHapticEffectStatus|SDL_GetJoystickBall|SDL_GetMasksForPixelFormat|SDL_GetPathInfo|SDL_GetRectUnion|SDL_GetRectUnionFloat|SDL_GetRenderClipRect|SDL_GetRenderColorScale|SDL_GetRenderDrawBlendMode|SDL_GetRenderDrawColor|SDL_GetRenderDrawColorFloat|SDL_GetRenderLogicalPresentation|SDL_GetRenderLogicalPresentationRect|SDL_GetRenderOutputSize|SDL_GetRenderSafeArea|SDL_GetRenderScale|SDL_GetRenderVSync|SDL_GetRenderViewport|SDL_GetSensorData|SDL_GetStorageFileSize|SDL_GetStoragePathInfo|SDL_GetSurfaceAlphaMod|SDL_GetSurfaceBlendMode|SDL_GetSurfaceClipRect|SDL_GetSurfaceColorKey|SDL_GetSurfaceColorMod|SDL_GetTextInputArea|SDL_GetTextureAlphaMod|SDL_GetTextureAlphaModFloat|SDL_GetTextureBlendMode|SDL_GetTextureColorMod|SDL_GetTextureColorModFloat|SDL_GetTextureScaleMode|SDL_GetTextureSize|SDL_GetWindowAspectRatio|SDL_GetWindowBordersSize|SDL_GetWindowMaximumSize|SDL_GetWindowMinimumSize|SDL_GetWindowPosition|SDL_GetWindowRelativeMouseMode|SDL_GetWindowSafeArea|SDL_GetWindowSize|SDL_GetWindowSizeInPixels|SDL_GetWindowSurfaceVSync|SDL_HideCursor|SDL_HideWindow|SDL_Init|SDL_InitHapticRumble|SDL_InitSubSystem|SDL_LoadWAV|SDL_LoadWAV_IO|SDL_LockAudioStream|SDL_LockProperties|SDL_LockSurface|SDL_LockTexture|SDL_LockTextureToSurface|SDL_MaximizeWindow|SDL_MinimizeWindow|SDL_MixAudio|SDL_OpenURL|SDL_OutOfMemory|SDL_PauseAudioDevice|SDL_PauseAudioStreamDevice|SDL_PauseHaptic|SDL_PlayHapticRumble|SDL_PremultiplyAlpha|SDL_PremultiplySurfaceAlpha|SDL_PushEvent|SDL_PutAudioStreamData|SDL_RaiseWindow|SDL_ReadStorageFile|SDL_ReadSurfacePixel|SDL_ReadSurfacePixelFloat|SDL_RegisterApp|SDL_ReloadGamepadMappings|SDL_RemovePath|SDL_RemoveStoragePath|SDL_RemoveTimer|SDL_RenamePath|SDL_RenameStoragePath|SDL_RenderClear|SDL_RenderCoordinatesFromWindow|SDL_RenderCoordinatesToWindow|SDL_RenderFillRect|SDL_RenderFillRects|SDL_RenderGeometry|SDL_RenderGeometryRaw|SDL_RenderLine|SDL_RenderLines|SDL_RenderPoint|SDL_RenderPoints|SDL_RenderPresent|SDL_RenderRect|SDL_RenderRects|SDL_RenderTexture|SDL_RenderTexture9Grid|SDL_RenderTextureRotated|SDL_RenderTextureTiled|SDL_RequestAndroidPermission|SDL_RestoreWindow|SDL_ResumeAudioDevice|SDL_ResumeAudioStreamDevice|SDL_ResumeHaptic|SDL_RumbleGamepad|SDL_RumbleGamepadTriggers|SDL_RumbleJoystick|SDL_RumbleJoystickTriggers|SDL_RunHapticEffect|SDL_SaveBMP|SDL_SaveBMP_IO|SDL_SendAndroidMessage|SDL_SendGamepadEffect|SDL_SendJoystickEffect|SDL_SendJoystickVirtualSensorData|SDL_SetAppMetadata|SDL_SetAppMetadataProperty|SDL_SetAudioDeviceGain|SDL_SetAudioPostmixCallback|SDL_SetAudioStreamFormat|SDL_SetAudioStreamFrequencyRatio|SDL_SetAudioStreamGain|SDL_SetAudioStreamGetCallback|SDL_SetAudioStreamInputChannelMap|SDL_SetAudioStreamOutputChannelMap|SDL_SetAudioStreamPutCallback|SDL_SetBooleanProperty|SDL_SetClipboardData|SDL_SetClipboardText|SDL_SetCursor|SDL_SetFloatProperty|SDL_SetGamepadLED|SDL_SetGamepadMapping|SDL_SetGamepadPlayerIndex|SDL_SetGamepadSensorEnabled|SDL_SetHapticAutocenter|SDL_SetHapticGain|SDL_SetJoystickLED|SDL_SetJoystickPlayerIndex|SDL_SetJoystickVirtualAxis|SDL_SetJoystickVirtualBall|SDL_SetJoystickVirtualButton|SDL_SetJoystickVirtualHat|SDL_SetJoystickVirtualTouchpad|SDL_SetLinuxThreadPriority|SDL_SetLinuxThreadPriorityAndPolicy|SDL_SetLogPriorityPrefix|SDL_SetMemoryFunctions|SDL_SetNumberProperty|SDL_SetPaletteColors|SDL_SetPointerProperty|SDL_SetPointerPropertyWithCleanup|SDL_SetPrimarySelectionText|SDL_SetRenderClipRect|SDL_SetRenderColorScale|SDL_SetRenderDrawBlendMode|SDL_SetRenderDrawColor|SDL_SetRenderDrawColorFloat|SDL_SetRenderLogicalPresentation|SDL_SetRenderScale|SDL_SetRenderTarget|SDL_SetRenderVSync|SDL_SetRenderViewport|SDL_SetScancodeName|SDL_SetStringProperty|SDL_SetSurfaceAlphaMod|SDL_SetSurfaceBlendMode|SDL_SetSurfaceColorKey|SDL_SetSurfaceColorMod|SDL_SetSurfaceColorspace|SDL_SetSurfacePalette|SDL_SetSurfaceRLE|SDL_SetTLS|SDL_SetTextInputArea|SDL_SetTextureAlphaMod|SDL_SetTextureAlphaModFloat|SDL_SetTextureBlendMode|SDL_SetTextureColorMod|SDL_SetTextureColorModFloat|SDL_SetTextureScaleMode|SDL_SetThreadPriority|SDL_SetWindowAlwaysOnTop|SDL_SetWindowAspectRatio|SDL_SetWindowBordered|SDL_SetWindowFocusable|SDL_SetWindowFullscreen|SDL_SetWindowFullscreenMode|SDL_SetWindowHitTest|SDL_SetWindowIcon|SDL_SetWindowKeyboardGrab|SDL_SetWindowMaximumSize|SDL_SetWindowMinimumSize|SDL_SetWindowModalFor|SDL_SetWindowMouseGrab|SDL_SetWindowMouseRect|SDL_SetWindowOpacity|SDL_SetWindowPosition|SDL_SetWindowRelativeMouseMode|SDL_SetWindowResizable|SDL_SetWindowShape|SDL_SetWindowSize|SDL_SetWindowSurfaceVSync|SDL_SetWindowTitle|SDL_SetiOSAnimationCallback|SDL_ShowAndroidToast|SDL_ShowCursor|SDL_ShowMessageBox|SDL_ShowSimpleMessageBox|SDL_ShowWindow|SDL_ShowWindowSystemMenu|SDL_StartTextInput|SDL_StartTextInputWithProperties|SDL_StopHapticEffect|SDL_StopHapticEffects|SDL_StopHapticRumble|SDL_StopTextInput|SDL_SyncWindow|SDL_TimeToDateTime|SDL_TryLockMutex|SDL_TryLockRWLockForReading|SDL_TryLockRWLockForWriting|SDL_TryWaitSemaphore|SDL_UnlockAudioStream|SDL_UpdateHapticEffect|SDL_UpdateNVTexture|SDL_UpdateTexture|SDL_UpdateWindowSurface|SDL_UpdateWindowSurfaceRects|SDL_UpdateYUVTexture|SDL_Vulkan_CreateSurface|SDL_Vulkan_LoadLibrary|SDL_WaitConditionTimeout|SDL_WaitSemaphoreTimeout|SDL_WarpMouseGlobal|SDL_WriteStorageFile|SDL_WriteSurfacePixel|SDL_WriteSurfacePixelFloat)$"; @@ ( func( ... ) - == 0 | - func( + !func( ... ) - < 0 | - func( + !func( ... ) - != 0 | - func( + !func( ... ) - == -1 )
2024-08-23 03:33:49 +03:00
if (!result) {
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
return TEST_ABORTED;
}
/* Call against new test window */
Remove more reserved identifiers (#6925)
2022-12-30 00:58:16 +03:00
window = createVideoSuiteTestWindow(title);
Pointer as bool (libsdl-org#7214)
2023-11-10 00:29:15 +03:00
if (!window) {
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
return TEST_ABORTED;
}
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
video: Implement asynchronous windowing SDL window size, state, and position functions have been considered immediate, with their effects assuming to have taken effect upon successful return of the function. However, several windowing systems handle these requests asynchronously, resulting in the functions blocking until the changes have taken effect, potentially for long periods of time. Additionally, some windowing systems treat these as requests, and can potentially deny or fulfill the request in a manner differently than the application expects, such as not allowing a window to be positioned or sized beyond desktop borders, prohibiting fullscreen, and so on. With these changes, applications can make requests of the window manager that do not block, with the understanding that an associated event will be sent if the request is fulfilled. Currently, size, position, maximize, minimize, and fullscreen calls are handled as asynchronous requests, with events being returned if the request is honored. If the application requires that the change take effect immediately, it can call the new SDL_SyncWindow function, which will attempt to block until the request is fulfilled, or some arbitrary timeout period elapses, the duration of which depends not only on the windowing system, but on the operation requested as well (e.g. a 100ms timeout is fine for most X11 events, but maximizing a window can take considerably longer for some reason). There is also a new hint 'SDL_VIDEO_SYNC_ALL_WINDOW_OPS' that will mimic the old behavior by synchronizing after every window operation with, again, the understanding that using this may result in the associated calls blocking for a relatively long period. The deferred model also results in the window size and position getters not reporting false coordinates anymore, as they only forward what the window manager reports vs allowing applications to set arbitrary values, and fullscreen enter/leave events that were initiated via the window manager update the window state appropriately, where they didn't before. Care was taken to ensure that order of operations is maintained, and that requests are not ignored or dropped. This does require some implicit internal synchronization in the various backends if many requests are made in a short period, as some state and behavior depends on other bits of state that need to be known at that particular point in time, but this isn't something that typical applications will hit, unless they are sending a lot of window state in a short time as the tests do. The automated tests developed to test the previous behavior also resulted in previously undefined behavior being defined and normalized across platforms, particularly when it comes to the sizing and positioning of windows when they are in a fixed-size state, such as maximized or fullscreen. Size and position requests made when the window is not in a movable or resizable state will be deferred until it can be applied, so no requests are lost. These changes fix another long-standing issue with renderers recreating maximized windows, where the original non-maximized size was lost, resulting in the window being restored to the wrong size. All automated video tests pass across all platforms. Overall, the "make a request/get an event" model better reflects how most windowing systems work, and some backends avoid spending significant time blocking while waiting for operations to complete.
2023-10-25 22:09:55 +03:00
SDL_GetWindowSize(window, &currentW, &currentH);
if (SDL_SetWindowSize(window, currentW, currentH)) {
SDLTest_Log("Skipping window resize tests: %s reports window resizing as unsupported", SDL_GetCurrentVideoDriver());
goto null_tests;
}
if (SDL_strcmp(SDL_GetCurrentVideoDriver(), "windows") == 0 ||
SDL_strcmp(SDL_GetCurrentVideoDriver(), "x11") == 0) {
/* Platform clips window size to screen size */
maxwVariation = 4;
maxhVariation = 4;
} else {
/* Platform allows window size >= screen size */
maxwVariation = 5;
maxhVariation = 5;
}
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
for (wVariation = 0; wVariation < maxwVariation; wVariation++) {
for (hVariation = 0; hVariation < maxhVariation; hVariation++) {
switch (wVariation) {
default:
case 0:
/* 1 Pixel Wide */
desiredW = 1;
break;
case 1:
/* Random width inside screen */
desiredW = SDLTest_RandomIntegerInRange(1, 100);
break;
case 2:
/* Width 1 pixel smaller than screen */
desiredW = display.w - 1;
break;
case 3:
/* Width at screen size */
desiredW = display.w;
break;
case 4:
/* Width 1 pixel larger than screen */
desiredW = display.w + 1;
break;
}
switch (hVariation) {
default:
case 0:
/* 1 Pixel High */
desiredH = 1;
break;
case 1:
/* Random height inside screen */
desiredH = SDLTest_RandomIntegerInRange(1, 100);
break;
case 2:
/* Height 1 pixel smaller than screen */
desiredH = display.h - 1;
break;
case 3:
/* Height at screen size */
desiredH = display.h;
break;
case 4:
/* Height 1 pixel larger than screen */
desiredH = display.h + 1;
break;
}
/* Set size */
SDL_SetWindowSize(window, desiredW, desiredH);
SDLTest_AssertPass("Call to SDL_SetWindowSize(...,%d,%d)", desiredW, desiredH);
video: Implement asynchronous windowing SDL window size, state, and position functions have been considered immediate, with their effects assuming to have taken effect upon successful return of the function. However, several windowing systems handle these requests asynchronously, resulting in the functions blocking until the changes have taken effect, potentially for long periods of time. Additionally, some windowing systems treat these as requests, and can potentially deny or fulfill the request in a manner differently than the application expects, such as not allowing a window to be positioned or sized beyond desktop borders, prohibiting fullscreen, and so on. With these changes, applications can make requests of the window manager that do not block, with the understanding that an associated event will be sent if the request is fulfilled. Currently, size, position, maximize, minimize, and fullscreen calls are handled as asynchronous requests, with events being returned if the request is honored. If the application requires that the change take effect immediately, it can call the new SDL_SyncWindow function, which will attempt to block until the request is fulfilled, or some arbitrary timeout period elapses, the duration of which depends not only on the windowing system, but on the operation requested as well (e.g. a 100ms timeout is fine for most X11 events, but maximizing a window can take considerably longer for some reason). There is also a new hint 'SDL_VIDEO_SYNC_ALL_WINDOW_OPS' that will mimic the old behavior by synchronizing after every window operation with, again, the understanding that using this may result in the associated calls blocking for a relatively long period. The deferred model also results in the window size and position getters not reporting false coordinates anymore, as they only forward what the window manager reports vs allowing applications to set arbitrary values, and fullscreen enter/leave events that were initiated via the window manager update the window state appropriately, where they didn't before. Care was taken to ensure that order of operations is maintained, and that requests are not ignored or dropped. This does require some implicit internal synchronization in the various backends if many requests are made in a short period, as some state and behavior depends on other bits of state that need to be known at that particular point in time, but this isn't something that typical applications will hit, unless they are sending a lot of window state in a short time as the tests do. The automated tests developed to test the previous behavior also resulted in previously undefined behavior being defined and normalized across platforms, particularly when it comes to the sizing and positioning of windows when they are in a fixed-size state, such as maximized or fullscreen. Size and position requests made when the window is not in a movable or resizable state will be deferred until it can be applied, so no requests are lost. These changes fix another long-standing issue with renderers recreating maximized windows, where the original non-maximized size was lost, resulting in the window being restored to the wrong size. All automated video tests pass across all platforms. Overall, the "make a request/get an event" model better reflects how most windowing systems work, and some backends avoid spending significant time blocking while waiting for operations to complete.
2023-10-25 22:09:55 +03:00
result = SDL_SyncWindow(window);
SDLTest_AssertPass("SDL_SyncWindow()");
Removed SDL_bool in favor of plain bool We require stdbool.h in the build environment, so we might as well use the plain bool type. If your environment doesn't have stdbool.h, this simple replacement will suffice: typedef signed char bool;
2024-09-18 17:52:28 +03:00
SDLTest_AssertCheck(result == true, "Verify return value; expected: true, got: %d", result);
video: Implement asynchronous windowing SDL window size, state, and position functions have been considered immediate, with their effects assuming to have taken effect upon successful return of the function. However, several windowing systems handle these requests asynchronously, resulting in the functions blocking until the changes have taken effect, potentially for long periods of time. Additionally, some windowing systems treat these as requests, and can potentially deny or fulfill the request in a manner differently than the application expects, such as not allowing a window to be positioned or sized beyond desktop borders, prohibiting fullscreen, and so on. With these changes, applications can make requests of the window manager that do not block, with the understanding that an associated event will be sent if the request is fulfilled. Currently, size, position, maximize, minimize, and fullscreen calls are handled as asynchronous requests, with events being returned if the request is honored. If the application requires that the change take effect immediately, it can call the new SDL_SyncWindow function, which will attempt to block until the request is fulfilled, or some arbitrary timeout period elapses, the duration of which depends not only on the windowing system, but on the operation requested as well (e.g. a 100ms timeout is fine for most X11 events, but maximizing a window can take considerably longer for some reason). There is also a new hint 'SDL_VIDEO_SYNC_ALL_WINDOW_OPS' that will mimic the old behavior by synchronizing after every window operation with, again, the understanding that using this may result in the associated calls blocking for a relatively long period. The deferred model also results in the window size and position getters not reporting false coordinates anymore, as they only forward what the window manager reports vs allowing applications to set arbitrary values, and fullscreen enter/leave events that were initiated via the window manager update the window state appropriately, where they didn't before. Care was taken to ensure that order of operations is maintained, and that requests are not ignored or dropped. This does require some implicit internal synchronization in the various backends if many requests are made in a short period, as some state and behavior depends on other bits of state that need to be known at that particular point in time, but this isn't something that typical applications will hit, unless they are sending a lot of window state in a short time as the tests do. The automated tests developed to test the previous behavior also resulted in previously undefined behavior being defined and normalized across platforms, particularly when it comes to the sizing and positioning of windows when they are in a fixed-size state, such as maximized or fullscreen. Size and position requests made when the window is not in a movable or resizable state will be deferred until it can be applied, so no requests are lost. These changes fix another long-standing issue with renderers recreating maximized windows, where the original non-maximized size was lost, resulting in the window being restored to the wrong size. All automated video tests pass across all platforms. Overall, the "make a request/get an event" model better reflects how most windowing systems work, and some backends avoid spending significant time blocking while waiting for operations to complete.
2023-10-25 22:09:55 +03:00
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
/* Get size */
currentW = desiredW + 1;
currentH = desiredH + 1;
SDL_GetWindowSize(window, &currentW, &currentH);
SDLTest_AssertPass("Call to SDL_GetWindowSize()");
testautomation_video: if SDL_SetWindowSize/Position isn't honored, we should check there is an event x11: send the events if various occasions
2023-03-20 13:25:10 +03:00
if (desiredW == currentW && desiredH == currentH) {
SDLTest_AssertCheck(desiredW == currentW, "Verify returned width; expected: %d, got: %d", desiredW, currentW);
SDLTest_AssertCheck(desiredH == currentH, "Verify returned height; expected: %d, got: %d", desiredH, currentH);
} else {
Removed SDL_bool in favor of plain bool We require stdbool.h in the build environment, so we might as well use the plain bool type. If your environment doesn't have stdbool.h, this simple replacement will suffice: typedef signed char bool;
2024-09-18 17:52:28 +03:00
bool hasEvent;
testautomation_video: if SDL_SetWindowSize/Position isn't honored, we should check there is an event x11: send the events if various occasions
2023-03-20 13:25:10 +03:00
/* SDL_SetWindowPosition() and SDL_SetWindowSize() will make requests of the window manager and set the internal position and size,
* and then we get events signaling what actually happened, and they get passed on to the application if they're not what we expect. */
video: Implement asynchronous windowing SDL window size, state, and position functions have been considered immediate, with their effects assuming to have taken effect upon successful return of the function. However, several windowing systems handle these requests asynchronously, resulting in the functions blocking until the changes have taken effect, potentially for long periods of time. Additionally, some windowing systems treat these as requests, and can potentially deny or fulfill the request in a manner differently than the application expects, such as not allowing a window to be positioned or sized beyond desktop borders, prohibiting fullscreen, and so on. With these changes, applications can make requests of the window manager that do not block, with the understanding that an associated event will be sent if the request is fulfilled. Currently, size, position, maximize, minimize, and fullscreen calls are handled as asynchronous requests, with events being returned if the request is honored. If the application requires that the change take effect immediately, it can call the new SDL_SyncWindow function, which will attempt to block until the request is fulfilled, or some arbitrary timeout period elapses, the duration of which depends not only on the windowing system, but on the operation requested as well (e.g. a 100ms timeout is fine for most X11 events, but maximizing a window can take considerably longer for some reason). There is also a new hint 'SDL_VIDEO_SYNC_ALL_WINDOW_OPS' that will mimic the old behavior by synchronizing after every window operation with, again, the understanding that using this may result in the associated calls blocking for a relatively long period. The deferred model also results in the window size and position getters not reporting false coordinates anymore, as they only forward what the window manager reports vs allowing applications to set arbitrary values, and fullscreen enter/leave events that were initiated via the window manager update the window state appropriately, where they didn't before. Care was taken to ensure that order of operations is maintained, and that requests are not ignored or dropped. This does require some implicit internal synchronization in the various backends if many requests are made in a short period, as some state and behavior depends on other bits of state that need to be known at that particular point in time, but this isn't something that typical applications will hit, unless they are sending a lot of window state in a short time as the tests do. The automated tests developed to test the previous behavior also resulted in previously undefined behavior being defined and normalized across platforms, particularly when it comes to the sizing and positioning of windows when they are in a fixed-size state, such as maximized or fullscreen. Size and position requests made when the window is not in a movable or resizable state will be deferred until it can be applied, so no requests are lost. These changes fix another long-standing issue with renderers recreating maximized windows, where the original non-maximized size was lost, resulting in the window being restored to the wrong size. All automated video tests pass across all platforms. Overall, the "make a request/get an event" model better reflects how most windowing systems work, and some backends avoid spending significant time blocking while waiting for operations to complete.
2023-10-25 22:09:55 +03:00
currentW = desiredW + 1;
currentH = desiredH + 1;
hasEvent = getSizeFromEvent(&currentW, &currentH);
Removed SDL_bool in favor of plain bool We require stdbool.h in the build environment, so we might as well use the plain bool type. If your environment doesn't have stdbool.h, this simple replacement will suffice: typedef signed char bool;
2024-09-18 17:52:28 +03:00
SDLTest_AssertCheck(hasEvent == true, "Changing size was not honored by WM, checking presence of SDL_EVENT_WINDOW_RESIZED");
testautomation_video: if SDL_SetWindowSize/Position isn't honored, we should check there is an event x11: send the events if various occasions
2023-03-20 13:25:10 +03:00
if (hasEvent) {
SDLTest_AssertCheck(desiredW == currentW, "Verify returned width is the one from SDL event; expected: %d, got: %d", desiredW, currentW);
SDLTest_AssertCheck(desiredH == currentH, "Verify returned height is the one from SDL event; expected: %d, got: %d", desiredH, currentH);
}
}
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
/* Get just width */
currentW = desiredW + 1;
SDL_GetWindowSize(window, &currentW, NULL);
SDLTest_AssertPass("Call to SDL_GetWindowSize(&h=NULL)");
SDLTest_AssertCheck(desiredW == currentW, "Verify returned width; expected: %d, got: %d", desiredW, currentW);
/* Get just height */
currentH = desiredH + 1;
SDL_GetWindowSize(window, NULL, &currentH);
SDLTest_AssertPass("Call to SDL_GetWindowSize(&w=NULL)");
SDLTest_AssertCheck(desiredH == currentH, "Verify returned height; expected: %d, got: %d", desiredH, currentH);
}
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
}
video: Implement asynchronous windowing SDL window size, state, and position functions have been considered immediate, with their effects assuming to have taken effect upon successful return of the function. However, several windowing systems handle these requests asynchronously, resulting in the functions blocking until the changes have taken effect, potentially for long periods of time. Additionally, some windowing systems treat these as requests, and can potentially deny or fulfill the request in a manner differently than the application expects, such as not allowing a window to be positioned or sized beyond desktop borders, prohibiting fullscreen, and so on. With these changes, applications can make requests of the window manager that do not block, with the understanding that an associated event will be sent if the request is fulfilled. Currently, size, position, maximize, minimize, and fullscreen calls are handled as asynchronous requests, with events being returned if the request is honored. If the application requires that the change take effect immediately, it can call the new SDL_SyncWindow function, which will attempt to block until the request is fulfilled, or some arbitrary timeout period elapses, the duration of which depends not only on the windowing system, but on the operation requested as well (e.g. a 100ms timeout is fine for most X11 events, but maximizing a window can take considerably longer for some reason). There is also a new hint 'SDL_VIDEO_SYNC_ALL_WINDOW_OPS' that will mimic the old behavior by synchronizing after every window operation with, again, the understanding that using this may result in the associated calls blocking for a relatively long period. The deferred model also results in the window size and position getters not reporting false coordinates anymore, as they only forward what the window manager reports vs allowing applications to set arbitrary values, and fullscreen enter/leave events that were initiated via the window manager update the window state appropriately, where they didn't before. Care was taken to ensure that order of operations is maintained, and that requests are not ignored or dropped. This does require some implicit internal synchronization in the various backends if many requests are made in a short period, as some state and behavior depends on other bits of state that need to be known at that particular point in time, but this isn't something that typical applications will hit, unless they are sending a lot of window state in a short time as the tests do. The automated tests developed to test the previous behavior also resulted in previously undefined behavior being defined and normalized across platforms, particularly when it comes to the sizing and positioning of windows when they are in a fixed-size state, such as maximized or fullscreen. Size and position requests made when the window is not in a movable or resizable state will be deferred until it can be applied, so no requests are lost. These changes fix another long-standing issue with renderers recreating maximized windows, where the original non-maximized size was lost, resulting in the window being restored to the wrong size. All automated video tests pass across all platforms. Overall, the "make a request/get an event" model better reflects how most windowing systems work, and some backends avoid spending significant time blocking while waiting for operations to complete.
2023-10-25 22:09:55 +03:00
null_tests:
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
/* Dummy call with both pointers NULL */
SDL_GetWindowSize(window, NULL, NULL);
SDLTest_AssertPass("Call to SDL_GetWindowSize(&w=NULL,&h=NULL)");
/* Negative tests for parameter input */
SDL_ClearError();
SDLTest_AssertPass("Call to SDL_ClearError()");
for (desiredH = -2; desiredH < 2; desiredH++) {
for (desiredW = -2; desiredW < 2; desiredW++) {
if (desiredW <= 0 || desiredH <= 0) {
SDL_SetWindowSize(window, desiredW, desiredH);
SDLTest_AssertPass("Call to SDL_SetWindowSize(...,%d,%d)", desiredW, desiredH);
Remove more reserved identifiers (#6925)
2022-12-30 00:58:16 +03:00
checkInvalidParameterError();
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
}
}
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
}
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
/* Clean up */
Remove more reserved identifiers (#6925)
2022-12-30 00:58:16 +03:00
destroyVideoSuiteTestWindow(window);
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
/* Set some 'magic' value for later check that nothing was changed */
referenceW = SDLTest_RandomSint32();
referenceH = SDLTest_RandomSint32();
currentW = referenceW;
currentH = referenceH;
desiredW = SDLTest_RandomSint32();
desiredH = SDLTest_RandomSint32();
/* Negative tests for window input */
SDL_ClearError();
SDLTest_AssertPass("Call to SDL_ClearError()");
SDL_GetWindowSize(NULL, &currentW, &currentH);
SDLTest_AssertPass("Call to SDL_GetWindowSize(window=NULL)");
SDLTest_AssertCheck(
currentW == referenceW && currentH == referenceH,
"Verify that content of W and H pointers has not been modified; expected: %d,%d; got: %d,%d",
referenceW, referenceH,
currentW, currentH);
Remove more reserved identifiers (#6925)
2022-12-30 00:58:16 +03:00
checkInvalidWindowError();
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
SDL_GetWindowSize(NULL, NULL, NULL);
SDLTest_AssertPass("Call to SDL_GetWindowSize(NULL, NULL, NULL)");
Remove more reserved identifiers (#6925)
2022-12-30 00:58:16 +03:00
checkInvalidWindowError();
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
SDL_SetWindowSize(NULL, desiredW, desiredH);
SDLTest_AssertPass("Call to SDL_SetWindowSize(window=NULL)");
Remove more reserved identifiers (#6925)
2022-12-30 00:58:16 +03:00
checkInvalidWindowError();
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
video: Implement asynchronous windowing SDL window size, state, and position functions have been considered immediate, with their effects assuming to have taken effect upon successful return of the function. However, several windowing systems handle these requests asynchronously, resulting in the functions blocking until the changes have taken effect, potentially for long periods of time. Additionally, some windowing systems treat these as requests, and can potentially deny or fulfill the request in a manner differently than the application expects, such as not allowing a window to be positioned or sized beyond desktop borders, prohibiting fullscreen, and so on. With these changes, applications can make requests of the window manager that do not block, with the understanding that an associated event will be sent if the request is fulfilled. Currently, size, position, maximize, minimize, and fullscreen calls are handled as asynchronous requests, with events being returned if the request is honored. If the application requires that the change take effect immediately, it can call the new SDL_SyncWindow function, which will attempt to block until the request is fulfilled, or some arbitrary timeout period elapses, the duration of which depends not only on the windowing system, but on the operation requested as well (e.g. a 100ms timeout is fine for most X11 events, but maximizing a window can take considerably longer for some reason). There is also a new hint 'SDL_VIDEO_SYNC_ALL_WINDOW_OPS' that will mimic the old behavior by synchronizing after every window operation with, again, the understanding that using this may result in the associated calls blocking for a relatively long period. The deferred model also results in the window size and position getters not reporting false coordinates anymore, as they only forward what the window manager reports vs allowing applications to set arbitrary values, and fullscreen enter/leave events that were initiated via the window manager update the window state appropriately, where they didn't before. Care was taken to ensure that order of operations is maintained, and that requests are not ignored or dropped. This does require some implicit internal synchronization in the various backends if many requests are made in a short period, as some state and behavior depends on other bits of state that need to be known at that particular point in time, but this isn't something that typical applications will hit, unless they are sending a lot of window state in a short time as the tests do. The automated tests developed to test the previous behavior also resulted in previously undefined behavior being defined and normalized across platforms, particularly when it comes to the sizing and positioning of windows when they are in a fixed-size state, such as maximized or fullscreen. Size and position requests made when the window is not in a movable or resizable state will be deferred until it can be applied, so no requests are lost. These changes fix another long-standing issue with renderers recreating maximized windows, where the original non-maximized size was lost, resulting in the window being restored to the wrong size. All automated video tests pass across all platforms. Overall, the "make a request/get an event" model better reflects how most windowing systems work, and some backends avoid spending significant time blocking while waiting for operations to complete.
2023-10-25 22:09:55 +03:00
/* Restore the hint to the previous value */
SDL_SetHint("SDL_BORDERLESS_RESIZABLE_STYLE", restoreHint ? "1" : "0");
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
return TEST_COMPLETED;
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
}
/**
docs: Remove Doxygen `\brief` tags. Doxygen and the wiki bridge don't need them; they'll both just use the first line/sentence instead. Fixes #8446.
2023-11-06 18:26:06 +03:00
* Tests call to SDL_GetWindowMinimumSize and SDL_SetWindowMinimumSize
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
*
*/
SDL_test: use SDLCALL calling convention This is needed when using a pre-built static SDL3_test library.
2024-09-06 04:21:13 +03:00
static int SDLCALL video_getSetWindowMinimumSize(void *arg)
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
{
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
const char *title = "video_getSetWindowMinimumSize Test Window";
SDL_Window *window;
int result;
SDL_Rect display;
int wVariation, hVariation;
int referenceW, referenceH;
int currentW, currentH;
int desiredW = 1;
int desiredH = 1;
/* Get display bounds for size range */
Displays are now referenced by instance ID instead of index
2023-01-30 00:30:55 +03:00
result = SDL_GetDisplayBounds(SDL_GetPrimaryDisplay(), &display);
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
SDLTest_AssertPass("SDL_GetDisplayBounds()");
Removed SDL_bool in favor of plain bool We require stdbool.h in the build environment, so we might as well use the plain bool type. If your environment doesn't have stdbool.h, this simple replacement will suffice: typedef signed char bool;
2024-09-18 17:52:28 +03:00
SDLTest_AssertCheck(result == true, "Verify return value; expected: true, got: %d", result);
Use SDL_bool instead an int return code in the SDL API Most SDL functions used to indicate success or failure using an int return code. These functions have been changed to return SDL_bool. Here is a coccinelle patch to change code that previously compared the return value to 0 and changes it to a boolean test: @ bool_return_type @ identifier func =~ "^(SDL_AddEventWatch|SDL_AddHintCallback|SDL_AddSurfaceAlternateImage|SDL_AddVulkanRenderSemaphores|SDL_BindAudioStream|SDL_BindAudioStreams|SDL_BlitSurface|SDL_BlitSurface9Grid|SDL_BlitSurfaceScaled|SDL_BlitSurfaceTiled|SDL_BlitSurfaceTiledWithScale|SDL_BlitSurfaceUnchecked|SDL_BlitSurfaceUncheckedScaled|SDL_CaptureMouse|SDL_ClearAudioStream|SDL_ClearClipboardData|SDL_ClearComposition|SDL_ClearError|SDL_ClearProperty|SDL_ClearSurface|SDL_CloseIO|SDL_CloseStorage|SDL_ConvertAudioSamples|SDL_ConvertEventToRenderCoordinates|SDL_ConvertPixels|SDL_ConvertPixelsAndColorspace|SDL_CopyFile|SDL_CopyProperties|SDL_CopyStorageFile|SDL_CreateDirectory|SDL_CreateStorageDirectory|SDL_CreateWindowAndRenderer|SDL_DateTimeToTime|SDL_DestroyWindowSurface|SDL_DetachVirtualJoystick|SDL_DisableScreenSaver|SDL_EnableScreenSaver|SDL_EnumerateDirectory|SDL_EnumerateProperties|SDL_EnumerateStorageDirectory|SDL_FillSurfaceRect|SDL_FillSurfaceRects|SDL_FlashWindow|SDL_FlipSurface|SDL_FlushAudioStream|SDL_FlushRenderer|SDL_GL_DestroyContext|SDL_GL_GetAttribute|SDL_GL_GetSwapInterval|SDL_GL_LoadLibrary|SDL_GL_MakeCurrent|SDL_GL_SetAttribute|SDL_GL_SetSwapInterval|SDL_GL_SwapWindow|SDL_GetAudioDeviceFormat|SDL_GetAudioStreamFormat|SDL_GetCameraFormat|SDL_GetClosestFullscreenDisplayMode|SDL_GetCurrentRenderOutputSize|SDL_GetCurrentTime|SDL_GetDXGIOutputInfo|SDL_GetDateTimeLocalePreferences|SDL_GetDisplayBounds|SDL_GetDisplayUsableBounds|SDL_GetGDKDefaultUser|SDL_GetGDKTaskQueue|SDL_GetGamepadSensorData|SDL_GetGamepadTouchpadFinger|SDL_GetHapticEffectStatus|SDL_GetJoystickBall|SDL_GetMasksForPixelFormat|SDL_GetPathInfo|SDL_GetRectUnion|SDL_GetRectUnionFloat|SDL_GetRenderClipRect|SDL_GetRenderColorScale|SDL_GetRenderDrawBlendMode|SDL_GetRenderDrawColor|SDL_GetRenderDrawColorFloat|SDL_GetRenderLogicalPresentation|SDL_GetRenderLogicalPresentationRect|SDL_GetRenderOutputSize|SDL_GetRenderSafeArea|SDL_GetRenderScale|SDL_GetRenderVSync|SDL_GetRenderViewport|SDL_GetSensorData|SDL_GetStorageFileSize|SDL_GetStoragePathInfo|SDL_GetSurfaceAlphaMod|SDL_GetSurfaceBlendMode|SDL_GetSurfaceClipRect|SDL_GetSurfaceColorKey|SDL_GetSurfaceColorMod|SDL_GetTextInputArea|SDL_GetTextureAlphaMod|SDL_GetTextureAlphaModFloat|SDL_GetTextureBlendMode|SDL_GetTextureColorMod|SDL_GetTextureColorModFloat|SDL_GetTextureScaleMode|SDL_GetTextureSize|SDL_GetWindowAspectRatio|SDL_GetWindowBordersSize|SDL_GetWindowMaximumSize|SDL_GetWindowMinimumSize|SDL_GetWindowPosition|SDL_GetWindowRelativeMouseMode|SDL_GetWindowSafeArea|SDL_GetWindowSize|SDL_GetWindowSizeInPixels|SDL_GetWindowSurfaceVSync|SDL_HideCursor|SDL_HideWindow|SDL_Init|SDL_InitHapticRumble|SDL_InitSubSystem|SDL_LoadWAV|SDL_LoadWAV_IO|SDL_LockAudioStream|SDL_LockProperties|SDL_LockSurface|SDL_LockTexture|SDL_LockTextureToSurface|SDL_MaximizeWindow|SDL_MinimizeWindow|SDL_MixAudio|SDL_OpenURL|SDL_OutOfMemory|SDL_PauseAudioDevice|SDL_PauseAudioStreamDevice|SDL_PauseHaptic|SDL_PlayHapticRumble|SDL_PremultiplyAlpha|SDL_PremultiplySurfaceAlpha|SDL_PushEvent|SDL_PutAudioStreamData|SDL_RaiseWindow|SDL_ReadStorageFile|SDL_ReadSurfacePixel|SDL_ReadSurfacePixelFloat|SDL_RegisterApp|SDL_ReloadGamepadMappings|SDL_RemovePath|SDL_RemoveStoragePath|SDL_RemoveTimer|SDL_RenamePath|SDL_RenameStoragePath|SDL_RenderClear|SDL_RenderCoordinatesFromWindow|SDL_RenderCoordinatesToWindow|SDL_RenderFillRect|SDL_RenderFillRects|SDL_RenderGeometry|SDL_RenderGeometryRaw|SDL_RenderLine|SDL_RenderLines|SDL_RenderPoint|SDL_RenderPoints|SDL_RenderPresent|SDL_RenderRect|SDL_RenderRects|SDL_RenderTexture|SDL_RenderTexture9Grid|SDL_RenderTextureRotated|SDL_RenderTextureTiled|SDL_RequestAndroidPermission|SDL_RestoreWindow|SDL_ResumeAudioDevice|SDL_ResumeAudioStreamDevice|SDL_ResumeHaptic|SDL_RumbleGamepad|SDL_RumbleGamepadTriggers|SDL_RumbleJoystick|SDL_RumbleJoystickTriggers|SDL_RunHapticEffect|SDL_SaveBMP|SDL_SaveBMP_IO|SDL_SendAndroidMessage|SDL_SendGamepadEffect|SDL_SendJoystickEffect|SDL_SendJoystickVirtualSensorData|SDL_SetAppMetadata|SDL_SetAppMetadataProperty|SDL_SetAudioDeviceGain|SDL_SetAudioPostmixCallback|SDL_SetAudioStreamFormat|SDL_SetAudioStreamFrequencyRatio|SDL_SetAudioStreamGain|SDL_SetAudioStreamGetCallback|SDL_SetAudioStreamInputChannelMap|SDL_SetAudioStreamOutputChannelMap|SDL_SetAudioStreamPutCallback|SDL_SetBooleanProperty|SDL_SetClipboardData|SDL_SetClipboardText|SDL_SetCursor|SDL_SetFloatProperty|SDL_SetGamepadLED|SDL_SetGamepadMapping|SDL_SetGamepadPlayerIndex|SDL_SetGamepadSensorEnabled|SDL_SetHapticAutocenter|SDL_SetHapticGain|SDL_SetJoystickLED|SDL_SetJoystickPlayerIndex|SDL_SetJoystickVirtualAxis|SDL_SetJoystickVirtualBall|SDL_SetJoystickVirtualButton|SDL_SetJoystickVirtualHat|SDL_SetJoystickVirtualTouchpad|SDL_SetLinuxThreadPriority|SDL_SetLinuxThreadPriorityAndPolicy|SDL_SetLogPriorityPrefix|SDL_SetMemoryFunctions|SDL_SetNumberProperty|SDL_SetPaletteColors|SDL_SetPointerProperty|SDL_SetPointerPropertyWithCleanup|SDL_SetPrimarySelectionText|SDL_SetRenderClipRect|SDL_SetRenderColorScale|SDL_SetRenderDrawBlendMode|SDL_SetRenderDrawColor|SDL_SetRenderDrawColorFloat|SDL_SetRenderLogicalPresentation|SDL_SetRenderScale|SDL_SetRenderTarget|SDL_SetRenderVSync|SDL_SetRenderViewport|SDL_SetScancodeName|SDL_SetStringProperty|SDL_SetSurfaceAlphaMod|SDL_SetSurfaceBlendMode|SDL_SetSurfaceColorKey|SDL_SetSurfaceColorMod|SDL_SetSurfaceColorspace|SDL_SetSurfacePalette|SDL_SetSurfaceRLE|SDL_SetTLS|SDL_SetTextInputArea|SDL_SetTextureAlphaMod|SDL_SetTextureAlphaModFloat|SDL_SetTextureBlendMode|SDL_SetTextureColorMod|SDL_SetTextureColorModFloat|SDL_SetTextureScaleMode|SDL_SetThreadPriority|SDL_SetWindowAlwaysOnTop|SDL_SetWindowAspectRatio|SDL_SetWindowBordered|SDL_SetWindowFocusable|SDL_SetWindowFullscreen|SDL_SetWindowFullscreenMode|SDL_SetWindowHitTest|SDL_SetWindowIcon|SDL_SetWindowKeyboardGrab|SDL_SetWindowMaximumSize|SDL_SetWindowMinimumSize|SDL_SetWindowModalFor|SDL_SetWindowMouseGrab|SDL_SetWindowMouseRect|SDL_SetWindowOpacity|SDL_SetWindowPosition|SDL_SetWindowRelativeMouseMode|SDL_SetWindowResizable|SDL_SetWindowShape|SDL_SetWindowSize|SDL_SetWindowSurfaceVSync|SDL_SetWindowTitle|SDL_SetiOSAnimationCallback|SDL_ShowAndroidToast|SDL_ShowCursor|SDL_ShowMessageBox|SDL_ShowSimpleMessageBox|SDL_ShowWindow|SDL_ShowWindowSystemMenu|SDL_StartTextInput|SDL_StartTextInputWithProperties|SDL_StopHapticEffect|SDL_StopHapticEffects|SDL_StopHapticRumble|SDL_StopTextInput|SDL_SyncWindow|SDL_TimeToDateTime|SDL_TryLockMutex|SDL_TryLockRWLockForReading|SDL_TryLockRWLockForWriting|SDL_TryWaitSemaphore|SDL_UnlockAudioStream|SDL_UpdateHapticEffect|SDL_UpdateNVTexture|SDL_UpdateTexture|SDL_UpdateWindowSurface|SDL_UpdateWindowSurfaceRects|SDL_UpdateYUVTexture|SDL_Vulkan_CreateSurface|SDL_Vulkan_LoadLibrary|SDL_WaitConditionTimeout|SDL_WaitSemaphoreTimeout|SDL_WarpMouseGlobal|SDL_WriteStorageFile|SDL_WriteSurfacePixel|SDL_WriteSurfacePixelFloat)$"; @@ ( func( ... ) - == 0 | - func( + !func( ... ) - < 0 | - func( + !func( ... ) - != 0 | - func( + !func( ... ) - == -1 )
2024-08-23 03:33:49 +03:00
if (!result) {
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
return TEST_ABORTED;
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
}
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
/* Call against new test window */
Remove more reserved identifiers (#6925)
2022-12-30 00:58:16 +03:00
window = createVideoSuiteTestWindow(title);
Pointer as bool (libsdl-org#7214)
2023-11-10 00:29:15 +03:00
if (!window) {
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
return TEST_ABORTED;
}
for (wVariation = 0; wVariation < 5; wVariation++) {
for (hVariation = 0; hVariation < 5; hVariation++) {
switch (wVariation) {
case 0:
/* 1 Pixel Wide */
desiredW = 1;
break;
case 1:
/* Random width inside screen */
desiredW = SDLTest_RandomIntegerInRange(2, display.w - 1);
break;
case 2:
/* Width at screen size */
desiredW = display.w;
break;
}
switch (hVariation) {
case 0:
/* 1 Pixel High */
desiredH = 1;
break;
case 1:
/* Random height inside screen */
desiredH = SDLTest_RandomIntegerInRange(2, display.h - 1);
break;
case 2:
/* Height at screen size */
desiredH = display.h;
break;
case 4:
/* Height 1 pixel larger than screen */
desiredH = display.h + 1;
break;
}
/* Set size */
SDL_SetWindowMinimumSize(window, desiredW, desiredH);
SDLTest_AssertPass("Call to SDL_SetWindowMinimumSize(...,%d,%d)", desiredW, desiredH);
/* Get size */
currentW = desiredW + 1;
currentH = desiredH + 1;
SDL_GetWindowMinimumSize(window, &currentW, &currentH);
SDLTest_AssertPass("Call to SDL_GetWindowMinimumSize()");
SDLTest_AssertCheck(desiredW == currentW, "Verify returned width; expected: %d, got: %d", desiredW, currentW);
SDLTest_AssertCheck(desiredH == currentH, "Verify returned height; expected: %d, got: %d", desiredH, currentH);
/* Get just width */
currentW = desiredW + 1;
SDL_GetWindowMinimumSize(window, &currentW, NULL);
SDLTest_AssertPass("Call to SDL_GetWindowMinimumSize(&h=NULL)");
SDLTest_AssertCheck(desiredW == currentW, "Verify returned width; expected: %d, got: %d", desiredW, currentH);
/* Get just height */
currentH = desiredH + 1;
SDL_GetWindowMinimumSize(window, NULL, &currentH);
SDLTest_AssertPass("Call to SDL_GetWindowMinimumSize(&w=NULL)");
SDLTest_AssertCheck(desiredH == currentH, "Verify returned height; expected: %d, got: %d", desiredW, currentH);
}
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
}
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
/* Dummy call with both pointers NULL */
SDL_GetWindowMinimumSize(window, NULL, NULL);
SDLTest_AssertPass("Call to SDL_GetWindowMinimumSize(&w=NULL,&h=NULL)");
/* Negative tests for parameter input */
SDL_ClearError();
SDLTest_AssertPass("Call to SDL_ClearError()");
for (desiredH = -2; desiredH < 2; desiredH++) {
for (desiredW = -2; desiredW < 2; desiredW++) {
Allow passing 0 to SDL_SetWindowMinimumSize() and SDL_SetWindowMaximumSize() to indicate no limit
2023-03-16 02:13:56 +03:00
if (desiredW < 0 || desiredH < 0) {
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
SDL_SetWindowMinimumSize(window, desiredW, desiredH);
SDLTest_AssertPass("Call to SDL_SetWindowMinimumSize(...,%d,%d)", desiredW, desiredH);
Remove more reserved identifiers (#6925)
2022-12-30 00:58:16 +03:00
checkInvalidParameterError();
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
}
}
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
}
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
/* Clean up */
Remove more reserved identifiers (#6925)
2022-12-30 00:58:16 +03:00
destroyVideoSuiteTestWindow(window);
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
/* Set some 'magic' value for later check that nothing was changed */
referenceW = SDLTest_RandomSint32();
referenceH = SDLTest_RandomSint32();
currentW = referenceW;
currentH = referenceH;
desiredW = SDLTest_RandomSint32();
desiredH = SDLTest_RandomSint32();
/* Negative tests for window input */
SDL_ClearError();
SDLTest_AssertPass("Call to SDL_ClearError()");
SDL_GetWindowMinimumSize(NULL, &currentW, &currentH);
SDLTest_AssertPass("Call to SDL_GetWindowMinimumSize(window=NULL)");
SDLTest_AssertCheck(
currentW == referenceW && currentH == referenceH,
"Verify that content of W and H pointers has not been modified; expected: %d,%d; got: %d,%d",
referenceW, referenceH,
currentW, currentH);
Remove more reserved identifiers (#6925)
2022-12-30 00:58:16 +03:00
checkInvalidWindowError();
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
SDL_GetWindowMinimumSize(NULL, NULL, NULL);
SDLTest_AssertPass("Call to SDL_GetWindowMinimumSize(NULL, NULL, NULL)");
Remove more reserved identifiers (#6925)
2022-12-30 00:58:16 +03:00
checkInvalidWindowError();
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
SDL_SetWindowMinimumSize(NULL, desiredW, desiredH);
SDLTest_AssertPass("Call to SDL_SetWindowMinimumSize(window=NULL)");
Remove more reserved identifiers (#6925)
2022-12-30 00:58:16 +03:00
checkInvalidWindowError();
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
return TEST_COMPLETED;
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
}
/**
docs: Remove Doxygen `\brief` tags. Doxygen and the wiki bridge don't need them; they'll both just use the first line/sentence instead. Fixes #8446.
2023-11-06 18:26:06 +03:00
* Tests call to SDL_GetWindowMaximumSize and SDL_SetWindowMaximumSize
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
*
*/
SDL_test: use SDLCALL calling convention This is needed when using a pre-built static SDL3_test library.
2024-09-06 04:21:13 +03:00
static int SDLCALL video_getSetWindowMaximumSize(void *arg)
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
{
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
const char *title = "video_getSetWindowMaximumSize Test Window";
SDL_Window *window;
int result;
SDL_Rect display;
int wVariation, hVariation;
int referenceW, referenceH;
int currentW, currentH;
More fix warnings about static function and prototype
2023-03-08 18:12:45 +03:00
int desiredW = 0, desiredH = 0;
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
/* Get display bounds for size range */
Displays are now referenced by instance ID instead of index
2023-01-30 00:30:55 +03:00
result = SDL_GetDisplayBounds(SDL_GetPrimaryDisplay(), &display);
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
SDLTest_AssertPass("SDL_GetDisplayBounds()");
Removed SDL_bool in favor of plain bool We require stdbool.h in the build environment, so we might as well use the plain bool type. If your environment doesn't have stdbool.h, this simple replacement will suffice: typedef signed char bool;
2024-09-18 17:52:28 +03:00
SDLTest_AssertCheck(result == true, "Verify return value; expected: true, got: %d", result);
Use SDL_bool instead an int return code in the SDL API Most SDL functions used to indicate success or failure using an int return code. These functions have been changed to return SDL_bool. Here is a coccinelle patch to change code that previously compared the return value to 0 and changes it to a boolean test: @ bool_return_type @ identifier func =~ "^(SDL_AddEventWatch|SDL_AddHintCallback|SDL_AddSurfaceAlternateImage|SDL_AddVulkanRenderSemaphores|SDL_BindAudioStream|SDL_BindAudioStreams|SDL_BlitSurface|SDL_BlitSurface9Grid|SDL_BlitSurfaceScaled|SDL_BlitSurfaceTiled|SDL_BlitSurfaceTiledWithScale|SDL_BlitSurfaceUnchecked|SDL_BlitSurfaceUncheckedScaled|SDL_CaptureMouse|SDL_ClearAudioStream|SDL_ClearClipboardData|SDL_ClearComposition|SDL_ClearError|SDL_ClearProperty|SDL_ClearSurface|SDL_CloseIO|SDL_CloseStorage|SDL_ConvertAudioSamples|SDL_ConvertEventToRenderCoordinates|SDL_ConvertPixels|SDL_ConvertPixelsAndColorspace|SDL_CopyFile|SDL_CopyProperties|SDL_CopyStorageFile|SDL_CreateDirectory|SDL_CreateStorageDirectory|SDL_CreateWindowAndRenderer|SDL_DateTimeToTime|SDL_DestroyWindowSurface|SDL_DetachVirtualJoystick|SDL_DisableScreenSaver|SDL_EnableScreenSaver|SDL_EnumerateDirectory|SDL_EnumerateProperties|SDL_EnumerateStorageDirectory|SDL_FillSurfaceRect|SDL_FillSurfaceRects|SDL_FlashWindow|SDL_FlipSurface|SDL_FlushAudioStream|SDL_FlushRenderer|SDL_GL_DestroyContext|SDL_GL_GetAttribute|SDL_GL_GetSwapInterval|SDL_GL_LoadLibrary|SDL_GL_MakeCurrent|SDL_GL_SetAttribute|SDL_GL_SetSwapInterval|SDL_GL_SwapWindow|SDL_GetAudioDeviceFormat|SDL_GetAudioStreamFormat|SDL_GetCameraFormat|SDL_GetClosestFullscreenDisplayMode|SDL_GetCurrentRenderOutputSize|SDL_GetCurrentTime|SDL_GetDXGIOutputInfo|SDL_GetDateTimeLocalePreferences|SDL_GetDisplayBounds|SDL_GetDisplayUsableBounds|SDL_GetGDKDefaultUser|SDL_GetGDKTaskQueue|SDL_GetGamepadSensorData|SDL_GetGamepadTouchpadFinger|SDL_GetHapticEffectStatus|SDL_GetJoystickBall|SDL_GetMasksForPixelFormat|SDL_GetPathInfo|SDL_GetRectUnion|SDL_GetRectUnionFloat|SDL_GetRenderClipRect|SDL_GetRenderColorScale|SDL_GetRenderDrawBlendMode|SDL_GetRenderDrawColor|SDL_GetRenderDrawColorFloat|SDL_GetRenderLogicalPresentation|SDL_GetRenderLogicalPresentationRect|SDL_GetRenderOutputSize|SDL_GetRenderSafeArea|SDL_GetRenderScale|SDL_GetRenderVSync|SDL_GetRenderViewport|SDL_GetSensorData|SDL_GetStorageFileSize|SDL_GetStoragePathInfo|SDL_GetSurfaceAlphaMod|SDL_GetSurfaceBlendMode|SDL_GetSurfaceClipRect|SDL_GetSurfaceColorKey|SDL_GetSurfaceColorMod|SDL_GetTextInputArea|SDL_GetTextureAlphaMod|SDL_GetTextureAlphaModFloat|SDL_GetTextureBlendMode|SDL_GetTextureColorMod|SDL_GetTextureColorModFloat|SDL_GetTextureScaleMode|SDL_GetTextureSize|SDL_GetWindowAspectRatio|SDL_GetWindowBordersSize|SDL_GetWindowMaximumSize|SDL_GetWindowMinimumSize|SDL_GetWindowPosition|SDL_GetWindowRelativeMouseMode|SDL_GetWindowSafeArea|SDL_GetWindowSize|SDL_GetWindowSizeInPixels|SDL_GetWindowSurfaceVSync|SDL_HideCursor|SDL_HideWindow|SDL_Init|SDL_InitHapticRumble|SDL_InitSubSystem|SDL_LoadWAV|SDL_LoadWAV_IO|SDL_LockAudioStream|SDL_LockProperties|SDL_LockSurface|SDL_LockTexture|SDL_LockTextureToSurface|SDL_MaximizeWindow|SDL_MinimizeWindow|SDL_MixAudio|SDL_OpenURL|SDL_OutOfMemory|SDL_PauseAudioDevice|SDL_PauseAudioStreamDevice|SDL_PauseHaptic|SDL_PlayHapticRumble|SDL_PremultiplyAlpha|SDL_PremultiplySurfaceAlpha|SDL_PushEvent|SDL_PutAudioStreamData|SDL_RaiseWindow|SDL_ReadStorageFile|SDL_ReadSurfacePixel|SDL_ReadSurfacePixelFloat|SDL_RegisterApp|SDL_ReloadGamepadMappings|SDL_RemovePath|SDL_RemoveStoragePath|SDL_RemoveTimer|SDL_RenamePath|SDL_RenameStoragePath|SDL_RenderClear|SDL_RenderCoordinatesFromWindow|SDL_RenderCoordinatesToWindow|SDL_RenderFillRect|SDL_RenderFillRects|SDL_RenderGeometry|SDL_RenderGeometryRaw|SDL_RenderLine|SDL_RenderLines|SDL_RenderPoint|SDL_RenderPoints|SDL_RenderPresent|SDL_RenderRect|SDL_RenderRects|SDL_RenderTexture|SDL_RenderTexture9Grid|SDL_RenderTextureRotated|SDL_RenderTextureTiled|SDL_RequestAndroidPermission|SDL_RestoreWindow|SDL_ResumeAudioDevice|SDL_ResumeAudioStreamDevice|SDL_ResumeHaptic|SDL_RumbleGamepad|SDL_RumbleGamepadTriggers|SDL_RumbleJoystick|SDL_RumbleJoystickTriggers|SDL_RunHapticEffect|SDL_SaveBMP|SDL_SaveBMP_IO|SDL_SendAndroidMessage|SDL_SendGamepadEffect|SDL_SendJoystickEffect|SDL_SendJoystickVirtualSensorData|SDL_SetAppMetadata|SDL_SetAppMetadataProperty|SDL_SetAudioDeviceGain|SDL_SetAudioPostmixCallback|SDL_SetAudioStreamFormat|SDL_SetAudioStreamFrequencyRatio|SDL_SetAudioStreamGain|SDL_SetAudioStreamGetCallback|SDL_SetAudioStreamInputChannelMap|SDL_SetAudioStreamOutputChannelMap|SDL_SetAudioStreamPutCallback|SDL_SetBooleanProperty|SDL_SetClipboardData|SDL_SetClipboardText|SDL_SetCursor|SDL_SetFloatProperty|SDL_SetGamepadLED|SDL_SetGamepadMapping|SDL_SetGamepadPlayerIndex|SDL_SetGamepadSensorEnabled|SDL_SetHapticAutocenter|SDL_SetHapticGain|SDL_SetJoystickLED|SDL_SetJoystickPlayerIndex|SDL_SetJoystickVirtualAxis|SDL_SetJoystickVirtualBall|SDL_SetJoystickVirtualButton|SDL_SetJoystickVirtualHat|SDL_SetJoystickVirtualTouchpad|SDL_SetLinuxThreadPriority|SDL_SetLinuxThreadPriorityAndPolicy|SDL_SetLogPriorityPrefix|SDL_SetMemoryFunctions|SDL_SetNumberProperty|SDL_SetPaletteColors|SDL_SetPointerProperty|SDL_SetPointerPropertyWithCleanup|SDL_SetPrimarySelectionText|SDL_SetRenderClipRect|SDL_SetRenderColorScale|SDL_SetRenderDrawBlendMode|SDL_SetRenderDrawColor|SDL_SetRenderDrawColorFloat|SDL_SetRenderLogicalPresentation|SDL_SetRenderScale|SDL_SetRenderTarget|SDL_SetRenderVSync|SDL_SetRenderViewport|SDL_SetScancodeName|SDL_SetStringProperty|SDL_SetSurfaceAlphaMod|SDL_SetSurfaceBlendMode|SDL_SetSurfaceColorKey|SDL_SetSurfaceColorMod|SDL_SetSurfaceColorspace|SDL_SetSurfacePalette|SDL_SetSurfaceRLE|SDL_SetTLS|SDL_SetTextInputArea|SDL_SetTextureAlphaMod|SDL_SetTextureAlphaModFloat|SDL_SetTextureBlendMode|SDL_SetTextureColorMod|SDL_SetTextureColorModFloat|SDL_SetTextureScaleMode|SDL_SetThreadPriority|SDL_SetWindowAlwaysOnTop|SDL_SetWindowAspectRatio|SDL_SetWindowBordered|SDL_SetWindowFocusable|SDL_SetWindowFullscreen|SDL_SetWindowFullscreenMode|SDL_SetWindowHitTest|SDL_SetWindowIcon|SDL_SetWindowKeyboardGrab|SDL_SetWindowMaximumSize|SDL_SetWindowMinimumSize|SDL_SetWindowModalFor|SDL_SetWindowMouseGrab|SDL_SetWindowMouseRect|SDL_SetWindowOpacity|SDL_SetWindowPosition|SDL_SetWindowRelativeMouseMode|SDL_SetWindowResizable|SDL_SetWindowShape|SDL_SetWindowSize|SDL_SetWindowSurfaceVSync|SDL_SetWindowTitle|SDL_SetiOSAnimationCallback|SDL_ShowAndroidToast|SDL_ShowCursor|SDL_ShowMessageBox|SDL_ShowSimpleMessageBox|SDL_ShowWindow|SDL_ShowWindowSystemMenu|SDL_StartTextInput|SDL_StartTextInputWithProperties|SDL_StopHapticEffect|SDL_StopHapticEffects|SDL_StopHapticRumble|SDL_StopTextInput|SDL_SyncWindow|SDL_TimeToDateTime|SDL_TryLockMutex|SDL_TryLockRWLockForReading|SDL_TryLockRWLockForWriting|SDL_TryWaitSemaphore|SDL_UnlockAudioStream|SDL_UpdateHapticEffect|SDL_UpdateNVTexture|SDL_UpdateTexture|SDL_UpdateWindowSurface|SDL_UpdateWindowSurfaceRects|SDL_UpdateYUVTexture|SDL_Vulkan_CreateSurface|SDL_Vulkan_LoadLibrary|SDL_WaitConditionTimeout|SDL_WaitSemaphoreTimeout|SDL_WarpMouseGlobal|SDL_WriteStorageFile|SDL_WriteSurfacePixel|SDL_WriteSurfacePixelFloat)$"; @@ ( func( ... ) - == 0 | - func( + !func( ... ) - < 0 | - func( + !func( ... ) - != 0 | - func( + !func( ... ) - == -1 )
2024-08-23 03:33:49 +03:00
if (!result) {
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
return TEST_ABORTED;
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
}
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
/* Call against new test window */
Remove more reserved identifiers (#6925)
2022-12-30 00:58:16 +03:00
window = createVideoSuiteTestWindow(title);
Pointer as bool (libsdl-org#7214)
2023-11-10 00:29:15 +03:00
if (!window) {
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
return TEST_ABORTED;
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
}
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
for (wVariation = 0; wVariation < 3; wVariation++) {
for (hVariation = 0; hVariation < 3; hVariation++) {
switch (wVariation) {
case 0:
/* 1 Pixel Wide */
desiredW = 1;
break;
case 1:
/* Random width inside screen */
desiredW = SDLTest_RandomIntegerInRange(2, display.w - 1);
break;
case 2:
/* Width at screen size */
desiredW = display.w;
break;
}
switch (hVariation) {
case 0:
/* 1 Pixel High */
desiredH = 1;
break;
case 1:
/* Random height inside screen */
desiredH = SDLTest_RandomIntegerInRange(2, display.h - 1);
break;
case 2:
/* Height at screen size */
desiredH = display.h;
break;
}
/* Set size */
SDL_SetWindowMaximumSize(window, desiredW, desiredH);
SDLTest_AssertPass("Call to SDL_SetWindowMaximumSize(...,%d,%d)", desiredW, desiredH);
/* Get size */
currentW = desiredW + 1;
currentH = desiredH + 1;
SDL_GetWindowMaximumSize(window, &currentW, &currentH);
SDLTest_AssertPass("Call to SDL_GetWindowMaximumSize()");
SDLTest_AssertCheck(desiredW == currentW, "Verify returned width; expected: %d, got: %d", desiredW, currentW);
SDLTest_AssertCheck(desiredH == currentH, "Verify returned height; expected: %d, got: %d", desiredH, currentH);
/* Get just width */
currentW = desiredW + 1;
SDL_GetWindowMaximumSize(window, &currentW, NULL);
SDLTest_AssertPass("Call to SDL_GetWindowMaximumSize(&h=NULL)");
SDLTest_AssertCheck(desiredW == currentW, "Verify returned width; expected: %d, got: %d", desiredW, currentH);
/* Get just height */
currentH = desiredH + 1;
SDL_GetWindowMaximumSize(window, NULL, &currentH);
SDLTest_AssertPass("Call to SDL_GetWindowMaximumSize(&w=NULL)");
SDLTest_AssertCheck(desiredH == currentH, "Verify returned height; expected: %d, got: %d", desiredW, currentH);
}
}
/* Dummy call with both pointers NULL */
SDL_GetWindowMaximumSize(window, NULL, NULL);
SDLTest_AssertPass("Call to SDL_GetWindowMaximumSize(&w=NULL,&h=NULL)");
/* Negative tests for parameter input */
SDL_ClearError();
SDLTest_AssertPass("Call to SDL_ClearError()");
for (desiredH = -2; desiredH < 2; desiredH++) {
for (desiredW = -2; desiredW < 2; desiredW++) {
Allow passing 0 to SDL_SetWindowMinimumSize() and SDL_SetWindowMaximumSize() to indicate no limit
2023-03-16 02:13:56 +03:00
if (desiredW < 0 || desiredH < 0) {
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
SDL_SetWindowMaximumSize(window, desiredW, desiredH);
SDLTest_AssertPass("Call to SDL_SetWindowMaximumSize(...,%d,%d)", desiredW, desiredH);
Remove more reserved identifiers (#6925)
2022-12-30 00:58:16 +03:00
checkInvalidParameterError();
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
}
}
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
}
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
/* Clean up */
Remove more reserved identifiers (#6925)
2022-12-30 00:58:16 +03:00
destroyVideoSuiteTestWindow(window);
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
/* Set some 'magic' value for later check that nothing was changed */
referenceW = SDLTest_RandomSint32();
referenceH = SDLTest_RandomSint32();
currentW = referenceW;
currentH = referenceH;
desiredW = SDLTest_RandomSint32();
desiredH = SDLTest_RandomSint32();
/* Negative tests */
SDL_ClearError();
SDLTest_AssertPass("Call to SDL_ClearError()");
SDL_GetWindowMaximumSize(NULL, &currentW, &currentH);
SDLTest_AssertPass("Call to SDL_GetWindowMaximumSize(window=NULL)");
SDLTest_AssertCheck(
currentW == referenceW && currentH == referenceH,
"Verify that content of W and H pointers has not been modified; expected: %d,%d; got: %d,%d",
referenceW, referenceH,
currentW, currentH);
Remove more reserved identifiers (#6925)
2022-12-30 00:58:16 +03:00
checkInvalidWindowError();
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
SDL_GetWindowMaximumSize(NULL, NULL, NULL);
SDLTest_AssertPass("Call to SDL_GetWindowMaximumSize(NULL, NULL, NULL)");
Remove more reserved identifiers (#6925)
2022-12-30 00:58:16 +03:00
checkInvalidWindowError();
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
SDL_SetWindowMaximumSize(NULL, desiredW, desiredH);
SDLTest_AssertPass("Call to SDL_SetWindowMaximumSize(window=NULL)");
Remove more reserved identifiers (#6925)
2022-12-30 00:58:16 +03:00
checkInvalidWindowError();
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
return TEST_COMPLETED;
}
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
/**
docs: Remove Doxygen `\brief` tags. Doxygen and the wiki bridge don't need them; they'll both just use the first line/sentence instead. Fixes #8446.
2023-11-06 18:26:06 +03:00
* Tests call to SDL_SetWindowData and SDL_GetWindowData
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
*
test: don't use wiki urls for documentation comments Also make consistent use of \ as documentation escape character.
2023-02-02 02:21:53 +03:00
* \sa SDL_SetWindowData
* \sa SDL_GetWindowData
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
*/
SDL_test: use SDLCALL calling convention This is needed when using a pre-built static SDL3_test library.
2024-09-06 04:21:13 +03:00
static int SDLCALL video_getSetWindowData(void *arg)
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
{
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
int returnValue = TEST_COMPLETED;
const char *title = "video_setGetWindowData Test Window";
SDL_Window *window;
const char *referenceName = "TestName";
const char *name = "TestName";
const char *referenceName2 = "TestName2";
const char *name2 = "TestName2";
int datasize;
char *referenceUserdata = NULL;
char *userdata = NULL;
char *referenceUserdata2 = NULL;
char *userdata2 = NULL;
char *result;
int iteration;
/* Call against new test window */
Remove more reserved identifiers (#6925)
2022-12-30 00:58:16 +03:00
window = createVideoSuiteTestWindow(title);
Pointer as bool (libsdl-org#7214)
2023-11-10 00:29:15 +03:00
if (!window) {
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
return TEST_ABORTED;
}
/* Create testdata */
datasize = SDLTest_RandomIntegerInRange(1, 32);
referenceUserdata = SDLTest_RandomAsciiStringOfSize(datasize);
Pointer as bool (libsdl-org#7214)
2023-11-10 00:29:15 +03:00
if (!referenceUserdata) {
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
returnValue = TEST_ABORTED;
goto cleanup;
}
userdata = SDL_strdup(referenceUserdata);
Pointer as bool (libsdl-org#7214)
2023-11-10 00:29:15 +03:00
if (!userdata) {
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
returnValue = TEST_ABORTED;
goto cleanup;
}
datasize = SDLTest_RandomIntegerInRange(1, 32);
referenceUserdata2 = SDLTest_RandomAsciiStringOfSize(datasize);
Pointer as bool (libsdl-org#7214)
2023-11-10 00:29:15 +03:00
if (!referenceUserdata2) {
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
returnValue = TEST_ABORTED;
goto cleanup;
}
Clang-Tidy fixes (#6725)
2022-12-02 00:07:03 +03:00
userdata2 = SDL_strdup(referenceUserdata2);
Pointer as bool (libsdl-org#7214)
2023-11-10 00:29:15 +03:00
if (!userdata2) {
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
returnValue = TEST_ABORTED;
goto cleanup;
}
/* Get non-existent data */
Renamed SDL_Get/SetProperty() to SDL_Get/SetPointerProperty() This is consistent with the naming for the functions that affect other data types Fixes https://github.com/libsdl-org/SDL/issues/10241
2024-07-12 19:39:58 +03:00
result = (char *)SDL_GetPointerProperty(SDL_GetWindowProperties(window), name, NULL);
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
SDLTest_AssertPass("Call to SDL_GetWindowData(..,%s)", name);
SDLTest_AssertCheck(result == NULL, "Validate that result is NULL");
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
SDLTest_AssertCheck(SDL_strcmp(referenceName, name) == 0, "Validate that name was not changed, expected: %s, got: %s", referenceName, name);
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
/* Set data */
Renamed SDL_Get/SetProperty() to SDL_Get/SetPointerProperty() This is consistent with the naming for the functions that affect other data types Fixes https://github.com/libsdl-org/SDL/issues/10241
2024-07-12 19:39:58 +03:00
SDL_SetPointerProperty(SDL_GetWindowProperties(window), name, userdata);
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
SDLTest_AssertPass("Call to SDL_SetWindowData(...%s,%s)", name, userdata);
SDLTest_AssertCheck(SDL_strcmp(referenceName, name) == 0, "Validate that name was not changed, expected: %s, got: %s", referenceName, name);
SDLTest_AssertCheck(SDL_strcmp(referenceUserdata, userdata) == 0, "Validate that userdata was not changed, expected: %s, got: %s", referenceUserdata, userdata);
/* Get data (twice) */
for (iteration = 1; iteration <= 2; iteration++) {
Renamed SDL_Get/SetProperty() to SDL_Get/SetPointerProperty() This is consistent with the naming for the functions that affect other data types Fixes https://github.com/libsdl-org/SDL/issues/10241
2024-07-12 19:39:58 +03:00
result = (char *)SDL_GetPointerProperty(SDL_GetWindowProperties(window), name, NULL);
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
SDLTest_AssertPass("Call to SDL_GetWindowData(..,%s) [iteration %d]", name, iteration);
SDLTest_AssertCheck(SDL_strcmp(referenceUserdata, result) == 0, "Validate that correct result was returned; expected: %s, got: %s", referenceUserdata, result);
SDLTest_AssertCheck(SDL_strcmp(referenceName, name) == 0, "Validate that name was not changed, expected: %s, got: %s", referenceName, name);
}
/* Set data again twice */
for (iteration = 1; iteration <= 2; iteration++) {
Renamed SDL_Get/SetProperty() to SDL_Get/SetPointerProperty() This is consistent with the naming for the functions that affect other data types Fixes https://github.com/libsdl-org/SDL/issues/10241
2024-07-12 19:39:58 +03:00
SDL_SetPointerProperty(SDL_GetWindowProperties(window), name, userdata);
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
SDLTest_AssertPass("Call to SDL_SetWindowData(...%s,%s) [iteration %d]", name, userdata, iteration);
SDLTest_AssertCheck(SDL_strcmp(referenceName, name) == 0, "Validate that name was not changed, expected: %s, got: %s", referenceName, name);
SDLTest_AssertCheck(SDL_strcmp(referenceUserdata, userdata) == 0, "Validate that userdata was not changed, expected: %s, got: %s", referenceUserdata, userdata);
}
/* Get data again */
Renamed SDL_Get/SetProperty() to SDL_Get/SetPointerProperty() This is consistent with the naming for the functions that affect other data types Fixes https://github.com/libsdl-org/SDL/issues/10241
2024-07-12 19:39:58 +03:00
result = (char *)SDL_GetPointerProperty(SDL_GetWindowProperties(window), name, NULL);
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
SDLTest_AssertPass("Call to SDL_GetWindowData(..,%s) [again]", name);
SDLTest_AssertCheck(SDL_strcmp(referenceUserdata, result) == 0, "Validate that correct result was returned; expected: %s, got: %s", referenceUserdata, result);
SDLTest_AssertCheck(SDL_strcmp(referenceName, name) == 0, "Validate that name was not changed, expected: %s, got: %s", referenceName, name);
/* Set data with new data */
Renamed SDL_Get/SetProperty() to SDL_Get/SetPointerProperty() This is consistent with the naming for the functions that affect other data types Fixes https://github.com/libsdl-org/SDL/issues/10241
2024-07-12 19:39:58 +03:00
SDL_SetPointerProperty(SDL_GetWindowProperties(window), name, userdata2);
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
SDLTest_AssertPass("Call to SDL_SetWindowData(...%s,%s) [new userdata]", name, userdata2);
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
SDLTest_AssertCheck(SDL_strcmp(referenceName, name) == 0, "Validate that name was not changed, expected: %s, got: %s", referenceName, name);
SDLTest_AssertCheck(SDL_strcmp(referenceUserdata, userdata) == 0, "Validate that userdata was not changed, expected: %s, got: %s", referenceUserdata, userdata);
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
SDLTest_AssertCheck(SDL_strcmp(referenceUserdata2, userdata2) == 0, "Validate that userdata2 was not changed, expected: %s, got: %s", referenceUserdata2, userdata2);
/* Set data with new data again */
Renamed SDL_Get/SetProperty() to SDL_Get/SetPointerProperty() This is consistent with the naming for the functions that affect other data types Fixes https://github.com/libsdl-org/SDL/issues/10241
2024-07-12 19:39:58 +03:00
SDL_SetPointerProperty(SDL_GetWindowProperties(window), name, userdata2);
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
SDLTest_AssertPass("Call to SDL_SetWindowData(...%s,%s) [new userdata again]", name, userdata2);
SDLTest_AssertCheck(SDL_strcmp(referenceName, name) == 0, "Validate that name was not changed, expected: %s, got: %s", referenceName, name);
SDLTest_AssertCheck(SDL_strcmp(referenceUserdata, userdata) == 0, "Validate that userdata was not changed, expected: %s, got: %s", referenceUserdata, userdata);
SDLTest_AssertCheck(SDL_strcmp(referenceUserdata2, userdata2) == 0, "Validate that userdata2 was not changed, expected: %s, got: %s", referenceUserdata2, userdata2);
/* Get new data */
Renamed SDL_Get/SetProperty() to SDL_Get/SetPointerProperty() This is consistent with the naming for the functions that affect other data types Fixes https://github.com/libsdl-org/SDL/issues/10241
2024-07-12 19:39:58 +03:00
result = (char *)SDL_GetPointerProperty(SDL_GetWindowProperties(window), name, NULL);
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
SDLTest_AssertPass("Call to SDL_GetWindowData(..,%s)", name);
SDLTest_AssertCheck(SDL_strcmp(referenceUserdata2, result) == 0, "Validate that correct result was returned; expected: %s, got: %s", referenceUserdata2, result);
SDLTest_AssertCheck(SDL_strcmp(referenceName, name) == 0, "Validate that name was not changed, expected: %s, got: %s", referenceName, name);
/* Set data with NULL to clear */
Renamed SDL_Get/SetProperty() to SDL_Get/SetPointerProperty() This is consistent with the naming for the functions that affect other data types Fixes https://github.com/libsdl-org/SDL/issues/10241
2024-07-12 19:39:58 +03:00
SDL_SetPointerProperty(SDL_GetWindowProperties(window), name, NULL);
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
SDLTest_AssertPass("Call to SDL_SetWindowData(...%s,NULL)", name);
SDLTest_AssertCheck(SDL_strcmp(referenceName, name) == 0, "Validate that name was not changed, expected: %s, got: %s", referenceName, name);
SDLTest_AssertCheck(SDL_strcmp(referenceUserdata, userdata) == 0, "Validate that userdata was not changed, expected: %s, got: %s", referenceUserdata, userdata);
SDLTest_AssertCheck(SDL_strcmp(referenceUserdata2, userdata2) == 0, "Validate that userdata2 was not changed, expected: %s, got: %s", referenceUserdata2, userdata2);
/* Set data with NULL to clear again */
Renamed SDL_Get/SetProperty() to SDL_Get/SetPointerProperty() This is consistent with the naming for the functions that affect other data types Fixes https://github.com/libsdl-org/SDL/issues/10241
2024-07-12 19:39:58 +03:00
SDL_SetPointerProperty(SDL_GetWindowProperties(window), name, NULL);
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
SDLTest_AssertPass("Call to SDL_SetWindowData(...%s,NULL) [again]", name);
SDLTest_AssertCheck(SDL_strcmp(referenceName, name) == 0, "Validate that name was not changed, expected: %s, got: %s", referenceName, name);
SDLTest_AssertCheck(SDL_strcmp(referenceUserdata, userdata) == 0, "Validate that userdata was not changed, expected: %s, got: %s", referenceUserdata, userdata);
SDLTest_AssertCheck(SDL_strcmp(referenceUserdata2, userdata2) == 0, "Validate that userdata2 was not changed, expected: %s, got: %s", referenceUserdata2, userdata2);
/* Get non-existent data */
Renamed SDL_Get/SetProperty() to SDL_Get/SetPointerProperty() This is consistent with the naming for the functions that affect other data types Fixes https://github.com/libsdl-org/SDL/issues/10241
2024-07-12 19:39:58 +03:00
result = (char *)SDL_GetPointerProperty(SDL_GetWindowProperties(window), name, NULL);
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
SDLTest_AssertPass("Call to SDL_GetWindowData(..,%s)", name);
SDLTest_AssertCheck(result == NULL, "Validate that result is NULL");
SDLTest_AssertCheck(SDL_strcmp(referenceName, name) == 0, "Validate that name was not changed, expected: %s, got: %s", referenceName, name);
/* Get non-existent data new name */
Renamed SDL_Get/SetProperty() to SDL_Get/SetPointerProperty() This is consistent with the naming for the functions that affect other data types Fixes https://github.com/libsdl-org/SDL/issues/10241
2024-07-12 19:39:58 +03:00
result = (char *)SDL_GetPointerProperty(SDL_GetWindowProperties(window), name2, NULL);
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
SDLTest_AssertPass("Call to SDL_GetWindowData(..,%s)", name2);
SDLTest_AssertCheck(result == NULL, "Validate that result is NULL");
SDLTest_AssertCheck(SDL_strcmp(referenceName2, name2) == 0, "Validate that name2 was not changed, expected: %s, got: %s", referenceName2, name2);
/* Set data (again) */
Renamed SDL_Get/SetProperty() to SDL_Get/SetPointerProperty() This is consistent with the naming for the functions that affect other data types Fixes https://github.com/libsdl-org/SDL/issues/10241
2024-07-12 19:39:58 +03:00
SDL_SetPointerProperty(SDL_GetWindowProperties(window), name, userdata);
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
SDLTest_AssertPass("Call to SDL_SetWindowData(...%s,%s) [again, after clear]", name, userdata);
SDLTest_AssertCheck(SDL_strcmp(referenceName, name) == 0, "Validate that name was not changed, expected: %s, got: %s", referenceName, name);
SDLTest_AssertCheck(SDL_strcmp(referenceUserdata, userdata) == 0, "Validate that userdata was not changed, expected: %s, got: %s", referenceUserdata, userdata);
/* Get data (again) */
Renamed SDL_Get/SetProperty() to SDL_Get/SetPointerProperty() This is consistent with the naming for the functions that affect other data types Fixes https://github.com/libsdl-org/SDL/issues/10241
2024-07-12 19:39:58 +03:00
result = (char *)SDL_GetPointerProperty(SDL_GetWindowProperties(window), name, NULL);
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
SDLTest_AssertPass("Call to SDL_GetWindowData(..,%s) [again, after clear]", name);
SDLTest_AssertCheck(SDL_strcmp(referenceUserdata, result) == 0, "Validate that correct result was returned; expected: %s, got: %s", referenceUserdata, result);
SDLTest_AssertCheck(SDL_strcmp(referenceName, name) == 0, "Validate that name was not changed, expected: %s, got: %s", referenceName, name);
/* Set data with NULL name, valid userdata */
Renamed SDL_Get/SetProperty() to SDL_Get/SetPointerProperty() This is consistent with the naming for the functions that affect other data types Fixes https://github.com/libsdl-org/SDL/issues/10241
2024-07-12 19:39:58 +03:00
SDL_SetPointerProperty(SDL_GetWindowProperties(window), NULL, userdata);
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
SDLTest_AssertPass("Call to SDL_SetWindowData(name=NULL)");
Remove more reserved identifiers (#6925)
2022-12-30 00:58:16 +03:00
checkInvalidParameterError();
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
/* Set data with empty name, valid userdata */
Renamed SDL_Get/SetProperty() to SDL_Get/SetPointerProperty() This is consistent with the naming for the functions that affect other data types Fixes https://github.com/libsdl-org/SDL/issues/10241
2024-07-12 19:39:58 +03:00
SDL_SetPointerProperty(SDL_GetWindowProperties(window), "", userdata);
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
SDLTest_AssertPass("Call to SDL_SetWindowData(name='')");
Remove more reserved identifiers (#6925)
2022-12-30 00:58:16 +03:00
checkInvalidParameterError();
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
/* Set data with NULL name, NULL userdata */
Renamed SDL_Get/SetProperty() to SDL_Get/SetPointerProperty() This is consistent with the naming for the functions that affect other data types Fixes https://github.com/libsdl-org/SDL/issues/10241
2024-07-12 19:39:58 +03:00
SDL_SetPointerProperty(SDL_GetWindowProperties(window), NULL, NULL);
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
SDLTest_AssertPass("Call to SDL_SetWindowData(name=NULL,userdata=NULL)");
Remove more reserved identifiers (#6925)
2022-12-30 00:58:16 +03:00
checkInvalidParameterError();
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
/* Set data with empty name, NULL userdata */
Renamed SDL_Get/SetProperty() to SDL_Get/SetPointerProperty() This is consistent with the naming for the functions that affect other data types Fixes https://github.com/libsdl-org/SDL/issues/10241
2024-07-12 19:39:58 +03:00
SDL_SetPointerProperty(SDL_GetWindowProperties(window), "", NULL);
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
SDLTest_AssertPass("Call to SDL_SetWindowData(name='',userdata=NULL)");
Remove more reserved identifiers (#6925)
2022-12-30 00:58:16 +03:00
checkInvalidParameterError();
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
/* Get data with NULL name */
Renamed SDL_Get/SetProperty() to SDL_Get/SetPointerProperty() This is consistent with the naming for the functions that affect other data types Fixes https://github.com/libsdl-org/SDL/issues/10241
2024-07-12 19:39:58 +03:00
result = (char *)SDL_GetPointerProperty(SDL_GetWindowProperties(window), NULL, NULL);
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
SDLTest_AssertPass("Call to SDL_GetWindowData(name=NULL)");
SDLTest_AssertCheck(result == NULL, "Validate that result is NULL");
/* Get data with empty name */
Renamed SDL_Get/SetProperty() to SDL_Get/SetPointerProperty() This is consistent with the naming for the functions that affect other data types Fixes https://github.com/libsdl-org/SDL/issues/10241
2024-07-12 19:39:58 +03:00
result = (char *)SDL_GetPointerProperty(SDL_GetWindowProperties(window), "", NULL);
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
SDLTest_AssertPass("Call to SDL_GetWindowData(name='')");
SDLTest_AssertCheck(result == NULL, "Validate that result is NULL");
/* Clean up */
Remove more reserved identifiers (#6925)
2022-12-30 00:58:16 +03:00
destroyVideoSuiteTestWindow(window);
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
cleanup:
SDL_free(referenceUserdata);
SDL_free(referenceUserdata2);
SDL_free(userdata);
SDL_free(userdata2);
return returnValue;
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
}
testautomation: add video_setWindowCenteredOnDisplay test
2022-06-05 23:44:30 +03:00
/**
docs: Remove Doxygen `\brief` tags. Doxygen and the wiki bridge don't need them; they'll both just use the first line/sentence instead. Fixes #8446.
2023-11-06 18:26:06 +03:00
* Tests the functionality of the SDL_WINDOWPOS_CENTERED_DISPLAY along with SDL_WINDOW_FULLSCREEN.
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
*
Fix various typos (docs/build scripts/tests)
2023-07-03 18:29:42 +03:00
* Especially useful when run on a multi-monitor system with different DPI scales per monitor,
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
* to test that the window size is maintained when moving between monitors.
tests: Fix automated window grab and positioning tests under Wayland/XWayland When using the Wayland video driver or X11 under XWayland, create a renderer and present a frame as part of window creation, as Wayland requires that a frame be presented for the window to be fully mapped and displayed onscreen. This fixes the grab and expected window size tests. This also disables the window positioning tests when running under the Wayland driver, as Wayland does not allow application windows to position themselves in the desktop space, which renders the tests unreliable and subject to spurious failure.
2023-07-23 18:25:15 +03:00
*
* As the Wayland windowing protocol does not allow application windows to control their position in the
* desktop space, coupled with the general asynchronous nature of Wayland compositors, the positioning
* tests don't work in windowed mode and are unreliable in fullscreen mode, thus are disabled when using
* the Wayland video driver. All that can be done is check that the windows are the expected size.
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
*/
SDL_test: use SDLCALL calling convention This is needed when using a pre-built static SDL3_test library.
2024-09-06 04:21:13 +03:00
static int SDLCALL video_setWindowCenteredOnDisplay(void *arg)
testautomation: add video_setWindowCenteredOnDisplay test
2022-06-05 23:44:30 +03:00
{
Removed temporary memory from the API It was intended to make the API easier to use, but various automatic garbage collection all had flaws, and making the application periodically clean up temporary memory added cognitive load to using the API, and in many cases was it was difficult to restructure threaded code to handle this. So, we're largely going back to the original system, where the API returns allocated results and you free them. In addition, to solve the problems we originally wanted temporary memory for: * Short strings with a finite count, like device names, get stored in a per-thread string pool. * Events continue to use temporary memory internally, which is cleaned up on the next event processing cycle.
2024-07-27 04:57:18 +03:00
SDL_DisplayID *displays;
testautomation: add video_setWindowCenteredOnDisplay test
2022-06-05 23:44:30 +03:00
SDL_Window *window;
const char *title = "video_setWindowCenteredOnDisplay Test Window";
int x, y, w, h;
int xVariation, yVariation;
int displayNum;
int result;
SDL_Rect display0, display1;
testautomation: skip window resize and repositioning tests on Emscripten
2024-08-18 00:20:41 +03:00
const char *video_driver = SDL_GetCurrentVideoDriver();
Removed SDL_bool in favor of plain bool We require stdbool.h in the build environment, so we might as well use the plain bool type. If your environment doesn't have stdbool.h, this simple replacement will suffice: typedef signed char bool;
2024-09-18 17:52:28 +03:00
bool video_driver_is_wayland = SDL_strcmp(video_driver, "wayland") == 0;
bool video_driver_is_emscripten = SDL_strcmp(video_driver, "emscripten") == 0;
testautomation: add video_setWindowCenteredOnDisplay test
2022-06-05 23:44:30 +03:00
Displays are now referenced by instance ID instead of index
2023-01-30 00:30:55 +03:00
displays = SDL_GetDisplays(&displayNum);
if (displays) {
testautomation: add video_setWindowCenteredOnDisplay test
2022-06-05 23:44:30 +03:00
Displays are now referenced by instance ID instead of index
2023-01-30 00:30:55 +03:00
/* Get display bounds */
tests: Check the usable bounds for centered window positioning SDL now centers windows according to the usable display bounds, so the window centering test requires an update for the new behavior.
2024-05-21 19:26:40 +03:00
result = SDL_GetDisplayUsableBounds(displays[0 % displayNum], &display0);
SDLTest_AssertPass("SDL_GetDisplayUsableBounds()");
Removed SDL_bool in favor of plain bool We require stdbool.h in the build environment, so we might as well use the plain bool type. If your environment doesn't have stdbool.h, this simple replacement will suffice: typedef signed char bool;
2024-09-18 17:52:28 +03:00
SDLTest_AssertCheck(result == true, "Verify return value; expected: true, got: %d", result);
Use SDL_bool instead an int return code in the SDL API Most SDL functions used to indicate success or failure using an int return code. These functions have been changed to return SDL_bool. Here is a coccinelle patch to change code that previously compared the return value to 0 and changes it to a boolean test: @ bool_return_type @ identifier func =~ "^(SDL_AddEventWatch|SDL_AddHintCallback|SDL_AddSurfaceAlternateImage|SDL_AddVulkanRenderSemaphores|SDL_BindAudioStream|SDL_BindAudioStreams|SDL_BlitSurface|SDL_BlitSurface9Grid|SDL_BlitSurfaceScaled|SDL_BlitSurfaceTiled|SDL_BlitSurfaceTiledWithScale|SDL_BlitSurfaceUnchecked|SDL_BlitSurfaceUncheckedScaled|SDL_CaptureMouse|SDL_ClearAudioStream|SDL_ClearClipboardData|SDL_ClearComposition|SDL_ClearError|SDL_ClearProperty|SDL_ClearSurface|SDL_CloseIO|SDL_CloseStorage|SDL_ConvertAudioSamples|SDL_ConvertEventToRenderCoordinates|SDL_ConvertPixels|SDL_ConvertPixelsAndColorspace|SDL_CopyFile|SDL_CopyProperties|SDL_CopyStorageFile|SDL_CreateDirectory|SDL_CreateStorageDirectory|SDL_CreateWindowAndRenderer|SDL_DateTimeToTime|SDL_DestroyWindowSurface|SDL_DetachVirtualJoystick|SDL_DisableScreenSaver|SDL_EnableScreenSaver|SDL_EnumerateDirectory|SDL_EnumerateProperties|SDL_EnumerateStorageDirectory|SDL_FillSurfaceRect|SDL_FillSurfaceRects|SDL_FlashWindow|SDL_FlipSurface|SDL_FlushAudioStream|SDL_FlushRenderer|SDL_GL_DestroyContext|SDL_GL_GetAttribute|SDL_GL_GetSwapInterval|SDL_GL_LoadLibrary|SDL_GL_MakeCurrent|SDL_GL_SetAttribute|SDL_GL_SetSwapInterval|SDL_GL_SwapWindow|SDL_GetAudioDeviceFormat|SDL_GetAudioStreamFormat|SDL_GetCameraFormat|SDL_GetClosestFullscreenDisplayMode|SDL_GetCurrentRenderOutputSize|SDL_GetCurrentTime|SDL_GetDXGIOutputInfo|SDL_GetDateTimeLocalePreferences|SDL_GetDisplayBounds|SDL_GetDisplayUsableBounds|SDL_GetGDKDefaultUser|SDL_GetGDKTaskQueue|SDL_GetGamepadSensorData|SDL_GetGamepadTouchpadFinger|SDL_GetHapticEffectStatus|SDL_GetJoystickBall|SDL_GetMasksForPixelFormat|SDL_GetPathInfo|SDL_GetRectUnion|SDL_GetRectUnionFloat|SDL_GetRenderClipRect|SDL_GetRenderColorScale|SDL_GetRenderDrawBlendMode|SDL_GetRenderDrawColor|SDL_GetRenderDrawColorFloat|SDL_GetRenderLogicalPresentation|SDL_GetRenderLogicalPresentationRect|SDL_GetRenderOutputSize|SDL_GetRenderSafeArea|SDL_GetRenderScale|SDL_GetRenderVSync|SDL_GetRenderViewport|SDL_GetSensorData|SDL_GetStorageFileSize|SDL_GetStoragePathInfo|SDL_GetSurfaceAlphaMod|SDL_GetSurfaceBlendMode|SDL_GetSurfaceClipRect|SDL_GetSurfaceColorKey|SDL_GetSurfaceColorMod|SDL_GetTextInputArea|SDL_GetTextureAlphaMod|SDL_GetTextureAlphaModFloat|SDL_GetTextureBlendMode|SDL_GetTextureColorMod|SDL_GetTextureColorModFloat|SDL_GetTextureScaleMode|SDL_GetTextureSize|SDL_GetWindowAspectRatio|SDL_GetWindowBordersSize|SDL_GetWindowMaximumSize|SDL_GetWindowMinimumSize|SDL_GetWindowPosition|SDL_GetWindowRelativeMouseMode|SDL_GetWindowSafeArea|SDL_GetWindowSize|SDL_GetWindowSizeInPixels|SDL_GetWindowSurfaceVSync|SDL_HideCursor|SDL_HideWindow|SDL_Init|SDL_InitHapticRumble|SDL_InitSubSystem|SDL_LoadWAV|SDL_LoadWAV_IO|SDL_LockAudioStream|SDL_LockProperties|SDL_LockSurface|SDL_LockTexture|SDL_LockTextureToSurface|SDL_MaximizeWindow|SDL_MinimizeWindow|SDL_MixAudio|SDL_OpenURL|SDL_OutOfMemory|SDL_PauseAudioDevice|SDL_PauseAudioStreamDevice|SDL_PauseHaptic|SDL_PlayHapticRumble|SDL_PremultiplyAlpha|SDL_PremultiplySurfaceAlpha|SDL_PushEvent|SDL_PutAudioStreamData|SDL_RaiseWindow|SDL_ReadStorageFile|SDL_ReadSurfacePixel|SDL_ReadSurfacePixelFloat|SDL_RegisterApp|SDL_ReloadGamepadMappings|SDL_RemovePath|SDL_RemoveStoragePath|SDL_RemoveTimer|SDL_RenamePath|SDL_RenameStoragePath|SDL_RenderClear|SDL_RenderCoordinatesFromWindow|SDL_RenderCoordinatesToWindow|SDL_RenderFillRect|SDL_RenderFillRects|SDL_RenderGeometry|SDL_RenderGeometryRaw|SDL_RenderLine|SDL_RenderLines|SDL_RenderPoint|SDL_RenderPoints|SDL_RenderPresent|SDL_RenderRect|SDL_RenderRects|SDL_RenderTexture|SDL_RenderTexture9Grid|SDL_RenderTextureRotated|SDL_RenderTextureTiled|SDL_RequestAndroidPermission|SDL_RestoreWindow|SDL_ResumeAudioDevice|SDL_ResumeAudioStreamDevice|SDL_ResumeHaptic|SDL_RumbleGamepad|SDL_RumbleGamepadTriggers|SDL_RumbleJoystick|SDL_RumbleJoystickTriggers|SDL_RunHapticEffect|SDL_SaveBMP|SDL_SaveBMP_IO|SDL_SendAndroidMessage|SDL_SendGamepadEffect|SDL_SendJoystickEffect|SDL_SendJoystickVirtualSensorData|SDL_SetAppMetadata|SDL_SetAppMetadataProperty|SDL_SetAudioDeviceGain|SDL_SetAudioPostmixCallback|SDL_SetAudioStreamFormat|SDL_SetAudioStreamFrequencyRatio|SDL_SetAudioStreamGain|SDL_SetAudioStreamGetCallback|SDL_SetAudioStreamInputChannelMap|SDL_SetAudioStreamOutputChannelMap|SDL_SetAudioStreamPutCallback|SDL_SetBooleanProperty|SDL_SetClipboardData|SDL_SetClipboardText|SDL_SetCursor|SDL_SetFloatProperty|SDL_SetGamepadLED|SDL_SetGamepadMapping|SDL_SetGamepadPlayerIndex|SDL_SetGamepadSensorEnabled|SDL_SetHapticAutocenter|SDL_SetHapticGain|SDL_SetJoystickLED|SDL_SetJoystickPlayerIndex|SDL_SetJoystickVirtualAxis|SDL_SetJoystickVirtualBall|SDL_SetJoystickVirtualButton|SDL_SetJoystickVirtualHat|SDL_SetJoystickVirtualTouchpad|SDL_SetLinuxThreadPriority|SDL_SetLinuxThreadPriorityAndPolicy|SDL_SetLogPriorityPrefix|SDL_SetMemoryFunctions|SDL_SetNumberProperty|SDL_SetPaletteColors|SDL_SetPointerProperty|SDL_SetPointerPropertyWithCleanup|SDL_SetPrimarySelectionText|SDL_SetRenderClipRect|SDL_SetRenderColorScale|SDL_SetRenderDrawBlendMode|SDL_SetRenderDrawColor|SDL_SetRenderDrawColorFloat|SDL_SetRenderLogicalPresentation|SDL_SetRenderScale|SDL_SetRenderTarget|SDL_SetRenderVSync|SDL_SetRenderViewport|SDL_SetScancodeName|SDL_SetStringProperty|SDL_SetSurfaceAlphaMod|SDL_SetSurfaceBlendMode|SDL_SetSurfaceColorKey|SDL_SetSurfaceColorMod|SDL_SetSurfaceColorspace|SDL_SetSurfacePalette|SDL_SetSurfaceRLE|SDL_SetTLS|SDL_SetTextInputArea|SDL_SetTextureAlphaMod|SDL_SetTextureAlphaModFloat|SDL_SetTextureBlendMode|SDL_SetTextureColorMod|SDL_SetTextureColorModFloat|SDL_SetTextureScaleMode|SDL_SetThreadPriority|SDL_SetWindowAlwaysOnTop|SDL_SetWindowAspectRatio|SDL_SetWindowBordered|SDL_SetWindowFocusable|SDL_SetWindowFullscreen|SDL_SetWindowFullscreenMode|SDL_SetWindowHitTest|SDL_SetWindowIcon|SDL_SetWindowKeyboardGrab|SDL_SetWindowMaximumSize|SDL_SetWindowMinimumSize|SDL_SetWindowModalFor|SDL_SetWindowMouseGrab|SDL_SetWindowMouseRect|SDL_SetWindowOpacity|SDL_SetWindowPosition|SDL_SetWindowRelativeMouseMode|SDL_SetWindowResizable|SDL_SetWindowShape|SDL_SetWindowSize|SDL_SetWindowSurfaceVSync|SDL_SetWindowTitle|SDL_SetiOSAnimationCallback|SDL_ShowAndroidToast|SDL_ShowCursor|SDL_ShowMessageBox|SDL_ShowSimpleMessageBox|SDL_ShowWindow|SDL_ShowWindowSystemMenu|SDL_StartTextInput|SDL_StartTextInputWithProperties|SDL_StopHapticEffect|SDL_StopHapticEffects|SDL_StopHapticRumble|SDL_StopTextInput|SDL_SyncWindow|SDL_TimeToDateTime|SDL_TryLockMutex|SDL_TryLockRWLockForReading|SDL_TryLockRWLockForWriting|SDL_TryWaitSemaphore|SDL_UnlockAudioStream|SDL_UpdateHapticEffect|SDL_UpdateNVTexture|SDL_UpdateTexture|SDL_UpdateWindowSurface|SDL_UpdateWindowSurfaceRects|SDL_UpdateYUVTexture|SDL_Vulkan_CreateSurface|SDL_Vulkan_LoadLibrary|SDL_WaitConditionTimeout|SDL_WaitSemaphoreTimeout|SDL_WarpMouseGlobal|SDL_WriteStorageFile|SDL_WriteSurfacePixel|SDL_WriteSurfacePixelFloat)$"; @@ ( func( ... ) - == 0 | - func( + !func( ... ) - < 0 | - func( + !func( ... ) - != 0 | - func( + !func( ... ) - == -1 )
2024-08-23 03:33:49 +03:00
if (!result) {
Displays are now referenced by instance ID instead of index
2023-01-30 00:30:55 +03:00
return TEST_ABORTED;
}
testautomation: add video_setWindowCenteredOnDisplay test
2022-06-05 23:44:30 +03:00
tests: Check the usable bounds for centered window positioning SDL now centers windows according to the usable display bounds, so the window centering test requires an update for the new behavior.
2024-05-21 19:26:40 +03:00
result = SDL_GetDisplayUsableBounds(displays[1 % displayNum], &display1);
SDLTest_AssertPass("SDL_GetDisplayUsableBounds()");
Removed SDL_bool in favor of plain bool We require stdbool.h in the build environment, so we might as well use the plain bool type. If your environment doesn't have stdbool.h, this simple replacement will suffice: typedef signed char bool;
2024-09-18 17:52:28 +03:00
SDLTest_AssertCheck(result == true, "Verify return value; expected: true, got: %d", result);
Use SDL_bool instead an int return code in the SDL API Most SDL functions used to indicate success or failure using an int return code. These functions have been changed to return SDL_bool. Here is a coccinelle patch to change code that previously compared the return value to 0 and changes it to a boolean test: @ bool_return_type @ identifier func =~ "^(SDL_AddEventWatch|SDL_AddHintCallback|SDL_AddSurfaceAlternateImage|SDL_AddVulkanRenderSemaphores|SDL_BindAudioStream|SDL_BindAudioStreams|SDL_BlitSurface|SDL_BlitSurface9Grid|SDL_BlitSurfaceScaled|SDL_BlitSurfaceTiled|SDL_BlitSurfaceTiledWithScale|SDL_BlitSurfaceUnchecked|SDL_BlitSurfaceUncheckedScaled|SDL_CaptureMouse|SDL_ClearAudioStream|SDL_ClearClipboardData|SDL_ClearComposition|SDL_ClearError|SDL_ClearProperty|SDL_ClearSurface|SDL_CloseIO|SDL_CloseStorage|SDL_ConvertAudioSamples|SDL_ConvertEventToRenderCoordinates|SDL_ConvertPixels|SDL_ConvertPixelsAndColorspace|SDL_CopyFile|SDL_CopyProperties|SDL_CopyStorageFile|SDL_CreateDirectory|SDL_CreateStorageDirectory|SDL_CreateWindowAndRenderer|SDL_DateTimeToTime|SDL_DestroyWindowSurface|SDL_DetachVirtualJoystick|SDL_DisableScreenSaver|SDL_EnableScreenSaver|SDL_EnumerateDirectory|SDL_EnumerateProperties|SDL_EnumerateStorageDirectory|SDL_FillSurfaceRect|SDL_FillSurfaceRects|SDL_FlashWindow|SDL_FlipSurface|SDL_FlushAudioStream|SDL_FlushRenderer|SDL_GL_DestroyContext|SDL_GL_GetAttribute|SDL_GL_GetSwapInterval|SDL_GL_LoadLibrary|SDL_GL_MakeCurrent|SDL_GL_SetAttribute|SDL_GL_SetSwapInterval|SDL_GL_SwapWindow|SDL_GetAudioDeviceFormat|SDL_GetAudioStreamFormat|SDL_GetCameraFormat|SDL_GetClosestFullscreenDisplayMode|SDL_GetCurrentRenderOutputSize|SDL_GetCurrentTime|SDL_GetDXGIOutputInfo|SDL_GetDateTimeLocalePreferences|SDL_GetDisplayBounds|SDL_GetDisplayUsableBounds|SDL_GetGDKDefaultUser|SDL_GetGDKTaskQueue|SDL_GetGamepadSensorData|SDL_GetGamepadTouchpadFinger|SDL_GetHapticEffectStatus|SDL_GetJoystickBall|SDL_GetMasksForPixelFormat|SDL_GetPathInfo|SDL_GetRectUnion|SDL_GetRectUnionFloat|SDL_GetRenderClipRect|SDL_GetRenderColorScale|SDL_GetRenderDrawBlendMode|SDL_GetRenderDrawColor|SDL_GetRenderDrawColorFloat|SDL_GetRenderLogicalPresentation|SDL_GetRenderLogicalPresentationRect|SDL_GetRenderOutputSize|SDL_GetRenderSafeArea|SDL_GetRenderScale|SDL_GetRenderVSync|SDL_GetRenderViewport|SDL_GetSensorData|SDL_GetStorageFileSize|SDL_GetStoragePathInfo|SDL_GetSurfaceAlphaMod|SDL_GetSurfaceBlendMode|SDL_GetSurfaceClipRect|SDL_GetSurfaceColorKey|SDL_GetSurfaceColorMod|SDL_GetTextInputArea|SDL_GetTextureAlphaMod|SDL_GetTextureAlphaModFloat|SDL_GetTextureBlendMode|SDL_GetTextureColorMod|SDL_GetTextureColorModFloat|SDL_GetTextureScaleMode|SDL_GetTextureSize|SDL_GetWindowAspectRatio|SDL_GetWindowBordersSize|SDL_GetWindowMaximumSize|SDL_GetWindowMinimumSize|SDL_GetWindowPosition|SDL_GetWindowRelativeMouseMode|SDL_GetWindowSafeArea|SDL_GetWindowSize|SDL_GetWindowSizeInPixels|SDL_GetWindowSurfaceVSync|SDL_HideCursor|SDL_HideWindow|SDL_Init|SDL_InitHapticRumble|SDL_InitSubSystem|SDL_LoadWAV|SDL_LoadWAV_IO|SDL_LockAudioStream|SDL_LockProperties|SDL_LockSurface|SDL_LockTexture|SDL_LockTextureToSurface|SDL_MaximizeWindow|SDL_MinimizeWindow|SDL_MixAudio|SDL_OpenURL|SDL_OutOfMemory|SDL_PauseAudioDevice|SDL_PauseAudioStreamDevice|SDL_PauseHaptic|SDL_PlayHapticRumble|SDL_PremultiplyAlpha|SDL_PremultiplySurfaceAlpha|SDL_PushEvent|SDL_PutAudioStreamData|SDL_RaiseWindow|SDL_ReadStorageFile|SDL_ReadSurfacePixel|SDL_ReadSurfacePixelFloat|SDL_RegisterApp|SDL_ReloadGamepadMappings|SDL_RemovePath|SDL_RemoveStoragePath|SDL_RemoveTimer|SDL_RenamePath|SDL_RenameStoragePath|SDL_RenderClear|SDL_RenderCoordinatesFromWindow|SDL_RenderCoordinatesToWindow|SDL_RenderFillRect|SDL_RenderFillRects|SDL_RenderGeometry|SDL_RenderGeometryRaw|SDL_RenderLine|SDL_RenderLines|SDL_RenderPoint|SDL_RenderPoints|SDL_RenderPresent|SDL_RenderRect|SDL_RenderRects|SDL_RenderTexture|SDL_RenderTexture9Grid|SDL_RenderTextureRotated|SDL_RenderTextureTiled|SDL_RequestAndroidPermission|SDL_RestoreWindow|SDL_ResumeAudioDevice|SDL_ResumeAudioStreamDevice|SDL_ResumeHaptic|SDL_RumbleGamepad|SDL_RumbleGamepadTriggers|SDL_RumbleJoystick|SDL_RumbleJoystickTriggers|SDL_RunHapticEffect|SDL_SaveBMP|SDL_SaveBMP_IO|SDL_SendAndroidMessage|SDL_SendGamepadEffect|SDL_SendJoystickEffect|SDL_SendJoystickVirtualSensorData|SDL_SetAppMetadata|SDL_SetAppMetadataProperty|SDL_SetAudioDeviceGain|SDL_SetAudioPostmixCallback|SDL_SetAudioStreamFormat|SDL_SetAudioStreamFrequencyRatio|SDL_SetAudioStreamGain|SDL_SetAudioStreamGetCallback|SDL_SetAudioStreamInputChannelMap|SDL_SetAudioStreamOutputChannelMap|SDL_SetAudioStreamPutCallback|SDL_SetBooleanProperty|SDL_SetClipboardData|SDL_SetClipboardText|SDL_SetCursor|SDL_SetFloatProperty|SDL_SetGamepadLED|SDL_SetGamepadMapping|SDL_SetGamepadPlayerIndex|SDL_SetGamepadSensorEnabled|SDL_SetHapticAutocenter|SDL_SetHapticGain|SDL_SetJoystickLED|SDL_SetJoystickPlayerIndex|SDL_SetJoystickVirtualAxis|SDL_SetJoystickVirtualBall|SDL_SetJoystickVirtualButton|SDL_SetJoystickVirtualHat|SDL_SetJoystickVirtualTouchpad|SDL_SetLinuxThreadPriority|SDL_SetLinuxThreadPriorityAndPolicy|SDL_SetLogPriorityPrefix|SDL_SetMemoryFunctions|SDL_SetNumberProperty|SDL_SetPaletteColors|SDL_SetPointerProperty|SDL_SetPointerPropertyWithCleanup|SDL_SetPrimarySelectionText|SDL_SetRenderClipRect|SDL_SetRenderColorScale|SDL_SetRenderDrawBlendMode|SDL_SetRenderDrawColor|SDL_SetRenderDrawColorFloat|SDL_SetRenderLogicalPresentation|SDL_SetRenderScale|SDL_SetRenderTarget|SDL_SetRenderVSync|SDL_SetRenderViewport|SDL_SetScancodeName|SDL_SetStringProperty|SDL_SetSurfaceAlphaMod|SDL_SetSurfaceBlendMode|SDL_SetSurfaceColorKey|SDL_SetSurfaceColorMod|SDL_SetSurfaceColorspace|SDL_SetSurfacePalette|SDL_SetSurfaceRLE|SDL_SetTLS|SDL_SetTextInputArea|SDL_SetTextureAlphaMod|SDL_SetTextureAlphaModFloat|SDL_SetTextureBlendMode|SDL_SetTextureColorMod|SDL_SetTextureColorModFloat|SDL_SetTextureScaleMode|SDL_SetThreadPriority|SDL_SetWindowAlwaysOnTop|SDL_SetWindowAspectRatio|SDL_SetWindowBordered|SDL_SetWindowFocusable|SDL_SetWindowFullscreen|SDL_SetWindowFullscreenMode|SDL_SetWindowHitTest|SDL_SetWindowIcon|SDL_SetWindowKeyboardGrab|SDL_SetWindowMaximumSize|SDL_SetWindowMinimumSize|SDL_SetWindowModalFor|SDL_SetWindowMouseGrab|SDL_SetWindowMouseRect|SDL_SetWindowOpacity|SDL_SetWindowPosition|SDL_SetWindowRelativeMouseMode|SDL_SetWindowResizable|SDL_SetWindowShape|SDL_SetWindowSize|SDL_SetWindowSurfaceVSync|SDL_SetWindowTitle|SDL_SetiOSAnimationCallback|SDL_ShowAndroidToast|SDL_ShowCursor|SDL_ShowMessageBox|SDL_ShowSimpleMessageBox|SDL_ShowWindow|SDL_ShowWindowSystemMenu|SDL_StartTextInput|SDL_StartTextInputWithProperties|SDL_StopHapticEffect|SDL_StopHapticEffects|SDL_StopHapticRumble|SDL_StopTextInput|SDL_SyncWindow|SDL_TimeToDateTime|SDL_TryLockMutex|SDL_TryLockRWLockForReading|SDL_TryLockRWLockForWriting|SDL_TryWaitSemaphore|SDL_UnlockAudioStream|SDL_UpdateHapticEffect|SDL_UpdateNVTexture|SDL_UpdateTexture|SDL_UpdateWindowSurface|SDL_UpdateWindowSurfaceRects|SDL_UpdateYUVTexture|SDL_Vulkan_CreateSurface|SDL_Vulkan_LoadLibrary|SDL_WaitConditionTimeout|SDL_WaitSemaphoreTimeout|SDL_WarpMouseGlobal|SDL_WriteStorageFile|SDL_WriteSurfacePixel|SDL_WriteSurfacePixelFloat)$"; @@ ( func( ... ) - == 0 | - func( + !func( ... ) - < 0 | - func( + !func( ... ) - != 0 | - func( + !func( ... ) - == -1 )
2024-08-23 03:33:49 +03:00
if (!result) {
Displays are now referenced by instance ID instead of index
2023-01-30 00:30:55 +03:00
return TEST_ABORTED;
}
testautomation: add video_setWindowCenteredOnDisplay test
2022-06-05 23:44:30 +03:00
Displays are now referenced by instance ID instead of index
2023-01-30 00:30:55 +03:00
for (xVariation = 0; xVariation < 2; xVariation++) {
for (yVariation = 0; yVariation < 2; yVariation++) {
int currentX = 0, currentY = 0;
int currentW = 0, currentH = 0;
int expectedX = 0, expectedY = 0;
int currentDisplay;
int expectedDisplay;
tests: Fix window centering test with Wayland display scaling mode The usable fullscreen bounds need to be queried after window creation, as Wayland can send different usable bounds depending on the focused window's scaling mode.
2024-01-21 19:38:23 +03:00
SDL_Rect expectedDisplayRect, expectedFullscreenRect;
SDL_CreateWindowWithPosition() and SDL_CreateWindowFrom() have been replaced with SDL_CreateWindowWithProperties() This is a more general interface that can be extended in the future as needed.
2023-11-13 22:29:30 +03:00
SDL_PropertiesID props;
Displays are now referenced by instance ID instead of index
2023-01-30 00:30:55 +03:00
/* xVariation is the display we start on */
Fixed display ID check in video_setWindowCenteredOnDisplay
2023-03-16 02:12:54 +03:00
expectedDisplay = displays[xVariation % displayNum];
Displays are now referenced by instance ID instead of index
2023-01-30 00:30:55 +03:00
x = SDL_WINDOWPOS_CENTERED_DISPLAY(expectedDisplay);
y = SDL_WINDOWPOS_CENTERED_DISPLAY(expectedDisplay);
w = SDLTest_RandomIntegerInRange(640, 800);
h = SDLTest_RandomIntegerInRange(400, 600);
expectedDisplayRect = (xVariation == 0) ? display0 : display1;
expectedX = (expectedDisplayRect.x + ((expectedDisplayRect.w - w) / 2));
expectedY = (expectedDisplayRect.y + ((expectedDisplayRect.h - h) / 2));
SDL_CreateWindowWithPosition() and SDL_CreateWindowFrom() have been replaced with SDL_CreateWindowWithProperties() This is a more general interface that can be extended in the future as needed.
2023-11-13 22:29:30 +03:00
props = SDL_CreateProperties();
Shortened SDL property names
2024-01-28 00:00:37 +03:00
SDL_SetStringProperty(props, SDL_PROP_WINDOW_CREATE_TITLE_STRING, title);
SDL_SetNumberProperty(props, SDL_PROP_WINDOW_CREATE_X_NUMBER, x);
SDL_SetNumberProperty(props, SDL_PROP_WINDOW_CREATE_Y_NUMBER, y);
SDL_SetNumberProperty(props, SDL_PROP_WINDOW_CREATE_WIDTH_NUMBER, w);
SDL_SetNumberProperty(props, SDL_PROP_WINDOW_CREATE_HEIGHT_NUMBER, h);
Removed SDL_bool in favor of plain bool We require stdbool.h in the build environment, so we might as well use the plain bool type. If your environment doesn't have stdbool.h, this simple replacement will suffice: typedef signed char bool;
2024-09-18 17:52:28 +03:00
SDL_SetBooleanProperty(props, SDL_PROP_WINDOW_CREATE_BORDERLESS_BOOLEAN, true);
SDL_CreateWindowWithPosition() and SDL_CreateWindowFrom() have been replaced with SDL_CreateWindowWithProperties() This is a more general interface that can be extended in the future as needed.
2023-11-13 22:29:30 +03:00
window = SDL_CreateWindowWithProperties(props);
SDL_DestroyProperties(props);
Displays are now referenced by instance ID instead of index
2023-01-30 00:30:55 +03:00
SDLTest_AssertPass("Call to SDL_CreateWindow('Title',%d,%d,%d,%d,SHOWN)", x, y, w, h);
Clean up any renderer in SDL_DestroyWindowSurface() Also added an automated test to verify window surface functionality. Fixes https://github.com/libsdl-org/SDL/issues/10133
2024-06-29 12:34:17 +03:00
SDLTest_AssertCheck(window != NULL, "Validate that returned window is not NULL");
Displays are now referenced by instance ID instead of index
2023-01-30 00:30:55 +03:00
tests: Fix automated window grab and positioning tests under Wayland/XWayland When using the Wayland video driver or X11 under XWayland, create a renderer and present a frame as part of window creation, as Wayland requires that a frame be presented for the window to be fully mapped and displayed onscreen. This fixes the grab and expected window size tests. This also disables the window positioning tests when running under the Wayland driver, as Wayland does not allow application windows to position themselves in the desktop space, which renders the tests unreliable and subject to spurious failure.
2023-07-23 18:25:15 +03:00
/* Wayland windows require that a frame be presented before they are fully mapped and visible onscreen. */
if (video_driver_is_wayland) {
Removed SDL_RendererFlags The flags parameter has been removed from SDL_CreateRenderer() and SDL_RENDERER_PRESENTVSYNC has been replaced with SDL_PROP_RENDERER_CREATE_PRESENT_VSYNC_NUMBER during window creation and SDL_PROP_RENDERER_VSYNC_NUMBER after renderer creation. SDL_SetRenderVSync() now takes additional values besides 0 and 1. The maximum texture size has been removed from SDL_RendererInfo, replaced with SDL_PROP_RENDERER_MAX_TEXTURE_SIZE_NUMBER.
2024-05-13 20:31:55 +03:00
SDL_Renderer *renderer = SDL_CreateRenderer(window, NULL);
tests: Fix automated window grab and positioning tests under Wayland/XWayland When using the Wayland video driver or X11 under XWayland, create a renderer and present a frame as part of window creation, as Wayland requires that a frame be presented for the window to be fully mapped and displayed onscreen. This fixes the grab and expected window size tests. This also disables the window positioning tests when running under the Wayland driver, as Wayland does not allow application windows to position themselves in the desktop space, which renders the tests unreliable and subject to spurious failure.
2023-07-23 18:25:15 +03:00
if (renderer) {
SDL_SetRenderDrawColor(renderer, 0x00, 0x00, 0x00, 0xFF);
SDL_RenderClear(renderer);
SDL_RenderPresent(renderer);
/* Some desktops don't display the window immediately after presentation,
* so delay to give the window time to actually appear on the desktop.
*/
SDL_Delay(100);
} else {
SDLTest_Log("Unable to create a renderer, tests may fail under Wayland");
}
}
Displays are now referenced by instance ID instead of index
2023-01-30 00:30:55 +03:00
/* Check the window is centered on the requested display */
currentDisplay = SDL_GetDisplayForWindow(window);
SDL_GetWindowSize(window, &currentW, &currentH);
SDL_GetWindowPosition(window, &currentX, &currentY);
testautomation: skip window resize and repositioning tests on Emscripten
2024-08-18 00:20:41 +03:00
if (video_driver_is_wayland) {
SDLTest_Log("Skipping display ID validation: %s driver does not support window positioning", video_driver);
} else {
tests: Fix automated window grab and positioning tests under Wayland/XWayland When using the Wayland video driver or X11 under XWayland, create a renderer and present a frame as part of window creation, as Wayland requires that a frame be presented for the window to be fully mapped and displayed onscreen. This fixes the grab and expected window size tests. This also disables the window positioning tests when running under the Wayland driver, as Wayland does not allow application windows to position themselves in the desktop space, which renders the tests unreliable and subject to spurious failure.
2023-07-23 18:25:15 +03:00
SDLTest_AssertCheck(currentDisplay == expectedDisplay, "Validate display ID (current: %d, expected: %d)", currentDisplay, expectedDisplay);
testautomation: skip window resize and repositioning tests on Emscripten
2024-08-18 00:20:41 +03:00
}
if (video_driver_is_emscripten) {
SDLTest_Log("Skipping window size validation: %s driver does not support window resizing", video_driver);
tests: Fix automated window grab and positioning tests under Wayland/XWayland When using the Wayland video driver or X11 under XWayland, create a renderer and present a frame as part of window creation, as Wayland requires that a frame be presented for the window to be fully mapped and displayed onscreen. This fixes the grab and expected window size tests. This also disables the window positioning tests when running under the Wayland driver, as Wayland does not allow application windows to position themselves in the desktop space, which renders the tests unreliable and subject to spurious failure.
2023-07-23 18:25:15 +03:00
} else {
testautomation: skip window resize and repositioning tests on Emscripten
2024-08-18 00:20:41 +03:00
SDLTest_AssertCheck(currentW == w, "Validate width (current: %d, expected: %d)", currentW, w);
SDLTest_AssertCheck(currentH == h, "Validate height (current: %d, expected: %d)", currentH, h);
tests: Fix automated window grab and positioning tests under Wayland/XWayland When using the Wayland video driver or X11 under XWayland, create a renderer and present a frame as part of window creation, as Wayland requires that a frame be presented for the window to be fully mapped and displayed onscreen. This fixes the grab and expected window size tests. This also disables the window positioning tests when running under the Wayland driver, as Wayland does not allow application windows to position themselves in the desktop space, which renders the tests unreliable and subject to spurious failure.
2023-07-23 18:25:15 +03:00
}
testautomation: skip window resize and repositioning tests on Emscripten
2024-08-18 00:20:41 +03:00
if (video_driver_is_emscripten || video_driver_is_wayland) {
SDLTest_Log("Skipping window position validation: %s driver does not support window positioning", video_driver);
} else {
tests: Fix automated window grab and positioning tests under Wayland/XWayland When using the Wayland video driver or X11 under XWayland, create a renderer and present a frame as part of window creation, as Wayland requires that a frame be presented for the window to be fully mapped and displayed onscreen. This fixes the grab and expected window size tests. This also disables the window positioning tests when running under the Wayland driver, as Wayland does not allow application windows to position themselves in the desktop space, which renders the tests unreliable and subject to spurious failure.
2023-07-23 18:25:15 +03:00
SDLTest_AssertCheck(currentX == expectedX, "Validate x (current: %d, expected: %d)", currentX, expectedX);
SDLTest_AssertCheck(currentY == expectedY, "Validate y (current: %d, expected: %d)", currentY, expectedY);
}
Displays are now referenced by instance ID instead of index
2023-01-30 00:30:55 +03:00
/* Enter fullscreen desktop */
tests: Fix automated window grab and positioning tests under Wayland/XWayland When using the Wayland video driver or X11 under XWayland, create a renderer and present a frame as part of window creation, as Wayland requires that a frame be presented for the window to be fully mapped and displayed onscreen. This fixes the grab and expected window size tests. This also disables the window positioning tests when running under the Wayland driver, as Wayland does not allow application windows to position themselves in the desktop space, which renders the tests unreliable and subject to spurious failure.
2023-07-23 18:25:15 +03:00
SDL_SetWindowPosition(window, x, y);
Removed SDL_bool in favor of plain bool We require stdbool.h in the build environment, so we might as well use the plain bool type. If your environment doesn't have stdbool.h, this simple replacement will suffice: typedef signed char bool;
2024-09-18 17:52:28 +03:00
result = SDL_SetWindowFullscreen(window, true);
SDLTest_AssertCheck(result == true, "Verify return value; expected: true, got: %d", result);
Displays are now referenced by instance ID instead of index
2023-01-30 00:30:55 +03:00
video: Implement asynchronous windowing SDL window size, state, and position functions have been considered immediate, with their effects assuming to have taken effect upon successful return of the function. However, several windowing systems handle these requests asynchronously, resulting in the functions blocking until the changes have taken effect, potentially for long periods of time. Additionally, some windowing systems treat these as requests, and can potentially deny or fulfill the request in a manner differently than the application expects, such as not allowing a window to be positioned or sized beyond desktop borders, prohibiting fullscreen, and so on. With these changes, applications can make requests of the window manager that do not block, with the understanding that an associated event will be sent if the request is fulfilled. Currently, size, position, maximize, minimize, and fullscreen calls are handled as asynchronous requests, with events being returned if the request is honored. If the application requires that the change take effect immediately, it can call the new SDL_SyncWindow function, which will attempt to block until the request is fulfilled, or some arbitrary timeout period elapses, the duration of which depends not only on the windowing system, but on the operation requested as well (e.g. a 100ms timeout is fine for most X11 events, but maximizing a window can take considerably longer for some reason). There is also a new hint 'SDL_VIDEO_SYNC_ALL_WINDOW_OPS' that will mimic the old behavior by synchronizing after every window operation with, again, the understanding that using this may result in the associated calls blocking for a relatively long period. The deferred model also results in the window size and position getters not reporting false coordinates anymore, as they only forward what the window manager reports vs allowing applications to set arbitrary values, and fullscreen enter/leave events that were initiated via the window manager update the window state appropriately, where they didn't before. Care was taken to ensure that order of operations is maintained, and that requests are not ignored or dropped. This does require some implicit internal synchronization in the various backends if many requests are made in a short period, as some state and behavior depends on other bits of state that need to be known at that particular point in time, but this isn't something that typical applications will hit, unless they are sending a lot of window state in a short time as the tests do. The automated tests developed to test the previous behavior also resulted in previously undefined behavior being defined and normalized across platforms, particularly when it comes to the sizing and positioning of windows when they are in a fixed-size state, such as maximized or fullscreen. Size and position requests made when the window is not in a movable or resizable state will be deferred until it can be applied, so no requests are lost. These changes fix another long-standing issue with renderers recreating maximized windows, where the original non-maximized size was lost, resulting in the window being restored to the wrong size. All automated video tests pass across all platforms. Overall, the "make a request/get an event" model better reflects how most windowing systems work, and some backends avoid spending significant time blocking while waiting for operations to complete.
2023-10-25 22:09:55 +03:00
result = SDL_SyncWindow(window);
SDLTest_AssertPass("SDL_SyncWindow()");
Removed SDL_bool in favor of plain bool We require stdbool.h in the build environment, so we might as well use the plain bool type. If your environment doesn't have stdbool.h, this simple replacement will suffice: typedef signed char bool;
2024-09-18 17:52:28 +03:00
SDLTest_AssertCheck(result == true, "Verify return value; expected: true, got: %d", result);
video: Implement asynchronous windowing SDL window size, state, and position functions have been considered immediate, with their effects assuming to have taken effect upon successful return of the function. However, several windowing systems handle these requests asynchronously, resulting in the functions blocking until the changes have taken effect, potentially for long periods of time. Additionally, some windowing systems treat these as requests, and can potentially deny or fulfill the request in a manner differently than the application expects, such as not allowing a window to be positioned or sized beyond desktop borders, prohibiting fullscreen, and so on. With these changes, applications can make requests of the window manager that do not block, with the understanding that an associated event will be sent if the request is fulfilled. Currently, size, position, maximize, minimize, and fullscreen calls are handled as asynchronous requests, with events being returned if the request is honored. If the application requires that the change take effect immediately, it can call the new SDL_SyncWindow function, which will attempt to block until the request is fulfilled, or some arbitrary timeout period elapses, the duration of which depends not only on the windowing system, but on the operation requested as well (e.g. a 100ms timeout is fine for most X11 events, but maximizing a window can take considerably longer for some reason). There is also a new hint 'SDL_VIDEO_SYNC_ALL_WINDOW_OPS' that will mimic the old behavior by synchronizing after every window operation with, again, the understanding that using this may result in the associated calls blocking for a relatively long period. The deferred model also results in the window size and position getters not reporting false coordinates anymore, as they only forward what the window manager reports vs allowing applications to set arbitrary values, and fullscreen enter/leave events that were initiated via the window manager update the window state appropriately, where they didn't before. Care was taken to ensure that order of operations is maintained, and that requests are not ignored or dropped. This does require some implicit internal synchronization in the various backends if many requests are made in a short period, as some state and behavior depends on other bits of state that need to be known at that particular point in time, but this isn't something that typical applications will hit, unless they are sending a lot of window state in a short time as the tests do. The automated tests developed to test the previous behavior also resulted in previously undefined behavior being defined and normalized across platforms, particularly when it comes to the sizing and positioning of windows when they are in a fixed-size state, such as maximized or fullscreen. Size and position requests made when the window is not in a movable or resizable state will be deferred until it can be applied, so no requests are lost. These changes fix another long-standing issue with renderers recreating maximized windows, where the original non-maximized size was lost, resulting in the window being restored to the wrong size. All automated video tests pass across all platforms. Overall, the "make a request/get an event" model better reflects how most windowing systems work, and some backends avoid spending significant time blocking while waiting for operations to complete.
2023-10-25 22:09:55 +03:00
Displays are now referenced by instance ID instead of index
2023-01-30 00:30:55 +03:00
/* Check we are filling the full display */
currentDisplay = SDL_GetDisplayForWindow(window);
SDL_GetWindowSize(window, &currentW, &currentH);
SDL_GetWindowPosition(window, &currentX, &currentY);
tests: Fix window centering test with Wayland display scaling mode The usable fullscreen bounds need to be queried after window creation, as Wayland can send different usable bounds depending on the focused window's scaling mode.
2024-01-21 19:38:23 +03:00
/* Get the expected fullscreen rect.
* This needs to be queried after window creation and positioning as some drivers can alter the
* usable bounds based on the window scaling mode.
*/
result = SDL_GetDisplayBounds(expectedDisplay, &expectedFullscreenRect);
SDLTest_AssertPass("SDL_GetDisplayBounds()");
Removed SDL_bool in favor of plain bool We require stdbool.h in the build environment, so we might as well use the plain bool type. If your environment doesn't have stdbool.h, this simple replacement will suffice: typedef signed char bool;
2024-09-18 17:52:28 +03:00
SDLTest_AssertCheck(result == true, "Verify return value; expected: true, got: %d", result);
tests: Fix window centering test with Wayland display scaling mode The usable fullscreen bounds need to be queried after window creation, as Wayland can send different usable bounds depending on the focused window's scaling mode.
2024-01-21 19:38:23 +03:00
testautomation: skip window resize and repositioning tests on Emscripten
2024-08-18 00:20:41 +03:00
if (video_driver_is_wayland) {
SDLTest_Log("Skipping display ID validation: Wayland driver does not support window positioning");
} else {
tests: Fix automated window grab and positioning tests under Wayland/XWayland When using the Wayland video driver or X11 under XWayland, create a renderer and present a frame as part of window creation, as Wayland requires that a frame be presented for the window to be fully mapped and displayed onscreen. This fixes the grab and expected window size tests. This also disables the window positioning tests when running under the Wayland driver, as Wayland does not allow application windows to position themselves in the desktop space, which renders the tests unreliable and subject to spurious failure.
2023-07-23 18:25:15 +03:00
SDLTest_AssertCheck(currentDisplay == expectedDisplay, "Validate display ID (current: %d, expected: %d)", currentDisplay, expectedDisplay);
testautomation: skip window resize and repositioning tests on Emscripten
2024-08-18 00:20:41 +03:00
}
if (video_driver_is_emscripten) {
SDLTest_Log("Skipping window position validation: %s driver does not support window resizing", video_driver);
tests: Fix automated window grab and positioning tests under Wayland/XWayland When using the Wayland video driver or X11 under XWayland, create a renderer and present a frame as part of window creation, as Wayland requires that a frame be presented for the window to be fully mapped and displayed onscreen. This fixes the grab and expected window size tests. This also disables the window positioning tests when running under the Wayland driver, as Wayland does not allow application windows to position themselves in the desktop space, which renders the tests unreliable and subject to spurious failure.
2023-07-23 18:25:15 +03:00
} else {
testautomation: skip window resize and repositioning tests on Emscripten
2024-08-18 00:20:41 +03:00
SDLTest_AssertCheck(currentW == expectedFullscreenRect.w, "Validate width (current: %d, expected: %d)", currentW, expectedFullscreenRect.w);
SDLTest_AssertCheck(currentH == expectedFullscreenRect.h, "Validate height (current: %d, expected: %d)", currentH, expectedFullscreenRect.h);
tests: Fix automated window grab and positioning tests under Wayland/XWayland When using the Wayland video driver or X11 under XWayland, create a renderer and present a frame as part of window creation, as Wayland requires that a frame be presented for the window to be fully mapped and displayed onscreen. This fixes the grab and expected window size tests. This also disables the window positioning tests when running under the Wayland driver, as Wayland does not allow application windows to position themselves in the desktop space, which renders the tests unreliable and subject to spurious failure.
2023-07-23 18:25:15 +03:00
}
testautomation: skip window resize and repositioning tests on Emscripten
2024-08-18 00:20:41 +03:00
if (video_driver_is_emscripten || video_driver_is_wayland) {
SDLTest_Log("Skipping window position validation: %s driver does not support window positioning", video_driver);
} else {
tests: Fix window centering test with Wayland display scaling mode The usable fullscreen bounds need to be queried after window creation, as Wayland can send different usable bounds depending on the focused window's scaling mode.
2024-01-21 19:38:23 +03:00
SDLTest_AssertCheck(currentX == expectedFullscreenRect.x, "Validate x (current: %d, expected: %d)", currentX, expectedFullscreenRect.x);
SDLTest_AssertCheck(currentY == expectedFullscreenRect.y, "Validate y (current: %d, expected: %d)", currentY, expectedFullscreenRect.y);
tests: Fix automated window grab and positioning tests under Wayland/XWayland When using the Wayland video driver or X11 under XWayland, create a renderer and present a frame as part of window creation, as Wayland requires that a frame be presented for the window to be fully mapped and displayed onscreen. This fixes the grab and expected window size tests. This also disables the window positioning tests when running under the Wayland driver, as Wayland does not allow application windows to position themselves in the desktop space, which renders the tests unreliable and subject to spurious failure.
2023-07-23 18:25:15 +03:00
}
Displays are now referenced by instance ID instead of index
2023-01-30 00:30:55 +03:00
/* Leave fullscreen desktop */
Removed SDL_bool in favor of plain bool We require stdbool.h in the build environment, so we might as well use the plain bool type. If your environment doesn't have stdbool.h, this simple replacement will suffice: typedef signed char bool;
2024-09-18 17:52:28 +03:00
result = SDL_SetWindowFullscreen(window, false);
SDLTest_AssertCheck(result == true, "Verify return value; expected: true, got: %d", result);
Displays are now referenced by instance ID instead of index
2023-01-30 00:30:55 +03:00
video: Implement asynchronous windowing SDL window size, state, and position functions have been considered immediate, with their effects assuming to have taken effect upon successful return of the function. However, several windowing systems handle these requests asynchronously, resulting in the functions blocking until the changes have taken effect, potentially for long periods of time. Additionally, some windowing systems treat these as requests, and can potentially deny or fulfill the request in a manner differently than the application expects, such as not allowing a window to be positioned or sized beyond desktop borders, prohibiting fullscreen, and so on. With these changes, applications can make requests of the window manager that do not block, with the understanding that an associated event will be sent if the request is fulfilled. Currently, size, position, maximize, minimize, and fullscreen calls are handled as asynchronous requests, with events being returned if the request is honored. If the application requires that the change take effect immediately, it can call the new SDL_SyncWindow function, which will attempt to block until the request is fulfilled, or some arbitrary timeout period elapses, the duration of which depends not only on the windowing system, but on the operation requested as well (e.g. a 100ms timeout is fine for most X11 events, but maximizing a window can take considerably longer for some reason). There is also a new hint 'SDL_VIDEO_SYNC_ALL_WINDOW_OPS' that will mimic the old behavior by synchronizing after every window operation with, again, the understanding that using this may result in the associated calls blocking for a relatively long period. The deferred model also results in the window size and position getters not reporting false coordinates anymore, as they only forward what the window manager reports vs allowing applications to set arbitrary values, and fullscreen enter/leave events that were initiated via the window manager update the window state appropriately, where they didn't before. Care was taken to ensure that order of operations is maintained, and that requests are not ignored or dropped. This does require some implicit internal synchronization in the various backends if many requests are made in a short period, as some state and behavior depends on other bits of state that need to be known at that particular point in time, but this isn't something that typical applications will hit, unless they are sending a lot of window state in a short time as the tests do. The automated tests developed to test the previous behavior also resulted in previously undefined behavior being defined and normalized across platforms, particularly when it comes to the sizing and positioning of windows when they are in a fixed-size state, such as maximized or fullscreen. Size and position requests made when the window is not in a movable or resizable state will be deferred until it can be applied, so no requests are lost. These changes fix another long-standing issue with renderers recreating maximized windows, where the original non-maximized size was lost, resulting in the window being restored to the wrong size. All automated video tests pass across all platforms. Overall, the "make a request/get an event" model better reflects how most windowing systems work, and some backends avoid spending significant time blocking while waiting for operations to complete.
2023-10-25 22:09:55 +03:00
result = SDL_SyncWindow(window);
SDLTest_AssertPass("SDL_SyncWindow()");
Removed SDL_bool in favor of plain bool We require stdbool.h in the build environment, so we might as well use the plain bool type. If your environment doesn't have stdbool.h, this simple replacement will suffice: typedef signed char bool;
2024-09-18 17:52:28 +03:00
SDLTest_AssertCheck(result == true, "Verify return value; expected: true, got: %d", result);
video: Implement asynchronous windowing SDL window size, state, and position functions have been considered immediate, with their effects assuming to have taken effect upon successful return of the function. However, several windowing systems handle these requests asynchronously, resulting in the functions blocking until the changes have taken effect, potentially for long periods of time. Additionally, some windowing systems treat these as requests, and can potentially deny or fulfill the request in a manner differently than the application expects, such as not allowing a window to be positioned or sized beyond desktop borders, prohibiting fullscreen, and so on. With these changes, applications can make requests of the window manager that do not block, with the understanding that an associated event will be sent if the request is fulfilled. Currently, size, position, maximize, minimize, and fullscreen calls are handled as asynchronous requests, with events being returned if the request is honored. If the application requires that the change take effect immediately, it can call the new SDL_SyncWindow function, which will attempt to block until the request is fulfilled, or some arbitrary timeout period elapses, the duration of which depends not only on the windowing system, but on the operation requested as well (e.g. a 100ms timeout is fine for most X11 events, but maximizing a window can take considerably longer for some reason). There is also a new hint 'SDL_VIDEO_SYNC_ALL_WINDOW_OPS' that will mimic the old behavior by synchronizing after every window operation with, again, the understanding that using this may result in the associated calls blocking for a relatively long period. The deferred model also results in the window size and position getters not reporting false coordinates anymore, as they only forward what the window manager reports vs allowing applications to set arbitrary values, and fullscreen enter/leave events that were initiated via the window manager update the window state appropriately, where they didn't before. Care was taken to ensure that order of operations is maintained, and that requests are not ignored or dropped. This does require some implicit internal synchronization in the various backends if many requests are made in a short period, as some state and behavior depends on other bits of state that need to be known at that particular point in time, but this isn't something that typical applications will hit, unless they are sending a lot of window state in a short time as the tests do. The automated tests developed to test the previous behavior also resulted in previously undefined behavior being defined and normalized across platforms, particularly when it comes to the sizing and positioning of windows when they are in a fixed-size state, such as maximized or fullscreen. Size and position requests made when the window is not in a movable or resizable state will be deferred until it can be applied, so no requests are lost. These changes fix another long-standing issue with renderers recreating maximized windows, where the original non-maximized size was lost, resulting in the window being restored to the wrong size. All automated video tests pass across all platforms. Overall, the "make a request/get an event" model better reflects how most windowing systems work, and some backends avoid spending significant time blocking while waiting for operations to complete.
2023-10-25 22:09:55 +03:00
Displays are now referenced by instance ID instead of index
2023-01-30 00:30:55 +03:00
/* Check window was restored correctly */
currentDisplay = SDL_GetDisplayForWindow(window);
SDL_GetWindowSize(window, &currentW, &currentH);
SDL_GetWindowPosition(window, &currentX, &currentY);
testautomation: skip window resize and repositioning tests on Emscripten
2024-08-18 00:20:41 +03:00
if (video_driver_is_wayland) {
SDLTest_Log("Skipping display ID validation: %s driver does not support window positioning", video_driver);
} else {
tests: Fix automated window grab and positioning tests under Wayland/XWayland When using the Wayland video driver or X11 under XWayland, create a renderer and present a frame as part of window creation, as Wayland requires that a frame be presented for the window to be fully mapped and displayed onscreen. This fixes the grab and expected window size tests. This also disables the window positioning tests when running under the Wayland driver, as Wayland does not allow application windows to position themselves in the desktop space, which renders the tests unreliable and subject to spurious failure.
2023-07-23 18:25:15 +03:00
SDLTest_AssertCheck(currentDisplay == expectedDisplay, "Validate display index (current: %d, expected: %d)", currentDisplay, expectedDisplay);
testautomation: skip window resize and repositioning tests on Emscripten
2024-08-18 00:20:41 +03:00
}
if (video_driver_is_emscripten) {
SDLTest_Log("Skipping window position validation: %s driver does not support window resizing", video_driver);
tests: Fix automated window grab and positioning tests under Wayland/XWayland When using the Wayland video driver or X11 under XWayland, create a renderer and present a frame as part of window creation, as Wayland requires that a frame be presented for the window to be fully mapped and displayed onscreen. This fixes the grab and expected window size tests. This also disables the window positioning tests when running under the Wayland driver, as Wayland does not allow application windows to position themselves in the desktop space, which renders the tests unreliable and subject to spurious failure.
2023-07-23 18:25:15 +03:00
} else {
testautomation: skip window resize and repositioning tests on Emscripten
2024-08-18 00:20:41 +03:00
SDLTest_AssertCheck(currentW == w, "Validate width (current: %d, expected: %d)", currentW, w);
SDLTest_AssertCheck(currentH == h, "Validate height (current: %d, expected: %d)", currentH, h);
tests: Fix automated window grab and positioning tests under Wayland/XWayland When using the Wayland video driver or X11 under XWayland, create a renderer and present a frame as part of window creation, as Wayland requires that a frame be presented for the window to be fully mapped and displayed onscreen. This fixes the grab and expected window size tests. This also disables the window positioning tests when running under the Wayland driver, as Wayland does not allow application windows to position themselves in the desktop space, which renders the tests unreliable and subject to spurious failure.
2023-07-23 18:25:15 +03:00
}
testautomation: skip window resize and repositioning tests on Emscripten
2024-08-18 00:20:41 +03:00
if (video_driver_is_wayland) {
SDLTest_Log("Skipping window position validation: %s driver does not support window positioning", video_driver);
} else {
tests: Fix automated window grab and positioning tests under Wayland/XWayland When using the Wayland video driver or X11 under XWayland, create a renderer and present a frame as part of window creation, as Wayland requires that a frame be presented for the window to be fully mapped and displayed onscreen. This fixes the grab and expected window size tests. This also disables the window positioning tests when running under the Wayland driver, as Wayland does not allow application windows to position themselves in the desktop space, which renders the tests unreliable and subject to spurious failure.
2023-07-23 18:25:15 +03:00
SDLTest_AssertCheck(currentX == expectedX, "Validate x (current: %d, expected: %d)", currentX, expectedX);
SDLTest_AssertCheck(currentY == expectedY, "Validate y (current: %d, expected: %d)", currentY, expectedY);
}
Displays are now referenced by instance ID instead of index
2023-01-30 00:30:55 +03:00
/* Center on display yVariation, and check window properties */
Fixed display ID check in video_setWindowCenteredOnDisplay
2023-03-16 02:12:54 +03:00
expectedDisplay = displays[yVariation % displayNum];
Displays are now referenced by instance ID instead of index
2023-01-30 00:30:55 +03:00
x = SDL_WINDOWPOS_CENTERED_DISPLAY(expectedDisplay);
y = SDL_WINDOWPOS_CENTERED_DISPLAY(expectedDisplay);
Fixed display ID check in video_setWindowCenteredOnDisplay
2023-03-16 02:12:54 +03:00
expectedDisplayRect = (yVariation == 0) ? display0 : display1;
Displays are now referenced by instance ID instead of index
2023-01-30 00:30:55 +03:00
expectedX = (expectedDisplayRect.x + ((expectedDisplayRect.w - w) / 2));
expectedY = (expectedDisplayRect.y + ((expectedDisplayRect.h - h) / 2));
SDL_SetWindowPosition(window, x, y);
video: Implement asynchronous windowing SDL window size, state, and position functions have been considered immediate, with their effects assuming to have taken effect upon successful return of the function. However, several windowing systems handle these requests asynchronously, resulting in the functions blocking until the changes have taken effect, potentially for long periods of time. Additionally, some windowing systems treat these as requests, and can potentially deny or fulfill the request in a manner differently than the application expects, such as not allowing a window to be positioned or sized beyond desktop borders, prohibiting fullscreen, and so on. With these changes, applications can make requests of the window manager that do not block, with the understanding that an associated event will be sent if the request is fulfilled. Currently, size, position, maximize, minimize, and fullscreen calls are handled as asynchronous requests, with events being returned if the request is honored. If the application requires that the change take effect immediately, it can call the new SDL_SyncWindow function, which will attempt to block until the request is fulfilled, or some arbitrary timeout period elapses, the duration of which depends not only on the windowing system, but on the operation requested as well (e.g. a 100ms timeout is fine for most X11 events, but maximizing a window can take considerably longer for some reason). There is also a new hint 'SDL_VIDEO_SYNC_ALL_WINDOW_OPS' that will mimic the old behavior by synchronizing after every window operation with, again, the understanding that using this may result in the associated calls blocking for a relatively long period. The deferred model also results in the window size and position getters not reporting false coordinates anymore, as they only forward what the window manager reports vs allowing applications to set arbitrary values, and fullscreen enter/leave events that were initiated via the window manager update the window state appropriately, where they didn't before. Care was taken to ensure that order of operations is maintained, and that requests are not ignored or dropped. This does require some implicit internal synchronization in the various backends if many requests are made in a short period, as some state and behavior depends on other bits of state that need to be known at that particular point in time, but this isn't something that typical applications will hit, unless they are sending a lot of window state in a short time as the tests do. The automated tests developed to test the previous behavior also resulted in previously undefined behavior being defined and normalized across platforms, particularly when it comes to the sizing and positioning of windows when they are in a fixed-size state, such as maximized or fullscreen. Size and position requests made when the window is not in a movable or resizable state will be deferred until it can be applied, so no requests are lost. These changes fix another long-standing issue with renderers recreating maximized windows, where the original non-maximized size was lost, resulting in the window being restored to the wrong size. All automated video tests pass across all platforms. Overall, the "make a request/get an event" model better reflects how most windowing systems work, and some backends avoid spending significant time blocking while waiting for operations to complete.
2023-10-25 22:09:55 +03:00
result = SDL_SyncWindow(window);
SDLTest_AssertPass("SDL_SyncWindow()");
Removed SDL_bool in favor of plain bool We require stdbool.h in the build environment, so we might as well use the plain bool type. If your environment doesn't have stdbool.h, this simple replacement will suffice: typedef signed char bool;
2024-09-18 17:52:28 +03:00
SDLTest_AssertCheck(result == true, "Verify return value; expected: true, got: %d", result);
video: Implement asynchronous windowing SDL window size, state, and position functions have been considered immediate, with their effects assuming to have taken effect upon successful return of the function. However, several windowing systems handle these requests asynchronously, resulting in the functions blocking until the changes have taken effect, potentially for long periods of time. Additionally, some windowing systems treat these as requests, and can potentially deny or fulfill the request in a manner differently than the application expects, such as not allowing a window to be positioned or sized beyond desktop borders, prohibiting fullscreen, and so on. With these changes, applications can make requests of the window manager that do not block, with the understanding that an associated event will be sent if the request is fulfilled. Currently, size, position, maximize, minimize, and fullscreen calls are handled as asynchronous requests, with events being returned if the request is honored. If the application requires that the change take effect immediately, it can call the new SDL_SyncWindow function, which will attempt to block until the request is fulfilled, or some arbitrary timeout period elapses, the duration of which depends not only on the windowing system, but on the operation requested as well (e.g. a 100ms timeout is fine for most X11 events, but maximizing a window can take considerably longer for some reason). There is also a new hint 'SDL_VIDEO_SYNC_ALL_WINDOW_OPS' that will mimic the old behavior by synchronizing after every window operation with, again, the understanding that using this may result in the associated calls blocking for a relatively long period. The deferred model also results in the window size and position getters not reporting false coordinates anymore, as they only forward what the window manager reports vs allowing applications to set arbitrary values, and fullscreen enter/leave events that were initiated via the window manager update the window state appropriately, where they didn't before. Care was taken to ensure that order of operations is maintained, and that requests are not ignored or dropped. This does require some implicit internal synchronization in the various backends if many requests are made in a short period, as some state and behavior depends on other bits of state that need to be known at that particular point in time, but this isn't something that typical applications will hit, unless they are sending a lot of window state in a short time as the tests do. The automated tests developed to test the previous behavior also resulted in previously undefined behavior being defined and normalized across platforms, particularly when it comes to the sizing and positioning of windows when they are in a fixed-size state, such as maximized or fullscreen. Size and position requests made when the window is not in a movable or resizable state will be deferred until it can be applied, so no requests are lost. These changes fix another long-standing issue with renderers recreating maximized windows, where the original non-maximized size was lost, resulting in the window being restored to the wrong size. All automated video tests pass across all platforms. Overall, the "make a request/get an event" model better reflects how most windowing systems work, and some backends avoid spending significant time blocking while waiting for operations to complete.
2023-10-25 22:09:55 +03:00
Displays are now referenced by instance ID instead of index
2023-01-30 00:30:55 +03:00
currentDisplay = SDL_GetDisplayForWindow(window);
SDL_GetWindowSize(window, &currentW, &currentH);
SDL_GetWindowPosition(window, &currentX, &currentY);
testautomation: skip window resize and repositioning tests on Emscripten
2024-08-18 00:20:41 +03:00
if (video_driver_is_wayland) {
SDLTest_Log("Skipping display ID validation: %s driver does not support window positioning", video_driver);
} else {
tests: Fix automated window grab and positioning tests under Wayland/XWayland When using the Wayland video driver or X11 under XWayland, create a renderer and present a frame as part of window creation, as Wayland requires that a frame be presented for the window to be fully mapped and displayed onscreen. This fixes the grab and expected window size tests. This also disables the window positioning tests when running under the Wayland driver, as Wayland does not allow application windows to position themselves in the desktop space, which renders the tests unreliable and subject to spurious failure.
2023-07-23 18:25:15 +03:00
SDLTest_AssertCheck(currentDisplay == expectedDisplay, "Validate display ID (current: %d, expected: %d)", currentDisplay, expectedDisplay);
testautomation: skip window resize and repositioning tests on Emscripten
2024-08-18 00:20:41 +03:00
}
if (video_driver_is_emscripten) {
SDLTest_Log("Skipping window position validation: %s driver does not support window resizing", video_driver);
tests: Fix automated window grab and positioning tests under Wayland/XWayland When using the Wayland video driver or X11 under XWayland, create a renderer and present a frame as part of window creation, as Wayland requires that a frame be presented for the window to be fully mapped and displayed onscreen. This fixes the grab and expected window size tests. This also disables the window positioning tests when running under the Wayland driver, as Wayland does not allow application windows to position themselves in the desktop space, which renders the tests unreliable and subject to spurious failure.
2023-07-23 18:25:15 +03:00
} else {
testautomation: skip window resize and repositioning tests on Emscripten
2024-08-18 00:20:41 +03:00
SDLTest_AssertCheck(currentW == w, "Validate width (current: %d, expected: %d)", currentW, w);
SDLTest_AssertCheck(currentH == h, "Validate height (current: %d, expected: %d)", currentH, h);
tests: Fix automated window grab and positioning tests under Wayland/XWayland When using the Wayland video driver or X11 under XWayland, create a renderer and present a frame as part of window creation, as Wayland requires that a frame be presented for the window to be fully mapped and displayed onscreen. This fixes the grab and expected window size tests. This also disables the window positioning tests when running under the Wayland driver, as Wayland does not allow application windows to position themselves in the desktop space, which renders the tests unreliable and subject to spurious failure.
2023-07-23 18:25:15 +03:00
}
testautomation: skip window resize and repositioning tests on Emscripten
2024-08-18 00:20:41 +03:00
if (video_driver_is_wayland) {
SDLTest_Log("Skipping window position validation: %s driver does not support window positioning", video_driver);
} else {
tests: Fix automated window grab and positioning tests under Wayland/XWayland When using the Wayland video driver or X11 under XWayland, create a renderer and present a frame as part of window creation, as Wayland requires that a frame be presented for the window to be fully mapped and displayed onscreen. This fixes the grab and expected window size tests. This also disables the window positioning tests when running under the Wayland driver, as Wayland does not allow application windows to position themselves in the desktop space, which renders the tests unreliable and subject to spurious failure.
2023-07-23 18:25:15 +03:00
SDLTest_AssertCheck(currentX == expectedX, "Validate x (current: %d, expected: %d)", currentX, expectedX);
SDLTest_AssertCheck(currentY == expectedY, "Validate y (current: %d, expected: %d)", currentY, expectedY);
}
Displays are now referenced by instance ID instead of index
2023-01-30 00:30:55 +03:00
/* Clean up */
destroyVideoSuiteTestWindow(window);
}
testautomation: add video_setWindowCenteredOnDisplay test
2022-06-05 23:44:30 +03:00
}
Removed temporary memory from the API It was intended to make the API easier to use, but various automatic garbage collection all had flaws, and making the application periodically clean up temporary memory added cognitive load to using the API, and in many cases was it was difficult to restructure threaded code to handle this. So, we're largely going back to the original system, where the API returns allocated results and you free them. In addition, to solve the problems we originally wanted temporary memory for: * Short strings with a finite count, like device names, get stored in a per-thread string pool. * Events continue to use temporary memory internally, which is cleaned up on the next event processing cycle.
2024-07-27 04:57:18 +03:00
SDL_free(displays);
testautomation: add video_setWindowCenteredOnDisplay test
2022-06-05 23:44:30 +03:00
}
return TEST_COMPLETED;
}
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
video: Implement asynchronous windowing SDL window size, state, and position functions have been considered immediate, with their effects assuming to have taken effect upon successful return of the function. However, several windowing systems handle these requests asynchronously, resulting in the functions blocking until the changes have taken effect, potentially for long periods of time. Additionally, some windowing systems treat these as requests, and can potentially deny or fulfill the request in a manner differently than the application expects, such as not allowing a window to be positioned or sized beyond desktop borders, prohibiting fullscreen, and so on. With these changes, applications can make requests of the window manager that do not block, with the understanding that an associated event will be sent if the request is fulfilled. Currently, size, position, maximize, minimize, and fullscreen calls are handled as asynchronous requests, with events being returned if the request is honored. If the application requires that the change take effect immediately, it can call the new SDL_SyncWindow function, which will attempt to block until the request is fulfilled, or some arbitrary timeout period elapses, the duration of which depends not only on the windowing system, but on the operation requested as well (e.g. a 100ms timeout is fine for most X11 events, but maximizing a window can take considerably longer for some reason). There is also a new hint 'SDL_VIDEO_SYNC_ALL_WINDOW_OPS' that will mimic the old behavior by synchronizing after every window operation with, again, the understanding that using this may result in the associated calls blocking for a relatively long period. The deferred model also results in the window size and position getters not reporting false coordinates anymore, as they only forward what the window manager reports vs allowing applications to set arbitrary values, and fullscreen enter/leave events that were initiated via the window manager update the window state appropriately, where they didn't before. Care was taken to ensure that order of operations is maintained, and that requests are not ignored or dropped. This does require some implicit internal synchronization in the various backends if many requests are made in a short period, as some state and behavior depends on other bits of state that need to be known at that particular point in time, but this isn't something that typical applications will hit, unless they are sending a lot of window state in a short time as the tests do. The automated tests developed to test the previous behavior also resulted in previously undefined behavior being defined and normalized across platforms, particularly when it comes to the sizing and positioning of windows when they are in a fixed-size state, such as maximized or fullscreen. Size and position requests made when the window is not in a movable or resizable state will be deferred until it can be applied, so no requests are lost. These changes fix another long-standing issue with renderers recreating maximized windows, where the original non-maximized size was lost, resulting in the window being restored to the wrong size. All automated video tests pass across all platforms. Overall, the "make a request/get an event" model better reflects how most windowing systems work, and some backends avoid spending significant time blocking while waiting for operations to complete.
2023-10-25 22:09:55 +03:00
/**
* Tests calls to SDL_MaximizeWindow(), SDL_RestoreWindow(), and SDL_SetWindowFullscreen(),
* interspersed with calls to set the window size and position, and verifies the flags,
* sizes, and positions of maximized, fullscreen, and restored windows.
*
* NOTE: This test is good on Mac, Win32, GNOME, and KDE (Wayland and X11). Other *nix
* desktops, particularly tiling desktops, may not support the expected behavior,
* so don't be surprised if this fails.
*/
SDL_test: use SDLCALL calling convention This is needed when using a pre-built static SDL3_test library.
2024-09-06 04:21:13 +03:00
static int SDLCALL video_getSetWindowState(void *arg)
video: Implement asynchronous windowing SDL window size, state, and position functions have been considered immediate, with their effects assuming to have taken effect upon successful return of the function. However, several windowing systems handle these requests asynchronously, resulting in the functions blocking until the changes have taken effect, potentially for long periods of time. Additionally, some windowing systems treat these as requests, and can potentially deny or fulfill the request in a manner differently than the application expects, such as not allowing a window to be positioned or sized beyond desktop borders, prohibiting fullscreen, and so on. With these changes, applications can make requests of the window manager that do not block, with the understanding that an associated event will be sent if the request is fulfilled. Currently, size, position, maximize, minimize, and fullscreen calls are handled as asynchronous requests, with events being returned if the request is honored. If the application requires that the change take effect immediately, it can call the new SDL_SyncWindow function, which will attempt to block until the request is fulfilled, or some arbitrary timeout period elapses, the duration of which depends not only on the windowing system, but on the operation requested as well (e.g. a 100ms timeout is fine for most X11 events, but maximizing a window can take considerably longer for some reason). There is also a new hint 'SDL_VIDEO_SYNC_ALL_WINDOW_OPS' that will mimic the old behavior by synchronizing after every window operation with, again, the understanding that using this may result in the associated calls blocking for a relatively long period. The deferred model also results in the window size and position getters not reporting false coordinates anymore, as they only forward what the window manager reports vs allowing applications to set arbitrary values, and fullscreen enter/leave events that were initiated via the window manager update the window state appropriately, where they didn't before. Care was taken to ensure that order of operations is maintained, and that requests are not ignored or dropped. This does require some implicit internal synchronization in the various backends if many requests are made in a short period, as some state and behavior depends on other bits of state that need to be known at that particular point in time, but this isn't something that typical applications will hit, unless they are sending a lot of window state in a short time as the tests do. The automated tests developed to test the previous behavior also resulted in previously undefined behavior being defined and normalized across platforms, particularly when it comes to the sizing and positioning of windows when they are in a fixed-size state, such as maximized or fullscreen. Size and position requests made when the window is not in a movable or resizable state will be deferred until it can be applied, so no requests are lost. These changes fix another long-standing issue with renderers recreating maximized windows, where the original non-maximized size was lost, resulting in the window being restored to the wrong size. All automated video tests pass across all platforms. Overall, the "make a request/get an event" model better reflects how most windowing systems work, and some backends avoid spending significant time blocking while waiting for operations to complete.
2023-10-25 22:09:55 +03:00
{
const char *title = "video_getSetWindowState Test Window";
SDL_Window *window;
int result;
SDL_Rect display;
Update some internal files to use `SDL_WindowFlags`
2024-03-03 22:30:57 +03:00
SDL_WindowFlags flags;
video: Implement asynchronous windowing SDL window size, state, and position functions have been considered immediate, with their effects assuming to have taken effect upon successful return of the function. However, several windowing systems handle these requests asynchronously, resulting in the functions blocking until the changes have taken effect, potentially for long periods of time. Additionally, some windowing systems treat these as requests, and can potentially deny or fulfill the request in a manner differently than the application expects, such as not allowing a window to be positioned or sized beyond desktop borders, prohibiting fullscreen, and so on. With these changes, applications can make requests of the window manager that do not block, with the understanding that an associated event will be sent if the request is fulfilled. Currently, size, position, maximize, minimize, and fullscreen calls are handled as asynchronous requests, with events being returned if the request is honored. If the application requires that the change take effect immediately, it can call the new SDL_SyncWindow function, which will attempt to block until the request is fulfilled, or some arbitrary timeout period elapses, the duration of which depends not only on the windowing system, but on the operation requested as well (e.g. a 100ms timeout is fine for most X11 events, but maximizing a window can take considerably longer for some reason). There is also a new hint 'SDL_VIDEO_SYNC_ALL_WINDOW_OPS' that will mimic the old behavior by synchronizing after every window operation with, again, the understanding that using this may result in the associated calls blocking for a relatively long period. The deferred model also results in the window size and position getters not reporting false coordinates anymore, as they only forward what the window manager reports vs allowing applications to set arbitrary values, and fullscreen enter/leave events that were initiated via the window manager update the window state appropriately, where they didn't before. Care was taken to ensure that order of operations is maintained, and that requests are not ignored or dropped. This does require some implicit internal synchronization in the various backends if many requests are made in a short period, as some state and behavior depends on other bits of state that need to be known at that particular point in time, but this isn't something that typical applications will hit, unless they are sending a lot of window state in a short time as the tests do. The automated tests developed to test the previous behavior also resulted in previously undefined behavior being defined and normalized across platforms, particularly when it comes to the sizing and positioning of windows when they are in a fixed-size state, such as maximized or fullscreen. Size and position requests made when the window is not in a movable or resizable state will be deferred until it can be applied, so no requests are lost. These changes fix another long-standing issue with renderers recreating maximized windows, where the original non-maximized size was lost, resulting in the window being restored to the wrong size. All automated video tests pass across all platforms. Overall, the "make a request/get an event" model better reflects how most windowing systems work, and some backends avoid spending significant time blocking while waiting for operations to complete.
2023-10-25 22:09:55 +03:00
int windowedX, windowedY;
int currentX, currentY;
int desiredX = 0, desiredY = 0;
int windowedW, windowedH;
int currentW, currentH;
int desiredW = 0, desiredH = 0;
Update some internal files to use `SDL_WindowFlags`
2024-03-03 22:30:57 +03:00
SDL_WindowFlags skipFlags = 0;
Removed SDL_bool in favor of plain bool We require stdbool.h in the build environment, so we might as well use the plain bool type. If your environment doesn't have stdbool.h, this simple replacement will suffice: typedef signed char bool;
2024-09-18 17:52:28 +03:00
const bool restoreHint = SDL_GetHintBoolean("SDL_BORDERLESS_RESIZABLE_STYLE", true);
const bool skipPos = SDL_strcmp(SDL_GetCurrentVideoDriver(), "wayland") == 0;
video: Implement asynchronous windowing SDL window size, state, and position functions have been considered immediate, with their effects assuming to have taken effect upon successful return of the function. However, several windowing systems handle these requests asynchronously, resulting in the functions blocking until the changes have taken effect, potentially for long periods of time. Additionally, some windowing systems treat these as requests, and can potentially deny or fulfill the request in a manner differently than the application expects, such as not allowing a window to be positioned or sized beyond desktop borders, prohibiting fullscreen, and so on. With these changes, applications can make requests of the window manager that do not block, with the understanding that an associated event will be sent if the request is fulfilled. Currently, size, position, maximize, minimize, and fullscreen calls are handled as asynchronous requests, with events being returned if the request is honored. If the application requires that the change take effect immediately, it can call the new SDL_SyncWindow function, which will attempt to block until the request is fulfilled, or some arbitrary timeout period elapses, the duration of which depends not only on the windowing system, but on the operation requested as well (e.g. a 100ms timeout is fine for most X11 events, but maximizing a window can take considerably longer for some reason). There is also a new hint 'SDL_VIDEO_SYNC_ALL_WINDOW_OPS' that will mimic the old behavior by synchronizing after every window operation with, again, the understanding that using this may result in the associated calls blocking for a relatively long period. The deferred model also results in the window size and position getters not reporting false coordinates anymore, as they only forward what the window manager reports vs allowing applications to set arbitrary values, and fullscreen enter/leave events that were initiated via the window manager update the window state appropriately, where they didn't before. Care was taken to ensure that order of operations is maintained, and that requests are not ignored or dropped. This does require some implicit internal synchronization in the various backends if many requests are made in a short period, as some state and behavior depends on other bits of state that need to be known at that particular point in time, but this isn't something that typical applications will hit, unless they are sending a lot of window state in a short time as the tests do. The automated tests developed to test the previous behavior also resulted in previously undefined behavior being defined and normalized across platforms, particularly when it comes to the sizing and positioning of windows when they are in a fixed-size state, such as maximized or fullscreen. Size and position requests made when the window is not in a movable or resizable state will be deferred until it can be applied, so no requests are lost. These changes fix another long-standing issue with renderers recreating maximized windows, where the original non-maximized size was lost, resulting in the window being restored to the wrong size. All automated video tests pass across all platforms. Overall, the "make a request/get an event" model better reflects how most windowing systems work, and some backends avoid spending significant time blocking while waiting for operations to complete.
2023-10-25 22:09:55 +03:00
/* This test is known to be good only on GNOME and KDE. At the time of writing, Weston seems to have maximize related bugs
* that prevent it from running correctly (no configure events are received when unsetting maximize), and tiling window
* managers such as Sway have fundamental behavioral differences that conflict with it.
*
* Other desktops can be enabled in the future as required.
*/
if (SDL_strcmp(SDL_GetCurrentVideoDriver(), "wayland") == 0 || SDL_strcmp(SDL_GetCurrentVideoDriver(), "x11") == 0) {
Added thread-safe environment functions Also marked the existing functions as unsafe, as they can cause crashes if used in multi-threaded applications. As a bonus, since the new functions are hashtable based, hint environment lookups are much faster.
2024-09-14 03:00:15 +03:00
const char *desktop = SDL_GetEnvironmentVariable(SDL_GetEnvironment(), "XDG_CURRENT_DESKTOP");
video: Implement asynchronous windowing SDL window size, state, and position functions have been considered immediate, with their effects assuming to have taken effect upon successful return of the function. However, several windowing systems handle these requests asynchronously, resulting in the functions blocking until the changes have taken effect, potentially for long periods of time. Additionally, some windowing systems treat these as requests, and can potentially deny or fulfill the request in a manner differently than the application expects, such as not allowing a window to be positioned or sized beyond desktop borders, prohibiting fullscreen, and so on. With these changes, applications can make requests of the window manager that do not block, with the understanding that an associated event will be sent if the request is fulfilled. Currently, size, position, maximize, minimize, and fullscreen calls are handled as asynchronous requests, with events being returned if the request is honored. If the application requires that the change take effect immediately, it can call the new SDL_SyncWindow function, which will attempt to block until the request is fulfilled, or some arbitrary timeout period elapses, the duration of which depends not only on the windowing system, but on the operation requested as well (e.g. a 100ms timeout is fine for most X11 events, but maximizing a window can take considerably longer for some reason). There is also a new hint 'SDL_VIDEO_SYNC_ALL_WINDOW_OPS' that will mimic the old behavior by synchronizing after every window operation with, again, the understanding that using this may result in the associated calls blocking for a relatively long period. The deferred model also results in the window size and position getters not reporting false coordinates anymore, as they only forward what the window manager reports vs allowing applications to set arbitrary values, and fullscreen enter/leave events that were initiated via the window manager update the window state appropriately, where they didn't before. Care was taken to ensure that order of operations is maintained, and that requests are not ignored or dropped. This does require some implicit internal synchronization in the various backends if many requests are made in a short period, as some state and behavior depends on other bits of state that need to be known at that particular point in time, but this isn't something that typical applications will hit, unless they are sending a lot of window state in a short time as the tests do. The automated tests developed to test the previous behavior also resulted in previously undefined behavior being defined and normalized across platforms, particularly when it comes to the sizing and positioning of windows when they are in a fixed-size state, such as maximized or fullscreen. Size and position requests made when the window is not in a movable or resizable state will be deferred until it can be applied, so no requests are lost. These changes fix another long-standing issue with renderers recreating maximized windows, where the original non-maximized size was lost, resulting in the window being restored to the wrong size. All automated video tests pass across all platforms. Overall, the "make a request/get an event" model better reflects how most windowing systems work, and some backends avoid spending significant time blocking while waiting for operations to complete.
2023-10-25 22:09:55 +03:00
if (SDL_strcmp(desktop, "GNOME") != 0 && SDL_strcmp(desktop, "KDE") != 0) {
SDLTest_Log("Skipping test video_getSetWindowState: desktop environment %s not supported", desktop);
return TEST_SKIPPED;
}
}
/* Win32 borderless windows are not resizable by default and need this undocumented hint */
SDL_SetHint("SDL_BORDERLESS_RESIZABLE_STYLE", "1");
/* Call against new test window */
window = createVideoSuiteTestWindow(title);
if (!window) {
return TEST_ABORTED;
}
SDL_GetWindowSize(window, &windowedW, &windowedH);
SDLTest_AssertPass("SDL_GetWindowSize()");
Added constant definitions for SDL properties Fixes https://github.com/libsdl-org/SDL/issues/8622
2024-01-08 03:59:41 +03:00
video: Implement asynchronous windowing SDL window size, state, and position functions have been considered immediate, with their effects assuming to have taken effect upon successful return of the function. However, several windowing systems handle these requests asynchronously, resulting in the functions blocking until the changes have taken effect, potentially for long periods of time. Additionally, some windowing systems treat these as requests, and can potentially deny or fulfill the request in a manner differently than the application expects, such as not allowing a window to be positioned or sized beyond desktop borders, prohibiting fullscreen, and so on. With these changes, applications can make requests of the window manager that do not block, with the understanding that an associated event will be sent if the request is fulfilled. Currently, size, position, maximize, minimize, and fullscreen calls are handled as asynchronous requests, with events being returned if the request is honored. If the application requires that the change take effect immediately, it can call the new SDL_SyncWindow function, which will attempt to block until the request is fulfilled, or some arbitrary timeout period elapses, the duration of which depends not only on the windowing system, but on the operation requested as well (e.g. a 100ms timeout is fine for most X11 events, but maximizing a window can take considerably longer for some reason). There is also a new hint 'SDL_VIDEO_SYNC_ALL_WINDOW_OPS' that will mimic the old behavior by synchronizing after every window operation with, again, the understanding that using this may result in the associated calls blocking for a relatively long period. The deferred model also results in the window size and position getters not reporting false coordinates anymore, as they only forward what the window manager reports vs allowing applications to set arbitrary values, and fullscreen enter/leave events that were initiated via the window manager update the window state appropriately, where they didn't before. Care was taken to ensure that order of operations is maintained, and that requests are not ignored or dropped. This does require some implicit internal synchronization in the various backends if many requests are made in a short period, as some state and behavior depends on other bits of state that need to be known at that particular point in time, but this isn't something that typical applications will hit, unless they are sending a lot of window state in a short time as the tests do. The automated tests developed to test the previous behavior also resulted in previously undefined behavior being defined and normalized across platforms, particularly when it comes to the sizing and positioning of windows when they are in a fixed-size state, such as maximized or fullscreen. Size and position requests made when the window is not in a movable or resizable state will be deferred until it can be applied, so no requests are lost. These changes fix another long-standing issue with renderers recreating maximized windows, where the original non-maximized size was lost, resulting in the window being restored to the wrong size. All automated video tests pass across all platforms. Overall, the "make a request/get an event" model better reflects how most windowing systems work, and some backends avoid spending significant time blocking while waiting for operations to complete.
2023-10-25 22:09:55 +03:00
SDL_GetWindowPosition(window, &windowedX, &windowedY);
SDLTest_AssertPass("SDL_GetWindowPosition()");
if (skipPos) {
SDLTest_Log("Skipping positioning tests: %s reports window positioning as unsupported", SDL_GetCurrentVideoDriver());
}
/* Maximize and check the dimensions */
result = SDL_MaximizeWindow(window);
SDLTest_AssertPass("SDL_MaximizeWindow()");
Use SDL_bool instead an int return code in the SDL API Most SDL functions used to indicate success or failure using an int return code. These functions have been changed to return SDL_bool. Here is a coccinelle patch to change code that previously compared the return value to 0 and changes it to a boolean test: @ bool_return_type @ identifier func =~ "^(SDL_AddEventWatch|SDL_AddHintCallback|SDL_AddSurfaceAlternateImage|SDL_AddVulkanRenderSemaphores|SDL_BindAudioStream|SDL_BindAudioStreams|SDL_BlitSurface|SDL_BlitSurface9Grid|SDL_BlitSurfaceScaled|SDL_BlitSurfaceTiled|SDL_BlitSurfaceTiledWithScale|SDL_BlitSurfaceUnchecked|SDL_BlitSurfaceUncheckedScaled|SDL_CaptureMouse|SDL_ClearAudioStream|SDL_ClearClipboardData|SDL_ClearComposition|SDL_ClearError|SDL_ClearProperty|SDL_ClearSurface|SDL_CloseIO|SDL_CloseStorage|SDL_ConvertAudioSamples|SDL_ConvertEventToRenderCoordinates|SDL_ConvertPixels|SDL_ConvertPixelsAndColorspace|SDL_CopyFile|SDL_CopyProperties|SDL_CopyStorageFile|SDL_CreateDirectory|SDL_CreateStorageDirectory|SDL_CreateWindowAndRenderer|SDL_DateTimeToTime|SDL_DestroyWindowSurface|SDL_DetachVirtualJoystick|SDL_DisableScreenSaver|SDL_EnableScreenSaver|SDL_EnumerateDirectory|SDL_EnumerateProperties|SDL_EnumerateStorageDirectory|SDL_FillSurfaceRect|SDL_FillSurfaceRects|SDL_FlashWindow|SDL_FlipSurface|SDL_FlushAudioStream|SDL_FlushRenderer|SDL_GL_DestroyContext|SDL_GL_GetAttribute|SDL_GL_GetSwapInterval|SDL_GL_LoadLibrary|SDL_GL_MakeCurrent|SDL_GL_SetAttribute|SDL_GL_SetSwapInterval|SDL_GL_SwapWindow|SDL_GetAudioDeviceFormat|SDL_GetAudioStreamFormat|SDL_GetCameraFormat|SDL_GetClosestFullscreenDisplayMode|SDL_GetCurrentRenderOutputSize|SDL_GetCurrentTime|SDL_GetDXGIOutputInfo|SDL_GetDateTimeLocalePreferences|SDL_GetDisplayBounds|SDL_GetDisplayUsableBounds|SDL_GetGDKDefaultUser|SDL_GetGDKTaskQueue|SDL_GetGamepadSensorData|SDL_GetGamepadTouchpadFinger|SDL_GetHapticEffectStatus|SDL_GetJoystickBall|SDL_GetMasksForPixelFormat|SDL_GetPathInfo|SDL_GetRectUnion|SDL_GetRectUnionFloat|SDL_GetRenderClipRect|SDL_GetRenderColorScale|SDL_GetRenderDrawBlendMode|SDL_GetRenderDrawColor|SDL_GetRenderDrawColorFloat|SDL_GetRenderLogicalPresentation|SDL_GetRenderLogicalPresentationRect|SDL_GetRenderOutputSize|SDL_GetRenderSafeArea|SDL_GetRenderScale|SDL_GetRenderVSync|SDL_GetRenderViewport|SDL_GetSensorData|SDL_GetStorageFileSize|SDL_GetStoragePathInfo|SDL_GetSurfaceAlphaMod|SDL_GetSurfaceBlendMode|SDL_GetSurfaceClipRect|SDL_GetSurfaceColorKey|SDL_GetSurfaceColorMod|SDL_GetTextInputArea|SDL_GetTextureAlphaMod|SDL_GetTextureAlphaModFloat|SDL_GetTextureBlendMode|SDL_GetTextureColorMod|SDL_GetTextureColorModFloat|SDL_GetTextureScaleMode|SDL_GetTextureSize|SDL_GetWindowAspectRatio|SDL_GetWindowBordersSize|SDL_GetWindowMaximumSize|SDL_GetWindowMinimumSize|SDL_GetWindowPosition|SDL_GetWindowRelativeMouseMode|SDL_GetWindowSafeArea|SDL_GetWindowSize|SDL_GetWindowSizeInPixels|SDL_GetWindowSurfaceVSync|SDL_HideCursor|SDL_HideWindow|SDL_Init|SDL_InitHapticRumble|SDL_InitSubSystem|SDL_LoadWAV|SDL_LoadWAV_IO|SDL_LockAudioStream|SDL_LockProperties|SDL_LockSurface|SDL_LockTexture|SDL_LockTextureToSurface|SDL_MaximizeWindow|SDL_MinimizeWindow|SDL_MixAudio|SDL_OpenURL|SDL_OutOfMemory|SDL_PauseAudioDevice|SDL_PauseAudioStreamDevice|SDL_PauseHaptic|SDL_PlayHapticRumble|SDL_PremultiplyAlpha|SDL_PremultiplySurfaceAlpha|SDL_PushEvent|SDL_PutAudioStreamData|SDL_RaiseWindow|SDL_ReadStorageFile|SDL_ReadSurfacePixel|SDL_ReadSurfacePixelFloat|SDL_RegisterApp|SDL_ReloadGamepadMappings|SDL_RemovePath|SDL_RemoveStoragePath|SDL_RemoveTimer|SDL_RenamePath|SDL_RenameStoragePath|SDL_RenderClear|SDL_RenderCoordinatesFromWindow|SDL_RenderCoordinatesToWindow|SDL_RenderFillRect|SDL_RenderFillRects|SDL_RenderGeometry|SDL_RenderGeometryRaw|SDL_RenderLine|SDL_RenderLines|SDL_RenderPoint|SDL_RenderPoints|SDL_RenderPresent|SDL_RenderRect|SDL_RenderRects|SDL_RenderTexture|SDL_RenderTexture9Grid|SDL_RenderTextureRotated|SDL_RenderTextureTiled|SDL_RequestAndroidPermission|SDL_RestoreWindow|SDL_ResumeAudioDevice|SDL_ResumeAudioStreamDevice|SDL_ResumeHaptic|SDL_RumbleGamepad|SDL_RumbleGamepadTriggers|SDL_RumbleJoystick|SDL_RumbleJoystickTriggers|SDL_RunHapticEffect|SDL_SaveBMP|SDL_SaveBMP_IO|SDL_SendAndroidMessage|SDL_SendGamepadEffect|SDL_SendJoystickEffect|SDL_SendJoystickVirtualSensorData|SDL_SetAppMetadata|SDL_SetAppMetadataProperty|SDL_SetAudioDeviceGain|SDL_SetAudioPostmixCallback|SDL_SetAudioStreamFormat|SDL_SetAudioStreamFrequencyRatio|SDL_SetAudioStreamGain|SDL_SetAudioStreamGetCallback|SDL_SetAudioStreamInputChannelMap|SDL_SetAudioStreamOutputChannelMap|SDL_SetAudioStreamPutCallback|SDL_SetBooleanProperty|SDL_SetClipboardData|SDL_SetClipboardText|SDL_SetCursor|SDL_SetFloatProperty|SDL_SetGamepadLED|SDL_SetGamepadMapping|SDL_SetGamepadPlayerIndex|SDL_SetGamepadSensorEnabled|SDL_SetHapticAutocenter|SDL_SetHapticGain|SDL_SetJoystickLED|SDL_SetJoystickPlayerIndex|SDL_SetJoystickVirtualAxis|SDL_SetJoystickVirtualBall|SDL_SetJoystickVirtualButton|SDL_SetJoystickVirtualHat|SDL_SetJoystickVirtualTouchpad|SDL_SetLinuxThreadPriority|SDL_SetLinuxThreadPriorityAndPolicy|SDL_SetLogPriorityPrefix|SDL_SetMemoryFunctions|SDL_SetNumberProperty|SDL_SetPaletteColors|SDL_SetPointerProperty|SDL_SetPointerPropertyWithCleanup|SDL_SetPrimarySelectionText|SDL_SetRenderClipRect|SDL_SetRenderColorScale|SDL_SetRenderDrawBlendMode|SDL_SetRenderDrawColor|SDL_SetRenderDrawColorFloat|SDL_SetRenderLogicalPresentation|SDL_SetRenderScale|SDL_SetRenderTarget|SDL_SetRenderVSync|SDL_SetRenderViewport|SDL_SetScancodeName|SDL_SetStringProperty|SDL_SetSurfaceAlphaMod|SDL_SetSurfaceBlendMode|SDL_SetSurfaceColorKey|SDL_SetSurfaceColorMod|SDL_SetSurfaceColorspace|SDL_SetSurfacePalette|SDL_SetSurfaceRLE|SDL_SetTLS|SDL_SetTextInputArea|SDL_SetTextureAlphaMod|SDL_SetTextureAlphaModFloat|SDL_SetTextureBlendMode|SDL_SetTextureColorMod|SDL_SetTextureColorModFloat|SDL_SetTextureScaleMode|SDL_SetThreadPriority|SDL_SetWindowAlwaysOnTop|SDL_SetWindowAspectRatio|SDL_SetWindowBordered|SDL_SetWindowFocusable|SDL_SetWindowFullscreen|SDL_SetWindowFullscreenMode|SDL_SetWindowHitTest|SDL_SetWindowIcon|SDL_SetWindowKeyboardGrab|SDL_SetWindowMaximumSize|SDL_SetWindowMinimumSize|SDL_SetWindowModalFor|SDL_SetWindowMouseGrab|SDL_SetWindowMouseRect|SDL_SetWindowOpacity|SDL_SetWindowPosition|SDL_SetWindowRelativeMouseMode|SDL_SetWindowResizable|SDL_SetWindowShape|SDL_SetWindowSize|SDL_SetWindowSurfaceVSync|SDL_SetWindowTitle|SDL_SetiOSAnimationCallback|SDL_ShowAndroidToast|SDL_ShowCursor|SDL_ShowMessageBox|SDL_ShowSimpleMessageBox|SDL_ShowWindow|SDL_ShowWindowSystemMenu|SDL_StartTextInput|SDL_StartTextInputWithProperties|SDL_StopHapticEffect|SDL_StopHapticEffects|SDL_StopHapticRumble|SDL_StopTextInput|SDL_SyncWindow|SDL_TimeToDateTime|SDL_TryLockMutex|SDL_TryLockRWLockForReading|SDL_TryLockRWLockForWriting|SDL_TryWaitSemaphore|SDL_UnlockAudioStream|SDL_UpdateHapticEffect|SDL_UpdateNVTexture|SDL_UpdateTexture|SDL_UpdateWindowSurface|SDL_UpdateWindowSurfaceRects|SDL_UpdateYUVTexture|SDL_Vulkan_CreateSurface|SDL_Vulkan_LoadLibrary|SDL_WaitConditionTimeout|SDL_WaitSemaphoreTimeout|SDL_WarpMouseGlobal|SDL_WriteStorageFile|SDL_WriteSurfacePixel|SDL_WriteSurfacePixelFloat)$"; @@ ( func( ... ) - == 0 | - func( + !func( ... ) - < 0 | - func( + !func( ... ) - != 0 | - func( + !func( ... ) - == -1 )
2024-08-23 03:33:49 +03:00
if (!result) {
video: Implement asynchronous windowing SDL window size, state, and position functions have been considered immediate, with their effects assuming to have taken effect upon successful return of the function. However, several windowing systems handle these requests asynchronously, resulting in the functions blocking until the changes have taken effect, potentially for long periods of time. Additionally, some windowing systems treat these as requests, and can potentially deny or fulfill the request in a manner differently than the application expects, such as not allowing a window to be positioned or sized beyond desktop borders, prohibiting fullscreen, and so on. With these changes, applications can make requests of the window manager that do not block, with the understanding that an associated event will be sent if the request is fulfilled. Currently, size, position, maximize, minimize, and fullscreen calls are handled as asynchronous requests, with events being returned if the request is honored. If the application requires that the change take effect immediately, it can call the new SDL_SyncWindow function, which will attempt to block until the request is fulfilled, or some arbitrary timeout period elapses, the duration of which depends not only on the windowing system, but on the operation requested as well (e.g. a 100ms timeout is fine for most X11 events, but maximizing a window can take considerably longer for some reason). There is also a new hint 'SDL_VIDEO_SYNC_ALL_WINDOW_OPS' that will mimic the old behavior by synchronizing after every window operation with, again, the understanding that using this may result in the associated calls blocking for a relatively long period. The deferred model also results in the window size and position getters not reporting false coordinates anymore, as they only forward what the window manager reports vs allowing applications to set arbitrary values, and fullscreen enter/leave events that were initiated via the window manager update the window state appropriately, where they didn't before. Care was taken to ensure that order of operations is maintained, and that requests are not ignored or dropped. This does require some implicit internal synchronization in the various backends if many requests are made in a short period, as some state and behavior depends on other bits of state that need to be known at that particular point in time, but this isn't something that typical applications will hit, unless they are sending a lot of window state in a short time as the tests do. The automated tests developed to test the previous behavior also resulted in previously undefined behavior being defined and normalized across platforms, particularly when it comes to the sizing and positioning of windows when they are in a fixed-size state, such as maximized or fullscreen. Size and position requests made when the window is not in a movable or resizable state will be deferred until it can be applied, so no requests are lost. These changes fix another long-standing issue with renderers recreating maximized windows, where the original non-maximized size was lost, resulting in the window being restored to the wrong size. All automated video tests pass across all platforms. Overall, the "make a request/get an event" model better reflects how most windowing systems work, and some backends avoid spending significant time blocking while waiting for operations to complete.
2023-10-25 22:09:55 +03:00
SDLTest_Log("Skipping state transition tests: %s reports window maximizing as unsupported", SDL_GetCurrentVideoDriver());
skipFlags |= SDL_WINDOW_MAXIMIZED;
goto minimize_test;
}
result = SDL_SyncWindow(window);
SDLTest_AssertPass("SDL_SyncWindow()");
Removed SDL_bool in favor of plain bool We require stdbool.h in the build environment, so we might as well use the plain bool type. If your environment doesn't have stdbool.h, this simple replacement will suffice: typedef signed char bool;
2024-09-18 17:52:28 +03:00
SDLTest_AssertCheck(result == true, "Verify return value; expected: true, got: %d", result);
video: Implement asynchronous windowing SDL window size, state, and position functions have been considered immediate, with their effects assuming to have taken effect upon successful return of the function. However, several windowing systems handle these requests asynchronously, resulting in the functions blocking until the changes have taken effect, potentially for long periods of time. Additionally, some windowing systems treat these as requests, and can potentially deny or fulfill the request in a manner differently than the application expects, such as not allowing a window to be positioned or sized beyond desktop borders, prohibiting fullscreen, and so on. With these changes, applications can make requests of the window manager that do not block, with the understanding that an associated event will be sent if the request is fulfilled. Currently, size, position, maximize, minimize, and fullscreen calls are handled as asynchronous requests, with events being returned if the request is honored. If the application requires that the change take effect immediately, it can call the new SDL_SyncWindow function, which will attempt to block until the request is fulfilled, or some arbitrary timeout period elapses, the duration of which depends not only on the windowing system, but on the operation requested as well (e.g. a 100ms timeout is fine for most X11 events, but maximizing a window can take considerably longer for some reason). There is also a new hint 'SDL_VIDEO_SYNC_ALL_WINDOW_OPS' that will mimic the old behavior by synchronizing after every window operation with, again, the understanding that using this may result in the associated calls blocking for a relatively long period. The deferred model also results in the window size and position getters not reporting false coordinates anymore, as they only forward what the window manager reports vs allowing applications to set arbitrary values, and fullscreen enter/leave events that were initiated via the window manager update the window state appropriately, where they didn't before. Care was taken to ensure that order of operations is maintained, and that requests are not ignored or dropped. This does require some implicit internal synchronization in the various backends if many requests are made in a short period, as some state and behavior depends on other bits of state that need to be known at that particular point in time, but this isn't something that typical applications will hit, unless they are sending a lot of window state in a short time as the tests do. The automated tests developed to test the previous behavior also resulted in previously undefined behavior being defined and normalized across platforms, particularly when it comes to the sizing and positioning of windows when they are in a fixed-size state, such as maximized or fullscreen. Size and position requests made when the window is not in a movable or resizable state will be deferred until it can be applied, so no requests are lost. These changes fix another long-standing issue with renderers recreating maximized windows, where the original non-maximized size was lost, resulting in the window being restored to the wrong size. All automated video tests pass across all platforms. Overall, the "make a request/get an event" model better reflects how most windowing systems work, and some backends avoid spending significant time blocking while waiting for operations to complete.
2023-10-25 22:09:55 +03:00
flags = SDL_GetWindowFlags(window);
SDLTest_AssertPass("SDL_GetWindowFlags()");
SDLTest_AssertCheck(flags & SDL_WINDOW_MAXIMIZED, "Verify the `SDL_WINDOW_MAXIMIZED` flag is set: %s", (flags & SDL_WINDOW_MAXIMIZED) ? "true" : "false");
/* Check that the maximized window doesn't extend beyond the usable display bounds.
* FIXME: Maximizing Win32 borderless windows is broken, so this always fails.
* Skip it for now.
*/
if (SDL_strcmp(SDL_GetCurrentVideoDriver(), "windows") != 0) {
result = SDL_GetDisplayUsableBounds(SDL_GetDisplayForWindow(window), &display);
SDLTest_AssertPass("SDL_GetDisplayUsableBounds()");
Removed SDL_bool in favor of plain bool We require stdbool.h in the build environment, so we might as well use the plain bool type. If your environment doesn't have stdbool.h, this simple replacement will suffice: typedef signed char bool;
2024-09-18 17:52:28 +03:00
SDLTest_AssertCheck(result == true, "Verify return value; expected: true, got: %d", result);
video: Implement asynchronous windowing SDL window size, state, and position functions have been considered immediate, with their effects assuming to have taken effect upon successful return of the function. However, several windowing systems handle these requests asynchronously, resulting in the functions blocking until the changes have taken effect, potentially for long periods of time. Additionally, some windowing systems treat these as requests, and can potentially deny or fulfill the request in a manner differently than the application expects, such as not allowing a window to be positioned or sized beyond desktop borders, prohibiting fullscreen, and so on. With these changes, applications can make requests of the window manager that do not block, with the understanding that an associated event will be sent if the request is fulfilled. Currently, size, position, maximize, minimize, and fullscreen calls are handled as asynchronous requests, with events being returned if the request is honored. If the application requires that the change take effect immediately, it can call the new SDL_SyncWindow function, which will attempt to block until the request is fulfilled, or some arbitrary timeout period elapses, the duration of which depends not only on the windowing system, but on the operation requested as well (e.g. a 100ms timeout is fine for most X11 events, but maximizing a window can take considerably longer for some reason). There is also a new hint 'SDL_VIDEO_SYNC_ALL_WINDOW_OPS' that will mimic the old behavior by synchronizing after every window operation with, again, the understanding that using this may result in the associated calls blocking for a relatively long period. The deferred model also results in the window size and position getters not reporting false coordinates anymore, as they only forward what the window manager reports vs allowing applications to set arbitrary values, and fullscreen enter/leave events that were initiated via the window manager update the window state appropriately, where they didn't before. Care was taken to ensure that order of operations is maintained, and that requests are not ignored or dropped. This does require some implicit internal synchronization in the various backends if many requests are made in a short period, as some state and behavior depends on other bits of state that need to be known at that particular point in time, but this isn't something that typical applications will hit, unless they are sending a lot of window state in a short time as the tests do. The automated tests developed to test the previous behavior also resulted in previously undefined behavior being defined and normalized across platforms, particularly when it comes to the sizing and positioning of windows when they are in a fixed-size state, such as maximized or fullscreen. Size and position requests made when the window is not in a movable or resizable state will be deferred until it can be applied, so no requests are lost. These changes fix another long-standing issue with renderers recreating maximized windows, where the original non-maximized size was lost, resulting in the window being restored to the wrong size. All automated video tests pass across all platforms. Overall, the "make a request/get an event" model better reflects how most windowing systems work, and some backends avoid spending significant time blocking while waiting for operations to complete.
2023-10-25 22:09:55 +03:00
desiredW = display.w;
desiredH = display.h;
currentW = windowedW + 1;
currentH = windowedH + 1;
SDL_GetWindowSize(window, &currentW, &currentH);
SDLTest_AssertPass("Call to SDL_GetWindowSize()");
SDLTest_AssertCheck(currentW <= desiredW, "Verify returned width; expected: <= %d, got: %d", desiredW,
currentW);
SDLTest_AssertCheck(currentH <= desiredH, "Verify returned height; expected: <= %d, got: %d", desiredH,
currentH);
}
/* Restore and check the dimensions */
result = SDL_RestoreWindow(window);
SDLTest_AssertPass("SDL_RestoreWindow()");
Removed SDL_bool in favor of plain bool We require stdbool.h in the build environment, so we might as well use the plain bool type. If your environment doesn't have stdbool.h, this simple replacement will suffice: typedef signed char bool;
2024-09-18 17:52:28 +03:00
SDLTest_AssertCheck(result == true, "Verify return value; expected: true, got: %d", result);
video: Implement asynchronous windowing SDL window size, state, and position functions have been considered immediate, with their effects assuming to have taken effect upon successful return of the function. However, several windowing systems handle these requests asynchronously, resulting in the functions blocking until the changes have taken effect, potentially for long periods of time. Additionally, some windowing systems treat these as requests, and can potentially deny or fulfill the request in a manner differently than the application expects, such as not allowing a window to be positioned or sized beyond desktop borders, prohibiting fullscreen, and so on. With these changes, applications can make requests of the window manager that do not block, with the understanding that an associated event will be sent if the request is fulfilled. Currently, size, position, maximize, minimize, and fullscreen calls are handled as asynchronous requests, with events being returned if the request is honored. If the application requires that the change take effect immediately, it can call the new SDL_SyncWindow function, which will attempt to block until the request is fulfilled, or some arbitrary timeout period elapses, the duration of which depends not only on the windowing system, but on the operation requested as well (e.g. a 100ms timeout is fine for most X11 events, but maximizing a window can take considerably longer for some reason). There is also a new hint 'SDL_VIDEO_SYNC_ALL_WINDOW_OPS' that will mimic the old behavior by synchronizing after every window operation with, again, the understanding that using this may result in the associated calls blocking for a relatively long period. The deferred model also results in the window size and position getters not reporting false coordinates anymore, as they only forward what the window manager reports vs allowing applications to set arbitrary values, and fullscreen enter/leave events that were initiated via the window manager update the window state appropriately, where they didn't before. Care was taken to ensure that order of operations is maintained, and that requests are not ignored or dropped. This does require some implicit internal synchronization in the various backends if many requests are made in a short period, as some state and behavior depends on other bits of state that need to be known at that particular point in time, but this isn't something that typical applications will hit, unless they are sending a lot of window state in a short time as the tests do. The automated tests developed to test the previous behavior also resulted in previously undefined behavior being defined and normalized across platforms, particularly when it comes to the sizing and positioning of windows when they are in a fixed-size state, such as maximized or fullscreen. Size and position requests made when the window is not in a movable or resizable state will be deferred until it can be applied, so no requests are lost. These changes fix another long-standing issue with renderers recreating maximized windows, where the original non-maximized size was lost, resulting in the window being restored to the wrong size. All automated video tests pass across all platforms. Overall, the "make a request/get an event" model better reflects how most windowing systems work, and some backends avoid spending significant time blocking while waiting for operations to complete.
2023-10-25 22:09:55 +03:00
result = SDL_SyncWindow(window);
SDLTest_AssertPass("SDL_SyncWindow()");
Removed SDL_bool in favor of plain bool We require stdbool.h in the build environment, so we might as well use the plain bool type. If your environment doesn't have stdbool.h, this simple replacement will suffice: typedef signed char bool;
2024-09-18 17:52:28 +03:00
SDLTest_AssertCheck(result == true, "Verify return value; expected: true, got: %d", result);
video: Implement asynchronous windowing SDL window size, state, and position functions have been considered immediate, with their effects assuming to have taken effect upon successful return of the function. However, several windowing systems handle these requests asynchronously, resulting in the functions blocking until the changes have taken effect, potentially for long periods of time. Additionally, some windowing systems treat these as requests, and can potentially deny or fulfill the request in a manner differently than the application expects, such as not allowing a window to be positioned or sized beyond desktop borders, prohibiting fullscreen, and so on. With these changes, applications can make requests of the window manager that do not block, with the understanding that an associated event will be sent if the request is fulfilled. Currently, size, position, maximize, minimize, and fullscreen calls are handled as asynchronous requests, with events being returned if the request is honored. If the application requires that the change take effect immediately, it can call the new SDL_SyncWindow function, which will attempt to block until the request is fulfilled, or some arbitrary timeout period elapses, the duration of which depends not only on the windowing system, but on the operation requested as well (e.g. a 100ms timeout is fine for most X11 events, but maximizing a window can take considerably longer for some reason). There is also a new hint 'SDL_VIDEO_SYNC_ALL_WINDOW_OPS' that will mimic the old behavior by synchronizing after every window operation with, again, the understanding that using this may result in the associated calls blocking for a relatively long period. The deferred model also results in the window size and position getters not reporting false coordinates anymore, as they only forward what the window manager reports vs allowing applications to set arbitrary values, and fullscreen enter/leave events that were initiated via the window manager update the window state appropriately, where they didn't before. Care was taken to ensure that order of operations is maintained, and that requests are not ignored or dropped. This does require some implicit internal synchronization in the various backends if many requests are made in a short period, as some state and behavior depends on other bits of state that need to be known at that particular point in time, but this isn't something that typical applications will hit, unless they are sending a lot of window state in a short time as the tests do. The automated tests developed to test the previous behavior also resulted in previously undefined behavior being defined and normalized across platforms, particularly when it comes to the sizing and positioning of windows when they are in a fixed-size state, such as maximized or fullscreen. Size and position requests made when the window is not in a movable or resizable state will be deferred until it can be applied, so no requests are lost. These changes fix another long-standing issue with renderers recreating maximized windows, where the original non-maximized size was lost, resulting in the window being restored to the wrong size. All automated video tests pass across all platforms. Overall, the "make a request/get an event" model better reflects how most windowing systems work, and some backends avoid spending significant time blocking while waiting for operations to complete.
2023-10-25 22:09:55 +03:00
flags = SDL_GetWindowFlags(window);
SDLTest_AssertPass("SDL_GetWindowFlags()");
SDLTest_AssertCheck(!(flags & SDL_WINDOW_MAXIMIZED), "Verify that the `SDL_WINDOW_MAXIMIZED` flag is cleared: %s", !(flags & SDL_WINDOW_MAXIMIZED) ? "true" : "false");
if (!skipPos) {
currentX = windowedX + 1;
currentY = windowedY + 1;
SDL_GetWindowPosition(window, &currentX, &currentY);
SDLTest_AssertPass("Call to SDL_GetWindowPosition()");
SDLTest_AssertCheck(windowedX == currentX, "Verify returned X coordinate; expected: %d, got: %d", windowedX, currentX);
SDLTest_AssertCheck(windowedY == currentY, "Verify returned Y coordinate; expected: %d, got: %d", windowedY, currentY);
}
currentW = windowedW + 1;
currentH = windowedH + 1;
SDL_GetWindowSize(window, &currentW, &currentH);
SDLTest_AssertPass("Call to SDL_GetWindowSize()");
SDLTest_AssertCheck(windowedW == currentW, "Verify returned width; expected: %d, got: %d", windowedW, currentW);
SDLTest_AssertCheck(windowedH == currentH, "Verify returned height; expected: %d, got: %d", windowedH, currentH);
/* Maximize, then immediately restore */
result = SDL_MaximizeWindow(window);
SDLTest_AssertPass("SDL_MaximizeWindow()");
Removed SDL_bool in favor of plain bool We require stdbool.h in the build environment, so we might as well use the plain bool type. If your environment doesn't have stdbool.h, this simple replacement will suffice: typedef signed char bool;
2024-09-18 17:52:28 +03:00
SDLTest_AssertCheck(result == true, "Verify return value; expected: true, got: %d", result);
video: Implement asynchronous windowing SDL window size, state, and position functions have been considered immediate, with their effects assuming to have taken effect upon successful return of the function. However, several windowing systems handle these requests asynchronously, resulting in the functions blocking until the changes have taken effect, potentially for long periods of time. Additionally, some windowing systems treat these as requests, and can potentially deny or fulfill the request in a manner differently than the application expects, such as not allowing a window to be positioned or sized beyond desktop borders, prohibiting fullscreen, and so on. With these changes, applications can make requests of the window manager that do not block, with the understanding that an associated event will be sent if the request is fulfilled. Currently, size, position, maximize, minimize, and fullscreen calls are handled as asynchronous requests, with events being returned if the request is honored. If the application requires that the change take effect immediately, it can call the new SDL_SyncWindow function, which will attempt to block until the request is fulfilled, or some arbitrary timeout period elapses, the duration of which depends not only on the windowing system, but on the operation requested as well (e.g. a 100ms timeout is fine for most X11 events, but maximizing a window can take considerably longer for some reason). There is also a new hint 'SDL_VIDEO_SYNC_ALL_WINDOW_OPS' that will mimic the old behavior by synchronizing after every window operation with, again, the understanding that using this may result in the associated calls blocking for a relatively long period. The deferred model also results in the window size and position getters not reporting false coordinates anymore, as they only forward what the window manager reports vs allowing applications to set arbitrary values, and fullscreen enter/leave events that were initiated via the window manager update the window state appropriately, where they didn't before. Care was taken to ensure that order of operations is maintained, and that requests are not ignored or dropped. This does require some implicit internal synchronization in the various backends if many requests are made in a short period, as some state and behavior depends on other bits of state that need to be known at that particular point in time, but this isn't something that typical applications will hit, unless they are sending a lot of window state in a short time as the tests do. The automated tests developed to test the previous behavior also resulted in previously undefined behavior being defined and normalized across platforms, particularly when it comes to the sizing and positioning of windows when they are in a fixed-size state, such as maximized or fullscreen. Size and position requests made when the window is not in a movable or resizable state will be deferred until it can be applied, so no requests are lost. These changes fix another long-standing issue with renderers recreating maximized windows, where the original non-maximized size was lost, resulting in the window being restored to the wrong size. All automated video tests pass across all platforms. Overall, the "make a request/get an event" model better reflects how most windowing systems work, and some backends avoid spending significant time blocking while waiting for operations to complete.
2023-10-25 22:09:55 +03:00
result = SDL_RestoreWindow(window);
SDLTest_AssertPass("SDL_RestoreWindow()");
Removed SDL_bool in favor of plain bool We require stdbool.h in the build environment, so we might as well use the plain bool type. If your environment doesn't have stdbool.h, this simple replacement will suffice: typedef signed char bool;
2024-09-18 17:52:28 +03:00
SDLTest_AssertCheck(result == true, "Verify return value; expected: true, got: %d", result);
video: Implement asynchronous windowing SDL window size, state, and position functions have been considered immediate, with their effects assuming to have taken effect upon successful return of the function. However, several windowing systems handle these requests asynchronously, resulting in the functions blocking until the changes have taken effect, potentially for long periods of time. Additionally, some windowing systems treat these as requests, and can potentially deny or fulfill the request in a manner differently than the application expects, such as not allowing a window to be positioned or sized beyond desktop borders, prohibiting fullscreen, and so on. With these changes, applications can make requests of the window manager that do not block, with the understanding that an associated event will be sent if the request is fulfilled. Currently, size, position, maximize, minimize, and fullscreen calls are handled as asynchronous requests, with events being returned if the request is honored. If the application requires that the change take effect immediately, it can call the new SDL_SyncWindow function, which will attempt to block until the request is fulfilled, or some arbitrary timeout period elapses, the duration of which depends not only on the windowing system, but on the operation requested as well (e.g. a 100ms timeout is fine for most X11 events, but maximizing a window can take considerably longer for some reason). There is also a new hint 'SDL_VIDEO_SYNC_ALL_WINDOW_OPS' that will mimic the old behavior by synchronizing after every window operation with, again, the understanding that using this may result in the associated calls blocking for a relatively long period. The deferred model also results in the window size and position getters not reporting false coordinates anymore, as they only forward what the window manager reports vs allowing applications to set arbitrary values, and fullscreen enter/leave events that were initiated via the window manager update the window state appropriately, where they didn't before. Care was taken to ensure that order of operations is maintained, and that requests are not ignored or dropped. This does require some implicit internal synchronization in the various backends if many requests are made in a short period, as some state and behavior depends on other bits of state that need to be known at that particular point in time, but this isn't something that typical applications will hit, unless they are sending a lot of window state in a short time as the tests do. The automated tests developed to test the previous behavior also resulted in previously undefined behavior being defined and normalized across platforms, particularly when it comes to the sizing and positioning of windows when they are in a fixed-size state, such as maximized or fullscreen. Size and position requests made when the window is not in a movable or resizable state will be deferred until it can be applied, so no requests are lost. These changes fix another long-standing issue with renderers recreating maximized windows, where the original non-maximized size was lost, resulting in the window being restored to the wrong size. All automated video tests pass across all platforms. Overall, the "make a request/get an event" model better reflects how most windowing systems work, and some backends avoid spending significant time blocking while waiting for operations to complete.
2023-10-25 22:09:55 +03:00
result = SDL_SyncWindow(window);
SDLTest_AssertPass("SDL_SyncWindow()");
Removed SDL_bool in favor of plain bool We require stdbool.h in the build environment, so we might as well use the plain bool type. If your environment doesn't have stdbool.h, this simple replacement will suffice: typedef signed char bool;
2024-09-18 17:52:28 +03:00
SDLTest_AssertCheck(result == true, "Verify return value; expected: true, got: %d", result);
video: Implement asynchronous windowing SDL window size, state, and position functions have been considered immediate, with their effects assuming to have taken effect upon successful return of the function. However, several windowing systems handle these requests asynchronously, resulting in the functions blocking until the changes have taken effect, potentially for long periods of time. Additionally, some windowing systems treat these as requests, and can potentially deny or fulfill the request in a manner differently than the application expects, such as not allowing a window to be positioned or sized beyond desktop borders, prohibiting fullscreen, and so on. With these changes, applications can make requests of the window manager that do not block, with the understanding that an associated event will be sent if the request is fulfilled. Currently, size, position, maximize, minimize, and fullscreen calls are handled as asynchronous requests, with events being returned if the request is honored. If the application requires that the change take effect immediately, it can call the new SDL_SyncWindow function, which will attempt to block until the request is fulfilled, or some arbitrary timeout period elapses, the duration of which depends not only on the windowing system, but on the operation requested as well (e.g. a 100ms timeout is fine for most X11 events, but maximizing a window can take considerably longer for some reason). There is also a new hint 'SDL_VIDEO_SYNC_ALL_WINDOW_OPS' that will mimic the old behavior by synchronizing after every window operation with, again, the understanding that using this may result in the associated calls blocking for a relatively long period. The deferred model also results in the window size and position getters not reporting false coordinates anymore, as they only forward what the window manager reports vs allowing applications to set arbitrary values, and fullscreen enter/leave events that were initiated via the window manager update the window state appropriately, where they didn't before. Care was taken to ensure that order of operations is maintained, and that requests are not ignored or dropped. This does require some implicit internal synchronization in the various backends if many requests are made in a short period, as some state and behavior depends on other bits of state that need to be known at that particular point in time, but this isn't something that typical applications will hit, unless they are sending a lot of window state in a short time as the tests do. The automated tests developed to test the previous behavior also resulted in previously undefined behavior being defined and normalized across platforms, particularly when it comes to the sizing and positioning of windows when they are in a fixed-size state, such as maximized or fullscreen. Size and position requests made when the window is not in a movable or resizable state will be deferred until it can be applied, so no requests are lost. These changes fix another long-standing issue with renderers recreating maximized windows, where the original non-maximized size was lost, resulting in the window being restored to the wrong size. All automated video tests pass across all platforms. Overall, the "make a request/get an event" model better reflects how most windowing systems work, and some backends avoid spending significant time blocking while waiting for operations to complete.
2023-10-25 22:09:55 +03:00
flags = SDL_GetWindowFlags(window);
SDLTest_AssertPass("SDL_GetWindowFlags()");
SDLTest_AssertCheck(!(flags & SDL_WINDOW_MAXIMIZED), "Verify that the `SDL_WINDOW_MAXIMIZED` flag is cleared: %s", !(flags & SDL_WINDOW_MAXIMIZED) ? "true" : "false");
/* Make sure the restored size and position matches the original windowed size and position. */
if (!skipPos) {
currentX = windowedX + 1;
currentY = windowedY + 1;
SDL_GetWindowPosition(window, &currentX, &currentY);
SDLTest_AssertPass("Call to SDL_GetWindowPosition()");
SDLTest_AssertCheck(windowedX == currentX, "Verify returned X coordinate; expected: %d, got: %d", windowedX, currentX);
SDLTest_AssertCheck(windowedY == currentY, "Verify returned Y coordinate; expected: %d, got: %d", windowedY, currentY);
}
currentW = windowedW + 1;
currentH = windowedH + 1;
SDL_GetWindowSize(window, &currentW, &currentH);
SDLTest_AssertPass("Call to SDL_GetWindowSize()");
SDLTest_AssertCheck(windowedW == currentW, "Verify returned width; expected: %d, got: %d", windowedW, currentW);
SDLTest_AssertCheck(windowedH == currentH, "Verify returned height; expected: %d, got: %d", windowedH, currentH);
/* Maximize, then enter fullscreen */
result = SDL_MaximizeWindow(window);
SDLTest_AssertPass("SDL_MaximizeWindow()");
Removed SDL_bool in favor of plain bool We require stdbool.h in the build environment, so we might as well use the plain bool type. If your environment doesn't have stdbool.h, this simple replacement will suffice: typedef signed char bool;
2024-09-18 17:52:28 +03:00
SDLTest_AssertCheck(result == true, "Verify return value; expected: true, got: %d", result);
video: Implement asynchronous windowing SDL window size, state, and position functions have been considered immediate, with their effects assuming to have taken effect upon successful return of the function. However, several windowing systems handle these requests asynchronously, resulting in the functions blocking until the changes have taken effect, potentially for long periods of time. Additionally, some windowing systems treat these as requests, and can potentially deny or fulfill the request in a manner differently than the application expects, such as not allowing a window to be positioned or sized beyond desktop borders, prohibiting fullscreen, and so on. With these changes, applications can make requests of the window manager that do not block, with the understanding that an associated event will be sent if the request is fulfilled. Currently, size, position, maximize, minimize, and fullscreen calls are handled as asynchronous requests, with events being returned if the request is honored. If the application requires that the change take effect immediately, it can call the new SDL_SyncWindow function, which will attempt to block until the request is fulfilled, or some arbitrary timeout period elapses, the duration of which depends not only on the windowing system, but on the operation requested as well (e.g. a 100ms timeout is fine for most X11 events, but maximizing a window can take considerably longer for some reason). There is also a new hint 'SDL_VIDEO_SYNC_ALL_WINDOW_OPS' that will mimic the old behavior by synchronizing after every window operation with, again, the understanding that using this may result in the associated calls blocking for a relatively long period. The deferred model also results in the window size and position getters not reporting false coordinates anymore, as they only forward what the window manager reports vs allowing applications to set arbitrary values, and fullscreen enter/leave events that were initiated via the window manager update the window state appropriately, where they didn't before. Care was taken to ensure that order of operations is maintained, and that requests are not ignored or dropped. This does require some implicit internal synchronization in the various backends if many requests are made in a short period, as some state and behavior depends on other bits of state that need to be known at that particular point in time, but this isn't something that typical applications will hit, unless they are sending a lot of window state in a short time as the tests do. The automated tests developed to test the previous behavior also resulted in previously undefined behavior being defined and normalized across platforms, particularly when it comes to the sizing and positioning of windows when they are in a fixed-size state, such as maximized or fullscreen. Size and position requests made when the window is not in a movable or resizable state will be deferred until it can be applied, so no requests are lost. These changes fix another long-standing issue with renderers recreating maximized windows, where the original non-maximized size was lost, resulting in the window being restored to the wrong size. All automated video tests pass across all platforms. Overall, the "make a request/get an event" model better reflects how most windowing systems work, and some backends avoid spending significant time blocking while waiting for operations to complete.
2023-10-25 22:09:55 +03:00
Removed SDL_bool in favor of plain bool We require stdbool.h in the build environment, so we might as well use the plain bool type. If your environment doesn't have stdbool.h, this simple replacement will suffice: typedef signed char bool;
2024-09-18 17:52:28 +03:00
result = SDL_SetWindowFullscreen(window, true);
SDLTest_AssertPass("SDL_SetWindowFullscreen(true)");
SDLTest_AssertCheck(result == true, "Verify return value; expected: true, got: %d", result);
video: Implement asynchronous windowing SDL window size, state, and position functions have been considered immediate, with their effects assuming to have taken effect upon successful return of the function. However, several windowing systems handle these requests asynchronously, resulting in the functions blocking until the changes have taken effect, potentially for long periods of time. Additionally, some windowing systems treat these as requests, and can potentially deny or fulfill the request in a manner differently than the application expects, such as not allowing a window to be positioned or sized beyond desktop borders, prohibiting fullscreen, and so on. With these changes, applications can make requests of the window manager that do not block, with the understanding that an associated event will be sent if the request is fulfilled. Currently, size, position, maximize, minimize, and fullscreen calls are handled as asynchronous requests, with events being returned if the request is honored. If the application requires that the change take effect immediately, it can call the new SDL_SyncWindow function, which will attempt to block until the request is fulfilled, or some arbitrary timeout period elapses, the duration of which depends not only on the windowing system, but on the operation requested as well (e.g. a 100ms timeout is fine for most X11 events, but maximizing a window can take considerably longer for some reason). There is also a new hint 'SDL_VIDEO_SYNC_ALL_WINDOW_OPS' that will mimic the old behavior by synchronizing after every window operation with, again, the understanding that using this may result in the associated calls blocking for a relatively long period. The deferred model also results in the window size and position getters not reporting false coordinates anymore, as they only forward what the window manager reports vs allowing applications to set arbitrary values, and fullscreen enter/leave events that were initiated via the window manager update the window state appropriately, where they didn't before. Care was taken to ensure that order of operations is maintained, and that requests are not ignored or dropped. This does require some implicit internal synchronization in the various backends if many requests are made in a short period, as some state and behavior depends on other bits of state that need to be known at that particular point in time, but this isn't something that typical applications will hit, unless they are sending a lot of window state in a short time as the tests do. The automated tests developed to test the previous behavior also resulted in previously undefined behavior being defined and normalized across platforms, particularly when it comes to the sizing and positioning of windows when they are in a fixed-size state, such as maximized or fullscreen. Size and position requests made when the window is not in a movable or resizable state will be deferred until it can be applied, so no requests are lost. These changes fix another long-standing issue with renderers recreating maximized windows, where the original non-maximized size was lost, resulting in the window being restored to the wrong size. All automated video tests pass across all platforms. Overall, the "make a request/get an event" model better reflects how most windowing systems work, and some backends avoid spending significant time blocking while waiting for operations to complete.
2023-10-25 22:09:55 +03:00
result = SDL_SyncWindow(window);
SDLTest_AssertPass("SDL_SyncWindow()");
Removed SDL_bool in favor of plain bool We require stdbool.h in the build environment, so we might as well use the plain bool type. If your environment doesn't have stdbool.h, this simple replacement will suffice: typedef signed char bool;
2024-09-18 17:52:28 +03:00
SDLTest_AssertCheck(result == true, "Verify return value; expected: true, got: %d", result);
video: Implement asynchronous windowing SDL window size, state, and position functions have been considered immediate, with their effects assuming to have taken effect upon successful return of the function. However, several windowing systems handle these requests asynchronously, resulting in the functions blocking until the changes have taken effect, potentially for long periods of time. Additionally, some windowing systems treat these as requests, and can potentially deny or fulfill the request in a manner differently than the application expects, such as not allowing a window to be positioned or sized beyond desktop borders, prohibiting fullscreen, and so on. With these changes, applications can make requests of the window manager that do not block, with the understanding that an associated event will be sent if the request is fulfilled. Currently, size, position, maximize, minimize, and fullscreen calls are handled as asynchronous requests, with events being returned if the request is honored. If the application requires that the change take effect immediately, it can call the new SDL_SyncWindow function, which will attempt to block until the request is fulfilled, or some arbitrary timeout period elapses, the duration of which depends not only on the windowing system, but on the operation requested as well (e.g. a 100ms timeout is fine for most X11 events, but maximizing a window can take considerably longer for some reason). There is also a new hint 'SDL_VIDEO_SYNC_ALL_WINDOW_OPS' that will mimic the old behavior by synchronizing after every window operation with, again, the understanding that using this may result in the associated calls blocking for a relatively long period. The deferred model also results in the window size and position getters not reporting false coordinates anymore, as they only forward what the window manager reports vs allowing applications to set arbitrary values, and fullscreen enter/leave events that were initiated via the window manager update the window state appropriately, where they didn't before. Care was taken to ensure that order of operations is maintained, and that requests are not ignored or dropped. This does require some implicit internal synchronization in the various backends if many requests are made in a short period, as some state and behavior depends on other bits of state that need to be known at that particular point in time, but this isn't something that typical applications will hit, unless they are sending a lot of window state in a short time as the tests do. The automated tests developed to test the previous behavior also resulted in previously undefined behavior being defined and normalized across platforms, particularly when it comes to the sizing and positioning of windows when they are in a fixed-size state, such as maximized or fullscreen. Size and position requests made when the window is not in a movable or resizable state will be deferred until it can be applied, so no requests are lost. These changes fix another long-standing issue with renderers recreating maximized windows, where the original non-maximized size was lost, resulting in the window being restored to the wrong size. All automated video tests pass across all platforms. Overall, the "make a request/get an event" model better reflects how most windowing systems work, and some backends avoid spending significant time blocking while waiting for operations to complete.
2023-10-25 22:09:55 +03:00
flags = SDL_GetWindowFlags(window);
SDLTest_AssertPass("SDL_GetWindowFlags()");
SDLTest_AssertCheck(flags & SDL_WINDOW_FULLSCREEN, "Verify the `SDL_WINDOW_FULLSCREEN` flag is set: %s", (flags & SDL_WINDOW_FULLSCREEN) ? "true" : "false");
SDLTest_AssertCheck(!(flags & SDL_WINDOW_MAXIMIZED), "Verify the `SDL_WINDOW_MAXIMIZED` flag is cleared: %s", !(flags & SDL_WINDOW_MAXIMIZED) ? "true" : "false");
/* Verify the fullscreen size and position */
result = SDL_GetDisplayBounds(SDL_GetDisplayForWindow(window), &display);
SDLTest_AssertPass("SDL_GetDisplayBounds()");
Removed SDL_bool in favor of plain bool We require stdbool.h in the build environment, so we might as well use the plain bool type. If your environment doesn't have stdbool.h, this simple replacement will suffice: typedef signed char bool;
2024-09-18 17:52:28 +03:00
SDLTest_AssertCheck(result == true, "Verify return value; expected: true, got: %d", result);
video: Implement asynchronous windowing SDL window size, state, and position functions have been considered immediate, with their effects assuming to have taken effect upon successful return of the function. However, several windowing systems handle these requests asynchronously, resulting in the functions blocking until the changes have taken effect, potentially for long periods of time. Additionally, some windowing systems treat these as requests, and can potentially deny or fulfill the request in a manner differently than the application expects, such as not allowing a window to be positioned or sized beyond desktop borders, prohibiting fullscreen, and so on. With these changes, applications can make requests of the window manager that do not block, with the understanding that an associated event will be sent if the request is fulfilled. Currently, size, position, maximize, minimize, and fullscreen calls are handled as asynchronous requests, with events being returned if the request is honored. If the application requires that the change take effect immediately, it can call the new SDL_SyncWindow function, which will attempt to block until the request is fulfilled, or some arbitrary timeout period elapses, the duration of which depends not only on the windowing system, but on the operation requested as well (e.g. a 100ms timeout is fine for most X11 events, but maximizing a window can take considerably longer for some reason). There is also a new hint 'SDL_VIDEO_SYNC_ALL_WINDOW_OPS' that will mimic the old behavior by synchronizing after every window operation with, again, the understanding that using this may result in the associated calls blocking for a relatively long period. The deferred model also results in the window size and position getters not reporting false coordinates anymore, as they only forward what the window manager reports vs allowing applications to set arbitrary values, and fullscreen enter/leave events that were initiated via the window manager update the window state appropriately, where they didn't before. Care was taken to ensure that order of operations is maintained, and that requests are not ignored or dropped. This does require some implicit internal synchronization in the various backends if many requests are made in a short period, as some state and behavior depends on other bits of state that need to be known at that particular point in time, but this isn't something that typical applications will hit, unless they are sending a lot of window state in a short time as the tests do. The automated tests developed to test the previous behavior also resulted in previously undefined behavior being defined and normalized across platforms, particularly when it comes to the sizing and positioning of windows when they are in a fixed-size state, such as maximized or fullscreen. Size and position requests made when the window is not in a movable or resizable state will be deferred until it can be applied, so no requests are lost. These changes fix another long-standing issue with renderers recreating maximized windows, where the original non-maximized size was lost, resulting in the window being restored to the wrong size. All automated video tests pass across all platforms. Overall, the "make a request/get an event" model better reflects how most windowing systems work, and some backends avoid spending significant time blocking while waiting for operations to complete.
2023-10-25 22:09:55 +03:00
if (!skipPos) {
desiredX = display.x;
desiredY = display.y;
currentX = windowedX + 1;
currentY = windowedY + 1;
SDL_GetWindowPosition(window, &currentX, &currentY);
SDLTest_AssertPass("Call to SDL_GetWindowPosition()");
SDLTest_AssertCheck(desiredX == currentX, "Verify returned X coordinate; expected: %d, got: %d", desiredX, currentX);
SDLTest_AssertCheck(desiredY == currentY, "Verify returned Y coordinate; expected: %d, got: %d", desiredY, currentY);
}
desiredW = display.w;
desiredH = display.h;
currentW = windowedW + 1;
currentH = windowedH + 1;
SDL_GetWindowSize(window, &currentW, &currentH);
SDLTest_AssertPass("Call to SDL_GetWindowSize()");
SDLTest_AssertCheck(currentW == desiredW, "Verify returned width; expected: %d, got: %d", desiredW, currentW);
SDLTest_AssertCheck(currentH == desiredH, "Verify returned height; expected: %d, got: %d", desiredH, currentH);
/* Leave fullscreen and restore the window */
Removed SDL_bool in favor of plain bool We require stdbool.h in the build environment, so we might as well use the plain bool type. If your environment doesn't have stdbool.h, this simple replacement will suffice: typedef signed char bool;
2024-09-18 17:52:28 +03:00
result = SDL_SetWindowFullscreen(window, false);
SDLTest_AssertPass("SDL_SetWindowFullscreen(false)");
SDLTest_AssertCheck(result == true, "Verify return value; expected: true, got: %d", result);
video: Implement asynchronous windowing SDL window size, state, and position functions have been considered immediate, with their effects assuming to have taken effect upon successful return of the function. However, several windowing systems handle these requests asynchronously, resulting in the functions blocking until the changes have taken effect, potentially for long periods of time. Additionally, some windowing systems treat these as requests, and can potentially deny or fulfill the request in a manner differently than the application expects, such as not allowing a window to be positioned or sized beyond desktop borders, prohibiting fullscreen, and so on. With these changes, applications can make requests of the window manager that do not block, with the understanding that an associated event will be sent if the request is fulfilled. Currently, size, position, maximize, minimize, and fullscreen calls are handled as asynchronous requests, with events being returned if the request is honored. If the application requires that the change take effect immediately, it can call the new SDL_SyncWindow function, which will attempt to block until the request is fulfilled, or some arbitrary timeout period elapses, the duration of which depends not only on the windowing system, but on the operation requested as well (e.g. a 100ms timeout is fine for most X11 events, but maximizing a window can take considerably longer for some reason). There is also a new hint 'SDL_VIDEO_SYNC_ALL_WINDOW_OPS' that will mimic the old behavior by synchronizing after every window operation with, again, the understanding that using this may result in the associated calls blocking for a relatively long period. The deferred model also results in the window size and position getters not reporting false coordinates anymore, as they only forward what the window manager reports vs allowing applications to set arbitrary values, and fullscreen enter/leave events that were initiated via the window manager update the window state appropriately, where they didn't before. Care was taken to ensure that order of operations is maintained, and that requests are not ignored or dropped. This does require some implicit internal synchronization in the various backends if many requests are made in a short period, as some state and behavior depends on other bits of state that need to be known at that particular point in time, but this isn't something that typical applications will hit, unless they are sending a lot of window state in a short time as the tests do. The automated tests developed to test the previous behavior also resulted in previously undefined behavior being defined and normalized across platforms, particularly when it comes to the sizing and positioning of windows when they are in a fixed-size state, such as maximized or fullscreen. Size and position requests made when the window is not in a movable or resizable state will be deferred until it can be applied, so no requests are lost. These changes fix another long-standing issue with renderers recreating maximized windows, where the original non-maximized size was lost, resulting in the window being restored to the wrong size. All automated video tests pass across all platforms. Overall, the "make a request/get an event" model better reflects how most windowing systems work, and some backends avoid spending significant time blocking while waiting for operations to complete.
2023-10-25 22:09:55 +03:00
result = SDL_RestoreWindow(window);
SDLTest_AssertPass("SDL_RestoreWindow()");
Removed SDL_bool in favor of plain bool We require stdbool.h in the build environment, so we might as well use the plain bool type. If your environment doesn't have stdbool.h, this simple replacement will suffice: typedef signed char bool;
2024-09-18 17:52:28 +03:00
SDLTest_AssertCheck(result == true, "Verify return value; expected: true, got: %d", result);
video: Implement asynchronous windowing SDL window size, state, and position functions have been considered immediate, with their effects assuming to have taken effect upon successful return of the function. However, several windowing systems handle these requests asynchronously, resulting in the functions blocking until the changes have taken effect, potentially for long periods of time. Additionally, some windowing systems treat these as requests, and can potentially deny or fulfill the request in a manner differently than the application expects, such as not allowing a window to be positioned or sized beyond desktop borders, prohibiting fullscreen, and so on. With these changes, applications can make requests of the window manager that do not block, with the understanding that an associated event will be sent if the request is fulfilled. Currently, size, position, maximize, minimize, and fullscreen calls are handled as asynchronous requests, with events being returned if the request is honored. If the application requires that the change take effect immediately, it can call the new SDL_SyncWindow function, which will attempt to block until the request is fulfilled, or some arbitrary timeout period elapses, the duration of which depends not only on the windowing system, but on the operation requested as well (e.g. a 100ms timeout is fine for most X11 events, but maximizing a window can take considerably longer for some reason). There is also a new hint 'SDL_VIDEO_SYNC_ALL_WINDOW_OPS' that will mimic the old behavior by synchronizing after every window operation with, again, the understanding that using this may result in the associated calls blocking for a relatively long period. The deferred model also results in the window size and position getters not reporting false coordinates anymore, as they only forward what the window manager reports vs allowing applications to set arbitrary values, and fullscreen enter/leave events that were initiated via the window manager update the window state appropriately, where they didn't before. Care was taken to ensure that order of operations is maintained, and that requests are not ignored or dropped. This does require some implicit internal synchronization in the various backends if many requests are made in a short period, as some state and behavior depends on other bits of state that need to be known at that particular point in time, but this isn't something that typical applications will hit, unless they are sending a lot of window state in a short time as the tests do. The automated tests developed to test the previous behavior also resulted in previously undefined behavior being defined and normalized across platforms, particularly when it comes to the sizing and positioning of windows when they are in a fixed-size state, such as maximized or fullscreen. Size and position requests made when the window is not in a movable or resizable state will be deferred until it can be applied, so no requests are lost. These changes fix another long-standing issue with renderers recreating maximized windows, where the original non-maximized size was lost, resulting in the window being restored to the wrong size. All automated video tests pass across all platforms. Overall, the "make a request/get an event" model better reflects how most windowing systems work, and some backends avoid spending significant time blocking while waiting for operations to complete.
2023-10-25 22:09:55 +03:00
result = SDL_SyncWindow(window);
SDLTest_AssertPass("SDL_SyncWindow()");
Removed SDL_bool in favor of plain bool We require stdbool.h in the build environment, so we might as well use the plain bool type. If your environment doesn't have stdbool.h, this simple replacement will suffice: typedef signed char bool;
2024-09-18 17:52:28 +03:00
SDLTest_AssertCheck(result == true, "Verify return value; expected: true, got: %d", result);
video: Implement asynchronous windowing SDL window size, state, and position functions have been considered immediate, with their effects assuming to have taken effect upon successful return of the function. However, several windowing systems handle these requests asynchronously, resulting in the functions blocking until the changes have taken effect, potentially for long periods of time. Additionally, some windowing systems treat these as requests, and can potentially deny or fulfill the request in a manner differently than the application expects, such as not allowing a window to be positioned or sized beyond desktop borders, prohibiting fullscreen, and so on. With these changes, applications can make requests of the window manager that do not block, with the understanding that an associated event will be sent if the request is fulfilled. Currently, size, position, maximize, minimize, and fullscreen calls are handled as asynchronous requests, with events being returned if the request is honored. If the application requires that the change take effect immediately, it can call the new SDL_SyncWindow function, which will attempt to block until the request is fulfilled, or some arbitrary timeout period elapses, the duration of which depends not only on the windowing system, but on the operation requested as well (e.g. a 100ms timeout is fine for most X11 events, but maximizing a window can take considerably longer for some reason). There is also a new hint 'SDL_VIDEO_SYNC_ALL_WINDOW_OPS' that will mimic the old behavior by synchronizing after every window operation with, again, the understanding that using this may result in the associated calls blocking for a relatively long period. The deferred model also results in the window size and position getters not reporting false coordinates anymore, as they only forward what the window manager reports vs allowing applications to set arbitrary values, and fullscreen enter/leave events that were initiated via the window manager update the window state appropriately, where they didn't before. Care was taken to ensure that order of operations is maintained, and that requests are not ignored or dropped. This does require some implicit internal synchronization in the various backends if many requests are made in a short period, as some state and behavior depends on other bits of state that need to be known at that particular point in time, but this isn't something that typical applications will hit, unless they are sending a lot of window state in a short time as the tests do. The automated tests developed to test the previous behavior also resulted in previously undefined behavior being defined and normalized across platforms, particularly when it comes to the sizing and positioning of windows when they are in a fixed-size state, such as maximized or fullscreen. Size and position requests made when the window is not in a movable or resizable state will be deferred until it can be applied, so no requests are lost. These changes fix another long-standing issue with renderers recreating maximized windows, where the original non-maximized size was lost, resulting in the window being restored to the wrong size. All automated video tests pass across all platforms. Overall, the "make a request/get an event" model better reflects how most windowing systems work, and some backends avoid spending significant time blocking while waiting for operations to complete.
2023-10-25 22:09:55 +03:00
flags = SDL_GetWindowFlags(window);
SDLTest_AssertPass("SDL_GetWindowFlags()");
SDLTest_AssertCheck(!(flags & SDL_WINDOW_MAXIMIZED), "Verify that the `SDL_WINDOW_MAXIMIZED` flag is cleared: %s", !(flags & SDL_WINDOW_MAXIMIZED) ? "true" : "false");
/* Make sure the restored size and position matches the original windowed size and position. */
if (!skipPos) {
currentX = windowedX + 1;
currentY = windowedY + 1;
SDL_GetWindowPosition(window, &currentX, &currentY);
SDLTest_AssertPass("Call to SDL_GetWindowPosition()");
SDLTest_AssertCheck(windowedX == currentX, "Verify returned X coordinate; expected: %d, got: %d", windowedX, currentX);
SDLTest_AssertCheck(windowedY == currentY, "Verify returned Y coordinate; expected: %d, got: %d", windowedY, currentY);
}
currentW = windowedW + 1;
currentH = windowedH + 1;
SDL_GetWindowSize(window, &currentW, &currentH);
SDLTest_AssertPass("Call to SDL_GetWindowSize()");
SDLTest_AssertCheck(windowedW == currentW, "Verify returned width; expected: %d, got: %d", windowedW, currentW);
SDLTest_AssertCheck(windowedH == currentH, "Verify returned height; expected: %d, got: %d", windowedH, currentH);
/* Maximize, change size, and restore */
result = SDL_MaximizeWindow(window);
SDLTest_AssertPass("SDL_MaximizeWindow()");
Removed SDL_bool in favor of plain bool We require stdbool.h in the build environment, so we might as well use the plain bool type. If your environment doesn't have stdbool.h, this simple replacement will suffice: typedef signed char bool;
2024-09-18 17:52:28 +03:00
SDLTest_AssertCheck(result == true, "Verify return value; expected: true, got: %d", result);
video: Implement asynchronous windowing SDL window size, state, and position functions have been considered immediate, with their effects assuming to have taken effect upon successful return of the function. However, several windowing systems handle these requests asynchronously, resulting in the functions blocking until the changes have taken effect, potentially for long periods of time. Additionally, some windowing systems treat these as requests, and can potentially deny or fulfill the request in a manner differently than the application expects, such as not allowing a window to be positioned or sized beyond desktop borders, prohibiting fullscreen, and so on. With these changes, applications can make requests of the window manager that do not block, with the understanding that an associated event will be sent if the request is fulfilled. Currently, size, position, maximize, minimize, and fullscreen calls are handled as asynchronous requests, with events being returned if the request is honored. If the application requires that the change take effect immediately, it can call the new SDL_SyncWindow function, which will attempt to block until the request is fulfilled, or some arbitrary timeout period elapses, the duration of which depends not only on the windowing system, but on the operation requested as well (e.g. a 100ms timeout is fine for most X11 events, but maximizing a window can take considerably longer for some reason). There is also a new hint 'SDL_VIDEO_SYNC_ALL_WINDOW_OPS' that will mimic the old behavior by synchronizing after every window operation with, again, the understanding that using this may result in the associated calls blocking for a relatively long period. The deferred model also results in the window size and position getters not reporting false coordinates anymore, as they only forward what the window manager reports vs allowing applications to set arbitrary values, and fullscreen enter/leave events that were initiated via the window manager update the window state appropriately, where they didn't before. Care was taken to ensure that order of operations is maintained, and that requests are not ignored or dropped. This does require some implicit internal synchronization in the various backends if many requests are made in a short period, as some state and behavior depends on other bits of state that need to be known at that particular point in time, but this isn't something that typical applications will hit, unless they are sending a lot of window state in a short time as the tests do. The automated tests developed to test the previous behavior also resulted in previously undefined behavior being defined and normalized across platforms, particularly when it comes to the sizing and positioning of windows when they are in a fixed-size state, such as maximized or fullscreen. Size and position requests made when the window is not in a movable or resizable state will be deferred until it can be applied, so no requests are lost. These changes fix another long-standing issue with renderers recreating maximized windows, where the original non-maximized size was lost, resulting in the window being restored to the wrong size. All automated video tests pass across all platforms. Overall, the "make a request/get an event" model better reflects how most windowing systems work, and some backends avoid spending significant time blocking while waiting for operations to complete.
2023-10-25 22:09:55 +03:00
desiredW = windowedW + 10;
desiredH = windowedH + 10;
result = SDL_SetWindowSize(window, desiredW, desiredH);
SDLTest_AssertPass("SDL_SetWindowSize()");
Removed SDL_bool in favor of plain bool We require stdbool.h in the build environment, so we might as well use the plain bool type. If your environment doesn't have stdbool.h, this simple replacement will suffice: typedef signed char bool;
2024-09-18 17:52:28 +03:00
SDLTest_AssertCheck(result == true, "Verify return value; expected: true, got: %d", result);
video: Implement asynchronous windowing SDL window size, state, and position functions have been considered immediate, with their effects assuming to have taken effect upon successful return of the function. However, several windowing systems handle these requests asynchronously, resulting in the functions blocking until the changes have taken effect, potentially for long periods of time. Additionally, some windowing systems treat these as requests, and can potentially deny or fulfill the request in a manner differently than the application expects, such as not allowing a window to be positioned or sized beyond desktop borders, prohibiting fullscreen, and so on. With these changes, applications can make requests of the window manager that do not block, with the understanding that an associated event will be sent if the request is fulfilled. Currently, size, position, maximize, minimize, and fullscreen calls are handled as asynchronous requests, with events being returned if the request is honored. If the application requires that the change take effect immediately, it can call the new SDL_SyncWindow function, which will attempt to block until the request is fulfilled, or some arbitrary timeout period elapses, the duration of which depends not only on the windowing system, but on the operation requested as well (e.g. a 100ms timeout is fine for most X11 events, but maximizing a window can take considerably longer for some reason). There is also a new hint 'SDL_VIDEO_SYNC_ALL_WINDOW_OPS' that will mimic the old behavior by synchronizing after every window operation with, again, the understanding that using this may result in the associated calls blocking for a relatively long period. The deferred model also results in the window size and position getters not reporting false coordinates anymore, as they only forward what the window manager reports vs allowing applications to set arbitrary values, and fullscreen enter/leave events that were initiated via the window manager update the window state appropriately, where they didn't before. Care was taken to ensure that order of operations is maintained, and that requests are not ignored or dropped. This does require some implicit internal synchronization in the various backends if many requests are made in a short period, as some state and behavior depends on other bits of state that need to be known at that particular point in time, but this isn't something that typical applications will hit, unless they are sending a lot of window state in a short time as the tests do. The automated tests developed to test the previous behavior also resulted in previously undefined behavior being defined and normalized across platforms, particularly when it comes to the sizing and positioning of windows when they are in a fixed-size state, such as maximized or fullscreen. Size and position requests made when the window is not in a movable or resizable state will be deferred until it can be applied, so no requests are lost. These changes fix another long-standing issue with renderers recreating maximized windows, where the original non-maximized size was lost, resulting in the window being restored to the wrong size. All automated video tests pass across all platforms. Overall, the "make a request/get an event" model better reflects how most windowing systems work, and some backends avoid spending significant time blocking while waiting for operations to complete.
2023-10-25 22:09:55 +03:00
if (!skipPos) {
desiredX = windowedX + 10;
desiredY = windowedY + 10;
result = SDL_SetWindowPosition(window, desiredX, desiredY);
SDLTest_AssertPass("SDL_SetWindowPosition()");
Removed SDL_bool in favor of plain bool We require stdbool.h in the build environment, so we might as well use the plain bool type. If your environment doesn't have stdbool.h, this simple replacement will suffice: typedef signed char bool;
2024-09-18 17:52:28 +03:00
SDLTest_AssertCheck(result == true, "Verify return value; expected: true, got: %d", result);
video: Implement asynchronous windowing SDL window size, state, and position functions have been considered immediate, with their effects assuming to have taken effect upon successful return of the function. However, several windowing systems handle these requests asynchronously, resulting in the functions blocking until the changes have taken effect, potentially for long periods of time. Additionally, some windowing systems treat these as requests, and can potentially deny or fulfill the request in a manner differently than the application expects, such as not allowing a window to be positioned or sized beyond desktop borders, prohibiting fullscreen, and so on. With these changes, applications can make requests of the window manager that do not block, with the understanding that an associated event will be sent if the request is fulfilled. Currently, size, position, maximize, minimize, and fullscreen calls are handled as asynchronous requests, with events being returned if the request is honored. If the application requires that the change take effect immediately, it can call the new SDL_SyncWindow function, which will attempt to block until the request is fulfilled, or some arbitrary timeout period elapses, the duration of which depends not only on the windowing system, but on the operation requested as well (e.g. a 100ms timeout is fine for most X11 events, but maximizing a window can take considerably longer for some reason). There is also a new hint 'SDL_VIDEO_SYNC_ALL_WINDOW_OPS' that will mimic the old behavior by synchronizing after every window operation with, again, the understanding that using this may result in the associated calls blocking for a relatively long period. The deferred model also results in the window size and position getters not reporting false coordinates anymore, as they only forward what the window manager reports vs allowing applications to set arbitrary values, and fullscreen enter/leave events that were initiated via the window manager update the window state appropriately, where they didn't before. Care was taken to ensure that order of operations is maintained, and that requests are not ignored or dropped. This does require some implicit internal synchronization in the various backends if many requests are made in a short period, as some state and behavior depends on other bits of state that need to be known at that particular point in time, but this isn't something that typical applications will hit, unless they are sending a lot of window state in a short time as the tests do. The automated tests developed to test the previous behavior also resulted in previously undefined behavior being defined and normalized across platforms, particularly when it comes to the sizing and positioning of windows when they are in a fixed-size state, such as maximized or fullscreen. Size and position requests made when the window is not in a movable or resizable state will be deferred until it can be applied, so no requests are lost. These changes fix another long-standing issue with renderers recreating maximized windows, where the original non-maximized size was lost, resulting in the window being restored to the wrong size. All automated video tests pass across all platforms. Overall, the "make a request/get an event" model better reflects how most windowing systems work, and some backends avoid spending significant time blocking while waiting for operations to complete.
2023-10-25 22:09:55 +03:00
}
result = SDL_RestoreWindow(window);
SDLTest_AssertPass("SDL_RestoreWindow()");
Removed SDL_bool in favor of plain bool We require stdbool.h in the build environment, so we might as well use the plain bool type. If your environment doesn't have stdbool.h, this simple replacement will suffice: typedef signed char bool;
2024-09-18 17:52:28 +03:00
SDLTest_AssertCheck(result == true, "Verify return value; expected: true, got: %d", result);
video: Implement asynchronous windowing SDL window size, state, and position functions have been considered immediate, with their effects assuming to have taken effect upon successful return of the function. However, several windowing systems handle these requests asynchronously, resulting in the functions blocking until the changes have taken effect, potentially for long periods of time. Additionally, some windowing systems treat these as requests, and can potentially deny or fulfill the request in a manner differently than the application expects, such as not allowing a window to be positioned or sized beyond desktop borders, prohibiting fullscreen, and so on. With these changes, applications can make requests of the window manager that do not block, with the understanding that an associated event will be sent if the request is fulfilled. Currently, size, position, maximize, minimize, and fullscreen calls are handled as asynchronous requests, with events being returned if the request is honored. If the application requires that the change take effect immediately, it can call the new SDL_SyncWindow function, which will attempt to block until the request is fulfilled, or some arbitrary timeout period elapses, the duration of which depends not only on the windowing system, but on the operation requested as well (e.g. a 100ms timeout is fine for most X11 events, but maximizing a window can take considerably longer for some reason). There is also a new hint 'SDL_VIDEO_SYNC_ALL_WINDOW_OPS' that will mimic the old behavior by synchronizing after every window operation with, again, the understanding that using this may result in the associated calls blocking for a relatively long period. The deferred model also results in the window size and position getters not reporting false coordinates anymore, as they only forward what the window manager reports vs allowing applications to set arbitrary values, and fullscreen enter/leave events that were initiated via the window manager update the window state appropriately, where they didn't before. Care was taken to ensure that order of operations is maintained, and that requests are not ignored or dropped. This does require some implicit internal synchronization in the various backends if many requests are made in a short period, as some state and behavior depends on other bits of state that need to be known at that particular point in time, but this isn't something that typical applications will hit, unless they are sending a lot of window state in a short time as the tests do. The automated tests developed to test the previous behavior also resulted in previously undefined behavior being defined and normalized across platforms, particularly when it comes to the sizing and positioning of windows when they are in a fixed-size state, such as maximized or fullscreen. Size and position requests made when the window is not in a movable or resizable state will be deferred until it can be applied, so no requests are lost. These changes fix another long-standing issue with renderers recreating maximized windows, where the original non-maximized size was lost, resulting in the window being restored to the wrong size. All automated video tests pass across all platforms. Overall, the "make a request/get an event" model better reflects how most windowing systems work, and some backends avoid spending significant time blocking while waiting for operations to complete.
2023-10-25 22:09:55 +03:00
result = SDL_SyncWindow(window);
SDLTest_AssertPass("SDL_SyncWindow()");
Removed SDL_bool in favor of plain bool We require stdbool.h in the build environment, so we might as well use the plain bool type. If your environment doesn't have stdbool.h, this simple replacement will suffice: typedef signed char bool;
2024-09-18 17:52:28 +03:00
SDLTest_AssertCheck(result == true, "Verify return value; expected: true, got: %d", result);
video: Implement asynchronous windowing SDL window size, state, and position functions have been considered immediate, with their effects assuming to have taken effect upon successful return of the function. However, several windowing systems handle these requests asynchronously, resulting in the functions blocking until the changes have taken effect, potentially for long periods of time. Additionally, some windowing systems treat these as requests, and can potentially deny or fulfill the request in a manner differently than the application expects, such as not allowing a window to be positioned or sized beyond desktop borders, prohibiting fullscreen, and so on. With these changes, applications can make requests of the window manager that do not block, with the understanding that an associated event will be sent if the request is fulfilled. Currently, size, position, maximize, minimize, and fullscreen calls are handled as asynchronous requests, with events being returned if the request is honored. If the application requires that the change take effect immediately, it can call the new SDL_SyncWindow function, which will attempt to block until the request is fulfilled, or some arbitrary timeout period elapses, the duration of which depends not only on the windowing system, but on the operation requested as well (e.g. a 100ms timeout is fine for most X11 events, but maximizing a window can take considerably longer for some reason). There is also a new hint 'SDL_VIDEO_SYNC_ALL_WINDOW_OPS' that will mimic the old behavior by synchronizing after every window operation with, again, the understanding that using this may result in the associated calls blocking for a relatively long period. The deferred model also results in the window size and position getters not reporting false coordinates anymore, as they only forward what the window manager reports vs allowing applications to set arbitrary values, and fullscreen enter/leave events that were initiated via the window manager update the window state appropriately, where they didn't before. Care was taken to ensure that order of operations is maintained, and that requests are not ignored or dropped. This does require some implicit internal synchronization in the various backends if many requests are made in a short period, as some state and behavior depends on other bits of state that need to be known at that particular point in time, but this isn't something that typical applications will hit, unless they are sending a lot of window state in a short time as the tests do. The automated tests developed to test the previous behavior also resulted in previously undefined behavior being defined and normalized across platforms, particularly when it comes to the sizing and positioning of windows when they are in a fixed-size state, such as maximized or fullscreen. Size and position requests made when the window is not in a movable or resizable state will be deferred until it can be applied, so no requests are lost. These changes fix another long-standing issue with renderers recreating maximized windows, where the original non-maximized size was lost, resulting in the window being restored to the wrong size. All automated video tests pass across all platforms. Overall, the "make a request/get an event" model better reflects how most windowing systems work, and some backends avoid spending significant time blocking while waiting for operations to complete.
2023-10-25 22:09:55 +03:00
flags = SDL_GetWindowFlags(window);
SDLTest_AssertPass("SDL_GetWindowFlags()");
SDLTest_AssertCheck(!(flags & SDL_WINDOW_MAXIMIZED), "Verify that the `SDL_WINDOW_MAXIMIZED` flag is cleared: %s", !(flags & SDL_WINDOW_MAXIMIZED) ? "true" : "false");
if (!skipPos) {
currentX = desiredX + 1;
currentY = desiredY + 1;
SDL_GetWindowPosition(window, &currentX, &currentY);
SDLTest_AssertPass("Call to SDL_GetWindowPosition()");
SDLTest_AssertCheck(desiredX == currentX, "Verify returned X coordinate; expected: %d, got: %d", desiredX, currentX);
SDLTest_AssertCheck(desiredY == currentY, "Verify returned Y coordinate; expected: %d, got: %d", desiredY, currentY);
}
currentW = desiredW + 1;
currentH = desiredH + 1;
SDL_GetWindowSize(window, &currentW, &currentH);
SDLTest_AssertPass("Call to SDL_GetWindowSize()");
SDLTest_AssertCheck(desiredW == currentW, "Verify returned width; expected: %d, got: %d", desiredW, currentW);
SDLTest_AssertCheck(desiredH == currentH, "Verify returned height; expected: %d, got: %d", desiredH, currentH);
/* Change size and position, maximize and restore */
desiredW = windowedW - 5;
desiredH = windowedH - 5;
result = SDL_SetWindowSize(window, desiredW, desiredH);
SDLTest_AssertPass("SDL_SetWindowSize()");
Removed SDL_bool in favor of plain bool We require stdbool.h in the build environment, so we might as well use the plain bool type. If your environment doesn't have stdbool.h, this simple replacement will suffice: typedef signed char bool;
2024-09-18 17:52:28 +03:00
SDLTest_AssertCheck(result == true, "Verify return value; expected: true, got: %d", result);
video: Implement asynchronous windowing SDL window size, state, and position functions have been considered immediate, with their effects assuming to have taken effect upon successful return of the function. However, several windowing systems handle these requests asynchronously, resulting in the functions blocking until the changes have taken effect, potentially for long periods of time. Additionally, some windowing systems treat these as requests, and can potentially deny or fulfill the request in a manner differently than the application expects, such as not allowing a window to be positioned or sized beyond desktop borders, prohibiting fullscreen, and so on. With these changes, applications can make requests of the window manager that do not block, with the understanding that an associated event will be sent if the request is fulfilled. Currently, size, position, maximize, minimize, and fullscreen calls are handled as asynchronous requests, with events being returned if the request is honored. If the application requires that the change take effect immediately, it can call the new SDL_SyncWindow function, which will attempt to block until the request is fulfilled, or some arbitrary timeout period elapses, the duration of which depends not only on the windowing system, but on the operation requested as well (e.g. a 100ms timeout is fine for most X11 events, but maximizing a window can take considerably longer for some reason). There is also a new hint 'SDL_VIDEO_SYNC_ALL_WINDOW_OPS' that will mimic the old behavior by synchronizing after every window operation with, again, the understanding that using this may result in the associated calls blocking for a relatively long period. The deferred model also results in the window size and position getters not reporting false coordinates anymore, as they only forward what the window manager reports vs allowing applications to set arbitrary values, and fullscreen enter/leave events that were initiated via the window manager update the window state appropriately, where they didn't before. Care was taken to ensure that order of operations is maintained, and that requests are not ignored or dropped. This does require some implicit internal synchronization in the various backends if many requests are made in a short period, as some state and behavior depends on other bits of state that need to be known at that particular point in time, but this isn't something that typical applications will hit, unless they are sending a lot of window state in a short time as the tests do. The automated tests developed to test the previous behavior also resulted in previously undefined behavior being defined and normalized across platforms, particularly when it comes to the sizing and positioning of windows when they are in a fixed-size state, such as maximized or fullscreen. Size and position requests made when the window is not in a movable or resizable state will be deferred until it can be applied, so no requests are lost. These changes fix another long-standing issue with renderers recreating maximized windows, where the original non-maximized size was lost, resulting in the window being restored to the wrong size. All automated video tests pass across all platforms. Overall, the "make a request/get an event" model better reflects how most windowing systems work, and some backends avoid spending significant time blocking while waiting for operations to complete.
2023-10-25 22:09:55 +03:00
if (!skipPos) {
desiredX = windowedX + 5;
desiredY = windowedY + 5;
result = SDL_SetWindowPosition(window, desiredX, desiredY);
SDLTest_AssertPass("SDL_SetWindowPosition()");
Removed SDL_bool in favor of plain bool We require stdbool.h in the build environment, so we might as well use the plain bool type. If your environment doesn't have stdbool.h, this simple replacement will suffice: typedef signed char bool;
2024-09-18 17:52:28 +03:00
SDLTest_AssertCheck(result == true, "Verify return value; expected: true, got: %d", result);
video: Implement asynchronous windowing SDL window size, state, and position functions have been considered immediate, with their effects assuming to have taken effect upon successful return of the function. However, several windowing systems handle these requests asynchronously, resulting in the functions blocking until the changes have taken effect, potentially for long periods of time. Additionally, some windowing systems treat these as requests, and can potentially deny or fulfill the request in a manner differently than the application expects, such as not allowing a window to be positioned or sized beyond desktop borders, prohibiting fullscreen, and so on. With these changes, applications can make requests of the window manager that do not block, with the understanding that an associated event will be sent if the request is fulfilled. Currently, size, position, maximize, minimize, and fullscreen calls are handled as asynchronous requests, with events being returned if the request is honored. If the application requires that the change take effect immediately, it can call the new SDL_SyncWindow function, which will attempt to block until the request is fulfilled, or some arbitrary timeout period elapses, the duration of which depends not only on the windowing system, but on the operation requested as well (e.g. a 100ms timeout is fine for most X11 events, but maximizing a window can take considerably longer for some reason). There is also a new hint 'SDL_VIDEO_SYNC_ALL_WINDOW_OPS' that will mimic the old behavior by synchronizing after every window operation with, again, the understanding that using this may result in the associated calls blocking for a relatively long period. The deferred model also results in the window size and position getters not reporting false coordinates anymore, as they only forward what the window manager reports vs allowing applications to set arbitrary values, and fullscreen enter/leave events that were initiated via the window manager update the window state appropriately, where they didn't before. Care was taken to ensure that order of operations is maintained, and that requests are not ignored or dropped. This does require some implicit internal synchronization in the various backends if many requests are made in a short period, as some state and behavior depends on other bits of state that need to be known at that particular point in time, but this isn't something that typical applications will hit, unless they are sending a lot of window state in a short time as the tests do. The automated tests developed to test the previous behavior also resulted in previously undefined behavior being defined and normalized across platforms, particularly when it comes to the sizing and positioning of windows when they are in a fixed-size state, such as maximized or fullscreen. Size and position requests made when the window is not in a movable or resizable state will be deferred until it can be applied, so no requests are lost. These changes fix another long-standing issue with renderers recreating maximized windows, where the original non-maximized size was lost, resulting in the window being restored to the wrong size. All automated video tests pass across all platforms. Overall, the "make a request/get an event" model better reflects how most windowing systems work, and some backends avoid spending significant time blocking while waiting for operations to complete.
2023-10-25 22:09:55 +03:00
}
result = SDL_MaximizeWindow(window);
SDLTest_AssertPass("SDL_MaximizeWindow()");
Removed SDL_bool in favor of plain bool We require stdbool.h in the build environment, so we might as well use the plain bool type. If your environment doesn't have stdbool.h, this simple replacement will suffice: typedef signed char bool;
2024-09-18 17:52:28 +03:00
SDLTest_AssertCheck(result == true, "Verify return value; expected: true, got: %d", result);
video: Implement asynchronous windowing SDL window size, state, and position functions have been considered immediate, with their effects assuming to have taken effect upon successful return of the function. However, several windowing systems handle these requests asynchronously, resulting in the functions blocking until the changes have taken effect, potentially for long periods of time. Additionally, some windowing systems treat these as requests, and can potentially deny or fulfill the request in a manner differently than the application expects, such as not allowing a window to be positioned or sized beyond desktop borders, prohibiting fullscreen, and so on. With these changes, applications can make requests of the window manager that do not block, with the understanding that an associated event will be sent if the request is fulfilled. Currently, size, position, maximize, minimize, and fullscreen calls are handled as asynchronous requests, with events being returned if the request is honored. If the application requires that the change take effect immediately, it can call the new SDL_SyncWindow function, which will attempt to block until the request is fulfilled, or some arbitrary timeout period elapses, the duration of which depends not only on the windowing system, but on the operation requested as well (e.g. a 100ms timeout is fine for most X11 events, but maximizing a window can take considerably longer for some reason). There is also a new hint 'SDL_VIDEO_SYNC_ALL_WINDOW_OPS' that will mimic the old behavior by synchronizing after every window operation with, again, the understanding that using this may result in the associated calls blocking for a relatively long period. The deferred model also results in the window size and position getters not reporting false coordinates anymore, as they only forward what the window manager reports vs allowing applications to set arbitrary values, and fullscreen enter/leave events that were initiated via the window manager update the window state appropriately, where they didn't before. Care was taken to ensure that order of operations is maintained, and that requests are not ignored or dropped. This does require some implicit internal synchronization in the various backends if many requests are made in a short period, as some state and behavior depends on other bits of state that need to be known at that particular point in time, but this isn't something that typical applications will hit, unless they are sending a lot of window state in a short time as the tests do. The automated tests developed to test the previous behavior also resulted in previously undefined behavior being defined and normalized across platforms, particularly when it comes to the sizing and positioning of windows when they are in a fixed-size state, such as maximized or fullscreen. Size and position requests made when the window is not in a movable or resizable state will be deferred until it can be applied, so no requests are lost. These changes fix another long-standing issue with renderers recreating maximized windows, where the original non-maximized size was lost, resulting in the window being restored to the wrong size. All automated video tests pass across all platforms. Overall, the "make a request/get an event" model better reflects how most windowing systems work, and some backends avoid spending significant time blocking while waiting for operations to complete.
2023-10-25 22:09:55 +03:00
result = SDL_RestoreWindow(window);
SDLTest_AssertPass("SDL_RestoreWindow()");
Removed SDL_bool in favor of plain bool We require stdbool.h in the build environment, so we might as well use the plain bool type. If your environment doesn't have stdbool.h, this simple replacement will suffice: typedef signed char bool;
2024-09-18 17:52:28 +03:00
SDLTest_AssertCheck(result == true, "Verify return value; expected: true, got: %d", result);
video: Implement asynchronous windowing SDL window size, state, and position functions have been considered immediate, with their effects assuming to have taken effect upon successful return of the function. However, several windowing systems handle these requests asynchronously, resulting in the functions blocking until the changes have taken effect, potentially for long periods of time. Additionally, some windowing systems treat these as requests, and can potentially deny or fulfill the request in a manner differently than the application expects, such as not allowing a window to be positioned or sized beyond desktop borders, prohibiting fullscreen, and so on. With these changes, applications can make requests of the window manager that do not block, with the understanding that an associated event will be sent if the request is fulfilled. Currently, size, position, maximize, minimize, and fullscreen calls are handled as asynchronous requests, with events being returned if the request is honored. If the application requires that the change take effect immediately, it can call the new SDL_SyncWindow function, which will attempt to block until the request is fulfilled, or some arbitrary timeout period elapses, the duration of which depends not only on the windowing system, but on the operation requested as well (e.g. a 100ms timeout is fine for most X11 events, but maximizing a window can take considerably longer for some reason). There is also a new hint 'SDL_VIDEO_SYNC_ALL_WINDOW_OPS' that will mimic the old behavior by synchronizing after every window operation with, again, the understanding that using this may result in the associated calls blocking for a relatively long period. The deferred model also results in the window size and position getters not reporting false coordinates anymore, as they only forward what the window manager reports vs allowing applications to set arbitrary values, and fullscreen enter/leave events that were initiated via the window manager update the window state appropriately, where they didn't before. Care was taken to ensure that order of operations is maintained, and that requests are not ignored or dropped. This does require some implicit internal synchronization in the various backends if many requests are made in a short period, as some state and behavior depends on other bits of state that need to be known at that particular point in time, but this isn't something that typical applications will hit, unless they are sending a lot of window state in a short time as the tests do. The automated tests developed to test the previous behavior also resulted in previously undefined behavior being defined and normalized across platforms, particularly when it comes to the sizing and positioning of windows when they are in a fixed-size state, such as maximized or fullscreen. Size and position requests made when the window is not in a movable or resizable state will be deferred until it can be applied, so no requests are lost. These changes fix another long-standing issue with renderers recreating maximized windows, where the original non-maximized size was lost, resulting in the window being restored to the wrong size. All automated video tests pass across all platforms. Overall, the "make a request/get an event" model better reflects how most windowing systems work, and some backends avoid spending significant time blocking while waiting for operations to complete.
2023-10-25 22:09:55 +03:00
result = SDL_SyncWindow(window);
SDLTest_AssertPass("SDL_SyncWindow()");
Removed SDL_bool in favor of plain bool We require stdbool.h in the build environment, so we might as well use the plain bool type. If your environment doesn't have stdbool.h, this simple replacement will suffice: typedef signed char bool;
2024-09-18 17:52:28 +03:00
SDLTest_AssertCheck(result == true, "Verify return value; expected: true, got: %d", result);
video: Implement asynchronous windowing SDL window size, state, and position functions have been considered immediate, with their effects assuming to have taken effect upon successful return of the function. However, several windowing systems handle these requests asynchronously, resulting in the functions blocking until the changes have taken effect, potentially for long periods of time. Additionally, some windowing systems treat these as requests, and can potentially deny or fulfill the request in a manner differently than the application expects, such as not allowing a window to be positioned or sized beyond desktop borders, prohibiting fullscreen, and so on. With these changes, applications can make requests of the window manager that do not block, with the understanding that an associated event will be sent if the request is fulfilled. Currently, size, position, maximize, minimize, and fullscreen calls are handled as asynchronous requests, with events being returned if the request is honored. If the application requires that the change take effect immediately, it can call the new SDL_SyncWindow function, which will attempt to block until the request is fulfilled, or some arbitrary timeout period elapses, the duration of which depends not only on the windowing system, but on the operation requested as well (e.g. a 100ms timeout is fine for most X11 events, but maximizing a window can take considerably longer for some reason). There is also a new hint 'SDL_VIDEO_SYNC_ALL_WINDOW_OPS' that will mimic the old behavior by synchronizing after every window operation with, again, the understanding that using this may result in the associated calls blocking for a relatively long period. The deferred model also results in the window size and position getters not reporting false coordinates anymore, as they only forward what the window manager reports vs allowing applications to set arbitrary values, and fullscreen enter/leave events that were initiated via the window manager update the window state appropriately, where they didn't before. Care was taken to ensure that order of operations is maintained, and that requests are not ignored or dropped. This does require some implicit internal synchronization in the various backends if many requests are made in a short period, as some state and behavior depends on other bits of state that need to be known at that particular point in time, but this isn't something that typical applications will hit, unless they are sending a lot of window state in a short time as the tests do. The automated tests developed to test the previous behavior also resulted in previously undefined behavior being defined and normalized across platforms, particularly when it comes to the sizing and positioning of windows when they are in a fixed-size state, such as maximized or fullscreen. Size and position requests made when the window is not in a movable or resizable state will be deferred until it can be applied, so no requests are lost. These changes fix another long-standing issue with renderers recreating maximized windows, where the original non-maximized size was lost, resulting in the window being restored to the wrong size. All automated video tests pass across all platforms. Overall, the "make a request/get an event" model better reflects how most windowing systems work, and some backends avoid spending significant time blocking while waiting for operations to complete.
2023-10-25 22:09:55 +03:00
flags = SDL_GetWindowFlags(window);
SDLTest_AssertPass("SDL_GetWindowFlags()");
SDLTest_AssertCheck(!(flags & SDL_WINDOW_MAXIMIZED), "Verify that the `SDL_WINDOW_MAXIMIZED` flag is cleared: %s", !(flags & SDL_WINDOW_MAXIMIZED) ? "true" : "false");
if (!skipPos) {
currentX = desiredX + 1;
currentY = desiredY + 1;
SDL_GetWindowPosition(window, &currentX, &currentY);
SDLTest_AssertPass("Call to SDL_GetWindowPosition()");
SDLTest_AssertCheck(desiredX == currentX, "Verify returned X coordinate; expected: %d, got: %d", desiredX, currentX);
SDLTest_AssertCheck(desiredY == currentY, "Verify returned Y coordinate; expected: %d, got: %d", desiredY, currentY);
}
currentW = desiredW + 1;
currentH = desiredH + 1;
SDL_GetWindowSize(window, &currentW, &currentH);
SDLTest_AssertPass("Call to SDL_GetWindowSize()");
SDLTest_AssertCheck(desiredW == currentW, "Verify returned width; expected: %d, got: %d", desiredW, currentW);
SDLTest_AssertCheck(desiredH == currentH, "Verify returned height; expected: %d, got: %d", desiredH, currentH);
minimize_test:
/* Minimize */
result = SDL_MinimizeWindow(window);
Use SDL_bool instead an int return code in the SDL API Most SDL functions used to indicate success or failure using an int return code. These functions have been changed to return SDL_bool. Here is a coccinelle patch to change code that previously compared the return value to 0 and changes it to a boolean test: @ bool_return_type @ identifier func =~ "^(SDL_AddEventWatch|SDL_AddHintCallback|SDL_AddSurfaceAlternateImage|SDL_AddVulkanRenderSemaphores|SDL_BindAudioStream|SDL_BindAudioStreams|SDL_BlitSurface|SDL_BlitSurface9Grid|SDL_BlitSurfaceScaled|SDL_BlitSurfaceTiled|SDL_BlitSurfaceTiledWithScale|SDL_BlitSurfaceUnchecked|SDL_BlitSurfaceUncheckedScaled|SDL_CaptureMouse|SDL_ClearAudioStream|SDL_ClearClipboardData|SDL_ClearComposition|SDL_ClearError|SDL_ClearProperty|SDL_ClearSurface|SDL_CloseIO|SDL_CloseStorage|SDL_ConvertAudioSamples|SDL_ConvertEventToRenderCoordinates|SDL_ConvertPixels|SDL_ConvertPixelsAndColorspace|SDL_CopyFile|SDL_CopyProperties|SDL_CopyStorageFile|SDL_CreateDirectory|SDL_CreateStorageDirectory|SDL_CreateWindowAndRenderer|SDL_DateTimeToTime|SDL_DestroyWindowSurface|SDL_DetachVirtualJoystick|SDL_DisableScreenSaver|SDL_EnableScreenSaver|SDL_EnumerateDirectory|SDL_EnumerateProperties|SDL_EnumerateStorageDirectory|SDL_FillSurfaceRect|SDL_FillSurfaceRects|SDL_FlashWindow|SDL_FlipSurface|SDL_FlushAudioStream|SDL_FlushRenderer|SDL_GL_DestroyContext|SDL_GL_GetAttribute|SDL_GL_GetSwapInterval|SDL_GL_LoadLibrary|SDL_GL_MakeCurrent|SDL_GL_SetAttribute|SDL_GL_SetSwapInterval|SDL_GL_SwapWindow|SDL_GetAudioDeviceFormat|SDL_GetAudioStreamFormat|SDL_GetCameraFormat|SDL_GetClosestFullscreenDisplayMode|SDL_GetCurrentRenderOutputSize|SDL_GetCurrentTime|SDL_GetDXGIOutputInfo|SDL_GetDateTimeLocalePreferences|SDL_GetDisplayBounds|SDL_GetDisplayUsableBounds|SDL_GetGDKDefaultUser|SDL_GetGDKTaskQueue|SDL_GetGamepadSensorData|SDL_GetGamepadTouchpadFinger|SDL_GetHapticEffectStatus|SDL_GetJoystickBall|SDL_GetMasksForPixelFormat|SDL_GetPathInfo|SDL_GetRectUnion|SDL_GetRectUnionFloat|SDL_GetRenderClipRect|SDL_GetRenderColorScale|SDL_GetRenderDrawBlendMode|SDL_GetRenderDrawColor|SDL_GetRenderDrawColorFloat|SDL_GetRenderLogicalPresentation|SDL_GetRenderLogicalPresentationRect|SDL_GetRenderOutputSize|SDL_GetRenderSafeArea|SDL_GetRenderScale|SDL_GetRenderVSync|SDL_GetRenderViewport|SDL_GetSensorData|SDL_GetStorageFileSize|SDL_GetStoragePathInfo|SDL_GetSurfaceAlphaMod|SDL_GetSurfaceBlendMode|SDL_GetSurfaceClipRect|SDL_GetSurfaceColorKey|SDL_GetSurfaceColorMod|SDL_GetTextInputArea|SDL_GetTextureAlphaMod|SDL_GetTextureAlphaModFloat|SDL_GetTextureBlendMode|SDL_GetTextureColorMod|SDL_GetTextureColorModFloat|SDL_GetTextureScaleMode|SDL_GetTextureSize|SDL_GetWindowAspectRatio|SDL_GetWindowBordersSize|SDL_GetWindowMaximumSize|SDL_GetWindowMinimumSize|SDL_GetWindowPosition|SDL_GetWindowRelativeMouseMode|SDL_GetWindowSafeArea|SDL_GetWindowSize|SDL_GetWindowSizeInPixels|SDL_GetWindowSurfaceVSync|SDL_HideCursor|SDL_HideWindow|SDL_Init|SDL_InitHapticRumble|SDL_InitSubSystem|SDL_LoadWAV|SDL_LoadWAV_IO|SDL_LockAudioStream|SDL_LockProperties|SDL_LockSurface|SDL_LockTexture|SDL_LockTextureToSurface|SDL_MaximizeWindow|SDL_MinimizeWindow|SDL_MixAudio|SDL_OpenURL|SDL_OutOfMemory|SDL_PauseAudioDevice|SDL_PauseAudioStreamDevice|SDL_PauseHaptic|SDL_PlayHapticRumble|SDL_PremultiplyAlpha|SDL_PremultiplySurfaceAlpha|SDL_PushEvent|SDL_PutAudioStreamData|SDL_RaiseWindow|SDL_ReadStorageFile|SDL_ReadSurfacePixel|SDL_ReadSurfacePixelFloat|SDL_RegisterApp|SDL_ReloadGamepadMappings|SDL_RemovePath|SDL_RemoveStoragePath|SDL_RemoveTimer|SDL_RenamePath|SDL_RenameStoragePath|SDL_RenderClear|SDL_RenderCoordinatesFromWindow|SDL_RenderCoordinatesToWindow|SDL_RenderFillRect|SDL_RenderFillRects|SDL_RenderGeometry|SDL_RenderGeometryRaw|SDL_RenderLine|SDL_RenderLines|SDL_RenderPoint|SDL_RenderPoints|SDL_RenderPresent|SDL_RenderRect|SDL_RenderRects|SDL_RenderTexture|SDL_RenderTexture9Grid|SDL_RenderTextureRotated|SDL_RenderTextureTiled|SDL_RequestAndroidPermission|SDL_RestoreWindow|SDL_ResumeAudioDevice|SDL_ResumeAudioStreamDevice|SDL_ResumeHaptic|SDL_RumbleGamepad|SDL_RumbleGamepadTriggers|SDL_RumbleJoystick|SDL_RumbleJoystickTriggers|SDL_RunHapticEffect|SDL_SaveBMP|SDL_SaveBMP_IO|SDL_SendAndroidMessage|SDL_SendGamepadEffect|SDL_SendJoystickEffect|SDL_SendJoystickVirtualSensorData|SDL_SetAppMetadata|SDL_SetAppMetadataProperty|SDL_SetAudioDeviceGain|SDL_SetAudioPostmixCallback|SDL_SetAudioStreamFormat|SDL_SetAudioStreamFrequencyRatio|SDL_SetAudioStreamGain|SDL_SetAudioStreamGetCallback|SDL_SetAudioStreamInputChannelMap|SDL_SetAudioStreamOutputChannelMap|SDL_SetAudioStreamPutCallback|SDL_SetBooleanProperty|SDL_SetClipboardData|SDL_SetClipboardText|SDL_SetCursor|SDL_SetFloatProperty|SDL_SetGamepadLED|SDL_SetGamepadMapping|SDL_SetGamepadPlayerIndex|SDL_SetGamepadSensorEnabled|SDL_SetHapticAutocenter|SDL_SetHapticGain|SDL_SetJoystickLED|SDL_SetJoystickPlayerIndex|SDL_SetJoystickVirtualAxis|SDL_SetJoystickVirtualBall|SDL_SetJoystickVirtualButton|SDL_SetJoystickVirtualHat|SDL_SetJoystickVirtualTouchpad|SDL_SetLinuxThreadPriority|SDL_SetLinuxThreadPriorityAndPolicy|SDL_SetLogPriorityPrefix|SDL_SetMemoryFunctions|SDL_SetNumberProperty|SDL_SetPaletteColors|SDL_SetPointerProperty|SDL_SetPointerPropertyWithCleanup|SDL_SetPrimarySelectionText|SDL_SetRenderClipRect|SDL_SetRenderColorScale|SDL_SetRenderDrawBlendMode|SDL_SetRenderDrawColor|SDL_SetRenderDrawColorFloat|SDL_SetRenderLogicalPresentation|SDL_SetRenderScale|SDL_SetRenderTarget|SDL_SetRenderVSync|SDL_SetRenderViewport|SDL_SetScancodeName|SDL_SetStringProperty|SDL_SetSurfaceAlphaMod|SDL_SetSurfaceBlendMode|SDL_SetSurfaceColorKey|SDL_SetSurfaceColorMod|SDL_SetSurfaceColorspace|SDL_SetSurfacePalette|SDL_SetSurfaceRLE|SDL_SetTLS|SDL_SetTextInputArea|SDL_SetTextureAlphaMod|SDL_SetTextureAlphaModFloat|SDL_SetTextureBlendMode|SDL_SetTextureColorMod|SDL_SetTextureColorModFloat|SDL_SetTextureScaleMode|SDL_SetThreadPriority|SDL_SetWindowAlwaysOnTop|SDL_SetWindowAspectRatio|SDL_SetWindowBordered|SDL_SetWindowFocusable|SDL_SetWindowFullscreen|SDL_SetWindowFullscreenMode|SDL_SetWindowHitTest|SDL_SetWindowIcon|SDL_SetWindowKeyboardGrab|SDL_SetWindowMaximumSize|SDL_SetWindowMinimumSize|SDL_SetWindowModalFor|SDL_SetWindowMouseGrab|SDL_SetWindowMouseRect|SDL_SetWindowOpacity|SDL_SetWindowPosition|SDL_SetWindowRelativeMouseMode|SDL_SetWindowResizable|SDL_SetWindowShape|SDL_SetWindowSize|SDL_SetWindowSurfaceVSync|SDL_SetWindowTitle|SDL_SetiOSAnimationCallback|SDL_ShowAndroidToast|SDL_ShowCursor|SDL_ShowMessageBox|SDL_ShowSimpleMessageBox|SDL_ShowWindow|SDL_ShowWindowSystemMenu|SDL_StartTextInput|SDL_StartTextInputWithProperties|SDL_StopHapticEffect|SDL_StopHapticEffects|SDL_StopHapticRumble|SDL_StopTextInput|SDL_SyncWindow|SDL_TimeToDateTime|SDL_TryLockMutex|SDL_TryLockRWLockForReading|SDL_TryLockRWLockForWriting|SDL_TryWaitSemaphore|SDL_UnlockAudioStream|SDL_UpdateHapticEffect|SDL_UpdateNVTexture|SDL_UpdateTexture|SDL_UpdateWindowSurface|SDL_UpdateWindowSurfaceRects|SDL_UpdateYUVTexture|SDL_Vulkan_CreateSurface|SDL_Vulkan_LoadLibrary|SDL_WaitConditionTimeout|SDL_WaitSemaphoreTimeout|SDL_WarpMouseGlobal|SDL_WriteStorageFile|SDL_WriteSurfacePixel|SDL_WriteSurfacePixelFloat)$"; @@ ( func( ... ) - == 0 | - func( + !func( ... ) - < 0 | - func( + !func( ... ) - != 0 | - func( + !func( ... ) - == -1 )
2024-08-23 03:33:49 +03:00
if (result) {
video: Implement asynchronous windowing SDL window size, state, and position functions have been considered immediate, with their effects assuming to have taken effect upon successful return of the function. However, several windowing systems handle these requests asynchronously, resulting in the functions blocking until the changes have taken effect, potentially for long periods of time. Additionally, some windowing systems treat these as requests, and can potentially deny or fulfill the request in a manner differently than the application expects, such as not allowing a window to be positioned or sized beyond desktop borders, prohibiting fullscreen, and so on. With these changes, applications can make requests of the window manager that do not block, with the understanding that an associated event will be sent if the request is fulfilled. Currently, size, position, maximize, minimize, and fullscreen calls are handled as asynchronous requests, with events being returned if the request is honored. If the application requires that the change take effect immediately, it can call the new SDL_SyncWindow function, which will attempt to block until the request is fulfilled, or some arbitrary timeout period elapses, the duration of which depends not only on the windowing system, but on the operation requested as well (e.g. a 100ms timeout is fine for most X11 events, but maximizing a window can take considerably longer for some reason). There is also a new hint 'SDL_VIDEO_SYNC_ALL_WINDOW_OPS' that will mimic the old behavior by synchronizing after every window operation with, again, the understanding that using this may result in the associated calls blocking for a relatively long period. The deferred model also results in the window size and position getters not reporting false coordinates anymore, as they only forward what the window manager reports vs allowing applications to set arbitrary values, and fullscreen enter/leave events that were initiated via the window manager update the window state appropriately, where they didn't before. Care was taken to ensure that order of operations is maintained, and that requests are not ignored or dropped. This does require some implicit internal synchronization in the various backends if many requests are made in a short period, as some state and behavior depends on other bits of state that need to be known at that particular point in time, but this isn't something that typical applications will hit, unless they are sending a lot of window state in a short time as the tests do. The automated tests developed to test the previous behavior also resulted in previously undefined behavior being defined and normalized across platforms, particularly when it comes to the sizing and positioning of windows when they are in a fixed-size state, such as maximized or fullscreen. Size and position requests made when the window is not in a movable or resizable state will be deferred until it can be applied, so no requests are lost. These changes fix another long-standing issue with renderers recreating maximized windows, where the original non-maximized size was lost, resulting in the window being restored to the wrong size. All automated video tests pass across all platforms. Overall, the "make a request/get an event" model better reflects how most windowing systems work, and some backends avoid spending significant time blocking while waiting for operations to complete.
2023-10-25 22:09:55 +03:00
SDLTest_AssertPass("SDL_MinimizeWindow()");
Removed SDL_bool in favor of plain bool We require stdbool.h in the build environment, so we might as well use the plain bool type. If your environment doesn't have stdbool.h, this simple replacement will suffice: typedef signed char bool;
2024-09-18 17:52:28 +03:00
SDLTest_AssertCheck(result == true, "Verify return value; expected: true, got: %d", result);
video: Implement asynchronous windowing SDL window size, state, and position functions have been considered immediate, with their effects assuming to have taken effect upon successful return of the function. However, several windowing systems handle these requests asynchronously, resulting in the functions blocking until the changes have taken effect, potentially for long periods of time. Additionally, some windowing systems treat these as requests, and can potentially deny or fulfill the request in a manner differently than the application expects, such as not allowing a window to be positioned or sized beyond desktop borders, prohibiting fullscreen, and so on. With these changes, applications can make requests of the window manager that do not block, with the understanding that an associated event will be sent if the request is fulfilled. Currently, size, position, maximize, minimize, and fullscreen calls are handled as asynchronous requests, with events being returned if the request is honored. If the application requires that the change take effect immediately, it can call the new SDL_SyncWindow function, which will attempt to block until the request is fulfilled, or some arbitrary timeout period elapses, the duration of which depends not only on the windowing system, but on the operation requested as well (e.g. a 100ms timeout is fine for most X11 events, but maximizing a window can take considerably longer for some reason). There is also a new hint 'SDL_VIDEO_SYNC_ALL_WINDOW_OPS' that will mimic the old behavior by synchronizing after every window operation with, again, the understanding that using this may result in the associated calls blocking for a relatively long period. The deferred model also results in the window size and position getters not reporting false coordinates anymore, as they only forward what the window manager reports vs allowing applications to set arbitrary values, and fullscreen enter/leave events that were initiated via the window manager update the window state appropriately, where they didn't before. Care was taken to ensure that order of operations is maintained, and that requests are not ignored or dropped. This does require some implicit internal synchronization in the various backends if many requests are made in a short period, as some state and behavior depends on other bits of state that need to be known at that particular point in time, but this isn't something that typical applications will hit, unless they are sending a lot of window state in a short time as the tests do. The automated tests developed to test the previous behavior also resulted in previously undefined behavior being defined and normalized across platforms, particularly when it comes to the sizing and positioning of windows when they are in a fixed-size state, such as maximized or fullscreen. Size and position requests made when the window is not in a movable or resizable state will be deferred until it can be applied, so no requests are lost. These changes fix another long-standing issue with renderers recreating maximized windows, where the original non-maximized size was lost, resulting in the window being restored to the wrong size. All automated video tests pass across all platforms. Overall, the "make a request/get an event" model better reflects how most windowing systems work, and some backends avoid spending significant time blocking while waiting for operations to complete.
2023-10-25 22:09:55 +03:00
result = SDL_SyncWindow(window);
SDLTest_AssertPass("SDL_SyncWindow()");
Removed SDL_bool in favor of plain bool We require stdbool.h in the build environment, so we might as well use the plain bool type. If your environment doesn't have stdbool.h, this simple replacement will suffice: typedef signed char bool;
2024-09-18 17:52:28 +03:00
SDLTest_AssertCheck(result == true, "Verify return value; expected: true, got: %d", result);
video: Implement asynchronous windowing SDL window size, state, and position functions have been considered immediate, with their effects assuming to have taken effect upon successful return of the function. However, several windowing systems handle these requests asynchronously, resulting in the functions blocking until the changes have taken effect, potentially for long periods of time. Additionally, some windowing systems treat these as requests, and can potentially deny or fulfill the request in a manner differently than the application expects, such as not allowing a window to be positioned or sized beyond desktop borders, prohibiting fullscreen, and so on. With these changes, applications can make requests of the window manager that do not block, with the understanding that an associated event will be sent if the request is fulfilled. Currently, size, position, maximize, minimize, and fullscreen calls are handled as asynchronous requests, with events being returned if the request is honored. If the application requires that the change take effect immediately, it can call the new SDL_SyncWindow function, which will attempt to block until the request is fulfilled, or some arbitrary timeout period elapses, the duration of which depends not only on the windowing system, but on the operation requested as well (e.g. a 100ms timeout is fine for most X11 events, but maximizing a window can take considerably longer for some reason). There is also a new hint 'SDL_VIDEO_SYNC_ALL_WINDOW_OPS' that will mimic the old behavior by synchronizing after every window operation with, again, the understanding that using this may result in the associated calls blocking for a relatively long period. The deferred model also results in the window size and position getters not reporting false coordinates anymore, as they only forward what the window manager reports vs allowing applications to set arbitrary values, and fullscreen enter/leave events that were initiated via the window manager update the window state appropriately, where they didn't before. Care was taken to ensure that order of operations is maintained, and that requests are not ignored or dropped. This does require some implicit internal synchronization in the various backends if many requests are made in a short period, as some state and behavior depends on other bits of state that need to be known at that particular point in time, but this isn't something that typical applications will hit, unless they are sending a lot of window state in a short time as the tests do. The automated tests developed to test the previous behavior also resulted in previously undefined behavior being defined and normalized across platforms, particularly when it comes to the sizing and positioning of windows when they are in a fixed-size state, such as maximized or fullscreen. Size and position requests made when the window is not in a movable or resizable state will be deferred until it can be applied, so no requests are lost. These changes fix another long-standing issue with renderers recreating maximized windows, where the original non-maximized size was lost, resulting in the window being restored to the wrong size. All automated video tests pass across all platforms. Overall, the "make a request/get an event" model better reflects how most windowing systems work, and some backends avoid spending significant time blocking while waiting for operations to complete.
2023-10-25 22:09:55 +03:00
flags = SDL_GetWindowFlags(window);
SDLTest_AssertPass("SDL_GetWindowFlags()");
SDLTest_AssertCheck(flags & SDL_WINDOW_MINIMIZED, "Verify that the `SDL_WINDOW_MINIMIZED` flag is set: %s", (flags & SDL_WINDOW_MINIMIZED) ? "true" : "false");
} else {
SDLTest_Log("Skipping minimize test: %s reports window minimizing as unsupported", SDL_GetCurrentVideoDriver());
skipFlags |= SDL_WINDOW_MINIMIZED;
}
/* Clean up */
destroyVideoSuiteTestWindow(window);
/* Restore the hint to the previous value */
SDL_SetHint("SDL_BORDERLESS_RESIZABLE_STYLE", restoreHint ? "1" : "0");
return skipFlags != (SDL_WINDOW_MAXIMIZED | SDL_WINDOW_MINIMIZED) ? TEST_COMPLETED : TEST_SKIPPED;
}
SDL_test: use SDLCALL calling convention This is needed when using a pre-built static SDL3_test library.
2024-09-06 04:21:13 +03:00
static int SDLCALL video_createMinimized(void *arg)
Added an automated test to check creation of minimized and maximized windows
2024-01-27 09:53:07 +03:00
{
const char *title = "video_createMinimized Test Window";
int result;
SDL_Window *window;
int windowedX, windowedY;
int windowedW, windowedH;
/* Call against new test window */
window = SDL_CreateWindow(title, 320, 200, SDL_WINDOW_MINIMIZED);
if (!window) {
return TEST_ABORTED;
}
SDL_GetWindowSize(window, &windowedW, &windowedH);
SDLTest_AssertPass("SDL_GetWindowSize()");
SDLTest_AssertCheck(windowedW > 0 && windowedH > 0, "Verify return value; expected: 320x200, got: %dx%d", windowedW, windowedH);
SDL_GetWindowSizeInPixels(window, &windowedW, &windowedH);
SDLTest_AssertPass("SDL_GetWindowSizeInPixels()");
SDLTest_AssertCheck(windowedW > 0 && windowedH > 0, "Verify return value; expected: > 0, got: %dx%d", windowedW, windowedH);
SDL_GetWindowPosition(window, &windowedX, &windowedY);
SDLTest_AssertPass("SDL_GetWindowPosition()");
SDLTest_AssertCheck(windowedX >= 0 && windowedY >= 0, "Verify return value; expected: >= 0, got: %d,%d", windowedX, windowedY);
if (SDL_GetWindowFlags(window) & SDL_WINDOW_MINIMIZED) {
result = SDL_RestoreWindow(window);
SDLTest_AssertPass("SDL_RestoreWindow()");
Removed SDL_bool in favor of plain bool We require stdbool.h in the build environment, so we might as well use the plain bool type. If your environment doesn't have stdbool.h, this simple replacement will suffice: typedef signed char bool;
2024-09-18 17:52:28 +03:00
SDLTest_AssertCheck(result == true, "Verify return value; expected: true, got: %d", result);
Added an automated test to check creation of minimized and maximized windows
2024-01-27 09:53:07 +03:00
} else {
SDLTest_Log("Requested minimized window on creation, but that isn't supported on this platform.");
}
SDL_DestroyWindow(window);
return TEST_COMPLETED;
}
SDL_test: use SDLCALL calling convention This is needed when using a pre-built static SDL3_test library.
2024-09-06 04:21:13 +03:00
static int SDLCALL video_createMaximized(void *arg)
Added an automated test to check creation of minimized and maximized windows
2024-01-27 09:53:07 +03:00
{
const char *title = "video_createMaximized Test Window";
int result;
SDL_Window *window;
int windowedX, windowedY;
int windowedW, windowedH;
/* Call against new test window */
window = SDL_CreateWindow(title, 320, 200, SDL_WINDOW_MAXIMIZED);
if (!window) {
return TEST_ABORTED;
}
SDL_GetWindowSize(window, &windowedW, &windowedH);
SDLTest_AssertPass("SDL_GetWindowSize()");
SDLTest_AssertCheck(windowedW > 0 && windowedH > 0, "Verify return value; expected: 320x200, got: %dx%d", windowedW, windowedH);
SDL_GetWindowSizeInPixels(window, &windowedW, &windowedH);
SDLTest_AssertPass("SDL_GetWindowSizeInPixels()");
SDLTest_AssertCheck(windowedW > 0 && windowedH > 0, "Verify return value; expected: > 0, got: %dx%d", windowedW, windowedH);
SDL_GetWindowPosition(window, &windowedX, &windowedY);
SDLTest_AssertPass("SDL_GetWindowPosition()");
SDLTest_AssertCheck(windowedX >= 0 && windowedY >= 0, "Verify return value; expected: >= 0, got: %d,%d", windowedX, windowedY);
if (SDL_GetWindowFlags(window) & SDL_WINDOW_MAXIMIZED) {
result = SDL_RestoreWindow(window);
SDLTest_AssertPass("SDL_RestoreWindow()");
Removed SDL_bool in favor of plain bool We require stdbool.h in the build environment, so we might as well use the plain bool type. If your environment doesn't have stdbool.h, this simple replacement will suffice: typedef signed char bool;
2024-09-18 17:52:28 +03:00
SDLTest_AssertCheck(result == true, "Verify return value; expected: true, got: %d", result);
Added an automated test to check creation of minimized and maximized windows
2024-01-27 09:53:07 +03:00
} else {
SDLTest_Log("Requested maximized window on creation, but that isn't supported on this platform.");
}
SDL_DestroyWindow(window);
return TEST_COMPLETED;
}
Clean up any renderer in SDL_DestroyWindowSurface() Also added an automated test to verify window surface functionality. Fixes https://github.com/libsdl-org/SDL/issues/10133
2024-06-29 12:34:17 +03:00
/**
* Tests window surface functionality
*/
SDL_test: use SDLCALL calling convention This is needed when using a pre-built static SDL3_test library.
2024-09-06 04:21:13 +03:00
static int SDLCALL video_getWindowSurface(void *arg)
Clean up any renderer in SDL_DestroyWindowSurface() Also added an automated test to verify window surface functionality. Fixes https://github.com/libsdl-org/SDL/issues/10133
2024-06-29 12:34:17 +03:00
{
const char *title = "video_getWindowSurface Test Window";
SDL_Window *window;
SDL_Surface *surface;
SDL_Renderer *renderer;
const char *renderer_name = NULL;
int result;
if (SDL_strcmp(SDL_GetCurrentVideoDriver(), "dummy") == 0) {
renderer_name = SDL_SOFTWARE_RENDERER;
}
/* Make sure we're testing interaction with an accelerated renderer */
SDL_SetHint(SDL_HINT_FRAMEBUFFER_ACCELERATION, "1");
window = SDL_CreateWindow(title, 320, 320, 0);
SDLTest_AssertPass("Call to SDL_CreateWindow('Title',320,320,0)");
SDLTest_AssertCheck(window != NULL, "Validate that returned window is not NULL");
surface = SDL_GetWindowSurface(window);
SDLTest_AssertPass("Call to SDL_GetWindowSurface(window)");
SDLTest_AssertCheck(surface != NULL, "Validate that returned surface is not NULL");
SDLTest_AssertCheck(SDL_WindowHasSurface(window), "Validate that window has a surface");
result = SDL_UpdateWindowSurface(window);
SDLTest_AssertPass("Call to SDL_UpdateWindowSurface(window)");
Removed SDL_bool in favor of plain bool We require stdbool.h in the build environment, so we might as well use the plain bool type. If your environment doesn't have stdbool.h, this simple replacement will suffice: typedef signed char bool;
2024-09-18 17:52:28 +03:00
SDLTest_AssertCheck(result == true, "Verify return value; expected: true, got: %d", result);
Clean up any renderer in SDL_DestroyWindowSurface() Also added an automated test to verify window surface functionality. Fixes https://github.com/libsdl-org/SDL/issues/10133
2024-06-29 12:34:17 +03:00
/* We shouldn't be able to create a renderer on a window with a surface */
renderer = SDL_CreateRenderer(window, renderer_name);
SDLTest_AssertPass("Call to SDL_CreateRenderer(window, %s)", renderer_name);
SDLTest_AssertCheck(renderer == NULL, "Validate that returned renderer is NULL");
result = SDL_DestroyWindowSurface(window);
SDLTest_AssertPass("Call to SDL_DestroyWindowSurface(window)");
Removed SDL_bool in favor of plain bool We require stdbool.h in the build environment, so we might as well use the plain bool type. If your environment doesn't have stdbool.h, this simple replacement will suffice: typedef signed char bool;
2024-09-18 17:52:28 +03:00
SDLTest_AssertCheck(result == true, "Verify return value; expected: true, got: %d", result);
Clean up any renderer in SDL_DestroyWindowSurface() Also added an automated test to verify window surface functionality. Fixes https://github.com/libsdl-org/SDL/issues/10133
2024-06-29 12:34:17 +03:00
SDLTest_AssertCheck(!SDL_WindowHasSurface(window), "Validate that window does not have a surface");
/* We should be able to create a renderer on the window now */
renderer = SDL_CreateRenderer(window, renderer_name);
SDLTest_AssertPass("Call to SDL_CreateRenderer(window, %s)", renderer_name);
SDLTest_AssertCheck(renderer != NULL, "Validate that returned renderer is not NULL");
/* We should not be able to create a window surface now, unless it was created by the renderer */
if (!SDL_WindowHasSurface(window)) {
surface = SDL_GetWindowSurface(window);
SDLTest_AssertPass("Call to SDL_GetWindowSurface(window)");
SDLTest_AssertCheck(surface == NULL, "Validate that returned surface is NULL");
}
SDL_DestroyRenderer(renderer);
SDLTest_AssertPass("Call to SDL_DestroyRenderer(renderer)");
SDLTest_AssertCheck(!SDL_WindowHasSurface(window), "Validate that window does not have a surface");
/* We should be able to create a window surface again */
surface = SDL_GetWindowSurface(window);
SDLTest_AssertPass("Call to SDL_GetWindowSurface(window)");
SDLTest_AssertCheck(surface != NULL, "Validate that returned surface is not NULL");
SDLTest_AssertCheck(SDL_WindowHasSurface(window), "Validate that window has a surface");
/* Clean up */
SDL_DestroyWindow(window);
return TEST_COMPLETED;
}
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
/* ================= Test References ================== */
/* Video test cases */
Added an automated test to check creation of minimized and maximized windows
2024-01-27 09:53:07 +03:00
static const SDLTest_TestCaseReference videoTestEnableDisableScreensaver = {
testautomation: skip window resize and repositioning tests on Emscripten
2024-08-18 00:20:41 +03:00
video_enableDisableScreensaver, "video_enableDisableScreensaver", "Enable and disable screenaver while checking state", TEST_ENABLED
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
};
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
Added an automated test to check creation of minimized and maximized windows
2024-01-27 09:53:07 +03:00
static const SDLTest_TestCaseReference videoTestCreateWindowVariousSizes = {
testautomation: skip window resize and repositioning tests on Emscripten
2024-08-18 00:20:41 +03:00
video_createWindowVariousSizes, "video_createWindowVariousSizes", "Create windows with various sizes", TEST_ENABLED
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
};
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
Added an automated test to check creation of minimized and maximized windows
2024-01-27 09:53:07 +03:00
static const SDLTest_TestCaseReference videoTestCreateWindowVariousFlags = {
testautomation: skip window resize and repositioning tests on Emscripten
2024-08-18 00:20:41 +03:00
video_createWindowVariousFlags, "video_createWindowVariousFlags", "Create windows using various flags", TEST_ENABLED
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
};
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
Added an automated test to check creation of minimized and maximized windows
2024-01-27 09:53:07 +03:00
static const SDLTest_TestCaseReference videoTestGetWindowFlags = {
testautomation: skip window resize and repositioning tests on Emscripten
2024-08-18 00:20:41 +03:00
video_getWindowFlags, "video_getWindowFlags", "Get window flags set during SDL_CreateWindow", TEST_ENABLED
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
};
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
Added an automated test to check creation of minimized and maximized windows
2024-01-27 09:53:07 +03:00
static const SDLTest_TestCaseReference videoTestGetFullscreenDisplayModes = {
testautomation: skip window resize and repositioning tests on Emscripten
2024-08-18 00:20:41 +03:00
video_getFullscreenDisplayModes, "video_getFullscreenDisplayModes", "Use SDL_GetFullscreenDisplayModes function to get number of display modes", TEST_ENABLED
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
};
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
Added an automated test to check creation of minimized and maximized windows
2024-01-27 09:53:07 +03:00
static const SDLTest_TestCaseReference videoTestGetClosestDisplayModeCurrentResolution = {
testautomation: skip window resize and repositioning tests on Emscripten
2024-08-18 00:20:41 +03:00
video_getClosestDisplayModeCurrentResolution, "video_getClosestDisplayModeCurrentResolution", "Use function to get closes match to requested display mode for current resolution", TEST_ENABLED
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
};
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
Added an automated test to check creation of minimized and maximized windows
2024-01-27 09:53:07 +03:00
static const SDLTest_TestCaseReference videoTestGetClosestDisplayModeRandomResolution = {
testautomation: skip window resize and repositioning tests on Emscripten
2024-08-18 00:20:41 +03:00
video_getClosestDisplayModeRandomResolution, "video_getClosestDisplayModeRandomResolution", "Use function to get closes match to requested display mode for random resolution", TEST_ENABLED
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
};
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
Added an automated test to check creation of minimized and maximized windows
2024-01-27 09:53:07 +03:00
static const SDLTest_TestCaseReference videoTestGetWindowDisplayMode = {
testautomation: skip window resize and repositioning tests on Emscripten
2024-08-18 00:20:41 +03:00
video_getWindowDisplayMode, "video_getWindowDisplayMode", "Get window display mode", TEST_ENABLED
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
};
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
Added an automated test to check creation of minimized and maximized windows
2024-01-27 09:53:07 +03:00
static const SDLTest_TestCaseReference videoTestGetWindowDisplayModeNegative = {
testautomation: skip window resize and repositioning tests on Emscripten
2024-08-18 00:20:41 +03:00
video_getWindowDisplayModeNegative, "video_getWindowDisplayModeNegative", "Get window display mode with invalid input", TEST_ENABLED
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
};
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
Added an automated test to check creation of minimized and maximized windows
2024-01-27 09:53:07 +03:00
static const SDLTest_TestCaseReference videoTestGetSetWindowGrab = {
testautomation: skip window resize and repositioning tests on Emscripten
2024-08-18 00:20:41 +03:00
video_getSetWindowGrab, "video_getSetWindowGrab", "Checks input grab positive and negative cases", TEST_ENABLED
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
};
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
Added an automated test to check creation of minimized and maximized windows
2024-01-27 09:53:07 +03:00
static const SDLTest_TestCaseReference videoTestGetWindowID = {
testautomation: skip window resize and repositioning tests on Emscripten
2024-08-18 00:20:41 +03:00
video_getWindowId, "video_getWindowId", "Checks SDL_GetWindowID and SDL_GetWindowFromID", TEST_ENABLED
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
};
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
Added an automated test to check creation of minimized and maximized windows
2024-01-27 09:53:07 +03:00
static const SDLTest_TestCaseReference videoTestGetWindowPixelFormat = {
testautomation: skip window resize and repositioning tests on Emscripten
2024-08-18 00:20:41 +03:00
video_getWindowPixelFormat, "video_getWindowPixelFormat", "Checks SDL_GetWindowPixelFormat", TEST_ENABLED
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
};
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
Added an automated test to check creation of minimized and maximized windows
2024-01-27 09:53:07 +03:00
static const SDLTest_TestCaseReference videoTestGetSetWindowPosition = {
testautomation: skip window resize and repositioning tests on Emscripten
2024-08-18 00:20:41 +03:00
video_getSetWindowPosition, "video_getSetWindowPosition", "Checks SDL_GetWindowPosition and SDL_SetWindowPosition positive and negative cases", TEST_ENABLED
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
};
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
Added an automated test to check creation of minimized and maximized windows
2024-01-27 09:53:07 +03:00
static const SDLTest_TestCaseReference videoTestGetSetWindowSize = {
testautomation: skip window resize and repositioning tests on Emscripten
2024-08-18 00:20:41 +03:00
video_getSetWindowSize, "video_getSetWindowSize", "Checks SDL_GetWindowSize and SDL_SetWindowSize positive and negative cases", TEST_ENABLED
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
};
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
Added an automated test to check creation of minimized and maximized windows
2024-01-27 09:53:07 +03:00
static const SDLTest_TestCaseReference videoTestGetSetWindowMinimumSize = {
testautomation: skip window resize and repositioning tests on Emscripten
2024-08-18 00:20:41 +03:00
video_getSetWindowMinimumSize, "video_getSetWindowMinimumSize", "Checks SDL_GetWindowMinimumSize and SDL_SetWindowMinimumSize positive and negative cases", TEST_ENABLED
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
};
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
Added an automated test to check creation of minimized and maximized windows
2024-01-27 09:53:07 +03:00
static const SDLTest_TestCaseReference videoTestGetSetWindowMaximumSize = {
testautomation: skip window resize and repositioning tests on Emscripten
2024-08-18 00:20:41 +03:00
video_getSetWindowMaximumSize, "video_getSetWindowMaximumSize", "Checks SDL_GetWindowMaximumSize and SDL_SetWindowMaximumSize positive and negative cases", TEST_ENABLED
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
};
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
Added an automated test to check creation of minimized and maximized windows
2024-01-27 09:53:07 +03:00
static const SDLTest_TestCaseReference videoTestGetSetWindowData = {
testautomation: skip window resize and repositioning tests on Emscripten
2024-08-18 00:20:41 +03:00
video_getSetWindowData, "video_getSetWindowData", "Checks SDL_SetWindowData and SDL_GetWindowData positive and negative cases", TEST_ENABLED
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
};
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
Added an automated test to check creation of minimized and maximized windows
2024-01-27 09:53:07 +03:00
static const SDLTest_TestCaseReference videoTestSetWindowCenteredOnDisplay = {
testautomation: skip window resize and repositioning tests on Emscripten
2024-08-18 00:20:41 +03:00
video_setWindowCenteredOnDisplay, "video_setWindowCenteredOnDisplay", "Checks using SDL_WINDOWPOS_CENTERED_DISPLAY centers the window on a display", TEST_ENABLED
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
};
testautomation: add video_setWindowCenteredOnDisplay test
2022-06-05 23:44:30 +03:00
Added an automated test to check creation of minimized and maximized windows
2024-01-27 09:53:07 +03:00
static const SDLTest_TestCaseReference videoTestGetSetWindowState = {
testautomation: skip window resize and repositioning tests on Emscripten
2024-08-18 00:20:41 +03:00
video_getSetWindowState, "video_getSetWindowState", "Checks transitioning between windowed, minimized, maximized, and fullscreen states", TEST_ENABLED
video: Implement asynchronous windowing SDL window size, state, and position functions have been considered immediate, with their effects assuming to have taken effect upon successful return of the function. However, several windowing systems handle these requests asynchronously, resulting in the functions blocking until the changes have taken effect, potentially for long periods of time. Additionally, some windowing systems treat these as requests, and can potentially deny or fulfill the request in a manner differently than the application expects, such as not allowing a window to be positioned or sized beyond desktop borders, prohibiting fullscreen, and so on. With these changes, applications can make requests of the window manager that do not block, with the understanding that an associated event will be sent if the request is fulfilled. Currently, size, position, maximize, minimize, and fullscreen calls are handled as asynchronous requests, with events being returned if the request is honored. If the application requires that the change take effect immediately, it can call the new SDL_SyncWindow function, which will attempt to block until the request is fulfilled, or some arbitrary timeout period elapses, the duration of which depends not only on the windowing system, but on the operation requested as well (e.g. a 100ms timeout is fine for most X11 events, but maximizing a window can take considerably longer for some reason). There is also a new hint 'SDL_VIDEO_SYNC_ALL_WINDOW_OPS' that will mimic the old behavior by synchronizing after every window operation with, again, the understanding that using this may result in the associated calls blocking for a relatively long period. The deferred model also results in the window size and position getters not reporting false coordinates anymore, as they only forward what the window manager reports vs allowing applications to set arbitrary values, and fullscreen enter/leave events that were initiated via the window manager update the window state appropriately, where they didn't before. Care was taken to ensure that order of operations is maintained, and that requests are not ignored or dropped. This does require some implicit internal synchronization in the various backends if many requests are made in a short period, as some state and behavior depends on other bits of state that need to be known at that particular point in time, but this isn't something that typical applications will hit, unless they are sending a lot of window state in a short time as the tests do. The automated tests developed to test the previous behavior also resulted in previously undefined behavior being defined and normalized across platforms, particularly when it comes to the sizing and positioning of windows when they are in a fixed-size state, such as maximized or fullscreen. Size and position requests made when the window is not in a movable or resizable state will be deferred until it can be applied, so no requests are lost. These changes fix another long-standing issue with renderers recreating maximized windows, where the original non-maximized size was lost, resulting in the window being restored to the wrong size. All automated video tests pass across all platforms. Overall, the "make a request/get an event" model better reflects how most windowing systems work, and some backends avoid spending significant time blocking while waiting for operations to complete.
2023-10-25 22:09:55 +03:00
};
Added an automated test to check creation of minimized and maximized windows
2024-01-27 09:53:07 +03:00
static const SDLTest_TestCaseReference videoTestCreateMinimized = {
testautomation: skip window resize and repositioning tests on Emscripten
2024-08-18 00:20:41 +03:00
video_createMinimized, "video_createMinimized", "Checks window state for windows created minimized", TEST_ENABLED
Added an automated test to check creation of minimized and maximized windows
2024-01-27 09:53:07 +03:00
};
static const SDLTest_TestCaseReference videoTestCreateMaximized = {
testautomation: skip window resize and repositioning tests on Emscripten
2024-08-18 00:20:41 +03:00
video_createMaximized, "video_createMaximized", "Checks window state for windows created maximized", TEST_ENABLED
Added an automated test to check creation of minimized and maximized windows
2024-01-27 09:53:07 +03:00
};
static const SDLTest_TestCaseReference videoTestGetWindowSurface = {
testautomation: skip window resize and repositioning tests on Emscripten
2024-08-18 00:20:41 +03:00
video_getWindowSurface, "video_getWindowSurface", "Checks window surface functionality", TEST_ENABLED
Clean up any renderer in SDL_DestroyWindowSurface() Also added an automated test to verify window surface functionality. Fixes https://github.com/libsdl-org/SDL/issues/10133
2024-06-29 12:34:17 +03:00
};
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
/* Sequence of Video test cases */
Update for SDL3 coding style (#6717) I updated .clang-format and ran clang-format 14 over the src and test directories to standardize the code base. In general I let clang-format have it's way, and added markup to prevent formatting of code that would break or be completely unreadable if formatted. The script I ran for the src directory is added as build-scripts/clang-format-src.sh This fixes: #6592 #6593 #6594
2022-11-30 23:51:59 +03:00
static const SDLTest_TestCaseReference *videoTests[] = {
Added an automated test to check creation of minimized and maximized windows
2024-01-27 09:53:07 +03:00
&videoTestEnableDisableScreensaver,
&videoTestCreateWindowVariousSizes,
&videoTestCreateWindowVariousFlags,
&videoTestGetWindowFlags,
&videoTestGetFullscreenDisplayModes,
&videoTestGetClosestDisplayModeCurrentResolution,
&videoTestGetClosestDisplayModeRandomResolution,
&videoTestGetWindowDisplayMode,
&videoTestGetWindowDisplayModeNegative,
&videoTestGetSetWindowGrab,
&videoTestGetWindowID,
&videoTestGetWindowPixelFormat,
&videoTestGetSetWindowPosition,
&videoTestGetSetWindowSize,
&videoTestGetSetWindowMinimumSize,
&videoTestGetSetWindowMaximumSize,
&videoTestGetSetWindowData,
&videoTestSetWindowCenteredOnDisplay,
&videoTestGetSetWindowState,
&videoTestCreateMinimized,
&videoTestCreateMaximized,
&videoTestGetWindowSurface,
NULL
Fixed crash if initialization of EGL failed but was tried again later. The internal function SDL_EGL_LoadLibrary() did not delete and remove a mostly uninitialized data structure if loading the library first failed. A later try to use EGL then skipped initialization and assumed it was previously successful because the data structure now already existed. This led to at least one crash in the internal function SDL_EGL_ChooseConfig() because a NULL pointer was dereferenced to make a call to eglBindAPI().
2015-06-21 18:33:46 +03:00
};
/* Video test suite (global) */
SDLTest_TestSuiteReference videoTestSuite = {
"Video",
NULL,
videoTests,
NULL
};