Doxygen uses a <div> with center alignment for images, hence
adding HTML center alignment for image is not only superfluous
but was also inconsistent.
Also removed some more unnecessary HTML tags, updated copyright year.
The code failed with 2 or more widgets when FLTK would trigger
partial redraws of the widgets, unless the Fl_Gl_Window was FL_DAMAGE_ALL.
This commit also adds a second FLTK widget to example/OpenGL3test to
check this multiple widget situation, and fixes an error where operator ! was used
instead of ~ .
This commit fixes " "drawing an fl_pie() with dimension 1x1".
But it does not change "drawing fl_line() with begin position = end position does not show"
because the same behaviour is observed with Windows and macOS.
This commit allows to switch between FL_DOUBLE / FL_SINGLE modes
in widget-containing GL3 windows.
Demo program examples/OpenGL3test is modified to show FLTK widgets
even if the platform does not support OpenGL 3.
Under non-macOS platforms, the key is to call glUseProgram(0); after having used OpenGL 3
which allows to then use OpenGL 1 and draw FLTK widgets over the OpenGL3 scene.
Under macOS, this is impossible because macOS GL3 contexts are not compatible
with GL1. The solution implemented here is to create an additional Fl_Gl_Window
placed above and sized as the GL3-based window, to give it a non opaque,
GL1-based context, and to put the FLTK widgets in that additional window.
Function fl_read_image() obliges to keep a minimal use of global variable fl_window
also under the Wayland platform, even if its type (Window) makes little sense for
the hybrid library because it has its X11 value (given by X11/X.h) which is not
meaningful for the Wayland leg of the hybrid platform.
Virtual member function Fl_Surface_Device::as_image_surface() becomes useless.
The special use context of fl_read_image() inside the draw() function of an
Fl_Double_Window, where the function should read inside the window's double buffer
rather than the on-screen window, no longer worked.
Removal of all uses of the fl_window global variable from platform-independent
code remains necessary, because this variable is ill-defined in the Wayland/X11
library.
1) add Wayland code that prevent subwindows from leaking outside their parent.
This does not cover GL subwindows.
2) add macOS code that prevent GL subwindows from leaking outside their parent.
This fixes issue #494 for the macOS platform.
N.B.: Wayland GL subwindows are not prevented from leaking because no solution
that would not require any change in client applications was found. Code that
would cover Wayland GL subwindows but would require client applications to always
use the FL_ALPHA flag is included in this commit in commented out form.
The old code would allocate memory on FL_KEYDOWN and FL_KEYUP events
and never free the latter.
The fix is to allocate memory only on FL_KEYDOWN events which will
be free'd later as intended.
Also fix two unrelated whitespace errors.
warning: ‘int xrender_supported()’ defined but not used
This happens only with certain configure options, particularly if
FLTK_USE_CAIRO is set, i.e. drawing with Cairo is enabled.
The problem is that the app sometimes stops after closing a file dialog window.
It results in fact from a bug in function draw_title_bar of libdecor-gtk.c that erroneously
calls libdecor_frame_set_max_content_size(). This bug is being proposed to be fixed
by a MR to libdecor.
But moving the libdecor_frame_set_min/max_content_size() calls after the capability changes
hides the bug, and the problem with file dialog doesn't show.
Conflicting demands arise in the implementation of class Fl_Xlib_Graphics_Driver
for drawing images with the XRender library :
1) Issue #163 leads to use a bilinear filter to draw-and-scale images.
2) This tends to blur the edges of drawn areas which is bad for tiled images
(that is because the edges get alpha values, even for an opaque source image).
This commit resolves the conflict adding a means to detect whether the library
is busy drawing a tiled image. If so, the bilinear filter is not applied, drawn areas
don't have blurred edges, resulting in a nice tiling.
With this commit, these test apps perform correctly:
- tiled_image is correct at all scaling factor values also when modified
to use a depth-3 or a depth-4 Fl_RGB_Image as tile;
- unittests - Drawing Images is correct at all scaling factor values;
- pixmap_browser scales correctly up and down JPEG and PNG images.
Modifying test/tiled_image to make it use a depth-3 image for
tiling (rather than an Fl_Pixmap) shows the filter does require
to use PictOpOver as the render op in the XRenderComposite call.
Support different margin sizes on all four edges. Default margin
and gap size is now 0 (compatible with Fl_Pack).
Doxygen: move the description from the constructor to the class
declaration which constitutes a "description".
Make some methods virtual and/or 'const'.
Clarify demo programs, make them even more "FLTK style".
This work is based on the repository and latest commit:
https://github.com/osen/FL_Flex.git
commit 36e4ed75a00daac825b87e81295818b4650991f5
Author: Karsten Pedersen <...>
Date: Fri Apr 23 12:06:16 2021 +0000
Added Fltk (LGPL) license.
This widget is similar to Fl_Pack and supports either one row or one
column of widgets but has some more features. Test and demo programs
are included:
test/flex_login.cxx: simple "login window" demo program
test/flex_demo.cxx: slightly more complex demo program
The original demo programs can still be compiled and built with
the new widget provided you '#include <FL/Fl_Flex.H>'.
Backwards compatible methods are included (except debug()).
The original widget has been modified to match FLTK standards and
enhanced in several ways, including:
- support box frames
- add HORIZONTAL and VERTICAL enum values (as in Fl_Pack)
- add horizontal() method (as in Fl_Pack)
- use type() rather than internal 'direction' variable
- add standard widget constructor (x, y, w, h, label)
- add margin and gap accessors rather than hard coding constants
- improve test and demo programs
- add documentation
- replace <vector> with array as required by FLTK CMP
- rename camelCase method names, keeping old names for compatibility:
- change 'setSize(Fl_Widget*, int)' to 'set_size(Fl_Widget*, int)'
- change 'bool isSetSize(Fl_Widget*)' to 'int set_size(Fl_Widget*)'
- remove debug() method
- add a way to "unset" fixed size: set_size(Fl_Widget *, 0)
- add layout() method to force recalculation of children
- unify resizeRow() and resizeCol() methods to avoid code duplication
- improve widget size calculation.
Please note that the parameter 'percent' of XBell(3).
Given the value of 100, the system defined setting
(normally specified via xset) is ignored and the
percent of 100 is used instead. When calling the
bell from fltk with FL_BEEP_DEFAULT I would expect
to get the default (system specified) percent setting.
The implemented approach is to create and use the pango_layout_ object only relatively
to an unscaled cairo context. With this, the pixel width of a drawn string equals
the sum of the widths of its characters.
This improves reading large selections and fixes one more bug.
- Fix reading selection data size of clients that don't send the size.
These clients don't respect the ICCCM. Using a default size instead.
- Improve memory allocation by pre-allocating at least 4 MB and
extending by 4 MB to reduce the number of realloc() operations.
This may waste up to 4 MB buffer space but this is only an issue
if clients don't send an appropriate size (see above).
- Limit the initial allocation to 200 MB in case clients send higher
values. This is very unlikely and might be a bug in these clients.
Default extension as above would apply anyway.
- Add more comments and optional debug statements (commented out).
Note: reading selections (clipboard) from other clients using the
INCR protocol is implemented but the opposite (providing large
selections (clipboard) by using the INCR protocol is not.
- Fix reading the size (aka "lower bound") of selection data.
- Use Fl::fatal() to terminate the process if memory for the selection
(aka clipboard) data can't be allocated. This should rarely
happen but if it does this is at least a "clean" exit and does
not overwrite arbitrary data waiting for later errors that are
hard to debug (as the old code would have done).
Todo: find a better solution because this can be caused by
another faulty process (the "selection owner"). It would be good
if we could ignore the transfer rather than killing the process.
- Continue processing the INCR protocol if another "unexpected" event
is received. Such events can definitely happen but the current code
can't deal with this because other events might cause recursions.
Hence such events are currently ignored.
Example: pressing and holding ctrl/v would trigger another clipboard
transfer while we're still processing one.
Todo: maybe process "other" events correctly while processing the INCR
protocol. The current processing is done inside a function and would
need to call fl_handle() with potential recursions, hence this would
likely need major refactoring.
This issue revealed that the macOS and Wayland platform processed events slightly differently
from other platforms.
Under X11 and Windows, Fl::wait() processes one event if one is present, and also all other events
present in the event queue immediately after the processing of that event.
Under macOS and Wayland, Fl::wait() processes one event if one is present and returns.
This commits makes the macOS and Wayland platforms behave as other platforms.
Fix Fl_Tiled_Image made from scaled source image.
Fix Fl_Shared_Image::update() to allow scaled source image.
Correct handling of default value (-1) of 3rd argument of 3-argument Fl_BMP_Image constructor.
This fixes some issues of the repeat timeout of Fl_Counter. It could
happen that the widget missed the FL_RELEASE event, for instance if
a modal window like fl_message() was opened in the callback but there
could be other reasons as well. See issue #450 for more info.
This commit prevents "timer loops" without intermediate event handling
if callback handling takes longer than the timer delay of repeating
timers.
For more details see GitHub issue #450.
The solution to avoid blurred graphics when the GUI is scaled is to turn off
cairo antialiasing when drawing horizontal and vertical lines and rectangles.
- emphasize that Fl_Image::copy(W, H) creates an image with
w() == data_w() == W and h() == data_h() == H
- clarify some more docs of Fl_Image methods (ensure that data_w()
and data_h() are used where appropriate rather than w() and h()
- improve wording, examples, and formatting of related docs.
Closes#431
Copying an image does not (and must not) change the original object,
hence copy() should always be 'const'.
This is *necessary* if the given Fl_Image object is 'const'.
- Add this include statement only where needed.
- Rename Fl_System_Driver::stat() to flstat().
This fixes an issue when using some (!) MinGW 64-bit build systems
that obviously '#define stat _stat64' or similar. This would
result in compiler problems if 'stat()' is a member function of
Fl_System_Driver.
Regenerated src/fl_cmap.h with less spaces, updated copyright year.
Only whitespace and comments are changed in this file.
Note that src/cmap.cxx is not compiled to build the FLTK libs,
it's only used to generate src/fl_cmap.h.
These two functions are fully defined in the platform specific driver
methods, there's no need to define them in the base class.
This quick fix returns 0 from the base class method. The main reason
was to avoid a compiler error of MSVC 2010 and earlier, i.e. before
MSVC 2012.
Todo: int Fl_WinAPI_System_Driver::clocale_sscanf() needs to be fixed
because '_vsscanf_l()' (with lower case 'L' for 'locale') does not
exist in MSVC 2010 and earlier. It is not clear yet in which version
it was added - current MS docs show it in MSVC 2015 and later.
Code like "void copy(const char */*stuff*/, ...)" would issue the
warning "'*/' seen outside comment" which is formally true but ...
These warnings don't appear with newer MS compilers, but I fixed them
anyway.
Remove false "optimization" that would return the same pointer
if the requested width and height are the same.
Note: copy() must always return a new image (this is the expected
behavior of copy). Otherwise deleting one of the objects (either
the source or the copy) would delete the other one as well.
- fix window size of GL windows as reported for test/glpuzzle.cxx
- fix window not resizing on ctrl/+/- (X11 only)
Unfortunately the previous commit "optimized" a necessary X11 message
away. This is now fixed.
Fixes#392.
