tests: thread-based client harness
This replaces the old test harness with a new one.
The old harness relied on fork()'ing each test which makes tests independent,
but makes debugging them harder. The new harness runs client code in a thread
instead of a new process. A side-effect of not fork()'ing anymore is that any
failure will stop running a test series short. Fortunately we do not have any
tests that are expected to crash or fail.
The old harness executed 'weston' from Meson, with lots of setup as both
command line options and environment variables. The new harness executes
wet_main() instead: the test program itself calls the compositor main function
to execute the compositor in-process. Command line arguments are configured in
the test program itself, not in meson.build. Environment variables aside, you
are able to run a test by simply executing the test program, even if it is a
plugin test.
The new harness adds a new type of iteration: fixtures. For now, fixtures are
used to set up the compositor for tests that need a compositor. If necessary, a
fixture setup may include a data array of arbitrary type for executing the test
series for each element in the array. This will be most useful for running
screenshooting tests with both Pixman- and GL-renderers.
The new harness outputs TAP formatted results into stdout. Meson is not
switched to consume TAP yet though, because it would require a Meson version
requirement bump and would not have any benefits at this time. OTOH outputting
TAP is trivial and sets up a clear precedent of random test chatter belonging
to stderr.
This commit migrates only few tests to actually make use of the new features:
roles is a basic client test, subsurface-shot is a client test that
demonstrates the fixture array, and plugin-registry is a plugin test. The rest
of the tests will be migrated later.
Once all tests are migrated, we can remove the test-specific setup from
meson.build, leaving only the actual build instructions in there.
The not migrated tests and stand-alone tests suffer only a minor change: they
no longer fork() for each TEST(), otherwise they keep running as before.
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.com>
2019-11-01 15:02:15 +03:00
|
|
|
/*
|
|
|
|
* Copyright 2019 Collabora, Ltd.
|
|
|
|
*
|
|
|
|
* Permission is hereby granted, free of charge, to any person obtaining
|
|
|
|
* a copy of this software and associated documentation files (the
|
|
|
|
* "Software"), to deal in the Software without restriction, including
|
|
|
|
* without limitation the rights to use, copy, modify, merge, publish,
|
|
|
|
* distribute, sublicense, and/or sell copies of the Software, and to
|
|
|
|
* permit persons to whom the Software is furnished to do so, subject to
|
|
|
|
* the following conditions:
|
|
|
|
*
|
|
|
|
* The above copyright notice and this permission notice (including the
|
|
|
|
* next paragraph) shall be included in all copies or substantial
|
|
|
|
* portions of the Software.
|
|
|
|
*
|
|
|
|
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
|
|
|
|
* EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
|
|
|
|
* MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
|
|
|
|
* NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
|
|
|
|
* BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
|
|
|
|
* ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
|
|
|
|
* CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
|
|
|
|
* SOFTWARE.
|
|
|
|
*/
|
|
|
|
|
|
|
|
#include "config.h"
|
|
|
|
|
|
|
|
#include <string.h>
|
|
|
|
#include <assert.h>
|
2020-05-06 17:21:03 +03:00
|
|
|
#include <libudev.h>
|
|
|
|
#include <unistd.h>
|
|
|
|
#include <sys/file.h>
|
2020-08-16 03:19:13 +03:00
|
|
|
#include <sys/types.h>
|
|
|
|
#include <sys/stat.h>
|
|
|
|
#include <stdarg.h>
|
|
|
|
#include <stdlib.h>
|
2020-05-06 17:21:03 +03:00
|
|
|
#include <errno.h>
|
tests: include fcntl.h for open(), O_RDWR, O_CLOEXEC and O_CREAT
musl libc (unlike glibc) requires explicitly incuding fcntl.h to define open(),
O_RDWR, O_CLOEXEC and O_CREAT. Otherwise the build fails with the errors:
| ../weston-9.0.0/tests/weston-test-fixture-compositor.c: In function 'wait_for_lock':
| ../weston-9.0.0/tests/weston-test-fixture-compositor.c:135:7: warning: implicit declaration of function 'open'; did you mean 'popen'? [-Wimplicit-function-declaration]
| 135 | fd = open(lock_path, O_RDWR | O_CLOEXEC | O_CREAT, 00700);
| | ^~~~
| | popen
| ../weston-9.0.0/tests/weston-test-fixture-compositor.c:135:23: error: 'O_RDWR' undeclared (first use in this function)
| 135 | fd = open(lock_path, O_RDWR | O_CLOEXEC | O_CREAT, 00700);
| | ^~~~~~
| ../weston-9.0.0/tests/weston-test-fixture-compositor.c:135:23: note: each undeclared identifier is reported only once for each function it appears in
| ../weston-9.0.0/tests/weston-test-fixture-compositor.c:135:32: error: 'O_CLOEXEC' undeclared (first use in this function)
| 135 | fd = open(lock_path, O_RDWR | O_CLOEXEC | O_CREAT, 00700);
| | ^~~~~~~~~
| ../weston-9.0.0/tests/weston-test-fixture-compositor.c:135:44: error: 'O_CREAT' undeclared (first use in this function)
| 135 | fd = open(lock_path, O_RDWR | O_CLOEXEC | O_CREAT, 00700);
| | ^~~~~~~
Signed-off-by: Denys Dmytriyenko <denys@ti.com>
2020-09-09 02:37:42 +03:00
|
|
|
#include <fcntl.h>
|
tests: thread-based client harness
This replaces the old test harness with a new one.
The old harness relied on fork()'ing each test which makes tests independent,
but makes debugging them harder. The new harness runs client code in a thread
instead of a new process. A side-effect of not fork()'ing anymore is that any
failure will stop running a test series short. Fortunately we do not have any
tests that are expected to crash or fail.
The old harness executed 'weston' from Meson, with lots of setup as both
command line options and environment variables. The new harness executes
wet_main() instead: the test program itself calls the compositor main function
to execute the compositor in-process. Command line arguments are configured in
the test program itself, not in meson.build. Environment variables aside, you
are able to run a test by simply executing the test program, even if it is a
plugin test.
The new harness adds a new type of iteration: fixtures. For now, fixtures are
used to set up the compositor for tests that need a compositor. If necessary, a
fixture setup may include a data array of arbitrary type for executing the test
series for each element in the array. This will be most useful for running
screenshooting tests with both Pixman- and GL-renderers.
The new harness outputs TAP formatted results into stdout. Meson is not
switched to consume TAP yet though, because it would require a Meson version
requirement bump and would not have any benefits at this time. OTOH outputting
TAP is trivial and sets up a clear precedent of random test chatter belonging
to stderr.
This commit migrates only few tests to actually make use of the new features:
roles is a basic client test, subsurface-shot is a client test that
demonstrates the fixture array, and plugin-registry is a plugin test. The rest
of the tests will be migrated later.
Once all tests are migrated, we can remove the test-specific setup from
meson.build, leaving only the actual build instructions in there.
The not migrated tests and stand-alone tests suffer only a minor change: they
no longer fork() for each TEST(), otherwise they keep running as before.
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.com>
2019-11-01 15:02:15 +03:00
|
|
|
|
|
|
|
#include "shared/helpers.h"
|
2022-03-08 16:30:04 +03:00
|
|
|
#include "shared/string-helpers.h"
|
tests: thread-based client harness
This replaces the old test harness with a new one.
The old harness relied on fork()'ing each test which makes tests independent,
but makes debugging them harder. The new harness runs client code in a thread
instead of a new process. A side-effect of not fork()'ing anymore is that any
failure will stop running a test series short. Fortunately we do not have any
tests that are expected to crash or fail.
The old harness executed 'weston' from Meson, with lots of setup as both
command line options and environment variables. The new harness executes
wet_main() instead: the test program itself calls the compositor main function
to execute the compositor in-process. Command line arguments are configured in
the test program itself, not in meson.build. Environment variables aside, you
are able to run a test by simply executing the test program, even if it is a
plugin test.
The new harness adds a new type of iteration: fixtures. For now, fixtures are
used to set up the compositor for tests that need a compositor. If necessary, a
fixture setup may include a data array of arbitrary type for executing the test
series for each element in the array. This will be most useful for running
screenshooting tests with both Pixman- and GL-renderers.
The new harness outputs TAP formatted results into stdout. Meson is not
switched to consume TAP yet though, because it would require a Meson version
requirement bump and would not have any benefits at this time. OTOH outputting
TAP is trivial and sets up a clear precedent of random test chatter belonging
to stderr.
This commit migrates only few tests to actually make use of the new features:
roles is a basic client test, subsurface-shot is a client test that
demonstrates the fixture array, and plugin-registry is a plugin test. The rest
of the tests will be migrated later.
Once all tests are migrated, we can remove the test-specific setup from
meson.build, leaving only the actual build instructions in there.
The not migrated tests and stand-alone tests suffer only a minor change: they
no longer fork() for each TEST(), otherwise they keep running as before.
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.com>
2019-11-01 15:02:15 +03:00
|
|
|
#include "weston-test-fixture-compositor.h"
|
|
|
|
#include "weston.h"
|
2019-11-01 16:39:30 +03:00
|
|
|
#include "test-config.h"
|
tests: thread-based client harness
This replaces the old test harness with a new one.
The old harness relied on fork()'ing each test which makes tests independent,
but makes debugging them harder. The new harness runs client code in a thread
instead of a new process. A side-effect of not fork()'ing anymore is that any
failure will stop running a test series short. Fortunately we do not have any
tests that are expected to crash or fail.
The old harness executed 'weston' from Meson, with lots of setup as both
command line options and environment variables. The new harness executes
wet_main() instead: the test program itself calls the compositor main function
to execute the compositor in-process. Command line arguments are configured in
the test program itself, not in meson.build. Environment variables aside, you
are able to run a test by simply executing the test program, even if it is a
plugin test.
The new harness adds a new type of iteration: fixtures. For now, fixtures are
used to set up the compositor for tests that need a compositor. If necessary, a
fixture setup may include a data array of arbitrary type for executing the test
series for each element in the array. This will be most useful for running
screenshooting tests with both Pixman- and GL-renderers.
The new harness outputs TAP formatted results into stdout. Meson is not
switched to consume TAP yet though, because it would require a Meson version
requirement bump and would not have any benefits at this time. OTOH outputting
TAP is trivial and sets up a clear precedent of random test chatter belonging
to stderr.
This commit migrates only few tests to actually make use of the new features:
roles is a basic client test, subsurface-shot is a client test that
demonstrates the fixture array, and plugin-registry is a plugin test. The rest
of the tests will be migrated later.
Once all tests are migrated, we can remove the test-specific setup from
meson.build, leaving only the actual build instructions in there.
The not migrated tests and stand-alone tests suffer only a minor change: they
no longer fork() for each TEST(), otherwise they keep running as before.
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.com>
2019-11-01 15:02:15 +03:00
|
|
|
|
2021-03-01 12:24:59 +03:00
|
|
|
static_assert(WET_MAIN_RET_MISSING_CAPS == RESULT_SKIP,
|
|
|
|
"wet_main() return value for skip is wrong");
|
|
|
|
|
tests: thread-based client harness
This replaces the old test harness with a new one.
The old harness relied on fork()'ing each test which makes tests independent,
but makes debugging them harder. The new harness runs client code in a thread
instead of a new process. A side-effect of not fork()'ing anymore is that any
failure will stop running a test series short. Fortunately we do not have any
tests that are expected to crash or fail.
The old harness executed 'weston' from Meson, with lots of setup as both
command line options and environment variables. The new harness executes
wet_main() instead: the test program itself calls the compositor main function
to execute the compositor in-process. Command line arguments are configured in
the test program itself, not in meson.build. Environment variables aside, you
are able to run a test by simply executing the test program, even if it is a
plugin test.
The new harness adds a new type of iteration: fixtures. For now, fixtures are
used to set up the compositor for tests that need a compositor. If necessary, a
fixture setup may include a data array of arbitrary type for executing the test
series for each element in the array. This will be most useful for running
screenshooting tests with both Pixman- and GL-renderers.
The new harness outputs TAP formatted results into stdout. Meson is not
switched to consume TAP yet though, because it would require a Meson version
requirement bump and would not have any benefits at this time. OTOH outputting
TAP is trivial and sets up a clear precedent of random test chatter belonging
to stderr.
This commit migrates only few tests to actually make use of the new features:
roles is a basic client test, subsurface-shot is a client test that
demonstrates the fixture array, and plugin-registry is a plugin test. The rest
of the tests will be migrated later.
Once all tests are migrated, we can remove the test-specific setup from
meson.build, leaving only the actual build instructions in there.
The not migrated tests and stand-alone tests suffer only a minor change: they
no longer fork() for each TEST(), otherwise they keep running as before.
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.com>
2019-11-01 15:02:15 +03:00
|
|
|
struct prog_args {
|
|
|
|
int argc;
|
|
|
|
char **argv;
|
|
|
|
char **saved;
|
|
|
|
int alloc;
|
|
|
|
};
|
|
|
|
|
|
|
|
static void
|
|
|
|
prog_args_init(struct prog_args *p)
|
|
|
|
{
|
|
|
|
memset(p, 0, sizeof(*p));
|
|
|
|
}
|
|
|
|
|
|
|
|
static void
|
|
|
|
prog_args_take(struct prog_args *p, char *arg)
|
|
|
|
{
|
|
|
|
assert(arg);
|
|
|
|
|
|
|
|
if (p->argc == p->alloc) {
|
|
|
|
p->alloc += 10;
|
|
|
|
p->argv = realloc(p->argv, sizeof(char *) * p->alloc);
|
|
|
|
assert(p->argv);
|
|
|
|
}
|
|
|
|
|
|
|
|
p->argv[p->argc++] = arg;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* The program to be executed will trample on argv, hence we need a copy to
|
|
|
|
* be able to free all our args.
|
|
|
|
*/
|
|
|
|
static void
|
|
|
|
prog_args_save(struct prog_args *p)
|
|
|
|
{
|
|
|
|
assert(p->saved == NULL);
|
|
|
|
|
|
|
|
p->saved = calloc(p->argc, sizeof(char *));
|
|
|
|
assert(p->saved);
|
|
|
|
|
|
|
|
memcpy(p->saved, p->argv, sizeof(char *) * p->argc);
|
|
|
|
}
|
|
|
|
|
|
|
|
static void
|
|
|
|
prog_args_fini(struct prog_args *p)
|
|
|
|
{
|
|
|
|
int i;
|
|
|
|
|
2022-06-24 15:38:35 +03:00
|
|
|
/* If our args have never been saved, then we haven't called the
|
|
|
|
* compositor, but we still need to free the args, not leak them. */
|
|
|
|
if (!p->saved)
|
|
|
|
prog_args_save(p);
|
|
|
|
|
2021-10-21 16:44:55 +03:00
|
|
|
if (p->saved) {
|
|
|
|
for (i = 0; i < p->argc; i++)
|
|
|
|
free(p->saved[i]);
|
|
|
|
free(p->saved);
|
|
|
|
}
|
tests: thread-based client harness
This replaces the old test harness with a new one.
The old harness relied on fork()'ing each test which makes tests independent,
but makes debugging them harder. The new harness runs client code in a thread
instead of a new process. A side-effect of not fork()'ing anymore is that any
failure will stop running a test series short. Fortunately we do not have any
tests that are expected to crash or fail.
The old harness executed 'weston' from Meson, with lots of setup as both
command line options and environment variables. The new harness executes
wet_main() instead: the test program itself calls the compositor main function
to execute the compositor in-process. Command line arguments are configured in
the test program itself, not in meson.build. Environment variables aside, you
are able to run a test by simply executing the test program, even if it is a
plugin test.
The new harness adds a new type of iteration: fixtures. For now, fixtures are
used to set up the compositor for tests that need a compositor. If necessary, a
fixture setup may include a data array of arbitrary type for executing the test
series for each element in the array. This will be most useful for running
screenshooting tests with both Pixman- and GL-renderers.
The new harness outputs TAP formatted results into stdout. Meson is not
switched to consume TAP yet though, because it would require a Meson version
requirement bump and would not have any benefits at this time. OTOH outputting
TAP is trivial and sets up a clear precedent of random test chatter belonging
to stderr.
This commit migrates only few tests to actually make use of the new features:
roles is a basic client test, subsurface-shot is a client test that
demonstrates the fixture array, and plugin-registry is a plugin test. The rest
of the tests will be migrated later.
Once all tests are migrated, we can remove the test-specific setup from
meson.build, leaving only the actual build instructions in there.
The not migrated tests and stand-alone tests suffer only a minor change: they
no longer fork() for each TEST(), otherwise they keep running as before.
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.com>
2019-11-01 15:02:15 +03:00
|
|
|
|
|
|
|
free(p->argv);
|
|
|
|
prog_args_init(p);
|
|
|
|
}
|
|
|
|
|
2020-05-06 17:21:03 +03:00
|
|
|
static char *
|
|
|
|
get_lock_path(void)
|
|
|
|
{
|
|
|
|
const char *env_path, *suffix;
|
|
|
|
char *lock_path;
|
|
|
|
|
|
|
|
suffix = "weston-test-suite-drm-lock";
|
|
|
|
env_path = getenv("XDG_RUNTIME_DIR");
|
|
|
|
if (!env_path) {
|
|
|
|
fprintf(stderr, "Failed to compute lock file path. " \
|
|
|
|
"XDG_RUNTIME_DIR is not set.\n");
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
|
2022-03-08 16:30:04 +03:00
|
|
|
str_printf(&lock_path, "%s/%s", env_path, suffix);
|
|
|
|
if (!lock_path)
|
2020-05-06 17:21:03 +03:00
|
|
|
return NULL;
|
|
|
|
|
|
|
|
return lock_path;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* DRM-backend tests need to be run sequentially, since there can only be
|
|
|
|
* one user at a time with master status in a DRM KMS device. Since the
|
|
|
|
* test suite runs the tests in parallel, there's a mechanism to assure
|
|
|
|
* only one DRM-backend test is running at a time: tests of this type keep
|
|
|
|
* waiting until they acquire a lock (which is hold until they end).
|
|
|
|
*
|
|
|
|
* Returns -1 in failure and fd in success.
|
|
|
|
*/
|
|
|
|
static int
|
|
|
|
wait_for_lock(void)
|
|
|
|
{
|
|
|
|
char *lock_path;
|
|
|
|
int fd;
|
|
|
|
|
|
|
|
lock_path = get_lock_path();
|
|
|
|
if (!lock_path)
|
|
|
|
goto err_path;
|
|
|
|
|
|
|
|
fd = open(lock_path, O_RDWR | O_CLOEXEC | O_CREAT, 00700);
|
|
|
|
if (fd == -1) {
|
|
|
|
fprintf(stderr, "Could not open lock file %s: %s\n", lock_path, strerror(errno));
|
|
|
|
goto err_open;
|
|
|
|
}
|
|
|
|
fprintf(stderr, "Waiting for lock on %s...\n", lock_path);
|
|
|
|
|
|
|
|
/* The call is blocking, so we don't need a 'while'. Also, as we
|
|
|
|
* have a timeout for each test, this won't get stuck waiting. */
|
|
|
|
if (flock(fd, LOCK_EX) == -1) {
|
|
|
|
fprintf(stderr, "Could not lock %s: %s\n", lock_path, strerror(errno));
|
|
|
|
goto err_lock;
|
|
|
|
}
|
|
|
|
fprintf(stderr, "Lock %s acquired.\n", lock_path);
|
|
|
|
free(lock_path);
|
|
|
|
|
|
|
|
return fd;
|
|
|
|
|
|
|
|
err_lock:
|
|
|
|
close(fd);
|
|
|
|
err_open:
|
|
|
|
free(lock_path);
|
|
|
|
err_path:
|
|
|
|
return -1;
|
|
|
|
}
|
|
|
|
|
tests: thread-based client harness
This replaces the old test harness with a new one.
The old harness relied on fork()'ing each test which makes tests independent,
but makes debugging them harder. The new harness runs client code in a thread
instead of a new process. A side-effect of not fork()'ing anymore is that any
failure will stop running a test series short. Fortunately we do not have any
tests that are expected to crash or fail.
The old harness executed 'weston' from Meson, with lots of setup as both
command line options and environment variables. The new harness executes
wet_main() instead: the test program itself calls the compositor main function
to execute the compositor in-process. Command line arguments are configured in
the test program itself, not in meson.build. Environment variables aside, you
are able to run a test by simply executing the test program, even if it is a
plugin test.
The new harness adds a new type of iteration: fixtures. For now, fixtures are
used to set up the compositor for tests that need a compositor. If necessary, a
fixture setup may include a data array of arbitrary type for executing the test
series for each element in the array. This will be most useful for running
screenshooting tests with both Pixman- and GL-renderers.
The new harness outputs TAP formatted results into stdout. Meson is not
switched to consume TAP yet though, because it would require a Meson version
requirement bump and would not have any benefits at this time. OTOH outputting
TAP is trivial and sets up a clear precedent of random test chatter belonging
to stderr.
This commit migrates only few tests to actually make use of the new features:
roles is a basic client test, subsurface-shot is a client test that
demonstrates the fixture array, and plugin-registry is a plugin test. The rest
of the tests will be migrated later.
Once all tests are migrated, we can remove the test-specific setup from
meson.build, leaving only the actual build instructions in there.
The not migrated tests and stand-alone tests suffer only a minor change: they
no longer fork() for each TEST(), otherwise they keep running as before.
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.com>
2019-11-01 15:02:15 +03:00
|
|
|
/** Initialize part of compositor setup
|
|
|
|
*
|
|
|
|
* \param setup The variable to initialize.
|
|
|
|
* \param testset_name Value for testset_name member.
|
|
|
|
*
|
|
|
|
* \ingroup testharness_private
|
|
|
|
*/
|
|
|
|
void
|
|
|
|
compositor_setup_defaults_(struct compositor_setup *setup,
|
|
|
|
const char *testset_name)
|
|
|
|
{
|
|
|
|
*setup = (struct compositor_setup) {
|
2020-10-19 22:06:22 +03:00
|
|
|
.test_quirks = (struct weston_testsuite_quirks){ },
|
tests: thread-based client harness
This replaces the old test harness with a new one.
The old harness relied on fork()'ing each test which makes tests independent,
but makes debugging them harder. The new harness runs client code in a thread
instead of a new process. A side-effect of not fork()'ing anymore is that any
failure will stop running a test series short. Fortunately we do not have any
tests that are expected to crash or fail.
The old harness executed 'weston' from Meson, with lots of setup as both
command line options and environment variables. The new harness executes
wet_main() instead: the test program itself calls the compositor main function
to execute the compositor in-process. Command line arguments are configured in
the test program itself, not in meson.build. Environment variables aside, you
are able to run a test by simply executing the test program, even if it is a
plugin test.
The new harness adds a new type of iteration: fixtures. For now, fixtures are
used to set up the compositor for tests that need a compositor. If necessary, a
fixture setup may include a data array of arbitrary type for executing the test
series for each element in the array. This will be most useful for running
screenshooting tests with both Pixman- and GL-renderers.
The new harness outputs TAP formatted results into stdout. Meson is not
switched to consume TAP yet though, because it would require a Meson version
requirement bump and would not have any benefits at this time. OTOH outputting
TAP is trivial and sets up a clear precedent of random test chatter belonging
to stderr.
This commit migrates only few tests to actually make use of the new features:
roles is a basic client test, subsurface-shot is a client test that
demonstrates the fixture array, and plugin-registry is a plugin test. The rest
of the tests will be migrated later.
Once all tests are migrated, we can remove the test-specific setup from
meson.build, leaving only the actual build instructions in there.
The not migrated tests and stand-alone tests suffer only a minor change: they
no longer fork() for each TEST(), otherwise they keep running as before.
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.com>
2019-11-01 15:02:15 +03:00
|
|
|
.backend = WESTON_BACKEND_HEADLESS,
|
2022-12-05 19:05:21 +03:00
|
|
|
.renderer = WESTON_RENDERER_NOOP,
|
2022-11-03 20:23:14 +03:00
|
|
|
.shell = SHELL_TEST_DESKTOP,
|
tests: thread-based client harness
This replaces the old test harness with a new one.
The old harness relied on fork()'ing each test which makes tests independent,
but makes debugging them harder. The new harness runs client code in a thread
instead of a new process. A side-effect of not fork()'ing anymore is that any
failure will stop running a test series short. Fortunately we do not have any
tests that are expected to crash or fail.
The old harness executed 'weston' from Meson, with lots of setup as both
command line options and environment variables. The new harness executes
wet_main() instead: the test program itself calls the compositor main function
to execute the compositor in-process. Command line arguments are configured in
the test program itself, not in meson.build. Environment variables aside, you
are able to run a test by simply executing the test program, even if it is a
plugin test.
The new harness adds a new type of iteration: fixtures. For now, fixtures are
used to set up the compositor for tests that need a compositor. If necessary, a
fixture setup may include a data array of arbitrary type for executing the test
series for each element in the array. This will be most useful for running
screenshooting tests with both Pixman- and GL-renderers.
The new harness outputs TAP formatted results into stdout. Meson is not
switched to consume TAP yet though, because it would require a Meson version
requirement bump and would not have any benefits at this time. OTOH outputting
TAP is trivial and sets up a clear precedent of random test chatter belonging
to stderr.
This commit migrates only few tests to actually make use of the new features:
roles is a basic client test, subsurface-shot is a client test that
demonstrates the fixture array, and plugin-registry is a plugin test. The rest
of the tests will be migrated later.
Once all tests are migrated, we can remove the test-specific setup from
meson.build, leaving only the actual build instructions in there.
The not migrated tests and stand-alone tests suffer only a minor change: they
no longer fork() for each TEST(), otherwise they keep running as before.
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.com>
2019-11-01 15:02:15 +03:00
|
|
|
.xwayland = false,
|
|
|
|
.width = 320,
|
|
|
|
.height = 240,
|
2020-01-22 12:53:12 +03:00
|
|
|
.scale = 1,
|
|
|
|
.transform = WL_OUTPUT_TRANSFORM_NORMAL,
|
tests: thread-based client harness
This replaces the old test harness with a new one.
The old harness relied on fork()'ing each test which makes tests independent,
but makes debugging them harder. The new harness runs client code in a thread
instead of a new process. A side-effect of not fork()'ing anymore is that any
failure will stop running a test series short. Fortunately we do not have any
tests that are expected to crash or fail.
The old harness executed 'weston' from Meson, with lots of setup as both
command line options and environment variables. The new harness executes
wet_main() instead: the test program itself calls the compositor main function
to execute the compositor in-process. Command line arguments are configured in
the test program itself, not in meson.build. Environment variables aside, you
are able to run a test by simply executing the test program, even if it is a
plugin test.
The new harness adds a new type of iteration: fixtures. For now, fixtures are
used to set up the compositor for tests that need a compositor. If necessary, a
fixture setup may include a data array of arbitrary type for executing the test
series for each element in the array. This will be most useful for running
screenshooting tests with both Pixman- and GL-renderers.
The new harness outputs TAP formatted results into stdout. Meson is not
switched to consume TAP yet though, because it would require a Meson version
requirement bump and would not have any benefits at this time. OTOH outputting
TAP is trivial and sets up a clear precedent of random test chatter belonging
to stderr.
This commit migrates only few tests to actually make use of the new features:
roles is a basic client test, subsurface-shot is a client test that
demonstrates the fixture array, and plugin-registry is a plugin test. The rest
of the tests will be migrated later.
Once all tests are migrated, we can remove the test-specific setup from
meson.build, leaving only the actual build instructions in there.
The not migrated tests and stand-alone tests suffer only a minor change: they
no longer fork() for each TEST(), otherwise they keep running as before.
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.com>
2019-11-01 15:02:15 +03:00
|
|
|
.config_file = NULL,
|
|
|
|
.extra_module = NULL,
|
|
|
|
.logging_scopes = NULL,
|
|
|
|
.testset_name = testset_name,
|
|
|
|
};
|
|
|
|
}
|
|
|
|
|
|
|
|
static const char *
|
|
|
|
backend_to_str(enum weston_compositor_backend b)
|
|
|
|
{
|
|
|
|
static const char * const names[] = {
|
2022-12-05 19:28:23 +03:00
|
|
|
[WESTON_BACKEND_DRM] = "drm",
|
|
|
|
[WESTON_BACKEND_HEADLESS] = "headless",
|
|
|
|
[WESTON_BACKEND_RDP] = "rdp",
|
|
|
|
[WESTON_BACKEND_VNC] = "vnc",
|
|
|
|
[WESTON_BACKEND_WAYLAND] = "wayland",
|
|
|
|
[WESTON_BACKEND_X11] = "x11",
|
tests: thread-based client harness
This replaces the old test harness with a new one.
The old harness relied on fork()'ing each test which makes tests independent,
but makes debugging them harder. The new harness runs client code in a thread
instead of a new process. A side-effect of not fork()'ing anymore is that any
failure will stop running a test series short. Fortunately we do not have any
tests that are expected to crash or fail.
The old harness executed 'weston' from Meson, with lots of setup as both
command line options and environment variables. The new harness executes
wet_main() instead: the test program itself calls the compositor main function
to execute the compositor in-process. Command line arguments are configured in
the test program itself, not in meson.build. Environment variables aside, you
are able to run a test by simply executing the test program, even if it is a
plugin test.
The new harness adds a new type of iteration: fixtures. For now, fixtures are
used to set up the compositor for tests that need a compositor. If necessary, a
fixture setup may include a data array of arbitrary type for executing the test
series for each element in the array. This will be most useful for running
screenshooting tests with both Pixman- and GL-renderers.
The new harness outputs TAP formatted results into stdout. Meson is not
switched to consume TAP yet though, because it would require a Meson version
requirement bump and would not have any benefits at this time. OTOH outputting
TAP is trivial and sets up a clear precedent of random test chatter belonging
to stderr.
This commit migrates only few tests to actually make use of the new features:
roles is a basic client test, subsurface-shot is a client test that
demonstrates the fixture array, and plugin-registry is a plugin test. The rest
of the tests will be migrated later.
Once all tests are migrated, we can remove the test-specific setup from
meson.build, leaving only the actual build instructions in there.
The not migrated tests and stand-alone tests suffer only a minor change: they
no longer fork() for each TEST(), otherwise they keep running as before.
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.com>
2019-11-01 15:02:15 +03:00
|
|
|
};
|
|
|
|
assert(b >= 0 && b < ARRAY_LENGTH(names));
|
|
|
|
return names[b];
|
|
|
|
}
|
|
|
|
|
|
|
|
static const char *
|
2022-12-05 19:05:21 +03:00
|
|
|
renderer_to_str(enum weston_renderer_type t)
|
tests: thread-based client harness
This replaces the old test harness with a new one.
The old harness relied on fork()'ing each test which makes tests independent,
but makes debugging them harder. The new harness runs client code in a thread
instead of a new process. A side-effect of not fork()'ing anymore is that any
failure will stop running a test series short. Fortunately we do not have any
tests that are expected to crash or fail.
The old harness executed 'weston' from Meson, with lots of setup as both
command line options and environment variables. The new harness executes
wet_main() instead: the test program itself calls the compositor main function
to execute the compositor in-process. Command line arguments are configured in
the test program itself, not in meson.build. Environment variables aside, you
are able to run a test by simply executing the test program, even if it is a
plugin test.
The new harness adds a new type of iteration: fixtures. For now, fixtures are
used to set up the compositor for tests that need a compositor. If necessary, a
fixture setup may include a data array of arbitrary type for executing the test
series for each element in the array. This will be most useful for running
screenshooting tests with both Pixman- and GL-renderers.
The new harness outputs TAP formatted results into stdout. Meson is not
switched to consume TAP yet though, because it would require a Meson version
requirement bump and would not have any benefits at this time. OTOH outputting
TAP is trivial and sets up a clear precedent of random test chatter belonging
to stderr.
This commit migrates only few tests to actually make use of the new features:
roles is a basic client test, subsurface-shot is a client test that
demonstrates the fixture array, and plugin-registry is a plugin test. The rest
of the tests will be migrated later.
Once all tests are migrated, we can remove the test-specific setup from
meson.build, leaving only the actual build instructions in there.
The not migrated tests and stand-alone tests suffer only a minor change: they
no longer fork() for each TEST(), otherwise they keep running as before.
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.com>
2019-11-01 15:02:15 +03:00
|
|
|
{
|
2022-12-30 00:12:24 +03:00
|
|
|
static const char * const names[] = {
|
2022-12-05 19:05:21 +03:00
|
|
|
[WESTON_RENDERER_NOOP] = "noop",
|
|
|
|
[WESTON_RENDERER_PIXMAN] = "pixman",
|
|
|
|
[WESTON_RENDERER_GL] = "gl",
|
2020-05-06 00:14:25 +03:00
|
|
|
};
|
2022-12-30 00:12:24 +03:00
|
|
|
assert(t >= 0 && t <= ARRAY_LENGTH(names));
|
|
|
|
return names[t];
|
tests: thread-based client harness
This replaces the old test harness with a new one.
The old harness relied on fork()'ing each test which makes tests independent,
but makes debugging them harder. The new harness runs client code in a thread
instead of a new process. A side-effect of not fork()'ing anymore is that any
failure will stop running a test series short. Fortunately we do not have any
tests that are expected to crash or fail.
The old harness executed 'weston' from Meson, with lots of setup as both
command line options and environment variables. The new harness executes
wet_main() instead: the test program itself calls the compositor main function
to execute the compositor in-process. Command line arguments are configured in
the test program itself, not in meson.build. Environment variables aside, you
are able to run a test by simply executing the test program, even if it is a
plugin test.
The new harness adds a new type of iteration: fixtures. For now, fixtures are
used to set up the compositor for tests that need a compositor. If necessary, a
fixture setup may include a data array of arbitrary type for executing the test
series for each element in the array. This will be most useful for running
screenshooting tests with both Pixman- and GL-renderers.
The new harness outputs TAP formatted results into stdout. Meson is not
switched to consume TAP yet though, because it would require a Meson version
requirement bump and would not have any benefits at this time. OTOH outputting
TAP is trivial and sets up a clear precedent of random test chatter belonging
to stderr.
This commit migrates only few tests to actually make use of the new features:
roles is a basic client test, subsurface-shot is a client test that
demonstrates the fixture array, and plugin-registry is a plugin test. The rest
of the tests will be migrated later.
Once all tests are migrated, we can remove the test-specific setup from
meson.build, leaving only the actual build instructions in there.
The not migrated tests and stand-alone tests suffer only a minor change: they
no longer fork() for each TEST(), otherwise they keep running as before.
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.com>
2019-11-01 15:02:15 +03:00
|
|
|
}
|
|
|
|
|
|
|
|
static const char *
|
|
|
|
shell_to_str(enum shell_type t)
|
|
|
|
{
|
|
|
|
static const char * const names[] = {
|
2022-12-29 18:38:37 +03:00
|
|
|
[SHELL_TEST_DESKTOP] = "weston-test-desktop",
|
|
|
|
[SHELL_DESKTOP] = "desktop",
|
|
|
|
[SHELL_FULLSCREEN] = "fullscreen",
|
|
|
|
[SHELL_IVI] = "ivi",
|
tests: thread-based client harness
This replaces the old test harness with a new one.
The old harness relied on fork()'ing each test which makes tests independent,
but makes debugging them harder. The new harness runs client code in a thread
instead of a new process. A side-effect of not fork()'ing anymore is that any
failure will stop running a test series short. Fortunately we do not have any
tests that are expected to crash or fail.
The old harness executed 'weston' from Meson, with lots of setup as both
command line options and environment variables. The new harness executes
wet_main() instead: the test program itself calls the compositor main function
to execute the compositor in-process. Command line arguments are configured in
the test program itself, not in meson.build. Environment variables aside, you
are able to run a test by simply executing the test program, even if it is a
plugin test.
The new harness adds a new type of iteration: fixtures. For now, fixtures are
used to set up the compositor for tests that need a compositor. If necessary, a
fixture setup may include a data array of arbitrary type for executing the test
series for each element in the array. This will be most useful for running
screenshooting tests with both Pixman- and GL-renderers.
The new harness outputs TAP formatted results into stdout. Meson is not
switched to consume TAP yet though, because it would require a Meson version
requirement bump and would not have any benefits at this time. OTOH outputting
TAP is trivial and sets up a clear precedent of random test chatter belonging
to stderr.
This commit migrates only few tests to actually make use of the new features:
roles is a basic client test, subsurface-shot is a client test that
demonstrates the fixture array, and plugin-registry is a plugin test. The rest
of the tests will be migrated later.
Once all tests are migrated, we can remove the test-specific setup from
meson.build, leaving only the actual build instructions in there.
The not migrated tests and stand-alone tests suffer only a minor change: they
no longer fork() for each TEST(), otherwise they keep running as before.
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.com>
2019-11-01 15:02:15 +03:00
|
|
|
};
|
|
|
|
assert(t >= 0 && t < ARRAY_LENGTH(names));
|
|
|
|
return names[t];
|
|
|
|
}
|
|
|
|
|
2020-01-22 12:53:12 +03:00
|
|
|
static const char *
|
|
|
|
transform_to_str(enum wl_output_transform t)
|
|
|
|
{
|
|
|
|
static const char * const names[] = {
|
|
|
|
[WL_OUTPUT_TRANSFORM_NORMAL] = "normal",
|
|
|
|
[WL_OUTPUT_TRANSFORM_90] = "rotate-90",
|
|
|
|
[WL_OUTPUT_TRANSFORM_180] = "rotate-180",
|
|
|
|
[WL_OUTPUT_TRANSFORM_270] = "rotate-270",
|
|
|
|
[WL_OUTPUT_TRANSFORM_FLIPPED] = "flipped",
|
|
|
|
[WL_OUTPUT_TRANSFORM_FLIPPED_90] = "flipped-rotate-90",
|
|
|
|
[WL_OUTPUT_TRANSFORM_FLIPPED_180] = "flipped-rotate-180",
|
|
|
|
[WL_OUTPUT_TRANSFORM_FLIPPED_270] = "flipped-rotate-270",
|
|
|
|
};
|
|
|
|
|
|
|
|
assert(t < ARRAY_LENGTH(names) && names[t]);
|
|
|
|
return names[t];
|
|
|
|
}
|
|
|
|
|
tests: thread-based client harness
This replaces the old test harness with a new one.
The old harness relied on fork()'ing each test which makes tests independent,
but makes debugging them harder. The new harness runs client code in a thread
instead of a new process. A side-effect of not fork()'ing anymore is that any
failure will stop running a test series short. Fortunately we do not have any
tests that are expected to crash or fail.
The old harness executed 'weston' from Meson, with lots of setup as both
command line options and environment variables. The new harness executes
wet_main() instead: the test program itself calls the compositor main function
to execute the compositor in-process. Command line arguments are configured in
the test program itself, not in meson.build. Environment variables aside, you
are able to run a test by simply executing the test program, even if it is a
plugin test.
The new harness adds a new type of iteration: fixtures. For now, fixtures are
used to set up the compositor for tests that need a compositor. If necessary, a
fixture setup may include a data array of arbitrary type for executing the test
series for each element in the array. This will be most useful for running
screenshooting tests with both Pixman- and GL-renderers.
The new harness outputs TAP formatted results into stdout. Meson is not
switched to consume TAP yet though, because it would require a Meson version
requirement bump and would not have any benefits at this time. OTOH outputting
TAP is trivial and sets up a clear precedent of random test chatter belonging
to stderr.
This commit migrates only few tests to actually make use of the new features:
roles is a basic client test, subsurface-shot is a client test that
demonstrates the fixture array, and plugin-registry is a plugin test. The rest
of the tests will be migrated later.
Once all tests are migrated, we can remove the test-specific setup from
meson.build, leaving only the actual build instructions in there.
The not migrated tests and stand-alone tests suffer only a minor change: they
no longer fork() for each TEST(), otherwise they keep running as before.
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.com>
2019-11-01 15:02:15 +03:00
|
|
|
/** Execute compositor
|
|
|
|
*
|
|
|
|
* Manufactures the compositor command line and calls wet_main().
|
|
|
|
*
|
|
|
|
* Returns RESULT_SKIP if the given setup contains features that were disabled
|
|
|
|
* in the build, e.g. GL-renderer or DRM-backend.
|
|
|
|
*
|
|
|
|
* \ingroup testharness_private
|
|
|
|
*/
|
|
|
|
int
|
|
|
|
execute_compositor(const struct compositor_setup *setup,
|
|
|
|
struct wet_testsuite_data *data)
|
|
|
|
{
|
2020-10-19 22:06:22 +03:00
|
|
|
struct weston_testsuite_data test_data;
|
tests: thread-based client harness
This replaces the old test harness with a new one.
The old harness relied on fork()'ing each test which makes tests independent,
but makes debugging them harder. The new harness runs client code in a thread
instead of a new process. A side-effect of not fork()'ing anymore is that any
failure will stop running a test series short. Fortunately we do not have any
tests that are expected to crash or fail.
The old harness executed 'weston' from Meson, with lots of setup as both
command line options and environment variables. The new harness executes
wet_main() instead: the test program itself calls the compositor main function
to execute the compositor in-process. Command line arguments are configured in
the test program itself, not in meson.build. Environment variables aside, you
are able to run a test by simply executing the test program, even if it is a
plugin test.
The new harness adds a new type of iteration: fixtures. For now, fixtures are
used to set up the compositor for tests that need a compositor. If necessary, a
fixture setup may include a data array of arbitrary type for executing the test
series for each element in the array. This will be most useful for running
screenshooting tests with both Pixman- and GL-renderers.
The new harness outputs TAP formatted results into stdout. Meson is not
switched to consume TAP yet though, because it would require a Meson version
requirement bump and would not have any benefits at this time. OTOH outputting
TAP is trivial and sets up a clear precedent of random test chatter belonging
to stderr.
This commit migrates only few tests to actually make use of the new features:
roles is a basic client test, subsurface-shot is a client test that
demonstrates the fixture array, and plugin-registry is a plugin test. The rest
of the tests will be migrated later.
Once all tests are migrated, we can remove the test-specific setup from
meson.build, leaving only the actual build instructions in there.
The not migrated tests and stand-alone tests suffer only a minor change: they
no longer fork() for each TEST(), otherwise they keep running as before.
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.com>
2019-11-01 15:02:15 +03:00
|
|
|
struct prog_args args;
|
|
|
|
char *tmp;
|
2022-12-30 00:12:24 +03:00
|
|
|
const char *drm_device;
|
2022-06-25 05:56:45 +03:00
|
|
|
int lock_fd = -1;
|
|
|
|
int ret = RESULT_OK;
|
tests: thread-based client harness
This replaces the old test harness with a new one.
The old harness relied on fork()'ing each test which makes tests independent,
but makes debugging them harder. The new harness runs client code in a thread
instead of a new process. A side-effect of not fork()'ing anymore is that any
failure will stop running a test series short. Fortunately we do not have any
tests that are expected to crash or fail.
The old harness executed 'weston' from Meson, with lots of setup as both
command line options and environment variables. The new harness executes
wet_main() instead: the test program itself calls the compositor main function
to execute the compositor in-process. Command line arguments are configured in
the test program itself, not in meson.build. Environment variables aside, you
are able to run a test by simply executing the test program, even if it is a
plugin test.
The new harness adds a new type of iteration: fixtures. For now, fixtures are
used to set up the compositor for tests that need a compositor. If necessary, a
fixture setup may include a data array of arbitrary type for executing the test
series for each element in the array. This will be most useful for running
screenshooting tests with both Pixman- and GL-renderers.
The new harness outputs TAP formatted results into stdout. Meson is not
switched to consume TAP yet though, because it would require a Meson version
requirement bump and would not have any benefits at this time. OTOH outputting
TAP is trivial and sets up a clear precedent of random test chatter belonging
to stderr.
This commit migrates only few tests to actually make use of the new features:
roles is a basic client test, subsurface-shot is a client test that
demonstrates the fixture array, and plugin-registry is a plugin test. The rest
of the tests will be migrated later.
Once all tests are migrated, we can remove the test-specific setup from
meson.build, leaving only the actual build instructions in there.
The not migrated tests and stand-alone tests suffer only a minor change: they
no longer fork() for each TEST(), otherwise they keep running as before.
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.com>
2019-11-01 15:02:15 +03:00
|
|
|
|
|
|
|
prog_args_init(&args);
|
|
|
|
|
|
|
|
/* argv[0] */
|
2022-03-08 16:30:04 +03:00
|
|
|
str_printf(&tmp, "weston-%s", setup->testset_name);
|
tests: thread-based client harness
This replaces the old test harness with a new one.
The old harness relied on fork()'ing each test which makes tests independent,
but makes debugging them harder. The new harness runs client code in a thread
instead of a new process. A side-effect of not fork()'ing anymore is that any
failure will stop running a test series short. Fortunately we do not have any
tests that are expected to crash or fail.
The old harness executed 'weston' from Meson, with lots of setup as both
command line options and environment variables. The new harness executes
wet_main() instead: the test program itself calls the compositor main function
to execute the compositor in-process. Command line arguments are configured in
the test program itself, not in meson.build. Environment variables aside, you
are able to run a test by simply executing the test program, even if it is a
plugin test.
The new harness adds a new type of iteration: fixtures. For now, fixtures are
used to set up the compositor for tests that need a compositor. If necessary, a
fixture setup may include a data array of arbitrary type for executing the test
series for each element in the array. This will be most useful for running
screenshooting tests with both Pixman- and GL-renderers.
The new harness outputs TAP formatted results into stdout. Meson is not
switched to consume TAP yet though, because it would require a Meson version
requirement bump and would not have any benefits at this time. OTOH outputting
TAP is trivial and sets up a clear precedent of random test chatter belonging
to stderr.
This commit migrates only few tests to actually make use of the new features:
roles is a basic client test, subsurface-shot is a client test that
demonstrates the fixture array, and plugin-registry is a plugin test. The rest
of the tests will be migrated later.
Once all tests are migrated, we can remove the test-specific setup from
meson.build, leaving only the actual build instructions in there.
The not migrated tests and stand-alone tests suffer only a minor change: they
no longer fork() for each TEST(), otherwise they keep running as before.
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.com>
2019-11-01 15:02:15 +03:00
|
|
|
prog_args_take(&args, tmp);
|
|
|
|
|
2022-03-08 16:30:04 +03:00
|
|
|
str_printf(&tmp, "--backend=%s", backend_to_str(setup->backend));
|
tests: thread-based client harness
This replaces the old test harness with a new one.
The old harness relied on fork()'ing each test which makes tests independent,
but makes debugging them harder. The new harness runs client code in a thread
instead of a new process. A side-effect of not fork()'ing anymore is that any
failure will stop running a test series short. Fortunately we do not have any
tests that are expected to crash or fail.
The old harness executed 'weston' from Meson, with lots of setup as both
command line options and environment variables. The new harness executes
wet_main() instead: the test program itself calls the compositor main function
to execute the compositor in-process. Command line arguments are configured in
the test program itself, not in meson.build. Environment variables aside, you
are able to run a test by simply executing the test program, even if it is a
plugin test.
The new harness adds a new type of iteration: fixtures. For now, fixtures are
used to set up the compositor for tests that need a compositor. If necessary, a
fixture setup may include a data array of arbitrary type for executing the test
series for each element in the array. This will be most useful for running
screenshooting tests with both Pixman- and GL-renderers.
The new harness outputs TAP formatted results into stdout. Meson is not
switched to consume TAP yet though, because it would require a Meson version
requirement bump and would not have any benefits at this time. OTOH outputting
TAP is trivial and sets up a clear precedent of random test chatter belonging
to stderr.
This commit migrates only few tests to actually make use of the new features:
roles is a basic client test, subsurface-shot is a client test that
demonstrates the fixture array, and plugin-registry is a plugin test. The rest
of the tests will be migrated later.
Once all tests are migrated, we can remove the test-specific setup from
meson.build, leaving only the actual build instructions in there.
The not migrated tests and stand-alone tests suffer only a minor change: they
no longer fork() for each TEST(), otherwise they keep running as before.
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.com>
2019-11-01 15:02:15 +03:00
|
|
|
prog_args_take(&args, tmp);
|
|
|
|
|
2020-05-06 17:21:03 +03:00
|
|
|
if (setup->backend == WESTON_BACKEND_DRM) {
|
|
|
|
|
|
|
|
drm_device = getenv("WESTON_TEST_SUITE_DRM_DEVICE");
|
|
|
|
if (!drm_device) {
|
|
|
|
fprintf(stderr, "Skipping DRM-backend tests because " \
|
|
|
|
"WESTON_TEST_SUITE_DRM_DEVICE is not set. " \
|
|
|
|
"See test suite documentation to learn how " \
|
|
|
|
"to run them.\n");
|
2021-10-21 16:44:55 +03:00
|
|
|
ret = RESULT_SKIP;
|
|
|
|
goto out;
|
2020-05-06 17:21:03 +03:00
|
|
|
}
|
2022-03-08 16:30:04 +03:00
|
|
|
str_printf(&tmp, "--drm-device=%s", drm_device);
|
2020-05-06 17:21:03 +03:00
|
|
|
prog_args_take(&args, tmp);
|
|
|
|
|
|
|
|
prog_args_take(&args, strdup("--seat=weston-test-seat"));
|
|
|
|
|
|
|
|
prog_args_take(&args, strdup("--continue-without-input"));
|
|
|
|
|
|
|
|
lock_fd = wait_for_lock();
|
2021-10-21 16:44:55 +03:00
|
|
|
if (lock_fd == -1) {
|
|
|
|
ret = RESULT_FAIL;
|
|
|
|
goto out;
|
|
|
|
}
|
2020-05-06 17:21:03 +03:00
|
|
|
}
|
|
|
|
|
2020-10-17 21:03:09 +03:00
|
|
|
/* Test suite needs the debug protocol to be able to take screenshots */
|
|
|
|
prog_args_take(&args, strdup("--debug"));
|
|
|
|
|
2022-03-08 16:30:04 +03:00
|
|
|
str_printf(&tmp, "--socket=%s", setup->testset_name);
|
tests: thread-based client harness
This replaces the old test harness with a new one.
The old harness relied on fork()'ing each test which makes tests independent,
but makes debugging them harder. The new harness runs client code in a thread
instead of a new process. A side-effect of not fork()'ing anymore is that any
failure will stop running a test series short. Fortunately we do not have any
tests that are expected to crash or fail.
The old harness executed 'weston' from Meson, with lots of setup as both
command line options and environment variables. The new harness executes
wet_main() instead: the test program itself calls the compositor main function
to execute the compositor in-process. Command line arguments are configured in
the test program itself, not in meson.build. Environment variables aside, you
are able to run a test by simply executing the test program, even if it is a
plugin test.
The new harness adds a new type of iteration: fixtures. For now, fixtures are
used to set up the compositor for tests that need a compositor. If necessary, a
fixture setup may include a data array of arbitrary type for executing the test
series for each element in the array. This will be most useful for running
screenshooting tests with both Pixman- and GL-renderers.
The new harness outputs TAP formatted results into stdout. Meson is not
switched to consume TAP yet though, because it would require a Meson version
requirement bump and would not have any benefits at this time. OTOH outputting
TAP is trivial and sets up a clear precedent of random test chatter belonging
to stderr.
This commit migrates only few tests to actually make use of the new features:
roles is a basic client test, subsurface-shot is a client test that
demonstrates the fixture array, and plugin-registry is a plugin test. The rest
of the tests will be migrated later.
Once all tests are migrated, we can remove the test-specific setup from
meson.build, leaving only the actual build instructions in there.
The not migrated tests and stand-alone tests suffer only a minor change: they
no longer fork() for each TEST(), otherwise they keep running as before.
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.com>
2019-11-01 15:02:15 +03:00
|
|
|
prog_args_take(&args, tmp);
|
|
|
|
|
2022-03-08 16:30:04 +03:00
|
|
|
str_printf(&tmp, "--modules=%s%s%s", TESTSUITE_PLUGIN_PATH,
|
|
|
|
setup->extra_module ? "," : "",
|
|
|
|
setup->extra_module ? setup->extra_module : "");
|
tests: thread-based client harness
This replaces the old test harness with a new one.
The old harness relied on fork()'ing each test which makes tests independent,
but makes debugging them harder. The new harness runs client code in a thread
instead of a new process. A side-effect of not fork()'ing anymore is that any
failure will stop running a test series short. Fortunately we do not have any
tests that are expected to crash or fail.
The old harness executed 'weston' from Meson, with lots of setup as both
command line options and environment variables. The new harness executes
wet_main() instead: the test program itself calls the compositor main function
to execute the compositor in-process. Command line arguments are configured in
the test program itself, not in meson.build. Environment variables aside, you
are able to run a test by simply executing the test program, even if it is a
plugin test.
The new harness adds a new type of iteration: fixtures. For now, fixtures are
used to set up the compositor for tests that need a compositor. If necessary, a
fixture setup may include a data array of arbitrary type for executing the test
series for each element in the array. This will be most useful for running
screenshooting tests with both Pixman- and GL-renderers.
The new harness outputs TAP formatted results into stdout. Meson is not
switched to consume TAP yet though, because it would require a Meson version
requirement bump and would not have any benefits at this time. OTOH outputting
TAP is trivial and sets up a clear precedent of random test chatter belonging
to stderr.
This commit migrates only few tests to actually make use of the new features:
roles is a basic client test, subsurface-shot is a client test that
demonstrates the fixture array, and plugin-registry is a plugin test. The rest
of the tests will be migrated later.
Once all tests are migrated, we can remove the test-specific setup from
meson.build, leaving only the actual build instructions in there.
The not migrated tests and stand-alone tests suffer only a minor change: they
no longer fork() for each TEST(), otherwise they keep running as before.
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.com>
2019-11-01 15:02:15 +03:00
|
|
|
prog_args_take(&args, tmp);
|
|
|
|
|
libweston: remove fbdev backend
Fbdev backend was deprecated in the Weston 10.0.0 release with
6338dbd5816689b2f08f48b359a972e16ff038d8. Before that, I suggested
already in 2019 to remove it, but it was too soon then. Now it seems the
final voices asking for fbdev to be kept have been satisfied, see the
linked issue.
Fbdev-backend uses a kernel graphics UAPI (fbdev) which is sub-par for a
Wayland compositor: you cannot do GPU accelerated graphics in any
reasonable way, no hotplug support, multi-output support is tedious, and
so on. Most importantly, Linux has deprecated fbdev a long time ago due
to the UAPI fitting modern systems and use cases very poorly, but cannot
get rid of it if any users remain. Let's do here what we can to reduce
fbdev usage.
I am doing color management related additions to libweston which require
adding checks to every backend. One backend less is less churn to write
and review.
Libweston major version has already been bumped to 11, so the next
release will be Weston 11, without fbdev. enum weston_compositor_backend
entries change their numerical values.
Fixes: https://gitlab.freedesktop.org/wayland/weston/-/issues/581
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.com>
2022-03-15 12:17:56 +03:00
|
|
|
if (setup->backend != WESTON_BACKEND_DRM) {
|
2022-03-08 16:30:04 +03:00
|
|
|
str_printf(&tmp, "--width=%d", setup->width);
|
2020-05-05 22:09:03 +03:00
|
|
|
prog_args_take(&args, tmp);
|
tests: thread-based client harness
This replaces the old test harness with a new one.
The old harness relied on fork()'ing each test which makes tests independent,
but makes debugging them harder. The new harness runs client code in a thread
instead of a new process. A side-effect of not fork()'ing anymore is that any
failure will stop running a test series short. Fortunately we do not have any
tests that are expected to crash or fail.
The old harness executed 'weston' from Meson, with lots of setup as both
command line options and environment variables. The new harness executes
wet_main() instead: the test program itself calls the compositor main function
to execute the compositor in-process. Command line arguments are configured in
the test program itself, not in meson.build. Environment variables aside, you
are able to run a test by simply executing the test program, even if it is a
plugin test.
The new harness adds a new type of iteration: fixtures. For now, fixtures are
used to set up the compositor for tests that need a compositor. If necessary, a
fixture setup may include a data array of arbitrary type for executing the test
series for each element in the array. This will be most useful for running
screenshooting tests with both Pixman- and GL-renderers.
The new harness outputs TAP formatted results into stdout. Meson is not
switched to consume TAP yet though, because it would require a Meson version
requirement bump and would not have any benefits at this time. OTOH outputting
TAP is trivial and sets up a clear precedent of random test chatter belonging
to stderr.
This commit migrates only few tests to actually make use of the new features:
roles is a basic client test, subsurface-shot is a client test that
demonstrates the fixture array, and plugin-registry is a plugin test. The rest
of the tests will be migrated later.
Once all tests are migrated, we can remove the test-specific setup from
meson.build, leaving only the actual build instructions in there.
The not migrated tests and stand-alone tests suffer only a minor change: they
no longer fork() for each TEST(), otherwise they keep running as before.
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.com>
2019-11-01 15:02:15 +03:00
|
|
|
|
2022-03-08 16:30:04 +03:00
|
|
|
str_printf(&tmp, "--height=%d", setup->height);
|
2020-05-05 22:09:03 +03:00
|
|
|
prog_args_take(&args, tmp);
|
|
|
|
}
|
tests: thread-based client harness
This replaces the old test harness with a new one.
The old harness relied on fork()'ing each test which makes tests independent,
but makes debugging them harder. The new harness runs client code in a thread
instead of a new process. A side-effect of not fork()'ing anymore is that any
failure will stop running a test series short. Fortunately we do not have any
tests that are expected to crash or fail.
The old harness executed 'weston' from Meson, with lots of setup as both
command line options and environment variables. The new harness executes
wet_main() instead: the test program itself calls the compositor main function
to execute the compositor in-process. Command line arguments are configured in
the test program itself, not in meson.build. Environment variables aside, you
are able to run a test by simply executing the test program, even if it is a
plugin test.
The new harness adds a new type of iteration: fixtures. For now, fixtures are
used to set up the compositor for tests that need a compositor. If necessary, a
fixture setup may include a data array of arbitrary type for executing the test
series for each element in the array. This will be most useful for running
screenshooting tests with both Pixman- and GL-renderers.
The new harness outputs TAP formatted results into stdout. Meson is not
switched to consume TAP yet though, because it would require a Meson version
requirement bump and would not have any benefits at this time. OTOH outputting
TAP is trivial and sets up a clear precedent of random test chatter belonging
to stderr.
This commit migrates only few tests to actually make use of the new features:
roles is a basic client test, subsurface-shot is a client test that
demonstrates the fixture array, and plugin-registry is a plugin test. The rest
of the tests will be migrated later.
Once all tests are migrated, we can remove the test-specific setup from
meson.build, leaving only the actual build instructions in there.
The not migrated tests and stand-alone tests suffer only a minor change: they
no longer fork() for each TEST(), otherwise they keep running as before.
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.com>
2019-11-01 15:02:15 +03:00
|
|
|
|
2020-01-22 12:53:12 +03:00
|
|
|
if (setup->scale != 1) {
|
2022-03-08 16:30:04 +03:00
|
|
|
str_printf(&tmp, "--scale=%d", setup->scale);
|
2020-01-22 12:53:12 +03:00
|
|
|
prog_args_take(&args, tmp);
|
|
|
|
}
|
|
|
|
|
|
|
|
if (setup->transform != WL_OUTPUT_TRANSFORM_NORMAL) {
|
2022-03-08 16:30:04 +03:00
|
|
|
str_printf(&tmp, "--transform=%s",
|
|
|
|
transform_to_str(setup->transform));
|
2020-01-22 12:53:12 +03:00
|
|
|
prog_args_take(&args, tmp);
|
|
|
|
}
|
|
|
|
|
tests: thread-based client harness
This replaces the old test harness with a new one.
The old harness relied on fork()'ing each test which makes tests independent,
but makes debugging them harder. The new harness runs client code in a thread
instead of a new process. A side-effect of not fork()'ing anymore is that any
failure will stop running a test series short. Fortunately we do not have any
tests that are expected to crash or fail.
The old harness executed 'weston' from Meson, with lots of setup as both
command line options and environment variables. The new harness executes
wet_main() instead: the test program itself calls the compositor main function
to execute the compositor in-process. Command line arguments are configured in
the test program itself, not in meson.build. Environment variables aside, you
are able to run a test by simply executing the test program, even if it is a
plugin test.
The new harness adds a new type of iteration: fixtures. For now, fixtures are
used to set up the compositor for tests that need a compositor. If necessary, a
fixture setup may include a data array of arbitrary type for executing the test
series for each element in the array. This will be most useful for running
screenshooting tests with both Pixman- and GL-renderers.
The new harness outputs TAP formatted results into stdout. Meson is not
switched to consume TAP yet though, because it would require a Meson version
requirement bump and would not have any benefits at this time. OTOH outputting
TAP is trivial and sets up a clear precedent of random test chatter belonging
to stderr.
This commit migrates only few tests to actually make use of the new features:
roles is a basic client test, subsurface-shot is a client test that
demonstrates the fixture array, and plugin-registry is a plugin test. The rest
of the tests will be migrated later.
Once all tests are migrated, we can remove the test-specific setup from
meson.build, leaving only the actual build instructions in there.
The not migrated tests and stand-alone tests suffer only a minor change: they
no longer fork() for each TEST(), otherwise they keep running as before.
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.com>
2019-11-01 15:02:15 +03:00
|
|
|
if (setup->config_file) {
|
2022-03-08 16:30:04 +03:00
|
|
|
str_printf(&tmp, "--config=%s", setup->config_file);
|
tests: thread-based client harness
This replaces the old test harness with a new one.
The old harness relied on fork()'ing each test which makes tests independent,
but makes debugging them harder. The new harness runs client code in a thread
instead of a new process. A side-effect of not fork()'ing anymore is that any
failure will stop running a test series short. Fortunately we do not have any
tests that are expected to crash or fail.
The old harness executed 'weston' from Meson, with lots of setup as both
command line options and environment variables. The new harness executes
wet_main() instead: the test program itself calls the compositor main function
to execute the compositor in-process. Command line arguments are configured in
the test program itself, not in meson.build. Environment variables aside, you
are able to run a test by simply executing the test program, even if it is a
plugin test.
The new harness adds a new type of iteration: fixtures. For now, fixtures are
used to set up the compositor for tests that need a compositor. If necessary, a
fixture setup may include a data array of arbitrary type for executing the test
series for each element in the array. This will be most useful for running
screenshooting tests with both Pixman- and GL-renderers.
The new harness outputs TAP formatted results into stdout. Meson is not
switched to consume TAP yet though, because it would require a Meson version
requirement bump and would not have any benefits at this time. OTOH outputting
TAP is trivial and sets up a clear precedent of random test chatter belonging
to stderr.
This commit migrates only few tests to actually make use of the new features:
roles is a basic client test, subsurface-shot is a client test that
demonstrates the fixture array, and plugin-registry is a plugin test. The rest
of the tests will be migrated later.
Once all tests are migrated, we can remove the test-specific setup from
meson.build, leaving only the actual build instructions in there.
The not migrated tests and stand-alone tests suffer only a minor change: they
no longer fork() for each TEST(), otherwise they keep running as before.
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.com>
2019-11-01 15:02:15 +03:00
|
|
|
prog_args_take(&args, tmp);
|
2020-08-16 03:19:13 +03:00
|
|
|
free(setup->config_file);
|
tests: thread-based client harness
This replaces the old test harness with a new one.
The old harness relied on fork()'ing each test which makes tests independent,
but makes debugging them harder. The new harness runs client code in a thread
instead of a new process. A side-effect of not fork()'ing anymore is that any
failure will stop running a test series short. Fortunately we do not have any
tests that are expected to crash or fail.
The old harness executed 'weston' from Meson, with lots of setup as both
command line options and environment variables. The new harness executes
wet_main() instead: the test program itself calls the compositor main function
to execute the compositor in-process. Command line arguments are configured in
the test program itself, not in meson.build. Environment variables aside, you
are able to run a test by simply executing the test program, even if it is a
plugin test.
The new harness adds a new type of iteration: fixtures. For now, fixtures are
used to set up the compositor for tests that need a compositor. If necessary, a
fixture setup may include a data array of arbitrary type for executing the test
series for each element in the array. This will be most useful for running
screenshooting tests with both Pixman- and GL-renderers.
The new harness outputs TAP formatted results into stdout. Meson is not
switched to consume TAP yet though, because it would require a Meson version
requirement bump and would not have any benefits at this time. OTOH outputting
TAP is trivial and sets up a clear precedent of random test chatter belonging
to stderr.
This commit migrates only few tests to actually make use of the new features:
roles is a basic client test, subsurface-shot is a client test that
demonstrates the fixture array, and plugin-registry is a plugin test. The rest
of the tests will be migrated later.
Once all tests are migrated, we can remove the test-specific setup from
meson.build, leaving only the actual build instructions in there.
The not migrated tests and stand-alone tests suffer only a minor change: they
no longer fork() for each TEST(), otherwise they keep running as before.
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.com>
2019-11-01 15:02:15 +03:00
|
|
|
} else {
|
|
|
|
prog_args_take(&args, strdup("--no-config"));
|
|
|
|
}
|
|
|
|
|
2022-12-30 00:12:24 +03:00
|
|
|
str_printf(&tmp, "--renderer=%s", renderer_to_str(setup->renderer));
|
|
|
|
prog_args_take(&args, strdup(tmp));
|
tests: thread-based client harness
This replaces the old test harness with a new one.
The old harness relied on fork()'ing each test which makes tests independent,
but makes debugging them harder. The new harness runs client code in a thread
instead of a new process. A side-effect of not fork()'ing anymore is that any
failure will stop running a test series short. Fortunately we do not have any
tests that are expected to crash or fail.
The old harness executed 'weston' from Meson, with lots of setup as both
command line options and environment variables. The new harness executes
wet_main() instead: the test program itself calls the compositor main function
to execute the compositor in-process. Command line arguments are configured in
the test program itself, not in meson.build. Environment variables aside, you
are able to run a test by simply executing the test program, even if it is a
plugin test.
The new harness adds a new type of iteration: fixtures. For now, fixtures are
used to set up the compositor for tests that need a compositor. If necessary, a
fixture setup may include a data array of arbitrary type for executing the test
series for each element in the array. This will be most useful for running
screenshooting tests with both Pixman- and GL-renderers.
The new harness outputs TAP formatted results into stdout. Meson is not
switched to consume TAP yet though, because it would require a Meson version
requirement bump and would not have any benefits at this time. OTOH outputting
TAP is trivial and sets up a clear precedent of random test chatter belonging
to stderr.
This commit migrates only few tests to actually make use of the new features:
roles is a basic client test, subsurface-shot is a client test that
demonstrates the fixture array, and plugin-registry is a plugin test. The rest
of the tests will be migrated later.
Once all tests are migrated, we can remove the test-specific setup from
meson.build, leaving only the actual build instructions in there.
The not migrated tests and stand-alone tests suffer only a minor change: they
no longer fork() for each TEST(), otherwise they keep running as before.
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.com>
2019-11-01 15:02:15 +03:00
|
|
|
|
2022-03-08 16:30:04 +03:00
|
|
|
str_printf(&tmp, "--shell=%s", shell_to_str(setup->shell));
|
tests: thread-based client harness
This replaces the old test harness with a new one.
The old harness relied on fork()'ing each test which makes tests independent,
but makes debugging them harder. The new harness runs client code in a thread
instead of a new process. A side-effect of not fork()'ing anymore is that any
failure will stop running a test series short. Fortunately we do not have any
tests that are expected to crash or fail.
The old harness executed 'weston' from Meson, with lots of setup as both
command line options and environment variables. The new harness executes
wet_main() instead: the test program itself calls the compositor main function
to execute the compositor in-process. Command line arguments are configured in
the test program itself, not in meson.build. Environment variables aside, you
are able to run a test by simply executing the test program, even if it is a
plugin test.
The new harness adds a new type of iteration: fixtures. For now, fixtures are
used to set up the compositor for tests that need a compositor. If necessary, a
fixture setup may include a data array of arbitrary type for executing the test
series for each element in the array. This will be most useful for running
screenshooting tests with both Pixman- and GL-renderers.
The new harness outputs TAP formatted results into stdout. Meson is not
switched to consume TAP yet though, because it would require a Meson version
requirement bump and would not have any benefits at this time. OTOH outputting
TAP is trivial and sets up a clear precedent of random test chatter belonging
to stderr.
This commit migrates only few tests to actually make use of the new features:
roles is a basic client test, subsurface-shot is a client test that
demonstrates the fixture array, and plugin-registry is a plugin test. The rest
of the tests will be migrated later.
Once all tests are migrated, we can remove the test-specific setup from
meson.build, leaving only the actual build instructions in there.
The not migrated tests and stand-alone tests suffer only a minor change: they
no longer fork() for each TEST(), otherwise they keep running as before.
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.com>
2019-11-01 15:02:15 +03:00
|
|
|
prog_args_take(&args, tmp);
|
|
|
|
|
|
|
|
if (setup->logging_scopes) {
|
2022-03-08 16:30:04 +03:00
|
|
|
str_printf(&tmp, "--logger-scopes=%s", setup->logging_scopes);
|
tests: thread-based client harness
This replaces the old test harness with a new one.
The old harness relied on fork()'ing each test which makes tests independent,
but makes debugging them harder. The new harness runs client code in a thread
instead of a new process. A side-effect of not fork()'ing anymore is that any
failure will stop running a test series short. Fortunately we do not have any
tests that are expected to crash or fail.
The old harness executed 'weston' from Meson, with lots of setup as both
command line options and environment variables. The new harness executes
wet_main() instead: the test program itself calls the compositor main function
to execute the compositor in-process. Command line arguments are configured in
the test program itself, not in meson.build. Environment variables aside, you
are able to run a test by simply executing the test program, even if it is a
plugin test.
The new harness adds a new type of iteration: fixtures. For now, fixtures are
used to set up the compositor for tests that need a compositor. If necessary, a
fixture setup may include a data array of arbitrary type for executing the test
series for each element in the array. This will be most useful for running
screenshooting tests with both Pixman- and GL-renderers.
The new harness outputs TAP formatted results into stdout. Meson is not
switched to consume TAP yet though, because it would require a Meson version
requirement bump and would not have any benefits at this time. OTOH outputting
TAP is trivial and sets up a clear precedent of random test chatter belonging
to stderr.
This commit migrates only few tests to actually make use of the new features:
roles is a basic client test, subsurface-shot is a client test that
demonstrates the fixture array, and plugin-registry is a plugin test. The rest
of the tests will be migrated later.
Once all tests are migrated, we can remove the test-specific setup from
meson.build, leaving only the actual build instructions in there.
The not migrated tests and stand-alone tests suffer only a minor change: they
no longer fork() for each TEST(), otherwise they keep running as before.
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.com>
2019-11-01 15:02:15 +03:00
|
|
|
prog_args_take(&args, tmp);
|
|
|
|
}
|
|
|
|
|
|
|
|
if (setup->xwayland)
|
|
|
|
prog_args_take(&args, strdup("--xwayland"));
|
|
|
|
|
2020-10-19 22:06:22 +03:00
|
|
|
test_data.test_quirks = setup->test_quirks;
|
|
|
|
test_data.test_private_data = data;
|
tests: thread-based client harness
This replaces the old test harness with a new one.
The old harness relied on fork()'ing each test which makes tests independent,
but makes debugging them harder. The new harness runs client code in a thread
instead of a new process. A side-effect of not fork()'ing anymore is that any
failure will stop running a test series short. Fortunately we do not have any
tests that are expected to crash or fail.
The old harness executed 'weston' from Meson, with lots of setup as both
command line options and environment variables. The new harness executes
wet_main() instead: the test program itself calls the compositor main function
to execute the compositor in-process. Command line arguments are configured in
the test program itself, not in meson.build. Environment variables aside, you
are able to run a test by simply executing the test program, even if it is a
plugin test.
The new harness adds a new type of iteration: fixtures. For now, fixtures are
used to set up the compositor for tests that need a compositor. If necessary, a
fixture setup may include a data array of arbitrary type for executing the test
series for each element in the array. This will be most useful for running
screenshooting tests with both Pixman- and GL-renderers.
The new harness outputs TAP formatted results into stdout. Meson is not
switched to consume TAP yet though, because it would require a Meson version
requirement bump and would not have any benefits at this time. OTOH outputting
TAP is trivial and sets up a clear precedent of random test chatter belonging
to stderr.
This commit migrates only few tests to actually make use of the new features:
roles is a basic client test, subsurface-shot is a client test that
demonstrates the fixture array, and plugin-registry is a plugin test. The rest
of the tests will be migrated later.
Once all tests are migrated, we can remove the test-specific setup from
meson.build, leaving only the actual build instructions in there.
The not migrated tests and stand-alone tests suffer only a minor change: they
no longer fork() for each TEST(), otherwise they keep running as before.
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.com>
2019-11-01 15:02:15 +03:00
|
|
|
prog_args_save(&args);
|
2022-06-25 05:56:45 +03:00
|
|
|
|
|
|
|
if (setenv("WESTON_MODULE_MAP", WESTON_MODULE_MAP, 0) < 0 ||
|
|
|
|
setenv("WESTON_DATA_DIR", WESTON_DATA_DIR, 0) < 0) {
|
|
|
|
fprintf(stderr, "Error: environment setup failed.\n");
|
|
|
|
ret = RESULT_HARD_ERROR;
|
|
|
|
}
|
|
|
|
|
|
|
|
#ifndef BUILD_DRM_COMPOSITOR
|
|
|
|
if (setup->backend == WESTON_BACKEND_DRM) {
|
|
|
|
fprintf(stderr, "DRM-backend required but not built, skipping.\n");
|
|
|
|
ret = RESULT_SKIP;
|
|
|
|
}
|
|
|
|
#endif
|
|
|
|
|
|
|
|
#ifndef BUILD_RDP_COMPOSITOR
|
|
|
|
if (setup->backend == WESTON_BACKEND_RDP) {
|
|
|
|
fprintf(stderr, "RDP-backend required but not built, skipping.\n");
|
|
|
|
ret = RESULT_SKIP;
|
|
|
|
}
|
|
|
|
#endif
|
|
|
|
|
|
|
|
#ifndef BUILD_WAYLAND_COMPOSITOR
|
|
|
|
if (setup->backend == WESTON_BACKEND_WAYLAND) {
|
|
|
|
fprintf(stderr, "wayland-backend required but not built, skipping.\n");
|
|
|
|
ret = RESULT_SKIP;
|
|
|
|
}
|
|
|
|
#endif
|
|
|
|
|
|
|
|
#ifndef BUILD_X11_COMPOSITOR
|
|
|
|
if (setup->backend == WESTON_BACKEND_X11) {
|
|
|
|
fprintf(stderr, "X11-backend required but not built, skipping.\n");
|
|
|
|
ret = RESULT_SKIP;
|
|
|
|
}
|
|
|
|
#endif
|
|
|
|
|
|
|
|
#ifndef ENABLE_EGL
|
2022-12-05 19:05:21 +03:00
|
|
|
if (setup->renderer == WESTON_RENDERER_GL) {
|
2022-06-25 05:56:45 +03:00
|
|
|
fprintf(stderr, "GL-renderer required but not built, skipping.\n");
|
|
|
|
ret = RESULT_SKIP;
|
|
|
|
}
|
|
|
|
#endif
|
|
|
|
|
|
|
|
if (ret == RESULT_OK)
|
|
|
|
ret = wet_main(args.argc, args.argv, &test_data);
|
tests: thread-based client harness
This replaces the old test harness with a new one.
The old harness relied on fork()'ing each test which makes tests independent,
but makes debugging them harder. The new harness runs client code in a thread
instead of a new process. A side-effect of not fork()'ing anymore is that any
failure will stop running a test series short. Fortunately we do not have any
tests that are expected to crash or fail.
The old harness executed 'weston' from Meson, with lots of setup as both
command line options and environment variables. The new harness executes
wet_main() instead: the test program itself calls the compositor main function
to execute the compositor in-process. Command line arguments are configured in
the test program itself, not in meson.build. Environment variables aside, you
are able to run a test by simply executing the test program, even if it is a
plugin test.
The new harness adds a new type of iteration: fixtures. For now, fixtures are
used to set up the compositor for tests that need a compositor. If necessary, a
fixture setup may include a data array of arbitrary type for executing the test
series for each element in the array. This will be most useful for running
screenshooting tests with both Pixman- and GL-renderers.
The new harness outputs TAP formatted results into stdout. Meson is not
switched to consume TAP yet though, because it would require a Meson version
requirement bump and would not have any benefits at this time. OTOH outputting
TAP is trivial and sets up a clear precedent of random test chatter belonging
to stderr.
This commit migrates only few tests to actually make use of the new features:
roles is a basic client test, subsurface-shot is a client test that
demonstrates the fixture array, and plugin-registry is a plugin test. The rest
of the tests will be migrated later.
Once all tests are migrated, we can remove the test-specific setup from
meson.build, leaving only the actual build instructions in there.
The not migrated tests and stand-alone tests suffer only a minor change: they
no longer fork() for each TEST(), otherwise they keep running as before.
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.com>
2019-11-01 15:02:15 +03:00
|
|
|
|
2021-10-21 16:44:55 +03:00
|
|
|
out:
|
tests: thread-based client harness
This replaces the old test harness with a new one.
The old harness relied on fork()'ing each test which makes tests independent,
but makes debugging them harder. The new harness runs client code in a thread
instead of a new process. A side-effect of not fork()'ing anymore is that any
failure will stop running a test series short. Fortunately we do not have any
tests that are expected to crash or fail.
The old harness executed 'weston' from Meson, with lots of setup as both
command line options and environment variables. The new harness executes
wet_main() instead: the test program itself calls the compositor main function
to execute the compositor in-process. Command line arguments are configured in
the test program itself, not in meson.build. Environment variables aside, you
are able to run a test by simply executing the test program, even if it is a
plugin test.
The new harness adds a new type of iteration: fixtures. For now, fixtures are
used to set up the compositor for tests that need a compositor. If necessary, a
fixture setup may include a data array of arbitrary type for executing the test
series for each element in the array. This will be most useful for running
screenshooting tests with both Pixman- and GL-renderers.
The new harness outputs TAP formatted results into stdout. Meson is not
switched to consume TAP yet though, because it would require a Meson version
requirement bump and would not have any benefits at this time. OTOH outputting
TAP is trivial and sets up a clear precedent of random test chatter belonging
to stderr.
This commit migrates only few tests to actually make use of the new features:
roles is a basic client test, subsurface-shot is a client test that
demonstrates the fixture array, and plugin-registry is a plugin test. The rest
of the tests will be migrated later.
Once all tests are migrated, we can remove the test-specific setup from
meson.build, leaving only the actual build instructions in there.
The not migrated tests and stand-alone tests suffer only a minor change: they
no longer fork() for each TEST(), otherwise they keep running as before.
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.com>
2019-11-01 15:02:15 +03:00
|
|
|
prog_args_fini(&args);
|
|
|
|
|
2020-05-06 17:21:03 +03:00
|
|
|
/* We acquired a lock (if this is a DRM-backend test) and now we can
|
|
|
|
* close its fd and release it, as it has already been run. */
|
|
|
|
if (lock_fd != -1)
|
|
|
|
close(lock_fd);
|
|
|
|
|
tests: thread-based client harness
This replaces the old test harness with a new one.
The old harness relied on fork()'ing each test which makes tests independent,
but makes debugging them harder. The new harness runs client code in a thread
instead of a new process. A side-effect of not fork()'ing anymore is that any
failure will stop running a test series short. Fortunately we do not have any
tests that are expected to crash or fail.
The old harness executed 'weston' from Meson, with lots of setup as both
command line options and environment variables. The new harness executes
wet_main() instead: the test program itself calls the compositor main function
to execute the compositor in-process. Command line arguments are configured in
the test program itself, not in meson.build. Environment variables aside, you
are able to run a test by simply executing the test program, even if it is a
plugin test.
The new harness adds a new type of iteration: fixtures. For now, fixtures are
used to set up the compositor for tests that need a compositor. If necessary, a
fixture setup may include a data array of arbitrary type for executing the test
series for each element in the array. This will be most useful for running
screenshooting tests with both Pixman- and GL-renderers.
The new harness outputs TAP formatted results into stdout. Meson is not
switched to consume TAP yet though, because it would require a Meson version
requirement bump and would not have any benefits at this time. OTOH outputting
TAP is trivial and sets up a clear precedent of random test chatter belonging
to stderr.
This commit migrates only few tests to actually make use of the new features:
roles is a basic client test, subsurface-shot is a client test that
demonstrates the fixture array, and plugin-registry is a plugin test. The rest
of the tests will be migrated later.
Once all tests are migrated, we can remove the test-specific setup from
meson.build, leaving only the actual build instructions in there.
The not migrated tests and stand-alone tests suffer only a minor change: they
no longer fork() for each TEST(), otherwise they keep running as before.
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.com>
2019-11-01 15:02:15 +03:00
|
|
|
return ret;
|
|
|
|
}
|
2020-08-16 03:19:13 +03:00
|
|
|
|
|
|
|
static void
|
|
|
|
write_cfg(va_list entry_list, FILE *weston_ini)
|
|
|
|
{
|
|
|
|
char *entry = va_arg(entry_list, char *);
|
|
|
|
int ret;
|
|
|
|
assert(entry);
|
|
|
|
|
|
|
|
while (entry) {
|
|
|
|
ret = fprintf(weston_ini, "%s\n", entry);
|
|
|
|
assert(ret >= 0);
|
|
|
|
free(entry);
|
|
|
|
entry = va_arg(entry_list, char *);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
static FILE *
|
|
|
|
open_ini_file(struct compositor_setup *setup)
|
|
|
|
{
|
|
|
|
char *wd, *tmp_path = NULL;
|
|
|
|
FILE *weston_ini = NULL;
|
|
|
|
|
|
|
|
assert(!setup->config_file);
|
|
|
|
|
|
|
|
wd = realpath(".", NULL);
|
|
|
|
assert(wd);
|
|
|
|
|
2022-03-08 16:30:04 +03:00
|
|
|
str_printf(&tmp_path, "%s/%s.ini", wd, setup->testset_name);
|
|
|
|
if (!tmp_path) {
|
2020-12-18 01:07:49 +03:00
|
|
|
fprintf(stderr, "Fail formatting Weston.ini file name.\n");
|
2020-08-16 03:19:13 +03:00
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
|
|
|
|
weston_ini = fopen(tmp_path, "w");
|
|
|
|
assert(weston_ini);
|
|
|
|
setup->config_file = tmp_path;
|
|
|
|
|
|
|
|
out:
|
|
|
|
free(wd);
|
|
|
|
return weston_ini;
|
|
|
|
}
|
|
|
|
|
|
|
|
void
|
|
|
|
weston_ini_setup_(struct compositor_setup *setup, ...)
|
|
|
|
{
|
|
|
|
FILE *weston_ini = NULL;
|
|
|
|
int ret;
|
|
|
|
va_list entry_list;
|
|
|
|
|
|
|
|
weston_ini = open_ini_file(setup);
|
|
|
|
assert(weston_ini);
|
|
|
|
|
|
|
|
va_start(entry_list, setup);
|
|
|
|
write_cfg(entry_list, weston_ini);
|
|
|
|
va_end(entry_list);
|
|
|
|
|
|
|
|
ret = fclose(weston_ini);
|
|
|
|
assert(ret != EOF);
|
|
|
|
}
|
|
|
|
|
|
|
|
char *
|
|
|
|
cfgln(const char *fmt, ...)
|
|
|
|
{
|
|
|
|
char *str;
|
|
|
|
int ret;
|
|
|
|
va_list ap;
|
|
|
|
|
|
|
|
va_start(ap, fmt);
|
|
|
|
ret = vasprintf(&str, fmt, ap);
|
|
|
|
assert(ret >= 0);
|
|
|
|
va_end(ap);
|
|
|
|
|
|
|
|
return str;
|
|
|
|
}
|
|
|
|
|