You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
weston/tests/ivi_layout-test-plugin.c

373 lines
11 KiB

tests: ivi_layout test infrastructure Testing the ivi_layout API requires two things: - the tests must be written as a controller module to access the API - the tests need a helper client to create some objects that can then be managed via the API This patch adds all the infrastructure and two different kinds of example tests. Internal ivi-shell (ivi_layout) API tests are listed as ivi-*.la files in TESTS in Makefile.am. Weston-tests-env detects these, and runs Weston with ivi-shell, and loads the given module as a controller module, not as a normal plugin. The test controller module ivi-*.la will launch a helper client. For ivi-layout-test.la the helper client is ivi-layout.ivi. The helper client uses the weston-test-runner framework to fork and exec each TEST with a fresh connection to the compositor. The actual test is triggered by the weston_test_runner protocol interface, a new addition to weston-test.xml. The helper client uses weston_test_runner to trigger a test, and the server side of the interface is implemented by the test controller module (ivi-layout-test.la). The server side of weston_test_runner uses the same trick as weston-test-runner.h to gather a list of defined tests. A test is defined with the RUNNER_TEST macro. If a test defined by RUNNER_TEST succeeds, an event is sent to the helper client that it can continue (or exit). If a test fails, a fatal protocol error is sent to the helper client. Once the helper client has iterated over all of its tests, it signals the batch success/failure via process exit code. That is cought in the test controller module, and forwarded as Weston's exit code. In summary: each ivi_layout test is a combination of a client side helper/setup and server side actual tests. v2: Load weston-test.so, because create_client() needs it. v3: add a comment about IVI_TEST_SURFACE_ID_BASE. v4: Rebased to upstream weston-tests-env changes. Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk> Reviewed-by: Derek Foreman <derekf@osg.samsung.com> (v2)
10 years ago
/*
* Copyright © 2012 Intel Corporation
* Copyright © 2013 DENSO CORPORATION
* Copyright © 2015 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:
tests: ivi_layout test infrastructure Testing the ivi_layout API requires two things: - the tests must be written as a controller module to access the API - the tests need a helper client to create some objects that can then be managed via the API This patch adds all the infrastructure and two different kinds of example tests. Internal ivi-shell (ivi_layout) API tests are listed as ivi-*.la files in TESTS in Makefile.am. Weston-tests-env detects these, and runs Weston with ivi-shell, and loads the given module as a controller module, not as a normal plugin. The test controller module ivi-*.la will launch a helper client. For ivi-layout-test.la the helper client is ivi-layout.ivi. The helper client uses the weston-test-runner framework to fork and exec each TEST with a fresh connection to the compositor. The actual test is triggered by the weston_test_runner protocol interface, a new addition to weston-test.xml. The helper client uses weston_test_runner to trigger a test, and the server side of the interface is implemented by the test controller module (ivi-layout-test.la). The server side of weston_test_runner uses the same trick as weston-test-runner.h to gather a list of defined tests. A test is defined with the RUNNER_TEST macro. If a test defined by RUNNER_TEST succeeds, an event is sent to the helper client that it can continue (or exit). If a test fails, a fatal protocol error is sent to the helper client. Once the helper client has iterated over all of its tests, it signals the batch success/failure via process exit code. That is cought in the test controller module, and forwarded as Weston's exit code. In summary: each ivi_layout test is a combination of a client side helper/setup and server side actual tests. v2: Load weston-test.so, because create_client() needs it. v3: add a comment about IVI_TEST_SURFACE_ID_BASE. v4: Rebased to upstream weston-tests-env changes. Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk> Reviewed-by: Derek Foreman <derekf@osg.samsung.com> (v2)
10 years ago
*
* 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.
tests: ivi_layout test infrastructure Testing the ivi_layout API requires two things: - the tests must be written as a controller module to access the API - the tests need a helper client to create some objects that can then be managed via the API This patch adds all the infrastructure and two different kinds of example tests. Internal ivi-shell (ivi_layout) API tests are listed as ivi-*.la files in TESTS in Makefile.am. Weston-tests-env detects these, and runs Weston with ivi-shell, and loads the given module as a controller module, not as a normal plugin. The test controller module ivi-*.la will launch a helper client. For ivi-layout-test.la the helper client is ivi-layout.ivi. The helper client uses the weston-test-runner framework to fork and exec each TEST with a fresh connection to the compositor. The actual test is triggered by the weston_test_runner protocol interface, a new addition to weston-test.xml. The helper client uses weston_test_runner to trigger a test, and the server side of the interface is implemented by the test controller module (ivi-layout-test.la). The server side of weston_test_runner uses the same trick as weston-test-runner.h to gather a list of defined tests. A test is defined with the RUNNER_TEST macro. If a test defined by RUNNER_TEST succeeds, an event is sent to the helper client that it can continue (or exit). If a test fails, a fatal protocol error is sent to the helper client. Once the helper client has iterated over all of its tests, it signals the batch success/failure via process exit code. That is cought in the test controller module, and forwarded as Weston's exit code. In summary: each ivi_layout test is a combination of a client side helper/setup and server side actual tests. v2: Load weston-test.so, because create_client() needs it. v3: add a comment about IVI_TEST_SURFACE_ID_BASE. v4: Rebased to upstream weston-tests-env changes. Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk> Reviewed-by: Derek Foreman <derekf@osg.samsung.com> (v2)
10 years ago
*/
#include "config.h"
#include <unistd.h>
#include <signal.h>
#include <string.h>
#include "src/compositor.h"
tests: ivi_layout test infrastructure Testing the ivi_layout API requires two things: - the tests must be written as a controller module to access the API - the tests need a helper client to create some objects that can then be managed via the API This patch adds all the infrastructure and two different kinds of example tests. Internal ivi-shell (ivi_layout) API tests are listed as ivi-*.la files in TESTS in Makefile.am. Weston-tests-env detects these, and runs Weston with ivi-shell, and loads the given module as a controller module, not as a normal plugin. The test controller module ivi-*.la will launch a helper client. For ivi-layout-test.la the helper client is ivi-layout.ivi. The helper client uses the weston-test-runner framework to fork and exec each TEST with a fresh connection to the compositor. The actual test is triggered by the weston_test_runner protocol interface, a new addition to weston-test.xml. The helper client uses weston_test_runner to trigger a test, and the server side of the interface is implemented by the test controller module (ivi-layout-test.la). The server side of weston_test_runner uses the same trick as weston-test-runner.h to gather a list of defined tests. A test is defined with the RUNNER_TEST macro. If a test defined by RUNNER_TEST succeeds, an event is sent to the helper client that it can continue (or exit). If a test fails, a fatal protocol error is sent to the helper client. Once the helper client has iterated over all of its tests, it signals the batch success/failure via process exit code. That is cought in the test controller module, and forwarded as Weston's exit code. In summary: each ivi_layout test is a combination of a client side helper/setup and server side actual tests. v2: Load weston-test.so, because create_client() needs it. v3: add a comment about IVI_TEST_SURFACE_ID_BASE. v4: Rebased to upstream weston-tests-env changes. Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk> Reviewed-by: Derek Foreman <derekf@osg.samsung.com> (v2)
10 years ago
#include "weston-test-server-protocol.h"
#include "ivi-test.h"
#include "ivi-shell/ivi-layout-export.h"
#include "shared/helpers.h"
tests: ivi_layout test infrastructure Testing the ivi_layout API requires two things: - the tests must be written as a controller module to access the API - the tests need a helper client to create some objects that can then be managed via the API This patch adds all the infrastructure and two different kinds of example tests. Internal ivi-shell (ivi_layout) API tests are listed as ivi-*.la files in TESTS in Makefile.am. Weston-tests-env detects these, and runs Weston with ivi-shell, and loads the given module as a controller module, not as a normal plugin. The test controller module ivi-*.la will launch a helper client. For ivi-layout-test.la the helper client is ivi-layout.ivi. The helper client uses the weston-test-runner framework to fork and exec each TEST with a fresh connection to the compositor. The actual test is triggered by the weston_test_runner protocol interface, a new addition to weston-test.xml. The helper client uses weston_test_runner to trigger a test, and the server side of the interface is implemented by the test controller module (ivi-layout-test.la). The server side of weston_test_runner uses the same trick as weston-test-runner.h to gather a list of defined tests. A test is defined with the RUNNER_TEST macro. If a test defined by RUNNER_TEST succeeds, an event is sent to the helper client that it can continue (or exit). If a test fails, a fatal protocol error is sent to the helper client. Once the helper client has iterated over all of its tests, it signals the batch success/failure via process exit code. That is cought in the test controller module, and forwarded as Weston's exit code. In summary: each ivi_layout test is a combination of a client side helper/setup and server side actual tests. v2: Load weston-test.so, because create_client() needs it. v3: add a comment about IVI_TEST_SURFACE_ID_BASE. v4: Rebased to upstream weston-tests-env changes. Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk> Reviewed-by: Derek Foreman <derekf@osg.samsung.com> (v2)
10 years ago
struct test_context;
struct runner_test {
const char *name;
void (*run)(struct test_context *);
} __attribute__ ((aligned (32)));
#define RUNNER_TEST(name) \
static void runner_func_##name(struct test_context *); \
\
const struct runner_test runner_test_##name \
__attribute__ ((section ("test_section"))) = \
{ \
#name, runner_func_##name \
}; \
\
static void runner_func_##name(struct test_context *ctx)
extern const struct runner_test __start_test_section;
extern const struct runner_test __stop_test_section;
static const struct runner_test *
find_runner_test(const char *name)
{
const struct runner_test *t;
for (t = &__start_test_section; t < &__stop_test_section; t++) {
if (strcmp(t->name, name) == 0)
return t;
}
return NULL;
}
struct test_launcher {
struct weston_compositor *compositor;
char exe[2048];
struct weston_process process;
const struct ivi_controller_interface *controller_interface;
};
static void
runner_destroy_handler(struct wl_client *client, struct wl_resource *resource)
{
wl_resource_destroy(resource);
}
struct test_context {
const struct ivi_controller_interface *controller_interface;
struct wl_resource *runner_resource;
};
static void
runner_run_handler(struct wl_client *client, struct wl_resource *resource,
const char *test_name)
{
struct test_launcher *launcher;
const struct runner_test *t;
struct test_context ctx;
launcher = wl_resource_get_user_data(resource);
ctx.controller_interface = launcher->controller_interface;
ctx.runner_resource = resource;
t = find_runner_test(test_name);
if (!t) {
weston_log("Error: runner test \"%s\" not found.\n",
test_name);
wl_resource_post_error(resource,
WESTON_TEST_RUNNER_ERROR_UNKNOWN_TEST,
"weston_test_runner: unknown: '%s'",
test_name);
return;
}
weston_log("weston_test_runner.run(\"%s\")\n", test_name);
t->run(&ctx);
weston_test_runner_send_finished(resource);
}
static const struct weston_test_runner_interface runner_implementation = {
runner_destroy_handler,
runner_run_handler
};
static void
bind_runner(struct wl_client *client, void *data,
uint32_t version, uint32_t id)
{
struct test_launcher *launcher = data;
struct wl_resource *resource;
resource = wl_resource_create(client, &weston_test_runner_interface,
1, id);
if (!resource) {
wl_client_post_no_memory(client);
return;
}
wl_resource_set_implementation(resource, &runner_implementation,
launcher, NULL);
}
static void
test_client_sigchld(struct weston_process *process, int status)
{
struct test_launcher *launcher =
container_of(process, struct test_launcher, process);
struct weston_compositor *c = launcher->compositor;
/* Chain up from weston-test-runner's exit code so that automake
* knows the exit status and can report e.g. skipped tests. */
if (WIFEXITED(status))
weston_compositor_exit_with_code(c, WEXITSTATUS(status));
else
weston_compositor_exit_with_code(c, EXIT_FAILURE);
}
static void
idle_launch_client(void *data)
{
struct test_launcher *launcher = data;
pid_t pid;
sigset_t allsigs;
pid = fork();
if (pid == -1) {
weston_log("fatal: failed to fork '%s': %m\n", launcher->exe);
weston_compositor_exit_with_code(launcher->compositor,
EXIT_FAILURE);
return;
}
if (pid == 0) {
sigfillset(&allsigs);
sigprocmask(SIG_UNBLOCK, &allsigs, NULL);
execl(launcher->exe, launcher->exe, NULL);
weston_log("compositor: executing '%s' failed: %m\n",
launcher->exe);
_exit(EXIT_FAILURE);
}
launcher->process.pid = pid;
launcher->process.cleanup = test_client_sigchld;
weston_watch_process(&launcher->process);
}
int
controller_module_init(struct weston_compositor *compositor,
int *argc, char *argv[],
const struct ivi_controller_interface *iface,
size_t iface_version);
WL_EXPORT int
controller_module_init(struct weston_compositor *compositor,
int *argc, char *argv[],
const struct ivi_controller_interface *iface,
size_t iface_version)
{
struct wl_event_loop *loop;
struct test_launcher *launcher;
const char *path;
/* strict check, since this is an internal test module */
if (iface_version != sizeof(*iface)) {
weston_log("fatal: controller interface mismatch\n");
return -1;
}
path = getenv("WESTON_BUILD_DIR");
if (!path) {
weston_log("test setup failure: WESTON_BUILD_DIR not set\n");
return -1;
}
launcher = zalloc(sizeof *launcher);
if (!launcher)
return -1;
launcher->compositor = compositor;
launcher->controller_interface = iface;
snprintf(launcher->exe, sizeof launcher->exe,
"%s/ivi-layout.ivi", path);
if (wl_global_create(compositor->wl_display,
&weston_test_runner_interface, 1,
launcher, bind_runner) == NULL)
return -1;
loop = wl_display_get_event_loop(compositor->wl_display);
wl_event_loop_add_idle(loop, idle_launch_client, launcher);
return 0;
}
static void
runner_assert_fail(const char *cond, const char *file, int line,
const char *func, struct test_context *ctx)
{
weston_log("Assert failure in %s:%d, %s: '%s'\n",
file, line, func, cond);
wl_resource_post_error(ctx->runner_resource,
WESTON_TEST_RUNNER_ERROR_TEST_FAILED,
"Assert failure in %s:%d, %s: '%s'\n",
file, line, func, cond);
}
#define runner_assert(cond) ({ \
bool b_ = (cond); \
if (!b_) \
runner_assert_fail(#cond, __FILE__, __LINE__, \
__func__, ctx); \
b_; \
})
#define runner_assert_or_return(cond) do { \
bool b_ = (cond); \
if (!b_) { \
runner_assert_fail(#cond, __FILE__, __LINE__, \
__func__, ctx); \
return; \
} \
} while (0)
/*************************** tests **********************************/
/*
* This is a controller module: a plugin to ivi-shell.so, i.e. a sub-plugin.
* This module is specially written to execute tests that target the
* ivi_layout API.
*
* This module is listed in TESTS in Makefile.am. weston-tests-env handles
* this module specially by loading it in ivi-shell.
*
* Once Weston init completes, this module launches one test program:
* ivi-layout.ivi (ivi_layout-test.c). That program uses the weston-test-runner
* framework to fork and exec each TEST() in ivi_layout-test.c with a fresh
* connection to the single compositor instance.
*
* Each TEST() in ivi_layout-test.c will bind to weston_test_runner global
* interface. A TEST() will set up the client state, and issue
* weston_test_runner.run request to execute the compositor-side of the test.
*
* The compositor-side parts of the tests are in this file. They are specified
* by RUNNER_TEST() macro, where the name argument matches the name string
* passed to weston_test_runner.run.
*
* A RUNNER_TEST() function simply returns when it succeeds. If it fails,
* a fatal protocol error is sent to the client from runner_assert() or
* runner_assert_or_return(). This module catches the test program exit
* code and passes it out of Weston to the test harness.
*
* A single TEST() in ivi_layout-test.c may use multiple RUNNER_TEST()s to
* achieve multiple test points over a client action sequence.
*/
RUNNER_TEST(surface_create_p1)
{
const struct ivi_controller_interface *ctl = ctx->controller_interface;
struct ivi_layout_surface *ivisurf[2];
uint32_t ivi_id;
ivisurf[0] = ctl->get_surface_from_id(IVI_TEST_SURFACE_ID(0));
runner_assert_or_return(ivisurf[0]);
ivisurf[1] = ctl->get_surface_from_id(IVI_TEST_SURFACE_ID(1));
runner_assert_or_return(ivisurf[1]);
ivi_id = ctl->get_id_of_surface(ivisurf[0]);
runner_assert_or_return(ivi_id == IVI_TEST_SURFACE_ID(0));
ivi_id = ctl->get_id_of_surface(ivisurf[1]);
runner_assert_or_return(ivi_id == IVI_TEST_SURFACE_ID(1));
}
RUNNER_TEST(surface_create_p2)
{
const struct ivi_controller_interface *ctl = ctx->controller_interface;
struct ivi_layout_surface *ivisurf;
/* the ivi_surface was destroyed by the client */
ivisurf = ctl->get_surface_from_id(IVI_TEST_SURFACE_ID(0));
runner_assert_or_return(ivisurf == NULL);
}
RUNNER_TEST(surface_visibility)
{
const struct ivi_controller_interface *ctl = ctx->controller_interface;
struct ivi_layout_surface *ivisurf;
int32_t ret;
bool visibility;
const struct ivi_layout_surface_properties *prop;
ivisurf = ctl->get_surface_from_id(IVI_TEST_SURFACE_ID(0));
runner_assert_or_return(ivisurf);
ret = ctl->surface_set_visibility(ivisurf, true);
runner_assert_or_return(ret == IVI_SUCCEEDED);
ctl->commit_changes();
visibility = ctl->surface_get_visibility(ivisurf);
runner_assert_or_return(visibility == true);
prop = ctl->get_properties_of_surface(ivisurf);
runner_assert_or_return(prop->visibility == true);
}
RUNNER_TEST(surface_opacity)
{
const struct ivi_controller_interface *ctl = ctx->controller_interface;
struct ivi_layout_surface *ivisurf;
int32_t ret;
wl_fixed_t opacity;
const struct ivi_layout_surface_properties *prop;
ivisurf = ctl->get_surface_from_id(IVI_TEST_SURFACE_ID(0));
runner_assert_or_return(ivisurf);
ret = ctl->surface_set_opacity(ivisurf, wl_fixed_from_double(0.5));
runner_assert_or_return(ret == IVI_SUCCEEDED);
ctl->commit_changes();
opacity = ctl->surface_get_opacity(ivisurf);
runner_assert_or_return(opacity == wl_fixed_from_double(0.5));
prop = ctl->get_properties_of_surface(ivisurf);
runner_assert_or_return(prop->opacity == wl_fixed_from_double(0.5));
}