luc14n0/glib
1
0
Fork 0
mirror of https://gitlab.gnome.org/GNOME/glib.git synced 2024-12-26 15:36:14 +01:00
Code Issues Packages Projects Releases Wiki Activity
309f5384ca
glib/gio/tests/task.c

2566 lines
74 KiB
C
Raw Normal View History

GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
/*
gio/tests/task: Run the worker indefinitely until it's cancelled Currently, the actual asynchronous work, represented by asynchronous_cancellation_run_task, was over before the GCancellable could be triggered. While that doesn't invalidate the purpose of the test, since it's fundamentally about cancellation, it would be nicer if the cancellation actually served some purpose instead of being a mere formality. https://gitlab.gnome.org/GNOME/glib/issues/1608
2019-02-10 20:11:22 +01:00
* Copyright 2012-2019 Red Hat, Inc.
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
*
tests: Add SPDX license headers automatically Add SPDX license (but not copyright) headers to all files which follow a certain pattern in their existing non-machine-readable header comment. This commit was entirely generated using the command: ``` git ls-files gio/tests/*.c | xargs perl -0777 -pi -e 's/\n \*\n \* This library is free software; you can redistribute it and\/or\n \* modify it under the terms of the GNU Lesser General Public/\n \*\n \* SPDX-License-Identifier: LGPL-2.1-or-later\n \*\n \* This library is free software; you can redistribute it and\/or\n \* modify it under the terms of the GNU Lesser General Public/igs' ``` Signed-off-by: Philip Withnall <pwithnall@endlessos.org> Helps: #1415
2022-05-18 10:20:07 +02:00
* SPDX-License-Identifier: LGPL-2.1-or-later
*
gio/tests/: LGPLv2+ -> LGPLv2.1+ A lot of tests in gio/tests/ don't have a license header. https://bugzilla.gnome.org/show_bug.cgi?id=776504
2017-05-27 17:19:21 +02:00
* This library is free software; you can redistribute it and/or
* modify it under the terms of the GNU Lesser General Public
* License as published by the Free Software Foundation; either
* version 2.1 of the License, or (at your option) any later version.
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
*
* See the included COPYING file for more information.
*/
#include <gio/gio.h>
gtask: remove hardcoded GTask thread-pool size GTask used a 10-thread thread pool for g_task_run_in_thread() / g_task_run_in_thread_sync(), but this ran into problems when task threads blocked waiting for another g_task_run_in_thread_sync() operation to complete. Previously there was a workaround for this, by bumping up the thread limit when that case was detected, but deadlocks could still happen if there were non-GTask threads involved. (Eg, task A sends a message to thread X and waits for a response, but thread X needs to complete task B in a thread before returning the response to task A.) So, allow GTask's thread pool to be expanded dynamically, by watching it from the glib worker thread, and growing it (at an exponentially-decreasing rate) if too much time passes without any tasks completing. This should solve the deadlocking problems without causing sudden breakage in apps that assume they can queue huge numbers of tasks at once without consequences. https://bugzilla.gnome.org/show_bug.cgi?id=687223
2015-03-09 21:33:16 +01:00
#include <string.h>
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
static GMainLoop *loop;
static GThread *main_thread;
static gssize magic;
/* We need objects for a few tests where we don't care what type
* they are, just that they're GObjects.
*/
#define g_dummy_object_new g_socket_client_new
gtask: Add a GTask:completed property This can be used to query whether the task has completed, in the sense that it has had a result set on it, and has already – or will soon – invoke its callback function. Notifications for this property are emitted immediately after the task’s main callback, in the same main context as that callback. This allows for multiple bits of code to listen for completion of the GTask, which opens the door for blocking on cancellation of the GTask and improved handling of ‘pending’ behaviour. https://bugzilla.gnome.org/show_bug.cgi?id=743636
2014-08-21 20:04:04 +02:00
static gboolean
idle_quit_loop (gpointer user_data)
{
g_main_loop_quit (loop);
return FALSE;
}
static void
completed_cb (GObject *gobject,
GParamSpec *pspec,
gpointer user_data)
{
gboolean *notification_emitted = user_data;
*notification_emitted = TRUE;
}
static void
wait_for_completed_notification (GTask *task)
{
gboolean notification_emitted = FALSE;
gboolean is_completed = FALSE;
/* Hold a ref. so we can check the :completed property afterwards. */
g_object_ref (task);
g_signal_connect (task, "notify::completed",
(GCallback) completed_cb, &notification_emitted);
g_idle_add (idle_quit_loop, NULL);
g_main_loop_run (loop);
g_assert_true (notification_emitted);
g_assert_true (g_task_get_completed (task));
g_object_get (G_OBJECT (task), "completed", &is_completed, NULL);
g_assert_true (is_completed);
g_object_unref (task);
}
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
/* test_basic */
static void
basic_callback (GObject *object,
GAsyncResult *result,
gpointer user_data)
{
gssize *result_out = user_data;
GError *error = NULL;
g_assert (object == NULL);
g_assert (g_task_is_valid (result, object));
g_assert (g_async_result_get_user_data (result) == user_data);
g_assert (!g_task_had_error (G_TASK (result)));
gtask: Add a GTask:completed property This can be used to query whether the task has completed, in the sense that it has had a result set on it, and has already – or will soon – invoke its callback function. Notifications for this property are emitted immediately after the task’s main callback, in the same main context as that callback. This allows for multiple bits of code to listen for completion of the GTask, which opens the door for blocking on cancellation of the GTask and improved handling of ‘pending’ behaviour. https://bugzilla.gnome.org/show_bug.cgi?id=743636
2014-08-21 20:04:04 +02:00
g_assert_false (g_task_get_completed (G_TASK (result)));
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
*result_out = g_task_propagate_int (G_TASK (result), &error);
g_assert_no_error (error);
gio/tests/task: Ensure that g_task_had_error doesn't forget the error https://bugzilla.gnome.org/show_bug.cgi?id=764163
2016-03-24 18:59:54 +01:00
g_assert (!g_task_had_error (G_TASK (result)));
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
g_main_loop_quit (loop);
}
static gboolean
basic_return (gpointer user_data)
{
GTask *task = user_data;
g_task_return_int (task, magic);
g_object_unref (task);
return FALSE;
}
static void
basic_destroy_notify (gpointer user_data)
{
gboolean *destroyed = user_data;
*destroyed = TRUE;
}
static void
test_basic (void)
{
GTask *task;
gssize result;
gboolean task_data_destroyed = FALSE;
gtask: Add a GTask:completed property This can be used to query whether the task has completed, in the sense that it has had a result set on it, and has already – or will soon – invoke its callback function. Notifications for this property are emitted immediately after the task’s main callback, in the same main context as that callback. This allows for multiple bits of code to listen for completion of the GTask, which opens the door for blocking on cancellation of the GTask and improved handling of ‘pending’ behaviour. https://bugzilla.gnome.org/show_bug.cgi?id=743636
2014-08-21 20:04:04 +02:00
gboolean notification_emitted = FALSE;
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
task = g_task_new (NULL, NULL, basic_callback, &result);
g_task_set_task_data (task, &task_data_destroyed, basic_destroy_notify);
g_object_add_weak_pointer (G_OBJECT (task), (gpointer *)&task);
gtask: Add a GTask:completed property This can be used to query whether the task has completed, in the sense that it has had a result set on it, and has already – or will soon – invoke its callback function. Notifications for this property are emitted immediately after the task’s main callback, in the same main context as that callback. This allows for multiple bits of code to listen for completion of the GTask, which opens the door for blocking on cancellation of the GTask and improved handling of ‘pending’ behaviour. https://bugzilla.gnome.org/show_bug.cgi?id=743636
2014-08-21 20:04:04 +02:00
g_signal_connect (task, "notify::completed",
(GCallback) completed_cb, &notification_emitted);
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
g_idle_add (basic_return, task);
g_main_loop_run (loop);
g_assert_cmpint (result, ==, magic);
g_assert (task_data_destroyed == TRUE);
gtask: Add a GTask:completed property This can be used to query whether the task has completed, in the sense that it has had a result set on it, and has already – or will soon – invoke its callback function. Notifications for this property are emitted immediately after the task’s main callback, in the same main context as that callback. This allows for multiple bits of code to listen for completion of the GTask, which opens the door for blocking on cancellation of the GTask and improved handling of ‘pending’ behaviour. https://bugzilla.gnome.org/show_bug.cgi?id=743636
2014-08-21 20:04:04 +02:00
g_assert_true (notification_emitted);
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
g_assert (task == NULL);
}
/* test_error */
static void
error_callback (GObject *object,
GAsyncResult *result,
gpointer user_data)
{
gssize *result_out = user_data;
GError *error = NULL;
g_assert (object == NULL);
g_assert (g_task_is_valid (result, object));
g_assert (g_async_result_get_user_data (result) == user_data);
g_assert (g_task_had_error (G_TASK (result)));
gtask: Add a GTask:completed property This can be used to query whether the task has completed, in the sense that it has had a result set on it, and has already – or will soon – invoke its callback function. Notifications for this property are emitted immediately after the task’s main callback, in the same main context as that callback. This allows for multiple bits of code to listen for completion of the GTask, which opens the door for blocking on cancellation of the GTask and improved handling of ‘pending’ behaviour. https://bugzilla.gnome.org/show_bug.cgi?id=743636
2014-08-21 20:04:04 +02:00
g_assert_false (g_task_get_completed (G_TASK (result)));
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
*result_out = g_task_propagate_int (G_TASK (result), &error);
g_assert_error (error, G_IO_ERROR, G_IO_ERROR_FAILED);
g_error_free (error);
gio/tests/task: Ensure that g_task_had_error doesn't forget the error https://bugzilla.gnome.org/show_bug.cgi?id=764163
2016-03-24 18:59:54 +01:00
g_assert (g_task_had_error (G_TASK (result)));
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
g_main_loop_quit (loop);
}
static gboolean
error_return (gpointer user_data)
{
GTask *task = user_data;
g_task_return_new_error (task,
G_IO_ERROR, G_IO_ERROR_FAILED,
"Failed");
g_object_unref (task);
return FALSE;
}
static void
error_destroy_notify (gpointer user_data)
{
gboolean *destroyed = user_data;
*destroyed = TRUE;
}
static void
test_error (void)
{
GTask *task;
gssize result;
gboolean first_task_data_destroyed = FALSE;
gboolean second_task_data_destroyed = FALSE;
gtask: Add a GTask:completed property This can be used to query whether the task has completed, in the sense that it has had a result set on it, and has already – or will soon – invoke its callback function. Notifications for this property are emitted immediately after the task’s main callback, in the same main context as that callback. This allows for multiple bits of code to listen for completion of the GTask, which opens the door for blocking on cancellation of the GTask and improved handling of ‘pending’ behaviour. https://bugzilla.gnome.org/show_bug.cgi?id=743636
2014-08-21 20:04:04 +02:00
gboolean notification_emitted = FALSE;
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
task = g_task_new (NULL, NULL, error_callback, &result);
g_object_add_weak_pointer (G_OBJECT (task), (gpointer *)&task);
gtask: Add a GTask:completed property This can be used to query whether the task has completed, in the sense that it has had a result set on it, and has already – or will soon – invoke its callback function. Notifications for this property are emitted immediately after the task’s main callback, in the same main context as that callback. This allows for multiple bits of code to listen for completion of the GTask, which opens the door for blocking on cancellation of the GTask and improved handling of ‘pending’ behaviour. https://bugzilla.gnome.org/show_bug.cgi?id=743636
2014-08-21 20:04:04 +02:00
g_signal_connect (task, "notify::completed",
(GCallback) completed_cb, &notification_emitted);
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
g_assert (first_task_data_destroyed == FALSE);
g_task_set_task_data (task, &first_task_data_destroyed, error_destroy_notify);
g_assert (first_task_data_destroyed == FALSE);
/* Calling g_task_set_task_data() again will destroy the first data */
g_task_set_task_data (task, &second_task_data_destroyed, error_destroy_notify);
g_assert (first_task_data_destroyed == TRUE);
g_assert (second_task_data_destroyed == FALSE);
g_idle_add (error_return, task);
g_main_loop_run (loop);
g_assert_cmpint (result, ==, -1);
g_assert (second_task_data_destroyed == TRUE);
gtask: Add a GTask:completed property This can be used to query whether the task has completed, in the sense that it has had a result set on it, and has already – or will soon – invoke its callback function. Notifications for this property are emitted immediately after the task’s main callback, in the same main context as that callback. This allows for multiple bits of code to listen for completion of the GTask, which opens the door for blocking on cancellation of the GTask and improved handling of ‘pending’ behaviour. https://bugzilla.gnome.org/show_bug.cgi?id=743636
2014-08-21 20:04:04 +02:00
g_assert_true (notification_emitted);
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
g_assert (task == NULL);
}
/* test_return_from_same_iteration: calling g_task_return_* from the
* loop iteration the task was created in defers completion until the
* next iteration.
*/
gboolean same_result = FALSE;
gtask: Add a GTask:completed property This can be used to query whether the task has completed, in the sense that it has had a result set on it, and has already – or will soon – invoke its callback function. Notifications for this property are emitted immediately after the task’s main callback, in the same main context as that callback. This allows for multiple bits of code to listen for completion of the GTask, which opens the door for blocking on cancellation of the GTask and improved handling of ‘pending’ behaviour. https://bugzilla.gnome.org/show_bug.cgi?id=743636
2014-08-21 20:04:04 +02:00
gboolean same_notification_emitted = FALSE;
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
static void
same_callback (GObject *object,
GAsyncResult *result,
gpointer user_data)
{
gboolean *result_out = user_data;
GError *error = NULL;
g_assert (object == NULL);
g_assert (g_task_is_valid (result, object));
g_assert (g_async_result_get_user_data (result) == user_data);
g_assert (!g_task_had_error (G_TASK (result)));
gtask: Add a GTask:completed property This can be used to query whether the task has completed, in the sense that it has had a result set on it, and has already – or will soon – invoke its callback function. Notifications for this property are emitted immediately after the task’s main callback, in the same main context as that callback. This allows for multiple bits of code to listen for completion of the GTask, which opens the door for blocking on cancellation of the GTask and improved handling of ‘pending’ behaviour. https://bugzilla.gnome.org/show_bug.cgi?id=743636
2014-08-21 20:04:04 +02:00
g_assert_false (g_task_get_completed (G_TASK (result)));
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
*result_out = g_task_propagate_boolean (G_TASK (result), &error);
g_assert_no_error (error);
gio/tests/task: Ensure that g_task_had_error doesn't forget the error https://bugzilla.gnome.org/show_bug.cgi?id=764163
2016-03-24 18:59:54 +01:00
g_assert (!g_task_had_error (G_TASK (result)));
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
g_main_loop_quit (loop);
}
static gboolean
same_start (gpointer user_data)
{
gpointer *weak_pointer = user_data;
GTask *task;
task = g_task_new (NULL, NULL, same_callback, &same_result);
*weak_pointer = task;
g_object_add_weak_pointer (G_OBJECT (task), weak_pointer);
gtask: Add a GTask:completed property This can be used to query whether the task has completed, in the sense that it has had a result set on it, and has already – or will soon – invoke its callback function. Notifications for this property are emitted immediately after the task’s main callback, in the same main context as that callback. This allows for multiple bits of code to listen for completion of the GTask, which opens the door for blocking on cancellation of the GTask and improved handling of ‘pending’ behaviour. https://bugzilla.gnome.org/show_bug.cgi?id=743636
2014-08-21 20:04:04 +02:00
g_signal_connect (task, "notify::completed",
(GCallback) completed_cb, &same_notification_emitted);
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
g_task_return_boolean (task, TRUE);
g_object_unref (task);
/* same_callback should not have been invoked yet */
g_assert (same_result == FALSE);
g_assert (*weak_pointer == task);
gtask: Add a GTask:completed property This can be used to query whether the task has completed, in the sense that it has had a result set on it, and has already – or will soon – invoke its callback function. Notifications for this property are emitted immediately after the task’s main callback, in the same main context as that callback. This allows for multiple bits of code to listen for completion of the GTask, which opens the door for blocking on cancellation of the GTask and improved handling of ‘pending’ behaviour. https://bugzilla.gnome.org/show_bug.cgi?id=743636
2014-08-21 20:04:04 +02:00
g_assert_false (same_notification_emitted);
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
return FALSE;
}
static void
test_return_from_same_iteration (void)
{
gpointer weak_pointer;
g_idle_add (same_start, &weak_pointer);
g_main_loop_run (loop);
g_assert (same_result == TRUE);
g_assert (weak_pointer == NULL);
gtask: Add a GTask:completed property This can be used to query whether the task has completed, in the sense that it has had a result set on it, and has already – or will soon – invoke its callback function. Notifications for this property are emitted immediately after the task’s main callback, in the same main context as that callback. This allows for multiple bits of code to listen for completion of the GTask, which opens the door for blocking on cancellation of the GTask and improved handling of ‘pending’ behaviour. https://bugzilla.gnome.org/show_bug.cgi?id=743636
2014-08-21 20:04:04 +02:00
g_assert_true (same_notification_emitted);
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
}
/* test_return_from_toplevel: calling g_task_return_* from outside any
* main loop completes the task inside the main loop.
*/
gtask: Add a GTask:completed property This can be used to query whether the task has completed, in the sense that it has had a result set on it, and has already – or will soon – invoke its callback function. Notifications for this property are emitted immediately after the task’s main callback, in the same main context as that callback. This allows for multiple bits of code to listen for completion of the GTask, which opens the door for blocking on cancellation of the GTask and improved handling of ‘pending’ behaviour. https://bugzilla.gnome.org/show_bug.cgi?id=743636
2014-08-21 20:04:04 +02:00
gboolean toplevel_notification_emitted = FALSE;
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
static void
toplevel_callback (GObject *object,
GAsyncResult *result,
gpointer user_data)
{
gboolean *result_out = user_data;
GError *error = NULL;
g_assert (object == NULL);
g_assert (g_task_is_valid (result, object));
g_assert (g_async_result_get_user_data (result) == user_data);
g_assert (!g_task_had_error (G_TASK (result)));
gtask: Add a GTask:completed property This can be used to query whether the task has completed, in the sense that it has had a result set on it, and has already – or will soon – invoke its callback function. Notifications for this property are emitted immediately after the task’s main callback, in the same main context as that callback. This allows for multiple bits of code to listen for completion of the GTask, which opens the door for blocking on cancellation of the GTask and improved handling of ‘pending’ behaviour. https://bugzilla.gnome.org/show_bug.cgi?id=743636
2014-08-21 20:04:04 +02:00
g_assert_false (g_task_get_completed (G_TASK (result)));
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
*result_out = g_task_propagate_boolean (G_TASK (result), &error);
g_assert_no_error (error);
gio/tests/task: Ensure that g_task_had_error doesn't forget the error https://bugzilla.gnome.org/show_bug.cgi?id=764163
2016-03-24 18:59:54 +01:00
g_assert (!g_task_had_error (G_TASK (result)));
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
g_main_loop_quit (loop);
}
static void
test_return_from_toplevel (void)
{
GTask *task;
gboolean result = FALSE;
task = g_task_new (NULL, NULL, toplevel_callback, &result);
g_object_add_weak_pointer (G_OBJECT (task), (gpointer *)&task);
gtask: Add a GTask:completed property This can be used to query whether the task has completed, in the sense that it has had a result set on it, and has already – or will soon – invoke its callback function. Notifications for this property are emitted immediately after the task’s main callback, in the same main context as that callback. This allows for multiple bits of code to listen for completion of the GTask, which opens the door for blocking on cancellation of the GTask and improved handling of ‘pending’ behaviour. https://bugzilla.gnome.org/show_bug.cgi?id=743636
2014-08-21 20:04:04 +02:00
g_signal_connect (task, "notify::completed",
(GCallback) completed_cb, &toplevel_notification_emitted);
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
g_task_return_boolean (task, TRUE);
g_object_unref (task);
/* toplevel_callback should not have been invoked yet */
g_assert (result == FALSE);
g_assert (task != NULL);
gtask: Add a GTask:completed property This can be used to query whether the task has completed, in the sense that it has had a result set on it, and has already – or will soon – invoke its callback function. Notifications for this property are emitted immediately after the task’s main callback, in the same main context as that callback. This allows for multiple bits of code to listen for completion of the GTask, which opens the door for blocking on cancellation of the GTask and improved handling of ‘pending’ behaviour. https://bugzilla.gnome.org/show_bug.cgi?id=743636
2014-08-21 20:04:04 +02:00
g_assert_false (toplevel_notification_emitted);
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
g_main_loop_run (loop);
g_assert (result == TRUE);
g_assert (task == NULL);
gtask: Add a GTask:completed property This can be used to query whether the task has completed, in the sense that it has had a result set on it, and has already – or will soon – invoke its callback function. Notifications for this property are emitted immediately after the task’s main callback, in the same main context as that callback. This allows for multiple bits of code to listen for completion of the GTask, which opens the door for blocking on cancellation of the GTask and improved handling of ‘pending’ behaviour. https://bugzilla.gnome.org/show_bug.cgi?id=743636
2014-08-21 20:04:04 +02:00
g_assert_true (toplevel_notification_emitted);
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
}
/* test_return_from_anon_thread: calling g_task_return_* from a
* thread with no thread-default main context will complete the
* task in the task's context/thread.
*/
gtask: Add a GTask:completed property This can be used to query whether the task has completed, in the sense that it has had a result set on it, and has already – or will soon – invoke its callback function. Notifications for this property are emitted immediately after the task’s main callback, in the same main context as that callback. This allows for multiple bits of code to listen for completion of the GTask, which opens the door for blocking on cancellation of the GTask and improved handling of ‘pending’ behaviour. https://bugzilla.gnome.org/show_bug.cgi?id=743636
2014-08-21 20:04:04 +02:00
gboolean anon_thread_notification_emitted = FALSE;
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
GThread *anon_thread;
static void
anon_callback (GObject *object,
GAsyncResult *result,
gpointer user_data)
{
gssize *result_out = user_data;
GError *error = NULL;
g_assert (object == NULL);
g_assert (g_task_is_valid (result, object));
g_assert (g_async_result_get_user_data (result) == user_data);
g_assert (!g_task_had_error (G_TASK (result)));
gtask: Add a GTask:completed property This can be used to query whether the task has completed, in the sense that it has had a result set on it, and has already – or will soon – invoke its callback function. Notifications for this property are emitted immediately after the task’s main callback, in the same main context as that callback. This allows for multiple bits of code to listen for completion of the GTask, which opens the door for blocking on cancellation of the GTask and improved handling of ‘pending’ behaviour. https://bugzilla.gnome.org/show_bug.cgi?id=743636
2014-08-21 20:04:04 +02:00
g_assert_false (g_task_get_completed (G_TASK (result)));
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
g_assert (g_thread_self () == main_thread);
*result_out = g_task_propagate_int (G_TASK (result), &error);
g_assert_no_error (error);
gio/tests/task: Ensure that g_task_had_error doesn't forget the error https://bugzilla.gnome.org/show_bug.cgi?id=764163
2016-03-24 18:59:54 +01:00
g_assert (!g_task_had_error (G_TASK (result)));
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
g_main_loop_quit (loop);
}
static gpointer
anon_thread_func (gpointer user_data)
{
GTask *task = user_data;
g_task_return_int (task, magic);
g_object_unref (task);
return NULL;
}
static gboolean
anon_start (gpointer user_data)
{
GTask *task = user_data;
anon_thread = g_thread_new ("test_return_from_anon_thread",
anon_thread_func, task);
return FALSE;
}
static void
test_return_from_anon_thread (void)
{
GTask *task;
gssize result = 0;
task = g_task_new (NULL, NULL, anon_callback, &result);
g_object_add_weak_pointer (G_OBJECT (task), (gpointer *)&task);
gtask: Add a GTask:completed property This can be used to query whether the task has completed, in the sense that it has had a result set on it, and has already – or will soon – invoke its callback function. Notifications for this property are emitted immediately after the task’s main callback, in the same main context as that callback. This allows for multiple bits of code to listen for completion of the GTask, which opens the door for blocking on cancellation of the GTask and improved handling of ‘pending’ behaviour. https://bugzilla.gnome.org/show_bug.cgi?id=743636
2014-08-21 20:04:04 +02:00
g_signal_connect (task, "notify::completed",
(GCallback) completed_cb,
&anon_thread_notification_emitted);
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
g_idle_add (anon_start, task);
g_main_loop_run (loop);
g_thread_join (anon_thread);
g_assert_cmpint (result, ==, magic);
g_assert (task == NULL);
gtask: Add a GTask:completed property This can be used to query whether the task has completed, in the sense that it has had a result set on it, and has already – or will soon – invoke its callback function. Notifications for this property are emitted immediately after the task’s main callback, in the same main context as that callback. This allows for multiple bits of code to listen for completion of the GTask, which opens the door for blocking on cancellation of the GTask and improved handling of ‘pending’ behaviour. https://bugzilla.gnome.org/show_bug.cgi?id=743636
2014-08-21 20:04:04 +02:00
g_assert_true (anon_thread_notification_emitted);
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
}
/* test_return_from_wrong_thread: calling g_task_return_* from a
* thread with its own thread-default main context will complete the
* task in the task's context/thread.
*/
gtask: Add a GTask:completed property This can be used to query whether the task has completed, in the sense that it has had a result set on it, and has already – or will soon – invoke its callback function. Notifications for this property are emitted immediately after the task’s main callback, in the same main context as that callback. This allows for multiple bits of code to listen for completion of the GTask, which opens the door for blocking on cancellation of the GTask and improved handling of ‘pending’ behaviour. https://bugzilla.gnome.org/show_bug.cgi?id=743636
2014-08-21 20:04:04 +02:00
gboolean wrong_thread_notification_emitted = FALSE;
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
GThread *wrong_thread;
static void
wrong_callback (GObject *object,
GAsyncResult *result,
gpointer user_data)
{
gssize *result_out = user_data;
GError *error = NULL;
g_assert (object == NULL);
g_assert (g_task_is_valid (result, object));
g_assert (g_async_result_get_user_data (result) == user_data);
g_assert (!g_task_had_error (G_TASK (result)));
gtask: Add a GTask:completed property This can be used to query whether the task has completed, in the sense that it has had a result set on it, and has already – or will soon – invoke its callback function. Notifications for this property are emitted immediately after the task’s main callback, in the same main context as that callback. This allows for multiple bits of code to listen for completion of the GTask, which opens the door for blocking on cancellation of the GTask and improved handling of ‘pending’ behaviour. https://bugzilla.gnome.org/show_bug.cgi?id=743636
2014-08-21 20:04:04 +02:00
g_assert_false (g_task_get_completed (G_TASK (result)));
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
g_assert (g_thread_self () == main_thread);
*result_out = g_task_propagate_int (G_TASK (result), &error);
g_assert_no_error (error);
gio/tests/task: Ensure that g_task_had_error doesn't forget the error https://bugzilla.gnome.org/show_bug.cgi?id=764163
2016-03-24 18:59:54 +01:00
g_assert (!g_task_had_error (G_TASK (result)));
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
g_main_loop_quit (loop);
}
static gpointer
wrong_thread_func (gpointer user_data)
{
GTask *task = user_data;
GMainContext *context;
context = g_main_context_new ();
g_main_context_push_thread_default (context);
g_assert (g_task_get_context (task) != context);
g_task_return_int (task, magic);
g_object_unref (task);
g_main_context_pop_thread_default (context);
g_main_context_unref (context);
return NULL;
}
static gboolean
wrong_start (gpointer user_data)
{
GTask *task = user_data;
wrong_thread = g_thread_new ("test_return_from_anon_thread",
wrong_thread_func, task);
return FALSE;
}
static void
test_return_from_wrong_thread (void)
{
GTask *task;
gssize result = 0;
task = g_task_new (NULL, NULL, wrong_callback, &result);
g_object_add_weak_pointer (G_OBJECT (task), (gpointer *)&task);
gtask: Add a GTask:completed property This can be used to query whether the task has completed, in the sense that it has had a result set on it, and has already – or will soon – invoke its callback function. Notifications for this property are emitted immediately after the task’s main callback, in the same main context as that callback. This allows for multiple bits of code to listen for completion of the GTask, which opens the door for blocking on cancellation of the GTask and improved handling of ‘pending’ behaviour. https://bugzilla.gnome.org/show_bug.cgi?id=743636
2014-08-21 20:04:04 +02:00
g_signal_connect (task, "notify::completed",
(GCallback) completed_cb,
&wrong_thread_notification_emitted);
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
g_idle_add (wrong_start, task);
g_main_loop_run (loop);
g_thread_join (wrong_thread);
g_assert_cmpint (result, ==, magic);
g_assert (task == NULL);
gtask: Add a GTask:completed property This can be used to query whether the task has completed, in the sense that it has had a result set on it, and has already – or will soon – invoke its callback function. Notifications for this property are emitted immediately after the task’s main callback, in the same main context as that callback. This allows for multiple bits of code to listen for completion of the GTask, which opens the door for blocking on cancellation of the GTask and improved handling of ‘pending’ behaviour. https://bugzilla.gnome.org/show_bug.cgi?id=743636
2014-08-21 20:04:04 +02:00
g_assert_true (wrong_thread_notification_emitted);
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
}
/* test_no_callback */
static void
test_no_callback (void)
{
GTask *task;
task = g_task_new (NULL, NULL, NULL, NULL);
g_object_add_weak_pointer (G_OBJECT (task), (gpointer *)&task);
g_task_return_boolean (task, TRUE);
g_object_unref (task);
gtask: Add a GTask:completed property This can be used to query whether the task has completed, in the sense that it has had a result set on it, and has already – or will soon – invoke its callback function. Notifications for this property are emitted immediately after the task’s main callback, in the same main context as that callback. This allows for multiple bits of code to listen for completion of the GTask, which opens the door for blocking on cancellation of the GTask and improved handling of ‘pending’ behaviour. https://bugzilla.gnome.org/show_bug.cgi?id=743636
2014-08-21 20:04:04 +02:00
/* Even though there’s no callback, the :completed notification has to
* happen in an idle handler. */
g_assert_nonnull (task);
wait_for_completed_notification (task);
g_assert_null (task);
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
}
/* test_report_error */
static void test_report_error (void);
gtask: Add a GTask:completed property This can be used to query whether the task has completed, in the sense that it has had a result set on it, and has already – or will soon – invoke its callback function. Notifications for this property are emitted immediately after the task’s main callback, in the same main context as that callback. This allows for multiple bits of code to listen for completion of the GTask, which opens the door for blocking on cancellation of the GTask and improved handling of ‘pending’ behaviour. https://bugzilla.gnome.org/show_bug.cgi?id=743636
2014-08-21 20:04:04 +02:00
gboolean error_notification_emitted = FALSE;
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
static void
report_callback (GObject *object,
GAsyncResult *result,
gpointer user_data)
{
gpointer *weak_pointer = user_data;
GError *error = NULL;
gssize ret;
g_assert (object == NULL);
g_assert (g_task_is_valid (result, object));
g_assert (g_async_result_get_user_data (result) == user_data);
g_assert (g_async_result_is_tagged (result, test_report_error));
g_assert (g_task_get_source_tag (G_TASK (result)) == test_report_error);
g_assert (g_task_had_error (G_TASK (result)));
gtask: Add a GTask:completed property This can be used to query whether the task has completed, in the sense that it has had a result set on it, and has already – or will soon – invoke its callback function. Notifications for this property are emitted immediately after the task’s main callback, in the same main context as that callback. This allows for multiple bits of code to listen for completion of the GTask, which opens the door for blocking on cancellation of the GTask and improved handling of ‘pending’ behaviour. https://bugzilla.gnome.org/show_bug.cgi?id=743636
2014-08-21 20:04:04 +02:00
g_assert_false (g_task_get_completed (G_TASK (result)));
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
ret = g_task_propagate_int (G_TASK (result), &error);
g_assert_error (error, G_IO_ERROR, G_IO_ERROR_FAILED);
g_assert_cmpint (ret, ==, -1);
g_error_free (error);
gio/tests/task: Ensure that g_task_had_error doesn't forget the error https://bugzilla.gnome.org/show_bug.cgi?id=764163
2016-03-24 18:59:54 +01:00
g_assert (g_task_had_error (G_TASK (result)));
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
*weak_pointer = result;
g_object_add_weak_pointer (G_OBJECT (result), weak_pointer);
gtask: Add a GTask:completed property This can be used to query whether the task has completed, in the sense that it has had a result set on it, and has already – or will soon – invoke its callback function. Notifications for this property are emitted immediately after the task’s main callback, in the same main context as that callback. This allows for multiple bits of code to listen for completion of the GTask, which opens the door for blocking on cancellation of the GTask and improved handling of ‘pending’ behaviour. https://bugzilla.gnome.org/show_bug.cgi?id=743636
2014-08-21 20:04:04 +02:00
g_signal_connect (result, "notify::completed",
(GCallback) completed_cb, &error_notification_emitted);
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
g_main_loop_quit (loop);
}
static void
test_report_error (void)
{
gpointer weak_pointer = (gpointer)-1;
g_task_report_new_error (NULL, report_callback, &weak_pointer,
test_report_error,
G_IO_ERROR, G_IO_ERROR_FAILED,
"Failed");
g_main_loop_run (loop);
g_assert (weak_pointer == NULL);
gtask: Add a GTask:completed property This can be used to query whether the task has completed, in the sense that it has had a result set on it, and has already – or will soon – invoke its callback function. Notifications for this property are emitted immediately after the task’s main callback, in the same main context as that callback. This allows for multiple bits of code to listen for completion of the GTask, which opens the door for blocking on cancellation of the GTask and improved handling of ‘pending’ behaviour. https://bugzilla.gnome.org/show_bug.cgi?id=743636
2014-08-21 20:04:04 +02:00
g_assert_true (error_notification_emitted);
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
}
/* test_priority: tasks complete in priority order */
static int counter = 0;
static void
priority_callback (GObject *object,
GAsyncResult *result,
gpointer user_data)
{
gssize *ret_out = user_data;
GError *error = NULL;
g_assert (object == NULL);
g_assert (g_task_is_valid (result, object));
g_assert (g_async_result_get_user_data (result) == user_data);
g_assert (!g_task_had_error (G_TASK (result)));
gtask: Add a GTask:completed property This can be used to query whether the task has completed, in the sense that it has had a result set on it, and has already – or will soon – invoke its callback function. Notifications for this property are emitted immediately after the task’s main callback, in the same main context as that callback. This allows for multiple bits of code to listen for completion of the GTask, which opens the door for blocking on cancellation of the GTask and improved handling of ‘pending’ behaviour. https://bugzilla.gnome.org/show_bug.cgi?id=743636
2014-08-21 20:04:04 +02:00
g_assert_false (g_task_get_completed (G_TASK (result)));
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
g_task_propagate_boolean (G_TASK (result), &error);
g_assert_no_error (error);
gio/tests/task: Ensure that g_task_had_error doesn't forget the error https://bugzilla.gnome.org/show_bug.cgi?id=764163
2016-03-24 18:59:54 +01:00
g_assert (!g_task_had_error (G_TASK (result)));
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
*ret_out = ++counter;
if (counter == 3)
g_main_loop_quit (loop);
}
static void
test_priority (void)
{
GTask *t1, *t2, *t3;
gssize ret1, ret2, ret3;
/* t2 has higher priority than either t1 or t3, so we can't
* accidentally pass the test just by completing the tasks in the
* order they were created (or in reverse order).
*/
t1 = g_task_new (NULL, NULL, priority_callback, &ret1);
g_task_set_priority (t1, G_PRIORITY_DEFAULT);
g_task_return_boolean (t1, TRUE);
g_object_unref (t1);
t2 = g_task_new (NULL, NULL, priority_callback, &ret2);
g_task_set_priority (t2, G_PRIORITY_HIGH);
g_task_return_boolean (t2, TRUE);
g_object_unref (t2);
t3 = g_task_new (NULL, NULL, priority_callback, &ret3);
g_task_set_priority (t3, G_PRIORITY_LOW);
g_task_return_boolean (t3, TRUE);
g_object_unref (t3);
g_main_loop_run (loop);
g_assert_cmpint (ret2, ==, 1);
g_assert_cmpint (ret1, ==, 2);
g_assert_cmpint (ret3, ==, 3);
}
gtask: Add a g_task_set_name() method Similarly to g_source_set_name(), this sets a name on a GTask for debugging and profiling. Importantly, this name is propagated to the GSource for idle callbacks for the GTask, ending the glorious reign of `[gio] complete_in_idle_cb`. The name can be queried using g_task_get_name(). Locking is avoided by only allowing the name to be set before the GTask is used from another thread. Includes tests. Signed-off-by: Philip Withnall <withnall@endlessm.com>
2018-10-07 18:37:58 +02:00
/* Test that getting and setting the task name works. */
static void name_callback (GObject *object,
GAsyncResult *result,
gpointer user_data);
static void
test_name (void)
{
GTask *t1 = NULL;
tests: Tweak the task tests Tweak the g_task_set_name test to use a non-literal string, so we don't test g_task_set_static_name twice.
2022-09-22 00:12:00 +02:00
char *orig = g_strdup ("some task");
gtask: Add a g_task_set_name() method Similarly to g_source_set_name(), this sets a name on a GTask for debugging and profiling. Importantly, this name is propagated to the GSource for idle callbacks for the GTask, ending the glorious reign of `[gio] complete_in_idle_cb`. The name can be queried using g_task_get_name(). Locking is avoided by only allowing the name to be set before the GTask is used from another thread. Includes tests. Signed-off-by: Philip Withnall <withnall@endlessm.com>
2018-10-07 18:37:58 +02:00
gchar *name1 = NULL;
t1 = g_task_new (NULL, NULL, name_callback, &name1);
tests: Tweak the task tests Tweak the g_task_set_name test to use a non-literal string, so we don't test g_task_set_static_name twice.
2022-09-22 00:12:00 +02:00
(g_task_set_name) (t1, orig);
gtask: Add a g_task_set_name() method Similarly to g_source_set_name(), this sets a name on a GTask for debugging and profiling. Importantly, this name is propagated to the GSource for idle callbacks for the GTask, ending the glorious reign of `[gio] complete_in_idle_cb`. The name can be queried using g_task_get_name(). Locking is avoided by only allowing the name to be set before the GTask is used from another thread. Includes tests. Signed-off-by: Philip Withnall <withnall@endlessm.com>
2018-10-07 18:37:58 +02:00
g_task_return_boolean (t1, TRUE);
g_object_unref (t1);
g_main_loop_run (loop);
tests: Tweak the task tests Tweak the g_task_set_name test to use a non-literal string, so we don't test g_task_set_static_name twice.
2022-09-22 00:12:00 +02:00
g_assert_cmpstr (name1, ==, orig);
gtask: Add a g_task_set_name() method Similarly to g_source_set_name(), this sets a name on a GTask for debugging and profiling. Importantly, this name is propagated to the GSource for idle callbacks for the GTask, ending the glorious reign of `[gio] complete_in_idle_cb`. The name can be queried using g_task_get_name(). Locking is avoided by only allowing the name to be set before the GTask is used from another thread. Includes tests. Signed-off-by: Philip Withnall <withnall@endlessm.com>
2018-10-07 18:37:58 +02:00
g_free (name1);
tests: Tweak the task tests Tweak the g_task_set_name test to use a non-literal string, so we don't test g_task_set_static_name twice.
2022-09-22 00:12:00 +02:00
g_free (orig);
}
static void
test_name_macro_wrapper (void)
{
GTask *t1 = NULL;
char *orig = g_strdup ("some task");
gchar *name1 = NULL;
t1 = g_task_new (NULL, NULL, name_callback, &name1);
g_task_set_name (t1, orig);
g_task_return_boolean (t1, TRUE);
g_object_unref (t1);
g_main_loop_run (loop);
g_assert_cmpstr (name1, ==, orig);
g_free (name1);
g_free (orig);
gtask: Add a g_task_set_name() method Similarly to g_source_set_name(), this sets a name on a GTask for debugging and profiling. Importantly, this name is propagated to the GSource for idle callbacks for the GTask, ending the glorious reign of `[gio] complete_in_idle_cb`. The name can be queried using g_task_get_name(). Locking is avoided by only allowing the name to be set before the GTask is used from another thread. Includes tests. Signed-off-by: Philip Withnall <withnall@endlessm.com>
2018-10-07 18:37:58 +02:00
}
static void
name_callback (GObject *object,
GAsyncResult *result,
gpointer user_data)
{
gchar **name_out = user_data;
GError *local_error = NULL;
g_assert_null (*name_out);
*name_out = g_strdup (g_task_get_name (G_TASK (result)));
g_task_propagate_boolean (G_TASK (result), &local_error);
g_assert_no_error (local_error);
g_main_loop_quit (loop);
}
tests: Add a test for g_task_set_static_name() The test checks that the original string is returned by g_task_get_name().
2022-09-11 21:11:34 +02:00
static void static_name_callback (GObject *object,
GAsyncResult *result,
gpointer user_data);
static void
test_static_name (void)
{
GTask *t1 = NULL;
char *orig = "some task";
tests: Tweak the task tests Tweak the g_task_set_name test to use a non-literal string, so we don't test g_task_set_static_name twice.
2022-09-22 00:12:00 +02:00
char *name1 = NULL;
tests: Add a test for g_task_set_static_name() The test checks that the original string is returned by g_task_get_name().
2022-09-11 21:11:34 +02:00
t1 = g_task_new (NULL, NULL, static_name_callback, &name1);
g_task_set_static_name (t1, orig);
g_task_return_boolean (t1, TRUE);
g_object_unref (t1);
g_main_loop_run (loop);
g_assert_true (name1 == orig);
}
static void
static_name_callback (GObject *object,
GAsyncResult *result,
gpointer user_data)
{
const char **name_out = user_data;
GError *local_error = NULL;
g_assert_null (*name_out);
*name_out = g_task_get_name (G_TASK (result));
g_task_propagate_boolean (G_TASK (result), &local_error);
g_assert_no_error (local_error);
g_main_loop_quit (loop);
}
gio/tests/task: Ensure that cancelled tasks are returned asynchronously Closes https://gitlab.gnome.org/GNOME/glib/issues/1608
2018-12-04 19:54:10 +01:00
/* test_asynchronous_cancellation: cancelled tasks are returned
* asynchronously, i.e. not from inside the GCancellable::cancelled
* handler.
gio/tests/task: Add comments documenting how the test works https://gitlab.gnome.org/GNOME/glib/issues/1608
2019-02-11 12:30:23 +01:00
*
* The test is set up further below in test_asynchronous_cancellation.
gio/tests/task: Ensure that cancelled tasks are returned asynchronously Closes https://gitlab.gnome.org/GNOME/glib/issues/1608
2018-12-04 19:54:10 +01:00
*/
gio/tests/task: Add comments documenting how the test works https://gitlab.gnome.org/GNOME/glib/issues/1608
2019-02-11 12:30:23 +01:00
/* asynchronous_cancellation_callback represents the callback that the
* caller of a typical asynchronous API would have passed. See
* test_asynchronous_cancellation.
*/
gio/tests/task: Ensure that cancelled tasks are returned asynchronously Closes https://gitlab.gnome.org/GNOME/glib/issues/1608
2018-12-04 19:54:10 +01:00
static void
asynchronous_cancellation_callback (GObject *object,
GAsyncResult *result,
gpointer user_data)
{
GError *error = NULL;
gio/tests/task: Run the worker indefinitely until it's cancelled Currently, the actual asynchronous work, represented by asynchronous_cancellation_run_task, was over before the GCancellable could be triggered. While that doesn't invalidate the purpose of the test, since it's fundamentally about cancellation, it would be nicer if the cancellation actually served some purpose instead of being a mere formality. https://gitlab.gnome.org/GNOME/glib/issues/1608
2019-02-10 20:11:22 +01:00
guint run_task_id;
gio/tests/task: Ensure that cancelled tasks are returned asynchronously Closes https://gitlab.gnome.org/GNOME/glib/issues/1608
2018-12-04 19:54:10 +01:00
g_assert_null (object);
g_assert_true (g_task_is_valid (result, object));
g_assert_true (g_async_result_get_user_data (result) == user_data);
g_assert_true (g_task_had_error (G_TASK (result)));
g_assert_false (g_task_get_completed (G_TASK (result)));
gio/tests/task: Run the worker indefinitely until it's cancelled Currently, the actual asynchronous work, represented by asynchronous_cancellation_run_task, was over before the GCancellable could be triggered. While that doesn't invalidate the purpose of the test, since it's fundamentally about cancellation, it would be nicer if the cancellation actually served some purpose instead of being a mere formality. https://gitlab.gnome.org/GNOME/glib/issues/1608
2019-02-10 20:11:22 +01:00
run_task_id = GPOINTER_TO_UINT (g_task_get_task_data (G_TASK (result)));
g_assert_cmpuint (run_task_id, ==, 0);
gio/tests/task: Ensure that cancelled tasks are returned asynchronously Closes https://gitlab.gnome.org/GNOME/glib/issues/1608
2018-12-04 19:54:10 +01:00
g_task_propagate_boolean (G_TASK (result), &error);
g_assert_error (error, G_IO_ERROR, G_IO_ERROR_CANCELLED);
g_clear_error (&error);
g_assert_true (g_task_had_error (G_TASK (result)));
g_main_loop_quit (loop);
}
gio/tests/task: Add comments documenting how the test works https://gitlab.gnome.org/GNOME/glib/issues/1608
2019-02-11 12:30:23 +01:00
/* asynchronous_cancellation_cancel_task represents a user cancelling
* the ongoing operation. To make it somewhat realistic it is delayed
* by 50ms via a timeout GSource. See test_asynchronous_cancellation.
*/
gio/tests/task: Ensure that cancelled tasks are returned asynchronously Closes https://gitlab.gnome.org/GNOME/glib/issues/1608
2018-12-04 19:54:10 +01:00
static gboolean
asynchronous_cancellation_cancel_task (gpointer user_data)
{
GCancellable *cancellable;
GTask *task = G_TASK (user_data);
cancellable = g_task_get_cancellable (task);
g_assert_true (G_IS_CANCELLABLE (cancellable));
g_cancellable_cancel (cancellable);
g_assert_false (g_task_get_completed (task));
return G_SOURCE_REMOVE;
}
gio/tests/task: Add comments documenting how the test works https://gitlab.gnome.org/GNOME/glib/issues/1608
2019-02-11 12:30:23 +01:00
/* asynchronous_cancellation_cancelled is the GCancellable::cancelled
* handler that's used by the asynchronous implementation for
* cancelling itself.
*/
gio/tests/task: Ensure that cancelled tasks are returned asynchronously Closes https://gitlab.gnome.org/GNOME/glib/issues/1608
2018-12-04 19:54:10 +01:00
static void
asynchronous_cancellation_cancelled (GCancellable *cancellable,
gpointer user_data)
{
GTask *task = G_TASK (user_data);
guint run_task_id;
g_assert_true (cancellable == g_task_get_cancellable (task));
run_task_id = GPOINTER_TO_UINT (g_task_get_task_data (task));
gio/tests/task: Run the worker indefinitely until it's cancelled Currently, the actual asynchronous work, represented by asynchronous_cancellation_run_task, was over before the GCancellable could be triggered. While that doesn't invalidate the purpose of the test, since it's fundamentally about cancellation, it would be nicer if the cancellation actually served some purpose instead of being a mere formality. https://gitlab.gnome.org/GNOME/glib/issues/1608
2019-02-10 20:11:22 +01:00
g_assert_cmpuint (run_task_id, !=, 0);
g_source_remove (run_task_id);
g_task_set_task_data (task, GUINT_TO_POINTER (0), NULL);
gio/tests/task: Ensure that cancelled tasks are returned asynchronously Closes https://gitlab.gnome.org/GNOME/glib/issues/1608
2018-12-04 19:54:10 +01:00
g_task_return_boolean (task, FALSE);
g_assert_false (g_task_get_completed (task));
}
gio/tests/task: Add comments documenting how the test works https://gitlab.gnome.org/GNOME/glib/issues/1608
2019-02-11 12:30:23 +01:00
/* asynchronous_cancellation_run_task represents the actual
* asynchronous work being done in an idle GSource as was mentioned
* above. This is effectively meant to be an infinite loop so that
* the only way to break out of it is via cancellation.
*/
gio/tests/task: Ensure that cancelled tasks are returned asynchronously Closes https://gitlab.gnome.org/GNOME/glib/issues/1608
2018-12-04 19:54:10 +01:00
static gboolean
asynchronous_cancellation_run_task (gpointer user_data)
{
GCancellable *cancellable;
GTask *task = G_TASK (user_data);
cancellable = g_task_get_cancellable (task);
g_assert_true (G_IS_CANCELLABLE (cancellable));
g_assert_false (g_cancellable_is_cancelled (cancellable));
gio/tests/task: Run the worker indefinitely until it's cancelled Currently, the actual asynchronous work, represented by asynchronous_cancellation_run_task, was over before the GCancellable could be triggered. While that doesn't invalidate the purpose of the test, since it's fundamentally about cancellation, it would be nicer if the cancellation actually served some purpose instead of being a mere formality. https://gitlab.gnome.org/GNOME/glib/issues/1608
2019-02-10 20:11:22 +01:00
return G_SOURCE_CONTINUE;
gio/tests/task: Ensure that cancelled tasks are returned asynchronously Closes https://gitlab.gnome.org/GNOME/glib/issues/1608
2018-12-04 19:54:10 +01:00
}
/* Test that cancellation is always asynchronous. The completion callback for
gio/tests/task: Add comments documenting how the test works https://gitlab.gnome.org/GNOME/glib/issues/1608
2019-02-11 12:30:23 +01:00
* a #GTask must not be called from inside the cancellation handler.
*
* The body of the loop inside test_asynchronous_cancellation
* represents what would have been a typical asynchronous API call,
* and its implementation. They are fused together without an API
* boundary. The actual work done by this asynchronous API is
* represented by an idle GSource.
*/
gio/tests/task: Ensure that cancelled tasks are returned asynchronously Closes https://gitlab.gnome.org/GNOME/glib/issues/1608
2018-12-04 19:54:10 +01:00
static void
test_asynchronous_cancellation (void)
{
guint i;
g_test_bug ("https://gitlab.gnome.org/GNOME/glib/issues/1608");
/* Run a few times to shake out any timing issues between the
* cancellation and task sources.
*/
for (i = 0; i < 5; i++)
{
GCancellable *cancellable;
GTask *task;
gboolean notification_emitted = FALSE;
guint run_task_id;
cancellable = g_cancellable_new ();
task = g_task_new (NULL, cancellable, asynchronous_cancellation_callback, NULL);
g_cancellable_connect (cancellable, (GCallback) asynchronous_cancellation_cancelled, task, NULL);
g_signal_connect (task, "notify::completed", (GCallback) completed_cb, &notification_emitted);
run_task_id = g_idle_add (asynchronous_cancellation_run_task, task);
g_source_set_name_by_id (run_task_id, "[test_asynchronous_cancellation] run_task");
g_task_set_task_data (task, GUINT_TO_POINTER (run_task_id), NULL);
g_timeout_add (50, asynchronous_cancellation_cancel_task, task);
g_main_loop_run (loop);
g_assert_true (g_task_get_completed (task));
g_assert_true (notification_emitted);
g_object_unref (cancellable);
g_object_unref (task);
}
}
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
/* test_check_cancellable: cancellation overrides return value */
enum {
CANCEL_BEFORE = (1 << 1),
CANCEL_AFTER = (1 << 2),
CHECK_CANCELLABLE = (1 << 3)
};
#define NUM_CANCEL_TESTS (CANCEL_BEFORE | CANCEL_AFTER | CHECK_CANCELLABLE)
static void
cancel_callback (GObject *object,
GAsyncResult *result,
gpointer user_data)
{
int state = GPOINTER_TO_INT (user_data);
GTask *task;
GCancellable *cancellable;
GError *error = NULL;
g_assert (object == NULL);
g_assert (g_task_is_valid (result, object));
g_assert (g_async_result_get_user_data (result) == user_data);
task = G_TASK (result);
cancellable = g_task_get_cancellable (task);
g_assert (G_IS_CANCELLABLE (cancellable));
if (state & (CANCEL_BEFORE | CANCEL_AFTER))
g_assert (g_cancellable_is_cancelled (cancellable));
else
g_assert (!g_cancellable_is_cancelled (cancellable));
if (state & CHECK_CANCELLABLE)
g_assert (g_task_get_check_cancellable (task));
else
g_assert (!g_task_get_check_cancellable (task));
if (g_task_propagate_boolean (task, &error))
{
g_assert (!g_cancellable_is_cancelled (cancellable) ||
!g_task_get_check_cancellable (task));
}
else
{
g_assert (g_cancellable_is_cancelled (cancellable) &&
g_task_get_check_cancellable (task));
g_assert_error (error, G_IO_ERROR, G_IO_ERROR_CANCELLED);
g_error_free (error);
}
g_main_loop_quit (loop);
}
static void
test_check_cancellable (void)
{
GTask *task;
GCancellable *cancellable;
int state;
cancellable = g_cancellable_new ();
for (state = 0; state <= NUM_CANCEL_TESTS; state++)
{
task = g_task_new (NULL, cancellable, cancel_callback,
GINT_TO_POINTER (state));
g_task_set_check_cancellable (task, (state & CHECK_CANCELLABLE) != 0);
if (state & CANCEL_BEFORE)
g_cancellable_cancel (cancellable);
g_task_return_boolean (task, TRUE);
if (state & CANCEL_AFTER)
g_cancellable_cancel (cancellable);
g_main_loop_run (loop);
g_object_unref (task);
g_cancellable_reset (cancellable);
}
g_object_unref (cancellable);
}
/* test_return_if_cancelled */
static void
return_if_cancelled_callback (GObject *object,
GAsyncResult *result,
gpointer user_data)
{
GError *error = NULL;
g_assert (object == NULL);
g_assert (g_task_is_valid (result, object));
g_assert (g_async_result_get_user_data (result) == user_data);
g_assert (g_task_had_error (G_TASK (result)));
gtask: Add a GTask:completed property This can be used to query whether the task has completed, in the sense that it has had a result set on it, and has already – or will soon – invoke its callback function. Notifications for this property are emitted immediately after the task’s main callback, in the same main context as that callback. This allows for multiple bits of code to listen for completion of the GTask, which opens the door for blocking on cancellation of the GTask and improved handling of ‘pending’ behaviour. https://bugzilla.gnome.org/show_bug.cgi?id=743636
2014-08-21 20:04:04 +02:00
g_assert_false (g_task_get_completed (G_TASK (result)));
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
g_task_propagate_boolean (G_TASK (result), &error);
g_assert_error (error, G_IO_ERROR, G_IO_ERROR_CANCELLED);
g_clear_error (&error);
gio/tests/task: Ensure that g_task_had_error doesn't forget the error https://bugzilla.gnome.org/show_bug.cgi?id=764163
2016-03-24 18:59:54 +01:00
g_assert (g_task_had_error (G_TASK (result)));
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
g_main_loop_quit (loop);
}
static void
test_return_if_cancelled (void)
{
GTask *task;
GCancellable *cancellable;
gboolean cancelled;
gtask: Add a GTask:completed property This can be used to query whether the task has completed, in the sense that it has had a result set on it, and has already – or will soon – invoke its callback function. Notifications for this property are emitted immediately after the task’s main callback, in the same main context as that callback. This allows for multiple bits of code to listen for completion of the GTask, which opens the door for blocking on cancellation of the GTask and improved handling of ‘pending’ behaviour. https://bugzilla.gnome.org/show_bug.cgi?id=743636
2014-08-21 20:04:04 +02:00
gboolean notification_emitted = FALSE;
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
cancellable = g_cancellable_new ();
task = g_task_new (NULL, cancellable, return_if_cancelled_callback, NULL);
gtask: Add a GTask:completed property This can be used to query whether the task has completed, in the sense that it has had a result set on it, and has already – or will soon – invoke its callback function. Notifications for this property are emitted immediately after the task’s main callback, in the same main context as that callback. This allows for multiple bits of code to listen for completion of the GTask, which opens the door for blocking on cancellation of the GTask and improved handling of ‘pending’ behaviour. https://bugzilla.gnome.org/show_bug.cgi?id=743636
2014-08-21 20:04:04 +02:00
g_signal_connect (task, "notify::completed",
(GCallback) completed_cb, &notification_emitted);
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
g_cancellable_cancel (cancellable);
cancelled = g_task_return_error_if_cancelled (task);
g_assert (cancelled);
gtask: Add a GTask:completed property This can be used to query whether the task has completed, in the sense that it has had a result set on it, and has already – or will soon – invoke its callback function. Notifications for this property are emitted immediately after the task’s main callback, in the same main context as that callback. This allows for multiple bits of code to listen for completion of the GTask, which opens the door for blocking on cancellation of the GTask and improved handling of ‘pending’ behaviour. https://bugzilla.gnome.org/show_bug.cgi?id=743636
2014-08-21 20:04:04 +02:00
g_assert_false (notification_emitted);
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
g_main_loop_run (loop);
g_object_unref (task);
gtask: Add a GTask:completed property This can be used to query whether the task has completed, in the sense that it has had a result set on it, and has already – or will soon – invoke its callback function. Notifications for this property are emitted immediately after the task’s main callback, in the same main context as that callback. This allows for multiple bits of code to listen for completion of the GTask, which opens the door for blocking on cancellation of the GTask and improved handling of ‘pending’ behaviour. https://bugzilla.gnome.org/show_bug.cgi?id=743636
2014-08-21 20:04:04 +02:00
g_assert_true (notification_emitted);
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
g_cancellable_reset (cancellable);
gtask: Add a GTask:completed property This can be used to query whether the task has completed, in the sense that it has had a result set on it, and has already – or will soon – invoke its callback function. Notifications for this property are emitted immediately after the task’s main callback, in the same main context as that callback. This allows for multiple bits of code to listen for completion of the GTask, which opens the door for blocking on cancellation of the GTask and improved handling of ‘pending’ behaviour. https://bugzilla.gnome.org/show_bug.cgi?id=743636
2014-08-21 20:04:04 +02:00
notification_emitted = FALSE;
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
task = g_task_new (NULL, cancellable, return_if_cancelled_callback, NULL);
gtask: Add a GTask:completed property This can be used to query whether the task has completed, in the sense that it has had a result set on it, and has already – or will soon – invoke its callback function. Notifications for this property are emitted immediately after the task’s main callback, in the same main context as that callback. This allows for multiple bits of code to listen for completion of the GTask, which opens the door for blocking on cancellation of the GTask and improved handling of ‘pending’ behaviour. https://bugzilla.gnome.org/show_bug.cgi?id=743636
2014-08-21 20:04:04 +02:00
g_signal_connect (task, "notify::completed",
(GCallback) completed_cb, &notification_emitted);
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
g_task_set_check_cancellable (task, FALSE);
g_cancellable_cancel (cancellable);
cancelled = g_task_return_error_if_cancelled (task);
g_assert (cancelled);
gtask: Add a GTask:completed property This can be used to query whether the task has completed, in the sense that it has had a result set on it, and has already – or will soon – invoke its callback function. Notifications for this property are emitted immediately after the task’s main callback, in the same main context as that callback. This allows for multiple bits of code to listen for completion of the GTask, which opens the door for blocking on cancellation of the GTask and improved handling of ‘pending’ behaviour. https://bugzilla.gnome.org/show_bug.cgi?id=743636
2014-08-21 20:04:04 +02:00
g_assert_false (notification_emitted);
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
g_main_loop_run (loop);
g_object_unref (task);
gtask: Add a GTask:completed property This can be used to query whether the task has completed, in the sense that it has had a result set on it, and has already – or will soon – invoke its callback function. Notifications for this property are emitted immediately after the task’s main callback, in the same main context as that callback. This allows for multiple bits of code to listen for completion of the GTask, which opens the door for blocking on cancellation of the GTask and improved handling of ‘pending’ behaviour. https://bugzilla.gnome.org/show_bug.cgi?id=743636
2014-08-21 20:04:04 +02:00
g_assert_true (notification_emitted);
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
g_object_unref (cancellable);
}
/* test_run_in_thread */
gio/tests/task: fix a race condition in test_run_in_thread() When running a task in a thread, GTask may still be internally holding a ref on the task in that thread even after the callback is called in the original thread (depending on thread scheduling). Fix the test to handle that by using a weak notify that signals a GCond, and wait for that GCond from the main thread. (And add a corresponding check to test_return_on_cancel().) https://bugzilla.gnome.org/show_bug.cgi?id=705152
2013-08-15 17:30:59 +02:00
static GMutex run_in_thread_mutex;
static GCond run_in_thread_cond;
static void
task_weak_notify (gpointer user_data,
GObject *ex_task)
{
gboolean *weak_notify_ran = user_data;
g_mutex_lock (&run_in_thread_mutex);
tests: Fix non-atomic access to some shared variables And drop the `volatile` qualifier from the variables, as that doesn’t help with thread safety. Signed-off-by: Philip Withnall <pwithnall@endlessos.org> Helps: #600
2020-11-11 19:21:00 +01:00
g_atomic_int_set (weak_notify_ran, TRUE);
gio/tests/task: fix a race condition in test_run_in_thread() When running a task in a thread, GTask may still be internally holding a ref on the task in that thread even after the callback is called in the original thread (depending on thread scheduling). Fix the test to handle that by using a weak notify that signals a GCond, and wait for that GCond from the main thread. (And add a corresponding check to test_return_on_cancel().) https://bugzilla.gnome.org/show_bug.cgi?id=705152
2013-08-15 17:30:59 +02:00
g_cond_signal (&run_in_thread_cond);
g_mutex_unlock (&run_in_thread_mutex);
}
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
static void
run_in_thread_callback (GObject *object,
GAsyncResult *result,
gpointer user_data)
{
gboolean *done = user_data;
GError *error = NULL;
gssize ret;
g_assert (g_thread_self () == main_thread);
g_assert (object == NULL);
g_assert (g_task_is_valid (result, object));
g_assert (g_async_result_get_user_data (result) == user_data);
g_assert (!g_task_had_error (G_TASK (result)));
gtask: Add a GTask:completed property This can be used to query whether the task has completed, in the sense that it has had a result set on it, and has already – or will soon – invoke its callback function. Notifications for this property are emitted immediately after the task’s main callback, in the same main context as that callback. This allows for multiple bits of code to listen for completion of the GTask, which opens the door for blocking on cancellation of the GTask and improved handling of ‘pending’ behaviour. https://bugzilla.gnome.org/show_bug.cgi?id=743636
2014-08-21 20:04:04 +02:00
g_assert_false (g_task_get_completed (G_TASK (result)));
gtask: Add a g_task_set_name() method Similarly to g_source_set_name(), this sets a name on a GTask for debugging and profiling. Importantly, this name is propagated to the GSource for idle callbacks for the GTask, ending the glorious reign of `[gio] complete_in_idle_cb`. The name can be queried using g_task_get_name(). Locking is avoided by only allowing the name to be set before the GTask is used from another thread. Includes tests. Signed-off-by: Philip Withnall <withnall@endlessm.com>
2018-10-07 18:37:58 +02:00
g_assert_cmpstr (g_task_get_name (G_TASK (result)), ==, "test_run_in_thread name");
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
ret = g_task_propagate_int (G_TASK (result), &error);
g_assert_no_error (error);
g_assert_cmpint (ret, ==, magic);
gio/tests/task: Ensure that g_task_had_error doesn't forget the error https://bugzilla.gnome.org/show_bug.cgi?id=764163
2016-03-24 18:59:54 +01:00
g_assert (!g_task_had_error (G_TASK (result)));
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
*done = TRUE;
g_main_loop_quit (loop);
}
static void
run_in_thread_thread (GTask *task,
gpointer source_object,
gpointer task_data,
GCancellable *cancellable)
{
gboolean *thread_ran = task_data;
g_assert (source_object == g_task_get_source_object (task));
g_assert (task_data == g_task_get_task_data (task));
g_assert (cancellable == g_task_get_cancellable (task));
gtask: Add a GTask:completed property This can be used to query whether the task has completed, in the sense that it has had a result set on it, and has already – or will soon – invoke its callback function. Notifications for this property are emitted immediately after the task’s main callback, in the same main context as that callback. This allows for multiple bits of code to listen for completion of the GTask, which opens the door for blocking on cancellation of the GTask and improved handling of ‘pending’ behaviour. https://bugzilla.gnome.org/show_bug.cgi?id=743636
2014-08-21 20:04:04 +02:00
g_assert_false (g_task_get_completed (task));
gtask: Add a g_task_set_name() method Similarly to g_source_set_name(), this sets a name on a GTask for debugging and profiling. Importantly, this name is propagated to the GSource for idle callbacks for the GTask, ending the glorious reign of `[gio] complete_in_idle_cb`. The name can be queried using g_task_get_name(). Locking is avoided by only allowing the name to be set before the GTask is used from another thread. Includes tests. Signed-off-by: Philip Withnall <withnall@endlessm.com>
2018-10-07 18:37:58 +02:00
g_assert_cmpstr (g_task_get_name (task), ==, "test_run_in_thread name");
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
g_assert (g_thread_self () != main_thread);
gio/tests/task: fix a race condition in test_run_in_thread() When running a task in a thread, GTask may still be internally holding a ref on the task in that thread even after the callback is called in the original thread (depending on thread scheduling). Fix the test to handle that by using a weak notify that signals a GCond, and wait for that GCond from the main thread. (And add a corresponding check to test_return_on_cancel().) https://bugzilla.gnome.org/show_bug.cgi?id=705152
2013-08-15 17:30:59 +02:00
g_mutex_lock (&run_in_thread_mutex);
tests: Fix non-atomic access to some shared variables And drop the `volatile` qualifier from the variables, as that doesn’t help with thread safety. Signed-off-by: Philip Withnall <pwithnall@endlessos.org> Helps: #600
2020-11-11 19:21:00 +01:00
g_atomic_int_set (thread_ran, TRUE);
gio/tests/task: fix a race condition in test_run_in_thread() When running a task in a thread, GTask may still be internally holding a ref on the task in that thread even after the callback is called in the original thread (depending on thread scheduling). Fix the test to handle that by using a weak notify that signals a GCond, and wait for that GCond from the main thread. (And add a corresponding check to test_return_on_cancel().) https://bugzilla.gnome.org/show_bug.cgi?id=705152
2013-08-15 17:30:59 +02:00
g_cond_signal (&run_in_thread_cond);
g_mutex_unlock (&run_in_thread_mutex);
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
g_task_return_int (task, magic);
}
static void
test_run_in_thread (void)
{
GTask *task;
tests: Fix non-atomic access to some shared variables And drop the `volatile` qualifier from the variables, as that doesn’t help with thread safety. Signed-off-by: Philip Withnall <pwithnall@endlessos.org> Helps: #600
2020-11-11 19:21:00 +01:00
gboolean thread_ran = FALSE; /* (atomic) */
gboolean weak_notify_ran = FALSE; /* (atomic) */
gtask: Add a GTask:completed property This can be used to query whether the task has completed, in the sense that it has had a result set on it, and has already – or will soon – invoke its callback function. Notifications for this property are emitted immediately after the task’s main callback, in the same main context as that callback. This allows for multiple bits of code to listen for completion of the GTask, which opens the door for blocking on cancellation of the GTask and improved handling of ‘pending’ behaviour. https://bugzilla.gnome.org/show_bug.cgi?id=743636
2014-08-21 20:04:04 +02:00
gboolean notification_emitted = FALSE;
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
gboolean done = FALSE;
task = g_task_new (NULL, NULL, run_in_thread_callback, &done);
gtask: Add a g_task_set_name() method Similarly to g_source_set_name(), this sets a name on a GTask for debugging and profiling. Importantly, this name is propagated to the GSource for idle callbacks for the GTask, ending the glorious reign of `[gio] complete_in_idle_cb`. The name can be queried using g_task_get_name(). Locking is avoided by only allowing the name to be set before the GTask is used from another thread. Includes tests. Signed-off-by: Philip Withnall <withnall@endlessm.com>
2018-10-07 18:37:58 +02:00
g_task_set_name (task, "test_run_in_thread name");
gio/tests/task: fix a race condition in test_run_in_thread() When running a task in a thread, GTask may still be internally holding a ref on the task in that thread even after the callback is called in the original thread (depending on thread scheduling). Fix the test to handle that by using a weak notify that signals a GCond, and wait for that GCond from the main thread. (And add a corresponding check to test_return_on_cancel().) https://bugzilla.gnome.org/show_bug.cgi?id=705152
2013-08-15 17:30:59 +02:00
g_object_weak_ref (G_OBJECT (task), task_weak_notify, (gpointer)&weak_notify_ran);
gtask: Add a GTask:completed property This can be used to query whether the task has completed, in the sense that it has had a result set on it, and has already – or will soon – invoke its callback function. Notifications for this property are emitted immediately after the task’s main callback, in the same main context as that callback. This allows for multiple bits of code to listen for completion of the GTask, which opens the door for blocking on cancellation of the GTask and improved handling of ‘pending’ behaviour. https://bugzilla.gnome.org/show_bug.cgi?id=743636
2014-08-21 20:04:04 +02:00
g_signal_connect (task, "notify::completed",
(GCallback) completed_cb, &notification_emitted);
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
g_task_set_task_data (task, (gpointer)&thread_ran, NULL);
g_task_run_in_thread (task, run_in_thread_thread);
gio/tests/task: fix a race condition in test_run_in_thread() When running a task in a thread, GTask may still be internally holding a ref on the task in that thread even after the callback is called in the original thread (depending on thread scheduling). Fix the test to handle that by using a weak notify that signals a GCond, and wait for that GCond from the main thread. (And add a corresponding check to test_return_on_cancel().) https://bugzilla.gnome.org/show_bug.cgi?id=705152
2013-08-15 17:30:59 +02:00
g_mutex_lock (&run_in_thread_mutex);
tests: Fix non-atomic access to some shared variables And drop the `volatile` qualifier from the variables, as that doesn’t help with thread safety. Signed-off-by: Philip Withnall <pwithnall@endlessos.org> Helps: #600
2020-11-11 19:21:00 +01:00
while (!g_atomic_int_get (&thread_ran))
gio/tests/task: fix a race condition in test_run_in_thread() When running a task in a thread, GTask may still be internally holding a ref on the task in that thread even after the callback is called in the original thread (depending on thread scheduling). Fix the test to handle that by using a weak notify that signals a GCond, and wait for that GCond from the main thread. (And add a corresponding check to test_return_on_cancel().) https://bugzilla.gnome.org/show_bug.cgi?id=705152
2013-08-15 17:30:59 +02:00
g_cond_wait (&run_in_thread_cond, &run_in_thread_mutex);
g_mutex_unlock (&run_in_thread_mutex);
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
g_assert (done == FALSE);
tests: Fix non-atomic access to some shared variables And drop the `volatile` qualifier from the variables, as that doesn’t help with thread safety. Signed-off-by: Philip Withnall <pwithnall@endlessos.org> Helps: #600
2020-11-11 19:21:00 +01:00
g_assert_false (g_atomic_int_get (&weak_notify_ran));
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
g_main_loop_run (loop);
g_assert (done == TRUE);
gtask: Add a GTask:completed property This can be used to query whether the task has completed, in the sense that it has had a result set on it, and has already – or will soon – invoke its callback function. Notifications for this property are emitted immediately after the task’s main callback, in the same main context as that callback. This allows for multiple bits of code to listen for completion of the GTask, which opens the door for blocking on cancellation of the GTask and improved handling of ‘pending’ behaviour. https://bugzilla.gnome.org/show_bug.cgi?id=743636
2014-08-21 20:04:04 +02:00
g_assert_true (notification_emitted);
gio/tests/task: fix a race condition in test_run_in_thread() When running a task in a thread, GTask may still be internally holding a ref on the task in that thread even after the callback is called in the original thread (depending on thread scheduling). Fix the test to handle that by using a weak notify that signals a GCond, and wait for that GCond from the main thread. (And add a corresponding check to test_return_on_cancel().) https://bugzilla.gnome.org/show_bug.cgi?id=705152
2013-08-15 17:30:59 +02:00
gtask: Add a g_task_set_name() method Similarly to g_source_set_name(), this sets a name on a GTask for debugging and profiling. Importantly, this name is propagated to the GSource for idle callbacks for the GTask, ending the glorious reign of `[gio] complete_in_idle_cb`. The name can be queried using g_task_get_name(). Locking is avoided by only allowing the name to be set before the GTask is used from another thread. Includes tests. Signed-off-by: Philip Withnall <withnall@endlessm.com>
2018-10-07 18:37:58 +02:00
g_assert_cmpstr (g_task_get_name (task), ==, "test_run_in_thread name");
g_object_unref (task);
gio/tests/task: fix a race condition in test_run_in_thread() When running a task in a thread, GTask may still be internally holding a ref on the task in that thread even after the callback is called in the original thread (depending on thread scheduling). Fix the test to handle that by using a weak notify that signals a GCond, and wait for that GCond from the main thread. (And add a corresponding check to test_return_on_cancel().) https://bugzilla.gnome.org/show_bug.cgi?id=705152
2013-08-15 17:30:59 +02:00
g_mutex_lock (&run_in_thread_mutex);
tests: Fix non-atomic access to some shared variables And drop the `volatile` qualifier from the variables, as that doesn’t help with thread safety. Signed-off-by: Philip Withnall <pwithnall@endlessos.org> Helps: #600
2020-11-11 19:21:00 +01:00
while (!g_atomic_int_get (&weak_notify_ran))
gio/tests/task: fix a race condition in test_run_in_thread() When running a task in a thread, GTask may still be internally holding a ref on the task in that thread even after the callback is called in the original thread (depending on thread scheduling). Fix the test to handle that by using a weak notify that signals a GCond, and wait for that GCond from the main thread. (And add a corresponding check to test_return_on_cancel().) https://bugzilla.gnome.org/show_bug.cgi?id=705152
2013-08-15 17:30:59 +02:00
g_cond_wait (&run_in_thread_cond, &run_in_thread_mutex);
g_mutex_unlock (&run_in_thread_mutex);
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
}
/* test_run_in_thread_sync */
static void
run_in_thread_sync_callback (GObject *object,
GAsyncResult *result,
gpointer user_data)
{
/* g_task_run_in_thread_sync() does not invoke the task's callback */
g_assert_not_reached ();
}
static void
run_in_thread_sync_thread (GTask *task,
gpointer source_object,
gpointer task_data,
GCancellable *cancellable)
{
gboolean *thread_ran = task_data;
g_assert (source_object == g_task_get_source_object (task));
g_assert (task_data == g_task_get_task_data (task));
g_assert (cancellable == g_task_get_cancellable (task));
gtask: Add a GTask:completed property This can be used to query whether the task has completed, in the sense that it has had a result set on it, and has already – or will soon – invoke its callback function. Notifications for this property are emitted immediately after the task’s main callback, in the same main context as that callback. This allows for multiple bits of code to listen for completion of the GTask, which opens the door for blocking on cancellation of the GTask and improved handling of ‘pending’ behaviour. https://bugzilla.gnome.org/show_bug.cgi?id=743636
2014-08-21 20:04:04 +02:00
g_assert_false (g_task_get_completed (task));
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
g_assert (g_thread_self () != main_thread);
tests: Fix non-atomic access to some shared variables And drop the `volatile` qualifier from the variables, as that doesn’t help with thread safety. Signed-off-by: Philip Withnall <pwithnall@endlessos.org> Helps: #600
2020-11-11 19:21:00 +01:00
g_atomic_int_set (thread_ran, TRUE);
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
g_task_return_int (task, magic);
}
static void
test_run_in_thread_sync (void)
{
GTask *task;
gboolean thread_ran = FALSE;
gssize ret;
gtask: Add a GTask:completed property This can be used to query whether the task has completed, in the sense that it has had a result set on it, and has already – or will soon – invoke its callback function. Notifications for this property are emitted immediately after the task’s main callback, in the same main context as that callback. This allows for multiple bits of code to listen for completion of the GTask, which opens the door for blocking on cancellation of the GTask and improved handling of ‘pending’ behaviour. https://bugzilla.gnome.org/show_bug.cgi?id=743636
2014-08-21 20:04:04 +02:00
gboolean notification_emitted = FALSE;
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
GError *error = NULL;
task = g_task_new (NULL, NULL, run_in_thread_sync_callback, NULL);
gtask: Add a GTask:completed property This can be used to query whether the task has completed, in the sense that it has had a result set on it, and has already – or will soon – invoke its callback function. Notifications for this property are emitted immediately after the task’s main callback, in the same main context as that callback. This allows for multiple bits of code to listen for completion of the GTask, which opens the door for blocking on cancellation of the GTask and improved handling of ‘pending’ behaviour. https://bugzilla.gnome.org/show_bug.cgi?id=743636
2014-08-21 20:04:04 +02:00
g_signal_connect (task, "notify::completed",
(GCallback) completed_cb,
&notification_emitted);
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
g_task_set_task_data (task, &thread_ran, NULL);
g_task_run_in_thread_sync (task, run_in_thread_sync_thread);
tests: Fix non-atomic access to some shared variables And drop the `volatile` qualifier from the variables, as that doesn’t help with thread safety. Signed-off-by: Philip Withnall <pwithnall@endlessos.org> Helps: #600
2020-11-11 19:21:00 +01:00
g_assert_true (g_atomic_int_get (&thread_ran));
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
g_assert (task != NULL);
g_assert (!g_task_had_error (task));
gtask: Add a GTask:completed property This can be used to query whether the task has completed, in the sense that it has had a result set on it, and has already – or will soon – invoke its callback function. Notifications for this property are emitted immediately after the task’s main callback, in the same main context as that callback. This allows for multiple bits of code to listen for completion of the GTask, which opens the door for blocking on cancellation of the GTask and improved handling of ‘pending’ behaviour. https://bugzilla.gnome.org/show_bug.cgi?id=743636
2014-08-21 20:04:04 +02:00
g_assert_true (g_task_get_completed (task));
g_assert_true (notification_emitted);
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
ret = g_task_propagate_int (task, &error);
g_assert_no_error (error);
g_assert_cmpint (ret, ==, magic);
gio/tests/task: Ensure that g_task_had_error doesn't forget the error https://bugzilla.gnome.org/show_bug.cgi?id=764163
2016-03-24 18:59:54 +01:00
g_assert (!g_task_had_error (task));
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
g_object_unref (task);
}
/* test_run_in_thread_priority */
static GMutex fake_task_mutex, last_fake_task_mutex;
static gint sequence_number = 0;
static void
quit_main_loop_callback (GObject *object,
GAsyncResult *result,
gpointer user_data)
{
GError *error = NULL;
gboolean ret;
g_assert (g_thread_self () == main_thread);
g_assert (object == NULL);
g_assert (g_task_is_valid (result, object));
g_assert (g_async_result_get_user_data (result) == user_data);
g_assert (!g_task_had_error (G_TASK (result)));
gtask: Add a GTask:completed property This can be used to query whether the task has completed, in the sense that it has had a result set on it, and has already – or will soon – invoke its callback function. Notifications for this property are emitted immediately after the task’s main callback, in the same main context as that callback. This allows for multiple bits of code to listen for completion of the GTask, which opens the door for blocking on cancellation of the GTask and improved handling of ‘pending’ behaviour. https://bugzilla.gnome.org/show_bug.cgi?id=743636
2014-08-21 20:04:04 +02:00
g_assert_false (g_task_get_completed (G_TASK (result)));
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
ret = g_task_propagate_boolean (G_TASK (result), &error);
g_assert_no_error (error);
g_assert_cmpint (ret, ==, TRUE);
gio/tests/task: Ensure that g_task_had_error doesn't forget the error https://bugzilla.gnome.org/show_bug.cgi?id=764163
2016-03-24 18:59:54 +01:00
g_assert (!g_task_had_error (G_TASK (result)));
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
g_main_loop_quit (loop);
}
static void
set_sequence_number_thread (GTask *task,
gpointer source_object,
gpointer task_data,
GCancellable *cancellable)
{
gint *seq_no_p = task_data;
*seq_no_p = ++sequence_number;
g_task_return_boolean (task, TRUE);
}
static void
fake_task_thread (GTask *task,
gpointer source_object,
gpointer task_data,
GCancellable *cancellable)
{
GMutex *mutex = task_data;
g_mutex_lock (mutex);
g_mutex_unlock (mutex);
g_task_return_boolean (task, TRUE);
}
gtask: don't deadlock when tasks block on other tasks If tasks block waiting for other tasks to complete then the system can end up starved for threads. Avoid this by bumping up max-threads in that case. This also reverts 7b1f8c58 and reverts max-threads for GTask's GThreadPool back to 10. https://bugzilla.gnome.org/show_bug.cgi?id=687223
2012-12-15 17:44:59 +01:00
#define G_TASK_THREAD_POOL_SIZE 10
static int fake_tasks_running;
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
static void
gtask: don't deadlock when tasks block on other tasks If tasks block waiting for other tasks to complete then the system can end up starved for threads. Avoid this by bumping up max-threads in that case. This also reverts 7b1f8c58 and reverts max-threads for GTask's GThreadPool back to 10. https://bugzilla.gnome.org/show_bug.cgi?id=687223
2012-12-15 17:44:59 +01:00
fake_task_callback (GObject *source,
GAsyncResult *result,
gpointer user_data)
{
if (--fake_tasks_running == 0)
g_main_loop_quit (loop);
}
static void
clog_up_thread_pool (void)
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
{
GTask *task;
int i;
g_thread_pool_stop_unused_threads ();
g_mutex_lock (&fake_task_mutex);
for (i = 0; i < G_TASK_THREAD_POOL_SIZE - 1; i++)
{
gtask: don't deadlock when tasks block on other tasks If tasks block waiting for other tasks to complete then the system can end up starved for threads. Avoid this by bumping up max-threads in that case. This also reverts 7b1f8c58 and reverts max-threads for GTask's GThreadPool back to 10. https://bugzilla.gnome.org/show_bug.cgi?id=687223
2012-12-15 17:44:59 +01:00
task = g_task_new (NULL, NULL, fake_task_callback, NULL);
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
g_task_set_task_data (task, &fake_task_mutex, NULL);
g_assert_cmpint (g_task_get_priority (task), ==, G_PRIORITY_DEFAULT);
g_task_set_priority (task, G_PRIORITY_HIGH * 2);
g_assert_cmpint (g_task_get_priority (task), ==, G_PRIORITY_HIGH * 2);
g_task_run_in_thread (task, fake_task_thread);
g_object_unref (task);
gtask: don't deadlock when tasks block on other tasks If tasks block waiting for other tasks to complete then the system can end up starved for threads. Avoid this by bumping up max-threads in that case. This also reverts 7b1f8c58 and reverts max-threads for GTask's GThreadPool back to 10. https://bugzilla.gnome.org/show_bug.cgi?id=687223
2012-12-15 17:44:59 +01:00
fake_tasks_running++;
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
}
g_mutex_lock (&last_fake_task_mutex);
task = g_task_new (NULL, NULL, NULL, NULL);
g_task_set_task_data (task, &last_fake_task_mutex, NULL);
g_task_set_priority (task, G_PRIORITY_HIGH * 2);
g_task_run_in_thread (task, fake_task_thread);
g_object_unref (task);
gtask: don't deadlock when tasks block on other tasks If tasks block waiting for other tasks to complete then the system can end up starved for threads. Avoid this by bumping up max-threads in that case. This also reverts 7b1f8c58 and reverts max-threads for GTask's GThreadPool back to 10. https://bugzilla.gnome.org/show_bug.cgi?id=687223
2012-12-15 17:44:59 +01:00
}
static void
unclog_thread_pool (void)
{
g_mutex_unlock (&fake_task_mutex);
g_main_loop_run (loop);
}
static void
test_run_in_thread_priority (void)
{
GTask *task;
GCancellable *cancellable;
int seq_a, seq_b, seq_c, seq_d;
clog_up_thread_pool ();
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
/* Queue three more tasks that we'll arrange to have run serially */
task = g_task_new (NULL, NULL, NULL, NULL);
g_task_set_task_data (task, &seq_a, NULL);
g_task_run_in_thread (task, set_sequence_number_thread);
g_object_unref (task);
task = g_task_new (NULL, NULL, quit_main_loop_callback, NULL);
g_task_set_task_data (task, &seq_b, NULL);
g_task_set_priority (task, G_PRIORITY_LOW);
g_task_run_in_thread (task, set_sequence_number_thread);
g_object_unref (task);
task = g_task_new (NULL, NULL, NULL, NULL);
g_task_set_task_data (task, &seq_c, NULL);
g_task_set_priority (task, G_PRIORITY_HIGH);
g_task_run_in_thread (task, set_sequence_number_thread);
g_object_unref (task);
cancellable = g_cancellable_new ();
task = g_task_new (NULL, cancellable, NULL, NULL);
g_task_set_task_data (task, &seq_d, NULL);
g_task_run_in_thread (task, set_sequence_number_thread);
g_cancellable_cancel (cancellable);
g_object_unref (cancellable);
g_object_unref (task);
/* Let the last fake task complete; the four other tasks will then
* complete serially, in the order D, C, A, B, and B will quit the
* main loop.
*/
g_mutex_unlock (&last_fake_task_mutex);
g_main_loop_run (loop);
g_assert_cmpint (seq_d, ==, 1);
g_assert_cmpint (seq_c, ==, 2);
g_assert_cmpint (seq_a, ==, 3);
g_assert_cmpint (seq_b, ==, 4);
gtask: don't deadlock when tasks block on other tasks If tasks block waiting for other tasks to complete then the system can end up starved for threads. Avoid this by bumping up max-threads in that case. This also reverts 7b1f8c58 and reverts max-threads for GTask's GThreadPool back to 10. https://bugzilla.gnome.org/show_bug.cgi?id=687223
2012-12-15 17:44:59 +01:00
unclog_thread_pool ();
}
/* test_run_in_thread_nested: task threads that block waiting on
* other task threads will not cause the thread pool to starve.
*/
static void
run_nested_task_thread (GTask *task,
gpointer source_object,
gpointer task_data,
GCancellable *cancellable)
{
GTask *nested;
int *nested_tasks_left = task_data;
if ((*nested_tasks_left)--)
{
nested = g_task_new (NULL, NULL, NULL, NULL);
g_task_set_task_data (nested, nested_tasks_left, NULL);
g_task_run_in_thread_sync (nested, run_nested_task_thread);
g_object_unref (nested);
}
g_task_return_boolean (task, TRUE);
}
static void
test_run_in_thread_nested (void)
{
GTask *task;
int nested_tasks_left = 2;
clog_up_thread_pool ();
task = g_task_new (NULL, NULL, quit_main_loop_callback, NULL);
g_task_set_task_data (task, &nested_tasks_left, NULL);
g_task_run_in_thread (task, run_nested_task_thread);
g_object_unref (task);
g_mutex_unlock (&last_fake_task_mutex);
g_main_loop_run (loop);
unclog_thread_pool ();
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
}
gtask: remove hardcoded GTask thread-pool size GTask used a 10-thread thread pool for g_task_run_in_thread() / g_task_run_in_thread_sync(), but this ran into problems when task threads blocked waiting for another g_task_run_in_thread_sync() operation to complete. Previously there was a workaround for this, by bumping up the thread limit when that case was detected, but deadlocks could still happen if there were non-GTask threads involved. (Eg, task A sends a message to thread X and waits for a response, but thread X needs to complete task B in a thread before returning the response to task A.) So, allow GTask's thread pool to be expanded dynamically, by watching it from the glib worker thread, and growing it (at an exponentially-decreasing rate) if too much time passes without any tasks completing. This should solve the deadlocking problems without causing sudden breakage in apps that assume they can queue huge numbers of tasks at once without consequences. https://bugzilla.gnome.org/show_bug.cgi?id=687223
2015-03-09 21:33:16 +01:00
/* test_run_in_thread_overflow: if you queue lots and lots and lots of
* tasks, they won't all run at once.
*/
static GMutex overflow_mutex;
tests: Fix data race in task test Ensure that all tasks have already completed before accessing buf array from main thread to avoid conflicting data access from multiple threads.
2019-02-27 01:00:00 +01:00
static guint overflow_completed;
gtask: remove hardcoded GTask thread-pool size GTask used a 10-thread thread pool for g_task_run_in_thread() / g_task_run_in_thread_sync(), but this ran into problems when task threads blocked waiting for another g_task_run_in_thread_sync() operation to complete. Previously there was a workaround for this, by bumping up the thread limit when that case was detected, but deadlocks could still happen if there were non-GTask threads involved. (Eg, task A sends a message to thread X and waits for a response, but thread X needs to complete task B in a thread before returning the response to task A.) So, allow GTask's thread pool to be expanded dynamically, by watching it from the glib worker thread, and growing it (at an exponentially-decreasing rate) if too much time passes without any tasks completing. This should solve the deadlocking problems without causing sudden breakage in apps that assume they can queue huge numbers of tasks at once without consequences. https://bugzilla.gnome.org/show_bug.cgi?id=687223
2015-03-09 21:33:16 +01:00
static void
run_overflow_task_thread (GTask *task,
gpointer source_object,
gpointer task_data,
GCancellable *cancellable)
{
gchar *result = task_data;
if (g_task_return_error_if_cancelled (task))
{
*result = 'X';
}
tests: Fix data race in task test Ensure that all tasks have already completed before accessing buf array from main thread to avoid conflicting data access from multiple threads.
2019-02-27 01:00:00 +01:00
else
{
/* Block until the main thread is ready. */
g_mutex_lock (&overflow_mutex);
g_mutex_unlock (&overflow_mutex);
gtask: remove hardcoded GTask thread-pool size GTask used a 10-thread thread pool for g_task_run_in_thread() / g_task_run_in_thread_sync(), but this ran into problems when task threads blocked waiting for another g_task_run_in_thread_sync() operation to complete. Previously there was a workaround for this, by bumping up the thread limit when that case was detected, but deadlocks could still happen if there were non-GTask threads involved. (Eg, task A sends a message to thread X and waits for a response, but thread X needs to complete task B in a thread before returning the response to task A.) So, allow GTask's thread pool to be expanded dynamically, by watching it from the glib worker thread, and growing it (at an exponentially-decreasing rate) if too much time passes without any tasks completing. This should solve the deadlocking problems without causing sudden breakage in apps that assume they can queue huge numbers of tasks at once without consequences. https://bugzilla.gnome.org/show_bug.cgi?id=687223
2015-03-09 21:33:16 +01:00
tests: Fix data race in task test Ensure that all tasks have already completed before accessing buf array from main thread to avoid conflicting data access from multiple threads.
2019-02-27 01:00:00 +01:00
*result = '.';
gtask: remove hardcoded GTask thread-pool size GTask used a 10-thread thread pool for g_task_run_in_thread() / g_task_run_in_thread_sync(), but this ran into problems when task threads blocked waiting for another g_task_run_in_thread_sync() operation to complete. Previously there was a workaround for this, by bumping up the thread limit when that case was detected, but deadlocks could still happen if there were non-GTask threads involved. (Eg, task A sends a message to thread X and waits for a response, but thread X needs to complete task B in a thread before returning the response to task A.) So, allow GTask's thread pool to be expanded dynamically, by watching it from the glib worker thread, and growing it (at an exponentially-decreasing rate) if too much time passes without any tasks completing. This should solve the deadlocking problems without causing sudden breakage in apps that assume they can queue huge numbers of tasks at once without consequences. https://bugzilla.gnome.org/show_bug.cgi?id=687223
2015-03-09 21:33:16 +01:00
tests: Fix data race in task test Ensure that all tasks have already completed before accessing buf array from main thread to avoid conflicting data access from multiple threads.
2019-02-27 01:00:00 +01:00
g_task_return_boolean (task, TRUE);
}
gtask: remove hardcoded GTask thread-pool size GTask used a 10-thread thread pool for g_task_run_in_thread() / g_task_run_in_thread_sync(), but this ran into problems when task threads blocked waiting for another g_task_run_in_thread_sync() operation to complete. Previously there was a workaround for this, by bumping up the thread limit when that case was detected, but deadlocks could still happen if there were non-GTask threads involved. (Eg, task A sends a message to thread X and waits for a response, but thread X needs to complete task B in a thread before returning the response to task A.) So, allow GTask's thread pool to be expanded dynamically, by watching it from the glib worker thread, and growing it (at an exponentially-decreasing rate) if too much time passes without any tasks completing. This should solve the deadlocking problems without causing sudden breakage in apps that assume they can queue huge numbers of tasks at once without consequences. https://bugzilla.gnome.org/show_bug.cgi?id=687223
2015-03-09 21:33:16 +01:00
tests: Fix data race in task test Ensure that all tasks have already completed before accessing buf array from main thread to avoid conflicting data access from multiple threads.
2019-02-27 01:00:00 +01:00
g_atomic_int_inc (&overflow_completed);
gtask: remove hardcoded GTask thread-pool size GTask used a 10-thread thread pool for g_task_run_in_thread() / g_task_run_in_thread_sync(), but this ran into problems when task threads blocked waiting for another g_task_run_in_thread_sync() operation to complete. Previously there was a workaround for this, by bumping up the thread limit when that case was detected, but deadlocks could still happen if there were non-GTask threads involved. (Eg, task A sends a message to thread X and waits for a response, but thread X needs to complete task B in a thread before returning the response to task A.) So, allow GTask's thread pool to be expanded dynamically, by watching it from the glib worker thread, and growing it (at an exponentially-decreasing rate) if too much time passes without any tasks completing. This should solve the deadlocking problems without causing sudden breakage in apps that assume they can queue huge numbers of tasks at once without consequences. https://bugzilla.gnome.org/show_bug.cgi?id=687223
2015-03-09 21:33:16 +01:00
}
#define NUM_OVERFLOW_TASKS 1024
static void
test_run_in_thread_overflow (void)
{
GCancellable *cancellable;
GTask *task;
gchar buf[NUM_OVERFLOW_TASKS + 1];
gint i;
/* Queue way too many tasks and then sleep for a bit. The first 10
* tasks will be dispatched to threads and will then block on
* overflow_mutex, so more threads will be created while this thread
* is sleeping. Then we cancel the cancellable, unlock the mutex,
* wait for all of the tasks to complete, and make sure that we got
* the behavior we expected.
*/
memset (buf, 0, sizeof (buf));
cancellable = g_cancellable_new ();
g_mutex_lock (&overflow_mutex);
for (i = 0; i < NUM_OVERFLOW_TASKS; i++)
{
task = g_task_new (NULL, cancellable, NULL, NULL);
g_task_set_task_data (task, buf + i, NULL);
g_task_run_in_thread (task, run_overflow_task_thread);
g_object_unref (task);
}
if (g_test_slow ())
g_usleep (5000000); /* 5 s */
else
g_usleep (500000); /* 0.5 s */
g_cancellable_cancel (cancellable);
g_object_unref (cancellable);
g_mutex_unlock (&overflow_mutex);
/* Wait for all tasks to complete. */
tests: Fix data race in task test Ensure that all tasks have already completed before accessing buf array from main thread to avoid conflicting data access from multiple threads.
2019-02-27 01:00:00 +01:00
while (g_atomic_int_get (&overflow_completed) != NUM_OVERFLOW_TASKS)
gtask: remove hardcoded GTask thread-pool size GTask used a 10-thread thread pool for g_task_run_in_thread() / g_task_run_in_thread_sync(), but this ran into problems when task threads blocked waiting for another g_task_run_in_thread_sync() operation to complete. Previously there was a workaround for this, by bumping up the thread limit when that case was detected, but deadlocks could still happen if there were non-GTask threads involved. (Eg, task A sends a message to thread X and waits for a response, but thread X needs to complete task B in a thread before returning the response to task A.) So, allow GTask's thread pool to be expanded dynamically, by watching it from the glib worker thread, and growing it (at an exponentially-decreasing rate) if too much time passes without any tasks completing. This should solve the deadlocking problems without causing sudden breakage in apps that assume they can queue huge numbers of tasks at once without consequences. https://bugzilla.gnome.org/show_bug.cgi?id=687223
2015-03-09 21:33:16 +01:00
g_usleep (1000);
tests: Fix data race in task test Ensure that all tasks have already completed before accessing buf array from main thread to avoid conflicting data access from multiple threads.
2019-02-27 01:00:00 +01:00
g_assert_cmpint (strlen (buf), ==, NUM_OVERFLOW_TASKS);
gtask: remove hardcoded GTask thread-pool size GTask used a 10-thread thread pool for g_task_run_in_thread() / g_task_run_in_thread_sync(), but this ran into problems when task threads blocked waiting for another g_task_run_in_thread_sync() operation to complete. Previously there was a workaround for this, by bumping up the thread limit when that case was detected, but deadlocks could still happen if there were non-GTask threads involved. (Eg, task A sends a message to thread X and waits for a response, but thread X needs to complete task B in a thread before returning the response to task A.) So, allow GTask's thread pool to be expanded dynamically, by watching it from the glib worker thread, and growing it (at an exponentially-decreasing rate) if too much time passes without any tasks completing. This should solve the deadlocking problems without causing sudden breakage in apps that assume they can queue huge numbers of tasks at once without consequences. https://bugzilla.gnome.org/show_bug.cgi?id=687223
2015-03-09 21:33:16 +01:00
i = strspn (buf, ".");
/* Given the sleep times above, i should be 14 for normal, 40 for
* slow. But if the machine is too slow/busy then the scheduling
* might get messed up and we'll get more or fewer threads than
* expected. But there are limits to how messed up it could
* plausibly get (and we hope that if gtask is actually broken then
* it will exceed those limits).
*/
g_assert_cmpint (i, >=, 10);
if (g_test_slow ())
g_assert_cmpint (i, <, 50);
else
g_assert_cmpint (i, <, 20);
g_assert_cmpint (i + strspn (buf + i, "X"), ==, NUM_OVERFLOW_TASKS);
}
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
/* test_return_on_cancel */
GMutex roc_init_mutex, roc_finish_mutex;
GCond roc_init_cond, roc_finish_cond;
typedef enum {
THREAD_STARTING,
THREAD_RUNNING,
THREAD_CANCELLED,
THREAD_COMPLETED
} ThreadState;
static void
return_on_cancel_callback (GObject *object,
GAsyncResult *result,
gpointer user_data)
{
gboolean *callback_ran = user_data;
GError *error = NULL;
gssize ret;
g_assert (g_thread_self () == main_thread);
g_assert (object == NULL);
g_assert (g_task_is_valid (result, object));
g_assert (g_async_result_get_user_data (result) == user_data);
g_assert (g_task_had_error (G_TASK (result)));
gtask: Add a GTask:completed property This can be used to query whether the task has completed, in the sense that it has had a result set on it, and has already – or will soon – invoke its callback function. Notifications for this property are emitted immediately after the task’s main callback, in the same main context as that callback. This allows for multiple bits of code to listen for completion of the GTask, which opens the door for blocking on cancellation of the GTask and improved handling of ‘pending’ behaviour. https://bugzilla.gnome.org/show_bug.cgi?id=743636
2014-08-21 20:04:04 +02:00
g_assert_false (g_task_get_completed (G_TASK (result)));
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
ret = g_task_propagate_int (G_TASK (result), &error);
g_assert_error (error, G_IO_ERROR, G_IO_ERROR_CANCELLED);
g_clear_error (&error);
g_assert_cmpint (ret, ==, -1);
gio/tests/task: Ensure that g_task_had_error doesn't forget the error https://bugzilla.gnome.org/show_bug.cgi?id=764163
2016-03-24 18:59:54 +01:00
g_assert (g_task_had_error (G_TASK (result)));
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
*callback_ran = TRUE;
g_main_loop_quit (loop);
}
static void
return_on_cancel_thread (GTask *task,
gpointer source_object,
gpointer task_data,
GCancellable *cancellable)
{
ThreadState *state = task_data;
g_assert (source_object == g_task_get_source_object (task));
g_assert (task_data == g_task_get_task_data (task));
g_assert (cancellable == g_task_get_cancellable (task));
g_assert (g_thread_self () != main_thread);
g_mutex_lock (&roc_init_mutex);
*state = THREAD_RUNNING;
g_cond_signal (&roc_init_cond);
g_mutex_unlock (&roc_init_mutex);
g_mutex_lock (&roc_finish_mutex);
if (!g_task_get_return_on_cancel (task) ||
g_task_set_return_on_cancel (task, FALSE))
{
*state = THREAD_COMPLETED;
g_task_return_int (task, magic);
}
else
*state = THREAD_CANCELLED;
g_cond_signal (&roc_finish_cond);
g_mutex_unlock (&roc_finish_mutex);
}
static void
test_return_on_cancel (void)
{
GTask *task;
GCancellable *cancellable;
tests: Fix non-atomic access to some shared variables And drop the `volatile` qualifier from the variables, as that doesn’t help with thread safety. Signed-off-by: Philip Withnall <pwithnall@endlessos.org> Helps: #600
2020-11-11 19:21:00 +01:00
ThreadState thread_state; /* (atomic) */
gboolean weak_notify_ran = FALSE; /* (atomic) */
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
gboolean callback_ran;
gtask: Add a GTask:completed property This can be used to query whether the task has completed, in the sense that it has had a result set on it, and has already – or will soon – invoke its callback function. Notifications for this property are emitted immediately after the task’s main callback, in the same main context as that callback. This allows for multiple bits of code to listen for completion of the GTask, which opens the door for blocking on cancellation of the GTask and improved handling of ‘pending’ behaviour. https://bugzilla.gnome.org/show_bug.cgi?id=743636
2014-08-21 20:04:04 +02:00
gboolean notification_emitted = FALSE;
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
cancellable = g_cancellable_new ();
/* If return-on-cancel is FALSE (default), the task does not return
* early.
*/
callback_ran = FALSE;
tests: Fix non-atomic access to some shared variables And drop the `volatile` qualifier from the variables, as that doesn’t help with thread safety. Signed-off-by: Philip Withnall <pwithnall@endlessos.org> Helps: #600
2020-11-11 19:21:00 +01:00
g_atomic_int_set (&thread_state, THREAD_STARTING);
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
task = g_task_new (NULL, cancellable, return_on_cancel_callback, &callback_ran);
gtask: Add a GTask:completed property This can be used to query whether the task has completed, in the sense that it has had a result set on it, and has already – or will soon – invoke its callback function. Notifications for this property are emitted immediately after the task’s main callback, in the same main context as that callback. This allows for multiple bits of code to listen for completion of the GTask, which opens the door for blocking on cancellation of the GTask and improved handling of ‘pending’ behaviour. https://bugzilla.gnome.org/show_bug.cgi?id=743636
2014-08-21 20:04:04 +02:00
g_signal_connect (task, "notify::completed",
(GCallback) completed_cb, &notification_emitted);
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
g_task_set_task_data (task, (gpointer)&thread_state, NULL);
g_mutex_lock (&roc_init_mutex);
g_mutex_lock (&roc_finish_mutex);
g_task_run_in_thread (task, return_on_cancel_thread);
g_object_unref (task);
tests: Fix non-atomic access to some shared variables And drop the `volatile` qualifier from the variables, as that doesn’t help with thread safety. Signed-off-by: Philip Withnall <pwithnall@endlessos.org> Helps: #600
2020-11-11 19:21:00 +01:00
while (g_atomic_int_get (&thread_state) == THREAD_STARTING)
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
g_cond_wait (&roc_init_cond, &roc_init_mutex);
g_mutex_unlock (&roc_init_mutex);
tests: Fix non-atomic access to some shared variables And drop the `volatile` qualifier from the variables, as that doesn’t help with thread safety. Signed-off-by: Philip Withnall <pwithnall@endlessos.org> Helps: #600
2020-11-11 19:21:00 +01:00
g_assert_cmpint (g_atomic_int_get (&thread_state), ==, THREAD_RUNNING);
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
g_assert (callback_ran == FALSE);
g_cancellable_cancel (cancellable);
g_mutex_unlock (&roc_finish_mutex);
g_main_loop_run (loop);
tests: Fix non-atomic access to some shared variables And drop the `volatile` qualifier from the variables, as that doesn’t help with thread safety. Signed-off-by: Philip Withnall <pwithnall@endlessos.org> Helps: #600
2020-11-11 19:21:00 +01:00
g_assert_cmpint (g_atomic_int_get (&thread_state), ==, THREAD_COMPLETED);
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
g_assert (callback_ran == TRUE);
gtask: Add a GTask:completed property This can be used to query whether the task has completed, in the sense that it has had a result set on it, and has already – or will soon – invoke its callback function. Notifications for this property are emitted immediately after the task’s main callback, in the same main context as that callback. This allows for multiple bits of code to listen for completion of the GTask, which opens the door for blocking on cancellation of the GTask and improved handling of ‘pending’ behaviour. https://bugzilla.gnome.org/show_bug.cgi?id=743636
2014-08-21 20:04:04 +02:00
g_assert_true (notification_emitted);
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
g_cancellable_reset (cancellable);
/* If return-on-cancel is TRUE, it does return early */
callback_ran = FALSE;
gtask: Add a GTask:completed property This can be used to query whether the task has completed, in the sense that it has had a result set on it, and has already – or will soon – invoke its callback function. Notifications for this property are emitted immediately after the task’s main callback, in the same main context as that callback. This allows for multiple bits of code to listen for completion of the GTask, which opens the door for blocking on cancellation of the GTask and improved handling of ‘pending’ behaviour. https://bugzilla.gnome.org/show_bug.cgi?id=743636
2014-08-21 20:04:04 +02:00
notification_emitted = FALSE;
tests: Fix non-atomic access to some shared variables And drop the `volatile` qualifier from the variables, as that doesn’t help with thread safety. Signed-off-by: Philip Withnall <pwithnall@endlessos.org> Helps: #600
2020-11-11 19:21:00 +01:00
g_atomic_int_set (&thread_state, THREAD_STARTING);
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
task = g_task_new (NULL, cancellable, return_on_cancel_callback, &callback_ran);
gio/tests/task: fix a race condition in test_run_in_thread() When running a task in a thread, GTask may still be internally holding a ref on the task in that thread even after the callback is called in the original thread (depending on thread scheduling). Fix the test to handle that by using a weak notify that signals a GCond, and wait for that GCond from the main thread. (And add a corresponding check to test_return_on_cancel().) https://bugzilla.gnome.org/show_bug.cgi?id=705152
2013-08-15 17:30:59 +02:00
g_object_weak_ref (G_OBJECT (task), task_weak_notify, (gpointer)&weak_notify_ran);
gtask: Add a GTask:completed property This can be used to query whether the task has completed, in the sense that it has had a result set on it, and has already – or will soon – invoke its callback function. Notifications for this property are emitted immediately after the task’s main callback, in the same main context as that callback. This allows for multiple bits of code to listen for completion of the GTask, which opens the door for blocking on cancellation of the GTask and improved handling of ‘pending’ behaviour. https://bugzilla.gnome.org/show_bug.cgi?id=743636
2014-08-21 20:04:04 +02:00
g_signal_connect (task, "notify::completed",
(GCallback) completed_cb, &notification_emitted);
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
g_task_set_return_on_cancel (task, TRUE);
g_task_set_task_data (task, (gpointer)&thread_state, NULL);
g_mutex_lock (&roc_init_mutex);
g_mutex_lock (&roc_finish_mutex);
g_task_run_in_thread (task, return_on_cancel_thread);
g_object_unref (task);
tests: Fix non-atomic access to some shared variables And drop the `volatile` qualifier from the variables, as that doesn’t help with thread safety. Signed-off-by: Philip Withnall <pwithnall@endlessos.org> Helps: #600
2020-11-11 19:21:00 +01:00
while (g_atomic_int_get (&thread_state) == THREAD_STARTING)
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
g_cond_wait (&roc_init_cond, &roc_init_mutex);
g_mutex_unlock (&roc_init_mutex);
tests: Fix non-atomic access to some shared variables And drop the `volatile` qualifier from the variables, as that doesn’t help with thread safety. Signed-off-by: Philip Withnall <pwithnall@endlessos.org> Helps: #600
2020-11-11 19:21:00 +01:00
g_assert_cmpint (g_atomic_int_get (&thread_state), ==, THREAD_RUNNING);
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
g_assert (callback_ran == FALSE);
g_cancellable_cancel (cancellable);
g_main_loop_run (loop);
tests: Fix non-atomic access to some shared variables And drop the `volatile` qualifier from the variables, as that doesn’t help with thread safety. Signed-off-by: Philip Withnall <pwithnall@endlessos.org> Helps: #600
2020-11-11 19:21:00 +01:00
g_assert_cmpint (g_atomic_int_get (&thread_state), ==, THREAD_RUNNING);
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
g_assert (callback_ran == TRUE);
tests: Fix non-atomic access to some shared variables And drop the `volatile` qualifier from the variables, as that doesn’t help with thread safety. Signed-off-by: Philip Withnall <pwithnall@endlessos.org> Helps: #600
2020-11-11 19:21:00 +01:00
g_assert_false (g_atomic_int_get (&weak_notify_ran));
gio/tests/task: fix a race condition in test_run_in_thread() When running a task in a thread, GTask may still be internally holding a ref on the task in that thread even after the callback is called in the original thread (depending on thread scheduling). Fix the test to handle that by using a weak notify that signals a GCond, and wait for that GCond from the main thread. (And add a corresponding check to test_return_on_cancel().) https://bugzilla.gnome.org/show_bug.cgi?id=705152
2013-08-15 17:30:59 +02:00
tests: Fix non-atomic access to some shared variables And drop the `volatile` qualifier from the variables, as that doesn’t help with thread safety. Signed-off-by: Philip Withnall <pwithnall@endlessos.org> Helps: #600
2020-11-11 19:21:00 +01:00
while (g_atomic_int_get (&thread_state) == THREAD_RUNNING)
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
g_cond_wait (&roc_finish_cond, &roc_finish_mutex);
g_mutex_unlock (&roc_finish_mutex);
tests: Fix non-atomic access to some shared variables And drop the `volatile` qualifier from the variables, as that doesn’t help with thread safety. Signed-off-by: Philip Withnall <pwithnall@endlessos.org> Helps: #600
2020-11-11 19:21:00 +01:00
g_assert_cmpint (g_atomic_int_get (&thread_state), ==, THREAD_CANCELLED);
gio/tests/task: fix a race condition in test_run_in_thread() When running a task in a thread, GTask may still be internally holding a ref on the task in that thread even after the callback is called in the original thread (depending on thread scheduling). Fix the test to handle that by using a weak notify that signals a GCond, and wait for that GCond from the main thread. (And add a corresponding check to test_return_on_cancel().) https://bugzilla.gnome.org/show_bug.cgi?id=705152
2013-08-15 17:30:59 +02:00
g_mutex_lock (&run_in_thread_mutex);
tests: Fix non-atomic access to some shared variables And drop the `volatile` qualifier from the variables, as that doesn’t help with thread safety. Signed-off-by: Philip Withnall <pwithnall@endlessos.org> Helps: #600
2020-11-11 19:21:00 +01:00
while (!g_atomic_int_get (&weak_notify_ran))
gio/tests/task: fix a race condition in test_run_in_thread() When running a task in a thread, GTask may still be internally holding a ref on the task in that thread even after the callback is called in the original thread (depending on thread scheduling). Fix the test to handle that by using a weak notify that signals a GCond, and wait for that GCond from the main thread. (And add a corresponding check to test_return_on_cancel().) https://bugzilla.gnome.org/show_bug.cgi?id=705152
2013-08-15 17:30:59 +02:00
g_cond_wait (&run_in_thread_cond, &run_in_thread_mutex);
g_mutex_unlock (&run_in_thread_mutex);
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
gtask: Add a GTask:completed property This can be used to query whether the task has completed, in the sense that it has had a result set on it, and has already – or will soon – invoke its callback function. Notifications for this property are emitted immediately after the task’s main callback, in the same main context as that callback. This allows for multiple bits of code to listen for completion of the GTask, which opens the door for blocking on cancellation of the GTask and improved handling of ‘pending’ behaviour. https://bugzilla.gnome.org/show_bug.cgi?id=743636
2014-08-21 20:04:04 +02:00
g_assert_true (notification_emitted);
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
g_cancellable_reset (cancellable);
/* If the task is already cancelled before it starts, it returns
* immediately, but the thread func still runs.
*/
callback_ran = FALSE;
gtask: Add a GTask:completed property This can be used to query whether the task has completed, in the sense that it has had a result set on it, and has already – or will soon – invoke its callback function. Notifications for this property are emitted immediately after the task’s main callback, in the same main context as that callback. This allows for multiple bits of code to listen for completion of the GTask, which opens the door for blocking on cancellation of the GTask and improved handling of ‘pending’ behaviour. https://bugzilla.gnome.org/show_bug.cgi?id=743636
2014-08-21 20:04:04 +02:00
notification_emitted = FALSE;
tests: Fix non-atomic access to some shared variables And drop the `volatile` qualifier from the variables, as that doesn’t help with thread safety. Signed-off-by: Philip Withnall <pwithnall@endlessos.org> Helps: #600
2020-11-11 19:21:00 +01:00
g_atomic_int_set (&thread_state, THREAD_STARTING);
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
task = g_task_new (NULL, cancellable, return_on_cancel_callback, &callback_ran);
gtask: Add a GTask:completed property This can be used to query whether the task has completed, in the sense that it has had a result set on it, and has already – or will soon – invoke its callback function. Notifications for this property are emitted immediately after the task’s main callback, in the same main context as that callback. This allows for multiple bits of code to listen for completion of the GTask, which opens the door for blocking on cancellation of the GTask and improved handling of ‘pending’ behaviour. https://bugzilla.gnome.org/show_bug.cgi?id=743636
2014-08-21 20:04:04 +02:00
g_signal_connect (task, "notify::completed",
(GCallback) completed_cb, &notification_emitted);
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
g_task_set_return_on_cancel (task, TRUE);
g_cancellable_cancel (cancellable);
g_task_set_task_data (task, (gpointer)&thread_state, NULL);
g_mutex_lock (&roc_init_mutex);
g_mutex_lock (&roc_finish_mutex);
g_task_run_in_thread (task, return_on_cancel_thread);
g_object_unref (task);
g_main_loop_run (loop);
g_assert (callback_ran == TRUE);
tests: Fix non-atomic access to some shared variables And drop the `volatile` qualifier from the variables, as that doesn’t help with thread safety. Signed-off-by: Philip Withnall <pwithnall@endlessos.org> Helps: #600
2020-11-11 19:21:00 +01:00
while (g_atomic_int_get (&thread_state) == THREAD_STARTING)
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
g_cond_wait (&roc_init_cond, &roc_init_mutex);
g_mutex_unlock (&roc_init_mutex);
tests: Fix non-atomic access to some shared variables And drop the `volatile` qualifier from the variables, as that doesn’t help with thread safety. Signed-off-by: Philip Withnall <pwithnall@endlessos.org> Helps: #600
2020-11-11 19:21:00 +01:00
g_assert_cmpint (g_atomic_int_get (&thread_state), ==, THREAD_RUNNING);
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
tests: Fix non-atomic access to some shared variables And drop the `volatile` qualifier from the variables, as that doesn’t help with thread safety. Signed-off-by: Philip Withnall <pwithnall@endlessos.org> Helps: #600
2020-11-11 19:21:00 +01:00
while (g_atomic_int_get (&thread_state) == THREAD_RUNNING)
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
g_cond_wait (&roc_finish_cond, &roc_finish_mutex);
g_mutex_unlock (&roc_finish_mutex);
tests: Fix non-atomic access to some shared variables And drop the `volatile` qualifier from the variables, as that doesn’t help with thread safety. Signed-off-by: Philip Withnall <pwithnall@endlessos.org> Helps: #600
2020-11-11 19:21:00 +01:00
g_assert_cmpint (g_atomic_int_get (&thread_state), ==, THREAD_CANCELLED);
gtask: Add a GTask:completed property This can be used to query whether the task has completed, in the sense that it has had a result set on it, and has already – or will soon – invoke its callback function. Notifications for this property are emitted immediately after the task’s main callback, in the same main context as that callback. This allows for multiple bits of code to listen for completion of the GTask, which opens the door for blocking on cancellation of the GTask and improved handling of ‘pending’ behaviour. https://bugzilla.gnome.org/show_bug.cgi?id=743636
2014-08-21 20:04:04 +02:00
g_assert_true (notification_emitted);
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
g_object_unref (cancellable);
}
/* test_return_on_cancel_sync */
static gpointer
cancel_sync_runner_thread (gpointer task)
{
g_task_run_in_thread_sync (task, return_on_cancel_thread);
return NULL;
}
static void
test_return_on_cancel_sync (void)
{
GTask *task;
GCancellable *cancellable;
tests: Fix non-atomic access to some shared variables And drop the `volatile` qualifier from the variables, as that doesn’t help with thread safety. Signed-off-by: Philip Withnall <pwithnall@endlessos.org> Helps: #600
2020-11-11 19:21:00 +01:00
ThreadState thread_state; /* (atomic) */
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
GThread *runner_thread;
gssize ret;
GError *error = NULL;
cancellable = g_cancellable_new ();
/* If return-on-cancel is FALSE, the task does not return early.
*/
tests: Fix non-atomic access to some shared variables And drop the `volatile` qualifier from the variables, as that doesn’t help with thread safety. Signed-off-by: Philip Withnall <pwithnall@endlessos.org> Helps: #600
2020-11-11 19:21:00 +01:00
g_atomic_int_set (&thread_state, THREAD_STARTING);
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
task = g_task_new (NULL, cancellable, run_in_thread_sync_callback, NULL);
g_task_set_task_data (task, (gpointer)&thread_state, NULL);
g_mutex_lock (&roc_init_mutex);
g_mutex_lock (&roc_finish_mutex);
runner_thread = g_thread_new ("return-on-cancel-sync runner thread",
cancel_sync_runner_thread, task);
tests: Fix non-atomic access to some shared variables And drop the `volatile` qualifier from the variables, as that doesn’t help with thread safety. Signed-off-by: Philip Withnall <pwithnall@endlessos.org> Helps: #600
2020-11-11 19:21:00 +01:00
while (g_atomic_int_get (&thread_state) == THREAD_STARTING)
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
g_cond_wait (&roc_init_cond, &roc_init_mutex);
g_mutex_unlock (&roc_init_mutex);
tests: Fix non-atomic access to some shared variables And drop the `volatile` qualifier from the variables, as that doesn’t help with thread safety. Signed-off-by: Philip Withnall <pwithnall@endlessos.org> Helps: #600
2020-11-11 19:21:00 +01:00
g_assert_cmpint (g_atomic_int_get (&thread_state), ==, THREAD_RUNNING);
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
g_cancellable_cancel (cancellable);
g_mutex_unlock (&roc_finish_mutex);
g_thread_join (runner_thread);
tests: Fix non-atomic access to some shared variables And drop the `volatile` qualifier from the variables, as that doesn’t help with thread safety. Signed-off-by: Philip Withnall <pwithnall@endlessos.org> Helps: #600
2020-11-11 19:21:00 +01:00
g_assert_cmpint (g_atomic_int_get (&thread_state), ==, THREAD_COMPLETED);
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
ret = g_task_propagate_int (task, &error);
g_assert_error (error, G_IO_ERROR, G_IO_ERROR_CANCELLED);
g_clear_error (&error);
g_assert_cmpint (ret, ==, -1);
g_object_unref (task);
g_cancellable_reset (cancellable);
/* If return-on-cancel is TRUE, it does return early */
tests: Fix non-atomic access to some shared variables And drop the `volatile` qualifier from the variables, as that doesn’t help with thread safety. Signed-off-by: Philip Withnall <pwithnall@endlessos.org> Helps: #600
2020-11-11 19:21:00 +01:00
g_atomic_int_set (&thread_state, THREAD_STARTING);
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
task = g_task_new (NULL, cancellable, run_in_thread_sync_callback, NULL);
g_task_set_return_on_cancel (task, TRUE);
g_task_set_task_data (task, (gpointer)&thread_state, NULL);
g_mutex_lock (&roc_init_mutex);
g_mutex_lock (&roc_finish_mutex);
runner_thread = g_thread_new ("return-on-cancel-sync runner thread",
cancel_sync_runner_thread, task);
tests: Fix non-atomic access to some shared variables And drop the `volatile` qualifier from the variables, as that doesn’t help with thread safety. Signed-off-by: Philip Withnall <pwithnall@endlessos.org> Helps: #600
2020-11-11 19:21:00 +01:00
while (g_atomic_int_get (&thread_state) == THREAD_STARTING)
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
g_cond_wait (&roc_init_cond, &roc_init_mutex);
g_mutex_unlock (&roc_init_mutex);
tests: Fix non-atomic access to some shared variables And drop the `volatile` qualifier from the variables, as that doesn’t help with thread safety. Signed-off-by: Philip Withnall <pwithnall@endlessos.org> Helps: #600
2020-11-11 19:21:00 +01:00
g_assert_cmpint (g_atomic_int_get (&thread_state), ==, THREAD_RUNNING);
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
g_cancellable_cancel (cancellable);
g_thread_join (runner_thread);
tests: Fix non-atomic access to some shared variables And drop the `volatile` qualifier from the variables, as that doesn’t help with thread safety. Signed-off-by: Philip Withnall <pwithnall@endlessos.org> Helps: #600
2020-11-11 19:21:00 +01:00
g_assert_cmpint (g_atomic_int_get (&thread_state), ==, THREAD_RUNNING);
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
ret = g_task_propagate_int (task, &error);
g_assert_error (error, G_IO_ERROR, G_IO_ERROR_CANCELLED);
g_clear_error (&error);
g_assert_cmpint (ret, ==, -1);
g_object_unref (task);
tests: Fix non-atomic access to some shared variables And drop the `volatile` qualifier from the variables, as that doesn’t help with thread safety. Signed-off-by: Philip Withnall <pwithnall@endlessos.org> Helps: #600
2020-11-11 19:21:00 +01:00
while (g_atomic_int_get (&thread_state) == THREAD_RUNNING)
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
g_cond_wait (&roc_finish_cond, &roc_finish_mutex);
g_mutex_unlock (&roc_finish_mutex);
tests: Fix non-atomic access to some shared variables And drop the `volatile` qualifier from the variables, as that doesn’t help with thread safety. Signed-off-by: Philip Withnall <pwithnall@endlessos.org> Helps: #600
2020-11-11 19:21:00 +01:00
g_assert_cmpint (g_atomic_int_get (&thread_state), ==, THREAD_CANCELLED);
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
g_cancellable_reset (cancellable);
/* If the task is already cancelled before it starts, it returns
* immediately, but the thread func still runs.
*/
tests: Fix non-atomic access to some shared variables And drop the `volatile` qualifier from the variables, as that doesn’t help with thread safety. Signed-off-by: Philip Withnall <pwithnall@endlessos.org> Helps: #600
2020-11-11 19:21:00 +01:00
g_atomic_int_set (&thread_state, THREAD_STARTING);
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
task = g_task_new (NULL, cancellable, run_in_thread_sync_callback, NULL);
g_task_set_return_on_cancel (task, TRUE);
g_cancellable_cancel (cancellable);
g_task_set_task_data (task, (gpointer)&thread_state, NULL);
g_mutex_lock (&roc_init_mutex);
g_mutex_lock (&roc_finish_mutex);
runner_thread = g_thread_new ("return-on-cancel-sync runner thread",
cancel_sync_runner_thread, task);
g_thread_join (runner_thread);
tests: Fix non-atomic access to some shared variables And drop the `volatile` qualifier from the variables, as that doesn’t help with thread safety. Signed-off-by: Philip Withnall <pwithnall@endlessos.org> Helps: #600
2020-11-11 19:21:00 +01:00
g_assert_cmpint (g_atomic_int_get (&thread_state), ==, THREAD_STARTING);
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
ret = g_task_propagate_int (task, &error);
g_assert_error (error, G_IO_ERROR, G_IO_ERROR_CANCELLED);
g_clear_error (&error);
g_assert_cmpint (ret, ==, -1);
g_object_unref (task);
tests: Fix non-atomic access to some shared variables And drop the `volatile` qualifier from the variables, as that doesn’t help with thread safety. Signed-off-by: Philip Withnall <pwithnall@endlessos.org> Helps: #600
2020-11-11 19:21:00 +01:00
while (g_atomic_int_get (&thread_state) == THREAD_STARTING)
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
g_cond_wait (&roc_init_cond, &roc_init_mutex);
g_mutex_unlock (&roc_init_mutex);
tests: Fix non-atomic access to some shared variables And drop the `volatile` qualifier from the variables, as that doesn’t help with thread safety. Signed-off-by: Philip Withnall <pwithnall@endlessos.org> Helps: #600
2020-11-11 19:21:00 +01:00
g_assert_cmpint (g_atomic_int_get (&thread_state), ==, THREAD_RUNNING);
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
tests: Fix non-atomic access to some shared variables And drop the `volatile` qualifier from the variables, as that doesn’t help with thread safety. Signed-off-by: Philip Withnall <pwithnall@endlessos.org> Helps: #600
2020-11-11 19:21:00 +01:00
while (g_atomic_int_get (&thread_state) == THREAD_RUNNING)
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
g_cond_wait (&roc_finish_cond, &roc_finish_mutex);
g_mutex_unlock (&roc_finish_mutex);
tests: Fix non-atomic access to some shared variables And drop the `volatile` qualifier from the variables, as that doesn’t help with thread safety. Signed-off-by: Philip Withnall <pwithnall@endlessos.org> Helps: #600
2020-11-11 19:21:00 +01:00
g_assert_cmpint (g_atomic_int_get (&thread_state), ==, THREAD_CANCELLED);
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
g_object_unref (cancellable);
}
/* test_return_on_cancel_atomic: turning return-on-cancel on/off is
* non-racy
*/
GMutex roca_mutex_1, roca_mutex_2;
GCond roca_cond_1, roca_cond_2;
static void
return_on_cancel_atomic_callback (GObject *object,
GAsyncResult *result,
gpointer user_data)
{
gboolean *callback_ran = user_data;
GError *error = NULL;
gssize ret;
g_assert (g_thread_self () == main_thread);
g_assert (object == NULL);
g_assert (g_task_is_valid (result, object));
g_assert (g_async_result_get_user_data (result) == user_data);
g_assert (g_task_had_error (G_TASK (result)));
gtask: Add a GTask:completed property This can be used to query whether the task has completed, in the sense that it has had a result set on it, and has already – or will soon – invoke its callback function. Notifications for this property are emitted immediately after the task’s main callback, in the same main context as that callback. This allows for multiple bits of code to listen for completion of the GTask, which opens the door for blocking on cancellation of the GTask and improved handling of ‘pending’ behaviour. https://bugzilla.gnome.org/show_bug.cgi?id=743636
2014-08-21 20:04:04 +02:00
g_assert_false (g_task_get_completed (G_TASK (result)));
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
ret = g_task_propagate_int (G_TASK (result), &error);
g_assert_error (error, G_IO_ERROR, G_IO_ERROR_CANCELLED);
g_clear_error (&error);
g_assert_cmpint (ret, ==, -1);
gio/tests/task: Ensure that g_task_had_error doesn't forget the error https://bugzilla.gnome.org/show_bug.cgi?id=764163
2016-03-24 18:59:54 +01:00
g_assert (g_task_had_error (G_TASK (result)));
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
*callback_ran = TRUE;
g_main_loop_quit (loop);
}
static void
return_on_cancel_atomic_thread (GTask *task,
gpointer source_object,
gpointer task_data,
GCancellable *cancellable)
{
tests: Fix non-atomic access to some shared variables And drop the `volatile` qualifier from the variables, as that doesn’t help with thread safety. Signed-off-by: Philip Withnall <pwithnall@endlessos.org> Helps: #600
2020-11-11 19:21:00 +01:00
gint *state = task_data; /* (atomic) */
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
g_assert (source_object == g_task_get_source_object (task));
g_assert (task_data == g_task_get_task_data (task));
g_assert (cancellable == g_task_get_cancellable (task));
gtask: Add a GTask:completed property This can be used to query whether the task has completed, in the sense that it has had a result set on it, and has already – or will soon – invoke its callback function. Notifications for this property are emitted immediately after the task’s main callback, in the same main context as that callback. This allows for multiple bits of code to listen for completion of the GTask, which opens the door for blocking on cancellation of the GTask and improved handling of ‘pending’ behaviour. https://bugzilla.gnome.org/show_bug.cgi?id=743636
2014-08-21 20:04:04 +02:00
g_assert_false (g_task_get_completed (task));
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
g_assert (g_thread_self () != main_thread);
tests: Fix non-atomic access to some shared variables And drop the `volatile` qualifier from the variables, as that doesn’t help with thread safety. Signed-off-by: Philip Withnall <pwithnall@endlessos.org> Helps: #600
2020-11-11 19:21:00 +01:00
g_assert_cmpint (g_atomic_int_get (state), ==, 0);
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
g_mutex_lock (&roca_mutex_1);
tests: Fix non-atomic access to some shared variables And drop the `volatile` qualifier from the variables, as that doesn’t help with thread safety. Signed-off-by: Philip Withnall <pwithnall@endlessos.org> Helps: #600
2020-11-11 19:21:00 +01:00
g_atomic_int_set (state, 1);
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
g_cond_signal (&roca_cond_1);
g_mutex_unlock (&roca_mutex_1);
g_mutex_lock (&roca_mutex_2);
if (g_task_set_return_on_cancel (task, FALSE))
tests: Fix non-atomic access to some shared variables And drop the `volatile` qualifier from the variables, as that doesn’t help with thread safety. Signed-off-by: Philip Withnall <pwithnall@endlessos.org> Helps: #600
2020-11-11 19:21:00 +01:00
g_atomic_int_set (state, 2);
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
else
tests: Fix non-atomic access to some shared variables And drop the `volatile` qualifier from the variables, as that doesn’t help with thread safety. Signed-off-by: Philip Withnall <pwithnall@endlessos.org> Helps: #600
2020-11-11 19:21:00 +01:00
g_atomic_int_set (state, 3);
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
g_cond_signal (&roca_cond_2);
g_mutex_unlock (&roca_mutex_2);
g_mutex_lock (&roca_mutex_1);
if (g_task_set_return_on_cancel (task, TRUE))
tests: Fix non-atomic access to some shared variables And drop the `volatile` qualifier from the variables, as that doesn’t help with thread safety. Signed-off-by: Philip Withnall <pwithnall@endlessos.org> Helps: #600
2020-11-11 19:21:00 +01:00
g_atomic_int_set (state, 4);
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
else
tests: Fix non-atomic access to some shared variables And drop the `volatile` qualifier from the variables, as that doesn’t help with thread safety. Signed-off-by: Philip Withnall <pwithnall@endlessos.org> Helps: #600
2020-11-11 19:21:00 +01:00
g_atomic_int_set (state, 5);
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
g_cond_signal (&roca_cond_1);
g_mutex_unlock (&roca_mutex_1);
g_mutex_lock (&roca_mutex_2);
if (g_task_set_return_on_cancel (task, TRUE))
tests: Fix non-atomic access to some shared variables And drop the `volatile` qualifier from the variables, as that doesn’t help with thread safety. Signed-off-by: Philip Withnall <pwithnall@endlessos.org> Helps: #600
2020-11-11 19:21:00 +01:00
g_atomic_int_set (state, 6);
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
else
tests: Fix non-atomic access to some shared variables And drop the `volatile` qualifier from the variables, as that doesn’t help with thread safety. Signed-off-by: Philip Withnall <pwithnall@endlessos.org> Helps: #600
2020-11-11 19:21:00 +01:00
g_atomic_int_set (state, 7);
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
g_cond_signal (&roca_cond_2);
g_mutex_unlock (&roca_mutex_2);
g_task_return_int (task, magic);
}
static void
test_return_on_cancel_atomic (void)
{
GTask *task;
GCancellable *cancellable;
tests: Fix non-atomic access to some shared variables And drop the `volatile` qualifier from the variables, as that doesn’t help with thread safety. Signed-off-by: Philip Withnall <pwithnall@endlessos.org> Helps: #600
2020-11-11 19:21:00 +01:00
gint state; /* (atomic) */
gtask: Add a GTask:completed property This can be used to query whether the task has completed, in the sense that it has had a result set on it, and has already – or will soon – invoke its callback function. Notifications for this property are emitted immediately after the task’s main callback, in the same main context as that callback. This allows for multiple bits of code to listen for completion of the GTask, which opens the door for blocking on cancellation of the GTask and improved handling of ‘pending’ behaviour. https://bugzilla.gnome.org/show_bug.cgi?id=743636
2014-08-21 20:04:04 +02:00
gboolean notification_emitted = FALSE;
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
gboolean callback_ran;
cancellable = g_cancellable_new ();
g_mutex_lock (&roca_mutex_1);
g_mutex_lock (&roca_mutex_2);
/* If we don't cancel it, each set_return_on_cancel() call will succeed */
tests: Fix non-atomic access to some shared variables And drop the `volatile` qualifier from the variables, as that doesn’t help with thread safety. Signed-off-by: Philip Withnall <pwithnall@endlessos.org> Helps: #600
2020-11-11 19:21:00 +01:00
g_atomic_int_set (&state, 0);
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
callback_ran = FALSE;
task = g_task_new (NULL, cancellable, return_on_cancel_atomic_callback, &callback_ran);
g_task_set_return_on_cancel (task, TRUE);
gtask: Add a GTask:completed property This can be used to query whether the task has completed, in the sense that it has had a result set on it, and has already – or will soon – invoke its callback function. Notifications for this property are emitted immediately after the task’s main callback, in the same main context as that callback. This allows for multiple bits of code to listen for completion of the GTask, which opens the door for blocking on cancellation of the GTask and improved handling of ‘pending’ behaviour. https://bugzilla.gnome.org/show_bug.cgi?id=743636
2014-08-21 20:04:04 +02:00
g_signal_connect (task, "notify::completed",
(GCallback) completed_cb, &notification_emitted);
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
g_task_set_task_data (task, (gpointer)&state, NULL);
g_task_run_in_thread (task, return_on_cancel_atomic_thread);
g_object_unref (task);
tests: Fix non-atomic access to some shared variables And drop the `volatile` qualifier from the variables, as that doesn’t help with thread safety. Signed-off-by: Philip Withnall <pwithnall@endlessos.org> Helps: #600
2020-11-11 19:21:00 +01:00
g_assert_cmpint (g_atomic_int_get (&state), ==, 0);
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
tests: Fix non-atomic access to some shared variables And drop the `volatile` qualifier from the variables, as that doesn’t help with thread safety. Signed-off-by: Philip Withnall <pwithnall@endlessos.org> Helps: #600
2020-11-11 19:21:00 +01:00
while (g_atomic_int_get (&state) == 0)
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
g_cond_wait (&roca_cond_1, &roca_mutex_1);
tests: Fix non-atomic access to some shared variables And drop the `volatile` qualifier from the variables, as that doesn’t help with thread safety. Signed-off-by: Philip Withnall <pwithnall@endlessos.org> Helps: #600
2020-11-11 19:21:00 +01:00
g_assert_cmpint (g_atomic_int_get (&state), ==, 1);
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
tests: Fix non-atomic access to some shared variables And drop the `volatile` qualifier from the variables, as that doesn’t help with thread safety. Signed-off-by: Philip Withnall <pwithnall@endlessos.org> Helps: #600
2020-11-11 19:21:00 +01:00
while (g_atomic_int_get (&state) == 1)
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
g_cond_wait (&roca_cond_2, &roca_mutex_2);
tests: Fix non-atomic access to some shared variables And drop the `volatile` qualifier from the variables, as that doesn’t help with thread safety. Signed-off-by: Philip Withnall <pwithnall@endlessos.org> Helps: #600
2020-11-11 19:21:00 +01:00
g_assert_cmpint (g_atomic_int_get (&state), ==, 2);
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
tests: Fix non-atomic access to some shared variables And drop the `volatile` qualifier from the variables, as that doesn’t help with thread safety. Signed-off-by: Philip Withnall <pwithnall@endlessos.org> Helps: #600
2020-11-11 19:21:00 +01:00
while (g_atomic_int_get (&state) == 2)
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
g_cond_wait (&roca_cond_1, &roca_mutex_1);
tests: Fix non-atomic access to some shared variables And drop the `volatile` qualifier from the variables, as that doesn’t help with thread safety. Signed-off-by: Philip Withnall <pwithnall@endlessos.org> Helps: #600
2020-11-11 19:21:00 +01:00
g_assert_cmpint (g_atomic_int_get (&state), ==, 4);
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
tests: Fix non-atomic access to some shared variables And drop the `volatile` qualifier from the variables, as that doesn’t help with thread safety. Signed-off-by: Philip Withnall <pwithnall@endlessos.org> Helps: #600
2020-11-11 19:21:00 +01:00
while (g_atomic_int_get (&state) == 4)
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
g_cond_wait (&roca_cond_2, &roca_mutex_2);
tests: Fix non-atomic access to some shared variables And drop the `volatile` qualifier from the variables, as that doesn’t help with thread safety. Signed-off-by: Philip Withnall <pwithnall@endlessos.org> Helps: #600
2020-11-11 19:21:00 +01:00
g_assert_cmpint (g_atomic_int_get (&state), ==, 6);
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
/* callback assumes there'll be a cancelled error */
g_cancellable_cancel (cancellable);
g_assert (callback_ran == FALSE);
g_main_loop_run (loop);
g_assert (callback_ran == TRUE);
gtask: Add a GTask:completed property This can be used to query whether the task has completed, in the sense that it has had a result set on it, and has already – or will soon – invoke its callback function. Notifications for this property are emitted immediately after the task’s main callback, in the same main context as that callback. This allows for multiple bits of code to listen for completion of the GTask, which opens the door for blocking on cancellation of the GTask and improved handling of ‘pending’ behaviour. https://bugzilla.gnome.org/show_bug.cgi?id=743636
2014-08-21 20:04:04 +02:00
g_assert_true (notification_emitted);
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
g_cancellable_reset (cancellable);
/* If we cancel while it's temporarily not return-on-cancel, the
* task won't complete right away, and further
* g_task_set_return_on_cancel() calls will return FALSE.
*/
tests: Fix non-atomic access to some shared variables And drop the `volatile` qualifier from the variables, as that doesn’t help with thread safety. Signed-off-by: Philip Withnall <pwithnall@endlessos.org> Helps: #600
2020-11-11 19:21:00 +01:00
g_atomic_int_set (&state, 0);
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
callback_ran = FALSE;
gtask: Add a GTask:completed property This can be used to query whether the task has completed, in the sense that it has had a result set on it, and has already – or will soon – invoke its callback function. Notifications for this property are emitted immediately after the task’s main callback, in the same main context as that callback. This allows for multiple bits of code to listen for completion of the GTask, which opens the door for blocking on cancellation of the GTask and improved handling of ‘pending’ behaviour. https://bugzilla.gnome.org/show_bug.cgi?id=743636
2014-08-21 20:04:04 +02:00
notification_emitted = FALSE;
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
task = g_task_new (NULL, cancellable, return_on_cancel_atomic_callback, &callback_ran);
g_task_set_return_on_cancel (task, TRUE);
gtask: Add a GTask:completed property This can be used to query whether the task has completed, in the sense that it has had a result set on it, and has already – or will soon – invoke its callback function. Notifications for this property are emitted immediately after the task’s main callback, in the same main context as that callback. This allows for multiple bits of code to listen for completion of the GTask, which opens the door for blocking on cancellation of the GTask and improved handling of ‘pending’ behaviour. https://bugzilla.gnome.org/show_bug.cgi?id=743636
2014-08-21 20:04:04 +02:00
g_signal_connect (task, "notify::completed",
(GCallback) completed_cb, &notification_emitted);
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
g_task_set_task_data (task, (gpointer)&state, NULL);
g_task_run_in_thread (task, return_on_cancel_atomic_thread);
tests: Fix non-atomic access to some shared variables And drop the `volatile` qualifier from the variables, as that doesn’t help with thread safety. Signed-off-by: Philip Withnall <pwithnall@endlessos.org> Helps: #600
2020-11-11 19:21:00 +01:00
g_assert_cmpint (g_atomic_int_get (&state), ==, 0);
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
tests: Fix non-atomic access to some shared variables And drop the `volatile` qualifier from the variables, as that doesn’t help with thread safety. Signed-off-by: Philip Withnall <pwithnall@endlessos.org> Helps: #600
2020-11-11 19:21:00 +01:00
while (g_atomic_int_get (&state) == 0)
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
g_cond_wait (&roca_cond_1, &roca_mutex_1);
tests: Fix non-atomic access to some shared variables And drop the `volatile` qualifier from the variables, as that doesn’t help with thread safety. Signed-off-by: Philip Withnall <pwithnall@endlessos.org> Helps: #600
2020-11-11 19:21:00 +01:00
g_assert_cmpint (g_atomic_int_get (&state), ==, 1);
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
g_assert (g_task_get_return_on_cancel (task));
tests: Fix non-atomic access to some shared variables And drop the `volatile` qualifier from the variables, as that doesn’t help with thread safety. Signed-off-by: Philip Withnall <pwithnall@endlessos.org> Helps: #600
2020-11-11 19:21:00 +01:00
while (g_atomic_int_get (&state) == 1)
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
g_cond_wait (&roca_cond_2, &roca_mutex_2);
tests: Fix non-atomic access to some shared variables And drop the `volatile` qualifier from the variables, as that doesn’t help with thread safety. Signed-off-by: Philip Withnall <pwithnall@endlessos.org> Helps: #600
2020-11-11 19:21:00 +01:00
g_assert_cmpint (g_atomic_int_get (&state), ==, 2);
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
g_assert (!g_task_get_return_on_cancel (task));
g_cancellable_cancel (cancellable);
g_idle_add (idle_quit_loop, NULL);
g_main_loop_run (loop);
g_assert (callback_ran == FALSE);
tests: Fix non-atomic access to some shared variables And drop the `volatile` qualifier from the variables, as that doesn’t help with thread safety. Signed-off-by: Philip Withnall <pwithnall@endlessos.org> Helps: #600
2020-11-11 19:21:00 +01:00
while (g_atomic_int_get (&state) == 2)
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
g_cond_wait (&roca_cond_1, &roca_mutex_1);
tests: Fix non-atomic access to some shared variables And drop the `volatile` qualifier from the variables, as that doesn’t help with thread safety. Signed-off-by: Philip Withnall <pwithnall@endlessos.org> Helps: #600
2020-11-11 19:21:00 +01:00
g_assert_cmpint (g_atomic_int_get (&state), ==, 5);
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
g_assert (!g_task_get_return_on_cancel (task));
g_main_loop_run (loop);
g_assert (callback_ran == TRUE);
gtask: Add a GTask:completed property This can be used to query whether the task has completed, in the sense that it has had a result set on it, and has already – or will soon – invoke its callback function. Notifications for this property are emitted immediately after the task’s main callback, in the same main context as that callback. This allows for multiple bits of code to listen for completion of the GTask, which opens the door for blocking on cancellation of the GTask and improved handling of ‘pending’ behaviour. https://bugzilla.gnome.org/show_bug.cgi?id=743636
2014-08-21 20:04:04 +02:00
g_assert_true (notification_emitted);
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
tests: Fix non-atomic access to some shared variables And drop the `volatile` qualifier from the variables, as that doesn’t help with thread safety. Signed-off-by: Philip Withnall <pwithnall@endlessos.org> Helps: #600
2020-11-11 19:21:00 +01:00
while (g_atomic_int_get (&state) == 5)
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
g_cond_wait (&roca_cond_2, &roca_mutex_2);
tests: Fix non-atomic access to some shared variables And drop the `volatile` qualifier from the variables, as that doesn’t help with thread safety. Signed-off-by: Philip Withnall <pwithnall@endlessos.org> Helps: #600
2020-11-11 19:21:00 +01:00
g_assert_cmpint (g_atomic_int_get (&state), ==, 7);
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
g_object_unref (cancellable);
g_mutex_unlock (&roca_mutex_1);
g_mutex_unlock (&roca_mutex_2);
tests: Fix location of an unref in the GTask tests This method drops the last reference *it* owns to the GTask, but then continues to call methods on the GTask. This wasn’t resulting in failures because a ref in another thread kept the GTask alive, but that’s quite dodgy. Signed-off-by: Philip Withnall <withnall@endlessm.com>
2018-10-01 22:44:01 +02:00
g_object_unref (task);
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
}
/* test_return_pointer: memory management of pointer returns */
static void
test_return_pointer (void)
{
GObject *object, *ret;
GTask *task;
GCancellable *cancellable;
GError *error = NULL;
/* If we don't read back the return value, the task will
* run its destroy notify.
*/
object = (GObject *)g_dummy_object_new ();
g_assert_cmpint (object->ref_count, ==, 1);
g_object_add_weak_pointer (object, (gpointer *)&object);
task = g_task_new (NULL, NULL, NULL, NULL);
g_object_add_weak_pointer (G_OBJECT (task), (gpointer *)&task);
g_task_return_pointer (task, object, g_object_unref);
g_assert_cmpint (object->ref_count, ==, 1);
gtask: Add a GTask:completed property This can be used to query whether the task has completed, in the sense that it has had a result set on it, and has already – or will soon – invoke its callback function. Notifications for this property are emitted immediately after the task’s main callback, in the same main context as that callback. This allows for multiple bits of code to listen for completion of the GTask, which opens the door for blocking on cancellation of the GTask and improved handling of ‘pending’ behaviour. https://bugzilla.gnome.org/show_bug.cgi?id=743636
2014-08-21 20:04:04 +02:00
/* Task and object are reffed until the :completed notification in idle. */
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
g_object_unref (task);
gtask: Add a GTask:completed property This can be used to query whether the task has completed, in the sense that it has had a result set on it, and has already – or will soon – invoke its callback function. Notifications for this property are emitted immediately after the task’s main callback, in the same main context as that callback. This allows for multiple bits of code to listen for completion of the GTask, which opens the door for blocking on cancellation of the GTask and improved handling of ‘pending’ behaviour. https://bugzilla.gnome.org/show_bug.cgi?id=743636
2014-08-21 20:04:04 +02:00
g_assert_nonnull (task);
g_assert_nonnull (object);
wait_for_completed_notification (task);
g_assert_null (task);
g_assert_null (object);
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
/* Likewise, if the return value is overwritten by an error */
object = (GObject *)g_dummy_object_new ();
g_assert_cmpint (object->ref_count, ==, 1);
g_object_add_weak_pointer (object, (gpointer *)&object);
cancellable = g_cancellable_new ();
task = g_task_new (NULL, cancellable, NULL, NULL);
g_object_add_weak_pointer (G_OBJECT (task), (gpointer *)&task);
g_task_return_pointer (task, object, g_object_unref);
g_assert_cmpint (object->ref_count, ==, 1);
g_cancellable_cancel (cancellable);
g_assert_cmpint (object->ref_count, ==, 1);
ret = g_task_propagate_pointer (task, &error);
g_assert (ret == NULL);
g_assert_error (error, G_IO_ERROR, G_IO_ERROR_CANCELLED);
g_clear_error (&error);
g_assert_cmpint (object->ref_count, ==, 1);
g_object_unref (task);
g_object_unref (cancellable);
gtask: Add a GTask:completed property This can be used to query whether the task has completed, in the sense that it has had a result set on it, and has already – or will soon – invoke its callback function. Notifications for this property are emitted immediately after the task’s main callback, in the same main context as that callback. This allows for multiple bits of code to listen for completion of the GTask, which opens the door for blocking on cancellation of the GTask and improved handling of ‘pending’ behaviour. https://bugzilla.gnome.org/show_bug.cgi?id=743636
2014-08-21 20:04:04 +02:00
g_assert_nonnull (task);
g_assert_nonnull (object);
wait_for_completed_notification (task);
g_assert_null (task);
g_assert_null (object);
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
/* If we read back the return value, we steal its ref */
object = (GObject *)g_dummy_object_new ();
g_assert_cmpint (object->ref_count, ==, 1);
g_object_add_weak_pointer (object, (gpointer *)&object);
task = g_task_new (NULL, NULL, NULL, NULL);
g_object_add_weak_pointer (G_OBJECT (task), (gpointer *)&task);
g_task_return_pointer (task, object, g_object_unref);
g_assert_cmpint (object->ref_count, ==, 1);
ret = g_task_propagate_pointer (task, &error);
g_assert_no_error (error);
g_assert (ret == object);
g_assert_cmpint (object->ref_count, ==, 1);
g_object_unref (task);
gtask: Add a GTask:completed property This can be used to query whether the task has completed, in the sense that it has had a result set on it, and has already – or will soon – invoke its callback function. Notifications for this property are emitted immediately after the task’s main callback, in the same main context as that callback. This allows for multiple bits of code to listen for completion of the GTask, which opens the door for blocking on cancellation of the GTask and improved handling of ‘pending’ behaviour. https://bugzilla.gnome.org/show_bug.cgi?id=743636
2014-08-21 20:04:04 +02:00
g_assert_nonnull (task);
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
g_assert_cmpint (object->ref_count, ==, 1);
g_object_unref (object);
g_assert (object == NULL);
gtask: Add a GTask:completed property This can be used to query whether the task has completed, in the sense that it has had a result set on it, and has already – or will soon – invoke its callback function. Notifications for this property are emitted immediately after the task’s main callback, in the same main context as that callback. This allows for multiple bits of code to listen for completion of the GTask, which opens the door for blocking on cancellation of the GTask and improved handling of ‘pending’ behaviour. https://bugzilla.gnome.org/show_bug.cgi?id=743636
2014-08-21 20:04:04 +02:00
wait_for_completed_notification (task);
g_assert_null (task);
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
}
task: Add return/propagate API for GValue This is useful for bindings. Related: https://gitlab.gnome.org/GNOME/glib/issues/668
2019-11-05 09:54:01 +01:00
static void
test_return_value (void)
{
GObject *object;
GValue value = G_VALUE_INIT;
GValue ret = G_VALUE_INIT;
GTask *task;
GError *error = NULL;
object = (GObject *)g_dummy_object_new ();
g_assert_cmpint (object->ref_count, ==, 1);
g_object_add_weak_pointer (object, (gpointer *)&object);
g_value_init (&value, G_TYPE_OBJECT);
g_value_set_object (&value, object);
g_assert_cmpint (object->ref_count, ==, 2);
task = g_task_new (NULL, NULL, NULL, NULL);
g_object_add_weak_pointer (G_OBJECT (task), (gpointer *)&task);
g_task_return_value (task, &value);
g_assert_cmpint (object->ref_count, ==, 3);
g_assert_true (g_task_propagate_value (task, &ret, &error));
g_assert_no_error (error);
g_assert_true (g_value_get_object (&ret) == object);
g_assert_cmpint (object->ref_count, ==, 3);
g_object_unref (task);
g_assert_nonnull (task);
wait_for_completed_notification (task);
g_assert_null (task);
g_assert_cmpint (object->ref_count, ==, 3);
g_value_unset (&ret);
g_assert_cmpint (object->ref_count, ==, 2);
g_value_unset (&value);
g_assert_cmpint (object->ref_count, ==, 1);
g_object_unref (object);
g_assert_null (object);
}
gtask: Add g_task_return_prefixed_error() An equivalent to g_propagate_prefixed_error, but for errors sent though GTask, allowing tasks to easily provide extra context to callers.
2023-10-31 04:49:28 +01:00
static void
test_return_prefixed_error (void)
{
GTask *task;
GError *original_error = NULL;
GError *error = NULL;
g_set_error (&original_error, G_IO_ERROR, G_IO_ERROR_UNKNOWN, "oh no!");
task = g_task_new (NULL, NULL, NULL, NULL);
g_task_return_prefixed_error (task, original_error, "task %s: ", "failed");
wait_for_completed_notification (task);
g_assert_null (g_task_propagate_pointer (task, &error));
g_assert_error (error, G_IO_ERROR, G_IO_ERROR_UNKNOWN);
g_assert_cmpstr (error->message, ==, "task failed: oh no!");
g_assert_finalize_object (task);
g_clear_error (&error);
}
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
/* test_object_keepalive: GTask takes a ref on its source object */
static GObject *keepalive_object;
static void
keepalive_callback (GObject *object,
GAsyncResult *result,
gpointer user_data)
{
gssize *result_out = user_data;
GError *error = NULL;
g_assert (object == keepalive_object);
g_assert (g_task_is_valid (result, object));
g_assert (g_async_result_get_user_data (result) == user_data);
g_assert (!g_task_had_error (G_TASK (result)));
gtask: Add a GTask:completed property This can be used to query whether the task has completed, in the sense that it has had a result set on it, and has already – or will soon – invoke its callback function. Notifications for this property are emitted immediately after the task’s main callback, in the same main context as that callback. This allows for multiple bits of code to listen for completion of the GTask, which opens the door for blocking on cancellation of the GTask and improved handling of ‘pending’ behaviour. https://bugzilla.gnome.org/show_bug.cgi?id=743636
2014-08-21 20:04:04 +02:00
g_assert_false (g_task_get_completed (G_TASK (result)));
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
*result_out = g_task_propagate_int (G_TASK (result), &error);
g_assert_no_error (error);
gio/tests/task: Ensure that g_task_had_error doesn't forget the error https://bugzilla.gnome.org/show_bug.cgi?id=764163
2016-03-24 18:59:54 +01:00
g_assert (!g_task_had_error (G_TASK (result)));
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
g_main_loop_quit (loop);
}
static void
test_object_keepalive (void)
{
GObject *object;
GTask *task;
gssize result;
int ref_count;
gtask: Add a GTask:completed property This can be used to query whether the task has completed, in the sense that it has had a result set on it, and has already – or will soon – invoke its callback function. Notifications for this property are emitted immediately after the task’s main callback, in the same main context as that callback. This allows for multiple bits of code to listen for completion of the GTask, which opens the door for blocking on cancellation of the GTask and improved handling of ‘pending’ behaviour. https://bugzilla.gnome.org/show_bug.cgi?id=743636
2014-08-21 20:04:04 +02:00
gboolean notification_emitted = FALSE;
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
keepalive_object = object = (GObject *)g_dummy_object_new ();
g_object_add_weak_pointer (object, (gpointer *)&object);
task = g_task_new (object, NULL, keepalive_callback, &result);
g_object_add_weak_pointer (G_OBJECT (task), (gpointer *)&task);
gtask: Add a GTask:completed property This can be used to query whether the task has completed, in the sense that it has had a result set on it, and has already – or will soon – invoke its callback function. Notifications for this property are emitted immediately after the task’s main callback, in the same main context as that callback. This allows for multiple bits of code to listen for completion of the GTask, which opens the door for blocking on cancellation of the GTask and improved handling of ‘pending’ behaviour. https://bugzilla.gnome.org/show_bug.cgi?id=743636
2014-08-21 20:04:04 +02:00
g_signal_connect (task, "notify::completed",
(GCallback) completed_cb, &notification_emitted);
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
ref_count = object->ref_count;
g_assert_cmpint (ref_count, >, 1);
g_assert (g_task_get_source_object (task) == object);
g_assert (g_async_result_get_source_object (G_ASYNC_RESULT (task)) == object);
g_assert_cmpint (object->ref_count, ==, ref_count + 1);
g_object_unref (object);
g_object_unref (object);
g_assert (object != NULL);
g_task_return_int (task, magic);
g_main_loop_run (loop);
g_assert (object != NULL);
g_assert_cmpint (result, ==, magic);
gtask: Add a GTask:completed property This can be used to query whether the task has completed, in the sense that it has had a result set on it, and has already – or will soon – invoke its callback function. Notifications for this property are emitted immediately after the task’s main callback, in the same main context as that callback. This allows for multiple bits of code to listen for completion of the GTask, which opens the door for blocking on cancellation of the GTask and improved handling of ‘pending’ behaviour. https://bugzilla.gnome.org/show_bug.cgi?id=743636
2014-08-21 20:04:04 +02:00
g_assert_true (notification_emitted);
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
g_object_unref (task);
g_assert (task == NULL);
g_assert (object == NULL);
}
/* test_legacy_error: legacy GSimpleAsyncResult handling */
static void test_legacy_error (void);
static void
legacy_error_callback (GObject *object,
GAsyncResult *result,
gpointer user_data)
{
gssize *result_out = user_data;
GError *error = NULL;
g_assert (object == NULL);
g_assert (g_async_result_is_tagged (result, test_legacy_error));
g_assert (g_async_result_get_user_data (result) == user_data);
if (g_async_result_legacy_propagate_error (result, &error))
{
g_assert (!g_task_is_valid (result, object));
G_GNUC_BEGIN_IGNORE_DEPRECATIONS;
g_assert (g_simple_async_result_is_valid (result, object, test_legacy_error));
G_GNUC_END_IGNORE_DEPRECATIONS;
g_assert_error (error, G_IO_ERROR, G_IO_ERROR_FAILED);
*result_out = -2;
tests: Various minor leak fixes in the GIO tests Signed-off-by: Philip Withnall <withnall@endlessm.com>
2018-06-27 10:57:21 +02:00
g_clear_error (&error);
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
}
else
{
g_assert (g_task_is_valid (result, object));
*result_out = g_task_propagate_int (G_TASK (result), NULL);
/* Might be error, might not */
}
g_main_loop_quit (loop);
}
static gboolean
legacy_error_return (gpointer user_data)
{
if (G_IS_TASK (user_data))
{
GTask *task = user_data;
g_task_return_int (task, magic);
g_object_unref (task);
}
else
{
gio/tests: Add a dynamic type check assertion This shuts up a static analysis false positive, and adds some extra checking. Found by scan-build. https://bugzilla.gnome.org/show_bug.cgi?id=113075
2013-11-26 12:21:37 +01:00
GSimpleAsyncResult *simple = G_SIMPLE_ASYNC_RESULT (user_data);
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
G_GNUC_BEGIN_IGNORE_DEPRECATIONS;
g_simple_async_result_set_error (simple,
G_IO_ERROR,
G_IO_ERROR_FAILED,
"Failed");
g_simple_async_result_complete (simple);
G_GNUC_END_IGNORE_DEPRECATIONS;
g_object_unref (simple);
}
return FALSE;
}
static void
test_legacy_error (void)
{
GTask *task;
GSimpleAsyncResult *simple;
gssize result;
/* GTask success */
task = g_task_new (NULL, NULL, legacy_error_callback, &result);
g_task_set_source_tag (task, test_legacy_error);
g_object_add_weak_pointer (G_OBJECT (task), (gpointer *)&task);
g_idle_add (legacy_error_return, task);
g_main_loop_run (loop);
g_assert_cmpint (result, ==, magic);
g_assert (task == NULL);
/* GTask error */
task = g_task_new (NULL, NULL, legacy_error_callback, &result);
g_task_set_source_tag (task, test_legacy_error);
g_object_add_weak_pointer (G_OBJECT (task), (gpointer *)&task);
g_task_return_new_error (task, G_IO_ERROR, G_IO_ERROR_FAILED,
"Failed");
g_object_unref (task);
g_main_loop_run (loop);
g_assert_cmpint (result, ==, -1);
g_assert (task == NULL);
/* GSimpleAsyncResult error */
G_GNUC_BEGIN_IGNORE_DEPRECATIONS;
simple = g_simple_async_result_new (NULL, legacy_error_callback, &result,
test_legacy_error);
G_GNUC_END_IGNORE_DEPRECATIONS;
g_object_add_weak_pointer (G_OBJECT (simple), (gpointer *)&simple);
g_idle_add (legacy_error_return, simple);
g_main_loop_run (loop);
g_assert_cmpint (result, ==, -2);
g_assert (simple == NULL);
}
tests: Add return ordering tests for GTask These are based on an example test program provided by Will Thompson in issue #1525. Signed-off-by: Philip Withnall <withnall@endlessm.com> https://gitlab.gnome.org/GNOME/glib/issues/1525
2018-10-01 22:45:04 +02:00
/* Various helper functions for the return tests below. */
static void
task_complete_cb (GObject *source,
GAsyncResult *result,
gpointer user_data)
{
GTask *task = G_TASK (result);
guint *calls = user_data;
g_assert_cmpint (++*calls, <=, 1);
/* Propagate the result, so it’s removed from the task’s internal state. */
g_task_propagate_boolean (task, NULL);
}
static void
return_twice (GTask *task)
{
gboolean error_first = GPOINTER_TO_UINT (g_task_get_task_data (task));
if (error_first)
{
g_task_return_new_error (task, G_IO_ERROR, G_IO_ERROR_UNKNOWN, "oh no");
g_task_return_boolean (task, TRUE);
}
else
{
g_task_return_boolean (task, TRUE);
g_task_return_new_error (task, G_IO_ERROR, G_IO_ERROR_UNKNOWN, "oh no");
}
}
static gboolean
idle_cb (gpointer user_data)
{
GTask *task = user_data;
return_twice (task);
g_object_unref (task);
return G_SOURCE_REMOVE;
}
static void
test_return_permutation (gboolean error_first,
gboolean return_in_idle)
{
guint calls = 0;
GTask *task = NULL;
g_test_bug ("https://gitlab.gnome.org/GNOME/glib/issues/1525");
task = g_task_new (NULL, NULL, task_complete_cb, &calls);
g_task_set_task_data (task, GUINT_TO_POINTER (error_first), NULL);
if (return_in_idle)
g_idle_add (idle_cb, g_object_ref (task));
else
return_twice (task);
while (calls == 0)
g_main_context_iteration (NULL, TRUE);
g_assert_cmpint (calls, ==, 1);
g_object_unref (task);
}
/* Test that calling g_task_return_boolean() after g_task_return_error(), when
* returning in an idle callback, correctly results in a critical warning. */
static void
test_return_in_idle_error_first (void)
{
if (g_test_subprocess ())
{
test_return_permutation (TRUE, TRUE);
return;
}
gtestutils: Add G_TEST_SUBPROCESS_DEFAULT, G_TEST_TRAP_DEFAULT This makes calls to test subprocesses with default behaviour more self-documenting. Signed-off-by: Simon McVittie <smcv@collabora.com>
2022-03-31 15:30:43 +02:00
g_test_trap_subprocess (NULL, 0, G_TEST_SUBPROCESS_DEFAULT);
tests: Add return ordering tests for GTask These are based on an example test program provided by Will Thompson in issue #1525. Signed-off-by: Philip Withnall <withnall@endlessm.com> https://gitlab.gnome.org/GNOME/glib/issues/1525
2018-10-01 22:45:04 +02:00
g_test_trap_assert_failed ();
g_test_trap_assert_stderr ("*CRITICAL*assertion '!task->ever_returned' failed*");
}
/* Test that calling g_task_return_error() after g_task_return_boolean(), when
* returning in an idle callback, correctly results in a critical warning. */
static void
test_return_in_idle_value_first (void)
{
if (g_test_subprocess ())
{
test_return_permutation (FALSE, TRUE);
return;
}
gtestutils: Add G_TEST_SUBPROCESS_DEFAULT, G_TEST_TRAP_DEFAULT This makes calls to test subprocesses with default behaviour more self-documenting. Signed-off-by: Simon McVittie <smcv@collabora.com>
2022-03-31 15:30:43 +02:00
g_test_trap_subprocess (NULL, 0, G_TEST_SUBPROCESS_DEFAULT);
tests: Add return ordering tests for GTask These are based on an example test program provided by Will Thompson in issue #1525. Signed-off-by: Philip Withnall <withnall@endlessm.com> https://gitlab.gnome.org/GNOME/glib/issues/1525
2018-10-01 22:45:04 +02:00
g_test_trap_assert_failed ();
g_test_trap_assert_stderr ("*CRITICAL*assertion '!task->ever_returned' failed*");
}
/* Test that calling g_task_return_boolean() after g_task_return_error(), when
* returning synchronously, correctly results in a critical warning. */
static void
test_return_error_first (void)
{
if (g_test_subprocess ())
{
test_return_permutation (TRUE, FALSE);
return;
}
gtestutils: Add G_TEST_SUBPROCESS_DEFAULT, G_TEST_TRAP_DEFAULT This makes calls to test subprocesses with default behaviour more self-documenting. Signed-off-by: Simon McVittie <smcv@collabora.com>
2022-03-31 15:30:43 +02:00
g_test_trap_subprocess (NULL, 0, G_TEST_SUBPROCESS_DEFAULT);
tests: Add return ordering tests for GTask These are based on an example test program provided by Will Thompson in issue #1525. Signed-off-by: Philip Withnall <withnall@endlessm.com> https://gitlab.gnome.org/GNOME/glib/issues/1525
2018-10-01 22:45:04 +02:00
g_test_trap_assert_failed ();
g_test_trap_assert_stderr ("*CRITICAL*assertion '!task->ever_returned' failed*");
}
/* Test that calling g_task_return_error() after g_task_return_boolean(), when
* returning synchronously, correctly results in a critical warning. */
static void
test_return_value_first (void)
{
if (g_test_subprocess ())
{
test_return_permutation (FALSE, FALSE);
return;
}
gtestutils: Add G_TEST_SUBPROCESS_DEFAULT, G_TEST_TRAP_DEFAULT This makes calls to test subprocesses with default behaviour more self-documenting. Signed-off-by: Simon McVittie <smcv@collabora.com>
2022-03-31 15:30:43 +02:00
g_test_trap_subprocess (NULL, 0, G_TEST_SUBPROCESS_DEFAULT);
tests: Add return ordering tests for GTask These are based on an example test program provided by Will Thompson in issue #1525. Signed-off-by: Philip Withnall <withnall@endlessm.com> https://gitlab.gnome.org/GNOME/glib/issues/1525
2018-10-01 22:45:04 +02:00
g_test_trap_assert_failed ();
g_test_trap_assert_stderr ("*CRITICAL*assertion '!task->ever_returned' failed*");
}
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
gtask: Don't overwrite source names Only set a name in g_task_attach_source if the source does not already have one. Including a new test by Philip Withnall.
2022-08-28 03:32:46 +02:00
static gboolean
source_cb (gpointer user_data)
{
return G_SOURCE_REMOVE;
}
static void
test_attach_source_set_name (void)
{
guint calls = 0;
GTask *task = NULL;
GSource *source = NULL;
GSourceFuncs source_funcs = { NULL, NULL, NULL, NULL, NULL, NULL };
g_test_summary ("Test that attaching a source to a task will set the source’s name if unset");
task = g_task_new (NULL, NULL, task_complete_cb, &calls);
g_task_set_name (task, "test name");
source = g_source_new (&source_funcs, sizeof (GSource));
g_task_attach_source (task, source, source_cb);
g_assert_cmpstr (g_source_get_name (source), ==, "test name");
g_source_unref (source);
source = g_source_new (&source_funcs, sizeof (GSource));
g_source_set_name (source, "not the task name");
g_task_attach_source (task, source, source_cb);
g_assert_cmpstr (g_source_get_name (source), ==, "not the task name");
g_source_unref (source);
g_object_unref (task);
}
gtask: Emit a debug message if a GTask is finalised without returning This typically indicates a bug in the program, where a GTask has been created, but a bug in the control flow has caused it to not return a value. There is one situation where it might be legitimate to finalise a GTask without returning: if an error happens in your *_async() start function after you’ve created a GTask, but before the async operation returns to the main loop; and you report the error using g_task_report_*error() rather than reporting it using the newly constructed GTask. Another situation is where you are just using GTask as a convenient way to move some work to another thread, without the complexity of creating and running your own thread pool. GDBus does this with g_dbus_interface_method_dispatch_helper(), for example. In most other cases, it’s a bug. Emit a debug message about it, but not a full-blown warning, as that would create noise in the legitimate cases. Signed-off-by: Philip Withnall <withnall@endlessm.com>
2018-10-07 18:15:49 +02:00
static void
test_finalize_without_return (void)
{
GTask *task = NULL;
guint n_calls = 0;
/* With a callback set. */
task = g_task_new (NULL, NULL, task_complete_cb, &n_calls);
g_test_expect_message (G_LOG_DOMAIN, G_LOG_LEVEL_CRITICAL,
"GTask * (source object: *, source tag: *) finalized without "
"ever returning (using g_task_return_*()). This potentially "
"indicates a bug in the program.");
g_object_unref (task);
g_test_assert_expected_messages ();
/* With a callback and task name set. */
task = g_task_new (NULL, NULL, task_complete_cb, &n_calls);
g_task_set_static_name (task, "oogly boogly");
g_test_expect_message (G_LOG_DOMAIN, G_LOG_LEVEL_CRITICAL,
"GTask oogly boogly (source object: *, source tag: *) finalized without "
"ever returning (using g_task_return_*()). This potentially "
"indicates a bug in the program.");
g_object_unref (task);
g_test_assert_expected_messages ();
/* Without a callback set. */
task = g_task_new (NULL, NULL, NULL, NULL);
g_test_expect_message (G_LOG_DOMAIN, G_LOG_LEVEL_DEBUG,
"GTask * (source object: *, source tag: *) finalized without "
"ever returning (using g_task_return_*()). This potentially "
"indicates a bug in the program.");
g_object_unref (task);
g_test_assert_expected_messages ();
}
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
int
main (int argc, char **argv)
{
int ret;
g_test_init (&argc, &argv, NULL);
loop = g_main_loop_new (NULL, FALSE);
main_thread = g_thread_self ();
magic = g_get_monotonic_time ();
g_test_add_func ("/gtask/basic", test_basic);
g_test_add_func ("/gtask/error", test_error);
g_test_add_func ("/gtask/return-from-same-iteration", test_return_from_same_iteration);
g_test_add_func ("/gtask/return-from-toplevel", test_return_from_toplevel);
g_test_add_func ("/gtask/return-from-anon-thread", test_return_from_anon_thread);
g_test_add_func ("/gtask/return-from-wrong-thread", test_return_from_wrong_thread);
g_test_add_func ("/gtask/no-callback", test_no_callback);
g_test_add_func ("/gtask/report-error", test_report_error);
g_test_add_func ("/gtask/priority", test_priority);
gtask: Add a g_task_set_name() method Similarly to g_source_set_name(), this sets a name on a GTask for debugging and profiling. Importantly, this name is propagated to the GSource for idle callbacks for the GTask, ending the glorious reign of `[gio] complete_in_idle_cb`. The name can be queried using g_task_get_name(). Locking is avoided by only allowing the name to be set before the GTask is used from another thread. Includes tests. Signed-off-by: Philip Withnall <withnall@endlessm.com>
2018-10-07 18:37:58 +02:00
g_test_add_func ("/gtask/name", test_name);
tests: Tweak the task tests Tweak the g_task_set_name test to use a non-literal string, so we don't test g_task_set_static_name twice.
2022-09-22 00:12:00 +02:00
g_test_add_func ("/gtask/name/macro-wrapper", test_name_macro_wrapper);
tests: Add a test for g_task_set_static_name() The test checks that the original string is returned by g_task_get_name().
2022-09-11 21:11:34 +02:00
g_test_add_func ("/gtask/static-name", test_static_name);
gio/tests/task: Ensure that cancelled tasks are returned asynchronously Closes https://gitlab.gnome.org/GNOME/glib/issues/1608
2018-12-04 19:54:10 +01:00
g_test_add_func ("/gtask/asynchronous-cancellation", test_asynchronous_cancellation);
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
g_test_add_func ("/gtask/check-cancellable", test_check_cancellable);
g_test_add_func ("/gtask/return-if-cancelled", test_return_if_cancelled);
g_test_add_func ("/gtask/run-in-thread", test_run_in_thread);
g_test_add_func ("/gtask/run-in-thread-sync", test_run_in_thread_sync);
g_test_add_func ("/gtask/run-in-thread-priority", test_run_in_thread_priority);
gtask: don't deadlock when tasks block on other tasks If tasks block waiting for other tasks to complete then the system can end up starved for threads. Avoid this by bumping up max-threads in that case. This also reverts 7b1f8c58 and reverts max-threads for GTask's GThreadPool back to 10. https://bugzilla.gnome.org/show_bug.cgi?id=687223
2012-12-15 17:44:59 +01:00
g_test_add_func ("/gtask/run-in-thread-nested", test_run_in_thread_nested);
gtask: remove hardcoded GTask thread-pool size GTask used a 10-thread thread pool for g_task_run_in_thread() / g_task_run_in_thread_sync(), but this ran into problems when task threads blocked waiting for another g_task_run_in_thread_sync() operation to complete. Previously there was a workaround for this, by bumping up the thread limit when that case was detected, but deadlocks could still happen if there were non-GTask threads involved. (Eg, task A sends a message to thread X and waits for a response, but thread X needs to complete task B in a thread before returning the response to task A.) So, allow GTask's thread pool to be expanded dynamically, by watching it from the glib worker thread, and growing it (at an exponentially-decreasing rate) if too much time passes without any tasks completing. This should solve the deadlocking problems without causing sudden breakage in apps that assume they can queue huge numbers of tasks at once without consequences. https://bugzilla.gnome.org/show_bug.cgi?id=687223
2015-03-09 21:33:16 +01:00
g_test_add_func ("/gtask/run-in-thread-overflow", test_run_in_thread_overflow);
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
g_test_add_func ("/gtask/return-on-cancel", test_return_on_cancel);
g_test_add_func ("/gtask/return-on-cancel-sync", test_return_on_cancel_sync);
g_test_add_func ("/gtask/return-on-cancel-atomic", test_return_on_cancel_atomic);
g_test_add_func ("/gtask/return-pointer", test_return_pointer);
task: Add return/propagate API for GValue This is useful for bindings. Related: https://gitlab.gnome.org/GNOME/glib/issues/668
2019-11-05 09:54:01 +01:00
g_test_add_func ("/gtask/return-value", test_return_value);
gtask: Add g_task_return_prefixed_error() An equivalent to g_propagate_prefixed_error, but for errors sent though GTask, allowing tasks to easily provide extra context to callers.
2023-10-31 04:49:28 +01:00
g_test_add_func ("/gtask/return-prefixed-error", test_return_prefixed_error);
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
g_test_add_func ("/gtask/object-keepalive", test_object_keepalive);
g_test_add_func ("/gtask/legacy-error", test_legacy_error);
tests: Add return ordering tests for GTask These are based on an example test program provided by Will Thompson in issue #1525. Signed-off-by: Philip Withnall <withnall@endlessm.com> https://gitlab.gnome.org/GNOME/glib/issues/1525
2018-10-01 22:45:04 +02:00
g_test_add_func ("/gtask/return/in-idle/error-first", test_return_in_idle_error_first);
g_test_add_func ("/gtask/return/in-idle/value-first", test_return_in_idle_value_first);
g_test_add_func ("/gtask/return/error-first", test_return_error_first);
g_test_add_func ("/gtask/return/value-first", test_return_value_first);
gtask: Don't overwrite source names Only set a name in g_task_attach_source if the source does not already have one. Including a new test by Philip Withnall.
2022-08-28 03:32:46 +02:00
g_test_add_func ("/gtask/attach-source/set-name", test_attach_source_set_name);
gtask: Emit a debug message if a GTask is finalised without returning This typically indicates a bug in the program, where a GTask has been created, but a bug in the control flow has caused it to not return a value. There is one situation where it might be legitimate to finalise a GTask without returning: if an error happens in your *_async() start function after you’ve created a GTask, but before the async operation returns to the main loop; and you report the error using g_task_report_*error() rather than reporting it using the newly constructed GTask. Another situation is where you are just using GTask as a convenient way to move some work to another thread, without the complexity of creating and running your own thread pool. GDBus does this with g_dbus_interface_method_dispatch_helper(), for example. In most other cases, it’s a bug. Emit a debug message about it, but not a full-blown warning, as that would create noise in the legitimate cases. Signed-off-by: Philip Withnall <withnall@endlessm.com>
2018-10-07 18:15:49 +02:00
g_test_add_func ("/gtask/finalize-without-return", test_finalize_without_return);
GTask: new GAsyncResult implementation / threaded task manager GTask is a replacement for GSimpleAsyncResult and GIOScheduler, that also allows for making cancellable wrappers around non-cancellable functions (as in GThreadedResolver). https://bugzilla.gnome.org/show_bug.cgi?id=661767
2011-10-05 16:05:50 +02:00
ret = g_test_run();
g_main_loop_unref (loop);
return ret;
}
Reference in New Issue Copy Permalink