Commit 8951f96c authored by Owen Taylor's avatar Owen Taylor Committed by Owen Taylor

Make some changes to the way that GMainContext works:

Sat Jun 30 15:49:10 2001  Owen Taylor  <otaylor@redhat.com>

	* glib/gmain.[ch]: (Mostly patch from Sebastian Wilhemi)
	Make some changes to the way that GMainContext works:

	 - a GMainContext is no longer associated with a single
	   thread, but any thread can acquire ownership
	   of thread and iterate.

	 - There is a facility g_main_context_wait() for
	   non-owner-threads to wait either for ownership
	   or for a condition to be broadcast.

	 - For efficiency, GMainLoop just piggybacks of
	   of the loops mutex / condition instead of
	   having a separate mutex/condition for each
	   GMainLoop.

	* glib/gthread.[ch]: Remove hacks to store the thread's
	GMainContext in the GThread structures, since we
	no longer have the GMainContext <=> GThread correspondence.

        * glib/gmain.[ch]: Make g_main_context_wakeup() public
	so someone could completely duplicate GMainLoop
	with the public API.

	* tests/mainloop-test: Fix up to the new API. Decidedly
	doesn't work at the moment, but that may be the IO
	channel changes, or preexisting locking problems.
parent c88e2529
Sat Jun 30 15:49:10 2001 Owen Taylor <otaylor@redhat.com>
* glib/gmain.[ch]: (Mostly patch from Sebastian Wilhemi)
Make some changes to the way that GMainContext works:
- a GMainContext is no longer associated with a single
thread, but any thread can acquire ownership
of thread and iterate.
- There is a facility g_main_context_wait() for
non-owner-threads to wait either for ownership
or for a condition to be broadcast.
- For efficiency, GMainLoop just piggybacks of
of the loops mutex / condition instead of
having a separate mutex/condition for each
GMainLoop.
* glib/gthread.[ch]: Remove hacks to store the thread's
GMainContext in the GThread structures, since we
no longer have the GMainContext <=> GThread correspondence.
* glib/gmain.[ch]: Make g_main_context_wakeup() public
so someone could completely duplicate GMainLoop
with the public API.
* tests/mainloop-test: Fix up to the new API. Decidedly
doesn't work at the moment, but that may be the IO
channel changes, or preexisting locking problems.
Sat Jun 30 13:18:28 2001 Owen Taylor <otaylor@redhat.com>
* glib/gstrfuncs.c glib/gstring.h: Try compiling
......
Sat Jun 30 15:49:10 2001 Owen Taylor <otaylor@redhat.com>
* glib/gmain.[ch]: (Mostly patch from Sebastian Wilhemi)
Make some changes to the way that GMainContext works:
- a GMainContext is no longer associated with a single
thread, but any thread can acquire ownership
of thread and iterate.
- There is a facility g_main_context_wait() for
non-owner-threads to wait either for ownership
or for a condition to be broadcast.
- For efficiency, GMainLoop just piggybacks of
of the loops mutex / condition instead of
having a separate mutex/condition for each
GMainLoop.
* glib/gthread.[ch]: Remove hacks to store the thread's
GMainContext in the GThread structures, since we
no longer have the GMainContext <=> GThread correspondence.
* glib/gmain.[ch]: Make g_main_context_wakeup() public
so someone could completely duplicate GMainLoop
with the public API.
* tests/mainloop-test: Fix up to the new API. Decidedly
doesn't work at the moment, but that may be the IO
channel changes, or preexisting locking problems.
Sat Jun 30 13:18:28 2001 Owen Taylor <otaylor@redhat.com>
* glib/gstrfuncs.c glib/gstring.h: Try compiling
......
Sat Jun 30 15:49:10 2001 Owen Taylor <otaylor@redhat.com>
* glib/gmain.[ch]: (Mostly patch from Sebastian Wilhemi)
Make some changes to the way that GMainContext works:
- a GMainContext is no longer associated with a single
thread, but any thread can acquire ownership
of thread and iterate.
- There is a facility g_main_context_wait() for
non-owner-threads to wait either for ownership
or for a condition to be broadcast.
- For efficiency, GMainLoop just piggybacks of
of the loops mutex / condition instead of
having a separate mutex/condition for each
GMainLoop.
* glib/gthread.[ch]: Remove hacks to store the thread's
GMainContext in the GThread structures, since we
no longer have the GMainContext <=> GThread correspondence.
* glib/gmain.[ch]: Make g_main_context_wakeup() public
so someone could completely duplicate GMainLoop
with the public API.
* tests/mainloop-test: Fix up to the new API. Decidedly
doesn't work at the moment, but that may be the IO
channel changes, or preexisting locking problems.
Sat Jun 30 13:18:28 2001 Owen Taylor <otaylor@redhat.com>
* glib/gstrfuncs.c glib/gstring.h: Try compiling
......
Sat Jun 30 15:49:10 2001 Owen Taylor <otaylor@redhat.com>
* glib/gmain.[ch]: (Mostly patch from Sebastian Wilhemi)
Make some changes to the way that GMainContext works:
- a GMainContext is no longer associated with a single
thread, but any thread can acquire ownership
of thread and iterate.
- There is a facility g_main_context_wait() for
non-owner-threads to wait either for ownership
or for a condition to be broadcast.
- For efficiency, GMainLoop just piggybacks of
of the loops mutex / condition instead of
having a separate mutex/condition for each
GMainLoop.
* glib/gthread.[ch]: Remove hacks to store the thread's
GMainContext in the GThread structures, since we
no longer have the GMainContext <=> GThread correspondence.
* glib/gmain.[ch]: Make g_main_context_wakeup() public
so someone could completely duplicate GMainLoop
with the public API.
* tests/mainloop-test: Fix up to the new API. Decidedly
doesn't work at the moment, but that may be the IO
channel changes, or preexisting locking problems.
Sat Jun 30 13:18:28 2001 Owen Taylor <otaylor@redhat.com>
* glib/gstrfuncs.c glib/gstring.h: Try compiling
......
Sat Jun 30 15:49:10 2001 Owen Taylor <otaylor@redhat.com>
* glib/gmain.[ch]: (Mostly patch from Sebastian Wilhemi)
Make some changes to the way that GMainContext works:
- a GMainContext is no longer associated with a single
thread, but any thread can acquire ownership
of thread and iterate.
- There is a facility g_main_context_wait() for
non-owner-threads to wait either for ownership
or for a condition to be broadcast.
- For efficiency, GMainLoop just piggybacks of
of the loops mutex / condition instead of
having a separate mutex/condition for each
GMainLoop.
* glib/gthread.[ch]: Remove hacks to store the thread's
GMainContext in the GThread structures, since we
no longer have the GMainContext <=> GThread correspondence.
* glib/gmain.[ch]: Make g_main_context_wakeup() public
so someone could completely duplicate GMainLoop
with the public API.
* tests/mainloop-test: Fix up to the new API. Decidedly
doesn't work at the moment, but that may be the IO
channel changes, or preexisting locking problems.
Sat Jun 30 13:18:28 2001 Owen Taylor <otaylor@redhat.com>
* glib/gstrfuncs.c glib/gstring.h: Try compiling
......
Sat Jun 30 15:49:10 2001 Owen Taylor <otaylor@redhat.com>
* glib/gmain.[ch]: (Mostly patch from Sebastian Wilhemi)
Make some changes to the way that GMainContext works:
- a GMainContext is no longer associated with a single
thread, but any thread can acquire ownership
of thread and iterate.
- There is a facility g_main_context_wait() for
non-owner-threads to wait either for ownership
or for a condition to be broadcast.
- For efficiency, GMainLoop just piggybacks of
of the loops mutex / condition instead of
having a separate mutex/condition for each
GMainLoop.
* glib/gthread.[ch]: Remove hacks to store the thread's
GMainContext in the GThread structures, since we
no longer have the GMainContext <=> GThread correspondence.
* glib/gmain.[ch]: Make g_main_context_wakeup() public
so someone could completely duplicate GMainLoop
with the public API.
* tests/mainloop-test: Fix up to the new API. Decidedly
doesn't work at the moment, but that may be the IO
channel changes, or preexisting locking problems.
Sat Jun 30 13:18:28 2001 Owen Taylor <otaylor@redhat.com>
* glib/gstrfuncs.c glib/gstring.h: Try compiling
......
Sat Jun 30 15:49:10 2001 Owen Taylor <otaylor@redhat.com>
* glib/gmain.[ch]: (Mostly patch from Sebastian Wilhemi)
Make some changes to the way that GMainContext works:
- a GMainContext is no longer associated with a single
thread, but any thread can acquire ownership
of thread and iterate.
- There is a facility g_main_context_wait() for
non-owner-threads to wait either for ownership
or for a condition to be broadcast.
- For efficiency, GMainLoop just piggybacks of
of the loops mutex / condition instead of
having a separate mutex/condition for each
GMainLoop.
* glib/gthread.[ch]: Remove hacks to store the thread's
GMainContext in the GThread structures, since we
no longer have the GMainContext <=> GThread correspondence.
* glib/gmain.[ch]: Make g_main_context_wakeup() public
so someone could completely duplicate GMainLoop
with the public API.
* tests/mainloop-test: Fix up to the new API. Decidedly
doesn't work at the moment, but that may be the IO
channel changes, or preexisting locking problems.
Sat Jun 30 13:18:28 2001 Owen Taylor <otaylor@redhat.com>
* glib/gstrfuncs.c glib/gstring.h: Try compiling
......
Sat Jun 30 15:49:10 2001 Owen Taylor <otaylor@redhat.com>
* glib/gmain.[ch]: (Mostly patch from Sebastian Wilhemi)
Make some changes to the way that GMainContext works:
- a GMainContext is no longer associated with a single
thread, but any thread can acquire ownership
of thread and iterate.
- There is a facility g_main_context_wait() for
non-owner-threads to wait either for ownership
or for a condition to be broadcast.
- For efficiency, GMainLoop just piggybacks of
of the loops mutex / condition instead of
having a separate mutex/condition for each
GMainLoop.
* glib/gthread.[ch]: Remove hacks to store the thread's
GMainContext in the GThread structures, since we
no longer have the GMainContext <=> GThread correspondence.
* glib/gmain.[ch]: Make g_main_context_wakeup() public
so someone could completely duplicate GMainLoop
with the public API.
* tests/mainloop-test: Fix up to the new API. Decidedly
doesn't work at the moment, but that may be the IO
channel changes, or preexisting locking problems.
Sat Jun 30 13:18:28 2001 Owen Taylor <otaylor@redhat.com>
* glib/gstrfuncs.c glib/gstring.h: Try compiling
......
This diff is collapsed.
......@@ -124,7 +124,9 @@ struct _GPollFD
/* GMainContext: */
GMainContext *g_main_context_get (GThread *thread);
GMainContext *g_main_context_new (void);
void g_main_context_ref (GMainContext *context);
void g_main_context_unref (GMainContext *context);
GMainContext *g_main_context_default (void);
gboolean g_main_context_iteration (GMainContext *context,
......@@ -143,6 +145,13 @@ GSource *g_main_context_find_source_by_funcs_user_data (GMainContext *conte
/* Low level functions for implementing custom main loops.
*/
void g_main_context_wakeup (GMainContext *context);
gboolean g_main_context_acquire (GMainContext *context);
void g_main_context_release (GMainContext *context);
gboolean g_main_context_wait (GMainContext *context,
GCond *cond,
GMutex *mutex);
gboolean g_main_context_prepare (GMainContext *context,
gint *priority);
gint g_main_context_query (GMainContext *context,
......@@ -177,6 +186,7 @@ void g_main_loop_quit (GMainLoop *loop);
GMainLoop *g_main_loop_ref (GMainLoop *loop);
void g_main_loop_unref (GMainLoop *loop);
gboolean g_main_loop_is_running (GMainLoop *loop);
GMainContext *g_main_loop_get_context (GMainLoop *loop);
/* GSource: */
......
......@@ -74,7 +74,6 @@ typedef struct _GRealThread GRealThread;
struct _GRealThread
{
GThread thread;
GMainContext *context;
gpointer private_data;
gpointer retval;
GSystemThread system_thread;
......@@ -465,8 +464,6 @@ g_static_private_free (GStaticPrivate *private_key)
G_UNLOCK (g_thread);
}
void g_main_context_destroy (GMainContext *context);
static void
g_thread_cleanup (gpointer data)
{
......@@ -487,8 +484,6 @@ g_thread_cleanup (gpointer data)
}
g_array_free (array, TRUE);
}
if (thread->context)
g_main_context_destroy (thread->context);
/* We only free the thread structure, if it isn't joinable. If
it is, the structure is freed in g_thread_join */
......@@ -560,7 +555,6 @@ g_thread_create_full (GThreadFunc func,
result->thread.func = func;
result->thread.data = data;
result->private_data = NULL;
result->context = NULL;
G_LOCK (g_thread);
G_THREAD_UF (thread_create, (g_thread_create_proxy, result,
stack_size, joinable, bound, priority,
......@@ -657,7 +651,6 @@ g_thread_self (void)
thread->thread.func = NULL;
thread->thread.data = NULL;
thread->private_data = NULL;
thread->context = NULL;
if (g_thread_supported ())
G_THREAD_UF (thread_self, (&thread->system_thread));
......
......@@ -133,7 +133,7 @@ adder_thread (gpointer data)
GIOChannel **channels = data;
AddrData addr_data;
context = g_main_context_get (g_thread_self());
context = g_main_context_new ();
g_mutex_lock (context_array_mutex);
......
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment