Re: missing GMainContext methods ... (fwd)

[Tim Janik <timj imendio com>]

On Thu, 22 Dec 2005, michael meeks wrote:

> ---------- Forwarded message ----------
> Date: Tue, 20 Dec 2005 17:21:07 +0000
> From: michael meeks <michael meeks novell com>
> To: Federico Mena Quintero <federico novell com>, Tim Janik 
> <timj imendio com>
> Cc: Gtk Hackers <gtk-devel-list gnome org>
> Subject: missing GMainContext methods ...
>
> So,
>
> 	I sent this a while back - but didn't see it on the list - perhaps
> broken subscription information (?).
>
> 	I append a patch implementing these two to some level.
>
> On Thu, 2005-11-17 at 16:57 +0000, michael meeks wrote:
> > 	So - I've been trying to use the GMainContext to fix a rather tricky
> > issue in using unsafe single-threaded code accessed via ORBit2 from
> > multiple OO.o threads in a safe & reliable way. This is somewhat
> > involved, for various reasons, but made particularly unpleasant due to 2
> > missing methods:
> >
> > 	a) gboolean g_main_context_is_owner (GMainContext *context);
> > 		+ this would tell you if the current thread owns the
> > 		  g_main_context. NB. this is subtly different from
> > 		  something like:
> > 			if (g_main_context_acquire ()) {
> > 				g_main_context_release();
> > 				... I own it - horay ...
> > 			}
> > 		+ since that actually transiently takes ownership of
> > 		  the thread but doesn't tell you if you (now) contine
> > 		  to own it [ due to some (much) higher stack frame
> > 		  having taken that lock ].

i think this should go in, with one modification:

+gboolean
+g_main_context_is_owner (GMainContext *context)
+{
+  gboolean is_owner;
+
+  if (!context)
+    context = g_main_context_default ();
+
+#ifdef G_THREADS_ENABLED
+  LOCK_CONTEXT (context);
+  is_owner = context->owner == G_THREAD_SELF;
+  UNLOCK_CONTEXT (context);
+#else
+  is_owner = TRUE;
+#endif
+
+  return is_owner;
+}

if (!context), the function should simply return FALSE, instead of
forcefully create a context. following the basic principle of eliminating
side-effects on getters/queries.

> > 	b) void g_main_context_acquire_with_wakeup (GMainContext *context);
> > 		+ this cunning method - would be the analogue of
> > 		  g_main_context_wait () - except instead of sitting
> > 		  around hoping that the other thread doing the poll
> > 		  will wake-up, it does a (safe) g_main_context_wakeup
> > 		  with the relevant locks held;
> > 		+ currently it appears impossible/acutely-ugly to
> > 		  get ownership of the GMainContext from another thread
> > 		  if the main thread is in it's poll.

this looks good to me as is:

+void
+g_main_context_acquire_with_wakeup (GMainContext *context)
+{
+  gboolean got_ownership;
+
+  if (!context)
+    context = g_main_context_default ();
+
+  got_ownership = g_main_context_acquire (context);
+  while (!got_ownership)
+    {
+      LOCK_CONTEXT (context);
+      g_main_context_wakeup_unlocked (context);
+      got_ownership = g_main_context_wait (context,
+                                          context->cond,
+                                          g_static_mutex_get_mutex (&context->mutex));
+      UNLOCK_CONTEXT (context);
+    }
+}

> > 	The basic reasoning here is that - ~all existing event-driven gtk+
> > code, typically happens from the default GMainContext, and/or at least
> > that provides a reasonable back-compatible way to create a backwards
> > compatible 'apartment' (by holding the GMainContext lock) to execute old
> > code in in-line.
>
> 	I attach a simple patch; the g_main_context_is_owner impl. is of 
> course
> trivial - it'd be great to be able to commit that by itself. The
> 'acquire_with_wakeup' is also relatively trivial, cf. above.
>
> 	There is no bugzilla number, is that a problem ?

i don't think so. unless owen vetos in, i'll apply them tomorrow with
the change outlined above.

> 	Thanks,
>
> 		Michael.
>
> --
> michael meeks novell com  <><, Pseudo Engineer, itinerant idiot

---
ciaoTJ