On Fri, Sep 14, 2012 at 11:11 AM, Emmanuele Bassi <ebassi@xxxxxxxxx> wrote:
[ .... a lot of very confusing stuff ... ]
are you saying that g_idle_add() is not thread-safe? because if it IS threadsafe, and the rest of program either
(a) only uses g_idle_add() to get work done in the main context
OR
(b) religiously uses enter/leave to do GDK/GTK work in other threads
then i don't see what the problem is. i'm fairly sure that we've been given assurances before that g_idle_add() is thread safe, and by implication, the callback it queues up will ALWAYS be executed inside the main context. you seem to be suggesting that executing within the main context is not enough to protect against other threads using enter/leave to do GDK/GTK work ... is that your point?
[ .... a lot of very confusing stuff ... ]
are you saying that g_idle_add() is not thread-safe? because if it IS threadsafe, and the rest of program either
(a) only uses g_idle_add() to get work done in the main context
OR
(b) religiously uses enter/leave to do GDK/GTK work in other threads
then i don't see what the problem is. i'm fairly sure that we've been given assurances before that g_idle_add() is thread safe, and by implication, the callback it queues up will ALWAYS be executed inside the main context. you seem to be suggesting that executing within the main context is not enough to protect against other threads using enter/leave to do GDK/GTK work ... is that your point?
_______________________________________________ gtk-list mailing list gtk-list@xxxxxxxxx https://mail.gnome.org/mailman/listinfo/gtk-list
