Re: GTK app misbehaving. (main thread blocking)

[Alexander Markley <alex cyberMalex com>]

> The simplest would be something like this:
>
> gboolean
> status_msg (gchar *msgdup)
> {
>  .... update status bar ...
>   g_free(msgdup);
>   return FALSE; /* only call once */
> }
>
>
> work_thread_func ()
> {
>    ....
>    g_idle_add (status_msg, g_strdup_printf("current value = %u", 99);
>    ....
> }

Excellent. That fixed the whole problem.

Thanks a bunch.
--Alex

Tim Müller wrote:
> On Saturday 24 April 2004 20:28, you wrote:
>
>
> > I know, and I tried. The problem is, of course, that things like status
> > bar push and pop fail when called from the work thread. (Now that I know
> > that GTK is not thread safe, that explains the erratic/unreliable
> > behavior that came from trying that.)
> >
> > I wonder if I could simply force my work thread to emit signals which
> > cause the GTK thread to change the statusbar?
> >
> > Perhaps not.
>
>
> The simplest would be something like this:
>
> gboolean
> status_msg (gchar *msgdup)
> {
>  .... update status bar ...
>   g_free(msgdup);
>   return FALSE; /* only call once */
> }
>
>
> work_thread_func ()
> {
>    ....
>    g_idle_add (status_msg, g_strdup_printf("current value = %u", 99);
>    ....
> }
>
> GLib, in particular the GLib main loop stuff, is thread-safe, so when you do a 
> simple g_idle_add() or g_timeout_add() from your work thread, the dispatch 
> will actually be happening from the main context/thread. That's the easiest 
> way to get stuff across to the main thread.
>
> Cheers
> -Tim