On 30 April 2013 10:30, Patrick Ohly <patrick.ohly@xxxxxxxxx> wrote:
> > In my case, I'm normally processing some data in a thread using library
> > calls, then when it finishes I want to display the data to the user.
>
> That implies going fully multithreaded, including explicit passing of
> information back and forth between threads. I was hoping to avoid that.
I agree with the other posters in the thread. Keep the glib main loop
always running in a thread somewhere, don't use nested glib main loops
(g_main_context_iteration() and friends), do all the heavy work in
other threads, and have them communicate via the glib thread with
g_idle_add().
You can do this very simple and reliably. For example:
worker()
{
char *str;
for(;;) {
str = g_strdup("hello world!\n");
g_idle_add(from_worker_cb, str);
sleep(1);
}
}
gboolean
from_worker(char *str)
{
update_textview(str);
g_free(str);
return FALSE;
}
I maintain a fairly large (>250kloc) gtk program done in this style,
calling into heavily-threaded, asynchronous libraries, and it works
well.
John
_______________________________________________
gtk-list mailing list
gtk-list@xxxxxxxxx
https://mail.gnome.org/mailman/listinfo/gtk-list
