Re: glib hash table assertion failures: need help in debuging

["Jasper St. Pierre" <jstpierre mecheye net>]

On Mon, Jun 4, 2012 at 2:18 AM, Akash Rawal <argmlac gmail com> wrote:
> Hi all,
>
> I have a GHashTable related question.
>
> I am writing a volume monitor for GIO.
>
> I have a code like this:
>
> 186     void pup_conv_mgr_close_all_convs(PupConvMgr *cmgr,
> PupConvCloseFlags flags)
> 187     {
> 188             GHashTableIter iter;
> 189             PupConv *oneconv;
> 190             g_hash_table_iter_init(&iter, cmgr->self_convs);
> 191             while(g_hash_table_iter_next(&iter, NULL, (gpointer
> *)&oneconv))
> 192             {
> 193                     pup_conv_close(oneconv, flags | PUP_CONV_NOREMOVE);
> 194                     if (flags&  PUP_CONV_FREE)
> 195                             g_hash_table_iter_remove(&iter);
> 196             }
> 197             g_hash_table_iter_init(&iter, cmgr->other_convs);
> 198             while(g_hash_table_iter_next(&iter, NULL, (gpointer
> *)&oneconv))
> 199             {
> 200                     pup_conv_close(oneconv, flags | PUP_CONV_NOREMOVE);
> 201                     if (flags&  PUP_CONV_FREE)
> 202                             g_hash_table_iter_remove(&iter);
> 203             }
> 204     }
>
> Exactly at line 198 I get a warning like:
>
> (pup-volume-monitor:4276): GLib-CRITICAL **: g_hash_table_iter_next:
> assertion `ri->version == ri->hash_table->version' failed

When you do g_hash_table_iter_remove, it increments a field on the
iterator and the hash table to ensure that the two don't get out of
sync. Removing while iterating is sort of a difficult problem. Are you
removing something else from the same hash table in your destroy
function that you passed to g_hash_table_init ?

> Although nothing worse happens, but I am curious to know why I get this.
>
> Thanks in advance.
>
> _______________________________________________
> gtk-list mailing list
> gtk-list gnome org
> https://mail.gnome.org/mailman/listinfo/gtk-list



-- 
  Jasper