Hello everyone. I made a small patch to correct the odd 'apply' behaviour on the GTK client, mainly that you can't unapply a container (switches from applied/opened to applied). What i did is simple (i'll explain even if code is pretty straightforward): I split gtk/gx11.c:close_container into 2 functions, close_container & do_close_container The first one is the reply to the server event 'container closed', sent to signal a container goes from opened to closed (and unapplied, actually). This one just clears locally the display list. The second handles the mouse click on the 'close' button, and just sends a client command 'apply <container>' to the server to ask for container close (this will generate a server event 'container closed', thus close_container gets called right after that). Ok, important point: I couldn't test that patch. Running under Windows, and i didn't want to mess with mingw/cygwin to try to compile the code. I do think my patch is valid (even some tabs may be weird), though. If i'm wrong on that, just kick me :-) Oh, and something else: if you look at the X11 client code, you'll see the 'close_container' is exactly how i made the GTK like, just local changes. Also the Gnome client prolly has the same bug as the GTK, since the close_container handles both the server event & the client button click. Finally, i'm sorry if the patch is weird, i couldn't find the correct diff option to strip first directory name... Thanks to all developers for the game ^_- Nicolas 'Ryo' -------------- next part -------------- diff -r crossfire-cvs-original/client/gtk/gx11.c crossfire-cvs/client/gtk/gx11.c 2761c2761 < void close_container (item *op) --- > void do_close_container (item *op) 2764a2765,2769 > } > } > > void close_container (item *op) > { 2768d2772 < } 3044c3048 < GTK_SIGNAL_FUNC(close_container), --- > GTK_SIGNAL_FUNC(do_close_container),