Outstanding issues

[Owen Taylor <otaylor redhat com>]

Here's what's left on blocking issues for 1.3.11:

 * GTK_OBJECT_DESTROYED

     http://mail.gnome.org/archives/gtk-devel-list/2001-November/msg00477.html

   Trivial to do, though if we do it it will cause some disruption
   downstream as people are forced to fix code now rather than
   spending days debugging it later.

 * Alex's Pango patch

     http://mail.gnome.org/archives/gtk-devel-list/2001-November/msg00329.html
   
   Ready to commit

 * TreeViewColumn size mode change

     http://mail.gnome.org/archives/gtk-devel-list/2001-November/msg00488.html

 * Remaining finalization of accel group changes from earlier this week

     http://mail.gnome.org/archives/gtk-devel-list/2001-November/msg00439.html

   There are basically two parts of this:

    1) Some trivial renaming action, small API fixes
 
    2) The possibility of reworking the connection between GtkAccelGroup
       and GtkAccelMap to avoid leaking out into the public API.

 * Changing GType to gsize to allow avoiding locking on
   TypeNode lookups.

   The API change here is simply to change types so sizeof(GType) ==
   sizeof(gpointer) and remove some not useful macros; this
   is the minimum change for 1.3.11.

   The actual implementation of hiding GTypeNode pointers inside
   GType ids is more complicated and needs to be redone due to
   changes in GObject.

 * closure/default handler chaining for language bindings

   The API is there, the implementation isn't. We're most likely
   going to have to let the implementation slide past 1.3.11 to not
   hold everything up. (Tim estimates end of the week for
   the implementation)

If we "cheat" on the last two issues, then part 2) of the accel group
changes is the only thing that involves significant work.  I'll take a
look at that later today; if it looks like it will be more than a few
hours work or just doesn't work out, we can forget it and go with 
the current way.

Regards,
                                        Owen