Possible emergency gtk+ release

[Jonathan Blandford <jrb redhat com>]

Hi Tim,

Just before 2.0.1 was released, a patch went in that inadvertently broke
GtkTreeModelSort.  This is resulting in lots of bugs in the current
GNOME Beta from people who use that object.  We can do two things:

1) Quickly put out a 2.0.1.1 that fixes this bug

2) Put out a 2.0.2

Owen is running out to catch a flight to GUADEC, and didn't have a
strong opinion which option to choose.  He did not want to increment
GLib's version number without changes though, so if we change to 2.0.2
we'll be out of sync with GLib.

Tim, if we make a 2.0.1.1 release, is changing VERSION to explicitly be
2.0.1.1 sufficient in configure.in (while keeping GTK_MAJOR_VERSION,
GTK_MINOR_VERSION, and GTK_MICRO_VERSION at 2, 0, 1).  Will this cause
other problems?

For those who cannot wait for a new release, and unwilling to get a fix
from CVS, here's the patch to fix the problem.

Thanks,
-Jonathan

--- gtk+-2.0.1/gtk/gtktreemodelsort.c.orig	Mon Apr  1 23:02:06 2002
+++ gtk+-2.0.1/gtk/gtktreemodelsort.c	Mon Apr  1 23:03:36 2002
@@ -804,6 +804,7 @@
 
       gtk_tree_model_sort_sort_level (tree_model_sort, level,
 				      FALSE, FALSE);
+      gtk_tree_model_sort_increment_stamp (tree_model_sort);
 
       if (gtk_tree_path_get_depth (path))
 	{
@@ -1511,6 +1512,7 @@
 
   if (emit_reordered)
     {
+      gtk_tree_model_sort_increment_stamp (tree_model_sort);
       if (level->parent_elt)
 	{
 	  iter.stamp = tree_model_sort->stamp;
@@ -1533,8 +1535,6 @@
 
       gtk_tree_path_free (path);
     }
-
-  gtk_tree_model_sort_increment_stamp (tree_model_sort);
 
   /* recurse, if possible */
   if (recurse)