Re: GNOME CVS: gtk+ hp

[Havoc Pennington <hp redhat com>]

Havoc Pennington <hp redhat com> writes:
> jacob berkman <jacob ximian com> writes:
> > On Wed, 2001-11-28 at 15:42, Gnome CVS User wrote:
> > > 
> > > CVSROOT:	/cvs/gnome
> > > Module name:	gtk+
> > > Changes by:	hp	01/11/28 15:42:15
> > > 
> > > Modified files:
> > > 	.              : ChangeLog 
> > > 	gtk            : gtkrange.c gtktextview.c 
> > > 
> > > Log message:
> > > 2001-11-28  Havoc Pennington  <hp redhat com>
> > > 
> > > * gtk/gtktextview.c (gtk_text_view_paint): put the warning
> > > explaining the validation error in front of the G_BREAKPOINT
> > > 
> > > * gtk/gtkrange.c (gtk_range_adjustment_value_changed): call
> > > process updates with include_children = FALSE instead of TRUE, so
> > > we redraw the range only, not the whole darn window.  Making
> > > GtkRange a NO_WINDOW widget had weirdly caused the text widget
> > > Trace/breakpoint trap bug to become much easier to trigger.
> > 
> > this did not fix it for me - i still get the failed assertion in
> > bug-buddy when i hit backspace in the text view.
> > 
> 
> bug-buddy doesn't crash for me if I hit backspace... urgh.
> 

Are you sure you have the G_BREAKPOINT() in gtktextview.c, vs. this
assertion described here:
 
  http://bugzilla.gnome.org/show_bug.cgi?id=65731

I was fixing the gtktextview.c thing yesterday, that's what I thought
you were talking about.

Vitaly's report this morning is the first one that seems to track down
what's different when the bug happens vs. when it doesn't in testtext.

Havoc