Re: breaking hard code freeze for bug-buddy

["Elijah Newren" <newren gmail com>]

On 9/1/06, Fernando Herrera <fherrera onirica com> wrote:
> Hello,
>
> We have an strange bug in bug-buddy: When the GtkTextView is shown
> automatically after gdb is finished, the internal bin->window pointer
> is not pointing to the mapped window, and calling
> gtk_widget_grab_focus on it makes gtk asserts (and crashing on
> 2.15.x).
>
> I can not reproduce this problem with a testcase for gtk+, so I'm not
> sure where the problem is.

Does this mean you have been unable to reproduce at all, or just
unable to reproduce outside bug-buddy?

> However I have a simple (and safe) workaround for that:
> calling the grab_focus function after the widget has been realized.
>
> This is the bug:
> http://bugzilla.gnome.org/show_bug.cgi?id=353348
>
> as you can see there are a lot of duplicates of it, so please, approve
> this break.

Yes, there are a lot of duplicates, so we need to watch this bug
closely.  I'm kind of curious whether you know the problem is actually
in blink_cb, though.  I know all the stack traces say that, but they
also claim that line 7590 of gtktextview.c is part of blink_cb, which
is about 3000 lines off.  I think the stack traces may be corrupt
here.  Could the problem be elsewhere?

Elijah