Re: cursor refcounts

[Havoc Pennington <hp redhat com>]

Tim Janik <timj@gtk.org> writes:
 
> that's because you code didn't include the assertments we have
> everywhere

There was a cursor bug which triggered your g_return_if_fail (I just
checked in a fix that removes the warnings) but I still get the
segfaults, and I can't trace them to my patch. It is in the drawable
code and there's a big old FIXME comment right in front of the
segfault.

Here is where it breaks:

322       /* FIXME: this doesn't work because bitmaps don't have visuals */
323       if (gdk_drawable_get_visual (src)->depth == 1)

If you can give me even a guess what this has to do with my cursor
patch I will be happy to fix it. But as it is I maintain that you or
Owen checked in segfaulting code. ;-)

> it'd be nice if you could test new code at least to a point where it doesn't
> segfault/warn anymore before comitting.
> 

I did test it, but the segfault I got didn't look (and still doesn't
look) related to my patch. And I got no warnings since the
g_return_if_fail weren't there. And at least one of the GTK
maintainers told me to commit it without suggesting adding the
checks. So there. :-P

Havoc