GNU bug report logs -
#20802
Segfault when showing non-GTK+ tooltip
Previous Next
Reported by: Tobias Getzner <tobias.getzner <at> gmx.de>
Date: Sat, 13 Jun 2015 09:19:02 UTC
Severity: normal
Tags: moreinfo
Done: Lars Ingebrigtsen <larsi <at> gnus.org>
Bug is archived. No further changes may be made.
Full log
Message #38 received at 20802 <at> debbugs.gnu.org (full text, mbox):
> Date: Mon, 15 Jun 2015 10:22:40 +0200
> From: martin rudalics <rudalics <at> gmx.at>
> CC: tobias.getzner <at> gmx.de, 20802 <at> debbugs.gnu.org
>
> > Thanks. Now I know why this cannot be reproduced on Windows: this is
> > bug#17524 coming back to haunt us. That bug was reported on Windows,
> > I fixed it on Windows, then suggested a similar fix for X, but was
> > told it didn't help there.
> >
> > So now please try making a fix on X similar to commit ebdc80316, and
> > if that indeed doesn't help with this crash, perhaps some simple
> > variation of that will.
>
> Sorry. I must be missing something very elementary here. Your "shadow"
> refcounts
>
> #ifdef GLYPH_DEBUG
> static int image_cache_refcount, dpyinfo_refcount;
> #endif
>
> are defined iff you have GLYPH_DEBUG defined. How are these supposed to
> work when GLYPH_DEBUG is not defined?
It cannot, obviously. Bug#17524 was reported in the GLYPH_DEBUG code:
the assertion there segfaulted when it tried to access the image
cache. So the code was fixed only for that situation.
Doing that for a non-GLYPH_DEBUG code will require the shadow variable
to come out of that condition as well. Or use some other flag
variable to indicate that x_free_frame_resources is about to be called
when the refcount was not yet incremented.
IOW, that bug was mentioned as a source of ideas, not as something to
copy verbatim to xfns.c. Sorry if I was unclear about that.
This bug report was last modified 3 years and 32 days ago.
Previous Next
GNU bug tracking system
Copyright (C) 1999 Darren O. Benham,
1997,2003 nCipher Corporation Ltd,
1994-97 Ian Jackson.