Bug#315083: libgtk2.0-0: breaks mozilla-firefox on amd64

[Loïc Minier]

        Hi,

On Tue, Jun 28, 2005, A Costa wrote:
> For example, it breaks 'gnome-terminal':
>     % gnome-terminal
>     (gnome-terminal:8130): Gdk-WARNING **: gdk_property_get(): length value has wrapped in calculation (did you pass G_MAXLONG?)
> ...the above error message repeats for 7 screens, while GT pops up blank,
> and has to be killed.  Another example:

 That's a libvte bug, fixed in vte 1:0.11.13-3:
vte (1:0.11.13-3) unstable; urgency=low

  * Fix Gdk warnings with a patch from upstream. (Closes: #316940, #318549)
    [debian/patches/20_gdk-warning.patch]

>     % querybts -w gnome-terminal
>     /usr/bin/sensible-browser
>     (x-www-browser:4270): Gdk-WARNING **: gdk_property_get(): length value has wrapped in calculation (did you pass G_MAXLONG?)
> ...(on my system 'x-www-browser' is 'galeon'.
>     % galeon ; echo $?
>     (galeon:8153): Gdk-WARNING **: gdk_property_get(): length value has wrapped in calculation (did you pass G_MAXLONG?)

 That's a mozilla bug (galeon uses mozilla).  See #318904, and #318779.

> ...'epiphany' is also broken:
>     % epiphany
>     (epiphany:8213): Gdk-WARNING **: gdk_property_get(): length value has wrapped in calculation (did you pass G_MAXLONG?)
> ...note the same line number in G & E, "8213".  'firefox' does
> start up, though with a similar but not identical error:

 Epiphany uses mozilla too.

>     % firefox 
>     (firefox-bin:8222): Gdk-WARNING **: gdk_property_get(): length value has wrapped in calculation (did you pass G_MAXLONG?)

 Firefox has been fixed in 1.0.6-1:
mozilla-firefox (1.0.6-1) unstable; urgency=low
  ...
  * gfx/src/gtk/nsScreenGtk.cpp: Patch from Loic Minier to fix
    gdk_property_get warnings.
  * widget/src/gtk2/mozdrawingarea.c: Patch from Loic Minier to fix
    crashes under GTK 2.7. (Closes: #318903)

> Worst of all, I can't start 'X' with 'gnome-sawfish' -- it's just a
> blank screen, no icons, capplets, menus, and the mouse buttons don't
> work.  (I feel grateful to have other window managers to fall back
> on...)
> A web search for similar errors turns up this:
>     #315462 sunclock: Segfaults
>     http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=315462
> ...which might be merged with this bug.  Of course the severity should
> be changed.  If it turns out to be hard to fix, maybe it should be
> cloned to the above packages.

 If you're willing to fix these issues too, have a look in depth at the
 upstream programs and check whether it's the same bug as in the other
 fixed apps, and fix them.  Do not merge back here.

   Cheers,

-- 
Loïc Minier <lool at dooz.org>