Gentoo Websites Logo
Go to: Gentoo Home Documentation Forums Lists Bugs Planet Store Wiki Get Gentoo!
Bug 11880 - pango-1.1.13 and perhaps xft2 making gnome2 apps crash
Summary: pango-1.1.13 and perhaps xft2 making gnome2 apps crash
Status: RESOLVED FIXED
Alias: None
Product: Gentoo Linux
Classification: Unclassified
Component: [OLD] GNOME (show other bugs)
Hardware: All Linux
: High normal
Assignee: Gentoo Linux Gnome Desktop Team
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2002-12-09 23:43 UTC by Jim Bray
Modified: 2002-12-10 09:40 UTC (History)
0 users

See Also:
Package list:
Runtime testing required: ---


Attachments

Note You need to log in before you can comment on or make changes to this bug.
Description Jim Bray 2002-12-09 23:43:50 UTC
I'm running gentoo/unstable, and recently all of gnome-2 became unusable. 
It looks like a pango-1.1.13 and xft problem. 
 
 I tried going back to older pango, but no luck: 
 
x11-libs/pango-1.0.5 * 
 
#gnome-terminal 
gnome-terminal: error while loading shared libraries: 
libpangoxft-1.0.so.0: cannot open shared object file: No such file or 
directory 
 
With current pango: 
 
pango-1.1.13 
 
#gnome-terminal 
 
** (gnome-terminal:23039): CRITICAL **: file pango-fontset.c: line 84 
(pango_fontset_get_font): assertion fontset != NULL' failed 
 
 I looked around the pango code a little, and found a possible bug, reported 
here: 
 
http://bugzilla.gnome.org/show_bug.cgi?id=100801 
 
and described below. 
 
 
 
bt: 
 
#0  0x40a426c1 in kill () from /lib/libc.so.6 
#1  0x4096d1be in pthread_kill () from /lib/libpthread.so.0 
#2  0x4096d503 in raise () from /lib/libpthread.so.0 
#3  0x40a43985 in abort () from /lib/libc.so.6 
#4  0x409dd3ee in g_logv () from /usr/lib/libglib-2.0.so.0 
#5  0x409dd422 in g_log () from /usr/lib/libglib-2.0.so.0 
#6  0x4060fce3 in pango_shape () from /usr/lib/libpango-1.0.so.0 
....... 
 
Actual failure point (pango-fontset.c): 
 
PangoFont * 
pango_fontset_get_font (PangoFontset  *fontset, 
                                               guint          wc) 
 
g_return_val_if_fail (fontset != NULL, NULL); 
 
  return PANGO_FONTSET_GET_CLASS (fontset)->get_font (fontset, wc); 
  } 
 
#### 
pango-context.c: 
 
        if (current_fonts) 
                          g_object_unref (current_fonts); 
 
 
current_fonts = pango_font_map_load_fontset (context->font_map, 
  context, 
                                                           current_desc, 
  language); 
........ 
######### 
NOTE: current-fonts apparently never NULL-tested, but can be NULL! 
######### 
analysis->font = pango_fontset_get_font (current_fonts, wc); 
 
###### 
pango-fontmap.c: 
 
 * 
  *   **/ 
   PangoFontset * 
   pango_font_map_load_fontset (PangoFontMap                 *fontmap, 
  PangoContext                 *context, 
                                                     const 
PangoFontDescription   *desc, 
                                     PangoLanguage                *language) 
 
 Returns the fontset, or %NULL if no font matched.
Comment 1 foser (RETIRED) gentoo-dev 2002-12-10 07:01:10 UTC
This is exactly what we don't want, you taking your local problem to upstream
while i know what the problem is. Please, do not file bugs upstream unless you
are _certain_ it is not Gentoo related and have consulted our bugzilla. I'm
quite certain there are bugs around with the same problems here.

Anyway, as owen said. You probably don't have any fonts configured . So try and
run 'fc-cache -v' see if any fonts are added. fc-list to check your current
fonts. If still not, attach your /etc/fonts/fonts.conf .
Comment 2 Jim Bray 2002-12-10 09:24:18 UTC
 The lack of a NULL-return check in the Pango source code was an upstream problem, not a Gentoo problem. Upstream says the current CVS pango gives a useful error message instead of crashing. Maybe you should fetch their CVS for gentoo-unstable.   The bug was legitimate, and had at least two other reporters.  http://bugzilla.gnome.org/show_bug.cgi?id=98672    Running fc-cache -v fixes the problem.    
Comment 3 foser (RETIRED) gentoo-dev 2002-12-10 09:40:40 UTC
The point im trying to make is that you would be better of asking/informing us
first. We would take it upstream if it turned out to be a problem in the source,
instead of a dozen of our users fileing the same bug there (hypotetical).

Using a CVS version for a bit more descriptive error msg, not a good idea imo.
It might not officialy crash then, it still won't run. It is in the end mostly
just a configuration problem, you should never have 0 fonts.