GNOME Bugzilla – Bug 82321
sawfish-ui exits when browsing fonts
Last modified: 2009-08-16 15:13:28 UTC
I attempted to browse a different font from: Appearance - Appearance - Default font. When pressing the ``browse'' window, sawfish-ui crashed with the following messages: -------------------------- ** (rep:18259): WARNING **: Couldn't load font "fixed Medium Semi-Condensed 0" falling back to "Sans Medium Semi-Condensed 0" ** (rep:18259): WARNING **: Couldn't load font "Sans Medium Semi-Condensed 0" falling back to "Sans 0" ** (rep:18259): WARNING **: All font failbacks failed!!!! -------------------------- sawfish runs OK on my machine (I only wanted to change the font...), so something is weird that it has no fallbacks.
An actual crash is critical/high but without a way to reproduce this, I can't really imagine that this is very useful for anyone. :/
I can reproduce this 100% with garnome 0.10.1 How can I help debug it ? I also get many gtk-critical assertions with sawfish-ui
hello ?? What do you need ? xlsfonts, what ?
sawfish can only load traditional X fonts. This looks like a pango-style font description. Maybe that's the problem
So why is it crashing ?
as far as i know it isn't crashing
*** Bug 87878 has been marked as a duplicate of this bug. ***
I'm at a loss how you could conclude that given the summary is "sawfish-ui crashes when browsing fonts" and the comment "sawfish-ui crashed", and my comment "I can reproduce this 100%". I couldn't find a core file though. If you're sure that the bug has been /fixed/ then sure, close it (I'm afraid I cannot test a more recent gnome version). I would hope it had been fixed before gnome 2 got released anyway.
when I tried it sawfish-ui popped up a window telling me it couldn't load the font I'm sure that the bug exists, and I'm sure that it hasn't been fixed. But I don't think it crashes sawfish, and probably not event the config tool
Nobody said it crashed sawfish itself. And I don't find denying the facts to be a profitable way to triage bugs, but maybe that's just me. Would you prefer "attempting to do this causes error messages, and sawfish config tool to disappear from the screen" ? Or does that "probably not happen" too ?
I think we're arguing at cross purposes
I just re-checked the whole thing (under GNOME 2.0), and reached the conclusion that indeed there is no crash (in the usual sense). Running: gnome/bin/rep sawfish-ui under gdb gave the following: ------------------------- Starting program: /home/arielt/software/gnome/bin/rep sawfish-ui ** (rep:20301): WARNING **: Couldn't load font "fixed Medium Semi-Condensed 0" falling back to "Sans Medium Semi-Condensed 0" ** (rep:20301): WARNING **: Couldn't load font "Sans Medium Semi-Condensed 0" falling back to "Sans 0" ** (rep:20301): WARNING **: All font failbacks failed!!!! Program exited with code 01. (gdb) ------------------------- Sorry for not being accurate in the terms of this bug report. Anyway, I suggest we adopt the description "attempting to do this causes error messages, and sawfish config tool to disappear from the screen". I think we all understand that there is a bug and it must be fixed (I think an ``exit 1;'' in such a situation is at least conceptual bug, even if the code line is there on purpose.)
I think this is a different bug to the one I am thinking about. My bug is that sawfish can never successfully use the font names that the gtk2 font dialog returns, since they are in the new style that pango uses, not the traditional X style that sawfish uses But this bug looks like the sawfish-ui program can't load the font itself, which makes me think perhaps the font configuration is wrong When I run sawfish-ui under gtk2 and select a font, I get a dialog telling me that sawfish couldn't load the font, I don't get a crash/exit
Hi, Here is a debug output from a debian user. >sawfish rep: received fatal signal: Segmentation fault struct debug_buf common: Backtrace in `fatal_signal_handler': <(null)+1076840056> <__libc_free+167> <(null)+134580255> <Fgarbage_collect+645> <rep_on_idle+113> <(null)+1074085998> <rep_event_loop+152> <Frecursive_edit+38>
why did you paste that into this bug? it doesn't look to be at all related. If it's reproducable please open a new bug including a proper (gdb with symbols) backtrace
Yes, this bug is realted to the font. http://bugs.debian.org/152530 Quotting this bug : Shortly after choosing a font, sawfish crashes (changing focus seems to trigger it)
these are two different issues
Then can you tell me how to solve this one ?
which one?
The one I've forwarded
As I said before, if you can get a good backtrace from when the sawfish program crashes, please open a new bug with it. I can't reproduce this myself, and the backtrace above doesn't have enough information to be useful, so there's nothing I can do currently
*** This bug has been marked as a duplicate of 86107 ***