Closed Bug 667197 Opened 14 years ago Closed 10 years ago

Crash [@ libfontconfig.so.1.4.4@0x1a942 ]

Categories

(Core :: Layout: Text and Fonts, defect)

Product:
Core
Component:
Layout: Text and Fonts
Platform:
x86
Linux
Type:
defect
Priority:
Not set
Severity:
critical

Tracking

()

Status:
RESOLVED INCOMPLETE

People

(Reporter: dao, Unassigned)

References

Details

(Keywords: crash)

Crash Data

This has caused dozens of crashes for me in the last couple of days. It seems pretty severe on my system. I'm happy to help debug but have no idea about libfontconfig at all
Thanks, Jeff. What version of Firefox are you running? And what version of libfontconfig is installed? Have you updated Firefox, or something else on your system in the last couple of days? Would you be able to get a stack trace with gdb, please? Perhaps the links here may give helpful instructions: https://developer.mozilla.org/En/How_to_get_a_stacktrace_for_a_bug_report#Linux
I am on nightly up to date as of today (so yes,I have updated nightly, usually every day). According to synaptic, I have libfontconfig 2.8.0-2.1ubuntu . I also have the -dev package but not the -dbg package. I will try to get a stacktrace soon
If FcPatternCreate fails, we are not handling the null pattern properly, but crash reports have small uptimes, suggesting this is not an out of memory problem. (In reply to Jeff Hammel [:jhammel] from comment #4) > According to synaptic, I have libfontconfig 2.8.0-2.1ubuntu . I also have > the -dev package but not the -dbg package. > > I will try to get a stacktrace soon Thanks. Please do install the libfontconfig-dbg package, as it is the fontconfig part of the stack that we don't have.
According to the debug identifier bp-af6ba814-5718-4408-a235-8c8a52110611 has libfontconfig.so.1.4.4 from libfontconfig1 (2.8.0-2.1ubuntu3) http://packages.ubuntu.com/natty/i386/libfontconfig1/download % addr2line -if -e usr/lib/debug/usr/lib/libfontconfig.so.1.4.4 0x1a942 FcStrCaseWalkerNext /build/buildd/fontconfig-2.8.0/src/fcstr.c:160 % addr2line -if -e usr/lib/debug/usr/lib/libfontconfig.so.1.4.4 0x1aced IA__FcStrCmpIgnoreCase /build/buildd/fontconfig-2.8.0/src/fcstr.c:225 % addr2line -if -e usr/lib/debug/usr/lib/libfontconfig.so.1.4.4 0x760f FcConfigCompareValue /build/buildd/fontconfig-2.8.0/src/fccfg.c:754 % addr2line -if -e usr/lib/debug/usr/lib/libfontconfig.so.1.4.4 0x8381 FcConfigMatchValueList /build/buildd/fontconfig-2.8.0/src/fccfg.c:1160 IA__FcConfigSubstituteWithPat /build/buildd/fontconfig-2.8.0/src/fccfg.c:1430 % addr2line -if -e usr/lib/debug/usr/lib/libfontconfig.so.1.4.4 0x88e6 IA__FcConfigSubstitute /build/buildd/fontconfig-2.8.0/src/fccfg.c:1584
Stack is the same as https://bugs.freedesktop.org/show_bug.cgi?id=20445, the first comment of which blames odd addresses returned by malloc. jemalloc has always returned even addresses and, with the patch in Bug 691003 landing, will return word-aligned addresses.
See Also: → https://bugs.freedesktop.org/show_bug.cgi?id=20445
Still seeing this in today's nightly.
I wonder whether it is possible that a corrupt fontconfig cache file might be involved. fc-cache -r will rebuild caches. I'm not sure but it may be necessary to also run this as root (and as the user) to ensure that there are no old files (from system directories) still used.
Unfortunately, i lost the system where this occured regularly (the OS, that is), but if I see it again, I will ping here
This still occurs in (very) low volume according to crash-stats, e.g.: bp-508cc0ae-8494-40e1-bcbe-384b52150521 It's unlikely to be something we can fix in Gecko, so I don't think this is worth tracking on our side. FTR, https://bugs.freedesktop.org/show_bug.cgi?id=20445 was resolved WONTFIX.
Status: NEW → RESOLVED
Closed: 10 years ago
Resolution: --- → INCOMPLETE
You need to log in before you can comment on or make changes to this bug.