GNOME Bugzilla – Bug 164266
svg loader causes segfault on IRIX
Last modified: 2006-09-29 19:25:34 UTC
Distribution: Irix Package: librsvg Severity: critical Version: GNOME2.9.4 unspecified Gnome-Distributor: GNOME.Org Bugzilla-Product: librsvg Bugzilla-Component: general Bugzilla-Version: 2.8.1 Description: Description of the crash: Hi, I'm using IRIX 6.5 and have a problem since before when I'm trying to install librsvg after compiling, gdk-pixbuf-query-loaders always segfaults with core dump, but this core is useless even though I compiled it with -g flag as follows. The following is the my system log at that time: unix: |$(0x6dd)ALERT: Process [gdk-pixbuf-query-loaders] 1785035 generated trap, but has signal 11 held or ignored unix: epc 0xae9ed70 ra 0xfb60630 badvaddr 0xae9ed70 unix: process has been terminated to prevent infinite loop Additional Information: The version of related programs is, gcc 3.4.3, gtk+/glib 2.6.1, libxml 2.6.9, libcroco 0.6.0, libgsf 1.11.1 and librsvg 2.8.1. Debugging Information: Backtrace was generated from '/usr/local/bin/gdk-pixbuf-query-loaders' Core was generated by `gdk-pixbuf-query-loaders'. Program terminated with signal 11, Segmentation fault. GDB is unable to find the start of the function at 0xae9ed70 and thus can't determine the size of that function's stack frame. This means that GDB may be unable to access that stack frame, or the frames below it. This problem is most likely caused by an invalid program counter or stack pointer. However, if you think GDB should simply search farther back from 0xae9ed70 for code which looks like the beginning of a function, you can increase the range of the search using the `set heuristic-fence-post' command.
+ Trace 54576
-- Daichi ------- Bug moved to this database by unknown@bugzilla.gnome.org 2005-01-16 12:59 ------- Unknown version 2.8.1 in product librsvg. Setting version to "unspecified". Unknown platform unknown. Setting to default platform "Other". Unknown milestone "unknown" in product "librsvg". Setting to default milestone for this product, '---' The original reporter of this bug does not have an account here. Reassigning to the person who moved it here, unknown@bugzilla.gnome.org. Previous reporter was daichik@users.sourceforge.net. Setting to default status "UNCONFIRMED". Setting qa contact to the default for this product. This bug either had no qa contact or an invalid one.
I've updated gcc 4.0.0, gtk+ 2.6.7, glib 2.6.4, libxml 2.6.17 on IRIX 6.5.22m, however it've been still causing segfault at registration for svg_loader.so. I'd like to know where problem could be existed.
if you had a better backtrace or something, that would be quite helpful. i suggest that you upgrade to the librsvg 2.12.x version and try that out. the registration functon doesn't do anything interesting, and this might be exposing a bug in GTK+ and not librsvg. thanks.
Created attachment 56058 [details] Runtime debugging log Hi, still not working here (on IRIX 6.5.22m): gcc 4.0.1, gtk+ 2.9.0, glib 2.9.2, librsvg 2.13.0. I could only get a dim light when I have had runtime debug with dbx, as you can see, the crash occurred in the pthread library and as far as I know, the codes below seem only reated with pthread, In gdk-pixbuf-loader/io-svg.c: 188 /* this is present only in GTK+ 2.6 and later. we want librsvg to work with older versions too. 189 we can't define this flag yet because Pango isn't threadsafe. */ 190 #ifndef GDK_PIXBUF_FORMAT_THREADSAFE 191 #define GDK_PIXBUF_FORMAT_THREADSAFE (1 << 2) 192 #endif but I have really no clue... probably as you said, this may not be concerned with librsvg, thank you for your replying anyway.
Have you had any better luck with this? Maybe in the 2.14.3 version?
Closing this bug report as no further information has been provided, sorry... Please feel free to reopen this bug if you can provide more information. Thanks a lot in advance!