Description of problem: Running 'inkscape' causes immediate segfault. Judging by feedback I received on fedora-devel-list this would appear to only happen on x86-64, not i386. Version-Release number of selected component (if applicable): $ rpm -qf /usr/bin/inkscape inkscape-0.45.1-1.fc7 $ rpm -q glibc glibc-2.6-4 glibc-2.6-4 $ rpm -q glibmm24 glibmm24-2.13.6-3.fc8 (all other components are up to date) How reproducible: 100%, on 2 different x86-64 machines. Steps to Reproduce: 1. yum update; yum install inkscape 2. inkscape 3. Actual results: Immediate segmentation fault. Expected results: Run inkscape. Additional info: $ strace inkscape [...] open("/etc/selinux/config", O_RDONLY) = 3 fstat(3, {st_mode=S_IFREG|0644, st_size=512, ...}) = 0 mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x2aaaaaac6000 read(3, "# This file controls the state o"..., 4096) = 512 read(3, "", 4096) = 0 close(3) = 0 munmap(0x2aaaaaac6000, 4096) = 0 statfs("/selinux", {f_type=0xf97cff8c, f_bsize=4096, f_blocks=0, f_bfree=0, f_bavail=0, f_files=0, f_ffree=0, f_fsid={0, 0}, f_namelen=255, f_frsize=4096}) = 0 stat("/selinux/class", 0x7fffb5522ae0) = -1 ENOENT (No such file or directory) open("/selinux/mls", O_RDONLY) = 3 read(3, "1", 19) = 1 close(3) = 0 futex(0x2aaaac590b40, FUTEX_WAKE, 2147483647) = 0 brk(0xdb9000) = 0xdb9000 brk(0xdda000) = 0xdda000 gettimeofday({1184683408, 102077}, NULL) = 0 --- SIGSEGV (Segmentation fault) @ 0 (0) --- +++ killed by SIGSEGV +++ $ gdb /usr/bin/inkscape [...] 0x00002aaab118aec9 in GC_malloc () from /usr/lib64/libgc.so.1 (gdb) bt #0 0x00002aaab118aec9 in GC_malloc () from /usr/lib64/libgc.so.1 #1 0x0000000000976628 in sp_repr_new (name=<value optimized out>) at ./gc-core.h:74 #2 0x0000000000972a4c in sp_repr_svg_read_node (node=0xd8f160, default_ns=0x0, prefix_map=0xd8e400) at xml/repr-io.cpp:472 #3 0x0000000000974150 in sp_repr_do_read (doc=<value optimized out>, default_ns=0x0) at xml/repr-io.cpp:382 #4 0x0000000000974272 in sp_repr_read_mem (buffer=<value optimized out>, length=10198, default_ns=0x0) at xml/repr-io.cpp:306 #5 0x00000000005a9b25 in Inkscape::Preferences::loadSkeleton () at preferences.cpp:34 #6 0x000000000068f1d5 in Application (this=0x2aaab13930b0, argc=14215168, argv=0x7fff4c3ecbc8, use_gui=11, new_gui=208) at application/application.cpp:45 #7 0x0000000000598976 in main (argc=1, argv=0x7fff4c3ecbc8) at main.cpp:489
Please pull gc 7.0 from updates-testing
Created attachment 159844 [details] Fixes 7.0 abi break The problem stems from a missing call to GC_init() (because the GC_INIT macro is different). This is a textbook example of how the abuse of macros is just terrible programming. Anyways, here's a simple patch that seems to be good enough to fix inkscape and w3m.
Thanks, will take a look today.
%changelog * Tue Jul 24 2007 Rex Dieter <rdieter[AT]fedoraproject.org> 7.0-2 - gcinit patch, ABI compatibility (#248700)
gc-7.0-2.fc7 has been pushed to the Fedora 7 testing repository. If problems still persist, please make note of it in this bug report.
gc-7.0-6.fc7 has been pushed to the Fedora 7 testing repository. If problems still persist, please make note of it in this bug report.
gc-7.0-6.fc7 has been pushed to the Fedora 7 stable repository. If problems still persist, please make note of it in this bug report.
FYI, this affects inkscape on EL-5 (which I've just built from FC6 SRPMs), where gc is still at version 6.8: % rpm -q gc gc-6.8-3.el5.x86_64 However, installing gc-7.0-6 doesn't fix the crash, so it might be another problem. And by the way, rpmlint finds some problems with the latest version: % rpmlint . gc.src: W: mixed-use-of-spaces-and-tabs (spaces: line 3, tab: line 64) gc.x86_64: E: binary-or-shlib-defines-rpath /usr/lib64/libcord.so.1.0.3 ['/usr/lib64'] gc.x86_64: E: binary-or-shlib-defines-rpath /usr/lib64/libgccpp.so.1.0.3 ['/usr/lib64'] gc-debuginfo.x86_64: W: spurious-executable-perm /usr/src/debug/gc-7.0/misc.c
Package Change Request ====================== Package Name: inkscape New Branches: EL-4 EL-5
nevermind