|Summary:||dev-util/cvs-1.12.12-r4 fails when compiled with -D_FORTIFY_SOURCE=2|
|Product:||Gentoo Linux||Reporter:||Martin Väth <martin>|
|Component:||[OLD] Games||Assignee:||Maintainers for cvs, and cvs related tools (the version control system) [OBSOLETE] <cvs-utils+obsolete>|
|Severity:||normal||CC:||atoth, rose, zorry|
|Package list:||Runtime testing required:||---|
|Bug Depends on:|
Description Martin Väth 2008-04-20 12:58:09 UTC
emerging ultimatestunts gives immediately the following strange error which is apparently due to the "autopoint -f" command in the ebuild: *** %n in writable segment detected *** *** %n in writable segment detected *** [several hundred times repeated] *** %n in writable segment detected *** Segmentation fault autopoint: *** infrastructure files for version 0.11.2 not found; this is autopoint from GNU gettext-tools 0.17 autopoint: *** Stop. I had compiled gettext with gcc-4.3 and "risky" CFLAGS, but also after recompiling with gcc-4.2.1 and CFLAGS="-march=pentium3m -O2 -fno-ident -pipe" the same result occurs. I had similar messages (including a segfault) also with an old (unsupported) xaralx ebuild which used "autoreconf" (not the newer eautoreconf).
Comment 1 Tomáš Chvátal (RETIRED) 2008-04-21 19:56:18 UTC
compiles fine on today ~x86. so you have problem somewhere else.
Comment 2 Martin Väth 2008-04-21 22:05:35 UTC
Thanks. I hadn't expected this, because I thought that autopoint is a rather self-consistent binary. But now I investigated further and found that it is essentially a shell script which calls cvs. It turned out that the culprit was cvs compiled with -D_FORTIFY_SOURCE=2. Probably this means that cvs will fail with hardened. Not sure whether I should close the bug as invalid or whether it should be reassigned to either the cvs or hardened people. You decide...
Comment 3 Jeroen Roovers (RETIRED) 2008-04-22 18:05:24 UTC
emerge --info, please.
Comment 4 Matt 2008-06-10 15:44:56 UTC
confirmed ! this happens approx. from gcc-4.2.0 to gcc-4.3.1 with and cvs compiled with -D_FORTIFY_SOURCE=2 enabled in cflags/cxxflags if -D_FORTIFY_SOURCE=2 is disabled there's no such problem >*** %n in writable segment detected *** >*** %n in writable segment detected *** >[several hundred times repeated] this does so much load on the x-server / system, that working isn't possible anymore ... only killing via magic sysrq key is possible (as last resort) system: Portage 18.104.22.168 (default/linux/amd64/2008.0, gcc-4.3.1, glibc-2.8_p20080602-r0, 2.6.26-rc5-zen1.0 x86_64) ================================================================= System uname: 2.6.26-rc5-zen1.0 x86_64 Intel(R) Core(TM)2 CPU 6600 @ 2.40GHz Timestamp of tree: Tue, 10 Jun 2008 12:16:01 +0000 app-shells/bash: 3.2_p39 dev-java/java-config: 1.3.7, 2.1.6 dev-lang/python: 2.4.4-r9, 2.5.2-r4 dev-python/pycrypto: 2.0.1-r6 sys-apps/baselayout: 2.0.0 sys-apps/openrc: 9999 sys-apps/sandbox: 22.214.171.124-r2 sys-devel/autoconf: 2.13, 2.62 sys-devel/automake: 1.4_p6, 1.5, 1.6.3, 1.7.9-r1, 1.8.5-r3, 1.9.6-r2, 1.10.1-r1 sys-devel/binutils: 2.18-r1, 126.96.36.199.4, 188.8.131.52.5, 184.108.40.206.6, 220.127.116.11.7 sys-devel/gcc-config: 1.4.1 sys-devel/libtool: 1.5.26 virtual/os-headers: 2.6.25-r4 ACCEPT_KEYWORDS="amd64 ~amd64" CBUILD="x86_64-pc-linux-gnu"
Comment 5 Magnus Granberg 2009-01-24 11:31:04 UTC
Comment 6 Patrick Lauer 2009-02-01 18:04:35 UTC
+ 01 Feb 2009; Patrick Lauer <firstname.lastname@example.org> + +files/cvs-1.12.12-cvs-gnulib-vasnprintf.patch, +cvs-1.12.12-r6.ebuild: + Fixing cvs to work with gcc 4.3.3 which enables FORTIFY_SOURCES. Fixes + #230849. Patch from Bruno Haible <email@example.com>, thanks to Magnus + Granberg <firstname.lastname@example.org> for some hints and explanations.
Comment 7 Patrick Lauer 2009-02-01 18:05:06 UTC
*** Bug 230849 has been marked as a duplicate of this bug. ***