As seen on http://savannah.gnu.org/bugs/?27475 libsigsegv's behavior broke with the latest F11 GCC update. The upstream fixes are 54b612e and 4f14ef8 in git://git.sv.gnu.org/libsigsegv.git and they should be cherry-picked into the F11 and rawhide packages.
funky, I'll take a look.
Paolo, thanks by the way for all the help here, would you be interested in helping maintain this package (ie, give you commit access to fix items)?
Woo, good news bad news. Here's a scratch build with adjusted patches: http://koji.fedoraproject.org/koji/taskinfo?taskID=1698646 Good news, build on x86_64 works Bad news, ppc64 fails now, 'make check' failures: Executing(%check): /bin/sh -e /var/tmp/rpm-tmp.zj7Ujq + umask 022 + cd /builddir/build/BUILD + cd libsigsegv-2.6 + unset DISPLAY + make check Making check in src make[1]: Entering directory `/builddir/build/BUILD/libsigsegv-2.6/src' make[1]: Nothing to be done for `check'. make[1]: Leaving directory `/builddir/build/BUILD/libsigsegv-2.6/src' Making check in tests make[1]: Entering directory `/builddir/build/BUILD/libsigsegv-2.6/tests' make check-TESTS make[2]: Entering directory `/builddir/build/BUILD/libsigsegv-2.6/tests' Test passed. PASS: sigsegv1 Test passed. PASS: sigsegv2 Doing SIGSEGV pass 1. Stack overflow 1 caught. Doing SIGSEGV pass 2. Stack overflow 2 caught. Test passed. PASS: sigsegv3 /bin/sh: line 5: 12410 Segmentation fault ${dir}$tst FAIL: stackoverflow1 /bin/sh: line 5: 12434 Segmentation fault ${dir}$tst FAIL: stackoverflow2 =================== 2 of 5 tests failed =================== Gonna (re)try the build, to make sure it's reproducible and not just gremlins.
Meh, looks reproducible, http://koji.fedoraproject.org/koji/taskinfo?taskID=1699058
OK, looks like currently(1) there's nothing on ppc64 using libsigsegv, so we could (at least temporarily) get away with an ExcludeArch: ppc64 (1) except for bug #511443 where gawk wants to start using (system) libsigsegv too
Noticed one thing, but the .m4 modifications (from libsigsegv-2.6 anyway) use char mystack[2 * SIGSTKSZ]; whereas the tests (still) use: char mystack[SIGSTKSZ]; is that kosher/intentional?
interesting, adjusting the patch so the tests use char mystack[2 * SIGSTKSZ]; too, makes the ppc64 test segfaults go away.
What's the status here? We're running out of time, is this something we'd slip the release for?
Should be resolved as of %changelog * Tue Sep 22 2009 Rex Dieter <rdieter> - 2.6-6 - respin mystack patch * Tue Sep 22 2009 Rex Dieter <rdieter> - 2.6-5 - libsigsegv allocates alternate stack on the main stack (#524796) I guess I meant to issue on F-11 update, and seem to have put it off, will take care of that asap Paolo, any feedback on comment #7 ?
libsigsegv-2.6-6.fc11 has been submitted as an update for Fedora 11. http://admin.fedoraproject.org/updates/libsigsegv-2.6-6.fc11
> interesting, adjusting the patch so the tests use > char mystack[2 * SIGSTKSZ]; > too, makes the ppc64 test segfaults go away. I would have to debug it, but I see no reason not to do that and remove the exclusion.
libsigsegv-2.6-6.fc11 has been pushed to the Fedora 11 testing repository. If problems still persist, please make note of it in this bug report. If you want to test the update, you can install it with su -c 'yum --enablerepo=updates-testing update libsigsegv'. You can provide feedback for this update here: http://admin.fedoraproject.org/updates/F11/FEDORA-2009-10762
no reason for this to block either of the two bugs it was blocking any more, they're both fixed. -- Fedora Bugzappers volunteer triage team https://fedoraproject.org/wiki/BugZappers
libsigsegv-2.6-6.fc11 has been pushed to the Fedora 11 stable repository. If problems still persist, please make note of it in this bug report.