[svn.haxx.se] · SVN Dev · SVN Users · SVN Org · TSVN Dev · TSVN Users · Subclipse Dev · Subclipse Users · this month's index

Re: Segfault in SVN 1.7.4

From: Ashod Nakashian <ashodnakashian_at_yahoo.com>
Date: Tue, 13 Mar 2012 05:01:42 -0700 (PDT)

> From: Stefan Sperling <stsp_at_elego.de>
>To: svn_at_feb17.org
>Cc: Philip Martin <philip.martin_at_wandisco.com>; dev_at_subversion.apache.org
>Sent: Tuesday, March 13, 2012 3:44 AM
>Subject: Re: Segfault in SVN 1.7.4
>On Mon, Mar 12, 2012 at 10:37:43PM -0700, svn_at_feb17.org wrote:
>> On Mon, Mar 12, 2012 at 08:30:42PM -0700, svn_at_feb17.org wrote:
>> > > svn_at_feb17.org writes:
>> > >
>> > > > to realize it was segfaulting on the linux side.  I've upgraded to
>> > > > 1.7.4 and the problem remains.  I finally ran the operation with
>> > > > apache runnning in GDB and it's a strlen operation that fails.  A lot
>> > > > of vars are optimized out so I didn't get much more information except
>> > > > that it appears to be trying to report a problem but then dying while
>> > > > reporting.
>> > >
>> >
>> A little more information on this.  I have probably rebuilt svn about 20 times tonight from scratch, with
>> latest source from repository, and after upgrading libtools, autoconf, apr etc.  It failed the make check
>> tests every time.  I finally tried a minimal configure with just the packages needed to get it to build,
>> e.g.
>> ./configure --with-apr=/tools/apr --with-apr-util=/tools/httpd-2.2.19-src/srclib/apr-util/ --with-sqlite=/tools/sqlite-3.7.10
>> versus:
>> ./configure --prefix=/tools/subersion-1.7.4 --with-apr=/tools/apr --with-apr-util=/tools/httpd-2.2.19-src/srclib/apr-util/ --with-apxs=/tools/httpd-2.2.19/bin/apxs --with-neon=/tools/neon-0.29.0 --with-sqlite=/tools/sqlite-3.7.10 --with-sasl=/tools/cyrus-sasl-2.1.25
>> and it seems to be passing its tests fine now (it's not very useful built this way, but that narrows down the problem).  I'll add back in the extra
>> packages tomorrow till I find which one is breaking things.  Thanks for your patience and again if you have any ideas what is causing the problem,
>> I'd be very happy to know.  What started as a simple attempt to check in some code has cost a lot of time http://xkcd.com/349/
>> Darren
>Why aren't you using pre-compiled binaries to save time?
>See http://subversion.apache.org/packages.html
>Given that you're building on 64bit Linux and you get crashes during
>format string parsing this could be a 32bit vs. 64bit library issue.
>I have no better ideas.

I'm suffering the same fate. I'm also building trunk on 64bit Linux and getting test failures.

./configure --enable-maintainer-mode --without-serf


At least one test XFAILED, checking /root/subversion/tests.log
XFAIL: lt-diff-diff3-test 13: difference at the start of a 128KB window
XFAIL: lt-fs-test 18: merging commit
       [[needs to be written to match new merge() algorithm expectations]]
XFAIL: lt-mergeinfo-test 16: merge of rangelists
XFAIL: lt-op-depth-test 21: test_shadowed_update
XFAIL: lt-op-depth-test 31: move_in_copy
XFAIL: lt-op-depth-test 32: move_in_replace
XFAIL: lt-op-depth-test 36: move_on_move
XFAIL: lt-op-depth-test 37: move_on_move2
XFAIL: lt-op-depth-test 38: move_added
XFAIL: blame_tests.py 15: blame -g handles changes from empty mergeinfo
XFAIL: copy_tests.py 103: copy and move conflicts
XFAIL: diff_tests.py 22: diff a file that has been renamed
XFAIL: diff_tests.py 49: diff URL against working copy with local mods
XFAIL: diff_tests.py 50: diff -r1 of dir with removed-then-readded file
XFAIL: diff_tests.py 61: diff WC-WC shows the correct base rev num
XFAIL: export_tests.py 11: export working copy at base revision
XFAIL: externals_tests.py 37: commit --include-externals --depth=immediates
XFAIL: externals_tests.py 38: external shadows an existing dir
XFAIL: merge_tests.py 49: avoid repeated merges for cyclic merging
XFAIL: merge_tests.py 64: merge target with non inheritable mergeinfo
XFAIL: merge_tests.py 114: don't inherit bogus mergeinfo
XFAIL: merge_tests.py 115: don't inherit bogus working mergeinfo
XFAIL: merge_tests.py 125: svnmucc: merge a replacement
XFAIL: merge_tests.py 126: replacement in merge source not notified correctly
XFAIL: merge_tree_conflict_tests.py 12: tree conflicts 5.1: leaf edit, tree del
XFAIL: merge_tree_conflict_tests.py 13: tree conflicts 5.2: leaf del, tree del
XFAIL: merge_tree_conflict_tests.py 18: tree conflicts 5.2: leaf del (no ci), tree del
XFAIL: merge_tree_conflict_tests.py 23: replace vs. delete tree-conflicts
XFAIL: merge_tree_conflict_tests.py 24: merge replace on local delete fails
XFAIL: resolve_tests.py 2: resolving prop conflicts
XFAIL: revert_tests.py 25: revert a copy with depth=files
XFAIL: revert_tests.py 26: revert a nested add with depth=immediates
XFAIL: special_tests.py 18: merge symlink-add from foreign repos
XFAIL: special_tests.py 22: replace symlinks
XFAIL: special_tests.py 24: cat added symlink
XFAIL: stat_tests.py 36: modified before translation, unmodified after
XFAIL: svnrdump_tests.py 28: dump: subdirectory
XFAIL: svnrdump_tests.py 29: dump: subdirectory with changes on root
XFAIL: switch_tests.py 9: switch a file to a dir and back to the file
XFAIL: tree_conflict_tests.py 14: merge dir: del/rpl/mv onto not-same
XFAIL: update_tests.py 65: update locally moved dir with incoming file move
XFAIL: update_tests.py 66: update locally modified file to equal versions
Summary of test results:
  1594 tests PASSED
  47 tests SKIPPED

I'm interested in finding the last revision that pass *all* tests on 64bit Linux, if there is such a revision. If I can help fix these failures I'd be happy.

Received on 2012-03-13 13:28:55 CET

This is an archived mail posted to the Subversion Dev mailing list.