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

Re: Test failures in the wc upgrade code

From: Greg Stein <gstein_at_gmail.com>
Date: Mon, 24 Aug 2009 20:03:21 +0000

This is undoubtedly due to my tightening of auto/manual version
detection and upgrading. You really can't get very far if it thinks
you're working with a wc-1 working copy.

The particular problem here may be a false signal, and there may be a
way for me to catch/relax the condition. I'll investigate.


On Mon, Aug 24, 2009 at 19:09, Hyrum K.
Wright<hyrum_wright_at_mail.utexas.edu> wrote:
> Greg,
> I've got a trunk repository, checked out with a 1.6.x client.  When running
> the tests in a clean working copy at r38921, I get 104 test failures, all
> with failures of the form:
> subversion/svn/util.c:898: (apr_err=155036)
> subversion/libsvn_client/merge.c:9296: (apr_err=155036)
> subversion/libsvn_client/merge.c:7904: (apr_err=155036)
> subversion/libsvn_client/merge.c:7454: (apr_err=155036)
> subversion/libsvn_client/merge.c:3917: (apr_err=155036)
> subversion/libsvn_client/merge.c:3022: (apr_err=155036)
> subversion/libsvn_client/mergeinfo.c:388: (apr_err=155036)
> subversion/libsvn_client/mergeinfo.c:252: (apr_err=155036)
> subversion/libsvn_wc/entries.c:1403: (apr_err=155036)
> subversion/libsvn_wc/entries.c:1075: (apr_err=155036)
> subversion/libsvn_wc/wc_db.c:1520: (apr_err=155036)
> subversion/libsvn_wc/wc_db.c:919: (apr_err=155036)
> subversion/libsvn_wc/upgrade.c:465: (apr_err=155036)
> svn: Working copy format of
> '/Users/Hyrum/dev/svn-trunk5/subversion/tests/cmdline' is too old (10);
> please run 'svn upgrade'
> A partial list of failures includes:
> FAIL:  autoprop_tests.py 7: import: config=no,  commandline=none
> FAIL:  autoprop_tests.py 8: import: config=yes, commandline=none
> FAIL:  autoprop_tests.py 9: import: config=no,  commandline=yes
> FAIL:  autoprop_tests.py 10: import: config=yes, commandline=yes
> FAIL:  autoprop_tests.py 11: import: config=no,  commandline=no
> FAIL:  autoprop_tests.py 12: import: config=yes, commandline=no
> FAIL:  blame_tests.py 14: blame -g output with inserted lines
> FAIL:  log_tests.py 28: log -g and simple propmod to merge-inheriting path
> FAIL:  merge_authz_tests.py 2: merge fails if subtree is deleted on src
> FAIL:  merge_tests.py 1: performing a merge, with mixed results
> FAIL:  merge_tests.py 2: merge and add new files/dirs with history
> FAIL:  merge_tests.py 4: some simple property merges
> FAIL:  merge_tests.py 12: merge change into unchanged binary file
>  ... [ most of merge_tests.py ]
> FAIL:  mergeinfo_tests.py 2: 'mergeinfo' on a path with mergeinfo
> FAIL:  mergeinfo_tests.py 5: mergeinfo of an unknown url should return error
> FAIL:  mergeinfo_tests.py 6: non-inheritable mergeinfo shows as merged
> FAIL:  resolved_tests.py 2: resolved on deleted item
> FAIL:  revert_tests.py 6: revert a merge replacement of file with history
> FAIL:  revert_tests.py 15: revert a committed replace-with-history == no-op
> FAIL:  revert_tests.py 17: status after replace+, revert, and local rm
> FAIL:  special_tests.py 7: merge symlink into file
> FAIL:  special_tests.py 8: merge file into symlink
> FAIL:  stat_tests.py 22: status on conflicted added file
> FAIL:  update_tests.py 36: mergeinfo does not elide after update
> I'm guessing this is related your recent upgrade work, but I haven't dug far
> enough to tell.  Is there anything obvious here which could be the result of
> your change?  (For one, I'm wondering why the test suite is attempting to
> open subversion/tests/cmdline/)
> -Hyrum

Received on 2009-08-27 00:06:45 CEST

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

This site is subject to the Apache Privacy Policy and the Apache Public Forum Archive Policy.