David Glasser wrote:
> This test fails for me under RA-local, when my working copy is checked
> out via HTTPS.
>
> CMD: svn mergeinfo
> file:///opt/svn/checkouts/in-memory-cache/subversion/tests/cmdline/svn-test-work/repositories/mergeinfo_tests-4/iota
> --config-dir /opt/svn/checkouts/in-memory-cache/subversion/tests/cmdline/svn-test-work/local_tmp/config
> --password rayjandom --no-auth-cache --username jrandom <TIME =
> 0.318293>
>
..
>
> The problem here is that the URL you specified is being parsed as the
> source URL (is that what you intended), and the target is being parsed
> from the WC. I guess you should specify both?
>
Yes, you're right. Fixed it in r30297.
Now, there's one remaining question, and that's why 'svn mergeinfo
--show-revs merged' on an unknown url doesn't always fail.
If I take my working copy of Subversion trunk and run
$ svn mergeinfo http://svn.collab.net/repos/svn/blabla .
I get:
subversion/libsvn_ra_serf/getlocationsegments.c:230: (apr_err=175007)
svn: '/repos/svn/!svn/bc/30297/blabla' path not found
But if I do the same thing in a small test repository I don't get any
error at all:
$ svnadmin create /tmp/repos
$ svn co file:///tmp/repos /tmp/wc
$ echo "blabla" > /tmp/wc/a
$ svn add /tmp/wc/a
$ svn ci -m "" /tmp/wc/a
$ svn mergeinfo file:///tmp/repos/blabla /tmp/wc
$
It's very confusing that the 'path is valid' check is only done in some
situations, and not in others.
Lieven
PS. just a reminder, can you not add svn_at_s.t.o in CC when you reply to
commit messages?
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe_at_subversion.tigris.org
For additional commands, e-mail: dev-help_at_subversion.tigris.org
Received on 2008-04-04 23:59:34 CEST