"Arild Fines" <arild.fines@broadpark.no> writes:
> Is this a regression, or was it intentional?
>
> J:\tmp\wc-47>svn st
> C Form.cs
> ? Form.cs.mine
> ? Form.cs.r6
> ? Form.cs.r7
>
> J:\tmp\wc-47>svn --version
> svn, version 0.22.0 (dev build)
>
> J:\tmp\wc-47>svn21 --version
> svn, version 0.21.0 (dev build)
>
> J:\tmp\wc-47>svn21 ci . -m "" -N
> svn: A conflict in the working copy obstructs the current operation
> svn: Commit failed (details follow):
> svn: Aborting commit: 'J:/tmp/wc-47/Form.cs' remains in conflict.
>
> J:\tmp\wc-47>svn ci . -m "" -N
> (No output)
>
> As you see, a non-recursive commit on a conflicted directory in 0.22 is a
> no-op, while in 0.21 it generated an error message.
This is intentional. 'svn commit -N' behaves a bit differently than
other nonrecursive operations. For commit, -N means depth 0
(directory only). Most other subcommands use -N to mean depth 1
(directory plus all non-directory children of that directory).
So, in 0.21, commit -N shouldn't have been worrying its little self
about conflicts on items that aren't part of the commit. In 0.22, it
doesn't. :-)
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Fri May 9 01:25:31 2003