On Thu, Nov 18, 2010 at 7:11 PM, <s...@apache.org> wrote:
> Author: stsp
> Date: Fri Nov 19 00:11:15 2010
> New Revision: 1036686
>
> URL: http://svn.apache.org/viewvc?rev=1036686&view=rev
> Log:
> * STATUS: Downgrade my vote for r1036429.
>
> Modified:
> subversion/branches/1.6.x/STATUS
>
> Modified: subversion/branches/1.6.x/STATUS
> URL:
> http://svn.apache.org/viewvc/subversion/branches/1.6.x/STATUS?rev=1036686&r1=1036685&r2=1036686&view=diff
> ==============================================================================
> --- subversion/branches/1.6.x/STATUS (original)
> +++ subversion/branches/1.6.x/STATUS Fri Nov 19 00:11:15 2010
> @@ -111,7 +111,9 @@ Candidate changes:
> Justification:
> Avoids an assert in the server.
> Votes:
> - +1: philip, stsp
> + +1: philip
> + +0: stsp (it fixes the test, but pburba asked on IRC how this change
> + will affect replaced directories and I'm not sure about that)
Hi Philip,
My question was what happens if, in svnsync test 29, if in r3,
^/trunk/H/B is not a replacement resulting from a copy, but is a
replacement without history from svn mkdir. I momentarily thought
that we'd end up in replay.c:add_subdir and would hit this block:
if (! change->copyfrom_known)
{
SVN_ERR(svn_fs_copied_from(&change->copyfrom_rev,
&change->copyfrom_path,
target_root, new_path, pool));
change->copyfrom_known = TRUE;
}
and call svn_fs_copied_from() for a path-rev that wasn't actually
copied. But when looking closer into this that seems impossible,
because we'd never be in add_subdir() in the first place...I
*think*...this part of the code is quite unfamiliar to me :-\
Paul
> * r1036534
> Allow perl bindings to compile within a symlinked working copy.
>
>
>
