> -----Original Message-----
> From: julianf...@apache.org [mailto:julianf...@apache.org]
> Sent: donderdag 21 augustus 2014 17:48
> To: comm...@subversion.apache.org
> Subject: svn commit: r1619452 - in /subversion/trunk/subversion:
> libsvn_client/diff.c tests/cmdline/diff_tests.py
>
> Author: julianfoad
> Date: Thu Aug 21 15:47:37 2014
> New Revision: 1619452
>
> URL: http://svn.apache.org/r1619452
> Log:
> Fix the copy-from revision number reported in a diff header. It previously
> reported 'nonexistent' instead of the copy-from revision in some cases.
>
> This bug seems to be a regression; 1.8.x releases show the correct revision.
1.8.x just reported the original/left revision in a lot of cases where it
didn't know the revision.
Especially on directories where 1.8.x where the old internal diff callbacks
didn't even report any revision... where the diff output just guessed the
revision based on the revision of the complete diff (or whatever it expected it
to be).
I don't think we ever reported the copy from revision, as the actual left-src
revision of the diff...
And I'm not sure if reporting copy from is really valid either, without
reporting that it is a copy and probably more importantly where the copy is
from.
The copy can be from any repository path at the specified revision...
So now you can no longer trust the revision to just specify that the path
existed at the current path in that revision.
Bert
