On 22-6-2009 16:29, Hyrum K. Wright wrote:
> Hmmm, I wonder if this is really a problem with the underlying
> backward compat C code. The test presumably passed when it was
> originally added, but when the underlying APIs got rewritten, there
> may be some bugs in the backward compatibility wrapper. If that's the
> case, we should fix the bugs there, not in the python testsuite.
>
> Caveat: I haven't looked at the history of this test, or the
> underlying code, so I could be completely making stuff up. But it's a
> thought.
Even the very first version of get_status_editor uses 'target' as
argument
(http://svn.collab.net/viewvc/svn/branches/1.0.x/subversion/include/svn_wc.h?revision=8524&view=markup,
line 1174). But I do wonder as well how the testcase could ever pass.
Just a shot in the dark, but it could pass on a linux box if self.path
is './', not sure though...
Regards,
Jens
------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=2364265
Received on 2009-06-22 21:40:56 CEST