Karl Fogel wrote:
>Greg Stein <gstein@lyra.org> writes:
>
>>>Well, in that case there's a bug. If all committing and updating
>>>happens from the same single working copy, Subversion should never
>>>flag an out-of-date error.
>>>
>>>Is there a reproduction recipe?
>>>
>>We've had a repro for ages...
>>
>>I call it a bug, Ben says it is desired behavior. More on that in another
>>response.
>>
>
>Huh? Are you sure we're talking about the same thing?
>
>He's got one working copy in the whole world for this repository, yet
>he's managing to get out-of-date errors (and as far as I know, he's
>never downdating, only committing and updating). That doesn't sound
>like the bug we know about. It doesn't sound like intentional
>behavior at all, unless I'm missing some scenario... ?
>
Of course there's a common scenario. We don't bump the wc revision on
the files at commit, only at update. All you have to do is commit just
one file to get a mixed-revision working copy. Here's an example (this
is a real wc I've been using to develp a small library):
[brane@SILMARIL utfopt]$ svn st -v
_ 3 ? ? .
_ 3 ? ? ./test
_ 11 ? ? ./test/Makefile
__ 3 ? ? ./test/in-ucs2.txt
__ 3 ? ? ./test/in-utf8.txt
__ 12 ? ? ./test/opts1.txt
M 10 ? ? ./test/parser.c
_ 3 ? ? ./test/u2u_common.h
_ 3 ? ? ./test/ucs2utf.c
_ 3 ? ? ./test/utf2ucs.c
M 10 ? ? ./utfopt.c
_ 10 ? ? ./utfopt.h
_ 3 ? ? ./utfopt_impl.h
This is the only working copy of that repository I ever made. It's the
same bug^H^H^Hmisfeature.
--
Brane Čibej <brane_at_xbc.nu> http://www.xbc.nu/brane/
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Sat Oct 21 14:37:01 2006