[svn.haxx.se] · SVN Dev · SVN Users · SVN Org · TSVN Dev · TSVN Users · Subclipse Dev · Subclipse Users · this month's index

Re: Multi-location commit - allowed?

From: Ben Collins-Sussman <sussman_at_collab.net>
Date: 2005-06-08 23:01:11 CEST

On Jun 8, 2005, at 12:02 PM, Scott Palmer wrote:

>
> On Jun 8, 2005, at 2:34 PM, Dale Worley wrote:
>
>
>>> From: Michael Hipp [mailto:Michael@Hipp.com]
>>>
>>> Yes, it says something like "C: is not a working copy" which
>>> seems an
>>> odd error message.
>>>
>>>
>>
>> When Svn does a check-in, it abstracts from the list of files
>> mentioned
>> their nearest common ancestor. That ancestor tells the repository
>> into
>> which the checkin will be done. In your case, the nearest common
>> ancestor
>> was C:, and evidently it wasn't the root of a WC.
>>
>
> Nor should it need to be. Why can't svn just check the "Repository
> UUID" from the WC metadata, see that they match,
> and proceed with the commit?
>

We used to have this behavior, but had to regress it to fix some
other bug. I think it may be a filed issue.

(In other words, 'svn commit target1 target2 ... targetN' used to
group all changes by repos UUID, and then make one atomic commit to
each repository involved.)

---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org
Received on Wed Jun 8 23:06:19 2005

This is an archived mail posted to the Subversion Users mailing list.