Mats Nilsson wrote:
> Hi,
>
> Max Bowsher wrote:
>> Mats Nilsson wrote:
>>> Hmm. Now, this leads me to think that mks2cvs didn't manage to
>>> correctly create branches according to the equivalent MKS construct
>>> (variant). I guess it only created tags, not branch tags.
>>>
>>> Well. Nevertheless, the ,v file contains explicit symbol definitions
>>> like this:
>>>
>>> head 1.16;
>>> Access ;
>>> symbols
>>> ..
>>> ..
>>> REL_4_2_1:1.16
>>> REL_4_1_8:1.15.1.2
>>> REL_4_2_0:1.16
>>> ..
>>> ..
>>
>> Something feels wrong about the above, but I would need more
>> context to say
>> for certain.
>
> I was correct in my suspicion that mks2cvs doesn't create labels for
> branches. If I doctor this file manually to include such labels (e.g.
> BRANCH:1.15.0.1), the cvs2svn conversion succeeds.
>
> Maybe cvs2svn should refrain from trying to convert revisions on RCS
> branches in the absence of corresponding branch labels.
It would be nice to at least fail with an informative error. Feel like
writing a patch? :-)
>> Is the data in the repository private? How big is the converted cvs
>> repository?
>>
>> If you can make it available, I will try to help determine if
>> the problem is caused by mks2cvs or cvs2svn.
>
> The repository is private, but I guess I could isolate and anonymize the
> culprit file and publish it if there is an interest. Personally, I'm now
> convinced that mks2cvs doesn't fully understand the mks project file
> format and that this is the crucial point.
Since you have isolated the problem to mks2cvs, there is no need to see your
repository.
Should you decide to attempt to enhance mks2cvs, you might like to take note
that CVS branches are supposed to use an *even* branch number unless they
are vendor branches - i.e. 1.15.0.2, not 1.
Max.
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org
Received on Mon Sep 1 14:00:42 2003