You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@subversion.apache.org by Scott Johnson <sj...@keypress.com> on 2006/07/12 23:16:58 UTC

CVS and Subversion sharing source code

Hi all--

I work with a development team that is attempting to coordinate  
software development with another team of developers at another  
company. The other developers have a CVS repository set up and we  
would like to take their code, put it in our own Subversion  
repository so we can track and maintain our changes, and then commit  
it back to the main CVS repository at various intervals.

Does anybody have any particular advice about how to carry this out,  
or advice AGAINST doing this, or any particular knowledge about  
whether this is possible? It is not clear to us what sort of gotchas  
we are going to run into, but we are hoping that if we take the code  
checked out of CVS, and use it to create a Subversion repository  
WITHOUT the CVS folders (advice on a painless way to do this would be  
helpful!), then we can proceed and both CVS and Subversion will see  
any changes as being no different from any users changes. We just  
have to be careful never to add the CVS folders to the Subversion  
repository, and vice-versa. Right?

Thanks for any help you can provide.

Scott

---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org

Re: CVS and Subversion sharing source code

Posted by Bob Proulx <bo...@proulx.com>.
Scott Johnson wrote:
> I work with a development team that is attempting to coordinate  
> software development with another team of developers at another  
> company. The other developers have a CVS repository set up and we  
> would like to take their code, put it in our own Subversion  
> repository so we can track and maintain our changes, and then commit  
> it back to the main CVS repository at various intervals.

There are now several different projects trying to do sync'ing between
different version control systems.  You might want to look at Tailor.

  http://www.darcs.net/DarcsWiki/Tailor

I have found Tailor quite slow but still better than most other
options for what you want to do.  Other options are to use something
like 'cvsps' but that will take more glue effort.  I have recently
heard of a project 'parsecvs'.

All of these are a work in progress.  All will need some amount of
effort to pull all of the pieces together.  But the need is there so I
expect the effort to be worth it.

> Does anybody have any particular advice about how to carry this out,  
> or advice AGAINST doing this, or any particular knowledge about  
> whether this is possible? It is not clear to us what sort of gotchas  
> we are going to run into, but we are hoping that if we take the code  
> checked out of CVS, and use it to create a Subversion repository  
> WITHOUT the CVS folders (advice on a painless way to do this would be  
> helpful!), then we can proceed and both CVS and Subversion will see  
> any changes as being no different from any users changes. We just  
> have to be careful never to add the CVS folders to the Subversion  
> repository, and vice-versa. Right?

The other time honored method is to import the other code into a
vendor branch.  Both CVS and Subversion document handling of vendor
branches.

  http://svnbook.red-bean.com/nightly/en/svn.advanced.vendorbr.html

Bob

---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org

RE: CVS and Subversion sharing source code

Posted by "L. Wayne Johnson" <wa...@zk.com>.
 

>> -----Original Message-----
>> From: Scott Johnson [mailto:sjohnson@keypress.com] 
>> Sent: Wednesday, July 12, 2006 4:17 PM
>> To: users@subversion.tigris.org
>> Subject: CVS and Subversion sharing source code
>> 
>> Hi all--
>> 
>> I work with a development team that is attempting to 
>> coordinate software development with another team of 
>> developers at another company. The other developers have a 
>> CVS repository set up and we would like to take their code, 
>> put it in our own Subversion repository so we can track and 
>> maintain our changes, and then commit it back to the main 
>> CVS repository at various intervals.
>> 
>> Does anybody have any particular advice about how to carry 
>> this out, or advice AGAINST doing this, or any particular 
>> knowledge about whether this is possible? It is not clear to 
>> us what sort of gotchas we are going to run into, but we are 
>> hoping that if we take the code checked out of CVS, and use 
>> it to create a Subversion repository WITHOUT the CVS folders 
>> (advice on a painless way to do this would be helpful!), 
>> then we can proceed and both CVS and Subversion will see any 
>> changes as being no different from any users changes. We 
>> just have to be careful never to add the CVS folders to the 
>> Subversion repository, and vice-versa. Right?
>> 
>> Thanks for any help you can provide.
>> 

I do something similar to this. Our company uses an old version of Source
Integrity. Because of our current policies it can be months between
check-ins (don't even ask!) I run a local SVN server to tack my work and
later check stuff into our official source code control. I am on the only
one using the repository so quite a few scenarios never come up. 

Currently I have a working copy that is dual source code controled called
xxxx_base. I use this working copy to move stuff between the 2. I never do
any work in the working copy.

I create another working copy when I do any work. Whether I create the wc
based on the trunk or a branch is related to the how soon I think I can
check in the code. When I am ready to check the code in to our official
source code control I move the changes into the xxxx_base working copy and
check it in from there. At strategic (?) points I update the xxxx_base from
our official source code control and check it back into SVN.

There may be a better way to do this. Even though I am the only user it can
get messy really quick if I am not careful. The biggets pain is that we use
both $Header$  and $Log$ expansion on so as soon as I move stuff from SVN
into Source Integrity I have to make sure to check the expansion changes
back into svn.


---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org