Re: cvs2svn: Excluding CVS folders

On 2013-04-19 00:30, David Morris wrote:
> Well I'm glad I'm not crazy -- I asked the same questions of our
> development teams.  I too was under the impression that CVS folders should
> only exist in a working copy.  When I found them in the first repo I
> migrated, I chalked it up to a check-in mishap at some point.  But now I
> find them in every repo so far (and we have dozens of repos).  I'm talking
> repos from different teams across the organization, managed with different
> client tools, from Eclipse to Netbeans to TortoiseCVS to command-line.
>  They're also in both Windows and Linux CVS servers.  I have no idea how
> they got in there, but they are there, and I don't want to migrate them.
> 
> And no, there was never a migration from RCS in the past.  Most of the
> projects were born in CVS, while a few old ones were manually migrated from
> old SourceSafe which was strictly a top-skim type of checkout/checkin
> migration.
> 

Those folders are created by Intelli J.

If I remember correct they contain a file fileattr.xml with some
meta informations (owner).

Look inside the the xml file and you see the information is useless.

Since you work (hopefully) on a copy of your repo remove them.
$> find $REPO -type d -name CVS | xargs rm -rf



> 
> On Thu, Apr 18, 2013 at 5:28 PM, Michael Haggerty <[email protected]>wrote:
> 
>> On 04/18/2013 10:37 PM, David Morris wrote:
>>> We're using cvs2svn in our migration from CVS to Subversion.  It's
>>> working great so far, but one thing that bugs me is that it's migrating
>>> all the 'CVS' folders in the repository.   I could find no way to stop
>>> this, and was frankly surprised, as these directories should be useless
>>> in a SVN scenario.  Why does cvs2svn not exclude them automatically, or
>>> at least provide an option?  Is there a significance to keeping these
>>> folders in SVN that I'm completely missing?  Some of our larger
>>> repositories have upwards of 1000 subdirectories, each of which contains
>>> a CVS directory.  I've searched and find it surprising that there isn't
>>> more demand for this (I've not found a single question about it).  I
>>> always assumed CVS didn't check these in either as they are workspace
>>> control files, but every one of our CVS repositories so far has them
>>> checked in.
>>
>> I'm confused.  The "CVS" directories are usually found in a CVS
>> checked-out working copy, not in the CVS repository itself.  Since
>> cvs2svn needs to access the repo and not a working copy, it usually
>> never *sees* any CVS directories.
>>
>> Why do you have "CVS" directories checked into your repository?
>>
>> RCS used to store file histories within the working copy; is it possible
>> that this is an artifact of some strange migration from RCS in the
>> distant past?
>>
>> Michael
>>
>> --
>> Michael Haggerty
>> [email protected]
>> http://softwareswirl.blogspot.com/
>>
> 
> ------------------------------------------------------
> http://cvs2svn.tigris.org/ds/viewMessage.do?dsForumId=1670&dsMessageId=3053693
> 
> To unsubscribe from this discussion, e-mail: [[email protected]].
>

------------------------------------------------------
http://cvs2svn.tigris.org/ds/viewMessage.do?dsForumId=1670&dsMessageId=3053701

To unsubscribe from this discussion, e-mail: [[email protected]].