Re: Corrupted binary files after converting cvs repo to git repo

On 05/12/2014 09:14 AM, Christian Anke wrote:
> i have tried several aproaches to fix this issue, but there seems to be no
> light.
> 
> As the subject told, the binary files in my resulting git repo are
> defective.
> This is caused by the keyword expansion feature. The binary files
> are jars which contain the expanded $Id$ keyword (which is the original
> state). But in the resulting git repo those keywords will be collapsed and
> the file is currupted.

The very first question is: are you sure that the files are not already
corrupt in CVS?  Sometimes people don't notice corruption in their CVS
repositories but only notice it after the conversion when they are
checking their data extra carefully.

Also, please note that CVSNT marks files as binary differently than CVS.
 If you are using CVSNT, please read the FAQ on that subject because
converting from CVSNT is not officially supported by cvs2git.

If the data are OK in CVS and the files are marked binary (-kb) in CVS,
then everything should work using the default settings.  If they don't,
then it is a bug in cvs2git and we would appreciate a bug report with
reproducible test case.  The FAQ describes how to create a test case
with minimal contents.

> Is this a general issue with cvs2git?

No, certainly not.  As far as I can remember, such problems have
invariably been caused either by problems in the CVS repository (some of
which cvs2git can help you with) or some kind of user mistake.  But we
have open minds, and either way will try to help.

Michael

-- 
Michael Haggerty
[email protected]
http://softwareswirl.blogspot.com/

------------------------------------------------------
http://cvs2svn.tigris.org/ds/viewMessage.do?dsForumId=1670&dsMessageId=3078134

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