Re: cvs2git initially, then git cvsimport

On 07/17/2012 05:40 PM, Luc Bourhis wrote:
>> cvs2git and "git cvsimport" use different mechanisms for inferring
>> changesets from the CVS history, and they would probably not match
>> up on the boundary. Please note that "git cvsimport" is not very
>> reliable, and it can silently produce incorrect content (even of
>> the HEAD revision).  The "ISSUES" section of its manpage lists some
>> of its known bugs.
>
> Thanks for that definite answer. So basically, the only reliable way
> is the one linked from the cvs2git documentation, i.e. running
> cvs2git once a day e.g. There is a point I do not understand though:
> won't cvs2git generate different sha1 at each run? It will make
> rebasing the working branch more difficult, won't it?

The procedure linked from the docs is a hack.  cvs2git may or may not 
generate the same SHA1 for different runs, depending on the phase of the 
moon and the current price of cut tulips.  If by chance it produces the 
same SHA1 twice in a row, then (if I understand correctly) a rebase 
would not be necessary.  If it produces different SHA1s, then a rebase 
is necessary.  Either way, you have to do a complete run of cvs2git 
every time you want to bring your git repo up to date; depending on the 
size of the repository, this can be prohibitive.

But I guess anything's better than having to use CVS :-)

Michael

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

------------------------------------------------------
http://cvs2svn.tigris.org/ds/viewMessage.do?dsForumId=1670&dsMessageId=2991293

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