[svn.haxx.se] · SVN Dev · SVN Users · SVN Org · TSVN Dev · TSVN Users · Subclipse Dev · Subclipse Users · this month's index

Re: Another cvs2svn conversion error

From: Max Bowsher <maxb_at_ukf.net>
Date: 2003-11-11 15:14:35 CET

On Tue, 11 Nov 2003, Mats Nilsson wrote:
> The attached file (~1 kb) is a anonymized repository of two files
> (a.txt,v and b.txt,v). It demonstrates a problem that I couldn't find in
> the issue tracker.
> Contrary to my expectations, cvs2svn.py (r7672) produces a dumpfile
> where at least one tag (REL_418_BUILD_842) lacks file a.txt, and the tag
> is created from /trunk instead of from /branches/BRANCH_41X, despite
> that there exists a corresponding CVS branch tag. And then the a.txt
> file is deleted from the tag.
> File a.txt,v contains, among a lot of other stuff:
> symbols
> REL_418_BUILD_842:
> Now the fun part:
> Removing the other file b.txt,v from the repository makes the problem go
> away, and tag/REL_418_BUILD_842 contains the right version of a.txt.
> This is an anonymized version of a CVS archive converted from an MKS-SI
> archive using a locally modified version of the mks2cvs.pl (google)
> version. We have tweaked mks2cvs to map MKS "variants" to CVS
> "branches". To the best of my knowledge, this is a valid CVS repository.
> Please prove me wrong.
> Just bringing it up here for some discussion, before I file an issue.

I think I recognize this bug. Similar behaviour is also exhibited by
cvs2svn/test-data/split-branch-cvsrepos, BUT the associated test doesn't
fail, because it only tests for successful exit, not for correct output.

Have a look at issue 1540 - if you agree that it is the same bug, please
attach your test data there.


To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Tue Nov 11 15:16:01 2003

This is an archived mail posted to the Subversion Dev mailing list.

This site is subject to the Apache Privacy Policy and the Apache Public Forum Archive Policy.