# git-svn checksum mismatch

2 messages from 2010-02-18 to 2010-02-18. Participants: Jason Baker, Eric Wong.
Thread: https://gitlist.dev/t/22712

## Jason Baker, 2010-02-18 14:55

Subject: git-svn checksum mismatch
Message-ID: <fa34f1941002180655n79430e79yc8658ce17e4e1093@mail.gmail.com>
URL: https://gitlist.dev/e/fa34f1941002180655n79430e79yc8658ce17e4e1093%40mail.gmail.com
In-Reply-To: <fa34f1941002180652wb91c36bt8feba5c7f653701d@mail.gmail.com>

```
On Wed, Feb 17, 2010 at 6:53 PM, Eric Wong <normalperson@yhbt.net> wrote:
>
> Jason Baker <jbaker@zeomega.com> wrote:
> > Hey Eric,
> >
> > I'm emailing you because you're listed on the git-svn manpage.  I'm trying
> > to import an svn repository using git svn, and I'm running into a bit of a
> > problem:
> >
> > r57269 = f5bbaec2f7e8d866d794b421f68706baf81b16ec
> > (jiva_model_0.6-Jiva-4.5.0)
> >     M    src/jiva_model/schema/episode.py
> >     M    src/jiva_model/schema/group.py
> > Checksum mismatch: src/jiva_model/schema/member.py
> > expected: d6f1b614f3ec9c757ab21ebdfd01e011
> >     got: a90dbcbdb3082deadecca9d1d00879bd
> >
> > I'm running git version 1.6.3.3 and git-svn version 1.6.3.3-2 on Ubuntu
> > Karmic.
> >
> > I'm able to check out these files at this revision normally through svn.
> > Does this mean that there's a problem somewhere in git svn?  Or do I need to
> > do something more complex to reproduce it outside of git svn?  It is
> > entirely possible that something's wrong with our svn repository.
>
> Hi Jason,
>
> It could be a problem with either the repo or git svn.  Is the repo
> public by any chance?  If it's a smallish repo I'd be glad to try
> reproducing it on my end.

Nope, it's a private repo and it's rather largeish.

>
> The checksum SVN relies on is just an MD5, so it should be easy to
> test arbitrary revisions of it (svn cat -r$rev $URL@$rev | md5sum)
> and see if there's an error for that repository.

This gives me something different from either value.  :-/

$ svn cat -r57269
$URL/jiva_model_0.6-Jiva-4.5.0/src/jiva_model/schema/member.py@57269 |
md5sum
b5a6b9473d602027b1fe9b5b9d993cc0  -

 (URL being the path to the root of the repo of course)

>
> You could also try updating between two revisions:
>
>  svn co -r57269 $URL FOO # (57269 was the last successful one)
>
>  cd FOO
>  svn up -r$NEXT_REV      # see if this errors out

This works:

$ svn co $URL/branches/jiva_model_0.6-Jiva-4.5.0/src/jiva_model/schema/
-r 57269
Checked out revision 57269.

$ cd schema

$ svn up -r 57270
At revision 57270.

So does this mean there's a problem on my side somewhere?

>
> > Also, is there a mailing list I should send questions like this to?
>
> The normal git mailing list, git@vger.kernel.org, but feel free to Cc me
> to increase my chances of seeing it.

Added that list to this email.  Thanks!

```

## Eric Wong, 2010-02-18 23:08

Subject: Re: git-svn checksum mismatch
Message-ID: <20100218230812.GA20034@dcvr.yhbt.net>
URL: https://gitlist.dev/e/20100218230812.GA20034%40dcvr.yhbt.net
In-Reply-To: <fa34f1941002180652wb91c36bt8feba5c7f653701d@mail.gmail.com>

```
Jason Baker <jbaker@zeomega.com> wrote:
> This works:
> 
> $ svn co $URL/branches/jiva_model_0.6-Jiva-4.5.0/src/jiva_model/schema/ -r
> 57269
> Checked out revision 57269.
> 
> $ cd schema
> 
> $ svn up -r 57270
> At revision 57270.
> 
> So does this mean there's a problem on my side somewhere?

Did r57270 change anything at all in the directories you're tracking?
You might need to step the revision up to the next one you're
tracking...

Not sure what else could be the cause...  Did you run out of disk space
or encounter any hardware failures at any time during the fetches?

-- 
Eric Wong

```