On Thursday 22 January 2004 09:42, Marc Haisenko wrote:
> Sorry for spam'ing you, John :-) Next try:
> Didn't CC: the list, here is what I wrote to John in reply:
> -----
>
> On Thursday 22 January 2004 14:14, you wrote:
> > > [snip]
> > >
> > > > > Have you tried "svnadmin verify" ?
> > > >
> > > > Actually, "verify" is "dump". :-) Seriously, the only
> > >
> > > thing it does it not
> > >
> > > > pipe the data to stdout.
> > >
> > > I've tried it anyways :-) And it really finishes without problems.
> >
> > You're kidding? Seriously, there is almost zero difference between
> > "verify" and "dump". There should not have been any difference. Did you
> > run something else in the meantime? If you're truly seeing this
> > behavior, this means one of 2 things: "dump" is broken, or "verify" is
> > broken. If "svnadmin dump" still doesn't work properly, is there any
> > chance that I could get a copy of your repository to test with?
>
> I'm currently trying to do a complete dump again, just for testing
> purposes, I'll mail you again when it's done.
>
> And sorry, there are slim chances for you get hands on our repository. For
> one it's quite big, I'd have to burn it onto a DVD. And since it contains
> critical data (like our product we're developing for nearly for year) it
> would at least require you to sign an NDA or something if I'd manage to get
> my boss to agree to giving you access to our repository :-)
Understood.
> But if it really turns out to be a severe bug I can invest some time into
> debugging.
Okay.
> > > So I should do an svnadmin dump -r335:336 or something like that ?
> >
> > Yep.
> >
> > -John
>
> I'll do that after the complete dump. But what should I do with it ?
From your other mail it looks like you ran into the same problem, so it
definitely sounds like a problem with the repository. I don't know if you
can test this or not, but can you checkout that revision 336 and verify that
the contents of that revision are correct (i.e, there are no corrupted source
files, images, etc.). The dumpfile should show which files were modified in
that rev, as should 'svn log -r336'. I'm curious as to whether there is
actually corrupted data, or that the wrong checksum got stored somehow. It's
more than likely the former.
-John
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Fri Jan 23 00:34:14 2004