On Tue, 2005-03-22 at 12:17 -0500, Greg Hudson wrote:
> On Tue, 2005-03-22 at 08:56, Max Bowsher wrote:
> > 24 hours ago, I would have agreed.
> >
> > However, having stumbled across the problem I filed here:
> > http://subversion.tigris.org/issues/show_bug.cgi?id=2253
> >
> > I wonder if perhaps FSFS isn't sufficiently mature yet.
>
> While this bug and the bug Jani mentioned (assuming it's an actual code
> bug) should certainly be fixed, I think they're both essentially
> irrelevant to this decision: firstly, becuase BDB's limitations and
> foibles bite people way more often than either of these bugs could ever
> affect people, and second, because the most important thing about
> checksums is detecting corruption (via svnadmin verify) in an existing
> repository, not detecting errors in a dumpfile.
>
I can also vouch for the fact that svnadmin verify does indeed work on
fsfs and verifies md5sum, in case someone wants to claim "it might be
broken too".
It caught the xdelta off by one error when i svnadmin verified the gcc
repo because the stored checksum from the fulltext didn't match the
checksum after applying the svn delta.
--Dan
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Tue Mar 22 20:59:28 2005