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

Re: Issue 1074: Need a svnadmin command to verify that the repository is not corrupted

From: <cmpilato_at_collab.net>
Date: 2003-07-19 19:31:40 CEST

John Szakmeister <john@szakmeister.net> writes:

> Following your guidance, I've gone back and done a few things. I'm now
> reading all of the reps using the cursor method. And I've implemented that
> walk routine in bdb/reps-table.c like you suggested.
>
> I went to do the second part of this, and noticed a few things.
> First, there is a function called svn_fs__rep_contents that's
> actually exported from reps-strings.c. It will actually grab the
> contents of a rep and compute the checksum already. Can I count on
> that as expected behavior? Since it's an exported function, I
> figured the best thing to do is create a top level function (i.e.,
> svn_fs_verify) that sits in fs.c and kick of this process of walking
> the tree and verifying checksums using svn_fs__retry_txn(). Does
> that seem reasonable?

Reasonable, and preferable. Put the cursor-walker down in the
libsvn_fs/bdb directory, and have it use a callback for each row it
hits. Drive that with code in an upper layer (either in
reps-strings.c or higher), where the callback is also implemented, and
use the reps-strings.c functions to do the dirty work.

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Sat Jul 19 19:32:52 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.