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

Re: reparent bug in svnserve

From: Peter Lundblad <plundblad_at_google.com>
Date: 2006-11-27 09:38:42 CET

David Glasser writes:
> On 11/26/06, David Glasser <glasser@mit.edu> wrote:
> > On 11/23/06, Alexander Sinyushkin <Alexander.Sinyushkin@svnkit.com> wrote:
> > > This is what I noticed while running 'externals' python tests.
> >
What do you mean? Can you trigger this using the python bindings, or are you
speaking the svn protocol directly?

> Actually, hmm, I think I see what the real issue is. Turns out that
> unlike most of the stubs in libsvn_ra/ra_loader.c, svn_ra_reparent
> actually does some common work --- it calls svn_ra_get_repos_root and
> does the appropriate check with svn_path_is_ancestor. So what you
> describe should not be a problem for anything that actually uses the C
> client implementation of svn_ra_reparent.

Or doesn't abuse the (underdocumented) svn protocol.

Note also that some other commands (diff, switch and link-path in the reporter)
also suffer from this bug.
>
> I'd recommend that SVNKit make the same client-side check as the C
> libraries do, but we should still probably fix this (using
> svn_path_is_ancestor).

+1.

  (Plus, there could hypothetically be a race
> condition or something with the client-side check.)
>

I don't think we support relocating the repository while it is online, if
that's what you're referring to...

Thanks,
//Peter

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Mon Nov 27 09:45:48 2006

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