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

Re: svnadmin dump | svnadmin load breaks peg revisions?

From: Dan Stromberg <dstromberglists_at_gmail.com>
Date: Mon, 10 Aug 2009 14:15:34 -0700

On Fri, Aug 7, 2009 at 2:40 PM, Les Mikesell<lesmikesell_at_gmail.com> wrote:
> Dan Stromberg wrote:
>> Am I correct in thinking that my options are:
>> 1) Plow ahead with the APR and apache (and subversion, with the new
>> APR) compiles
>> 2) Go back to the old subversion, and just live with the potentially
>> corrupted repo that doesn't seem to be preventing modern builds
>> 3) Do a tar | ssh tar despite the word size and slight operating
>> system difference?
> Before any of those, I'd look around for pre-built packages of subversion
> and mod_dav_svn for your OS target. Maybe someone else has done the work
> (that's one of the advantage of running RHEL or CentOS - you can find
> working RPMs for about everything).

That's advice to live by, but it seems it just won't work in this
case. We're stuck with openSUSE 10.3, and... I suppose there could
be something prebuilt for openSUSE 10.3 somewhere that has recent
apache and subversion and apr, but if there is I sure didn't find it.

I totally agree that when baseline-ing the OS for a product, it's
better to use something with a longer lifetime, but it's too late to
go back and change that now. We actually were on CentOS for a while,
but one of our partners wanted us to go Novell, and for cost reasons
we went part openSUSE and part SLES. For this, we pretty much have to
go openSUSE though.

Anyway, I've got things built now - I've upgraded to subversion 1.6.4
and apache 2.2.13. It was a little more complicated than usual,
because the apache build had problems with both openssl 0.9.8k and
openssl 1.0.0 beta 3 until I got the next version of apache (which
wasn't available when I started) and threw in a cpp symbol I shouldn't
have needed to set manually. Then apache built against openssl 1.0.0
beta 3.

I got bit by the change in semantics of svn checkout -N, but replacing
that with --ignore-externals appears to be giving the old -N behavior
:) So it wasn't too hard to get past.

-=BUT=-, I've still got the bad peg revisions problem. :(

I'm beginning to think I should take this to the dev list, but it
seems like a difficult to replicate issue unless you happen to have a
copy of our repo, which gets into the issue of whether we can share it
long enough to get a fix.

Can anyone on the list confirm that this is a new issue before I bug
the dev's? It sounds like they're tired of simple and known issues
coming up on-list.



To unsubscribe from this discussion, e-mail: [users-unsubscribe_at_subversion.tigris.org].
Received on 2009-08-10 23:16:36 CEST

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

This site is subject to the Apache Privacy Policy and the Apache Public Forum Archive Policy.