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

Re: post-commit question

From: Jeff Lanzarotta <delux256-subversion_at_yahoo.com>
Date: 2007-09-19 17:18:52 CEST

How would a script fork itself into the background?

John Peacock <jpeacock@tigris.org> wrote: Stefan Sperling wrote:
> The commit has a write lock on the repo, and your script waits
> for the write lock to be released so it can grab a read lock
> to do the update.

I think this theory is highly unlikely (or even wrong ;-), since all
post-commit hooks that expect to be able to update a working copy would
suffer from this problem otherwise. It should be simple to check,
however, simply replace that post-commit with one that forks into the
background and does a sleep(1). That way, the transaction will be
"completed" by the time the update happens.

I suspect that, even though the OP claims to have tested with various
users, this is actually a rights issue. I'm especially suspicious of
the fact that this appears to be a remote filesystem of some type:


FWIW, I am the author of SVN::Notify::Mirror, which was written
specifically to facilitate both local and remote WC updates in a
post-commit hook. Check CPAN for details...


To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org
Received on Wed Sep 19 17:19:44 2007

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.