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

Re: FSFS hotcopy issues

From: Max Bowsher <maxb_at_ukf.net>
Date: 2005-09-24 23:53:09 CEST

Greg Hudson wrote:
> On Sat, 2005-09-24 at 22:33 +0100, Max Bowsher wrote:
>> However, hotcopy is a corner case. Current FSFS hotcopy neither takes the
>> write-lock, nor does it refrain from creating the format file until the
>> FS
>> is complete in every other way.
>
> So the destination could wind up looking like a repository when it's not
> complete, I guess. Not sure if that's a real concern.

A very minor concern, I guess. I'll fix it when I fix the equivalent bug in
libsvn_repos.

>> On a related note, I think I may have discovered a hotcopy race condition
>> related to locking:
>
>> OOPS! Now there is a lock in the hotcopied repository, for a path which
>> does NOT exist in the hotcopied repository.
>
> Is that a bug? I thought we could lock nonexistent paths.

The design might be capable of that, but the client certainly disallows it.
Regardless of that, it is defiinitely a bug to propagate a lock to a
hotcopy, if the lock never existed on the revision which was selected as the
head revision of the hotcopy.

Max.

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Sat Sep 24 23:53:58 2005

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.