Daniel Rall wrote:
> On Fri, 12 Jan 2007, Nathan Kidd wrote:
>
>> Malcolm Rowe wrote:
>>
>>> I'd rather we try to make it impossible to have orphaned locks in the
>>> first place :-) Is that a hard problem?
>
>> Any repo of mainly non-mergable data that uses a 'svn:needs-lock on all
>> files' model is extremely likely to have these orphaned locks. There
>> needs to be a way to clean old/existing repositories. Of course
>> remove-zombie-locks.py can do this now, it's a question of whether this
>> is such a core problem that it should be built into svnadmin or not.
>
> Hmm, I've been wondering the same thing as Malcolm. Why don't
> operations which could cause orphaned locks -- 'delete', 'move' (?),
> etc. -- check whether the target and child paths have any of the lock
> properties set on them, and take the appropriate action?
I guess my reply wasn't too clear. You're right, of course, that these
operations should be fixed, that's what Issue #2507 is about. Fixing
it, however, only means that no new orphans will be created in the
future. It doesn't do anything about already-existing orphans. That's
what this thread is about: how should the user remove pre-existing
orphaned locks.
remove-zombie-locks.py already can do this, but since the problem will
trivially occur by default under normal usage until such time as #2507
is fixed and people upgrade to this fixed version, it has potential to
affect many, many people. cmpilato suggested, I think rightly, that
given the potential widespread nature of the problem it makes sense to
teach svnadmin how to fix the problem.
-Nathan
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Mon Jan 15 21:53:39 2007