Hi Robert,

On Wed, 2007-06-27 at 04:20 -0700, Robert Hudson wrote:
> It is possible that a bug in Test Director can be “locked” by some-one
> updating it.  This means that when we do a commit the “read” checks in
> the pre-commit hook pass OK, but then when the post-commit hook tried
> to “update” the bug it fails to get the write lock required and so it
> doesn’t get the updates.

Are you saying that the lock changes between the time the
activity_verify is issued and the time the activity_commit is issued ?

Or are you just saying that the activity_verify simply doesn't check for
the lock ?
 
> What I’d like to add is a new method called something like
> “check_bug_lock” which can be called as part of the pre-commit hook to

Let's call it "integration_bug_lock_available"

>  see if the bug is in a state that will enable it to be updated.
> Ideally I’d like to add this to the integration layer so that it can
> be implemented by any bug-tracking system (obviously default will be
> to return not locked if bugs are never locked in the specific system).
> 
>  
> 
> What do you think?

If you were referring to the latter I think it makes perfect sense the
way you describe it:

http://bugzilla.mkgnu.net/show_bug.cgi?id=1015

> (I wanted to avoid fixing it specifically for Test Director in case it
> may also be an issue for other systems)

Good thinking. Go for it.

Attachment: signature.asc
Description: This is a digitally signed message part

_______________________________________________
scmbug-users mailing list
[email protected]
http://lists.mkgnu.net/cgi-bin/mailman/listinfo/scmbug-users

Reply via email to