Hi Oliver,

On Mon, 2007-12-24 at 09:12 +0100, Oliver Schäfer wrote:
> I have one more question about the synchronous approach in Scmbug,  
> how do you guarantee that there wasn't an external modification to  
> the bugtracker system? You can't, so we aren't real 'thread safe',  
> are we? E.g, what about a version which was added to a product while  
> we are in a labeling operation?

This is certainly flawed.

This consistency of committing dirty data back to the database should be
provided by the Bugzilla API really. We won't get into the business of
implementing that, too. So for Bugzilla versions prior to 2.22.x we just
don't do it, and manually reimplement access to the database, and for
Bugzilla 3.x, where the API is more modular and reusable, we rely on the
API to do it.

> One more thought: What if the user hasn't the permission to add a  
> version to a product? This is very common for bugzilla.

I'm guessing that Bugzilla has some "status_resolution_change" function,
or something similar that at least reports back an error when there are
errors (including a permission error). If the error code is propagated
correctly back to us, we can provide a much more informative message. If
not, well, that's something we need to alert the Bugzilla developers
about and wait for them to fix 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