751863 - Cancelling requests of people who are no longer involved, and requests on non-open bugs

Status: RESOLVED WONTFIX

(bugzilla.mozilla.org :: Administration, task)

Description

[Gervase Markham [:gerv]]

This blog post proposes further steps to deal with the backlog of unaddressed reviews: 
http://blog.gerv.net/2012/03/old-requests-next-steps/

I would like to implement some of those by writing a script to cancel certain types of old request, while adding an explanatory comment. Here are the suggestions, with expanded info:

1) Cancel existing requests of requestees who have not been active in Bugzilla for > N days (suggested N: 60). As in bug 751862, this requires a backport of bug 240437, or updating to a version containing it.

2) Cancel existing requests on non-open bugs (are there any which could validly be on such bugs?) There are at least 264 of these, some very old:

https://bugzilla.mozilla.org/buglist.cgi?list_id=3021588;field0-0-0=flagtypes.name;cmdtype=doit;remtype=asdefault;query_format=advanced;bug_status=RESOLVED;bug_status=VERIFIED;bug_status=CLOSED;type0-0-0=equals;value0-0-0=review%3F

Gerv

Comment 1

[:Gavin Sharp [email: [e-mail]]]

(In reply to Gervase Markham [:gerv] from comment #0)
> 2) Cancel existing requests on non-open bugs (are there any which could
> validly be on such bugs?)

Yes, there are work flows that require making requests on closed bugs (bugs are marked FIXED when the patch lands on trunk, but often branch/backport patches require additional review), so we can't just indiscriminately clear them all.

Comment 2

[Gervase Markham [:gerv]]

gavin: is there a threshold after which such reviews are highly likely to be bogus? One cycle (6 weeks)? 2 cycles (12 weeks)?

Gerv

Comment 3

[Gervase Markham [:gerv]]

gavin: (reping) is there a threshold after which reviews on closed bugs are highly likely to be bogus? One cycle (6 weeks)? 2 cycles (12 weeks)?

Gerv

Comment 4

[:Gavin Sharp [email: [e-mail]]]

I don't know. What's the distribution of ages of existing requests on closed bugs?

Comment 5

[Gervase Markham [:gerv]]

Gavin: how does the answer to that question help you answer my question?

Getting your answer is tricky. Firstly, let me make sure I understand what you want. Is it:  a graph of days-since-bug-was-closed against number-of-bugs for all review-? requests in the whole of BMO?

Gerv

Comment 6

[:Gavin Sharp [email: [e-mail]]]

It wouldn't help answer your question, but it might help determine how much value we'd get from cancelling these requests. If it's a hard answer to get, don't bother; I'm not sure it's a great use of your or anyone else's time to pursue cancelling requests on closed bugs specifically (as opposed to dealing with old requests in general, ignoring the bug's resolution).

Comment 7

[Gervase Markham [:gerv]]

This got bogged down on the question above. Getting the data is not trivial. I think it is safe to do 1), and to do 2) for bugs closed more than 12 weeks ago.

Gerv

Comment 8

[:Gavin Sharp [email: [e-mail]]]

I don't think you should use the bug's status as a factor in determining the time threshold that you'll use to cancel requests. If you're not willing to cancel requests for open bugs older than N days, you shouldn't be cancelling requests on closed bugs either. I.e. I don't think doing 2) is worth it at all.

Comment 9

[Mark Côté [:mcote]]

No action on this bug for a year, and I heard a lot of negative feedback on automatically closing reviews.  I'm going to WONTFIX unless we want to reopen the discussion.