stale MR triaging - you can help!

Harald Sitter sitter at
Fri Oct 16 10:47:06 BST 2020

On 16.10.20 10:00, Albert Astals Cid wrote:
> El dimarts, 13 d’octubre de 2020, a les 15:53:10 CEST, Harald Sitter va escriure:
>> Moin
>> As promised in a previous mail I've set up a system that helps us find
>> stale MRs to ensure patches don't fall by the wayside.
>> Lists are filed as issues on invent. To help out you can hit the bell to
>> subscribe to `New issue` events at
>> First list of stale MRs here:
>> To deal with a stale MR you'll want to check the box and either make
>> sure someone takes a look at the MR (e.g. by @mentioning someone who may
>> have experience) or doing the review yourself.
>> Unlike previously described I've refined the behavior a bit:
>> a) MRs are not linked but listed with verbatim url, it sucks a bit but
>> it prevents MRs from updating due to the back-reference an actual
>> mention would cause
>> b) stale MRs are now MRs that have not seen activity within 4 weeks (up
>> from 2 - was woefully too short a time frame)
>> c) MRs without subscribers after 2+ weeks are recorded separately
>> d) MRs without subscribers after 1+ week where the author is not a KDE
>> dev are also recorded (currently a bit bugged as it turns out)
>> The motivation with c and d is to prevent MRs from going stale to begin
>> with, ideally anyway.
> Can we get the issues in each of those 4 categories in different sections of the issue you create?
> It makes more sense for me to look at d) than at some MR whose author is Aleix (for example), if it's stale, he knows how to do how to un-stale it, while a newbie may not know how to.

Indeed. That's actually how it works, we've already dealt with the
important ones this week though. So all that's left is the mountain of
otherwise randomly stale MRs ;)


More information about the kde-devel mailing list