> I actually like it.
> If "Evil approval bot" mailed you warning that it will merge 12 pending
> changesets in two days if there's no action from your part, that would
> force some promptly action by you.
>

Not a chance in hell we'll ever allow this for the operations repo or
the mediawiki deployment branches.

> I recently had a trivial patch to operations/puppet waiting for more
> than a month. When I noticed I hadn't added any Reviewer, and added to
> it, the changeset was fixed in the same day. But that also shows that
> nobody looked for new changes there.
>

That's because ops tends to be swamped with requests constantly. We
work mostly on an interruption model. I doubt we'll change our
workflow any time soon to check for changes that haven't asked for a
review.

> I have also seen people approving their own commits to core, something
> I'm not comfortable with.
>

I dislike this, very much. The workflow in ops requires this, but
there's no reason it should happen in mediawiki core.

> I was also recently unhappy when I discovered that one patch I thought I
> had open, had been abandoned without explanation. There can be good
> reasons for doing that, this is a bad idea, no longer needed, fixed in a
> different way in I123456... or even "closing because it has been waiting
> for a new patch for too long and I don't like seeing this open" (which I
> suspect was the case), but *please*, if you're closing another people's
> patch, leave an explanation!
>

Indeed. Like bugs, something should never be dropped without saying why.

- Ryan

_______________________________________________
Wikitech-l mailing list
[email protected]
https://lists.wikimedia.org/mailman/listinfo/wikitech-l

Reply via email to