Hi,
On 2022-07-18 13:34:52 -0700, Jacob Champion wrote:
> On 7/18/22 12:32, Andres Freund wrote:
> > I'm not following - I'm talking about the patch author needing a while to
> > address the higher level feedback given by a reviewer. The author might put
> > out a couple new versions, which each might still benefit from review. In that
> > - pretty common imo - situation I don't think it's useful for the reviewer
> > that provided the higher level feedback to be removed from the patch.
>
> Okay, I think I get it now. Thanks.
>
> There's still something off in that case that I can't quite
> articulate... Is it your intent to use Reviewer as a signal that "I'll
> come back to this eventually"?
That, and as a way to find out what I possible should look at again.
> As a signal to other prospective reviewers that you're handling the patch?
Definitely not. I think no reviewer on a patch should be taken as
that. There's often many angles to a patch, and leaving trivial patches aside,
no reviewer is an expert in all of them.
> How should a CFM move things forward when they come to a patch that's been
> responded to by the author but the sole Reviewer has been silent?
Ping the reviewer and/or thread, ensure the patch is needs-review state. I
don't think removing reviewers in the CF app would help with that anyway -
often some reviewers explicitly state that they're only reviewing a specific
part of the patch, or that looked at everything but lack expertise to be
confident in their positions etc. Such reviewers might do more rounds of
feedback to newer patches, but the patch might still need more feedback.
ISTM that you're trying to get patches to have zero reviewers if they need
more reviewers, because that can serve as a signal in the CF app. But to me
that's a bad proxy.
Greetings,
Andres Freund