Re: [PATCH] CF app: add "Returned: Needs more interest" - Mailing list pgsql-hackers

Hi Andres,

My intention had not quite been for this to be a referendum on the
decision for every patch -- we can do that if it helps, but I don't
think we necessarily have to have unanimity on the bucketing for every
patch in order for the new state to be useful.

On 8/3/22 12:46, Andres Freund wrote:
>> - https://commitfest.postgresql.org/38/2482/
> 
> Hm - "Returned: Needs more interest" doesn't seem like it'd have been more
> descriptive? It was split off a patchset that was committed at the tail end of
> 15 (and which still has *severe* code quality issues). Imo having a CF entry
> before the rest of the jsonpath stuff made it in doesn't seem like a good
> idea
There were no comments about code quality issues on the thread that I
can see, and there were three people who independently said "I don't
know why this isn't getting review." Seems like a shoe-in for "needs
more interest".

>> - https://commitfest.postgresql.org/38/3338/
> 
> Here it'd have fit.

Okay. That's one.

>> - https://commitfest.postgresql.org/38/3181/
> 
> FWIW, I mentioned at least once that I didn't think this was worth pursuing.

(I don't see that comment on that thread? You mentioned it needed a rebase.)

IMO, mentioning that something is not worth pursuing is not actionable
feedback. It's a declaration of non-interest in the mildest case, and a
Rejection in the strongest case. But let's please not say "meh" and then
Return with Feedback; an author can't do anything with that.

>> - https://commitfest.postgresql.org/38/2918/
> 
> Hm, certainly not a lot of review activity.

That's two.

>> - https://commitfest.postgresql.org/38/2710/
> 
> A good bit of this was committed in some form with a decent amount of review
> activity for a while.

But then the rest of it stalled. Something has to be done with the open
entry.

>> - https://commitfest.postgresql.org/38/2266/ (this one was particularly
>> miscommunicated during the first RwF)
> 
> I'd say misunderstanding than miscommunication...

The CFM sending it said, "It seems there has been no activity since last
version of the patch so I don't think RwF is correct" [1], and then the
email sent said "you are encouraged to send a new patch [...] with the
suggested changes." But there were no suggested changes left to make.

This really highlights, for me, why the two states should not be
combined into one.

> It seems partially stalled due to the potential better approach based on
> https://www.postgresql.org/message-id/flat/15848.1576515643%40sss.pgh.pa.us ?
> In which case RwF doesn't seem to inappropriate.

Those comments are, as far as I can tell, not in the thread. (And the
new thread you linked is also stalled.)

>> - https://commitfest.postgresql.org/38/2218/
> 
> Yep.

That's three.

>> - https://commitfest.postgresql.org/38/3256/
> 
> Yep.

That's four.

>> - https://commitfest.postgresql.org/38/3310/
> 
> I don't really understand why this has been RwF'd, doesn't seem that long
> since the last review leading to changes.

Eight months without feedback, when we expect authors to turn around a
patch in two weeks or less to avoid being RwF'd, is a long time IMHO. I
don't think a patch should sit motionless in CF for eight months; it's
not at all fair to the author.

>> - https://commitfest.postgresql.org/38/3050/
> 
> Given that a non-author did a revision of the patch, listed a number of TODO
> items and said "I'll create regression tests firstly." - I don't think "lacks
> interest" would have been appropriate, and RwF is?

That was six months ago, and prior to that there was another six month
silence. I'd say that lacks interest, and I don't feel like it's
currently reviewable in CF.

>> (Even if they'd all received skeptical feedback, if the author replies in
>> good faith and is met with silence for months, we need to not keep stringing
>> them along.)
> 
> I agree very much with that - just am doubtful that "lacks interest" is a good
> way of dealing with it, unless we just want to treat it as a nicer sounding
> "rejected".
Tom summed up my position well: there's a difference between those two
that is both meaningful and actionable for contributors. Is there an
alternative you'd prefer?

Thanks for the discussion!
--Jacob

[1] https://www.postgresql.org/message-id/20211004071249.GA6304%40ahch-to