Patch frustration

[Adrian Buehlmann]

On 2012-06-07 02:31, Bryan O'Sullivan wrote:
> On Wed, Jun 6, 2012 at 5:19 PM, Adrian Buehlmann <adrian at cadifra.com
> <mailto:adrian at cadifra.com>> wrote:
> 
>     The current system of contributing and reviewing patches is an
>     absolute pain.
> 
> 
> I wouldn't go that far, but it is a bit of a problem, and I've heard
> other people mention it too.
> 
> Back in the day, there was the crew repository for small
> maintenance-oriented patches. It seems to have fallen into disuse, which
> is a shame.
> 
> Matt, for small, well contained patches, would you consider accepting
> stuff via crew (and maybe crew-stable) again? It would take some of the
> pressure off you, which might be helpful.

I'm not sure this can really be solved by a second-class repo, where
people push things that can't be commented on. It also shouldn't be
needed to bypass the official review channel in any way.

The crew repo is there, and Matt pulls from it. The thing is, people
have stopped pushing there. And I think it is because it is so
inflexible. And Matt also needs to pull and merge. Those merges are
often a bit meaningless.

What I have noticed, is a bit of a habit that patches that have a
comment, get dropped "because there is a comment". Of course, I think it
is probably difficult for Matt to know when a discussion has ended and
the patch has "died" without a resolution.

It's sometimes unclear in what state a patch is. Is it still cooking?

I have to admit that I have sometimes fallen in the trap of sometimes
getting annoyed by getting a comment on a patch. Yes, I know, it's
stupid. Perhaps a reviewer could say if a comment is blocking or just a
remark and the patch could be taken anyway.

Maybe reviewers could also say that a patch can be taken, even if it
needs an additional polishing fix.

FWIW, the thing I hate absoluteley the most is if I get a comment about
a code part that is totally unrelated to what I'm patching (hasn't
happened lately, luckily).

Another thing that comes to mind is that I think that discussions about
some topics can get a bit heated. Those involved are in the middle of
that thing and are working on getting a resolution of discussion items.
There are topics that are going on for many months. For outsiders of the
debate, it can be difficult to understand in detail what the state of
the discussion about obscure topic X is and who said what not so nice
thing to whom. I'm pretty sure people not interested in the boring
details about the testsuite can get annoyed by reading Mads and me
banging our heads on a problem. On the other hand, we don't have
anything to hide.