Content-type: text/plain; charset=us-ascii
Precedence: list
MIME-version: 1.0 (Mac OS X Mail 11.5 \(3445.9.1\))
Subject: Re: [alpine-devel] How to improve quality control for patch reviews
From: Chris Ely <tcely@icloud.com>
In-reply-to: <1532897225.933218.1456675080.62DDC693@webmail.messagingengine.com>
Date: Mon, 30 Jul 2018 15:15:01 -0400
Cc: alpine-devel@lists.alpinelinux.org
Content-transfer-encoding: quoted-printable
Message-id: <AACD560B-34B5-4C8F-BAB1-D2B21A324597@icloud.com>
References: <20180726141558.2d451763@ncopa-desktop.copa.dup.pw>
 <1532897225.933218.1456675080.62DDC693@webmail.messagingengine.com>
To: Jean-Louis Fuchs <ganwell@fangorn.ch>

> On Jul 29, 2018, at 4:47 PM, Jean-Louis Fuchs <ganwell@fangorn.ch> =
wrote:
>=20
> Hi
>=20
> On Thu, Jul 26, 2018, at 05:15, Natanael Copa wrote:
>> - give more people push access. Look for people that are candidates =
to
>>   get push access. Help them to improve. Follow up when they are =
"good
>>   enough".
>=20
> In theory you don't need push access to review a PR:
>=20
> - If there is a problem anybody can state it on github or the =
mailinglist
> - If one thinks the PR is ok one can write "looks good to me"

The problem I've seen with this approach is that PRs don't change state =
until someone with "write" access to the aports repository on GitHub =
comes along. If the set of people with  "push" access to =
git.alpinelinux.org is relatively small and the subset of those people =
with write access on GitHub is even smaller, I don't see how keeping up =
with aports changes is going to remain possible.

> Idea: maybe it is possible that more people are allowed to add a =
label: "looks good", but without push access. So people with push access =
can work faster.

This is the idea behind pull request approvals. I've suggested using =
https://github.com/zalando/zappr for this in a previous email to this =
list.

We'll need to determine the best way to expand the pool of of people who =
can change a PRs labels / assignments, etc. on GitHub. I think =
collaborators is a good way, but that hasn't been tested yet, as far as =
I know. Ultimately, an accepted PR can't really be "complete" until it =
gets merged into master back on git.alpinelinux.org. Expanding the =
number of people with "push" access to git.alpinelinux.org might just be =
unavoidable here as the amount of work increases with every new active =
software package.=


---
Unsubscribe:  alpine-devel+unsubscribe@lists.alpinelinux.org
Help:         alpine-devel+help@lists.alpinelinux.org
---