Re: Steward opinions on gerrit?

[Carlos O'Donell <carlos at redhat dot com>]

On 11/27/19 5:47 PM, Joseph Myers wrote:
> On Wed, 27 Nov 2019, Carlos O'Donell wrote:
> 
>> As stewards for glibc I wanted to ask each of you if you had
>> any opinions on gerrit, and thoughts on or the fundamental
>> requirements of a review system for a project like glibc.
>>
>> Joseph, 
>>
>> You write and review a lot of patches. Does using gerrit
>> interest you? Does it solve any problems?
> 
> I'm not convinced the experiment is advanced enough to tell whether it 
> helps much.  The main problem I see patch review systems as being for is 
> to track changes that are under consideration but not yet approved, and 
> it's hard to judge when there are only a few changes in there.  (I suspect 
> any system would gradually accumulate patches that were submitted by 
> one-off contributors, issues raised, but never revised and never 
> definitively rejected either.
Gerrit does track changes under consideration, but falls down because it
doesn't accept email input at various stages and is non-robust in this case
if you clip the mail footer. I'm not happy about that.

I'd like to try Patchwork v2.0 and do a comparison.

> There's potential for various kinds of CI integration with patch tracking, 
> but I don't think we've tried that.
 
We have not yet tried that. We haven't gotten past making the new patch
tracking work well enough that we aren't doing too much manual work.

The problem with Patchwork v1.0 seemed to be that we needed a lot of manual
work to keep the patch queue clean and so we didn't do it.

-- 
Cheers,
Carlos.