Re: FYI: Status of gdb usage of gerrit

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

On 11/25/19 2:12 PM, Florian Weimer wrote:
> * Carlos O'Donell:
> 
>>> I can post a detailed review, but given the low level of interest we
>>> seem to have in Gerrit, I'm not sure if that is necessary.
>>  
>> Please do. I would like to hear that and have it written down for the
>> record.
> 
> I'm not sure if this would be a good use of my time because at this
> point, it does not seem particularly likely to me that we are going to
> switch to Gerrit.
> 

Why?

If we don't discuss this then we can't learn what areas gerrit was bad 
at and apply the same critera of evaluation for the next tool.

(1) Email.

For example I think several of us have unspoken assumptions about what
we want from a code-review / CI / CD system.

Firstly, *nothing* will integrate with email in any serious way that
also allows:
* comment tracking
* side-by-side diffs
* quick version diffing with side-by-side diffs

If email is a success criteria then we have lost, and the only real option
is to use something like patchwork 2.0 which should have:

* same old email review workflow
* patch tracking
* CI/CD

In which case I need to:
* manually track comments from previous reviewers [time lost]
* use local tooling for side-by-side diffs (suggestions?)
* use local tooling for version-by-version changes (suggestions?)

I'm OK doing manual comment tracking.

(2) All or nothing.

Do we all need to switch to gerrit or only those that want to?
- Is that exclusionary of people in the community that don't
  want to use gerrit to review a particularly patch?
  - They could ask the review to be moved to email?

Would it be wrong if we had two ways to submit patches?

I just did a *bunch* of reviews in gerrit becuase it was easy
to see what's ready, merged/still applies, needs comments
answered, etc.

I can review by email too, it just requires more local tooling
to be setup.

There is some experiments going on with kernel + gerrit:
https://twitter.com/dvyukov/status/1161246539051937792
https://linux.googlesource.com/Documentation#gerrit-code-reviews-for-the-linux-kernel

-- 
Cheers,
Carlos.