[Reviewers] Heads up: new accept process is live

[Pierre-Yves David]

On 05/13/2016 08:23 PM, Matt Mackall wrote:
> The new accept process we discussed at the sprint is now up and running. And to
> make sure it gets tested, I'm going to focus on tooling rather than review for a
> while. So to keep patches flowing, you'll want to dig into this sooner rather
> than later.
>
> https://www.mercurial-scm.org/wiki/AcceptProcess

I think I sense a small pathological case in the current state:

If I understand the process correctly, If there is a 30 patches stack
pushed by other people (so "reviewed" by them) and I've to fix a
typo/small error in the first one. I'll repush the whole 40 to
hg-commited and they will be marked as reviewed by me despite the fact
if only looked at the first one. This seems like a problem.

In its review tool, Logilab have some basic code to detect rebase that
did not alter a patch (and carry its review property unaltered), we can
probably reuse it:

https://hg.logilab.org/review/cubes/vcreview/file/tip/__init__.py

-- 
Pierre-Yves David