Re: [PATCH 00/11] Moving global state into the repository object (part 2)

From: Jonathan Nieder <jrnieder@gmail.com>
Date: Feb 28, 2018, 19:19 UTC
Message-ID: <20180228191931.GF251290@aiede.svl.corp.google.com>
In-Reply-To: <CAGZ79kaDVrn6Qw_xsdyBST6XfJMXs43qDWx8jY2J9VsXwDKm5Q@mail.gmail.com>

Hi,

Stefan Beller wrote:

> On Wed, Feb 28, 2018 at 9:57 AM, Junio C Hamano <gitster@pobox.com> wrote:

Show 8 quoted lines

>> Wait a minute.  Is that topic ever shown to work well together with
>> other topics in flight and are now ready to be built upon?  I had an
>> impression that it is just starting to get serious reviews.
>
> And I had the impression the serious reviews were done and fine;
> the only issue would be demonstrating its working fine with other
> series, where I was also worrying about conflicts with
> brians series. And to address that, I'd just send series in small sizes.

Some of the patches looked cooked to me and others still do not look cooked yet. I marked the former with Reviewed-by. In general, a few things can help to make the process easier for me:

 1. Giving a quick reply to a review to say how the comments were
    resolved, sometimes even with a resend of that one patch to
    illustrate.  That way the conversation can continue and the
    individual patch can get to a reviewed state faster, without
    having to chase between different rerolls of the entire series.

    This also has an effect of making the review process more
    collaborative: perhaps after seeing how you address their
    comments, a reviewer may have another idea that they suggest via a
    patch to squash in, etc.

 2. In a reroll, summarizing the result of previous reviews by
    including acks as appropriate and Reviewed-by if a reviewer
    granted it.  This helps with reviewing the reroll since it tells
    people where to focus their attention.

[...]

> Is there anything that a contributor can help with that eases
> refactoring series in flight?

For helping reviewers, see above.

For helping Junio, what I've seen people occasionally do is to locally run a "trial merge" against next and pu and see what semantic or lexical conflicts arise. In the cover letter you can describe these and give Junio advice to make applying the patch easier for him.

[...]

> Sorry for the miscommunication, though,

FWIW, even though part 1 doesn't look done to me yet, it looks *close* to done, and I was happy to see the sneak peek at part 2.

Thanks, Jonathan