Re: [PATCH 1/2] xread: retry after poll on EAGAIN/EWOULDBLOCK

From: Jeff King <peff@peff.net>
Date: Jun 27, 2016, 19:17 UTC
Message-ID: <20160627191720.GE9594@sigill.intra.peff.net>
In-Reply-To: <CAGZ79kZ94PaOfq3GimWiHULbTE7ihMzL9S=Y+npQ4F5gGwFrsA@mail.gmail.com>

On Mon, Jun 27, 2016 at 09:49:06AM -0700, Stefan Beller wrote:

Show 8 quoted lines

> Quite a while ago, when I started doing code reviews professionally, I
> wondered if the code review procedure can be semi-automated, as
> automation helps keeping the error rate low. By that I mean having a
> check list which I can check off each point for each patch. That seems
> to be very good in theory, but when trying it I was finding myself
> doing a lot of unneeded work as some points of such a check list just
> do not apply for a specific patch. So I did not follow through with
> that.

I have wondered, too, if we could have better tooling to help us with reviews. But one of the things I really _like_ about doing reviews for git (versus other projects) is that doing review via email is unconstrained. The primary recipient is human, and I can format and say whatever I like in the way that best communicates to the human, without worrying about fitting my comments into a pre-made form.

That being said, I suspect one could go a long way by picking out basic patterns from emailed responses. For example, you could imagine a system that makes a todo list of review comments (one comment per response to a quoted section) and associates them with given bits of the code (by seeing what's in the quoted section). That todo list can become a checklist when sending out the next revision, or could even be used interactively to see what happened to each code spot (did you fix it? How? In which commit?). That would help reviewers, but also would help submitters send out the cover letter for the next version (by reminding them what to mention).

Of course, none of that would have helped my comment, which was in a "PS" several emails deep in a discussion thread. ;)

-Peff