[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Code reviews

From: Stefan Monnier
Subject: Re: Code reviews
Date: Tue, 08 Mar 2016 00:30:31 -0500
User-agent: Gnus/5.13 (Gnus v5.13) Emacs/25.1.50 (gnu/linux)

> Something like this is what I meant with "Don't give commit access to
> the golden branches to everyone". Anyways, you can't expect having high
> quality commit logs (or VC history at all, take a look at the DAG to see
> what I mean) and give write access liberally.

Indeed (and yes, the DAG shape is another one of those problems, made
worse by the fact that Git doesn't help us).

> Having ChangeLogs as a compromise is absurd: you have to proof-read
> and fix them anyway.  Correct or reject the real thing (the
> commit) instead.

Agreed.  My choice has been to live with unfixed (and hence
lower-quality) commit messages [even if some people here do make the
effort to fix them in the ChangeLog, I don't benefit from it because
I don't use the ChangeLog].

> I guess that asking for a review queue is out of the question, although
> I'm afraid that it would not turn to be a good thing since some people
> here tend to be quite picky when reviewing foreign contributions

True enough on both counts.  A code review system should allow both to
"reject a commit and ask the author to provide a new patch" and to "fix
the commit ourselves" since in many cases it's easier both for the
contributors and the reviewer.


reply via email to

[Prev in Thread] Current Thread [Next in Thread]