[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <20160424181837.GA32104@earth>
Date: Sun, 24 Apr 2016 20:18:37 +0200
From: Sebastian Reichel <sre@...nel.org>
To: Joe Perches <joe@...ches.com>
Cc: changbin.du@...el.com, apw@...onical.com,
linux-kernel@...r.kernel.org
Subject: Re: [PATCH] checkpatch.pl: add support for checking patch from git
repository
Hi,
On Sun, Apr 24, 2016 at 10:43:44AM -0700, Joe Perches wrote:
> On Sun, 2016-04-24 at 19:30 +0200, Sebastian Reichel wrote:
> > On Sun, Apr 24, 2016 at 09:07:21AM -0700, Joe Perches wrote:
> > > On Sun, 2016-04-24 at 17:10 +0200, Sebastian Reichel wrote:
> > > > On Sun, Apr 24, 2016 at 04:22:45AM -0700, Joe Perches wrote:
> > > > > On Sun, 2016-04-24 at 18:40 +0800, changbin.du@...el.com wrote:
> > > > > > This patch add "-g, --git" option that tread FILE as git commits
> > > > > > expression. You can specify the git commit hash ID expressions,
> > > > > > then these commits from your git repository will be checked.
> > > > > Why would anyone want to use checkpatch on commits already in git?
> > > > It may be in some non-public development branch. Usually when I
> > > > write patches I open a file, change it and commit the result or even
> > > > interim result to have backups and other git features available as
> > > > soon as possible. All testing is done later.
> > > >
> > > > So IMHO this is a really useful feature.
> > > I think it would be a more useful feature for
> > > something like a git pull request rather than
> > > a local git repository.
> > There are basically two places, where one wants to check patches:
> >
> > 1. When one creates/modifies patches
> > 2. When one wants to apply patches in some tree
>
> 3. when one wants to accept patches from a pull request.
>
> > I'm perfectly happy with checkpatch's current behaviour for
> > the second task. OTOH during development I would find it useful
> > if I can do something like "checkpatch --git HEAD~3..HEAD".
>
> So you can rework the patches that are already applied?
> What would you do if it showed errors/defects?
>
> Encouraging rework seems inefficient.
No, but I can do it with patches I'm currently writing. My workflow
when working on patches I write some stuff and commit it. When I
have reached a status where the result looks ok to me, I start to
prepare it for sending it to the mailinglists.
That involves squashing some patches, changing some descriptions,
running checkpatch, maybe reorder some patches. IMHO having a --git
option is not an encouragement, but makes it less annoying to run it
in this use case.
-- Sebastian
Download attachment "signature.asc" of type "application/pgp-signature" (820 bytes)
Powered by blists - more mailing lists