linux-kernel - Re: Moratorium on coding style patches (was Re: [PATCH] include/linux/kernel.h: fixed coding style issues)

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite for Android: free password hash cracker in your pocket

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <yw1xtwa6tm4g.fsf@unicorn.mansr.com>
Date:   Wed, 14 Dec 2016 18:13:19 +0000
From:   Måns Rullgård <mans@...sr.com>
To:     Alexey Dobriyan <adobriyan@...il.com>
Cc:     piotrgregor@...ncme.org,
        Linus Torvalds <torvalds@...ux-foundation.org>,
        Linux Kernel <linux-kernel@...r.kernel.org>
Subject: Re: Moratorium on coding style patches (was Re: [PATCH] include/linux/kernel.h: fixed coding style issues)

Alexey Dobriyan <adobriyan@...il.com> writes:

> OK, someone needs to say it.
>
> These type of patches are advertised by some people as a good way
> to enter Linux kernel development. You know, to learn how the process
> works, how to setup email client pipeline, to get initial feedback.
> And it is true. What those people aren't saying is that the above is
> about ~0.01% of what is kernel or any other project development is about.
> It is the easy part.
>
> But the patches also create problems for those who are already in.
> The very immediate is that "git-blame" stops working. It simply points
> to the irrelevant commit and developer is forced to either search
> manually through the history or search the web for "git-blame" options.
> (maybe there is such an option but that's not the point).
>
> And "git-blame" usually happens in important cases: when developer
> searches for a possible bugfix or wondering who wrote that crap.
>
> Au contraire, coding style patch is something unimportant.
> Whitespace here, whitespace there, who cares. On the grand scale,
> coding style compliance is important but in my experience Linux
> kernel CS compliance is top notch for the project of Linux's size.
>
> So the tradeoff is not in the patch favour and all you need to follow
> coding style is basically "indent -kr -i8 -l80" for new code.
> It is then becomes non problem because editors defaults are doing
> the job.
>
> And they create rejects against other non-coding style patches,
> again slowing down people who need to fix real problems.

I agree with all of that, yet I still sometimes create pure style
patches.  This tends to happen when I can't look at the code I'm trying
to debug without wanting to claw my eyes out.  In such cases, an initial
cleanup patch often makes the actual fix much easier to comprehend.
There's not a lot such code in the kernel, thankfully, but it does
exist.

> Said that, I call for a tree wide moratorium on pure coding style changes.

In light of the above, I'd relax this slightly and say don't do style
patches unless as a prelude to some real work.

-- 
Måns Rullgård