Re: [PATCH] checkpatch: don't encourage new code to use "networking" style comments

[Brian Norris]

On Fri, Apr 28, 2017 at 03:31:33PM -0400, David Miller wrote:
> From: Brian Norris <briannorris@xxxxxxxxxxxx>
> Date: Fri, 28 Apr 2017 12:27:22 -0700
> 
> > On Fri, Apr 28, 2017 at 11:24:18AM -0700, Joe Perches wrote:
> >> On Fri, 2017-04-28 at 10:55 -0700, Brian Norris wrote:
> >> I believe the only person that actually cares about
> >> the networking
> >> comment style is David Miller.
> > 
> > Which is why I've CC'd him. If even *he* doesn't care about having
> > this warning in checkpatch, then why should anyone else?
> 
> Well it potentially saves one round trip for patch submissions.
> 
> What I'm not going to do is let people start using different comment
> style even for new code in files like net/core/whatever.c after I've
> spent nearly two decades getting them to be one way so far.

Well, that sounds like mixed messages to me, but I suppose you're the
(networking) boss.

FWIW, I don't see this consistently applied at all. Unless my regexes
are completely wrong [*], it's roughly 50/50 in drivers/net/ and net/,
and roughly 40/60 (favoring "net" style) in net/core/.

But if that's still the rule, then I guess I'll let this patch drop. And
hope that I'm not the next one Linus notices using the net style and
yells at.

Brian

[*] Which they quite likely are. But here goes:

  Multiline comment with '/*' on first line:
  \/\*$
  Potential (doesn't catch all) multiline comment with '/*' followed by
  text:
  \/\*[^\*][^\*]*$
  FWIW, kerneldoc gets counted for neither.