Re: CodingStyle vs checkpatch for block comments

[Joe Perches <joe@perches.com>]

On Tue, 2012-04-03 at 13:25 -0400, Chris Metcalf wrote:
> (Adding linux-kernel to the cc's and restating the question...)
> 
> I accepted a patch from Kautuk Consul and adjusted his comment style, but
> he pointed out that he had changed his comment style based on a warning

It's a check not a warning.

> from checkpatch.
> 
> The Documentation/CodingStyle file says:
> 
>     The preferred style for long (multi-line) comments is:
> 
>        /*
>         * This is the preferred style for multi-line
>         * comments in the Linux kernel source code.
>         * Please use it consistently.
>         *
>         * Description:  A column of asterisks on the left side,
>         * with beginning and ending almost-blank lines.
>         */
> 
> However, a new change to "checkpatch --strict" by Joe Perches, commit
> aad4f61498, causes this construct to be flagged as a warning

checkpatch flags it with a "check" only when --strict
is on the checkpatch command line.

>  if (and only
> if) it is preceded by a blank line.  Joe said the change was to support
> David Miller's preferred style, but that he didn't much care one way or
> another.

I think the block comment style is less important than
the actual content of the block comment.  The actual
comment content is rather difficult for checkpatch to
parse though.

> The relevant code in checkpatch.pl is:
> 
>                 if ($rawline =~ /^\+[ \t]*\/\*[ \t]*$/ &&
>                     $prevrawline =~ /^\+[ \t]*$/) {
>                         CHK("BLOCK_COMMENT_STYLE",
>                             "Don't begin block comments with only a /*
> line, use /* comment...\n" . $hereprev);
>                 }
> 
> So, my questions -
> 
> 1. I'm not sure what the regexps are really trying to avoid.  Presumably a
> blank line followed by a block comment is OK?  Certainly the kernel sources
> are full of this construct.

It emits a check message on
<blank line>
	/*

but not
<blank line>
	/* some actual comment

> 2. The actual warning message emitted seems to directly contradict the
> CodingStyle document, so presumably we should either clarify the message,
> or update CodingStyle if we're really trying to change the style.

Or just remove it or add a test for the patched file
to be in net/... or drivers/net... or something.