Re: [PATCH] checkpatch: Categorize some long line length checks

[Joe Perches <joe@perches.com>]

On Sat, 2015-05-23 at 13:32 +0200, Julia Lawall wrote:
> On Fri, 22 May 2015, Joe Perches wrote:
> > Many lines of code extend beyond the maximum line length.
> > Some of these are possibly justified by use type.
> >
> > For instance:
> >
> > structure definitions where comments are added per member like
> >
> > struct foo {
> > 	type member;		/* some long description */
> 
> I'm not super fond of the comment one.  Perhaps people could express
> themselves more concisely, or put the details elsewhere?

Concision is good, straining for brevity or bad
formatting isn't.

I've seen a lot of ugly patches lately to "fix"
code like this by making it worse.

By default, there is still a long_line warning for
this style.  It arguably could be appropriate to
keep some lines like this and this makes it easy
to tell people "add --ignore=<type>".

This patch shouldn't be applied right now anyway.

I think the idea is OK, but this implementation
could be improved and clarified by moving the
current exclusions before the classifications.

Anyone else have an opinion?

I'll send a V2 later unless there are more comments.