[Linux-kernel-mentees] [PATCH v5] checkpatch: add fix and improve warning msg for Non-standard signature
joe at perches.com
Tue Dec 1 18:39:24 UTC 2020
On Tue, 2020-12-01 at 19:21 +0100, Lukas Bulwahn wrote:
> On Tue, Dec 1, 2020 at 6:24 PM Joe Perches <joe at perches.com> wrote:
> > On Tue, 2020-12-01 at 16:59 +0530, Aditya Srivastava wrote:
> > > Currently, checkpatch.pl warns for BAD_SIGN_OFF on non-standard signature
> > > styles.
> > >
> > > This warning occurs because of incorrect use of signature tags,
> > > e.g. an evaluation on v4.13..v5.8 showed the use of following incorrect
> > > signature tags, which may seem correct, but are not standard:
> > I'm not a fan of this patch.
> > There is already a "non-standard" signature warning for
> > all of these cases since 2012, predating the range of this
> > retrospective evaluation by over 5 years and yet these
> > existing commits have been accepted.
> > The value in actual standardization and effectively
> > requiring specific signature style tags is quite low.
> > Anyone that signed a thing a particular way should be free
> > to sign the thing as they choose.
> > Most of these warnings would also still be in the tree in
> > the future in new patches as running checkpatch without
> > it emitting a message of any type isn't a requirement nor
> > should checkpatch use actually be required workflow.
> Can we scale this fixing feature down to the very obvious synonyms
> that simply do not add anything but confusion?
> Such as for those four here:
> Co-authored-by (count: 43) => Co-developed-by
I've never been a big fan of "Co-developed-by" as a signature tag,
but a "this should be that" here could be ok.
> Reviewed-off-by (count: 5) => Reviewed-by
I don't see value. If no one notices a BAD_SIGN_OFF
for the Reviewed-off-by:, I doubt this would add anything.
> Proposed-by (count: 5) => Suggested-by
> Suggestions-by (count: 3) => Suggested-by
Suggestions-by is not suggested-by as these suggestions could
have been in response to an initial patch proposal and the
author could have incorporated those suggestions.
> Then, we can probably also drop the rationale because it is pretty clear.
> Of course, the impact might be really zero, given that it is unclear
> if those authors did actually ever run checkpatch in the first place.
> Joe, if you see no value in even such a minimal fix feature, let us
> drop that idea and move on. There are enough other things to work on.
Maybe only add the Co-authored-by: -> Co-developed-by: check.
But IMO: none of this is particularly useful.
More information about the Linux-kernel-mentees