Re: [PATCH v3] checkpatch: add new warnings to author signoff checks.

[Lukas Bulwahn <lukas.bulwahn@gmail.com>]

[-- Attachment #1: Type: text/plain, Size: 3197 bytes --]



On Tue, 6 Oct 2020, Dwaipayan Ray wrote:

> On Tue, Oct 6, 2020 at 2:39 AM Joe Perches <joe@perches.com> wrote:
> >
> > On Tue, 2020-10-06 at 01:37 +0530, Dwaipayan Ray wrote:
> > > On Tue, Oct 6, 2020 at 1:07 AM Joe Perches <joe@perches.com> wrote:
> > > > On Tue, 2020-10-06 at 00:54 +0530, Dwaipayan Ray wrote:
> > > > > The author signed-off-by checks are currently very vague.
> > > > > Cases like same name or same address are not handled separately.
> > > >
> > > > When you run tests for this, how many mismatches are
> > > > caused by name formatting changes like:
> > > >
> > > > From: "Developer, J. Random" <jrd@bigcorp.com>
> > > > ...
> > > > Signed-off-by: "J. Random Developer" <jrd@bigcorp.com>?
> > > >
> > > > Should these differences generate a warning?
> > > >
> > >
> > > Hi,
> > > I ran my tests on non merge commits between v5.7 and v5.8.
> > >
> > > There were a total of 250 NO_AUTHOR_SIGN_OFF Warnings
> > >
> > > 203 of these were email address mismatches.
> > > 32 of these were name mismatches.
> > >
> > > So for the name mismatches, the typical cases are like:
> > >
> > > 'From: tannerlove <tannerlove@google.com>' != 'Signed-off-by: Tanner
> > > Love <tannerlove@google.com>'
> > > 'From: "朱灿灿" <zhucancan@vivo.com>' != 'Signed-off-by: zhucancan
> > > <zhucancan@vivo.com>'
> > > 'From: Yuval Basson <ybason@marvell.com>' != 'Signed-off-by: Yuval
> > > Bason <ybason@marvell.com>'
> > > 'From: allen <allen.chen@ite.com.tw>' != 'Signed-off-by: Allen Chen
> > > <allen.chen@ite.com.tw>'
> > >
> > > I didn't find the exact formatting change you mentioned in my commit range.
> > > But I did find something like:
> > >
> > > 'From: "Paul A. Clarke" <pc@us.ibm.com>' != 'Signed-off-by: Paul
> > > Clarke <pc@us.ibm.com>'
> > >
> > > So it's like some have parts of their names removed, some have language
> > > conflicts, and yet some have well different spellings, or initials,
> > > etc. It's like
> > > a wide variety of things happening here.
> > >
> > > I think considering these, it should be warned about, and let people know
> > > that there might be something wrong going on.
> > >
> > > What do you think?
> >
> > Except for comments and quotes like:
> >
> >         From: J. Random Developer (BigCorp) <jrd@bigcorp.com>
> >         Signed-off-by: "J. Random Developer" <jrd@bigcorp.com>
> >
> > I think any time there's a mismatch, there
> > should be a warning emitted.
> >
> > That includes any subaddress detail difference.
> >
> >
> Hi,
> Yeah these cases are being handled.
> 
> Comments and quotes don't generate any warning message but
> all the other mismatches do.
> 
> Only the check for subaddress generates a --strict check message.
> others are all WARN messages. It was followed from our discussion at
> https://lore.kernel.org/linux-kernel-mentees/7b52e085f0b69ad1742966f8eacd02deb9299b96.camel@perches.com/
> 
> So does it need to be changed to a WARN or is it fine like that?
>

I will repeat what I suggested before:

I think the complete mismatch where we cannot even find a name or an email 
match the author deserves to be reported as ERROR.

Dwaipayan, if Joe does not disagree, could you change that in your PATCH v4?

Lukas