[Linux-kernel-mentees] checkpatch.pl investigation: NO_AUTHOR_SIGN_OFF issues

Dwaipayan Ray dwaipayanray1 at gmail.com
Tue Sep 22 19:08:38 UTC 2020


>
> Generally, I think it is a good first proof of concept.
> I believe you that functionality 'basically' works; again, we might
> already want to run a full-scale evaluation on that. Just to see
> if there are some impacts we might not be aware of yet.
>
> As you already wrote, we, Joe, you and me, need to figure out
> now all the further details:
>
> - how can we avoid the duplicate code in checkpatch.pl and
> get_maintainers.pl?
>
> - what is performance impact, especially as AUTHOR_SIGN_OFF check is not
> triggered often, and there are many other rules in checkpatch.pl?
>
> - further details, such as why do we need the lk_path is the first place?
> and many more questions of that kind.
>
> Feel free to sketch a first commit message and create a PATCH RFC for the
> discussion with Joe.
>
>
> Lukas
>

Hi,
As for the lk_path, it can be removed easily. To me too, it didn't make much
sense since it was just a duplicate, as $root should contain the same.

But again due to some reason,
$root in checkpatch had the value ".",
while $lk_path in get_maintainer had the value "./"

I have no idea yet if this was a design decision or just different handling.

So, I can change the part where I referenced the mailmap file by adding
a trailing / with $root rather than $lk_path. That should do it.

And for the duplicate code part, Joe did mention that either I could copy
or place the read_mailmap code in a separate file and reference both
checkpatch and get_maintainers from there.

To me, copying seems much feasible because the referenced part of the
mailmap handling code here is very small as there are some minor
changes.

Thanks,
Dwaipayan.


More information about the Linux-kernel-mentees mailing list