On Mon, 2020-11-23 at 22:54 +0530, Aditya Srivastava wrote:
> Currently, checkpatch.pl warns for BAD_SIGN_OFF on non-standard signature
> styles.
I think this proposed change is unnecessary.
> 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:
Standards are useful, but standards are not constraints.
> 1) Requested-by (count: 48) => Suggested-by
> Rationale: In an open-source project, there are no 'requests', just
> 'suggestions' to convince a maintainer to accept your patch
There's nothing really wrong with some non-standard signatures.
And I think leaving humor like brown-paper-bag-by: is useful.
Just telling people that they are using a non-standard signature
I think is enough.
On 23/11/20 11:03 pm, Joe Perches wrote:
> On Mon, 2020-11-23 at 22:54 +0530, Aditya Srivastava wrote:
>> Currently, checkpatch.pl warns for BAD_SIGN_OFF on non-standard signature
>> styles.
>
> I think this proposed change is unnecessary.
>
>> 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:
>
> Standards are useful, but standards are not constraints.
>
>> 1) Requested-by (count: 48) => Suggested-by
>> Rationale: In an open-source project, there are no 'requests', just
>> 'suggestions' to convince a maintainer to accept your patch
>
> There's nothing really wrong with some non-standard signatures.
> And I think leaving humor like brown-paper-bag-by: is useful.
>
> Just telling people that they are using a non-standard signature
> I think is enough.
>
Hi Joe
Thanks for reviewing. We were also planning to provide fix for certain
non-standard signature warnings due to typo mistake in the signoffs,
using edit distance approach. These signatures were probably not
intended by the user.
E.g. for signatures like: 'Reviwed-by:', 'Singed-off-by:',etc.
Here is the list I have generated for non-standard signatures with
edit distance of 2 or less along with their count (over v4.13..v5.8).
(Among total 539 Non-standard signature warnings, warnings caused by
typo mistakes are 85 in number):
https://github.com/AdityaSrivast/kernel-tasks/blob/master/random/non_standard_signature/less_than2/signs_freq.txt
This is the predicted correct signoffs we are getting (for less than
or equal to 2):
https://github.com/AdityaSrivast/kernel-tasks/blob/master/random/non_standard_signature/less_than_3.txt
The reason I have chosen 2 as threshold is that count 3 onwards, the
results start deviating.
List for edit distance 3:
https://github.com/AdityaSrivast/kernel-tasks/blob/master/random/non_standard_signature/equal_3.txt
What do you think?
Thanks
Aditya
On Mon, Nov 23, 2020 at 6:33 PM Joe Perches <[email protected]> wrote:
>
> On Mon, 2020-11-23 at 22:54 +0530, Aditya Srivastava wrote:
> > Currently, checkpatch.pl warns for BAD_SIGN_OFF on non-standard signature
> > styles.
>
> I think this proposed change is unnecessary.
>
> > 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:
>
> Standards are useful, but standards are not constraints.
>
Agree, but we do try to create statistics and try to derive quality
statements from those tags (yes, empirical software engineering black
magic...).
Hence, I am in favor of suggesting to rewrite those tags that really
do not add anything at all. E.g., Suggestions-by: vs. Suggested-by, or
Coauthored-by vs. Co-developed-by.
Anyone can ignore checkpatch; so it is not a constraint unless
enforced by subsystem maintainers.
> > 1) Requested-by (count: 48) => Suggested-by
> > Rationale: In an open-source project, there are no 'requests', just
> > 'suggestions' to convince a maintainer to accept your patch
>
> There's nothing really wrong with some non-standard signatures.
> And I think leaving humor like brown-paper-bag-by: is useful.
>
I think we do not want to take the humor and fun away from patches.
So let us not suggest deleting the humorous and celebrating ones.
> Just telling people that they are using a non-standard signature
> I think is enough.
>
Maybe a patch reduced to the very obvious synonyms helps newcomers or
people with lousy memory to be reminded that it is called
"Co-developed-by:" not "Co-authored-by".
Lukas