[Piglit] Require Signed-off-by for patches?

[Aaron Watry]

On Fri, Nov 15, 2013 at 3:42 PM, Eric Anholt <eric at anholt.net> wrote:
> Jordan Justen <jljusten at gmail.com> writes:
>
>> On Thu, Nov 14, 2013 at 3:09 PM, Ian Romanick <idr at freedesktop.org> wrote:
>>> On 11/13/2013 05:12 PM, Matt Turner wrote:
>>>> On Wed, Nov 13, 2013 at 12:06 PM, Jordan Justen <jljusten at gmail.com> wrote:
>>>>> What are the arguments against just following the kernel's
>>>>> Signed-off-by practice?
>>>>
>>>> What are the arguments for it?
>>>
>>> Other than habit, there probably aren't any arguments in-favor for
>>> piglit.  The sorts of things that patch signing is designed protect
>>> against really aren't relevant for piglit.  The probability of someone
>>> distributing piglit (are there any?) being sued because a third party's
>>> IP somehow leaked into the project seems infinitesimal, at best.
>>
>> So, you're saying there is a stronger argument for adding this for
>> Mesa? (I agree. :)
>>
>> I agree to your point that piglit is less likely to have an issue that
>> Signed-off-by can help with. But I also think that once a project has
>> decides to adopt Reviewed-by, etc, then the extra step of
>> Signed-off-by is trivial. At that point it seems there is some (small)
>> benefit gained in consistency of process between open source projects.
>>
>> It also doesn't hurt that git makes Signed-off-by so easy. If I
>> actually had to type it out, then I probably would not think it was
>> worth it for piglit. :)
>>
>> Do we think Signed-off-by may cause people to have reservations about
>> contributing code to piglit?
>
> No, but I know from other projects that I *will* forget signed-off-by
> and get nagged about it.
>
> I think s-o-b is silly.  Whoever incorrectly put their s-o-b on a patch
> will just say "Oh, I had no idea I was agreeing to *that*", since most
> new people I see apply s-o-b at someone's else's request don't know
> about the developer's certificate of origin, or don't read it when
> pointed to it.  It's like a EULA we present to developers, and they just
> click right through.

I'd agree with this.  I guarantee that I will probably forget to attach a s-o-b
at some point.  At that point, we have some ambiguity about the status of
the patch and someone's either going to call me out on it, or the piglit history
will have yet another patch that's missing this piece.  Unless we have some
pre-commit hook enforcing this, things will slip through.

I can understand that we might want a signed-off-by statement when the patch
was originally contributed/written by someone other than oneself, but if I am
submitting and committing my own patches without a 3rd party being
involved, I see
it as an unnecessary hoop to jump through.

That being said, I'm just me, and I'm not a copyright lawyer... so yeah.

--Aaron

>
> That said, I'm not super opposed if other people are excited about it.
>
> _______________________________________________
> Piglit mailing list
> Piglit at lists.freedesktop.org
> http://lists.freedesktop.org/mailman/listinfo/piglit
>