[Mesa-dev] Patchwork review process (efficiency) questions

[Rob Clark]

On Fri, Jun 3, 2016 at 11:49 AM, ⚛ <0xe2.0x9a.0x9b at gmail.com> wrote:
> On Fri, Jun 3, 2016 at 2:08 PM, Nicolai Hähnle <nhaehnle at gmail.com> wrote:
>> On 03.06.2016 13:12, wrote:
>>>
>>> Situation: Looking at the content displayed by the web browser for URL
>>> http://patchwork.freedesktop.org/project/mesa/series and sub-pages
>>> accessible via the links.
>>
>>
>> Patchwork isn't really central to most people's workflow as far as I
>> understand it.
>
> Ok. But how does a developer know (better: gets notified) when a patch
> has been accepted and added to mainline mesa-git? Viewing all git log
> messages every day because - just in case - the patch _might_ have
> been added to mesa-git seems quite inefficient to me.
>
>> Most of your questions' answers naturally fall out from that.
>
> Not in my world.
>
>> This mailing list is what's important.
>
> In my opinion, now isn't the 1990-ties. (This is just my personal
> opinion, you do not need to agree with it.)

I will point out a couple notes/observations:

Kernel (drm/dri-devel), xorg, and other related projects use the same
process, and a lot of us do (or at least at some point have) been
active in 2 or more of these.

Also, I have seen/used some other processes (gerrit, github pulls,
etc).. and IMO on those projects the review process ended up being a
lot more rubber-stamping and less thorough review of the changes.
There is some value in not making things too "push-button"..

Possibly there is some room for improvement, like commit hooks that
send out "your patch has been merged" notifications.  I'm not sure to
what extent people would like vs be annoyed by such a thing.  For me,
it is not too useful since I rebase mesa frequently enough that it is
pretty obvious when patches have/have-not been pushed.  But otoh it is
easy enough to set up a mail filter if I don't want to see those msgs.

>>> - What is the influence of the default ordering (URL suffix
>>> "?ordering=-last_updated") on the behavior of reviewers?
>>
>> Probably zero, because I doubt people go via Patchwork.
>
> In that case, the process might be even less efficient than I though.
>
>>> - What about those patches on the 10th page from previous year? Why
>>> are they in the list?
>>
>> Nothing.
>>
>>> - Do patch submitters regularly clean up outdated patches?
>>
>> No.
>>
>>> - Does a patch submitter receive a notification email when he/she
>>> forgets about a patch over time?
>>
>> No.
>>
>>> It seems to me that the current review process isn't as efficient as it
>>> can be.
>>
>> To be blunt: that may be the case, but even if so, it's extremely unlikely
>> bordering on the impossible that comments from the sideline from somebody
>> who hasn't got an experience of contributing could ever be helpful.
>
> Outsiders and enemies. That is so ... cool.

keep in mind that mesa, like many open source projects is essentially
a meritocracy.. folks with more of a history of contributions[1] get
taken more seriously.  In a meritocracy, you have to earn your
position.  That is not trying to snub an "outsider", just trying to
explain how things work so new folks know how to work with the rest of
the community.

[1] patches, but also useful review comments.. see my earlier comment
about review-bandwidth being the actual bottleneck

BR,
-R