pull requests

Bartłomiej Rutkowski robak at freebsd.org
Wed Mar 1 12:16:56 UTC 2017


On Wed, Mar 1, 2017 at 12:07 PM, Kristof Provost <kp at freebsd.org> wrote:

> On 1 Mar 2017, at 13:00, Bartłomiej Rutkowski wrote:
>
>> Don't get me wrong, I am simply trying to understand what exactly would
>> that accomplish - see https://reviews.freebsd.org/D9649. Lets, for a fun
>> thought experiment, imagine it was an GH pull request that was transferred
>> to Phabric. What now? 2 weeks passed and nothing happened. Would any
>> policy
>> fix that?
>>
>> No, the best policy in the world isn’t going to fix lack of time.
>
> It might help a little if these at least all land on the same pile, but
> it’s
> not going to magic up reviewers.  I’m not sure anything would.
>
> A possibly policy answer would be that reviews are automatically closed
> after a
> while. It’s not great, but at least the submitter knows that nothing is
> going
> to happen. Leaving it open forever also accomplishes nothing.
>

Well, have you seen Bugzilla recently, or ever? :) We've THOUSANDS open
PR's (literally, ~2300 for ports, ~6800 for src, with oldest ones going
back as far as to 2001). Does it make any sense to waste time to write such
policies for Phabric but not for Bugzilla? And if you're about to write one
for both, good luck convincing people that bugs shouldn't just lay around
forever 'just in case'...

In any case, if you'd happen to me successful with that, I'd support it.
While 'just in case' argument makes some sense, I still think that keeping
more PR's open than we can really handle causes cognitive load and
discouragement in taking care of them just by sheer looking at the number
of PR's.

Kind regards,
Bartek Rutkowski


More information about the freebsd-git mailing list