Thread overview
[dmd-internals] Auto-warn authors followed by auto-close of very stale PRs
May 06, 2016
Andrej Mitrovic
May 07, 2016
Daniel Murphy
May 08, 2016
Walter Bright
May 08, 2016
Leandro Lucarella
May 09, 2016
Leandro Lucarella
May 09, 2016
Andrej Mitrovic
May 07, 2016
It seems that after open pulls page 4 or so on the DMD side of things a pretty large number of PRs are completely broken and haven't been touched in months, or worse in years. E.g. https://github.com/dlang/dmd/pull/4658

There is really no benefit to having PRs just sit idly like that.

I think we could automate the clean-up of stale pulls and we could even do it in stages:

After X amount of time we could make `dlang-bot` attempt to notify the author of his stale PR, then let some more time pass again, and if the PR is still in bad shape `dlang-bot` would close the PR, send a notification to the PR author about the pull being closed, but also CC to dmd-internals / phobos-internals / etc in case the author went AWOL so we have a chance of picking up the work from there.

Thoughts?
_______________________________________________
dmd-internals mailing list
dmd-internals@puremagic.com
http://lists.puremagic.com/mailman/listinfo/dmd-internals
May 07, 2016
What would that gain us?  I eventually come back to my open PRs...

On Sat, May 7, 2016 at 12:11 AM, Andrej Mitrovic via dmd-internals <dmd-internals@puremagic.com> wrote:
> It seems that after open pulls page 4 or so on the DMD side of things a pretty large number of PRs are completely broken and haven't been touched in months, or worse in years. E.g. https://github.com/dlang/dmd/pull/4658
>
> There is really no benefit to having PRs just sit idly like that.
>
> I think we could automate the clean-up of stale pulls and we could even do it in stages:
>
> After X amount of time we could make `dlang-bot` attempt to notify the author of his stale PR, then let some more time pass again, and if the PR is still in bad shape `dlang-bot` would close the PR, send a notification to the PR author about the pull being closed, but also CC to dmd-internals / phobos-internals / etc in case the author went AWOL so we have a chance of picking up the work from there.
>
> Thoughts?
> _______________________________________________
> dmd-internals mailing list
> dmd-internals@puremagic.com
> http://lists.puremagic.com/mailman/listinfo/dmd-internals
_______________________________________________
dmd-internals mailing list
dmd-internals@puremagic.com
http://lists.puremagic.com/mailman/listinfo/dmd-internals
May 08, 2016
I agree with Daniel. Closing PRs just because they are old has no benefit, and sometimes we do go back, fix one, and pull it, or use it as a guide to making a better PR.


On 5/7/2016 9:35 AM, Daniel Murphy via dmd-internals wrote:
> What would that gain us?  I eventually come back to my open PRs...
>
> On Sat, May 7, 2016 at 12:11 AM, Andrej Mitrovic via dmd-internals
> <dmd-internals@puremagic.com> wrote:
>> It seems that after open pulls page 4 or so on the DMD side of things
>> a pretty large number of PRs are completely broken and haven't been
>> touched in months, or worse in years. E.g.
>> https://github.com/dlang/dmd/pull/4658
>>
>> There is really no benefit to having PRs just sit idly like that.
>>
>> I think we could automate the clean-up of stale pulls and we could
>> even do it in stages:
>>
>> After X amount of time we could make `dlang-bot` attempt to notify the
>> author of his stale PR, then let some more time pass again, and if the
>> PR is still in bad shape `dlang-bot` would close the PR, send a
>> notification to the PR author about the pull being closed, but also CC
>> to dmd-internals / phobos-internals / etc in case the author went AWOL
>> so we have a chance of picking up the work from there.
>>
>> Thoughts?
>> _______________________________________________
>> dmd-internals mailing list
>> dmd-internals@puremagic.com
>> http://lists.puremagic.com/mailman/listinfo/dmd-internals
> _______________________________________________
> dmd-internals mailing list
> dmd-internals@puremagic.com
> http://lists.puremagic.com/mailman/listinfo/dmd-internals

_______________________________________________
dmd-internals mailing list
dmd-internals@puremagic.com
http://lists.puremagic.com/mailman/listinfo/dmd-internals
May 09, 2016
If a PR is extremely old and have rotten for ages, it's not really useful, so is just noise in the PR list. I think if you are conservative about the "X amount of time" it might help. You can also always come to old closed PRs (you could use a label to mark PR automatically closed).

Also, if you want a PR to stay alive, you just have to tell the bot that. I guess the bot wouldn't close the PR if it has any activity after the warning...

Walter Bright via dmd-internals, el  8 de May a las 07:55 me escribiste:
> I agree with Daniel. Closing PRs just because they are old has no benefit, and sometimes we do go back, fix one, and pull it, or use it as a guide to making a better PR.
> 
> 
> On 5/7/2016 9:35 AM, Daniel Murphy via dmd-internals wrote:
> >What would that gain us?  I eventually come back to my open PRs...
> >
> >On Sat, May 7, 2016 at 12:11 AM, Andrej Mitrovic via dmd-internals <dmd-internals@puremagic.com> wrote:
> >>It seems that after open pulls page 4 or so on the DMD side of things a pretty large number of PRs are completely broken and haven't been touched in months, or worse in years. E.g. https://github.com/dlang/dmd/pull/4658
> >>
> >>There is really no benefit to having PRs just sit idly like that.
> >>
> >>I think we could automate the clean-up of stale pulls and we could even do it in stages:
> >>
> >>After X amount of time we could make `dlang-bot` attempt to notify the author of his stale PR, then let some more time pass again, and if the PR is still in bad shape `dlang-bot` would close the PR, send a notification to the PR author about the pull being closed, but also CC to dmd-internals / phobos-internals / etc in case the author went AWOL so we have a chance of picking up the work from there.
> >>
> >>Thoughts?
> >>_______________________________________________
> >>dmd-internals mailing list
> >>dmd-internals@puremagic.com
> >>http://lists.puremagic.com/mailman/listinfo/dmd-internals
> >_______________________________________________
> >dmd-internals mailing list
> >dmd-internals@puremagic.com
> >http://lists.puremagic.com/mailman/listinfo/dmd-internals
> 
> _______________________________________________
> dmd-internals mailing list
> dmd-internals@puremagic.com
> http://lists.puremagic.com/mailman/listinfo/dmd-internals

-- 
Leandro Lucarella (AKA luca)                     http://llucax.com.ar/
----------------------------------------------------------------------
They love me like I was a brother
They protect me, listen to me
They dug me my very own garden
Gave me sunshine, made me happy

Nice dream, nice dream
Nice dream
_______________________________________________
dmd-internals mailing list
dmd-internals@puremagic.com
http://lists.puremagic.com/mailman/listinfo/dmd-internals
May 09, 2016
BTW, having old broken PRs consumes resources in the auto-tester, right? Once a PR have rotten it will be never magically fixed, so it doesn't make sense to keep testing it with the auto-tester.

Leandro Lucarella via dmd-internals, el  9 de May a las 00:15 me escribiste:
> If a PR is extremely old and have rotten for ages, it's not really useful, so is just noise in the PR list. I think if you are conservative about the "X amount of time" it might help. You can also always come to old closed PRs (you could use a label to mark PR automatically closed).
> 
> Also, if you want a PR to stay alive, you just have to tell the bot that. I guess the bot wouldn't close the PR if it has any activity after the warning...
> 
> Walter Bright via dmd-internals, el  8 de May a las 07:55 me escribiste:
> > I agree with Daniel. Closing PRs just because they are old has no benefit, and sometimes we do go back, fix one, and pull it, or use it as a guide to making a better PR.
> > 
> > 
> > On 5/7/2016 9:35 AM, Daniel Murphy via dmd-internals wrote:
> > >What would that gain us?  I eventually come back to my open PRs...
> > >
> > >On Sat, May 7, 2016 at 12:11 AM, Andrej Mitrovic via dmd-internals <dmd-internals@puremagic.com> wrote:
> > >>It seems that after open pulls page 4 or so on the DMD side of things a pretty large number of PRs are completely broken and haven't been touched in months, or worse in years. E.g. https://github.com/dlang/dmd/pull/4658
> > >>
> > >>There is really no benefit to having PRs just sit idly like that.
> > >>
> > >>I think we could automate the clean-up of stale pulls and we could even do it in stages:
> > >>
> > >>After X amount of time we could make `dlang-bot` attempt to notify the author of his stale PR, then let some more time pass again, and if the PR is still in bad shape `dlang-bot` would close the PR, send a notification to the PR author about the pull being closed, but also CC to dmd-internals / phobos-internals / etc in case the author went AWOL so we have a chance of picking up the work from there.
> > >>
> > >>Thoughts?
> > >>_______________________________________________
> > >>dmd-internals mailing list
> > >>dmd-internals@puremagic.com
> > >>http://lists.puremagic.com/mailman/listinfo/dmd-internals
> > >_______________________________________________
> > >dmd-internals mailing list
> > >dmd-internals@puremagic.com
> > >http://lists.puremagic.com/mailman/listinfo/dmd-internals
> > 
> > _______________________________________________
> > dmd-internals mailing list
> > dmd-internals@puremagic.com
> > http://lists.puremagic.com/mailman/listinfo/dmd-internals
> 
> -- 
> Leandro Lucarella (AKA luca)                     http://llucax.com.ar/
> ----------------------------------------------------------------------
> They love me like I was a brother
> They protect me, listen to me
> They dug me my very own garden
> Gave me sunshine, made me happy
> 
> Nice dream, nice dream
> Nice dream
> _______________________________________________
> dmd-internals mailing list
> dmd-internals@puremagic.com
> http://lists.puremagic.com/mailman/listinfo/dmd-internals

-- 
Leandro Lucarella (AKA luca)                     http://llucax.com.ar/
----------------------------------------------------------------------
Cada movimiento que no se hace, es un movimiento que se pierde.
Y cada movimiento que se pierde, se transforma en una mochila.
Y las mochilas nos alejan, de nuestros amigos y nuestras amigas.
Y nuestros amigos se transforman, en enemigos y en enemigas.
Cada movimiento que se hace, es una mochila que se deja.
_______________________________________________
dmd-internals mailing list
dmd-internals@puremagic.com
http://lists.puremagic.com/mailman/listinfo/dmd-internals
May 09, 2016
There's also the PR effect (public relations, not pull request). If people see there's hundreds of open PRs they might reach bad conclusions, like "my code will probably never be merged so I better not even try contributing".

On 5/9/16, Leandro Lucarella via dmd-internals <dmd-internals@puremagic.com> wrote:
> BTW, having old broken PRs consumes resources in the auto-tester, right? Once a PR have rotten it will be never magically fixed, so it doesn't make sense to keep testing it with the auto-tester.
>
> Leandro Lucarella via dmd-internals, el  9 de May a las 00:15 me escribiste:
>> If a PR is extremely old and have rotten for ages, it's not really useful, so is just noise in the PR list. I think if you are conservative about the "X amount of time" it might help. You can also always come to old closed PRs (you could use a label to mark PR automatically closed).
>>
>> Also, if you want a PR to stay alive, you just have to tell the bot that. I guess the bot wouldn't close the PR if it has any activity after the warning...
>>
>> Walter Bright via dmd-internals, el  8 de May a las 07:55 me escribiste:
>> > I agree with Daniel. Closing PRs just because they are old has no benefit, and sometimes we do go back, fix one, and pull it, or use it as a guide to making a better PR.
>> >
>> >
>> > On 5/7/2016 9:35 AM, Daniel Murphy via dmd-internals wrote:
>> > >What would that gain us?  I eventually come back to my open PRs...
>> > >
>> > >On Sat, May 7, 2016 at 12:11 AM, Andrej Mitrovic via dmd-internals <dmd-internals@puremagic.com> wrote:
>> > >>It seems that after open pulls page 4 or so on the DMD side of things a pretty large number of PRs are completely broken and haven't been touched in months, or worse in years. E.g. https://github.com/dlang/dmd/pull/4658
>> > >>
>> > >>There is really no benefit to having PRs just sit idly like that.
>> > >>
>> > >>I think we could automate the clean-up of stale pulls and we could even do it in stages:
>> > >>
>> > >>After X amount of time we could make `dlang-bot` attempt to notify the author of his stale PR, then let some more time pass again, and if the PR is still in bad shape `dlang-bot` would close the PR, send a notification to the PR author about the pull being closed, but also CC to dmd-internals / phobos-internals / etc in case the author went AWOL so we have a chance of picking up the work from there.
>> > >>
>> > >>Thoughts?
>> > >>_______________________________________________
>> > >>dmd-internals mailing list
>> > >>dmd-internals@puremagic.com
>> > >>http://lists.puremagic.com/mailman/listinfo/dmd-internals
>> > >_______________________________________________
>> > >dmd-internals mailing list
>> > >dmd-internals@puremagic.com
>> > >http://lists.puremagic.com/mailman/listinfo/dmd-internals
>> >
>> > _______________________________________________
>> > dmd-internals mailing list
>> > dmd-internals@puremagic.com
>> > http://lists.puremagic.com/mailman/listinfo/dmd-internals
>>
>> --
>> Leandro Lucarella (AKA luca)                     http://llucax.com.ar/
>> ----------------------------------------------------------------------
>> They love me like I was a brother
>> They protect me, listen to me
>> They dug me my very own garden
>> Gave me sunshine, made me happy
>>
>> Nice dream, nice dream
>> Nice dream
>> _______________________________________________
>> dmd-internals mailing list
>> dmd-internals@puremagic.com
>> http://lists.puremagic.com/mailman/listinfo/dmd-internals
>
> --
> Leandro Lucarella (AKA luca)                     http://llucax.com.ar/
> ----------------------------------------------------------------------
> Cada movimiento que no se hace, es un movimiento que se pierde.
> Y cada movimiento que se pierde, se transforma en una mochila.
> Y las mochilas nos alejan, de nuestros amigos y nuestras amigas.
> Y nuestros amigos se transforman, en enemigos y en enemigas.
> Cada movimiento que se hace, es una mochila que se deja.
> _______________________________________________
> dmd-internals mailing list
> dmd-internals@puremagic.com
> http://lists.puremagic.com/mailman/listinfo/dmd-internals
_______________________________________________
dmd-internals mailing list
dmd-internals@puremagic.com
http://lists.puremagic.com/mailman/listinfo/dmd-internals
May 09, 2016
> On May 9, 2016, at 9:01 AM, Leandro Lucarella via dmd-internals <dmd-internals@puremagic.com> wrote:
> 
> BTW, having old broken PRs consumes resources in the auto-tester, right? Once a PR have rotten it will be never magically fixed, so it doesn't make sense to keep testing it with the auto-tester.

This consumes nearly zero cycles on the auto tester (if it ever gets there, newer PRs are tested first), because each test fails instantly with not being able to merge.

I think we should leave them open. Note that I think a non-member cannot reopen a closed PR that was closed by a member. At least someone told me once that the PR I closed of theirs they could not reopen.

-Steve
_______________________________________________
dmd-internals mailing list
dmd-internals@puremagic.com
http://lists.puremagic.com/mailman/listinfo/dmd-internals