New Github Issues

[H. S. Teoh via Digitalmars-d]

On Mon, Jul 28, 2014 at 05:31:55PM +0000, Brad Anderson via Digitalmars-d wrote:
> Github updated their Issues system (which includes Pull Requests). You
> can read about it here:
> 
> https://github.com/blog/1866-the-new-github-issues
> 
> Bugzilla is here to stay but the newly added Labels feature could
> probably help organize Pull Requests. The usual "Enhancement", "Bug",
> etc. are nice but I'm wondering if anyone can think of a way they
> could be used to help deal with the pull request backlog.
[...]

We could adopt what we did with Phobos PRs over the last 2-3 weeks. :-)

After a rant about the PR backlog in one of those interminable rant
threads, a few of us decided to stop arguing about it and instead do
something about it: ping PRs that haven't been updated for a long time
(changing the sorting to 'least recently updated' or 'oldest' helps find
these stagnating PRs more easily), review PRs that have been sitting
around with no review comments, suggest PRs be closed if there's no hope
they will be merged, etc..  A few committers got a bit more aggressive
about pulling PRs -- with the view that if something was a mistake, we
could always revert it later, since after all that's what git is for!

The result was that the original Phobos PR backlog of about 90+ or so
dropped to around 70 by the end of the week, and this morning we broke
the 60 mark, and we're now down to 56. The PR list page used to be 4
pages, and now it's down to 3 and approaching 2. If this trend will
continue, we should soon be able to get things down to about 1 page or
so, and the situation will be much more controllable.

There was at least 1 revert that I know of from this recent effort, and
1 fixup PR to repair some flaws in a previous PR. But I see that as a
good thing: we're getting things moving, getting the feedback and
discussion going, and making progress in general, instead of waiting
around for PRs to become ideal flawless ivory towers that also never
materialize.

In contrast, the DMD PR list has breached 3 digits, and shows little
sign of slowing down. Perhaps we should consider adopting a similar
approach there as well? ;-)

On that note, though, the Phobos situation, while much improved, still
isn't quite there yet. I'd like to invite everyone here to chime in and
review PRs. As long as you can code in D, you're qualified to review
*something* -- not every PR involves rocket science. Sometimes even
trivial nitpick comments like minor Phobos style violations may help
awaken a dormant PR and get things moving again. You don't need commit
rights to contribute in this way -- I don't, for example -- but your
work will be much appreciated. And I have to say that it is very
gratifying to see the Phobos open PR count drop, and realize I was a
part of it.


T

-- 
The best compiler is between your ears. -- Michael Abrash