[development] What about reviewing patches?

Nathaniel Catchpole catch56 at googlemail.com
Thu Aug 14 18:00:42 UTC 2008

On Thu, Aug 14, 2008 at 6:29 PM, Angela Byron  wrote:
> True. But exactly *one* person can do something about the RTBC queue. 2,000+
> people can do something about the CNR queue.

I think the main issue /at the moment/ is that this is probably the
longest in several releases that there's only been one person able to
do anything about the RTBC queue. Drupal 5 was released on January
15th 2007 [1], Gabor was given commit access on April 17, 2007[2],
three months later. Drupal 6 was released on February 13, 2008 [3],
it's now 6 months later and there's no sign of a branch maintainer
yet. Additionally, the last core commit by Steven Wittens was July 26,
2007 [4], three weeks after code freeze [5] - so there's actually two
less people committing to HEAD than during the Drupal 6 code thaw, and
Gabor was employed full time working by Acquia just to work on core
commits from August [6] as well.

However, even with three core committers, or two and one person
employed full time, reviews have always been the primary bottleneck.

> Let's have a community outcry about the RTBC queue once the situation is
> reversed, and we have 325 patches waiting for core committers' blessing, and
> only 40 patches that need community review. Until then, we have work to do.
> -Angie

It's pretty easy to make a difference to the RTBC queue as well, at
least when it's as long as it is now, I just went through everything
that had been untouched for 5 weeks or more, bumped everything that
still applied, found 2-3 that needed re-rolling, and one that was in
entirely the wrong queue. Quite a few patches left would probably get
in easier if they had simpletests written for them too - so people
could find those and help them out if simpletests are missing. For
now, RTBC patches against D7 are down to ~ 30 from over twice that a
few days ago (there's been some commits as well recently, although 30
is still quite a lot considering many of them are heavily tested and
nearly all should be valid against HEAD). We'll see if it goes up or
down from here.

Like you said, a review queue of about 40 would be pretty good - as
long as the RTBC queue goes down by the same proportion (about 5-15).
Would be much, much less work nursing patches once the initial pain
was over with.


[1] http://drupal.org/node/109494
[2] http://buytaert.net/gabor-hojtsy
[3] http://drupal.org/node/221219
[4] http://drupal.org/cvs?commit=74849
[5] http://drupal.org/drupal-6.0-code-freeze
[6] http://hojtsy.hu/blog/2008-feb-15/drupal-6-maintainers-perspective

More information about the development mailing list