[development] Fwd: [drupal:unconed] /modules/taxonomy taxonomy.module

Karthik narakasura at gmail.com
Tue Apr 17 08:35:59 UTC 2007


On 14/04/07, Steven Wittens <steven at acko.net> wrote:
> This was a *community reviewed* patch that got committed only a week
> before, and which had some bad last minute changes in it that no-one
> noticed when it was committed. The review process exists not to
> satisfy your desire for bureaucracy, but to make sure the community

My desire for bureaucracy? hah! Please don't confuse bureaucracy with
process or work-flow. The results are plainly evident in the issue in
question, where Dries had no idea that the commit had already been
made and wasted X minutes of his life, and more importantly, review
and commit time. The question of patch triviality is of no relevance
here, considering that an issue exists.

As stated earlier, this is not the first commit that this has happened
to recently. Off hand, the last commit to tableheader.js also lacked
an nid or credit, and that was not a readily apparent patch.

> In the case of really trivial bugfixes (like this) or typo
> corrections, there is absolutely no reason to object to the change,
> and thus no reason to waste everyone's time with an issue and review.

If it is trivial, then, IMO, there is no need for a review. However,
please create an issue and mark it fixed directly and include the nid
in your commit message. What if Drumm or Killes want to back-port a
trivial patch? Or if what you thought was trivial ... wasn't? As
you've so eloquently put, processes involving people are inherently
flawed. You are currently leaving no avenue for any community feedback
altogether.

Moreover, this is CVS. If I check my CVS log six months from now, I
will have no idea whether you committed one file or ten thousand. I
will have to chase up the commit via the commit message to find out.
Instead, please spend the 15-30 seconds required to create an issue
and mark it fixed.

And on a related note, I would also like to request that you please
stick to the convention adopted by all other committers and include
attribution in your commit messages. I am well aware of your views on
this matter. However, whether it is right or wrong is besides the
point and is something that you guys can iron out between yourselves
and / or with the community. In the meantime, it will be appreciated
if all of you are consistent. I would also like to point out that your
preference to avoid attribution in commit messages is decidedly at
odds with your theme .info patch wherein you were / are pushing for an
attribution field on the theme page.

Please also understand that I'm not having a go at you personally; I
understand that a committer's job is difficult and generally
thankless. All I'm trying to establish here is consistent behaviour
that will save time and energy for all parties concerned by generally
reducing confusion all around.

Thanks,
-K


More information about the development mailing list