Despite the minor uptick in heat, I'm fairly happy with how this thread
is working out:

 - We've have a better idea how we should be using hlint within
   the test suite (to watch for errors only), which doesn't preclude
   using hlint as a style checker on the side)
 
 - We've identified some shortcomings with hlint (lack of cpp support)
   which Neil has very promptly fixed (thanks!)

 - And now we've got a chance to smooth out our policy on incoming
   patches
 
Perhaps this is chance for us to update
http://wiki.darcs.net/DeveloperGettingStarted so that folks know better
what to expect and it all goes smoother next time.  It's also a good
thing to do now that there's more of us committers so that we have a
more or less uniform idea of what our contract is with people who submit
patches.

On Fri, Aug 07, 2009 at 10:21:07 +0200, Petr Rockai wrote:
> > Is there any reason that this couldn't've been based on my patches?

> As for based on your patches
> -- I started that way, but then I realized that would create a number of
> versions that cannot pass the tests, without enough justification -- if the
> change is a single patch, both before and after the tests pass just fine, not
> disrupting trackdown. I would have done most of the changes I did anyway -- I
> just squashed them, together with your original patches, into a single patch.

As I understand it, Petr's patch adds new stuff (ratification) and
squashes old stuff together.  We agree about the new stuff, but we
disagree about the squashing.  Petr *did* make an initial attempt to
base the new stuff on the old stuff but then decided it would be cleaner
to just smoosh it all down.

Fair enough. I have no opinion on the smooshing, but I'd like to suggest
a small guideline for committers:

  Treat flaws in incoming patches as teachable moments.

I know that sounds cringe-worthy and American (with apologies to my
compatriots) but I hope that I am making you all groan for the greater
good.

If an incoming patch could be improved, I think we should resist the
temptation to make the improvement ourselves.  Instead, we should
communicate about what we'd like to see.  Now this may require more
effort and noise in the short term, lengthy threads about seemingly
trivial matters, but in the long term it will shape future patches from
the same submitter and also from the lurkers on this list (hi!).  And
that I think will make things more efficient, make it possible to accept
more patches with less unnecessary discussion.

Thanks!

-- 
Eric Kow <http://www.nltg.brighton.ac.uk/home/Eric.Kow>
PGP Key ID: 08AC04F9

Attachment: pgp6rFP9eBW3g.pgp
Description: PGP signature

_______________________________________________
darcs-users mailing list
[email protected]
http://lists.osuosl.org/mailman/listinfo/darcs-users

Reply via email to