[darcs-devel] Politics and procedure in darcs Was: darcs replace -- punctuation

David Roundy droundy at darcs.net
Sat Mar 3 07:22:20 PST 2007


On Sat, Mar 03, 2007 at 11:57:09AM +0100, Eric Y. Kow wrote:
> Hi Samuel,
> 
> > I've recently had a *VERY* bad experience with submitting a patch to
> > the Firefox folks.  I won't go into details here, but suffice it to
> > say that there was an exorbitant amount of politicking involved, and
> > frankly, I feel like total crap for even bothering to try and fix a
> > bug that is, oh, I dunno, five years old and counting.
> 
> Darcs is relatively free from politics, being a much smaller project
> that Firefox; however, we do suffer from a lack of time.

I'll just add to what Eric said that we *try* at darcs to be very friendly
and encouraging to newcomers.  However, we also try to only accept decent
quality code, which sometimes can be discouraging to people new to
Haskell.  This worked out better when I (and others) had more spare time,
as I really enjoy reviewing code and suggesting cleaner alternatives, which
works well with new developers who are patient and enjoy learning (which
really is the best kind to attract to the project anyhow).  Recently, we've
mostly been pretty swamped (myself in particular), so the quality of
critique has dropped.  Eric is doing great, but only reviews patches once a
week, which can be frustrating to new developers who want to get something
finished quickly, but need help in improving their code.

As Eric said, all patches go to the darcs-devel mailing list, including my
own.  Technically the darcs-unstable maintainer could apply their own
patches to the central repository directly (and in some cases they do, and
I used to do that all the time), but tradition (begun by igloo) dictates
that they first send them to the list like anyone else would do, and wait
for replies.  And I feel especially obligated to review those patches, as
it's pretty close to impossible to effectively review one's own code.  So
your contribution will go through the same process anyone else's will,
which helps a lot to reduce politicking.

For new features (or formats), it's best to propose them early, possibly
before writing code, but certainly before spending much time, to see if the
feature is deemed acceptable.  We're very hesitant to add new commands, and
pretty hesitant to add new flags, just to keep darcs easy to learn and to
use.  New formats (e.g. a new darcs replace format) will need discussion,
to be sure we won't need to change it again soon, and careful review to
make sure the transition is smooth.  Among other things, a new format
requires learning and using the RepoFormat mechanism (which allows
backwards and forwards compatibility, and avoids corruption).
Unfortunately, this mechanism, while successfully in place and believed
correct, hasn't been widely used, just because we haven't introduced many
new formats since _darcs/format has been introduced.

> > My question, therefore, is related to, "How *painful* will it
> > be for me to contribute back to darcs?"
> 
> I expect it to be relatively smooth sailing.  Do check the activity
> of the darcs-devel list to get an idea.  You might also want to see
> when patches get Rejected, or when I ask for resubmission.

I'll just add that I attribute much of darcs' success to the lack of pain
associated with contributing back--some of which is because we use darcs,
but a lot of it is just because we're a friendly crowd (which also attracts
friendly developers and motivates them to be friendly).
-- 
David Roundy
http://www.darcs.net
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: Digital signature
Url : http://lists.osuosl.org/pipermail/darcs-devel/attachments/20070303/5bc34b51/attachment-0001.pgp


More information about the darcs-devel mailing list