> On a side note, the Windows Installer framework has a version naming scheme > similar to hackage's, that requires that a version be a sequence of > decreasingly significant numbers. So even if we would make hackage accept > Debian-like ~alpha versions, we'd have to hack something up for the Windows > installers that we plan to make for darcs 2.5 and beyond.
First: I think everybody can agree on the principle that we should stick
to numbers, not because we like them but because it's the path of least
resistance
> > In that light, it does seem that this proposal would reduce to one version
> > name per source version. I'm okay with that then. Or at least, willing to
> > see if it works better than the current system.
>
> But then, we don't have to change the numbering policy to have only one name.
> We can also just abolish the alpa, beta and release candidate names without
> changing the numbering scheme. Is there anything in favor of actually
> changing
> the numbering scheme?
That's Jason's and Ben's original request. But I suspect we use those
names because they are really compelling. I think "Darcs 2.9 beta 1" is
clearer that "Darcs 2.8.99.1" because the later looks superficially like
part of the Darcs 2.8 line.
Now suppose Darcs 2.6 was the last release to go out. We want to
prepare its successor. How would we communicate things in both schemes?
============================= ======= ======== =================
task current odd/even comments
============================= ======= ======== =================
Name previous release 2.6 2.8
Name work in HEAD 2.6 2.7 [1]_
(before cutting the branch)
Name next release 2.7 2.8
Indicate that we have a 2.6.97 2.7.97 equivalent
soon-to-be-released version
Name release branch darcs 2.6.x 2.7.y odd/even is more
(making distinction against distinct
current release)
Name ongoing work in HEAD 2.7 2.9
after release branch has been
cut
============================== ======= ======== ================
.. [1] Under the current scheme, we could also say "2.7" in HEAD
but that would be a lie, because it's not 2.7 yet
I think the odd-even scheme makes certain distinctions come out a little
more clearly, but it's your call!
--
Eric Kow <http://www.nltg.brighton.ac.uk/home/Eric.Kow>
PGP Key ID: 08AC04F9
pgpieTX5LfraH.pgp
Description: PGP signature
_______________________________________________ darcs-users mailing list [email protected] http://lists.osuosl.org/mailman/listinfo/darcs-users
