On Jun 19, 2012 1:35 PM, "Mark Lentczner" <m...@glyphic.com> wrote:
> tl;dr:
> Squeezing a release out in the next six weeks would accomplish little.
A new early (beta?) release based on ghc-7.4.2 would be a good thing IMHO.
> Our mid-November 2012.4.0.0 actually looks pretty good given the lead
time that the distros need for their first 2013 release.
>
> On Fri, Jun 15, 2012 at 7:13 PM, Jens Petersen <juhpeter...@gmail.com>
wrote:
>>
>> How about a tight schedule for 2012.4 something like this (now that we
have the luxury of the ghc-7.4.2 release in our hands:)?:
>>
>> Alpha (2012.3.80.x): June 30
>> Beta (2012.3.90.x): July 15
>> Final (2012.4.0.x): August 1
Ok forget my crazy provocative schedule (on further reflection not even
sure an alpha would really be necessary this time).
> I don't see how we could make this schedule, or how that schedule could
result in anything other than 2012.2.0.0 but based on GHC 7.4.2. That is,
other than the version number bump, an 2012.4.0.0 based on such a release
schedule would be really no different for what ends up in the next Fedora
or Debian or Ubuntu release than if they included 2012.2.0.0 but bumped to
7.4.2 themselves. Since all of those distros already package up GHC
independently of the rest of the HP, I don't see an issue with them
releasing it such a form.
Sure we could/can, but the whole point of this discussion was to avoid that
so that it would be easier for Linux distros to follow Haskell Platform.
>
> Judging from the release schedule for Ubunutu 12.04 - the alpha for an
Ubuntu April release starts in the prior December. This would mean our
November release is well timed to that schedule. Given that this is a
normal time frame for distributions, they are always going to be shipping a
GHC and packages that are 5 to 6 months old. Since they, and we, are aiming
at stability, I don't see this as a big problem.
>
> We could move our release dates later, so as to ship a more current HP to
the distros... if the distros are willing to enter alpha while we are still
in beta. Personally, I'm not a fan of this, as I worry there would be no
wiggle room.
I am not sure on the best solution but I thought we had general consensus
earlier that the schedule could and should be improved for Linux distros.
I don't see why the ("baby") distros need to be "thrown out" together with
("the bath water") my over eager-schedule.
_______________________________________________
Haskell-platform mailing list
Haskell-platform@projects.haskell.org
http://projects.haskell.org/cgi-bin/mailman/listinfo/haskell-platform
