Hi,
thanks for the clarification.
Then I hope that you’ll switch nightly to aeson-0.11 before switching
to GHC-8. This will allow us to build on such a nightly snaptop. Not as
nice as an LTS release, but it serves the purpose.
Greetings,
Joachim
Am Donnerstag, den 17.03.2016, 09:44 -0700 schrieb Dan Burton:
> This is not the case.
>
> LTS 5 interrupted LTS 4 due to aeson-0.10 bugs, so it was LTS 4 that
> was not a full cycle. LTS 5 was intended to be a full cycle LTS
> release. We want every LTS release to be full cycle, and the only
> reason LTS 4 was not full cycle was due to bugs in aeson-0.10 that
> caused `stack install pandoc` to fail. We waited 4 months between the
> release of aeson-0.10 and including it in an LTS, and even then we
> discovered it was a mistake, so we're trying to be a little more
> cautious about gaining unanimous community support before upgrading.
>
> I, for one, anticipated that it would take some time for aeson-0.11
> to be released, and some more time after that for community adoption.
> I was wrong about the first part: aeson-0.11 came out pretty quickly
> once the spotlight was shining on aeson-0.10's bugs. As for the
> second part, we're still not quite to the point where nightly builds
> have switched over to aeson-0.11. Mainly, we're waiting on pandoc.
> Community adoption moved a lot faster compared to aeson-0.10, which
> means this version is probably good.
>
> -- Dan Burton
>
> On Thu, Mar 17, 2016 at 12:31 AM, Joachim Breitner <mail@joachim-brei
>
tner.de> wrote:
> > Hi,
> >
> > as I recall, when LTS 5 was released, it was said that its not
> > going to be a full cycle, but rather an intermediate thing to fix
> > the aeson issues, to be superseded as soon as the aeson bugs were
> > fixed and the dependencies were updated. So this is not the case?
> >
> > Greetings, Joachim
> >
> > Am 17. März 2016 07:29:08 MEZ, schrieb Dan Burton <danburton.email@
> >
gmail.com>: