GHC release timing and future build infrastructure

Herbert Valerio Riedel hvriedel at
Thu Aug 3 08:38:38 UTC 2017


On 2017-08-03 at 08:41:59 +0200, Ara Adkins wrote:


> I nevertheless see `stack` as a huge boon for easing adoption of new
> compiler versions (and hence new language features/extensions).

Since I totally disagree about Stack being beneficial to the quick
adoption of new features, would you care to elaborate what makes you
think that Stack was a "huge boon" in this particular context?

Just to name one example where Stack has been lagging behind Cabal for
several months already: Support for Backpack or convenience libraries --
there's already a handful of packages on Hackage which Stack (and
consequently Stackage) cannot install due to this. Being slow to adopt
new features IMO makes totally sense for Stack, as it's target audience
is practicioners who value stability, "reproducibility" and "long term
support", so it's actually a good thing that Stack stays behind until we
iron out issues with new bleeding edge features and Stack(age) adopts
them only when they're officially released and deemed ready for Stack's
target audience.

-- hvr

More information about the ghc-devs mailing list