Closed glasserc closed 9 years ago
ig has somewhat strict constraints on what it will compile against. In particular:
ig
attoparsec
conduit
1.0
1.1
.Internal
monad-control
This branch loosens those constraints by deleting the direct dependency on attoparsec (leaving our direct dependencies to deal with it), widening the constraint on conduit in the 1.1 series case, and adding support for the new monad-control. This is based on https://github.com/kazu-yamamoto/ghc-mod/pull/423/ and https://hackage.haskell.org/package/monad-control-0.3.3.1/docs/Control-Monad-Trans-Control.html#g:2, and they seem to compile, although I can't claim that I understand them thoroughly enough to promise that they work.
Awesome, thanks! Released as ig-0.2.2.
ig-0.2.2
ig
has somewhat strict constraints on what it will compile against. In particular:attoparsec
, which is not a direct dependency ofig
, has to be less than 0.12.conduit
can be either1.0
series or1.1
series. It seems from the changelog at https://github.com/snoyberg/conduit/commit/e03ee0e24212f358dfa376283125618e725c0156 that no breaking changes are expected unless you import.Internal
.monad-control
> 1.0 breaks everything.This branch loosens those constraints by deleting the direct dependency on
attoparsec
(leaving our direct dependencies to deal with it), widening the constraint onconduit
in the 1.1 series case, and adding support for the newmonad-control
. This is based on https://github.com/kazu-yamamoto/ghc-mod/pull/423/ and https://hackage.haskell.org/package/monad-control-0.3.3.1/docs/Control-Monad-Trans-Control.html#g:2, and they seem to compile, although I can't claim that I understand them thoroughly enough to promise that they work.