Hey, this is fixed upstream but the upcoming release isn't ready yet (see https://github.com/kazu-yamamoto/ghc-mod/milestones/Next%20release for a list of remaining issues). As soon as all of those are fixed I'll upload a new version to hackage and update the package in Debian as well. In the mean time you'll have to compile from source and hope nothing breaks too badly. I've mostly gotten positive responses so far but if anything does break just let us know. @Joachim can we upload the not quite ready yet version to experimental so people don't have to compile from source? --Daniel On Sat, Jun 27, 2015 at 11:47:22PM +0100, Andreas Reuleaux wrote: > I have been on this list for a quite a while, but quiet, > as I am really just a debian haskell user, not a debian developer > (ie. I develop Haskell code, but have never created any debian > packages so far). That being said, I am quite happy with > Haskell on Debian, and thanks a lot for all your great work. > > I try to rely on debian haskell packages as much a I can, > (and to resort to hackage only for the things that have not > been debianized) for I find them well maintained, > like to get occasional updates automatically, and don't want > to be bothered by installing everything by hand (from hackage). > > This is especially true, for my infrastructure / tool chain: > I have happily always used eg. ghc from debian, as well as > cabal-install, ghc-mod > > I am on stretch, and have thus recently received tons of updated > packages together with ghc 7.8. > > While the transition from ghc 7.6 to 7.8 has been smooth for me, > if I am not mistaken: shortly after this main transition > I got a newer cabal-install as well, now 1.22.0.1. > > If I am using ghc-mod currently (on any haskell file in a small project > with a cabal file etc.): > > $ ghc-mod check SigmaTypes.hs > /home/rx/etc/pire/dist/setup-config: hGetContents: invalid argument (invalid byte sequence) > > I am pretty sure that my cabal file is fine, and if I remember right, > there was a short period recently, where I had (from stretch): > > ghc 7.8 already > ghc-mod > but cabal-install still at 1.20 > > where I could rebuild my project, and ghc-mod would not complain. > > According to > https://github.com/kazu-yamamoto/ghc-mod/issues/417 > and > https://github.com/kazu-yamamoto/ghc-mod/wiki/InconsistentCabalVersions > > ghc-mod and cabal-install 1.22 don't play well together. > > But maybe it's just me experiencing these problems? Can anyone confirm > these issues? > * in any haskell / cabal project of yours: > * remove and recreate the sandbox > * rebuild, ie. cabal install things from hackage, that are not in > debian > * and see if ghc-mod check complains on any haskell file > > If it's only me, then I could of course give a more precise > details, paste some code etc > > There are some workaround suggested on > https://github.com/kazu-yamamoto/ghc-mod/wiki/InconsistentCabalVersions > like > cabal install --constraint "Cabal < 1.22" cabal-install > I assume in my sandbox, not system wide (?), which I would rather > want to avoid if possible. > > Anyway, as I expect ghc 7.8 to stay in stretch for a while, it would > be nice if the three: ghc, cabal-install, ghc-mod would play together > nicely. > > > Thanks, > Andreas > > > > > -- > To UNSUBSCRIBE, email to debian-haskell-request@lists.debian.org > with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org > Archive: [🔎] 87a8vkydj9.fsf@a-rx.info">https://lists.debian.org/[🔎] 87a8vkydj9.fsf@a-rx.info > >
Attachment:
signature.asc
Description: Digital signature