- Glasgow-haskell-users

7.8.1, template haskell, and dynamic libraries
by harry 19 Feb '14

19 Feb '14

The docs for 7.8.1 say "Template Haskell must now load dynamic object files, not static ones". Does this mean that, if I'm using Template Haskell, every package which the templates depend on have to be built with --enable-shared? -- View this message in context: http://haskell.1045720.n5.nabble.com/7-8-1-template-haskell-and-dynamic-lib… Sent from the Haskell - Glasgow-haskell-users mailing list archive at Nabble.com.

10 15

State of -XImpredicativeTypes
by Merijn Verstraaten 19 Feb '14

19 Feb '14

Lectori salutem, What is the actual state of ImpredicativeTypes? It appears documented as a "properly" finished GHC extension, but on IRC and other places I keep hearing it's poorly tested, buggy or incomplete. Is this true or just FUD? Cheers, Merijn

2 1

Re: ghc-7.8.1 and stm
by Joachim Breitner 19 Feb '14

19 Feb '14

Hi David, Am Dienstag, den 18.02.2014, 14:12 -0800 schrieb David Fox: > It seems to me that the stm library that is supposed to be built into > ghc-7.8.1 is missing. The deb provides and conflicts with > libghc-stm-dev , but does not provide libghc-stm-dev-2.4.2.1-abcde. this seems to be the result of a possible inadvertent change in the GHC source tarball creation: stm had been a "extra" lib already in 7.6 and before, but was not part of the tarball. Herbert or Austin, is the sdist code not doing the right thing here? Greetings, Joachim -- Joachim "nomeata" Breitner Debian Developer nomeata(a)debian.org | ICQ# 74513189 | GPG-Keyid: 4743206C JID: nomeata(a)joachim-breitner.de | http://people.debian.org/~nomeata

4 3

Re: 7.8.1-candidate fail
by Jan Stolarek 17 Feb '14

17 Feb '14

Thanks. I upgraded Squeeze to Wheezy recently just because of that issue, so it's kinda solved for me. Janek Dnia piątek, 14 lutego 2014, Tyler Huffman napisał: > I added a comment in the ticket, but it looks like Debian Squeeze is using > libgmp3-dev 2:4.3.2, which was the same version that I had when I was > running into this issue. > > Using an updated version of libgmp3-dev will fix the issue, but I'm not > entirely sure if we need to consider supporting the packages in Debian > Squeeze repository since Debian Wheezy is the new stable distribution, and > it uses libgmp-dev 2:5.0.5. Someone with more intimate knowledge of GHC > development might be able to speak on this issue. > > > Regards, > Tyler Huffman > > On Thu, Feb 6, 2014 at 1:50 PM, Jan Stolarek <jan.stolarek(a)p.lodz.pl> wrote: > > I had the same problem on Debian Squeeze: > > > > https://ghc.haskell.org/trac/ghc/ticket/8666 > > > > What is your distro? > > > > CCing ghc-devs. > > > > Janek > > > > Dnia czwartek, 6 lutego 2014, Sergei Meshveliani napisał: > > > Dear GHC team, > > > > > > I am trying to test ghc-7.8.20140130-src.tar.bz2 > > > > > > I make it from source with ghc-7.6.3 on Debian Linux (64 bit). > > > > > > ./configure looks all right. > > > > > > And `make' reports after 40 minutes: > > > > > > ------------------------------------------------------- > > > ... > > > ... > > > "inplace/bin/ghc-stage1" -optc-Ilibraries/integer-gmp/. > > > -optc-I'/home/mechvel/g..... > > > ... > > > ... > > > /usr/bin/ld: libraries/integer-gmp/gmp/objs/aors.o: relocation > > > R_X86_64_32 against `__gmpz_sub' can not be used when making a shared > > > object; recompile with -fPIC > > > libraries/integer-gmp/gmp/objs/aors.o: could not read symbols: Bad > > > value collect2: ld returned 1 exit status > > > make[1]: *** > > > [libraries/integer-gmp/dist-install/build/libHSinteger-gmp-0.5.1.0\ > > > -ghc7.8.20140130.so] Error 1 > > > ------------------------------------------------------- > > > > > > > > > What might this mean? Need I to install a fresher libgmp ? > > > > > > Thanks, > > > > > > ------ > > > Sergei > > > > > > _______________________________________________ > > > Glasgow-haskell-users mailing list > > > Glasgow-haskell-users(a)haskell.org > > > http://www.haskell.org/mailman/listinfo/glasgow-haskell-users > > > > _______________________________________________ > > ghc-devs mailing list > > ghc-devs(a)haskell.org > > http://www.haskell.org/mailman/listinfo/ghc-devs

2 1

Functional dependencies can "return" kinds, type families cannot
by José Pedro Magalhães 13 Feb '14

13 Feb '14

Hello, Maybe this is well known already (or maybe it's a bug), but lately I've again found that functional dependencies are more versatile than type families. In particular, they can be used to compute kinds from types, whereas type families cannot. Consider the code below, which implements a class for generic (working for types other than lists) map: {-# LANGUAGE DataKinds #-} {-# LANGUAGE PolyKinds #-} {-# LANGUAGE TypeFamilies #-} {-# LANGUAGE GADTs #-} {-# LANGUAGE FlexibleInstances #-} {-# LANGUAGE ScopedTypeVariables #-} {-# LANGUAGE TypeOperators #-} {-# LANGUAGE MultiParamTypeClasses #-} {-# LANGUAGE FunctionalDependencies #-} module MapN where -- (Type-level) Peano naturals data Nat = Ze | Su Nat -- Proxy data Proxy t = Proxy -- Singleton type for promoted lists data SList (l :: [*]) where SNil :: SList '[] SCons :: h -> SList t -> SList (h ': t) -- Apply a type to a list of variables type family AppVars (t :: k) (vs :: [*]) :: * type instance AppVars t '[] = t type instance AppVars t (v ': vs) = AppVars (t v) vs -- Domains of a list of functions type family Doms (funs :: [*]) :: [*] type instance Doms '[] = '[] type instance Doms ((a -> b) ': ts) = a ': Doms ts -- Codomains of a list of functions type family CoDoms (funs :: [*]) :: [*] type instance CoDoms '[] = '[] type instance CoDoms ((a -> b) ': ts) = b ': CoDoms ts -- Generalised map -- How can I get rid of the FD below? I know how to do so by using a proxy -- argument of type |Proxy t|, but I don't want an extra argument. The usual -- trick of replacing the FD by a TF doesn't work here because the TF would have -- to return a *kind*, not a type. class GMap (t :: k) (fs :: [*]) | fs -> k where gmap :: SList fs -> AppVars t (Doms fs) -> AppVars t (CoDoms fs) -- List instance instance GMap [] '[a -> b] where gmap _ [] = [] -- The line below does not typecheck without the above FD gmap fs@(SCons f SNil) (h:t) = f h : (gmap fs t) The most interesting part here is the functional dependency fs -> k, where kis a kind variable! If this is not a bug (and it does seem to work as I expect it to), then could we have type families return kinds too?... Cheers, Pedro

2 1

target audience for the binary distribution
by harry 11 Feb '14

11 Feb '14

I'm unable to install the binary distribution of 7.8.1 on RHEL because it requires libgmp.so.10 and GLIBC_2.15, which are much greater than the version available on Red Hat. (I'm on a shared system without root access, so upgrading libc is out of the question, even if it could be done.) If this is what most users would like, I certainly don't want to be the Luddite trying to hold them back. Who actually are "most users" for the bindist? Debian & derivatives have the latest GHC in the package repository, so it's presumably Red Hat & Co, and people without root. If the bindist is only relevant to rootless users on a modern Debian derived distro, is it targeting the right audience? -- View this message in context: http://haskell.1045720.n5.nabble.com/target-audience-for-the-binary-distrib… Sent from the Haskell - Glasgow-haskell-users mailing list archive at Nabble.com.

5 5

RHEL/EPEL 5 ghc packages
by Jens Petersen 10 Feb '14

10 Feb '14

Hi, I wanted to mention some newer ghc builds I have made for RHEL5. EPEL5 currently has ghc-6.12.3 in stable, but I have built ghc-7.0.4 which has been in EPEL5 testing now for over a month. The update also includes cabal-install. I am planning to push it to stable this month but I wanted to mention it first here as a heads-up. https://admin.fedoraproject.org/updates/FEDORA-EPEL-2013-12484/ghc-7.0.4-45… Please report any problems with this in bugzilla: https://bugzilla.redhat.com/enter_bug.cgi?product=Fedora%20EPEL&component=g… Jens ps Additionally I have a new ghc-7.4.2 repo for EPEL5 (which should also work on RHEL6) and includes haskell-platform: http://repos.fedorapeople.org/repos/petersen/ghc-7.4.2/ which is a backport from Fedora 19. Please report any problems with this directly to me. pps Later this year I plan to backport ghc-7.4.2 officially to EPEL6.

2 2

7.8.1-candidate fail
by Sergei Meshveliani 06 Feb '14

06 Feb '14

Dear GHC team, I am trying to test ghc-7.8.20140130-src.tar.bz2 I make it from source with ghc-7.6.3 on Debian Linux (64 bit). ./configure looks all right. And `make' reports after 40 minutes: ------------------------------------------------------- ... ... "inplace/bin/ghc-stage1" -optc-Ilibraries/integer-gmp/. -optc-I'/home/mechvel/g..... ... ... /usr/bin/ld: libraries/integer-gmp/gmp/objs/aors.o: relocation R_X86_64_32 against `__gmpz_sub' can not be used when making a shared object; recompile with -fPIC libraries/integer-gmp/gmp/objs/aors.o: could not read symbols: Bad value collect2: ld returned 1 exit status make[1]: *** [libraries/integer-gmp/dist-install/build/libHSinteger-gmp-0.5.1.0\ -ghc7.8.20140130.so] Error 1 ------------------------------------------------------- What might this mean? Need I to install a fresher libgmp ? Thanks, ------ Sergei

2 1

GHC iOS 7.8.1 RC1 for Device & Simulator
by Luke Iannini 06 Feb '14

06 Feb '14

Hi folks, RC builds of GHC iOS are now ready for devices and the iOS simulator. https://github.com/ghc-ios/ghc-ios-scripts/releases/download/7.8-rc1-device… https://github.com/ghc-ios/ghc-ios-scripts/releases/download/7.8-rc1-simula… You'll find complete instructions on their usage here, including compiling cabal packages: https://github.com/ghc-ios/ghc-ios-scripts/blob/master/README.md Please let me know how it goes! Cheers Luke

1 0

empty stable and current bindist directories
by harry 06 Feb '14

06 Feb '14

http://www.haskell.org/ghc/dist/current/dist/ and http://www.haskell.org/ghc/dist/stable/dist/ are empty. Is this correct? -- View this message in context: http://haskell.1045720.n5.nabble.com/empty-stable-and-current-bindist-direc… Sent from the Haskell - Glasgow-haskell-users mailing list archive at Nabble.com.

1 0