May 2014 - jhc - Haskell.org

The Wonders of generic applicative traversal routines
by John Meacham 13 May '14

13 May '14

Before http://repetae.net/drop/TypeSyns_old.hs and after http://repetae.net/drop/TypeSyns_new.hs that's right, the work of a 550 line complicated file done in a few lines. The magic is in FrontEnd.Syn.Traverse where I have a class that recurses over arbitrary source syntax and takes a data HsOps m = HsOps { opHsDecl :: HsDecl -> m HsDecl, opHsExp :: HsExp -> m HsExp, opHsPat :: HsPat -> m HsPat, opHsType :: HsType -> m HsType, opHsStmt :: HsStmt -> m HsStmt } as an argument, by recursively defining your ops by 'tying the knot' each of the routines will recurse down the others. quite handy. Now I have about 2500 lines of code to excise from FrontEnd/. Many improvements to the front end have been put off due to the sheer amount of traversal code that has to modified for every new syntax construct. good times. John -- John Meacham - http://notanumber.net/

2 8

Re: [jhc] darcs patch: src/Support/Inst.hs: don't define instances present in...
by John Meacham 13 May '14

13 May '14

Cool. thanks. I was actually wondering why those instances didn't exist as they seemed pretty obvious and I almost made them conditional. I'll modify your patch to be an actual feature check with the HAS_TRAVERSABLE_TUPLE check. On Tue, May 13, 2014 at 11:20 AM, Sergei Trofimovich <slyfox(a)community.haskell.org> wrote: > 1 patch for repository http://repetae.net/repos/jhc: > > Tue May 13 21:19:19 FET 2014 Sergei Trofimovich <slyfox(a)community.haskell.org> > * src/Support/Inst.hs: don't define instances present in ghc-7.8 > > [ 85 of 183] Compiling Support.Inst ( src/Support/Inst.hs, src/Support/Inst.o ) > > src/Support/Inst.hs:7:10: > Duplicate instance declarations: > instance [overlap ok] Foldable ((,) a) > -- Defined at src/Support/Inst.hs:7:10 > instance Foldable ((,) a) -- Defined in 'Data.Foldable' > > src/Support/Inst.hs:9:10: > Duplicate instance declarations: > instance [overlap ok] Traversable ((,) a) > -- Defined at src/Support/Inst.hs:9:10 > instance Traversable ((,) a) -- Defined in 'Data.Traversable' > > > [src/Support/Inst.hs: don't define instances present in ghc-7.8 > Sergei Trofimovich <slyfox(a)community.haskell.org>**20140513181919 > Ignore-this: 846e90c9a24c9eccc5c7358078529a72 > > [ 85 of 183] Compiling Support.Inst ( src/Support/Inst.hs, src/Support/Inst.o ) > > src/Support/Inst.hs:7:10: > Duplicate instance declarations: > instance [overlap ok] Foldable ((,) a) > -- Defined at src/Support/Inst.hs:7:10 > instance Foldable ((,) a) -- Defined in 'Data.Foldable' > > src/Support/Inst.hs:9:10: > Duplicate instance declarations: > instance [overlap ok] Traversable ((,) a) > -- Defined at src/Support/Inst.hs:9:10 > instance Traversable ((,) a) -- Defined in 'Data.Traversable' > ] hunk ./src/Support/Inst.hs 1 > +{-# LANGUAGE CPP #-} > module Support.Inst where > > import Control.Applicative > hunk ./src/Support/Inst.hs 8 > import Data.Foldable > import Data.Traversable > > +#if __GLASGOW_HASKELL__ < 708 > instance Foldable ((,) a) where > foldMap = foldMapDefault > instance Traversable ((,) a) where > hunk ./src/Support/Inst.hs 13 > traverse f (x,y) = (,) x <$> f y > +#endif > > instance Functor ((,,) a b) where > fmap = fmapDefault > > -- John Meacham - http://notanumber.net/

1 0

Re: [jhc] [Haskell] ANNOUNCE: jhc-0.8.1
by John Meacham 13 May '14

13 May '14

I modified it so all packages depended on are listed in only a single spot in the configuration to make it easier to change and give much better error messages. see the new http://repetae.net/repos/jhc/configure.ac and the NEEDS_PACKAGE macros. Possibly existing instances are pretty easy to work around, see USE_MONOID_DOC in the config file. I'll make the show instance for identity conditionally defined too, since that seems to be floating around. Having to specify specific versions in the script would be hacky, much better to actually identify the reasons for incompatibility and choose based on that. John On Mon, May 12, 2014 at 10:36 PM, Krzysztof Skrzętnicki <gtener(a)gmail.com> wrote: > Hmm, I'll give it a try, thanks! > As for the other errors for sure there is missing Binary instance for strict > ByteStrings in newest binary package. > Another one is instance for Show (Identity a) which in turn needed to be > commented out because it appeared in newer version of some other package, > don't know which one. > Lastly I get a lot of errors caused by mixing library versions. I *think* > that the problem is that, unlike Cabal, the Makefile specifies -package foo > without specific versions. The result is that, given the fact that I have > more than 1 version of many libraries installed, it tries to build with two > different versions of same library, or so would it seem: 1 is dependency of > other library, the other is from -package specification. I think this is the > case. > This is why I asked for specific versions of all the libraries involved. The > idea was to simply specify *all* of them on command line and/or Makefile. > > Right now I don't have time to replicate the errors, but I will try again > building jhc later. > > > Best regards, > Krzysztof > > > On Tue, May 13, 2014 at 5:09 AM, John Meacham <john(a)repetae.net> wrote: >> >> Yeah, there was a bug in the way it detected editline/readline which >> has been fixed in the repo. >> >> You can run configure with --disable-line to work around it. or change >> the word USE_NOLINE to USE_READLINE in src/Util/Interact.hs >> >> always some silly typo that works its way in somewhere. I should stop >> the version number shift and declare it 1.0.0 and use the third digit >> for actual point releases rather than keep the perpetual 0.x.y wasting >> the first digit. but then I can't hide behind the 'beta' shield >> anymore. :) >> >> John >> >> On Mon, May 12, 2014 at 7:56 PM, Jens Petersen >> <juhp(a)community.haskell.org> wrote: >> > Thank you for the new release. :) >> > >> > On 13 May 2014 04:40, John Meacham <john(a)repetae.net> wrote: >> >> >> >> as for the packages i've been testing with >> >> >> >> >> >> >> >> fgl,regex-compat,bytestring,binary,mtl,containers,unix,utf8-string,zlib,HsSyck,filepath,process,syb,old-time,pretty. >> > >> > >> > and editline ? >> > >> > For me it fails to build on Fedora 20 with the readline package but >> > completes with editline. >> > >> > Krzysztof: maybe try removing or hiding the readline package or posting >> > your >> > build error. :) >> > >> > Jens >> > >> >> >> >> -- >> John Meacham - http://notanumber.net/ > > -- John Meacham - http://notanumber.net/

2 5

Re: [jhc] [Haskell] ANNOUNCE: jhc-0.8.1
by John Meacham 13 May '14

13 May '14

Hi, I need to update that page, I compile it with the ubuntu ghc and the ubuntu packaged ones. Can you tell me what libraries you are having issues with? the ./configure should tell you the names of any that I expected might be an issue, I'll actually update it to check and report on everything consistently. even ones I expect to come with ghc. If there are compatibility issues between versions that's a bug I can fix. (usually just a 'hiding' or explicit import will fix any such issue) as for the packages i've been testing with fgl,regex-compat,bytestring,binary,mtl,containers,unix,utf8-string,zlib,HsSyck,filepath,process,syb,old-time,pretty. Specific versions should not matter from anything back in ghc 7.2 days to now, if there is a bug where it won't compile with a version expected to be found in the wild, please feel free to report it. John On Mon, May 12, 2014 at 12:10 PM, Krzysztof Skrzętnicki <gtener(a)gmail.com> wrote: > Hello, > > I tried compiling jhc from source (compiled version doesn't work on my > system) but after several attempts I just couldn't find a working set of > libraries for it. Can you specify which versions of libraries are known to > work for jhc? > > Best regards, > Krzysztof Skrzętnicki > > > On Sun, May 11, 2014 at 10:20 PM, John Meacham <john(a)repetae.net> wrote: >> >> After a hiatus, jhc 0.8.1 is released. >> >> http://repetae.net/computer/jhc >> >> - New license, jhc is now released under a permissive BSD style licence >> rather >> than the GPL. The license is compatible with that of ghc allowing code >> mixing >> between them. >> >> - New library layout based around the standards, there are now haskell98 >> and >> haskell2010 packages that are guarenteed to be future proof strictly >> compatible with the respective standards. A package haskell-extras >> contains >> the additonal libraries from ghc's base. >> >> - Native support for complex and vector SIMD primitives, exposed via type >> functions. for instance 'foo :: Complex_ Float32_' for hardware >> accelerated >> complex 32 bit floats for instance. These are unboxed only for now, full >> library Num support in the works. >> >> - support for android as a target, you must install the android NDK to use >> this. >> >> - Support for embedded ARM architectures imported from Kiwamu Okabe's >> branch >> allowing targeting bare hardware with no OS. >> >> - user defined kinds, introduced with the 'kind' keyword otherwise looking >> like >> 'type' declarations. >> >> - export/import lists now allow namespace qualifiers kind, class, type, or >> data >> to explicitly only import or export the specific named entity. As an >> extension allowed by this, classes and types no longer are in the same >> namespace and can share names. >> >> - ForeignPtr's now have working finalizers when collected by the RTS. >> >> - CTYPE pragma to allow promoting arbitrary C types to FFIable entities. >> _______________________________________________ >> Haskell mailing list >> Haskell(a)haskell.org >> http://www.haskell.org/mailman/listinfo/haskell > > -- John Meacham - http://notanumber.net/

2 2

Re: [jhc] darcs patch: Fixer/VMap: avoid clashing with Data.Typ... (and 1 more)
by John Meacham 12 May '14

12 May '14

Thanks! Patches have been applied and pushed. John On Mon, May 12, 2014 at 12:22 AM, Sergei Trofimovich <slyfox(a)community.haskell.org> wrote: > 2 patches for repository http://repetae.net/repos/jhc: > > Mon May 12 10:15:45 FET 2014 Sergei Trofimovich <slyfox(a)community.haskell.org> > * Fixer/VMap: avoid clashing with Data.Typeable.Proxy > > Fails on ghc-7.8 thusly: > [ 83 of 182] Compiling Fixer.VMap ( src/Fixer/VMap.hs, src/Fixer/VMap.o ) > > src/Fixer/VMap.hs:29:28: > Ambiguous occurrence Proxy > It could refer to either Fixer.VMap.Proxy , > defined at src/Fixer/VMap.hs:33:1 > or Data.Typeable.Proxy , > imported from Data.Typeable at src/Fixer/VMap.hs:16:1-20 > (and originally defined in Data.Proxy ) > > Mon May 12 10:21:14 FET 2014 Sergei Trofimovich <slyfox(a)community.haskell.org> > * Util.Interact: fix build failure against readline > > On ghc with both readline and editline enabled fails thusly: > [ 84 of 182] Compiling Util.Interact ( src/Util/Interact.hs, src/Util/Interact.o ) > > src/Util/Interact.hs:41:5: > Not in scope: setCompletionEntryFunction > > src/Util/Interact.hs:42:10: > Not in scope: readline > Perhaps you meant one of these: > readLine (line 40), readFile (imported from IO) > > src/Util/Interact.hs:46:19: Not in scope: addHistory > src/Util/Interact.hs:154:29: Not in scope: addHistory > src/Util/Interact.hs:188:19: Not in scope: addHistory > > > [Fixer/VMap: avoid clashing with Data.Typeable.Proxy > Sergei Trofimovich <slyfox(a)community.haskell.org>**20140512071545 > Ignore-this: 10d60bdf29789790a666b6231575f1b1 > > Fails on ghc-7.8 thusly: > [ 83 of 182] Compiling Fixer.VMap ( src/Fixer/VMap.hs, src/Fixer/VMap.o ) > > src/Fixer/VMap.hs:29:28: > Ambiguous occurrence 'Proxy' > It could refer to either 'Fixer.VMap.Proxy', > defined at src/Fixer/VMap.hs:33:1 > or 'Data.Typeable.Proxy', > imported from 'Data.Typeable' at src/Fixer/VMap.hs:16:1-20 > (and originally defined in 'Data.Proxy') > ] hunk ./src/Fixer/VMap.hs 16 > )where > > import Data.Monoid(Monoid(..)) > -import Data.Typeable > +import qualified Data.Typeable as T -- qualified to avoid clashing with T.Proxy > import List(intersperse) > import qualified Data.Map as Map > import qualified Data.Set as Set > hunk ./src/Fixer/VMap.hs 31 > vmapArgs :: Map.Map (n,Int) (VMap p n), > vmapNodes :: Either (Proxy p) (Set.Set n) > } > - deriving(Typeable) > + deriving(T.Typeable) > > data Proxy p = Proxy p | DepthExceeded > hunk ./src/Fixer/VMap.hs 34 > - deriving(Eq,Ord,Typeable) > + deriving(Eq,Ord,T.Typeable) > > instance Show p => Show (Proxy p) where > showsPrec n (Proxy p) = showsPrec n p > [Util.Interact: fix build failure against readline > Sergei Trofimovich <slyfox(a)community.haskell.org>**20140512072114 > Ignore-this: b064847e105d135da3358607654de267 > > On ghc with both readline and editline enabled fails thusly: > [ 84 of 182] Compiling Util.Interact ( src/Util/Interact.hs, src/Util/Interact.o ) > > src/Util/Interact.hs:41:5: > Not in scope: 'setCompletionEntryFunction' > > src/Util/Interact.hs:42:10: > Not in scope: 'readline' > Perhaps you meant one of these: > 'readLine' (line 40), 'readFile' (imported from IO) > > src/Util/Interact.hs:46:19: Not in scope: 'addHistory' > src/Util/Interact.hs:154:29: Not in scope: 'addHistory' > src/Util/Interact.hs:188:19: Not in scope: 'addHistory' > ] hunk ./src/Util/Interact.hs 26 > import IO > #if USE_EDITLINE > ;import System.Console.Editline.Readline > -#elif USE_NOLINE > +#elif USE_READLINE > ;import System.Console.Readline > #endif > > > -- John Meacham - http://notanumber.net/

1 0

Re: [jhc] [Haskell] ANNOUNCE: jhc-0.8.1
by John Meacham 11 May '14

11 May '14

On Sun, May 11, 2014 at 1:26 PM, Henning Thielemann <schlepptop(a)henning-thielemann.de> wrote: > Could you clarify the current relation between jhc and ajhc. Are they > independent from each other, or does ajhc still pull all features from jhc? I try to pull what I can back from ajhc, I had to take a break from developing jhc for a bit (grumble) so it was easier for kiwamu to branch on his own. As long as we don't diverge too much, there shouldn't be an issue sharing any patches between them. I don't see us diverging much as kiwamu is working mainly on the embedded targets, I'm in the process of revamping the intermediate representation to allow faster/better/easier optimizations and fixing some warts on the type system with the powers of hindsight and some inspiring papers by SPJ :) John

2 1

ANNOUNCE: jhc-0.8.1
by John Meacham 11 May '14

11 May '14

After a hiatus, jhc 0.8.1 is released. http://repetae.net/computer/jhc - New license, jhc is now released under a permissive BSD style licence rather than the GPL. The license is compatible with that of ghc allowing code mixing between them. - New library layout based around the standards, there are now haskell98 and haskell2010 packages that are guarenteed to be future proof strictly compatible with the respective standards. A package haskell-extras contains the additonal libraries from ghc's base. - Native support for complex and vector SIMD primitives, exposed via type functions. for instance 'foo :: Complex_ Float32_' for hardware accelerated complex 32 bit floats for instance. These are unboxed only for now, full library Num support in the works. - support for android as a target, you must install the android NDK to use this. - Support for embedded ARM architectures imported from Kiwamu Okabe's branch allowing targeting bare hardware with no OS. - user defined kinds, introduced with the 'kind' keyword otherwise looking like 'type' declarations. - export/import lists now allow namespace qualifiers kind, class, type, or data to explicitly only import or export the specific named entity. As an extension allowed by this, classes and types no longer are in the same namespace and can share names. - ForeignPtr's now have working finalizers when collected by the RTS. - CTYPE pragma to allow promoting arbitrary C types to FFIable entities.

2 1