- Glasgow-haskell-users

Template Haskell...
by MR K P SCHUPKE 30 Oct '03

30 Oct '03

Just a thought, but once you have a data structure representing reified code there is a couple of things it would be nice to be able to do: 1) use the Parser to read a source file (i'm assuming the parser produces the same data stuctures as template haskell) 2) run the reified code directly (using ghci?) It looks from inspection that there is a Syntax.hs and a THSyntax.hs ... Is it possible to unify the data-types from these? Can you use cmRunStmt to run a template instead of splicing it? Regards, Keean Schupke.

1 0

Template Haskell...
by MR K P SCHUPKE 29 Oct '03

29 Oct '03

Okay, I have got my template working, but its extremely unreadable... could anyone tell me how to make this look more like normal Haskell code? stringType :: String -> Q [Dec] stringType s = do x <- gensym "x" return [ Newtype [] s [] (Constr s [(NonStrict,Tcon (TconName "String"))]) [], Proto ("show"++s) (Tapp (Tapp (Tcon Arrow) (Tcon (TconName s))) (Tcon (TconName "ShowS"))), Fun ("show"++s) [Clause [Pcon s [Pvar x]] (Normal (App (Var "showString") (Var x))) []], Instance [] (Tapp (Tcon (TconName "Show")) (Tcon (TconName s))) [ Fun "showsPrec" [Clause [Pwild,Pvar x] (Normal (App (Var ("show"++s)) (Var x))) []]] ]

2 1

ghc-6.0.1 on OS X Panther build
by Gregory Wright 29 Oct '03

29 Oct '03

Hi Wolfgang, I tried building 6.0.1 on Panther using 6.0.1 built on Jaguar. (The build was run under darwinports.) It failed with the following: ------------------------------------------------------------------------ ==fptools== make all -r; in /Users/gwright/src/darwinports/dports/lang/ghc/work/ghc-6.0.1/ libraries/base ------------------------------------------------------------------------ rm -f GHC/Base.o; if [ ! -d GHC/Base_split ]; then mkdir GHC/Base_split; else /usr/bin/find GHC/Base_split -name '*.o' -print | xargs rm -f __rm_food; fi; ../../ghc/compiler/ghc-inplace -H16m -O -I/opt/local/include -L/opt/local/lib -pgmP "gcc3 -E -traditional" -fglasgow-exts -cpp -Iinclude -#include HsBase.h -funbox-strict-fields -package-name base -O -Rghc-timing -split-objs -c GHC/Base.lhs -o GHC/Base.o -ohi GHC/Base.hi /tmp/ghc3798.split__111.s:unknown:missing indirect symbols for section (__DATA,__nl_symbol_ptr) <<ghc: 138861204 bytes, 13 GCs, 715994/1423644 avg/max bytes residency (2 samples), 19M in use, 0.01 INIT (0.03 elapsed), 2.01 MUT (22.94 elapsed), 0.28 GC (0.31 elapsed) :ghc>> make[2]: *** [GHC/Base.o] Error 1 make[1]: *** [all] Error 1 make: *** [build] Error 1 Warning: the following items did not execute (for ghc): com.apple.build > Ever see anything like this before? Best Wishes, Greg

2 1

GHC and Kinds
by Steffen Mazanek 29 Oct '03

29 Oct '03

Hello again, I have browsed the GHC commentary and I could not find a lot of information about kinds. But kinds are very important for my implementation! The parser seems to call liftedTypeKind from TypeRep. But in TypeRep we have type Kind = Type What does this mean? I briefly explain the basics of my algorithm, so that you see how much it depends on the notion of Kinds. I have the following declaration: data Kind = KVar Kindvar --kind variables | KVarInv Kindvar --inversion of kind variables | Star --corresponds to * | Cov --corresponds to + | Contr --corresponds to - | Inv --corresponds to x | Kind :-> Kind --function kind deriving Eq The new atomic kinds Cov, Contr and Inv can be explained by some examples. In my system, the list type constructor, that is covariant in its argument, i.e., ([] a)<([] b) whenever a<b, is represented by (Tycon "[]" (Cov:->Star)). In contrast, the arrow type constructor's kind is given by (Contr:->(Cov:->Star)) due to its contravariance. If one tries to establish a subtype relationship, kinds are taken into consideration. How can I gently integrate this behavior in the GHC? Till now I have worked with "Typing Haskell in Haskell" and I really expected a declaration like data Kind = Star | Kfun Kind Kind Any help is greatly appreciated. Regards, Steffen Mazanek

1 0

Re: Marshalling functions was: Transmitting Haskell values
by C.Reinke 29 Oct '03

29 Oct '03

> >> Is [marshaling functions] something absolutely impossible in > >> Haskell and by what reason? Just because of strong typing (forgive > >> my stupidity ;)? Or are there some deeper theoretical limitations? If you're interested in some recent work here, have a look at Clean (similar enough to Haskell), specifically their work on first-class I/O. I gave a reference in another thread: http://www.haskell.org/pipermail/haskell-cafe/2003-October/005243.html Then follow the references. "orthogonal persistence" is another nice search key.. > > The big theoretical issue is whether it would provide an Eq or Show > > instance for -> by the backdoor. Careful API design could avoid the > > worst of this. once you can do IO, you can inspect the GHC heap for "equality", or map memory into Haskell data structures, so that's not really new .. > What's the problem with a Show instance for ->? (Or, put another way: > what makes the current Show instance in the Prelude harmless compared to > a fuller implementation?) have you tried to show a function?-) an equality based on that would be very rough, a proper equality would be undecidable, and there are whole worlds of other "equalities" in between. Cheers, Claus

1 0

RE: Transmitting Haskell values
by Simon Peyton-Jones 29 Oct '03

29 Oct '03

| > The GHC versions are simply more efficient versions. Show and Read are | > completely portable. | | AIIEE! Now they tell me -- I've optimized one of my programs, and it | now spends most of it's time 'read'ing. So the thing is just to | import GHC(read) instead of using the Prelude version? I guess it's | worth a try (although I expect I really need to go to binary IO) No, you get the efficient ones when you import the Prelude. GHC's Read class has two extra methods, readPrec and readListPrec, which read using a parser monad rather than the 'list of successes' that is (sadly) built into the definition of the standard Read class. When defining a new type, and making a Read instance for it, you'll get more efficient reading if you define GHC's Read methods (the normal readsPrec and readList will then be defined in terms of them). There are many examples in GHC.Base.Read The underlying parser technology is Koen Claessens's parallel parser http://www.cs.chalmers.se/~koen/publications.html This new parsing stuff is hardly documented at all (e.g. http://www.haskell.org/ghc/docs/latest/html/base/GHC.Read.html#readPrec) . If anyone would like to remedy this by writing Haddock documentation, it'd be great. (I'd be glad to hand-hold.) Simon

1 0

Templates...
by MR K P SCHUPKE 29 Oct '03

29 Oct '03

I would like to create a template to generate type declarations, for example: newtype "mytype" = "myconstructor" String I am struggling without an example... how do I do it? Regards, Keean Schupke.

1 0

unknown symbol `__stginit_List_'
by Josef Svenningsson 29 Oct '03

29 Oct '03

Hi! Here's a letter from someone who has next to no clue about what he is doing so have patience and don't assume any knowledge when answering... I'm trying to create a package for ghc. I've struggled quite a lot and gotten this far: (This is when invoking ghci -package yahu) Loading package base ... linking ... done. Loading package yahu ... linking ... /.../chalmers.se/fs/cab/cs/work/proj/multi/pub/lib/yahu/Yahu/YahuHaskell.o: unknown symbol `__stginit_List_' ghc-6.0.1: panic! (the `impossible' happened, GHC version 6.0.1): can't load package `yahu' I'm surprised by this. Why do I get this error message? From the name of the symbol I would expect that it can be found in package base but aparently not... Please advice. /Josef

3 3

GHC changes
by Steffen Mazanek 29 Oct '03

29 Oct '03

Hello GHC freaks, I have made some changes to the GHC source code (6.0.1) and I would really welcome your feedback, whether they are ok for my task (first step: introduce record expressions of the shape 'struct {x1=e1,x2=e2,...,xn=en}'): * added ("struct", ITstruct, bit glaExtsBit), to ghcExtensionKeywordsFM and ITstruct to Token in Lex.lhs * added | Struct [(id, HsExpr id)] to data type HsExpr id in HsExpr.lhs Please tell me, what is this id for? Can it be used in this context? What is FastString? * added 'struct' { ITstruct } to the token part of Parser.y (please tell me, why in the lexer we declare ITstruct to represent the keyword "struct" and in the parser specification we turn back to 'struct'), added structassignments : structassignments ',' structassignment { $3 : $1 } | structassignments ',' { $1 } | structassignment { [$1] } | {- empty -} { [] } structassignment : qvar '=' exp { ($1,$3) } as new production rules (what type has structassignments? I can not simply write [(id,HsExpr id)], is qvar '=' exp appropriate? I do not understand structassignments completely (I have tried to make it analog to fbinds)), extended exp10 by | 'struct' '{' structassignments '}' {Struct $3} | '{' structassignments '}' {Struct $2} and finally added | Struct [(id, HsExpr id)] to HsExpr How can I speed up the recompilation process? I just want to check, whether my changes are compilable or not, is it possible to switch off optimization and other time consuming transformations? At next I will try to master the following steps: * embed the remaining new production rules to the parser, e.g., record type declaration, selectors, particular kind annotations * adaptation of the type checker * reduction of struct expressions Thank you very much, Steffen Mazanek

1 0

File permissions of /tmp files
by Peter Simons 29 Oct '03

29 Oct '03

Hi, it appears that ghc (and ghci) create temporary files in /tmp during the un-lit phase, at least. It would be nice, if ghc would ... - create those files with read/write permission for the owner only, 0600. - set-up some atexit(3) job, or whatever, to make sure those files _are_ deleted when ghc exits. It might even be possible to create the file and then remove(3) it, right away. The file handle will continue to work, but there is never an entry in the file system. Just a suggestion. :-) Peter

3 2