- Glasgow-haskell-users

Problem Building Package
by Matthew Donadio 23 Apr '03

23 Apr '03

Hi all, I am having a bit of trouble builing local packages. I have three libraries, A, B, and C, that I want to make into local packages. A and B depend only on base, while C depends on base, A, and B. The souces gets built with ghc --make -package-name=A {list of files} ghc --make -package-name=B {list of files} ghc --make -package-name=C {list of files} and I create the archives libHS[A,B,C].a This all works fine. I have three simple .pkg. If I add the dependencies into package-deps, then ghc-pkg says that the dependencies do not exist. If I leave that blank, then I can create my local .conf file. If I do this, I cannot get ghc to use the .conf file. For example, if I try ghc --make -package-conf foo.conf -package C -o SomeApp SomeApp.hs Then ghc complains that it can't find the interface files, even though import-dirs is correct in foo.conf. Can anyone shed some light on this, or point me to a small library that builds a GHC package? I know I am doing something stupid, but I can't figure it out. I tried looking at the GHC Makefiles, but I am having trouble following them because of all the includes. Thanks. -- Matthew Donadio (m.p.donadio(a)ieee.org)

1 0

Immutable CStrings
by George Russell 23 Apr '03

23 Apr '03

How would one implement an CStringLen-style type that (a) was efficient, in particular could be read and written to Handles efficiently; (b) would get automatically deallocated by the Haskell garbage collector, when Haskell no longer referred to it; (c) was immutable, so that once created the character data could not be changed; (d) consequently had the property that conversion to and from String would be a pure operation; (e) could be passed to and from C using the FFI. (Of course C would need you to split the length and character data up; the character data would presumably have type "const char *".) ? It would be rather nice to have such a type.

2 1

RE: ghc-[45].X on mips-sgi-irix65 (not yet)
by Rafael Martinez Torres 23 Apr '03

23 Apr '03

Well , after 8 months of hard-hacking to port ghc into mips-sgi-irix65, this is my modest contribution to the Haskell community: just a stauts-report. :-| . I hope to be usefull for future portes, not only for those for mips-sgi-irix65, saving time on things I have spent months on... For the time beeing I can not upload it at http://segovia.sip.ucm.es/~rmartine , so the links to download the "unstable" patches will not be available...However, they would be usefull for porters, not for final users... But I hope them to be available in the next days... I will announce... If some one is interesting , please e-mail to rmartine(a)fdi.ucm.es, and I will send them...

1 1

Hdirect with ghc 5.04.3, building problems
by Petter Egesund 23 Apr '03

23 Apr '03

Hi, I am trying to build Hdirect with ghc 5.04.3. The sources are from http://galois.com/~sof/hdirect-0.19-src.tar.gz I follow the instructions in INSTALL. make boot works fine, but make lib gives this errs; make -C lib boot && make -C lib all make[1]: Entering directory `/cygdrive/h/download/hdirect-0.19/lib' ghc -M -optdep-f -optdep.depend -optdep-o -optdepo -cpp -DBEGIN_GHC_ONLY='-}' -DEND_GHC_ONLY='{-' -DBEGIN_NOT_FOR_GHC='{-' -DEND_NOT_FOR_GHC='-}' -DELSE_FOR_GHC='-}-}' -package-name hdirect -fglasgow-exts -fvia-C -Wall -fno-warn-deprecations -package lang -static Pointer.lhs HDirect.lhs PointerPrim.hs WideString.hs make[1]: Leaving directory `/cygdrive/h/download/hdirect-0.19/lib' make[1]: Entering directory `/cygdrive/h/download/hdirect-0.19/lib' ghc -cpp -DBEGIN_GHC_ONLY='-}' -DEND_GHC_ONLY='{-' -DBEGIN_NOT_FOR_GHC='{-' -DEND_NOT_FOR_GHC='-}' -DELSE_FOR_GHC='-}-}' -package-name hdirect -fglasgow-exts -fvia-C -Wall -fno-warn-deprecations -package lang -static -c HDirect.lhs -o HDirect.o -osuf o HDirect.lhs:1218: Variable not in scope: `getTag#' make[1]: *** [HDirect.o] Error 1 make[1]: Leaving directory `/cygdrive/h/download/hdirect-0.19/lib' make: *** [lib] Error 2 I have this setting in config.mk: HC=ghc TARGETPLATFORM=cygwin32 Any clue? Cheers, Petter Egesund Norway ########################################### This message has been scanned by F-Secure Anti-Virus for Microsoft Exchange. For more information, connect to http://www.F-Secure.com/

1 0

RE: Existentials
by Hans Aberg 22 Apr '03

22 Apr '03

At 10:34 +0100 2003/04/22, Simon Marlow wrote: >If the quantifier is outside the constructor, then it scopes over all >the fields of the constructor. So, if you write > > data T = forall a . C t1 .. tn > >what should this mean (assuming you want a non-existential >interpretation)? I think you have problem here because, you apply the quantifier to something which is not a predicate. This practise makes it hard to figure out what the semantics should be. For example, in the original example: >At 09:02 +0100 2003/04/17, Simon Peyton-Jones wrote: > data T = forall a. Foo a (a -> Int) if T is a well formed formula in a quantification theory there is an associated predicate P(a) = Foo a (a -> Int) from which one gets T := all a: P(a). One of the quantification axioms that I listed said: axiom Q4. t free for x in A |- (all x: A) => [x\t]A. It means that from T, one can plug in a value t (a term) [a\t]T = [a\t]P(a) to get Foo t (t -> Int) Is that not what you want? However, in math, quantifiers are only applied to predicates, not sets. For sets, one uses operations such as union, intersection and complement. It seems me that data types correspond to sets, not predicates. Thus, one should not use quantifiers "all" and "exist" at all on them, but something corresponding to sets instead. In naive set theory, one identifies the set S := { x | P(x) } with the predicate P(x). If one should get the set S from the closed formula T above, one first applies the axiom Q4 to get P(x) := [a\x]T, and then one can get S from P(x). In the example above, the set becomes S = { a | Foo a (a -> Int) } To begin with, this does not make logically sense, as Foo a (a -> Int) is supposed to return not logical values as is required by a predicate, but data of type T. It seems me that one should instead have someting like data T = union a. Foo a (a -> Int) or data T = disjointUnion a. Foo a (a -> Int) Then, when one instantiates, one will create another addition to the data type, which is how I think is what you really want. Right? Hans Aberg

1 0

DijitalCenter - Haftanın Mutlu Ürünleri
by DijitalCenter 22 Apr '03

22 Apr '03

1 0

RE: Existentials
by Simon Marlow 22 Apr '03

22 Apr '03

> "Simon Marlow" <simonmar(a)microsoft.com> writes: > > > > The example I tried failed, so I assumed it wasn't supported. > > > > You've written an existential constructor. For universal > > quantification, write it like this: > > > > data T = Foo (forall a . Enum a => a -> a) > > > > a good illustration of the confusion caused by the dual use > of forall, I > > guess :-) > > Exactly so. :-) If we change the syntax of existentials, would it > be possible to write my local universal example as I did originally > and have it work as expected? Or will it still be necessary to push > the quantifier inside the constructor? If the quantifier is outside the constructor, then it scopes over all the fields of the constructor. So, if you write data T = forall a . C t1 .. tn what should this mean (assuming you want a non-existential interpretation)? Perhaps the sensible meaning is that it is isomorphic to data T = C !(forall a . (t1,...,tn)) Cheers, Simon

1 0

RE: Classes and interfaces, a naming restriction
by Simon Peyton-Jones 22 Apr '03

22 Apr '03

| Aparantly there's a restriction on having a class and a data type with | the same name. A strange thing to want to do? Maybe not, but illegal in Haskell. I think the main reason is distinguishing classes from instances in export and import lists: module Foo( IStream(..) ) where ... What gets exported, the class IStream or the data type IStream? The best you can do is put them in separate modules. Then you can name them as M1.IStream and M2.IStream if you want. Simon | Here's what I'd like to do: | | class IStream stream char where | readStream :: stream -> Int -> IO [char] | | data IStream char = forall stream. IStream stream char => IStream stream | | instance IStream (IStream char) char where | readStream (IStream stream) = readStream stream | | So here the *type* IStream is the representative member of the *class* | IStream. An OOP analogy is that a particular 'stream' such that | (IStream stream char) => stream | is a particular concrete class that inherits from a virtual base class IStream, | but a 'stream' that is of type IStream is a pointer/reference to something that | inherits from the virtual base class (or a COM interface pointer) | | With the former we have (potentially) static binding, with the latter it is necessarily | dynamic binding. So whenever I need a hetrogenious collection of streams I can apply: | IStream :: forall c s. (IStream s c) => s -> IStream c | | Do I really have to use data IStreamWrapper = ... instead? :-( | | Duncan | _______________________________________________ | Glasgow-haskell-users mailing list | Glasgow-haskell-users(a)haskell.org | http://www.haskell.org/mailman/listinfo/glasgow-haskell-users

1 0

Classes and interfaces, a naming restriction
by Duncan Coutts 21 Apr '03

21 Apr '03

Hi, Aparantly there's a restriction on having a class and a data type with the same name. A strange thing to want to do? Here's what I'd like to do: class IStream stream char where readStream :: stream -> Int -> IO [char] data IStream char = forall stream. IStream stream char => IStream stream instance IStream (IStream char) char where readStream (IStream stream) = readStream stream So here the *type* IStream is the representative member of the *class* IStream. An OOP analogy is that a particular 'stream' such that (IStream stream char) => stream is a particular concrete class that inherits from a virtual base class IStream, but a 'stream' that is of type IStream is a pointer/reference to something that inherits from the virtual base class (or a COM interface pointer) With the former we have (potentially) static binding, with the latter it is necessarily dynamic binding. So whenever I need a hetrogenious collection of streams I can apply: IStream :: forall c s. (IStream s c) => s -> IStream c Do I really have to use data IStreamWrapper = ... instead? :-( Duncan

1 0

BLACKHOLE in heap profile
by Hal Daume III 19 Apr '03

19 Apr '03

My program which should use constant stack/heap is using a ton of heap. I profiled with -hd and the guy that's eating all the heap is BLACKHOLE. What does this mean? -- Hal Daume III | hdaume(a)isi.edu "Arrest this man, he talks in maths." | www.isi.edu/~hdaume

3 5