June 2003 - template-haskell

Reification of top level variables - was: Re: Problem with Labelled Fields
by Eric Offermann 27 Jun '03

27 Jun '03

> What you will actually > get is the code for applying "c_succ" to lift c (it is lifted as it is > an argument to the function). However, lifting c needs to use this > instance again so loops. > > Ian Ian, thanks a lot for pointing that out. This leads me to the following modification: {- spliceC.hs -} module spliceC where import Language.Haskell.THSyntax data C a = C {c_succ :: a -> a ,c_pred :: a -> a } instance Lift (C a) where lift c = recCon "C" [fieldExp "c_succ" [| c_succ $(var "sampleC") |] ,fieldExp "c_pred" [| c_pred $(var "sampleC") |] ] c_succS :: C a -> ExpQ c_succS c = [| \ a -> c_succ c a |] sampleC :: C Int sampleC = C {c_succ = \ n -> n + 1 ,c_pred = \ n -> n - 1 } Thus, the possibility of using a "reifyVar c :: Q String" in the instance of lift c yielding the original name of c would be useful to keep the instance decleration generic. Any hint on implementing that? Thanks, Eric -- +++ GMX - Mail, Messaging & more http://www.gmx.net +++ Bitte lächeln! Fotogalerie online mit GMX ohne eigene Homepage!

2 1

Problem with Labelled Fields
by Eric Offermann 26 Jun '03

26 Jun '03

Hi! Perhaps there's anybody out there who can help me with the following problem with labelled fields (using ghc 6.0). The easy example modules are: {- spliceC.hs -} module spliceC where import Language.Haskell.THSyntax data C a = C {c_succ :: a -> a ,c_pred :: a -> a } instance Lift (C a) where lift c = recCon "C" [fieldExp "c_succ" [| c_succ c |] ,fieldExp "c_pred" [| c_pred c |] ] c_succS :: C a -> ExpQ c_succS c = [| \ a -> c_succ c a |] sampleC :: C Int sampleC = C {c_succ = \ n -> n + 1 ,c_pred = \ n -> n - 1 } {- main.hs -} module Main where import spliceC sampleC_succ = $(c_succS sampleC) I suppose that this should bring a function sampleC_succ into scope, instantiated for sampleC. But in fact I get *** Exception: stack overflow Who's wrong? Thanks, Eric

2 1

RE: Templates in ghc-6.0 & --make
by Simon Peyton-Jones 06 Jun '03

06 Jun '03

[redirecting to template Haskell. Are you on that list, Christoph?] Excellent point below. The dependency tracker assumes that if the type of an imported thing is unchanged, you don't need to recompile. But here it's a change in the *value* of an imported thing that should force recompilation. Sigh. Don't hold your breath for a fix! Simon | -----Original Message----- | From: glasgow-haskell-bugs-admin(a)haskell.org [mailto:glasgow-haskell-bugs-admin@haskell.org] On | Behalf Of Ch. A. Herrmann | Sent: 05 June 2003 12:21 | To: glasgow-haskell-bugs(a)haskell.org | Subject: Templates in ghc-6.0 & --make | | Hi, | | first, many thanks to the implementors | for releasing Template Haskell. | | A little flaw which can be observed with the | example Printf/Main from the web users guide. | | Changing in "Printf.hs" the parse function | to deliver a different format is not | recognized by the dependency tracker of | ghc --make | causing Main.hs to be skipped in the compilation, | but metaprogramming is supposed to change the | code of Main.hs. | | Many thanks in advance | -- | Dr. Christoph Herrmann | Teaching and Research Assistant | University of Passau, Germany | http://www.fmi.uni-passau.de/~herrmann | | _______________________________________________ | Glasgow-haskell-bugs mailing list | Glasgow-haskell-bugs(a)haskell.org | http://www.haskell.org/mailman/listinfo/glasgow-haskell-bugs

1 0

Template Haskell webpage
by Ian Lynagh 03 Jun '03

03 Jun '03

Hi all, I have created a page on Template Haskell (based in part on Simon Peyton Jones' existing page) at http://www.haskell.org/th/ If you're using Template Haskell then please drop me an e-mail with a paragraph or so saying what you are doing and giving a link to a webpage if you have a suitable one. Likewise if you have written a paper on Template Haskell then let me know where it is, along with a BiBTeX entry for it, and I'll add it to the page. Thanks! Ian, immensely proud of his logo :-)

1 0