April 2004 - template-haskell

printf example out-of-sync
by Isaac Jones 07 Apr '04

07 Apr '04

Greetings! I noticed today that the printf example from the GHC 6.2 manual is out of sync with at least the GHC 6.2 and libraries on my machine. I updated the wiki page (for that example, but not the others, yet). Since the purpose of the example is to get newbies over the "confidence barrier", I thought I'd mention it ;) http://www.haskell.org/hawiki/TemplateHaskell peace, isaac

1 0

RE: mkName and qualified names
by Simon Peyton-Jones 07 Apr '04

07 Apr '04

1 0

FW: mkName and qualified names
by Simon Peyton-Jones 07 Apr '04

07 Apr '04

Suggestions from George -----Original Message----- From: George Russell [mailto:ger@informatik.uni-bremen.de] Sent: 07 April 2004 10:37 To: Simon Peyton-Jones Subject: Re: mkName and qualified names Hello Simon, incidentally you asked for priorities. Mine are (still) (1) spliced import declarations; (2) [ee| .. |] and [dd| .. |], which only attempt to resolve names when the splice is actually used. Both are important because I would like to be able to get rid of the remaining CPP directives in my code. best wishes, George

1 0

RE: mkName and qualified names
by Simon Peyton-Jones 06 Apr '04

06 Apr '04

Template Haskell folk I've fixed several bugs in Template Haskell, and made minor changes as follows * Two bugs reported by Duncan Coutts - do {d <- runQ $ [| let foo = bar where bar = 3 in foo |]; print d} used to panic - using newName "foo" twice gave rise to duplicate-binding warning * Make Language.Haskell.TH export dyn :: String -> ExpQ, as advertised in "note2" * Give an explicit export list to Language.Haskell.TH, so that it no longer exposes random helper functions from inside TH.Lib, say. I may have inadvertently hidden something useful, so please yell if so * Arrange that mkName (and hence dyn) understand qualified names. So dyn "Maybe.isJust" will now generate the qualified name, as you might expect. All of this in the HEAD only, of course. Keep pinging if there are bugs you trip over. And what is the top priority for improvement? Simon | -----Original Message----- | From: George Russell [mailto:ger@informatik.uni-bremen.de] | Sent: 02 April 2004 10:12 | To: Simon Peyton-Jones | Subject: Re: mkName and qualified names | | Simon Peyton-Jones wrote: | > Ah, indeed so. Apologies. | > | > The constructor Var is actually called VarE. | > | > And "dyn" should be defined but isn't. I was wrong to say the mkName | > won't do the job -- it will. But as you found, it doesn't work for | > qualified names. It probably should. (Or maybe "dyn" should but mkName | > should not, because dyn is only used for *occurrences* whereas mkName | > makes a name that you can use for binding too. It wouldn't make sense | > to say | > let n = mkName "A.x" in LamE (VarP n) (VarE n) | | Don't you bind a qualified name in a record or class declaration sometimes, | for example | x = M.A {M.b = True} | or | instance M.A B where | M.c = True | | Or is that something different? | | just checking, | | George

1 0

RE: mkName and qualified names
by Simon Peyton-Jones 02 Apr '04

02 Apr '04

Ah, indeed so. Apologies. The constructor Var is actually called VarE. And "dyn" should be defined but isn't. I was wrong to say the mkName won't do the job -- it will. But as you found, it doesn't work for qualified names. It probably should. (Or maybe "dyn" should but mkName should not, because dyn is only used for *occurrences* whereas mkName makes a name that you can use for binding too. It wouldn't make sense to say let n = mkName "A.x" in LamE (VarP n) (VarE n) ) So that's something I should fix. But meanwhile, the unqualified version works just fine mkTyConApp2 = $( varE . mkName $ "mkTyConApp" ) Simon | -----Original Message----- | From: George Russell [mailto:ger@informatik.uni-bremen.de] | Sent: 01 April 2004 16:03 | To: George Russell | Cc: Simon Peyton-Jones | Subject: Re: mkName and qualified names | | George Russell wrote: | > Simon Peyton-Jones wrote: | > | >> If you read "Notes about TH v2" you'll see that if you want something | >> that is *not* lexically scoped, you can use | >> $(dyn "x") | >> inside quotes. Section 4.4 dicssusses | > | > | > "dyn" doesn't seem to be exported by Language.Haskell.TH in the HEAD | > version. | | ... | | and while "Notes about TH v2" defines "dyn s = return (Var (mkName s))", | I can't do that because Language.Haskell.TH in the HEAD version doesn't | export Var either. I did use "varE" in my original example. | | best wishes, | | George

1 0