Dear Huggy people, I've just noticed that Hugs uses slightly different syntax than GHC for explicit for-alls in types. In Hugs you say forall a,b,c. ...type.... In GHC you say forall a b c. ...type... It's a pity to have unnecessary syntactic differences. Could we make them the same? I'd like to suggest that GHC is more consistent with the rest of Haskell. At the term level we don't use commas when we quantify: \ a b c -> .... not \ a,b,c -> ..... It's a pretty easy change to make. What think you? Simon
Simon, et. al, forall tends to be used with one argument, hence the reason its not really biting right now. If no-one objects, I'll change it this week at some point, to match GHC. The "lambda does not use comma" seems convincing to me. Andy -----Original Message----- From: hugs-bugs-admin@haskell.org [mailto:hugs-bugs-admin@haskell.org]On Behalf Of Simon Peyton-Jones Sent: Sunday, April 29, 2001 12:16 PM To: hugs-bugs@haskell.org Cc: simonpj@microsoft.com Subject: Forall syntax Dear Huggy people, I've just noticed that Hugs uses slightly different syntax than GHC for explicit for-alls in types. In Hugs you say forall a,b,c. ...type.... In GHC you say forall a b c. ...type... It's a pity to have unnecessary syntactic differences. Could we make them the same? I'd like to suggest that GHC is more consistent with the rest of Haskell. At the term level we don't use commas when we quantify: \ a b c -> .... not \ a,b,c -> ..... It's a pretty easy change to make. What think you? Simon _______________________________________________ Hugs-Bugs mailing list Hugs-Bugs@haskell.org http://www.haskell.org/mailman/listinfo/hugs-bugs
Hi Simon, Andy, et al. | [Hugs and GHC should agree on a syntax for forall expressions ... one | that doesn't require commas between variable names ...] | | forall tends to be used with one argument, hence the reason its not | really biting right now. The reason it's not biting is that the problem doesn't exist! :-) I have strong recollections that I noticed and fixed this inconsistency while traveling on a train through the French countryside in September 1999 ... even if my memory is failing me, the most recent Hugs seems to accept forall's without commas between arguments. Perhaps Simon is using an old version of Hugs, looking at some out of date documentation, or has found a specific instance where the change was missed? All the best, Mark
participants (3)
-
Andy Gill -
Mark P Jones -
Simon Peyton-Jones