
Seems there's some trouble with your mail server, I keep getting 'delivery failure'. Hopefully mails through the list will get through. Anyway, I've now merged my repo with the changes to haskell-src in CVS. There were no merge conflicts.
I'd also like to parameterize the parser over the Extensions type currently in Distribution.Extension (maybe it should move to Language.Haskell.Extension) so that users could selectively enable extensions.
I agree that the parser (and the lexer too) should be parametrized over extensions. The Extensions type in Distribution.Extension looks a bit messy, but I guess that's a different issue entirely, I suppose once that's stable we'll have no problems using it. I've started adding some infrastructure to my lib for doing this parametrization, it seems to be fairly straight-forward (but tedious). I'll keep working on it. There are a few non-conservative changes introduced in haskell-src-exts, so I guess it's best to wait until all the extension parametrization is there before adding my stuff to the official cvs. /Niklas

On the side: Arthur Baars has a full Haskell combinator parser. He made it work with ghc and could bootstrap itself, i.e. it parses all the Ghc sources and all of the fptools libraries. The speed of the parser seems pretty good: ghc compiles itself and fptools about 5-10% slower than with the standard haddock parser. I am not sure if he reads this list so you might want to contact him directly to get the sources. -- Daan Leijen. Niklas Broberg wrote:
Seems there's some trouble with your mail server, I keep getting 'delivery failure'. Hopefully mails through the list will get through.
Anyway, I've now merged my repo with the changes to haskell-src in CVS. There were no merge conflicts.
I'd also like to parameterize the parser over the Extensions type currently in Distribution.Extension (maybe it should move to Language.Haskell.Extension) so that users could selectively enable extensions.
I agree that the parser (and the lexer too) should be parametrized over extensions. The Extensions type in Distribution.Extension looks a bit messy, but I guess that's a different issue entirely, I suppose once that's stable we'll have no problems using it. I've started adding some infrastructure to my lib for doing this parametrization, it seems to be fairly straight-forward (but tedious). I'll keep working on it.
There are a few non-conservative changes introduced in haskell-src-exts, so I guess it's best to wait until all the extension parametrization is there before adding my stuff to the official cvs.
/Niklas _______________________________________________ Libraries mailing list Libraries@haskell.org http://www.haskell.org/mailman/listinfo/libraries

Niklas Broberg
I agree that the parser (and the lexer too) should be parametrized over extensions. The Extensions type in Distribution.Extension looks a bit messy, but I guess that's a different issue entirely, I suppose once that's stable we'll have no problems using it.
Actionable feedback on the Distribution library is always welcome. peace, isaac
participants (3)
-
Daan Leijen
-
Isaac Jones
-
Niklas Broberg