I believe that the limitation is that they use Programatica's parser to get an AST to run their refactorings on. I think they've looked several times at using ghc's apis to do this, but hit various problems. I think that the main problem is that no other parser preserves things like code layout and commenting, which is of course pretty critical to refactoring programs in a sane kind of way.
I think the main issue was that we looked at the port too early. I believe GHC would offer everything in terms of position information but at the time of the port there were issues such as Strafunski traversing over GHC's AST (now made possible with derivable instances); together with time restraints and the learning curve of GHC's internals. There is a technical report available that details some of the work: (Porting HaRe to the GHC API) http://www.cs.kent.ac.uk/pubs/2005/2266/ Thanks, Chris.
Thanks
Tom _______________________________________________ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe