> Wouldn't it be good to implement an LVM import/export feature in GHC, Hugs > and nhc98?
To read and write lvm bytecodes? That could be quite a lot of work because the bytecodes are specific to the abstract machine and the abstract machines are different. Unless the LVM is essentially just a G-machine (I suspect it isn't), generating LVM code in Hugs would require a new code generator. Reading LVM code is potentially even harder. IMO, the right way to do this is to read/write 'core' lambda calculus like the code GHC uses as its intermediate form. Having found a way to import/export code, there is a second problem to be overcome. Different compilers have different sets of primitive operations. I'm sure they all have the same arithmetic operations and, of course, they differ in whether or not they have concurrency operations but there are big differences in their implementation of the IO monad and exceptions, their array operations, etc. -- Alastair Reid www.haskell-consulting.com _______________________________________________ Haskell mailing list [EMAIL PROTECTED] http://www.haskell.org/mailman/listinfo/haskell
