Sven Panne writes:
>
> Moving to a standard is a good thing, but Green Card seems to move
> faster than existing code for it can be adapted. What is the schedule
> for GC3 and will there be something like a GC2->GC3 converter?
>
Hope to have a release ready soon. A translator that spits out IDL
specs given Green Card 2 input will not be supplied, as the two
approaches to describing bindings to foreign functionality are
fundamentally different. Green Card 2 starts with a Haskell type
signature and tries to derive the (proto)type of the external function
from it, with the programmer making up the difference in C.
The IDL compiler goes the other way, starting with an external
specification of a function/type. It is mapped to a corresponding
Haskell function/type, leaving the programmer to make up the
difference this time in Haskell rather than C.
Hmm..labelling both tools Green Card is probably more confusing than
helpful.
GC2 will still be bugfixed (at least I hope so), as I think it still
has a role to play, especially since there's now a number of Haskell
library bindings written using it.
--Sigbjorn
