Despite the fact that there are a few bugs which escaped 5.04.3, I think
we should concentrate our efforts on 5.06 (or 6.00) now. We can move
towards a release fairly quickly (say, a month or so). Comments?
I had hoped that there was time to get the threads issue finished before 6.00...
The threaded RTS without bound threads is fine, but due to it's inherent problems with some foreign libraries I think we still can't enable it by default (HOpenGL installation is difficult enough without having to recompile GHC).
Now I could probably implement a prototype of the current proposal in one weekend, and a finished version in two or three, but it'd still need some testing, and more discussion:
Daan Leijen has recently expressed fundamentel disagreement with the whole idea of the threaded RTS, and offered a (not-yet-fully-defined) counter-proposal (with which I, in turn, disagree fundamentally).
It would be a pity to have to wait another year for GHC to have proper thread support... Of course, it wouldn't bee to good either to hold up all of GHC just because of one feature that only some people really need...
Cheers,
Wolfgang
_______________________________________________ Cvs-ghc mailing list [EMAIL PROTECTED] http://www.haskell.org/mailman/listinfo/cvs-ghc
