Well, a breakout game does *not* work (yet) in most other FRP
implementations except Yampa, which do have firm theoretical foundations :-)
2009/4/15 Patai Gergely
I don't think using dirty tricks to implement FRP deserves flak, at all, from my POV, it sounds like complaining that the IO monad is implemented using C... meaning that if you're that close to bare thunks, you have every right to use any means necessary to make them behave properly. Dirtiness is not the problem, but the fact that it can leak out at the present moment. I want guarantees to exclude the possibility of undesired behaviour on the user side. Am I right thinking that the NOINLINE pragma on unsafeDupablePerformIO prevents the problem of multiple evaluation discussed yesterday? Or should I add NOINLINE to primitives in Elerea.Internal too? If that guaranteed sharing, it would certainly solve most of the problems we talked about. Apart from that, I'm still not sure that latching works the way intended all the time, but the fact that the breakout example works is an indication that at least it's not hopelessly broken.
Gergely
-- http://www.fastmail.fm - Access all of your messages and folders wherever you are
_______________________________________________ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe