[MLton-devel] yeah, that did the trick

Matthew Fluet fluet@CS.Cornell.EDU
Mon, 16 Sep 2002 17:08:42 -0400 (EDT)

> This is a bit scary.  What are the semantics supposed to be for infinite
> reals under comparison.  I.e., I am assuming that
> 	Math.pow (3.0, 1000000.0)
> is infinite, but isn't bit-identical to Real.posInf.  Is it allowed for
> the Real.== to return false?

I don't believe so.  My understanding is that all posInf values are bit
equal.  Steve could check the bits with the PackReal structure, although
that ensures that they will be moved out of the floating point stack.

This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
MLton-devel mailing list