[MLton-devel] cvs commit: (almost) full support for Real32

[Henry Cejtin]

I  don't  know  what the semantics requires about conversion to Real32.reals.
Some of it might be ok given that reals are no longer an eqtype.

Like I said, one thing that is definitely essential is that if  I  explicitly
test  that a Real32.real using isFinite or isNan or isNormal or class then it
can't be something where the answer would change after it go stored.  I think
that  this  also  has  to hold for Real32.real's, otherwise the functions are
rather useless.  The same goes for unordered.

I don't think that the efficiency concern  is  very  large  for  Real32.reals
because  the  only  real  reason  to  have  them  is for storing them (say in
arrays).  The case of Real64.real's is much  more  serious  a  concern  since
again  there  are  80-bit  reals  that, when converted to 64 bits will become
infinite (or subnormal).  Definitely  a  mess  since  it  would  be  horribly
expensive to have to store normal Real.real's after every operation.



-------------------------------------------------------
This SF.Net email sponsored by: Free pre-built ASP.NET sites including
Data Reports, E-commerce, Portals, and Forums are available now.
Download today and enter to win an XBOX or Visual Studio .NET.
http://aspnet.click-url.com/go/psa00100003ave/direct;at.aspnet_072303_01/01
_______________________________________________
MLton-devel mailing list
MLton-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/mlton-devel