The development version of racket includes a new feature for work with extended-precision (80-bit) floating-point numbers: http://pre.racket-lang.org/docs/html/reference/extflonums.html
If __SSE_MATH__ is defined, extflonums are enabled. On i386, I need add -msse and -mfpmath=sse to CPPFLAGS and CFLAGS for turn on the feature. I know the rules about extra CFLAGS in ports but can I make a exception here?. With or without extflonums enabled, racket will only run in CPUs with SSE, so I'm not limiting the user-base of the package. Cheers. -- Juan Francisco Cantero Hurtado http://juanfra.info
