From: Bil Kleb Date: 2002-02-01T23:14:47+09:00 Subject: Re: smallest non-zero number (and other limits) ts wrote: > > Well, you have the comments of matz about this > > > If we can provide portable default definition of these > > Any idea how to do this ? ANSI-compliant C compilers *have to* provide a float.h, viz, http://www.cs.bsu.edu/homepages/peb/cplusplus/float.html and since Ruby has to be compiled for each type of machine via a C compiler... Admittedly there might be some non-ANSI-compliant C compilers out there which might not have this header. These could be detected^[*] and the Float class constants depreciated (simply not set?) accordingly. However, I would not be surprised if the machine running such an ancient compiler (pre 1978 or so?) did not use punch cards which would diminish the Ruby experience anyway... ;) -- Bil *via adding float.h to autoconf's AC_CHECK_HEADERS macro in configure.in