INFINITY is C99, NAN is not. http://www.gnu.org/software/libtool/...y-and-NaN.html
[edit] Hmm, strike that. NAN looks like it might be C99 too.
http://std.dkuug.dk/jtc1/sc22/wg21/d...2002/n1372.txtC99 adds all sorts of stuff:
...
-- macros INFINITY, NAN for special floating-point values
Of course, the GCC page is likely more trustworthy than that source . . . . [/edit]