Regarding the definition of PDF_REAL_MAX: #define PDF_REAL_MAX (340282346638528859811704183484516925440.000000)
I think that it would be better to use the POSIX FLT_MAX constant instead of the raw number. I applied a patch that defines PDF_REAL_(MIN|MAX) in terms of FLT_(MIN|MAX). It also includes the float gnulib module. -- Jose E. Marchesi <[email protected]> http://www.jemarch.net GNU Project http://www.gnu.org
