This is the mail archive of the
libstdc++@sources.redhat.com
mailing list for the libstdc++ project.
Re: Numeric limits
- To: "Anthony Williams" <anthony at anthonyw dot cjb dot net>
- Subject: Re: Numeric limits
- From: Gabriel Dos Reis <Gabriel dot Dos-Reis at cmla dot ens-cachan dot fr>
- Date: 27 Jul 2000 12:06:37 +0200
- Cc: <libstdc++ at sourceware dot cygnus dot com>
- Organization: CMLA, ENS Cachan -- CNRS UMR 8536 (France)
- References: <000f01bff7a7$91a2ade0$5919cac2@derwent.co.uk>
"Anthony Williams" <anthony_w.geo@yahoo.com> writes:
| I was just looking at the numeric limits stuff, and noticed that
| numeric_limits<>::infinity(), quiet_NaN() and signaling_NaN() are all set to
| return zero for all the floating point types.
There are there just for transition. By no means, are they in their
frozen states.
|
| On Solaris there is a library (sunmath) that provides functions to obtain these
| values, which can then be used in gen_num_limits.cc to provide them without
| requiring this library.
I can, unfortunately, no longer get gcc (and therefore libstdc++-v3)
built on my solaris boxes (because of HUGE memory usage) -- just tried
this morning, and got seg-fault, sight. So I can't experiment myself.
However, I'm williing to cooperate wiith you to sort these things.
| Would you like the code I've written for gen_num_limits.cc which uses these on
| Solaris?
Yes, definitely.
-- Gaby