This is the mail archive of the
gcc-bugs@gcc.gnu.org
mailing list for the GCC project.
[Bug c/14621] libm tan failure for pi/2 argument
- From: "bangerth at dealii dot org" <gcc-bugzilla at gcc dot gnu dot org>
- To: gcc-bugs at gcc dot gnu dot org
- Date: 17 Mar 2004 21:18:11 -0000
- Subject: [Bug c/14621] libm tan failure for pi/2 argument
- References: <20040317202427.14621.gerald.evenden@verizon.net>
- Reply-to: gcc-bugzilla at gcc dot gnu dot org
------- Additional Comments From bangerth at dealii dot org 2004-03-17 21:18 -------
Of course tan(pi/2) is infinite. However, M_PI_2!=pi/2 (the former
being a finite precision floating point number, the latter a
transcendental non-rational number; it should be obvious from
these characterizations that there is no way they are equal).
That a range error _may_ occur is not a mandate that it _must_ occur.
In particular, the man pages on linux do not say anything about setting
errno at all. And still, even if they did, the implementation of tan()
is not under the control of gcc, but is part of libc. Please complain
to them if you still believe that you are right. You better guard against
people citing the C99 standard, though, which also does not say anything
about errno in section 7.12.4.7.
W.
--
What |Removed |Added
----------------------------------------------------------------------------
Status|UNCONFIRMED |RESOLVED
Resolution| |INVALID
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=14621