This is the mail archive of the
gcc-bugs@gcc.gnu.org
mailing list for the GCC project.
Re: Floating point bug in egcs and gcc-2.8.1
- To: Andreas Jaeger <aj at arthur dot rhein-neckar dot de>
- Subject: Re: Floating point bug in egcs and gcc-2.8.1
- From: Jim Wilson <wilson at cygnus dot com>
- Date: Tue, 05 May 1998 21:15:27 -0700
- cc: egcs-bugs at cygnus dot com, bug-gcc at prep dot ai dot mit dot edu
egcs-2.91.26 with -O0: broken
egcs-2.91.26 with -O99: seems to work (killed it still running at count=2057)
The difference between a -O0 compile, and a -O1 compile, is that the -O0
version calls the glibc math routines, and the -O1 version uses inline
expansions for the math routines. Since only the -O0 version is failing, that
would indicate a bug in the glibc math routines. The bug probably only
occurs with out-of-range values.
If I compile with -O99 -D__NO_MATH_INLINES then the testcase also fails.
Further evidence that the problem is in the actual glibc math routines.
Of course, it could be that gcc miscompiled the glibc math routines, but
I have no easy way to debug that as I don't know how to compile glibc offhand.
Jim