This is the mail archive of the
libc-hacker@sourceware.cygnus.com
mailing list for the glibc project.
Re: libm-i387 bug?
- To: libc-hacker@cygnus.com
- Subject: Re: libm-i387 bug?
- From: Zack Weinberg <zack@rabi.phys.columbia.edu>
- Date: Thu, 16 Jul 1998 10:57:39 -0400
On 16 Jul 1998 07:12:47 -0700, Ulrich Drepper wrote:
>Andreas Jaeger <aj@arthur.rhein-neckar.de> writes:
>
>> >>>>> Zack Weinberg writes:
>>
>> > This test program prints "-1.949464 0.000000". Shouldn't it print
>> > "0.000000 1.000000" ?
>> It should - but the problem is even more obscure. The appended
>> program prints:
>
>Which platform and which gas? I've fixed yesterday problems with the
>latest gas snapshot and the i686 implementation of fdim.
$ uname -ms
Linux i586
$ gcc -v
Reading specs from /usr/lib/gcc-lib/i486-linux/egcs-2.90.29/specs
gcc version egcs-2.90.29 980515 (egcs-1.0.3 release)
$ as -v
GNU assembler version 2.9.1 (i486-linux), using BFD version 2.9.1
$ ./a.out
-1.949464 0.000000
8192.000000 6144.000000
6144.000000 8192.000000
0.000000 6144.000000
0.000000 6144.000000
$
where a.out is AJ's test program. The C library is CVS from July 5 or
so.
If I turn optimization on I get different but still erroneous output.
zw