Re: gcc-4.2.2 - wrong result of tanh - was (Re: octave-2.9.17 from source - make check fails - please help!)

[juras]

Rafael Laboissiere wrote:
> 
> It is not an issue in gcc, but in libc.  This has been already discussed
> elsewhere [1].  What version of libc are you using?  This problem has been
> fixed over two years ago.
> 
> [1]
> http://www.mail-archive.com/address@hidden/msg91789.html
> 
> Cheers,
> 
> -- 
> Rafael Laboissière
> 
> 

Thank you
I have glibc-2.3.6.
If I understand correctly the definition of tanh is in temlated library.
Apart from that, I compiled this simple test program with gcc-3.3.6, which I
have in /opt directory and guess what? - It works OK:

address@hidden:~/download/numerics/test$ which g++
/opt/gcc-3.3.6/bin/g++
address@hidden:~/download/numerics/test$ g++ -v
Reading specs from /opt/gcc-3.3.6/lib/gcc-lib/i686-pc-linux-gnu/3.3.6/specs
Configured with: ../gcc-3.3.6/configure --prefix=/opt/gcc-3.3.6
--enable-shared --enable-languages=c,c++ --enable-threads=posix
Thread model: posix
gcc version 3.3.6
address@hidden:~/download/numerics/test$ g++ tanh.cc -o tanh-336
address@hidden:~/download/numerics/test$ ./tanh-336
x=(0,1.5708), tanh(x)  = (0,1.63318e+16)
x=(0,1.5708), sinh(x)  = (0,1)
x=(0,1.5708), cosh(x)  = (6.12303e-17,0)
x=(0,1.5708), sinh/cosh= (0,1.63318e+16)
x=(0,1.5708), sh/=ch   = (0,1.63318e+16)
x=(0,1.5708), sh/ch2   = (0,1.63318e+16)
address@hidden:~/download/numerics/test$  

-- 
View this message in context: 
http://www.nabble.com/octave-2.9.17-from-source---make-check-fails---please-help%21-tf4796444.html#a13727510
Sent from the Octave - Maintainers mailing list archive at Nabble.com.