svn commit: trunk/uClibc: include libc/sysdeps/linux/common/bits etc...

Bernhard Reutner-Fischer rep.dot.nop at gmail.com
Mon Feb 16 17:21:11 UTC 2009


On Mon, Feb 16, 2009 at 06:02:20PM +0100, Carmelo AMOROSO wrote:
>-----BEGIN PGP SIGNED MESSAGE-----
>Hash: SHA1
>
>Carmelo AMOROSO wrote:
>> aldot at uclibc.org wrote:
>>> Author: aldot
>>> Date: 2008-12-22 11:31:12 +0000 (Mon, 22 Dec 2008)
>>> New Revision: 24487
>> 
>>> Log:
>>> - fix fpclassify, signbit, isfinite, isnan, isinf macros for long double math support
>>> - add rule to create preprocessor output for float- and long double math wrapper
>> 
>> 
>>> Modified:
>>>    trunk/uClibc/Makerules
>>>    trunk/uClibc/include/math.h
>>>    trunk/uClibc/libc/sysdeps/linux/common/bits/mathcalls.h
>>>    trunk/uClibc/libm/Makefile.in
>>>    trunk/uClibc/libm/e_scalb.c
>>>    trunk/uClibc/libm/float_wrappers.c
>>>    trunk/uClibc/libm/ldouble_wrappers.c
>>>    trunk/uClibc/libm/s_finite.c
>> 
>> Hi Berharnd,
>> I'm having some problem to build ntop application with recent uclibc.
>> ntop is trying to access 'finite' and 'isinf' symbols instead of
>> __finite and __isinf.
>> Before this commit we had both (finite/isinf as weak alias of
>> __finite/__isinf as latest glibc has indeed).
>> I'm wondering the rationale to have removed the alias.
>
>Indeed from the man page notes I can read
>
>NOTES
>Note that these functions are obsolete.  C99 defines macros isfinite(),
>isinf(), and isnan() (for all types) replacing them.  Further note that
>the C99 isinf() has weaker guarantees on the return value.  See
>fpclas-sify(3).
>
>so probably ntop heck should be updated by using properly macros.

I just saw your mail now, sorry. As you see those apps should use
the correct macros (this "new" set of macros was already mandated by
SUSv3, fwiw).


More information about the uClibc-cvs mailing list