Re: Changes in source for different distributions
Galen Hazelwood <galenh@micron.net> writes:
> libieee.a is a dummy library, AFAIK. The linux libc has been
> ieee-complient for a long time. The reason that bo's libc5 lacks it is
> because it isn't necessary.
I don't think this is quite true. At least in some of the software
projects I work on we have had to use
#ifdef LINUX_KLUDGES
#include <fpu_control.h>
__setfpucw(_FPU_IEEE);
#endif
to get the IEEE specified behaviour for floating point exceptions.
In /usr/include/i386/fpu_control.h on a "bo" system the definition of
_FPU_DEFAULT is different from _FPU_IEEE.
> Douglas Bates wrote:
> > BTW, can anyone tell me where I could find documentation on that
> > library? I don't see much in /usr/doc/libc6-dev/* referring to that
> > library. Does linking it simply force the floating point arithmetic
> > to comply the IEEE floating point standard? (This would be a "good
> > thing".)
I should be more specific on what I meant by my question. Does
linking with -lieee on an i386 system result in setting the floating
point unit control word to the _FPU_IEEE value? There seems to have
been some changes in /usr/include/fpu_control.h that indicate that
IEEE behaviour is now the default.
--
Douglas Bates bates@stat.wisc.edu
Statistics Department 608/262-2598
University of Wisconsin - Madison http://www.stat.wisc.edu/~bates/
--
TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to
debian-devel-request@lists.debian.org .
Trouble? e-mail to templin@bucknell.edu .
Reply to: