>> Note that while you're adding the C99 math stuff, you
>> might want to fix up float.h, which is just wrong about long
>> doubles (see PR i386/38288).
> Right, I should have said no one has started committing C99 math
> functions.
(Reading this exchange reminded me that I promised fellow gcc
de
David Schultz <[EMAIL PROTECTED]> writes:
> Thus spake Mike Barcroft <[EMAIL PROTECTED]>:
> > No one has started work on any of the C99 math functions yet. I
> > think with the exception of the math functions we conform to C99.
>
> Actually, I hacked up some patches for fpclassify(), is*(), and
>
Thus spake Mike Barcroft <[EMAIL PROTECTED]>:
> No one has started work on any of the C99 math functions yet. I
> think with the exception of the math functions we conform to C99.
Actually, I hacked up some patches for fpclassify(), is*(), and
friends some time ago. But nobody was interested in
Mike Barcroft wrote:
> > > This is because we lack the
> > >
> > > long double fabsl(long double);
> > >
> > > in -lm and .
> >
> > OK, thanks for tracking it down. This looks like an important
> > omission that should be fixed for 5.0-R.
>
> No one has started work on any of the C99 math functio
Kris Kennaway <[EMAIL PROTECTED]> writes:
> On Tue, Oct 22, 2002 at 11:22:41AM +0300, Ruslan Ermilov wrote:
> > On Sat, Oct 19, 2002 at 07:54:01PM -0700, Kris Kennaway wrote:
> > > World is broken if you try and build with -fno-builtin:
> > >
> > > c++ -O -pipe -ggdb -fno-builtin -march=pentium3
libstdc++ is supposed to provide crude replacements for missing -l functions
in its libmath/stubs.c file. Apparently, missing fabsl is an omission and
should be fixed in FSF sources.
--
Alexander Kabaev
To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-current" in the body
On Tue, Oct 22, 2002 at 01:28:56PM -0700, Terry Lambert wrote:
> Stefan Farfeleder wrote:
> [ ... fabsl() ... ]
> > > What standard defines this thing, which g++ has as a built-in?
> > >
> > > Alternately, the use could avoid adding the "-fno-builtin", and
> > > the problem would go away.
> >
> >
Kris Kennaway wrote:
> Yes, because as it stands this function is supplied by gcc as a
> built-in function, and the source tree will not compile without it
> (e.g. with a non-gcc compiler).
>
> > Alternately, the use could avoid adding the "-fno-builtin", and
> > the problem would go away.
>
> -f
Stefan Farfeleder wrote:
[ ... fabsl() ... ]
> > What standard defines this thing, which g++ has as a built-in?
> >
> > Alternately, the use could avoid adding the "-fno-builtin", and
> > the problem would go away.
>
> ISO C99
>
> 7.12.7.2 The fabs functions
> Synopsis
> #include
> doubl
On Tue, Oct 22, 2002 at 12:40:38PM -0700, Terry Lambert wrote:
> Is it?
Yes, because as it stands this function is supplied by gcc as a
built-in function, and the source tree will not compile without it
(e.g. with a non-gcc compiler).
> Alternately, the use could avoid adding the "-fno-builtin",
On Tue, Oct 22, 2002 at 12:40:38PM -0700, Terry Lambert wrote:
> Kris Kennaway wrote:
> > > > /usr/src/contrib/gperf/src/bool-array.icc:81: warning: rand() does not produce
>high-quality random numbers and should not generally be used
> > > > /usr/obj/usr/src/i386/usr/lib/libstdc++.so: undefined r
Kris Kennaway wrote:
> > > /usr/src/contrib/gperf/src/bool-array.icc:81: warning: rand() does not produce
>high-quality random numbers and should not generally be used
> > > /usr/obj/usr/src/i386/usr/lib/libstdc++.so: undefined reference to `fabsl'
> > > *** Error code 1
> > >
> > This is because
On Tue, Oct 22, 2002 at 11:22:41AM +0300, Ruslan Ermilov wrote:
> On Sat, Oct 19, 2002 at 07:54:01PM -0700, Kris Kennaway wrote:
> > World is broken if you try and build with -fno-builtin:
> >
> > c++ -O -pipe -ggdb -fno-builtin -march=pentium3
>-I/usr/src/gnu/usr.bin/gperf/../../../contrib/g
On Sat, Oct 19, 2002 at 07:54:01PM -0700, Kris Kennaway wrote:
> World is broken if you try and build with -fno-builtin:
>
> c++ -O -pipe -ggdb -fno-builtin -march=pentium3
>-I/usr/src/gnu/usr.bin/gperf/../../../contrib/gperf/lib -I/usr/src/gnu/usr.bin/gperf
>-o gperf bool-array.o gen-perf.
14 matches
Mail list logo
