all: remove 'constant in conditional context' from LINTFLAGS

Lint no longer emits this message, as it produced too many false
positives.

libm: Add missing fma(3) and variants for hppa.

Fixes a build error on pkgsrc/lang/python312. Ok'ed by skrll@.
Should be pulled up to netbsd-10.

Pull in missing functions for MKSOFTFLOAT.

move a COPTS setting after bsd.own.mk.

the riscv64 compat build trips on this saying:
bsd.rv32.mk" line 25: while evaluating variable "COPTS": while evaluating condition " == "gcc"": Bad condition

branches: 1.238.2;
libm: Fix arm expected symbol list harder.

LIB is not defined by the time we need it in the m.expsym rule, so
just say m here like in the other lines.

libm: Fix mips and arm expected symbol lists.

Had forgotten that they were overriding LIB_EXPSYM. Use a new
intermediate variable EXPSYM_SRC_MD for that.

libm: Factor out common expected symbol list.

Should substantially reduce the maintenance burden.

libm: Fix powerpc64 build.

- Include fenv.c and fma(3) symbols (which just use the FMADD
instruction).
- Note the .FN symbols in libm for the asm functions. The FN symbols
point at the function _descriptors_; the .FN symbols point at the
first instruction of the function.

XXX Unclear why we have the .FN symbols for asm functions but not for
C functions. I'm not sure we should be exporting them.

libm: Add frexpf and frexpl on VAX.

These are trivial subroutines, not symbol aliases, for separate
reasons:

- frexpf has a different ABI from frexp (float vs double argument)

- frexp is defined in libc, not libm, so although long double is the
same as double, frexpl can't be an alias in libm of a symbol
defined in libc

libm: Split expected symbols in mips further by 32-bit vs 64-bit.

32-bit has binary64 long double, same as double; 64-bit has binary128
long double, which is implemented with a few more symbols in libm
(which should maybe be hidden internal symbols, but let's get this
diagnostic measure in the build working before we think about
possibly deleting private symbols).

libm: Split expected symbols on arm and mips by hard/soft float.

libm: long double is just double on hppa, so don't use s_rintl.c.

PR misc/58054

libm: No s_rintl.c on mips32.

This is only for machines where long double is not double.

Not 100% sure this is the right criterion, but there's a good chance
it is.

libm: long double is just double on arm, so don't use s_rintl.c.

PR misc/58054

libm: New s_rintl.S for x86.

Prompted by PR lib/58054.

libm/s_logl: suppress lint warnings

Seen on sparc64.

sinhf is in the same manual page as sinh not sin

Add the long double math functions recently added from FreeBSD to their related man pages

Bring in more long double functions from FreeBSD.

Fix building with M68040 and M68060 options.

libm Makefile: add fma() functions for mips.

This pushed the ramdisk size, so that needed to be bumped as well.
Otherwise, there were no build issues for either 32 or 64 bit mips.
Sort-of OK'ed by simonb@.

libm: Add fma(3) and friends to m68k

libm: Add missing fma(3) and friends for sh3.

libm: suppress lint warning about '>>' on signed integer

GCC has implementation-defined behavior for that case.

branches: 1.218.2;
Fix vax build.

Add sincos{,f,l} from FreeBSD

Use "fmadd" to implement fma(3) on ppc.

libm: build fma(3), fmaf(3), fmal(3) on powerpc.

For now this builds the C version, for completeness, so
that e.g. lang/ocaml can be configured & built. However,
googling reveals that powerpc does have a "fused multiply add"
instruction, ref.
https://www.ibm.com/docs/en/aix/7.1?topic=set-fmadd-fma-floating-multiply-add-instruction
so this could probably be taken advantage of for a more
optimized version.

libm: build fma(3), fmaf(3), fmal(3) on alpha

Alpha has no hardware-assisted fused multiply-add, so the fallback
C implementation must be used. A basic smoke test shows the
implementation works for
"x * x - 1.0f" vs "fmaf(x, x, -1.0f)" with "volatile float x = 1.0008f;"

Use ${MACHINE_MIPS64}

Update LLVM to 10.0.0git (01f3a59fb3e2542fce74c768718f594d0debd0da)

Add manpage links for frexpf and frexpl.

Manpage links were missing for frexpf and frexpl even thogh they were
mentioned in the manpage for frexp.

branches: 1.210.2;
Add a C99 symbol to libm: nexttowardl

It's an alias for an already existing symbol nextafterl.

Patch obtained from <mgorny>

Detected by the LLVM buildbot breakage in tests.

document cosl with MLINKS and in the man page

also want s_nexttoward.c on riscv. now tests/lib/libm links.

we need rintl (long double) to link binaries

Default our libm to -frounding-math, if built with GCC.

The long-standing GCC default is to not respect rounding mode.
it looks like GCC 7 optimizes rint to a builtin, causing our few
rounding mode tests to fail.

Fixes PR port-amd64/54000: FP tests failing on amd64 since gcc7 import

Disable x87 implementations of sin, cos, tan.

The x87 hardware uses a bad approximation to pi for argument
reduction, and consequently yields bad answers for inputs near pi or
pi/2.

Tweak one tanf test whose doubly rounded output is a little too far
from the correct answer on the doubly rounded input.

branches: 1.204.2;
Disable sanitizers for libm and librt

The sanitizers are designed to use pristine OS versions of these libraries,
install interceptors for public symbols in them and link with them in every
sanitized executable.

Having unmodified base libraries (c, m, rt, pthread) also makes the system
usable with programs linking with them and without sanitization, as there
are no fatal issues during the resolving of missing symbols.

The remaining libraries like libutil, libintl etc are prebuilt with
sanitization and all their users in the MKSANITIZER=yes distribution are
required to use the same sanitizer.

libc and libpthread are already marked with the NOSANITIZER flag.

branches: 1.203.2;
Add man page link for hypotl.3

add nexttoward for mips64 and aarch64

Actually... s_nexttowardf.c appears to be usable on soft-float as well;
it was for arm earlier and briefly for all mips, so just move it into
COMMON_SRCS.

Move the inclusion of s_nexttowardf.c into a common test for
MKSOFTFLOAT=="no". This brings us nexttowardf() for sparc, at least.

add s_rintl.c for aarch64. ok by ryo@.

In the non-softfloat m68k part, add s_nexttoward.c and s_nexttowardf.c,
so that we get nexttoward() and nexttowardf() implemented here as well.

rintl is required for archs that support long double.

Move s_nearbyint.c out of the cpu ifdefs and into COMMON_SRCS,
which is used everywhere except on vax. This is to simplify and
to ensure the function is included in the build, allowing us to
re-add the nearbyint() test.

The sparc port needs s_nearbyint.c as well.

Non-softfloat m68k needs s_rintl.c.

Add s_nearbyint.c to alpha, ia64, riscv, mips, powerpc, and sh3 ports.

Add some blank lines to separate the different CPU archs,
for a minimal readability improvement.

branches: 1.191.2;
Replace our rounding functions with the ones from FreeBSD

branches: 1.190.2;
sun2 just has to be different, of course.
it doesn't use the libc softfloat and thus doesn't set MKSOFTFLOAT=yes.
enable the fenv functions here explicitly for now.

provide a common softfloat fenv implemenation and use it for softfloat builds.
restore ABI compatibility with previous releases for ieeefp.h on sh3.
add namespace.h protection for all the fenv interfaces.
use MKSOFTFLOAT on sh3 instead of assuming softfloat.
standardize on comparing MKSOFTFLOAT with "no".
remove the arm-specific softfloat fenv code (which also had several bugs).
fix logic errors in the arm hardfloat feraiseexcept() and feupdateenv().

Add fenv for ia64, imported from FreeBSD, compile tested only.

branches: 1.187.2;
more stuff needs -mfpu=vfp

deal with namespace issues and missing atanhf for vax.

Actually get fma(3) and friends into ARM libm.

skip buggy assembly for libm_g.a

fix double merge.

do the FENV test consistently

more fenv for sh

fix ilogb*

add ilogb.

fenv.h for alpha

branches: 1.177.2;
No fenv for m68k with SOFTFLOAT for now - so at least it builds.

Added ldexpl to ldexp manpage since it is missing.
Also added ldexpf manpage link since it is missing.

Add fabsl to the manpage describing the fabs functions.

Add sqrtl

Add mips fenv.c (From FreeBSD)

Powerpc fenv.c from FreeBSD

Add manpage to the fma function in libm.
The manpage is taken from the FreeBSD project since the fma implementation is
based on the one in FreeBSD.

Add entry for nexttowardf(3) and a link for the function name.
Bump manual page date.

Add entries for rintl(3) and the nearbyint functions, and
make manual page links to the new functions.
Bump manual page date.

fenv(3) support for hppa

PR/49399: Henning Petersen: Add man pages for erfl and erfcl

Add RISC-V MD bits.

Adding extra dependencies does not affect the inference rule search anymore.
XXX[1]: should that be the case?
Simplify the code anyway via selective path search.
XXX[2]: Split arch portions to arch specific files, this is a mess

branches: 1.164.2;
revert parts of 1.163 that affect the x86 build.

possibly, we want to build the i386 fenv.c for compat build, but they
seem largely compatible and for now, just go back to what we had before
that wasn't breaking the build or generating bug reports.

MKCOMPAT fixes for when compat MACHINE_CPU != normal MACHINE_CPU

Add -mfpu=vfp for fenv.c, since it gets those instructions from
machdep/armreg.h

Changes to existing files to enable building AARCH64 userland.
evbarm64-el
This is clang only. While gcc4.8 supports aarch64, no netbsd support has
been written for aarch64 with gcc4.8.

Add modfl(3). From FreeBSD.

Re-enable alpha IEEE compiler options

Comment out the ieee fp flags for alpha for now - something (tm) is not
working right, see PR port-alpha/48782.

Tune COPTS for alpha

branches: 1.156.2;
add -g to LINTFLAGS

Include new/missing files for vax

Include compiler-rt in libc, libm and libkern.

Add various functions which need fenv support to arm.

Try to fix the build (again): move e_sqrtl.c to common soure, but
move the include of fenv.h into the #ifdef __HAVE_LONG_DOUBLE block
and also provide a -DHAVE_FENV_H from the makefile for archs that have
it - leaving out exception settings on architectures that do not provide
it.

e_sqrtl.c requires fenv.h, so only build it on platfomrs supporting that

Add cbrtl(3) and sqrtl(3), from FreeBSD.

truncl(3)

Add FreeBSD's truncl and minimal testing.

Add some fake long double -> double wrappers for various transcendental
functions. They are expected to be implement properly at a later time.

Initial version of fmodl from FreeBSD.
Basic test case for the fmod family.

Man page entries for ceill, floorl and roundl.

Add initial version of ceill, floorl and roundl from FreeBSD.

Hook complex operation support into libm when not using libgcc.

Update LLVM/Clang snapshot to r191105. This brings in a better detection
of unused static variables and functions. Disable this for some external
code and for ioconf.c in the kernel.

s_nexttoward.c does not make sense compiling for 32 bit sparc (double ==
long double), so ifdef it out in the sparc64 compat library build. I hope
the nextafter alias is set up correctly (where?) - and someone please
suggest a better way for this if clause.

Add nexttoward, nearbyint and rintl for sparc64

Add missing .go and fix lint rule.

Fix POLYD/Mariah fallout and include the scalbn sources (which do not
use polyd).
Fix global names and weak aliases.

Use MKSOFTFLOAT = no for including arm hardfloat files.
When arm and MKSOFTFLOAT = yes, make explicit dependency to .c files
so .S won't be used.

Drop s_scalbnl.c (and s_scalbln.S as side effect) and provide scalbnl in
s_scalbln.c.

Retire ldexp sources. scalbn has the aliases.

Pass a real file as MLINKS target; fix unnecessary (re)install.

Use __HAVE_LONG_DOUBLE instead of EXT_EXPBITS and include s_nextafterl.c
as well as s_frexpl.c in the global list of "common" sources - as we
seem to have no arch redefining the standard double variant, it seems
unlikely we will have on redefining the long double version in assembler.

any arch that has fenv.[ch] can have fma as well (since fma uses fenv).

Add VFP versions of lrint/lrintf/lrintl for earmhf*

add more functions required by fma

hook fma to the build

Hook to the build.

Make this work for NetBSD

Add support for the ARM hardfloat instructions vsqrt and vabs.

branches: 1.125.2;
Use COPTS instead of CFLAGS

Add tgamma{,f} from FreeBSD via rudolf, netbsd at eq dot cz

Disable new -Wstring-plus-int warning where needed for now.
Adjust various places that add GCC-only options to check for the active
compiler first.

These directories default to WARNS?=5

On alpha compile the math library with dynamic rounding mode instead
of fixed normal IEEE rounding mode. This makes function that depends on
the current rounding mode, such as rint(3), just work.

branches: 1.120.2; 1.120.4;
Split out pow(3) from exp(3).

Split out the logarithm functions from the perplex exp(3). Also remove some
useless notes in the latter (namely, no one cares that financial calculations
were done on Hewlett-Packard HP-71B, or that there is a LN1() in Pascal).

Deprecate ieee(3).

Split out the scalbn(3) family out from ieee(3) and properly document it.

Add logbl(3).

Add ilogbl(3).

Add scalbnl(3) mlinks.

Add scalbn{,f,l} for VAX.

Add scalbnl to libm.

remove most of the remaining HAVE_GCC tests that are always true in
the modern world.

branches: 1.110.2;
Add fenv support for sparc. Mostly copied from sparc64 and share with it.

Remove trailing '\'.

Split out the ilogb(3) family out of ieee(3) and properly document it.

Merge floor(3) to ceil(3).

Split out from ieee(3) also the remainder(3) and copysign(3) families.

Split the nextafter(3) family of functions to their own page and properly
document these.

remquo{,f} from FreeBSD via Stathis Kamperis

not everyone has fenv.c, pointed by Havard Eidnes

fenv support for sparc64 from Stathis Kamperis

branches: 1.101.2;
Remove NOMANDOC, the documents are at least parsed somewhat sensible.

Imlementations of fmax, fmaxf, fmin and fminf libm functions for VAX.

Add noieee versions for lround(), lroundf(), lrint(), lrintf(),
llrint() and llrintf(). Code copied from round(), roundf() and
rint() and modified for return values. Its possible this may not
do the right things in edge cases, but if so its likely to have
the same issues as the existing round(), roundf() and rint().

All this used by vax (only), and should allow xnest to complete
build.

Commit SoC long double support from Stathis Kamperis

two disgusting hacks:
-mk/bsd.lib.mk picks up a .S asm file behind our back (did it do so
always?). s_modf.S is incorrect; I'm undecided whether it makes sense
to fix it, so add a stupid rule to enforce the .c file to be used.
-on i386, s_tanh.c gets miscompiled. It works with -O0, so add
COPTS for exactly that file.

Fix typo in MLINKS, from Stathis Kamperis. Adapt set lists, and run
them through sort while there.

Add support for fenv.h interface for i386 and amd64.

Submitted by Stathis Kamperis as part of GSoC 2010 and ported from
FreeBSD.

-build ldexp/frexp/modf locally again rather than relying on libc
reaching over, to get things into a sane state,
-do the usual namespace protection game with scalbn(f) because it
is used internally
(more symbol renamings are necessary)

The weak_alias change for mc68881 is untested.

build s_frexp.c, s_ldexp.c and s_modf.c here for new platforms, since they
don't exist as compat functions in libc.

current list of new platforms: powerpc64.

from dennis.c.ferguson@gmail.com in PR#43042. idea to not put compat
functions into new platforms from drochner@netbsd.

Add exp2 and exp2m

Disable the use of mandoc for now until support for conditional
expression exists.

add f{min,max,dim}{,l,f} from FreeBSD

Change the use of formally undocumented features, which have now been
made to fail. Specifically, change
.ifdef(SYMBOL) -> .ifdef SYMBOL or .if defined(SYMBOL),
and corresponding for .ifndef.

Also correct one error in lib/libm/Makefile (.ifdef (${MKCOMPLEX} != "no")?!?).

branches: 1.88.4; 1.88.8;
Wrap compiler-specific flags with HAVE_GCC and HAVE_PCC as necessary. Add a few flags for PCC.

Check MKCOMPLEX variable to build complex support into libm.

- Fold libm387 into libm.
- Leave libm387 as an empty shell.

branches: 1.85.2;
Remove 3 functions that exist in libc. This does not break binary compatibility
since we always link with libc.

Put back USE_SHLIBDIR and remove the confusing lint comment that made me
comment it out in the first place.

avoid shadow variables, and enable lint.

branches: 1.82.4;
Add manpages for the complex math functions, originally from SUSv3,
roff source from the Linux documentation project.
Modifications before import:
-added NetBSD RCS ID
-removed Linux PROLOG and declarations with "long double"
-ran the "deshallify" script as required by The Open Group
Split out complex related things into an own Makefile fragment.
Thanks to hubertf for directions.

make all sun2 use -O0 and move most of the hacks out into just 3 files.

make sun2 build again with liberal use of -O0.

branches: 1.79.2;
Add C99 complex support, for double and float.
Most complex function implementations are from the "c9x-complex" library,
originating from the "cephes" math library, see
http://www.netlib.org/cephes/, from Stephen L. Moshier, incorporated and
redistributed with the NetBSD license by permission of the author.

Error behaviour and other boundary conditions (branch cuts)
need to be looked at.

For namespace sanity, I've done the rename/weak alias procedure to
most of the exported functions which are also used internally.
Didn't do so for sin/cos(f) yet because assembler implementations use
them directly, and renaming functions shared between the main libm
and the machine specific "overlay" might raise binary compatibility
issues.

-get cabs() and cabsf() out of public view - these are historical and
conflict with C99 functions which are builtins in newer gcc
(actually, the old cabs() is ABI compatible with the new _complex one
on i386, but this is purely accidental)
remove public prototypes and manpages, move the code into a compat
subdirectory as libc does so that binary compatibility is kept
-add a manpage for the isgreater() etc macros, borrowed from FreeBSD

move manpages for frexp/ldexp/modf to libm -- the (most) implementation is
already there

noieee src is not strict-aliasing compliant. apply -fno-strict-aliasing.

build w_cabs.c/w_cabsf.c with -fno-builtin-cabs[f]

XXX: our cabs()/cabsf() are different to others.

Add trunc() and truncf() C99 functions, from FreeBSD.

Build the sun versions of frexp(), ldexp() and modf().
While the versions in libc/gen are prettier to look at, having single
precision versions in libm and double precision versions in libc causes
a lot of confusion, and to favor libm gets us coprocessor support.

Add nan(3), nanf(3), and nanl(3); welcome libm.so.0.4.

C versions of non-IEEE round() / roundf(). Alas, using the VAX
cvtrdl; cvtld doesn't work for big numbers.

Implement round(3), so that jot(1) can link again.

Add man page for l{,l}rint{,f}(3), very loosely based on the FreeBSD one.
Requested by martin@
Reviewed by drochner@

Add i387 FPU routines for log2(3) and log2f(3).

ok christos@

Add wrappers for the newly added functions. Thanks Klaus!

Add log2 and log2f from FreeBSD PR 83845

branches: 1.65.2;
Add a conditional for proper m68k softfloat operation. It looks like I
forgot to commit this file back when this functionality was added.
Noted by Bruce O'Neel (the original author) in port-mac68k@.

lib/24252: Add C99 functions round(3) and roundf(3).
libm minor unchanged; ride a recent bump.

From FreeBSD.

also provide an alpha assembler implemtation of lrint()

enable lrint.S for amd64 (which doesn't build libm387),
and tighten the generic->machdep source distribution a bit to avoid
unwanted substitutions

being here, add the (l)lround(f) functions to libm
(almost the same like lrint)

Implement (l)lrint(f).
(Could get some optimization for LP64; this would need a change to
the float access macros in math_private.h.)

branches: 1.59.2;
Compile libm on VAX at WARNS=2; leave other architectures at WARNS=1

branches: 1.58.2;
Drink ALL the Kool-Aid.

(Move -mieee from lib/ to bsd.*.mk; new var MKIEEEFP defaults to yes.)

On Alpha, build libc and libm with -mieee ... A metric crapload
of 3rd party software expects this.

XXX TODO: consider libm-fast for alpha that is not -mieee. Possibly
determine -mieee or not at run-time based on a bit in the executable
headers.

* Add user-controlled mk.conf variables
- SHLIBDIR Location to install shared libraries if ${USE_SHLIBDIR}
is "yes". Defaults to "/usr/lib".

- USE_SHLIBDIR If "yes", install shared libraries in ${SHLIBDIR}
instead of ${LIBDIR}. Defaults to "no".
Sets ${_LIBSODIR} to the appropriate value.
This may be set by individual Makefiles as well.

- SHLINKDIR Location of shared linker. Defaults to "/usr/libexec".
If != "/usr/libexec", change the dynamic-linker
encoded in shared programs

* Set USE_SHLIBDIR for libraries used by /bin and /sbin:
libc libcrypt libcrypto libedit libipsec libkvm libm libmi387
libtermcap libutil libz

* If ${_LIBSODIR} != ${LIBDIR}, add symlinks from ${LIBDIR}/${LIB}.so*
to ${_LIBSODIR}/${LIB}.so* for compatibility.

* Always install /sbin/init statically (for now)


The net effect of these changes depends on how the variables are set:

1.) If nothing is set or changed, there is no change from the
current behaviour:
- Static /bin, /sbin, and bits of /usr/*
- Dynamic rest
- Shared linker is /usr/libexec/ld*so

2.) If the following make variables are set:
LDSTATIC=
SHLINKDIR=/lib
SHLIBDIR=/lib
Then the behaviour becomes:
- Dynamic tools
- .so libraries used by /bin and /sbin are installed to /lib,
with symlinks from /usr/lib/lib*so to -> /lib/lib*so
where appropriate
- Shared linker is /lib/ld*so

3.) As per 2.), but add the following variable:
USE_SHLIBDIR=yes
This forces all .so's to be instaleld in /lib (with compat
symlinks), not just those tagged by their Makefiles to be.
Again, compat symlinks are installed

Readability cleanups; MKfoo=no -> NOfoo.

explicitly set to WARNS?=1

Use arch/i387 for x86_64.

branches: 1.52.4;
If we are not including i387 support in libm, build a separate libi387.

branches: 1.51.4;
Use arch/m68k copies of the common functions even for mc68881 hardware.

Pull in Motorola 68060 floating point library support package based
libm code, .ifdef M68060.

Document isinff(3) and isnanf(3), cloned from the corresponding double-
precision manual page. It is kept separately since the latter are part of
libc and thus impose different linkage instructions.

Document reentrant log gamma interfaces.

Add single-precision versions of interfaces to their manual pages' NAME section
and install links as appropriate.

More trailing white space.

Quick and dirty hook to compile a non-6888[12] libm for 68040 and 68060.
This is switched to by the M68040 or M68060 defines.
XXX Eventually, a M060SP based libm should be built for M68060.

NOLINT -> MKLINT=no.

Fix PIC code and fix problem with obj dirs.

Actually build __setjmp.S/__sigsetjmp.S for vax. PICifiy the setjmp code.
Fix libm Makefile to handle obj directories.

The Mariah chip lacks the POLYD instruction, so some assembly coded
calculations must be done in C instead. Fast fix until the POLY{D,F,G,H}
insn's are written.

on i386, use FPU instructions in libm if I387_LIBM is set, from dan@geek.com.au

move NOLINT definition

Make sure the reason that NOLINT is set is documented. Additionally
in places where it should be set (or has to be set right now), do so.

use CPPFLAGS instead of CFLAGS

branches: 1.36.2;
kill tab on line by itself, keep contents within 80 columns

rcsid facism, WARNSify

force NOLINT on the Alpha, at least until we fix FP underflow/denorm handling

Non-IEEE assembler files wasn't handled correctly. Trouble pointed out by
Robin Birch, fix by Mats O Jansson/Christos Zoulas.

Add a continuation slash in the commented-out i387 section

Add new i387 math functions to ARCH_SRCS

Enable mc68881 version of fmod

enable alpha-specific libm files

Changed w_gamma.c, w_gamma_r.c, w_gammaf.c, w_gammaf_r.c to call
__ieee754_lgamma_r or __ieee754_lgammaf_r.
Removed e_gamma.c, e_gamma_r.c, e_gammaf.c, e_gammaf_r.c, e_lgamma.c,
e_lgamma_f.c.

Got Makefile to understand to use non-IEEE files on vax systems.

Add i387 "float" functions to the makefile

Enable mc68881 versions of remainder() and scalb().

specify man pages the new way.

Added s_isinff.c

Float versions of math functions. From Ian Taylor (ian@cygnus.com), with
minor changes by me.

Update for mc68881 ceil(), floor(), & rint(), and s_sqrt.S -> e_sqrt.S
naming mistake.

Added a ieee_test(3) man page (Like SunOS) derrived from an earlier copy
of ieee(3). This man page describes logb, scalb, and significand, which
are really only useful for verifying IEEE754 compliance.

infnan was a vaxism.

Added i387 version of significand().

Christos Zoulas' Makefile fixes.

Add cabs() & drem() for BSD libm compatibility.

Update manpages for new math library.

Compiled shared too.

Comment out all functions that are also provided by our C library: fabs(),
frexp(), isnan(), ldexp(), and modf().

isnan() was provided by both C and Math libraries.

Changes for fdlibm based math library

use .S not .s

update for MACHINE_ARCH

Use BYTE_ORDER == LITTLE_ENDIAN instead of -Dnational to select byte order
of floating point numbers.

The assumption that a processor's floats use the same order as its integers
is probably unwarranted, but it is better than requiring non-national CPUs
to #define national, as that drags in national-specific code too.
(According to a comment in the code, the national FPU does not do infinity)

Install cbrt link to sqrt manpage.

No PIC, you probably don't want the overhead in these functions.

put gamma.c into libm (because it's missing)...

Split log() and log__D() into separate files. Add logtab.c for logarithm
tables and log.h for common definitions.
Split exp() and exp__D() into separate files.

New code from uunet.

Add RCS identifiers.

branches: 1.1.1;
Initial revision