special care is made to avoid any inexact computations when either arg
is zero (in which case th...

the error status is required to be sticky after failure of dlopen or
dlsym until cleared by dler...

(tgamma must be thread-safe, signgam is for lgamma* functions)

this is necessary so that we can freely add macro versions of some of
the math/complex functions...

the old formula atan2(1,sqrt((1+x)/(1-x))) was faster but
could give nan result at x=1 when the ...

DECIMAL_DIG is not the same as LDBL_DIG

type_DIG is the maximimum number of decimal digits that...

this has not been tested heavily, but it's known to at least assemble
and run in basic usage cas...

some software apparently uses this and breaks with musl due to
mismatching definitions...

the main practical results of this change are
1. the regex code is no longer subject to LGPL; it...

old code saved/restored the fenv (the new code is only as slow
as that when inexact is not set b...

fix comments about special cases

fix special cases, use multiplication instead of scalbnl

the fscale instruction is slow everywhere, probably because it
involves a costly and unnecessary...

this is a lot more efficient and also what is generally wanted.
perhaps the bit shuffling could ...

exponents (base 2) near 16383 were broken due to (1) wrong cutoff, and
(2) inability to fit the ...

zero, one, two, half are replaced by const literals
The policy was to use the f suffix for float...

Underflow exception is only raised when the result is
invalid, but fmod is always exact. x87 has...

Some code assumed ldexp(x, 1) is faster than 2.0*x,
but ldexp is a wrapper around scalbn which u...

Some long double consts were stored in two doubles as a workaround
for x86_64 and i386 with the ...

this could perhaps use some additional testing for corner cases, but
it seems to be correct.

up to 30% faster exp2 by avoiding slow frndint and fscale functions.
expm1 also takes a much mor...

The old scalbn.c was wrong and slow, the new one is just slow.
(scalbn(0x1p+1023,-2097) should g...

unlike some implementations, these functions perform the equivalent of
gcc's -ffloat-store on th...

unlike trig functions, these are easy to do in asm because they do not
involve (arbitrary-precis...

infinities were getting converted into nans. the new code simply tests
for infinity and replaces...

correctly rounded double precision fma using extended
precision arithmetics for ld80 systems (x87)

these are functions that have direct fpu approaches to implementation
without problematic except...

Simple wrappers around round is enough because
spurious inexact exception is allowed.

A faster workaround for spurious inexact exceptions
when the result cannot be represented. The o...

it's not even provided in the library at the moment, but could easily
be provided with weak alia...

untested; may need followup-fixes.

if all exception flags will be cleared, we can avoid the expensive
store/reload of the environme...

Note that the new fesetround has slightly different semantics:

Storing the floating-point envir...

this is necessary to support archs where fenv is incomplete or
unavailable (presently arm). fma,...

otherwise, the standard C lgamma function will clobber a symbol in the
namespace reserved for th...

standard functions cannot depend on nonstandard symbols

a double precision nan, when converted to extended (80-bit) precision,
will never end in 0x400, ...

long double and float bessel functions are no longer xsi extensions

presumably broken gcc may generate calls to these, and it's said that
ffmpeg makes use of sincosf.