Is libc C99 compliant?

[Steve Kargl]

On Thu, Jan 31, 2019 at 09:32:11AM -0700, Warner Losh wrote:
> On Wed, Jan 30, 2019 at 11:33 PM Steve Kargl <
> sgk at troutmask.apl.washington.edu> wrote:
> 
> > Should/can TARGET_LIBC_HAS_FUNCTION be updated to at least
> > default_libc_has_function?  More importantly now that libm
> > contains sincos[fl], should FreeBSD gcc config file be updated
> > to use
> >
> > bool
> > bsd_libc_has_function (enum function_class fn_class)
> > {
> >   if (fn_class == function_c94
> >       || fn_class == function_c99_misc
> >       || fn_class == function_c99_math_complex
> >       || fn_class == function_sincos)
> >     return true;
> >
> >   return false;
> > }
> 
> On its surface, this seems sane to me. what does this control? And is this
> for our ancient 4.2 or current gcc...
> 

AFAICT, it controls the use of C99 compliant functions as builtins.
Unfortunately and I should have remembered, FreeBSD's C runtime
libraries (ie libc+libm) are not C99 compliant.  The problem (for me)
is that function_c99_math_complex indicates that libm includes
a complete set of C99 complex math function, which of course
it doesn't.  Testing with GCC trunk gives

1 default_libc_has_function  (C99 compliant libc+libm)
2 no_c99_libc_has_function   (FreeBSD current setting)
                                  1            2
                === gcc Summary ===
# of expected passes            134923       134887
# of unexpected failures        171          207   <-- This is good.    
# of unexpected successes       27           27
# of expected failures          550          550
# of unresolved testcases       14           14
# of unsupported tests          2222         2222

                === g++ Summary ===
# of expected passes            124009       124009
# of unexpected failures        41           41
# of expected failures          548          548
# of unsupported tests          5585         5585

                === gfortran Summary ===
# of expected passes            48992        48993
# of unexpected failures        2            1     <-- This is bad.
# of expected failures          130          130
# of unsupported tests          88           88

To fix the "This is bad" requires improvements to libm.  The one
individual, who spent 15+ years trying to improve libm, was forced
to hand in his commit bit due to the fiasco of jenkins spamming
the -current and -stable mailing lists.   Mysteriously, a month or
so after that individual stopped contributed to libm, jenkins
stopped spamming the mailing list.

-- 
Steve