RFC: avoiding "inexact" from x86 ceil / floor / trunc

[Joseph Myers]

On Wed, 25 May 2016, Szabolcs Nagy wrote:

> > My inclination is that handling the case of not caring about "inexact" is 
> > better done in the compiler, in which case these functions can be inlined 
> > rather than exporting separate versions from glibc.  Unfortunately GCC 
> > doesn't currently inline these functions in all the cases where it could 
> > (or have an option to select TS 18661-1 semantics):
> > 
> 
> yes, if the compiler can handle this (and doesn't care about
> inexact by default for -std=c99/c11) that sounds acceptable.

GCC will now inline these functions without caring about "inexact" unless 
-fno-fp-int-builtin-inexact is used (where the default is 
-ffp-int-builtin-inexact, allowing "inexact" for noninteger arguments, for 
all currently supported standard versions; if C2x incorporates the TS 
18661-1 requirements, the default would then change for -std=c2x / 
-std=gnu2x).

-- 
Joseph S. Myers
joseph@codesourcery.com