Mercurial > hg > xemacs-beta
diff src/floatfns.c @ 4865:6c0bb4d2c23a
Always use our rint(), for rounding consistency with the bignum code.
ChangeLog addition:
2010-01-16 Aidan Kehoe <kehoea@parhasard.net>
* configure: Regenerate, now we no longer look for rint().
src/ChangeLog addition:
2010-01-16 Aidan Kehoe <kehoea@parhasard.net>
* config.h.in:
* floatfns.c (emacs_rint):
Don't look for rint in configure, always use our own
implementation that rounds to the even number in the case of
ambiguity, for consistency with the bignum code.
| author | Aidan Kehoe <kehoea@parhasard.net> |
|---|---|
| date | Sat, 16 Jan 2010 19:04:52 +0000 |
| parents | f31c12360354 |
| children | 6ef8256a020a 19a72041c5ed |
line wrap: on
line diff
--- a/src/floatfns.c Thu Jan 14 20:55:10 2010 -0500 +++ b/src/floatfns.c Sat Jan 16 19:04:52 2010 +0000 @@ -26,7 +26,6 @@ Define HAVE_INVERSE_HYPERBOLIC if you have acosh, asinh, and atanh. Define HAVE_CBRT if you have cbrt(). - Define HAVE_RINT if you have rint(). If you don't define these, then the appropriate routines will be simulated. Define HAVE_MATHERR if on a system supporting the SysV matherr() callback. @@ -50,11 +49,8 @@ #include "syssignal.h" #include "sysfloat.h" -/* The code uses emacs_rint, so that it works to undefine HAVE_RINT - if `rint' exists but does not work right. */ -#ifdef HAVE_RINT -#define emacs_rint rint -#else +/* An implementation of rint that always rounds towards the even number in + the case of ambiguity. */ static double emacs_rint (double x) { @@ -65,7 +61,6 @@ r += r < x ? 1.0 : -1.0; return r; } -#endif /* Nonzero while executing in floating point. This tells float_error what to do. */