Re: Floating point modulus

[Peter Stephenson <p.w.stephenson@xxxxxxxxxxxx>]

On Sat, 10 Jan 2015 23:10:17 -0800
Bart Schaefer <schaefer@xxxxxxxxxxxxxxxx> wrote:
> The following patch would instead silently substitute the fmod() call
> when using the % operator on a float.

Seems OK --- it avoids the very unusual implicit (at the shell rather
than C level) cast from float back to integer.  The only query is
whether we need to test for HAVE_FMOD, though the manual page I can see
suggests only the fmodf and fmodl variants we don't need are likely to be
problematic.

pws