Gcc builtin review: isinf, insnan ...

[Joseph Myers]

On Fri, 29 May 2015, Wilco Dijkstra wrote:

> > It's certainly the case that substantial GCC work is needed to make
> > floating-point properly follow even C99/C11 Annex F regarding exceptions /
> > rounding modes, let alone make signaling NaNs follow TS 18661-1, though
> > the signaling NaNs fixes would probably be more straightforward (reviewing
> > lots of code and making local changes, rather than any tricky design
> > issues involved for other floating-point areas).  While I think such
> > support would be desirable, it also seems there's limited demand for it.
> 
> IMHO support for different rounding modes, trapping exceptions and signalling
> NaNs are not required by many (most ARM CPUs don't even support FP traps...), 

-ftrapping-math (default) is nothing to do with traps in the sense of 
SIGFPE or change of flow of control, only about the raising of exception 
flags that can be tested with <fenv.h> interfaces.  Annex F and TS 18661 
also have nothing to with the traps (part 5 deals with alternate exception 
handling, but I doubt processor-level traps will be particularly useful in 
implementing it).  -frounding-math and -fsignaling-nans are not enabled by 
default.

-- 
Joseph S. Myers
joseph@codesourcery.com