Update:

On Tue, Sep 18, 2012 at 10:45 PM, Eric Decker <[email protected]> wrote:
>
> try using the 4.6.3 compiler.   3.2.3 is really old.

OK, since I'm on cygwin, I had to settle with the 4.5.3, patched thru
2012-02-28 from here
http://tinyos.stanford.edu/tinyos/dists/cygwin/

That made a difference: the conversion
int64_t a = x;
flot f = (float)a;
int64_t b = (int64_t)f;

apart from its inherent 24-bit mantissa rounding errors, now simply
"clips" the result to around +/- (2^32) = 2 x int32_t range.

How does this reproduce on 4.6.3?

Regards,
Flemming Nyboe
_______________________________________________
Tinyos-help mailing list
[email protected]
https://www.millennium.berkeley.edu/cgi-bin/mailman/listinfo/tinyos-help

Reply via email to