On Thursday 24 October 2002 2:55 pm, Andre Poenitz wrote:
> On Thu, Oct 24, 2002 at 10:51:07PM +0900, Rob Lahaye wrote:
> > "If implemented correctly", indeed. We have to think
> > carefully now and do all sorts of complicated checks to be
> > sure it's really approximately zero. All that will be so
> > much easier when only check of a negative value is needed.
Perhaps.
if (float_equal(val, 0.0, 0.05) || val < 0)
val = 0.0;
> ... and the intention would be clearer if there was a boolean
> flag 'use_scale' instead of using the magic (float!, no less)
> value 0.0 to indicate that.
Someone who shall remain nameless changed that on 22 August
I believe (judging from development/FORMAT). Oops. Does that I mean
I said who? ;-)
Angus
