Re: math package floating point exceptions

[James Walker <jamesw@...>]

Alex Davies wrote:
> Hi James,
>
> What that means is that your host app has floating point exceptions 
> enabled. By default, the fpu should just produce propagating NaNs and 
> Infs. Perhaps you're running inside Delphi or similar? Anyway, to 
> disable floating point exceptions.
>
> On an x86 where __asm{} allows you to type assembler, this'll do it:
>
> __asm {
>  finit;
> }
>
> (Reinitializes the fpu to its default state)

Yes, it is a Delphi app.  We had tried using _controlfp to mask out 
floating point exceptions, but the copy protection package we're using 
on release builds seems to prevent that from working.

--
  James W. Walker, Innoventive Software LLC
  <http://www.frameforge3d.com/>