Re: [Freeglut-developer] glutSolidCone() "reloaded"

[Richard R. <sf...@ol...>]

On Tue, Sep 30, 2003 at 07:05:48PM +1000, Nigel Stewart and Fiona Smith wrote:
> >>...there also appears to be
> >>a problem with the solid cone...
> >
> >Is this the same bug with the cone about the base being drawn?
> 
> 	I've identified a number of issues with glutSolidCone()
 [...]

Re-reading your code, I misunderstood how much you reduced the
code.  I really didn't appreciate it.  (Partly because I didn't
read your code that carefully, but got hung-up over things
like your use of assert().  (^&)


I think that if you move the table initialization to a separate
function so that the other shape sine/cosine-using shapes can
exploit the work (and hide all of the allocation, error-checking,
error messages, and array filling in the one function), this would
be good.

I'm skeptical about its value as an optimization, but I do agree
that it seems to be an improvement to the code.

I would add one point, however:


> 		z0 = z1; z1 += zStep;
> 		r0 = r1; r1 -= rStep;

IMHO, it's better to do a multiplication than a summation.
It's not really any faster to do this, and at least in principle,
numerical errors build up.  (In practice, as long as n is tiny,
no one will care.  (^&)


-- 
  "I probably don't know what I'm talking about."  http://www.olib.org/~rkr/