Re: Minor cleanups in stencil-integral.cc (issue 579630043 by address@hidden)

[dak]

On 2020/04/24 21:33:12, Carl wrote:
> On 2020/04/24 21:19:57, dak wrote:
> > On 2020/04/24 21:18:12, dak wrote:
> > >
> >
>
https://codereview.appspot.com/579630043/diff/555740043/lily/stencil-integral.cc
> > > File lily/stencil-integral.cc (right):
> > > 
> > >
> >
>
https://codereview.appspot.com/579630043/diff/555740043/lily/stencil-integral.cc#newcode465
> > > lily/stencil-integral.cc:465: // more convoluted, but it's fairly
hot path.
> > > Sorry for not being clear: the question was not why this change
was
> effective
> > in
> > > saving time, but why it was valid.  When thickness is zero, you
only update
> > the
> > > upper skyline.  Why would the lower skyline no longer need
updating?
> > 
> > Well, other way round, but apart from that the question stands.
> 
> When thickness is zero, the upper and lower curves are the same. 
Either one
> completes the skyline.

Of course they are the same.  That is not the question.  The question is
why I am considering the identical curve for the minimum skyline while I
don't let it participate in the maximum skyline any more.

If everything is of thickness null, the maximum skyline will be
non-existent while the minimum skyline is there.  While the skylines
should be identical rather than only one of them being there.

https://codereview.appspot.com/579630043/