Re: [K3d-development] Empty Tracker [was:] Re: GSoC 2010 application rejected

[Bart J. <bar...@li...>]

On Fri, Mar 26, 2010 at 5:27 AM, Timothy M. Shead <ts...@k-...> wrote:
> Yeah, I have some doubts about this ... on deeper thought does it make sense
> to assume that points and attributes will always be interpolated in the same
> way?  I'm thinking that it doesn't - in general, a modifier has carte
> blanche to calculate attributes in whatever way "makes sense", which won't
> always be the same as new/existing points.

Nope, it doesn't, in general. I've refined the tests a little, so we
can decide for every test which point attributes need comparing. It
exposed bugs in all tests I've looked at so far :)

As part of these tests, I'm also outputting one point per line for the
arrays, since it's easier to tell the points apart that way, and the
line numbers can be used to find the point in the viewport. Maybe this
output method could be used for the mesh diff output as well? See i.e.

http://www.k-3d.org/cdash/testDetails.php?test=1093403&build=5083

vs.

http://www.k-3d.org/cdash/testDetails.php?test=1094063&build=5093

Cheers,

-- 
Bart