I suspect that you're doing something silly that's taking all that time.

This was done with a simplistic collision detection, just like that which
you describe, with no clever bounding or other tricks.
I didn't count placement failures, I just chose a sensible number of trees
to be placed by doing a previous run with a #debug that logs the number of
successful placements.

The parse time was 27 seconds on a 850MHz machine.
--

Post a reply to this message

Attachments:
Download 'fillbox2.jpg' (51 KB)

Preview of image 'fillbox2.jpg'