b2TimeOfImpact can give false positives when calling it manually. The
recent change you made culled a lot of these, but not all. I was working on
some test cases to give you when I stumbled onto a fix.
At lines 57 and 58 in b2TimeOfImpact.cpp you initialize xfA and xfB to the
sweep value from 0.0f, no matter what t1 was used. If you pass t1 in and
use that value instead, it works perfectly. (I've attached a patch with the
fix in case this was unclear.)
I ran into this working on the convex cast patch:
http://box2d.org/forum/viewtopic.php?f=3&t=4780
Original issue reported on code.google.com by NagelBa...@gmail.com on 17 Apr 2010 at 8:08
Original issue reported on code.google.com by
NagelBa...@gmail.com
on 17 Apr 2010 at 8:08Attachments: