Re: Patch for collision_trimesh_trimesh_new available
"Bram Stolk" <[email protected]>
| Newsgroups | gmane.comp.lib.ode |
|---|---|
| Message-ID | <[email protected]> |
I'm very busy this week, but I will try to get it in as soon as I can. Thank you for the patch. bram On 9/14/07, Oleh Derevenko <[email protected]> wrote: > > Hi > > I've submited patch #1794781 > ============ > Fixed: Bug in hashing algorithm when program did not search for key match > if hash table bucket was full > Fixed: Inefficiency of hash value to index conversion when only lower 8 > bist of 32 bits available were used > Improved: Hashing algorithm itself having removed code with no effect from > it, having improved input value preprocessing quality and having changed it > to generate better bit distribution in output > Fixed: Bug when different contacts could be merged together because of > hash collision > Fixed: Bug when extra contacts were ignored if all contact slots were > already occupied instead of trying to merge them with some of matching > contacts already recorded > Fixed: Bug with least depth contact being selected during contact merging > instead of most deep one > Discovered: Bug with opposite contacts being reported as one contact with > dummy normal of (1; 0; 0) instead of being annihilated > Fixed: floating point types and immediate constants to match ODE style > > How to merge: > 1) Checkout revision 1216 > 2) Unpack archive in trunk/ode/src/ > 3) Update from SVN > 4) Resolve conflicts (if any) > 5) Commit to SVN > > QA Status: > Compiled but not tested (do not have possibility to). If you have any > standard tests it would be appreciated if they be run before committing > this. > ============ > > I've made corrections mostly to PushNewContact() function (replacement for > GenerateContact() in old code). I did not review mathematics part (did not > want to spend time trying to understand all that). However, considering the > quality of hash table implementation, I would not trust the author too much > and would suggest somebody who is familiar with math used there to review > the rest of the code (FindTriangleTriangleCollision()). > > More changes by me are coming in next few days. > > Oleh Derevenko > -- Yahoo ID: oleh_derevenko > -- ICQ: 36361783 > > _______________________________________________ > ODE mailing list > [email protected] > http://ode.org/mailman/listinfo/ode > -- Zapp: Captain's log, stardate...er.. Kif: Ohhh. April 13th. Zapp: April 13th. Point 2. _______________________________________________ ODE mailing list [email protected] http://ode.org/mailman/listinfo/ode