| 03:50.58 | *** join/#brlcad nsd (~nicholas@c-24-0-153-224.hsd1.pa.comcast.net) |
| 06:07.26 | *** join/#brlcad merzo (~merzo@152-234-132-95.pool.ukrtel.net) |
| 06:56.34 | *** join/#brlcad jordisayol (~jordisayo@unaffiliated/jordisayol) |
| 08:27.18 | *** join/#brlcad abhi2011 (~chatzilla@ip170-79-211-87.adsl2.static.versatel.nl) |
| 10:24.56 | abhi2011 | Getting some weird cases like this now : http://postimage.org/image/lkepvy2s/full/ |
| 10:25.08 | abhi2011 | its due to recalculation of the bb each iteration |
| 10:28.03 | abhi2011 | but should be gone when I start calculating the points using rt |
| 10:28.26 | abhi2011 | the challenge now is to generate contact points between colliding objects based on the same criteria that bullet does |
| 10:36.20 | abhi2011 | apparently it generates points exactly along edges or to maximize the area covered by them |
| 11:40.23 | CIA-48 | BRL-CAD: 03abhi2011 * r47008 10/brlcad/trunk/src/libged/simulate/ (simcollisionalgo.cpp simphysics.cpp simulate.c): Added code to check the collision points during collision processing |
| 13:20.45 | CIA-48 | BRL-CAD: 03abhi2011 * r47009 10/brlcad/trunk/src/libged/simulate/ (simcollisionalgo.cpp simulate.c simulate.h): Added contact manifold related members to the sim_params struct to communicate them back to libged and display them |
| 13:53.30 | brlcad | abhi2011: looking great! |
| 13:54.00 | brlcad | I expected that the bounding box recalc would cause "wrong" collision detection, but it's right ;) |
| 13:54.42 | brlcad | so bullet can basically do preliminary intersection calculations applying velocity and rotation adjustments on objects that aren't touching without problem |
| 13:55.00 | ``Erik | bullet does 'sweep' cd, iirc |
| 13:55.14 | brlcad | until two bounding boxes intersect, at which point rays have to be fired to determine whether they "actually" intersect yet (and where those points are) |
| 13:58.08 | abhi2011 | brlcad: yes, well bullet also calculates the same bounding box as we do at the start of each iteration, so the results should be ok |
| 14:00.45 | abhi2011 | yes, so after every iteration I am going to fire a bunch of rays and report upto 4 points for every case where objects touch |
| 14:01.01 | abhi2011 | these have to have the maximum possible volume |
| 14:01.17 | abhi2011 | which should be easy to get, just get the points with maximum x,y |
| 14:01.24 | abhi2011 | and minimum |
| 14:04.55 | abhi2011 | so where can i find some code to help me draw arrows in mged |
| 14:05.32 | abhi2011 | i need to draw the contact normals to check if they are being calculated correctly, then reproduce the same through rt |
| 14:06.07 | abhi2011 | the normals have to point correctly so that forces are apllied in the right direction |
| 19:43.47 | abhi2011 | thats strange , I am trying to setup a linked list of manifold points, but I get an error after the function that allocates memory for each node has completed |
| 19:44.06 | abhi2011 | #0 0x00007fffe2e6a6eb in malloc_consolidate () from /lib64/libc.so.6 |
| 19:44.08 | abhi2011 | #1 0x00007fffe2e6be14 in _int_malloc () from /lib64/libc.so.6 |
| 19:44.09 | abhi2011 | #2 0x00007fffe2e6ed99 in malloc () from /lib64/libc.so.6 |
| 19:47.25 | CIA-48 | BRL-CAD: 03abhi2011 * r47010 10/brlcad/trunk/src/libged/simulate/ (simcollisionalgo.cpp simulate.c simulate.h): Trying to pass contact manifolds via a linked list back to libged |
| 20:13.56 | abhi2011 | are there any issues with bu_malloc() in c++ |
| 20:27.29 | *** join/#brlcad merzo (~merzo@177-241-132-95.pool.ukrtel.net) |
| 20:48.54 | CIA-48 | BRL-CAD: 03abhi2011 * r47011 10/brlcad/trunk/src/libged/simulate/ (simcollisionalgo.cpp simulate.c): Fixed some memory bugs |
| 20:49.00 | abhi2011 | ah shoot, it was a bug in my code...casting aspersion on bu_malloc() unnecessarily :P |
| 20:54.45 | abhi2011 | ah frack!!!....had the arguments for VMOVE() reversed, was therefore pushing weird data into the bullet simulation instead of getting points out and got some extremely strange results.....get a grip abhi !!! |
| 20:55.23 | CIA-48 | BRL-CAD: 03abhi2011 * r47012 10/brlcad/trunk/src/libged/simulate/simcollisionalgo.cpp: fixed order of arguments in the VMOVE() which retrieve the collision points |
| 21:04.12 | CIA-48 | BRL-CAD: 03abhi2011 * r47013 10/brlcad/trunk/src/libged/simulate/simulate.c: More bug fixes, forgot to uncomment the memory cleanup routine |
| 23:37.54 | *** join/#brlcad piksi (piksi@pi-xi.net) |