The following warnings occurred: | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Warning [2] Undefined array key "lockoutexpiry" - Line: 94 - File: global.php PHP 8.1.31 (Linux)
|
bullet dynamics - Printable Version +- Forums (https://www.vdrift.net/Forum) +-- Forum: Project (https://www.vdrift.net/Forum/forumdisplay.php?fid=4) +--- Forum: Development (https://www.vdrift.net/Forum/forumdisplay.php?fid=9) +--- Thread: bullet dynamics (/showthread.php?tid=1262) |
- NaN - 05-09-2010 There are 42 tracks in vdrift-data repo. That's a lot of work. - NaN - 05-09-2010 I did a diff on interlagos roads.trk. The changes after your fixes affect internal bezier patch points P[1,0]-P[2,3]. These are generated procedurally to guarantee a smooth surface. Have there been changes to this algorithm at some time in the past? We could eventually simply recalculate them? - alex25 - 05-09-2010 the algorithm was changed at some point to accommodate for the selection of 2,3 or 4 vertices at a time (there is a thread here discussing this). but the problem is even worse. i just checked ring2007 and it has holes in the bezier surface and it was done with the latest version of the track editor so even re-tracing it might not fix the problem. - NaN - 05-09-2010 Did re-tracing help on other tracks? There has to be a bug in the ray cast algorithm then. - alex25 - 05-09-2010 it did actually and the surfaces are different (do a diff to see what changed). there could also be a bug in the ray cast algorithm. an easy track to test is rouen as the hole is right under the start/finish line (i guess you need to make the changes i posted to see it) - NaN - 05-09-2010 Looks bad. Will do some debugging. - NaN - 05-09-2010 The front wheel rays report no collision after crossing the start line. - alex25 - 05-09-2010 the patch is there (shows up in the track editor). actually that one is really easy to fix as this is the last patch. but it will be good to find out why the wheels report no collision. --alex-- - NaN - 05-09-2010 I hacked the racing line code to draw the patches. There is a gap. - NaN - 05-09-2010 To draw the patches: Replace lines 38-41 in roadpatch.cpp with: Code: MATHVECTOR <float, 3> v0 = patch.GetPoint(0,0); Enable racingline. - NaN - 05-09-2010 Strange. Ruudskogen also has a gap but it doesn't affect the car. - NaN - 05-09-2010 To see the separate patches: Modify lines 74-76: Code: uvs[5] = 1;//(v2-v1).Magnitude(); - NaN - 05-09-2010 OK. There is no gap. The racing line code won't draw the last patch. Due to support for open ended tracks I guess. To change this add after line 134 in roadstrip.cpp Quote: else - NaN - 05-09-2010 I was able to fix the last patch at rouen by using the third point row as first. Original(third point row(lines 9-12) doesn't fit): Code: -583.278 43.9189 212.332 Modified: Code: -584.023 43.8964 208.364 The question is why the first row of the last patch is set as third? - alex25 - 05-09-2010 there is probably a bug in the track editor. joe is the expert at that. --alex-- |