I might be wrong here, but that ball looks really tiny. Could it be possible that HL rounds the collision boxes to make it easier to work with them? Or maybe HL requires a minimum size, because it might take too much processor power to calculate collisions for tiny boxes. When rouding or increasing size HL is also apparently not centering the object in the collision box, thus giving the appearance that 'left' walls are bad, but 'right' ones are good (and this also goes for forward/backward and up/down then).
Try changing the size of the ball to 'nice' numbers (like 8, 16, 32, 64, etc). Maybe that will solve the problem. Jeroen "ShadowLord" Bogers ----- Original Message ----- From: "Mark Gornall" <[EMAIL PROTECTED]> To: <[EMAIL PROTECTED]> Sent: Monday, June 23, 2003 23:37 Subject: [hlcoders] Fw: Re:origin / collision problem > Ok, I'll try one more time, then I'll give up (on the problem and probably > on this list!) > > The fist image (5K) shows the *closest* I can push the ball to the wall > without it stopping as if it's collided with something: > http://www.planethalflife.com/ios/collision_bad.jpg > > The second image (4K) shows how close I can move the ball to the wall on the > opposite side of the map: > http://www.planethalflife.com/ios/collision_good.jpg > > One thing I noticed, when getting pm_debug / PM_DrawPhysEntBBox to draw the > bounding box, the pe->model pointer is null, but pe->name has the correct > name for the ball and it also reports the correct min/max, which it then > draws. The null pointer seemed odd but the code checks for it so maybe it's > ok, I don't know. > > I thought it might be the map or clip brushes but I tried it in other maps > and the same thing happens on the same sides of the maps. > The map in the picture shows normal bsp walls, no clip, nothing fancy. > I have rebuilt the map with Zoners tools and some really old ones that come > with hammer, just in case it was a map/hull problem. Both gave identical > results. > > This doesn't make sense to me so I'd appreciate any help. > Thanks, > Mark. > > > _______________________________________________ > To unsubscribe, edit your list preferences, or view the list archives, please visit: > http://list.valvesoftware.com/mailman/listinfo/hlcoders > > > _______________________________________________ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlcoders
