Stupid Question: hitsphere evaluation efficiency

[MadRat]

Probably could save people alot of hassles with footprint settings if somehow Upswing could measure the model dimensions and display a "suggested footprint size".

[FLOZi]

upsPRing

[KDR_11k]

Probably could save people alot of hassles with footprint settings if somehow Upswing could measure the model dimensions and display a "suggested footprint size".

Just look at the grid it displays.

[FLOZi]

I have my own question regarding hitspheres; Any way of manipulating weapons and hitspheres so that the x,z passed to HitByWeapon are always from the side that is actually hit, rather than occasionally from the opposite side?

[KDR_11k]

That would require making weapons check for collision on the line between their old and new position (use the beamlaser code if in doubt) instead of only checking at the positions themselves. Would be great if that were implemented, it'd also avoid the problem of fast weapons travelling through small targets without connecting.

[FLOZi]

That would require making weapons check for collision on the line between their old and new position (use the beamlaser code if in doubt) instead of only checking at the positions themselves. Would be great if that were implemented, it'd also avoid the problem of fast weapons travelling through small targets without connecting.

Actually, it wouldn't, tracing the ray would be pretty pointless on a ballistic weapon as there could be another unit/feature between where it hits and where it collides.

You're on the right track though. It should be possible to determine the angle relative to the map which the weapon collided just from the start and end positions, no need to trace all points along the line.

[Argh]

Why not use a 3-point, two-ray approximation? That'd take care of curving shots as well as straight ones- put the top point at the top of the arc.