[2.0.23] Asteroid collectors don't pick up asteroids from the sides

[Cubickman]

Hello,

I've noticed that my asteroid collectors don't pick up chunks coming from the sides of the platform. However, they are neither programmed with filters nor connected to the logic network and their inventories are empty.

Am I doing something wrong or is this a bug ?

[Muche]

Could you attach your save file? I think it may be useful for determining what's happening.

[Loewchen]

And add the version to the title please.

[Cubickman]

The save file and the version :

2.0.23-0

[Cubickman]

I tried to reproduce this bug (because I saw that the collectors were taking again) and I succeeded.

It seems that the bug occurs when a platform is first stopped in orbit, so that the chunks come from the sides. And then, if it starts up and then quickly afterwards is told to go back in orbit and stop, or go back and change direction again. At that point, the collectors don't work.

[Muche]

I'd recheck collectors' behaviour in this case with v2.0.26, as they'll get performance improvements (see viewtopic.php?p=652937#p652937).
In particular, whether chunks' projected paths are detected by collectors when the platform changes speed/direction, after 300 ticks at the latest.

[Genhis]

It seems that the bug occurs when a platform is first stopped in orbit, so that the chunks come from the sides. And then, if it starts up and then quickly afterwards is told to go back in orbit and stop, or go back and change direction again. At that point, the collectors don't work.

Thanks for the report, this is a known issue. It happens because asteroid chunks do a raycast when they spawn and register on asteroid collectors in their path as a form of optimization. If their velocity vector changes, the new path doesn't match the projected one. I'm moving this to minor issues, most likely no action will be done.

[computeraddict]

Thanks for the report, this is a known issue. It happens because asteroid chunks do a raycast when they spawn and register on asteroid collectors in their path as a form of optimization. If their velocity vector changes, the new path doesn't match the projected one. I'm moving this to minor issues, most likely no action will be done.

This might explain why I've seen side-facing collectors on a speed throttled platform miss a lot of chunks on the edges of their ranges, as the speed is always changing a little bit. How bad would the performance hit be to do a couple of additional lines of collision detection based on the recent top and minimum speeds of the platform? Or to recalculate collision for, say, one velocity-modified chunk per platform per tick?

[Muche]

Last time I did some testing on this, changing circuit-controlled filters on an asteroid collector forced its recalculation.
So I cleared filters for 1 tick once per couple of seconds and it collected them all.

However, that was before 118736 Endgame trip to Shattered Planet is incredibly laggy of v2.0.26 and 124971 Asteroid chunk is sometimes ignored by collector of v2.0.30, so the behaviour might have changed.