Witchspaced nav buoys

Screet · Post by **Screet** » Mon Mar 16, 2009 2:12 pm

Hi,

occasionally a ship goes into witchspace at the nav buoy. This is not allowed for players, but NPCs can. Strange things happen:
- the buoy also is gone (thrown into witchspace, I guess)
- the witchspace bubble does not expire, shouldn't it after some seconds?

It's also always very big bubbles...as if their size is related to the size of the ship creating them.

Other problems arise from this:
- any ship launching or heading for the station hits the bubble and is thrown out of system, including GRS ones...
- escort ships tumbling close to the bubble, AI-state "LOOKING_FOR_BUSINESS"

Screet

Post by **Kaks** » Mon Mar 16, 2009 2:50 pm

Screet wrote:
- the witchspace bubble does not expire, shouldn't it after some seconds?

It's also always very big bubbles...as if their size is related to the size of the ship creating them.

It should! However, both bubble persistence and size are indeed proportional to the ship's mass. I'm not too sure what's supposed to happen if another ship adds to the first ship's mass though.

As for the other problems: IIRC, there's already a bug in BerliOS regarding NPCs not being mass locked & this seems to be another instance of that bug: definitely in the to-do list!

Cheers,

Kaks.

Website · Post by **DaddyHoggy** » Mon Mar 16, 2009 3:08 pm

If you have YAH installed then the nav buoy could be a Tetra-Screen and they're HUGE! - If one of them is included in the size/persistence calculation then no wonder you end up with a overly big and seemingly endless Witchspace "bubble"

Screet · Post by **Screet** » Mon Mar 16, 2009 3:10 pm

DaddyHoggy wrote:
If you have YAH installed then the nav buoy could be a Tetra-Screen and they're HUGE! - If one of them is included in the size/persistence calculation then no wonder you end up with a overly big and seemingly endless Witchspace "bubble"

Yes, YAH is installed...I didn't want to miss all those nice adds for a laugh now and then

Screet

Post by **Micha** » Mon Mar 16, 2009 3:19 pm

Witchspace bubbles are sized relative to the mass inside them, and decay over time. However, if mass is added, they stick around for longer - so continuously 'feeding' a witchspace bubble would keep it around forever. Kindof like a black hole. Would be impressive if it got big enough to swallow the Station!

As for the ship entering witchspace too early, I had the opposite problem in my WiP build - ships just kept cruising around behind the station at ~7km distance trying to witchspace out but can't because they never got far enough away. I've modded my exitingTraderAI slightly to cause them to start heading towards the Witchspace beacon, instead of sticking around near the station, which fixed that.

Post by **Eric Walch** » Mon Mar 16, 2009 4:56 pm

Micha wrote:
I've modded my exitingTraderAI slightly to cause them to start heading towards the Witchspace beacon, instead of sticking around near the station, which fixed that.

It does not fix it completely. I had the same problem with the grs station. That is also big and used a variation of your approach. However, in some systems the station was blocking the direct way to the witchpoint. In those rare systems my leaving ships often crash into the station.
This problem will be different with a rotating station were every ship diverges in an other direction depending on the rotation state during launch.

Post by **Micha** » Mon Mar 16, 2009 5:02 pm

I thought 'plotCourseToWitchpoint' (or whatever it was called) actually generates waypoints around obstacles if necessary? At least that's the impression I got from reading the code, haven't actually seen an incident where a ship was blocked from heading into deep space yet.

Post by **Eric Walch** » Mon Mar 16, 2009 5:13 pm

Micha wrote:
I thought 'plotCourseToWitchpoint' (or whatever it was called) actually generates waypoints around obstacles if necessary? At least that's the impression I got from reading the code, haven't actually seen an incident where a ship was blocked from heading into deep space yet.

You mean "checkCourseToDestination" Yes it does but is unreliable near large masses. Or it behaves bad near the planet (low FPS ratings). I have seen traders hitting the planet using that code