Help with a strange Ghost Bug (not the usual one)!

LittleBear · Post by **LittleBear** » Tue Aug 15, 2006 9:37 pm

I'm working on an OXP with Ramon & Heritic, which features some very big but killable ships. EG:-

The ship is big and uses a lot of sub-entries for its turrents and other bits of it. (bridge, shield generators and 4 turrets).

The ship works well in playtesting, but a stange effect keeps happening when the ship is killed. Although it dies and blows up most impressivley, a ghost of the ship is left in space.

Oolite knows the ship is dead (it death_actions trigger). It stops fireing and just spins about venting plasma. The ship is really a ghost as you can fly through it without harm.

This is very odd. But perhaps it has somthing to do with the number of sub-entries.

Can anybody think why this should happen?

At the moment the sub-entries all have energy values. Do they need this to work? Would they still work if I set them to zero? Could this be causing the bug?

Would giving the main and the sub-entries the death-actions become explosion help?

It is very odd as Oolite knows the ship its dead but leaves it's image there.
Any thoughts?

aegidian · Post by **aegidian** » Tue Aug 15, 2006 10:04 pm

This is the second instance I've seen of a large ship or carrier/station leaving a ghost when destroyed.

Which narrows down the search for this bug somewhat.

LittleBear · Post by **LittleBear** » Tue Aug 15, 2006 10:15 pm

If it helps, this ship is not a carrier (it dosen't need defence ships with 4 turrets, 2 military lasers, 2 thargoid lasers and 12 Hard Head Missiles!), but it is carrier size.

ramon · Post by **ramon** » Tue Aug 15, 2006 10:19 pm

and <key>frangible is set to true if this makes any difference, and all the subentities use NullAI.

LittleBear · Post by **LittleBear** » Tue Aug 15, 2006 10:22 pm

Should have mentioned that. But I think it is somthing to do with size, as the same problem happend when I made the Military Station destroyable to test the DeathBurger Script.

JensAyton · Post by **JensAyton** » Thu Aug 17, 2006 12:37 am

In that case, try scaling the model down and see what happens. :-)

LittleBear · Post by **LittleBear** » Thu Aug 17, 2006 8:21 am

I don't want to make it small though!

JensAyton · Post by **JensAyton** » Thu Aug 17, 2006 9:25 am

No, but it would test your hypothesis. (I suspect that the problem would remain, but if it didn’t, that would be valuable debugging information.)