Havok4 Beta Test Ideas

From Second Life Wiki
Jump to: navigation, search

Building and Linking

  • Construction tools and HUDs
  • manual linking/unlinking Linkability Rules
  • test link distance rules (for attachments too) Linkability Rules
  • Make sure that linked objects remain linked when you rez them in any orientation
    • Rez out large linked objects from your inventory and make sure they are still linked, e.g., buildings, ships, aircraft, cars
    • Rez out linked objects containing small prims, e.g., jewelry, furniture, attachments, HUD attachments.
    • Rez out linked objects that have both small and large prims, e.g., furniture, vehicles.
    • edit a prims of a linked object and move it as far from the other prims as possible. Take the linked object, rez it, and make sure it is still linked.
  • moving/rotating/scaling sub-parts of linked sets (via UI and scripting)

Scripted Movement

  • all script movement, actions (hover, vehicle, buoyancy etc)
  • sitting on objects, standing up Sitting test
  • interpenetration resolution when two dynamic objects are overlapping
  • collision events (script callbacks, especially callbacks on linked children)
  • avatar controls (flying, running, walking)

Collisions

  • collision correctness (do objects/avatars visually appear to collide/deflect at the right place?)
  • piles of colliding objects
  • creating objects from inventory
  • reading the mass of objects after shape/scale changes, especially attachments
  • scripted attachments that push the avatar or other objects, especially after shape/scale changes

Physics Overloads

  • "deep thinks"--regions slowness caused by Havok trying to untangle complex physical objects (which may still exist but it would be interesting to find out how things are different).
  • permission errors
  • rezzing objects via script

Vehicles

  • Fly or drive any vehicles you have, make sure they behave the same as on the main grid.
  • Fly or driver your vehicles across region boundaries. Verify they are no worse than on the main grid. (They may not be any better either)

Avatar Motion

  • Make sure Avatar controls feel the same
  • Walk, fly and run
  • Walk up and down stairs
  • Walk up and down hills
  • Walk, run and fly across region boundaries
  • Collide with objects
  • Collide with other avatars
  • Wear any flight assist attachments and verify your flight speed and control feels the same

Surf board/elevator/magic carpet

Test: A simple flat object with a script to move it using llMoveToTarget() at different speeds and directions. Can you ride it by simply standing on it?

Result: Havok4 is better than Havok1. In H1, the AV falls off the object at speeds around 10 m/s. In H4, speeds of 30 m/s or more seem to still work pretty well.

Two AVs on the same object seems to work fine. They can walk around on the object during motion.

Foot Shuffling: When the object moves, the avatar tends to turn and face the direction of movement, which is OK, but perhaps not desirable. The AV also shuffles its feet. I guess that because it is being caused to move by some outside agent, the agent thinks it must be walking. This is particularly bad when the elevator car is moving up or down. The legs act pretty weird, folding up into the body.

Scripted Behavior

  • Make sure script rezzing items work. e.g., guns.
  • Make sure push guns still push you, or cause damage. (get the other person's consent first!)
  • Test scripted linking and unlinking
  • Make lots of physical objects and drop them in a pile. Drive a vehicle through the pile.
  • Test phantom
  • Toggle phantom status with a script and verify phantom actually toggles
  • Test targetomega, targetomega child prims, and targetomega attachments and HUD objects.
  • Test mega prim rezzing and physics
  • Estate tools "Get Top Colliders"

Permissions

  • Walking, flying and driving vehicles into parcels you are banned from.
  • Walking, flying and driving vehicles into regions you are banned from.
  • Test being ejected from a parcel via right-clicking an avatar > More > Eject...