stop memleaker from leaking...... yes i see the irony
free was being blocked from running when unitialized
in shutdown reset our allocList allocCount and currentAllocId so the array can be written to again
Namespace Leaks: these leaks would have been freed on app kill but they were still contaminating the log so free em.
Con::shutdown: gGlobablVars again another leak that would of been freed on app kill
AssetManager onRemove should be clearing its lists
Scripting language needs a free for its strings these destructors free after the rule is executed so the data is cached internally and then freed. This may be needed for other types in the union (should not be used on nodes just raw data)
LEAK from G:\GameDev\GithubRepo\Torque3D\Engine\source\console/consoleObject.h:686
- leak around props remaining after the const_cast free it up
LEAK around sound asset
- assetEnumNameConcat was creating a leak.
This change makes the memory manager work again for detecting leaks, the built in one kept coming into de-ref and other bugs so this is the start of a refactor to get it working.
AIFlyingVehicleControllerData add flightfloor and cieling
resolvepitch from (portions of) the old aiflyingvehicle resource
no reversing for flyingvehicles, so bottom out resolvespeed at 0
for players this defaults to 0.001, for wheeledvehicles, 2.0, and for flyingvehicles, 200
fix naveditor cript not looking up the object.datablock.aicontrollerdata
fix AIWheeledVehicleControllerData not binding the relevant ::resolvespeed
also the relevant ::resolvespeed now lowers the throttle post-turning
add AIFlyingVehicleControllerData
aiwheeleedveiclecontrollerdata resolvespeed now only touches throttle
objects assigned aicontrollers now reflect that by thier objecttype
basic flocking
list aiControllerData's in the datablock. though the command is still required to set the controler and look up the relevant db for game specific logic
variable raylength for the rpath filter. same 0.001 for players to stop recalculating a path when jumping, but bump anything they're mounted to to a 2 unit check
bit of work towards parallel parking. or at least not ending up arcing back and forth infinitely in an arc
in order to call derivatives of AIControllerData datablocks *without* requiring an accompanying AIController subclass as well, leverage the fastdelegate system for our resolver callbacks
additionally, don't try and repath in mid air
aigoal: initialize inange/infirinrange to false. use those to filter callbacks