If the from and to vector were parallel this code would sometimes result
in an invalid rotation of the entity being upside down.
We've fixed it by instead relying on the WFMath code for the same
functionality.
* common/const.cpp, common/const.h, common/random.h,
physics/BBox.h, physics/Collision.cpp, physics/Collision.h,
physics/Quaternion.cpp, physics/Quaternion.h, physics/Vector3D.h,
rulesets/Movement.cpp, rulesets/Plant.cpp, rulesets/Python_API.cpp,
tests/collisiontest.cpp: Update variables relating to geometry
from double to float as this is the type used internall by
WFMath.
* physics/Vector3D.cpp, physics/Quaternion.cpp, physics/Quaternion.h:
Remove legacy code.
* server/CommPSQLSocket.cpp: Reformatting.
* server/WorldRouter.cpp: Eliminate unnecessary call to opEnumerate()
and enhance debugging output.
* server/Player.cpp: Correct the use of string.compare() in older
versions of libstdc++.
* common/types.h: Added enumeration for other operations, not
handled by the core, but considered valid. Added new type
for storing mapping between operation parents and their
enumerated type.
* common/op_switch.h: Separate handling of operations known to
be invalid from operations which are valid, but not explicitly
added to the core.
* common/inheritance.h, common/inheritance.cpp, common/custom.cpp:
Provide global operation mapping where enumeration of operations
can be looked up. Install all types handled by the core, and
types known to be used in scripts.
* common/BaseEntity.h, common/BaseEntity.cpp: Add object specific
operation mapping where operation types each object is subscribed
to can be enumerated, and others are rejected. Add mechanism
so that objects can add to the mapping, this subscribing to
operations. This makes enumerating operations much cheaper, and
eliminates the overhead of handling ops an object has no interest
in.
* rulesets/Thing.h, rulesets/Thing.cpp: Add mechanism so scripts
can subscribe to operations.
* rulesets/Character.h, rulesets/Character.cpp:
Add subscription mechanism for ops going to and from
the mind.
* rulesets/BaseMind.h, rulesets/BaseMind.cpp:
Add subscription mechanism for sight and sound perception
operations sub operations, and provide an interface to it
for scripts.
* server/Connection.cpp, server/Admin.cpp, server/Account.cpp,
rulesets/World.cpp, rulesets/World.h, rulesets/Thing.cpp,
rulesets/Stackable.cpp, rulesets/Plant.cpp, rulesets/Food.cpp,
rulesets/Character.cpp, rulesets/BaseMind.cpp:
Add subscriptions to operations for all core base classes.
* rulesets/Python_API.cpp: When script is created, go through
methods, and subscribe to the operations which are handled by
them.
* rulesets/Character.cpp: When calculating orientation of a walking
character, assume it is standing upright.
* physics/Vector3D.h: Add versions of mag() and distance() which
omit the square root calculations, so that when their values
are only used for comparison, the expensive square root
calculations can be ommited. Add new version of collision
and range checking functions which provide the optimal
interface to the application. Add versions of methods which
modify this rather than creating a new operation to eliminate
wasted copies.
* physics/Quaternion.cpp: Make use of faster interface provided
by Vector3D.
* physics/BBox.h: Add new version of collision checking function
which provides optimal interface to the application.
* modules/Location.h, modules/Location.cpp: Add new functions
to optimally calculate the distance between any two entities.
* modules/Location.h: Re-write collision and range functions to
make much more optimal use of Vector3D, BBox and distance
functionality.
* rulesets/Py_Vector3D.cpp, rulesets/Pedestrian.cpp,
rulesets/Movement.cpp, rulesets/Entity.cpp, rulesets/Character.cpp:
Eliminate implicit copies from vector arithmetic.
* rulesets/Pedestrian.cpp, rulesets/MemMap.cpp, common/const.h:
Re-write code which uses magnitude of vectors to use the square
of the magnitude instead, as it is much cheaper to calculate.
in the constructor which takes 2 vectors, and moved it out
of the header file. This appears to work now that the args
have been swapped round. See comments in Quaternion.cpp for
details.
Al Riddoch <alriddoch@zepler.org>