BarrowmanCalculator
Modify BarrowmanCalculator to actually sum fin nonaxial forces
Modify FinSetCalcTest to sum fins in test3Fin() and test4Fin()
Extra bonus: if at some point we allow fins to be other than radially
symmetrical, this code should still work (not tested)
Modify BarrowmanCalculator to use InstanceMap and InstanceContext to
iterate through all components. This addresses 481 by actually
computing nonaxial forces for all components, instead of computing a
single instanceable component and using it as a representative of all
instances.
Modify calculateNonAxialForces to accept a Transformation argument, so
they can use rotations from instances (only actually affects
FinSetCalc and TubeFinSetCalc)
- Fins may be attached to Transitions (and subclass NoseCones )
[Fix] FinSet now implements the Ring-Instanceable interface
[Refactor] Rocket inherits from ComponentAssembly instead of RocketComponent
[Fix][Refactor] Fin tabs are now correctly validated upon change
[Fix] Fin tabs are now corrected to be no-bigger-than their fins
[Refactor] FinSet.getBodyRadius(..) now requires an argument
[Fix] restricted fin tab positioning to be strictly top/middle/bottom
[Refactor] Reimplement FreeformFinSet.setPoint(...)
[Fix] Prevent Freeform Fins movement past parent's top/front
[bugfix] Fins are now addable to transitions from the GUI
[Fix] Fins, Transitions are now drawn correctly in fin-design window
[Minor] Added makeV2 rocket to TestRockets
[fix] getRootPoints() impl & test
(Effectively a re-write of the MassCalculation code)
- renamed some mass calculator accesor methods
- MassData, InertiaMatrix refactored into 'RigidBody' class
- refactors out cache code to separate wrapper class
- calculations now use the Transform class to translate masses, CM, and MOI
- new class: MassCalculation
- contains relevant context for a given calculation: tree-root, type, time, config
- contains most actual calculation code
- calculations are tracked with a context class: MassCalculation
This is a relatively major refactor / rewrite of the 3d rendering code.
- components geometries are rendered recursively
- components inherit parents' transformations ( translation, rotation)
- implemented Transformation#toGLMatrix()
-- openrocket transformations can be directly fed into Java OpenGL
- added: FinSet#getBoundingBox()
- improved documentation on RocketComponent Location methods
- Refactor RocketRenderer:
- render component trees recursively
- removed RocketRendere#isDrawn(c) -- return true in all implementations
- Refactor ComponentRenderer
- renamed variables to be more descriptive
- changed RocketComponent#toAbsolute(...) => RocketComponent#getComponentLocations()
- Adjust FinRender Code:
- Render Single Fin Instance at-a-time
- takes in an angle for the instance
- assumes the fin is already at it's desired position.
- renames 'fs' -> 'finSet'
- These two methods did the same thing, so this patch removes the former from RocketComponent and descendants
- Wherever used, "PositionValue(...)" was replace with "AxialOffset(...)"
- Removed some extraneous redirect definitions of PositionValue(...)
- Fixed some UI references to the old method
- Rocket total mass is split into 'dryMass' and 'propellantMass'
-- each mass type has a corresponding calculation method in MassCalculator
- Calculating Inertias and Center-of-Mass FOR MOTORS assume:
- time-invariant x-coordinate
- time-decreasing density
- mass correctly tracks propellant usage, as measured by thrustCurves
- elimated MassCalcType enum: was not actually solving a problem.
- simply use motorTime: 0 for launch Double.MAX_VALUE for burnout
- NO_MOTORS is represented by a configuration without attached motors
-- try: "rocket.setSelectedConfiguration( rocket.getEmptyConfiguration())"
- 'dry mass' vs 'total mass': 90% of the time, a caller wanted 'total mass'
- total_mass = dry_mass + propellant_mass
- mass @ simulation time wasn't represented by the enum, had existing overloads anyway
- get vs calculate methods:
-- gets revalidate the cache, then retreive specific information
-- calculate simple calculate the desired information, and ignore the cache.
--- I'm particularly confident about cache reliability:
particularly as it doesn't account for changing time during simulation.
- reduced / simplified debugging messages
- Cleaned up the ThrustCurveMotor api interface
-- Converted 'getTimePoints().length' -> 'getSampleSize()'
-- Unified ThrustCurveMotor gets to the form getXXXX(double motorTime)
-- Restricted access for several methods from public -> protected
- Updated ThrustCurveMotorTest.java
