Resets recursively, keeping track of children. Created tests, which pass. Computation terminates on graph cycles. Added comments.

Some refactoring, separating node type from node instance. Added parameters, and disabled ability to set inputs.

Modified test script and fixed setParam function

see cb_setparam, cb_getparam

- Realised that controller value may get confused with the actual value of the PID terms in the controller. To avoid this confusion in the future I have refactored things referencing the PID constant to now be called "control_parameter"
- There are quite a few instances of variables that changed.
- Sorry Jake

I was not copying the float value into the array correctly.
Hopefully I have it right now... will check with eric when he gets back from making tea.

- the desired format is in the comment above the funciton
- format should be controller ID, controller value ID, controller value
- should now be sending the data in the correct format

changed getcontrol to respond with the same message ID that was received from groundstation in the original getcontrol command

- FIXME: there are still a few issues making it so that setting and getting setpoints may not work

- added some comments in commands.c/h about usage of the file
- added a MAX_CONTROLLER_ID and MAX_CONTROL_VAL_ID to the enums in commands.h
- implemented cb_set_control in callbacks.c to work with the new parameter struct
- changes the parameter_struct to now contain an array of pid_controllers instead of a bunch of random PID_t members
- the new parameter_struct array can now be accessed by controller ID as defined in commands.h

added the datatype parameter back into MessageType struct in commands.h because Jake actually does need it.

Conflicts:
	groundStation/src/backend/commands.c
	quad/sw/modular_quad_pid/src/callbacks.c
	quad/sw/modular_quad_pid/src/initialize_components.c

- removed the "ID", datatype parameters from MessateType struct in commands.h
- made equivalent changes in commands.c

- removed every use of the "subtype" term
- made the "type" term 16 bits
- made sure that every reference to a message type is 16 bits and used correctly
- simplified commands enum names
- implemented setcontrol callback function that switch on the controller and P, I, D, setpoint value to be set and then set them
- partly implemented getcontrol callback function
- NOTE: setpoint is a double. this causes some problems..
- TODO: add setpoints for rate controllers
- TODO: figure out what message ID I should send back when I get a getpid command

- Added a comment saying that we may not need the ID parameter on the MessageType Struct anymore.
- Removed calls of the nature: "MessageTypes[TYPE_ID].ID" and replaced them with "TYPE_ID" because it is the same thing..
- Added function comments in callbacks.c
- Simplified the callbacks.c by having a more generic setval, getval, respval callback function instead of a billion of them for each PID value.

also added some more comments in both files

there is now no such thing as a message sub type. There is only a message type.
this simplifies the commands struct greatly.
the commands.c file was aprox 1400 lines and now it is about 150...

jpbush authored 8 years ago and dawehr committed 8 years ago

also reorganized the ordering of the commands in commands.c
added a enum for message types and subtype but this will get removed on the next commit