Added the Polars.calculateVMG function, which calculates the VMG from a given wind angle, wind speed, and angle to destination. In may not be completely correct at the moment, so it's a work in progress.
#story[873]
Fixed a bug where the polar parser would attempt to insert (wind speed, windspeed, windangle) instead of (windspeed, windangle, boatspeed).
Removed some unused constructors from RaceXMLReader, and fixed a bug where it wouldn't actually use the polar table correctly.
Boat's constructor now expects a Polars table.
Added some comments to the Polars class, and specified the HashMap's template parameters
Added PolarsTest - which needs to be implemented properly when the VMG calculations have been added.
#story[900]
Network:
Moved the Network message classes from the Utils package to the Messages package.
Renamed BoatLocationMessage to BoatLocation and BoatStatusMessage to BoatStatus to be consistent with other message classes.
Renamed the BoatStatus enumeration to BoatStatusEnum as it conflicted with BoatStatus (the message).
Moved the BoatStatusEnum and MessageType enumerations from the Utils package to the Messages/Enums package.
Changed the BoatStatusEnum and MessageType enumerations to use map look ups in the enum.fromByte(b) method - this means that there's less copy and pasted values, and the fromByte function doesn't need to be modified if new enumerations are added.
Added a sequenceNumber member to the Heartbeat class.
Added an InvalidMessageException in the package Networking/Exceptions. This is thrown when a message is read, but it is invalid in some way.
Refactored/tidied up the Networking/BinaryMessageEncoder and Decoder classes. The decoder throws InvalidMessageExceptions instead of returning null.
Visualiser:
VisualiserInput now wraps a DataInputStream around the socket. This provides the stream.readFully(buffer) function, which is a blocking read, removing the need for busy wait loops. Replaced the getBytes() function with getNextMessage() and getNextMessageBytes(). These read the next message from the socket, and return it as a message object and a byte array, respectively.
Changed the current heartbeat timeout to 10 seconds. Added some work-in-progress code to attempt to reconnect when connection is lost. It currently doesn't work. I think Fan-Wu was doing a proper implementation, however.
VisualiserInput also has a queue of received events. Currently not really used, but could be useful in the future. Events get added as they are read.
Changed VisualiserInputs main loop to use instanceOf instead of switch+case. Feedback wanted - are there any downsides to using this instanceOf method?
#story[778,782]
Also changed the while(visualiserInput.getRaceStatus() == null) loops to an if-then-return early block. They caused the program to freeze if the race data source wasn't sending any race status messages (currently, our mock doesn't send any once the race has finished).
#story[782]
Made a few quick fixes (to be tidied a bit later) to Mock.Race class. It now has this.startTime, instead of the AnimationTimers having their own starTime value. RaceStatus messages are now sent properly (still need a refactor though), so they actually send the race star time - this means that the visualizer using the mock data source displays the correct time instead of (midnight, jan 1, 1970).
Removed one of the constructors from RaceStatus - it allowed for constructing and sending essentially invalid RaceStatus messages, and simply wasn't needed.
Added a temporary title ("RaceVision - Team 7") to the visualiser - may be worth changing when we decide on a team name.
Fixed a bug with the order of operations in visualiser.StartController.countdownTimer(). Calling begin race before hiding panes caused them to never be hidden.
Mock.Event. Changed the scaleFactor to 5x for the time being - to easier testing of things like pre-race timers, countdowns, etc...
#story[778,782]
-This was necessary because one boat (the last one to be processed) would lag and jump, as sometimes there was not enough time for all location messages to be sent before the next lot were added.
- The solution was to add a boatOffset, which means while the boats are still processed with3 coming after 2 coming after 1 etc, the loop starts with boat 0, then the next run through it wiull start with boat 1 and so on.
-This eliminated the lag (or at least spread it evenly across all the boats)
#story[778]
Joseph Gardner
fjc40
hba56
David Wu
Erika Savell