vjurenka / bwmirror-generator Goto Github PK

It's difficult to know which files are input or output when everything is all in the parent directory. Some clean up task should make the workspace much friendlier.

unit.getUnitsInWeaponRange(weapon) consistently crashes the VM.

To my knowledge this can always be reproduced by calling the method as follows:
unit.getUnitsInWeaponRange(unit.getType().groundWeapon())

There are no instructions for setting up and running BWMirror-Generator. It's important to have instructions so that potential contributors can get up to speed with minimal effort.

Running CJavaPipeline in intelliJ gives me several errors, along the lines of...

Error on line 9 in /C:/Users/Adam/BWMirror-Generator/generated/bwta/BaseLocation.java:9: error: package bwapi does not exist
import bwapi.Position;
            ^Error on line 10 in /C:/Users/Adam/BWMirror-Generator/generated/bwta/BaseLocation.java:10: error: package bwapi does not exist
import bwapi.TilePosition;
            ^Error on line 11 in /C:/Users/Adam/BWMirror-Generator/generated/bwta/BaseLocation.java:11: error: package bwapi does not exist
import bwapi.Player;
            ^Error on line 12 in /C:/Users/Adam/BWMirror-Generator/generated/bwta/BaseLocation.java:12: error: package bwapi does not exist
import bwapi.Unit;
            ^Error on line 13 in /C:/Users/Adam/BWMirror-Generator/generated/bwta/BaseLocation.java:13: error: package bwapi does not exist

https://github.com/bwapi/bwapi/releases/tag/v4.2.0

if a unit a is in range of an enemy unit e, calling e.getUnitsInRadius(range) will return a different unit object b, such that a.equals(b) == false even though a.getID() == b.getID().

AFAIK this can be reproduced anytime.

Proposal: either maintain only one unit object per unit or override the equals method of the Unit class such that it returns true for objects with the same ID (unit1.getID() == unit2.getID()).

BWMirror currently supports only x86.
This is implementd as a check for the result of System.getProperty("os.arch").
Due to the bug in openjdk, the returned value can be "i386" too
https://bugs.openjdk.java.net/browse/JDK-4468998