Doesn't handle polymorphic / unpacking / etc programs

[JeremyWildsmith]

Consider the below code

set [someOtherProgram + 0], 0xac02 ;add a, 10
set [someOtherProgram + 1], 0xac02 ;add 1, 10
set [someOtherProgram + 2], 0x8781 ;set pc, 0

:someOtherProgram

A lot of kernels do this (i.e, for loading a program from disk to memory) however DCPU16-IDE fails this test due to the following line (694) of code in emulator.js on the first if statement.

	this.step = function() {
		if(this.PC.get() < this.program.length) {
			this.nextInstruction();
			
			if(this.attachedDebugger && this.paused)
				this.attachedDebugger.onStep(this.PC.get());
			
			// process one interrupt if we have one
			if(this.interruptQueueingEnabled == false && this.interruptQueue.length > 0) {
				this.processInterrupt(this.interruptQueue.shift());
			}
			
			return true;
		}
		else return false;
	};

The fix would be to simply remove the if statement. A DCPU program never really ends. If you change this, PC can grow to an invalid value (ie, outside the bounds of 16 bit register) due to the increments on step. So this fix also requires another change (most safely in Emulator.js line 12:

Register.prototype.set = function(val) { this.contents = val & 0xFFFF; }

[dangermccann]

You're certainly correct that checking the PC against program length is not the right approach. How do you suggest the emulator should behave for simple programs that do not manage the PC in some sort of event loop? For example, a very basic program such as this one:

SET A, 1
SET B, 2

will produce an error in the emulator after the PC advances past the second SET instruction (Invalid opcode: 0). The program will then halt, which technically is what a real computer would do with this input. Do you think a user may find that behavior counterintuitive? Or would you consider that to be the expected behavior of the emulator?