11:20
Raw stack traces will be very nice. I am going to do the resolution as late as possible so that way it is actually faster. Since throwables could just be initialized and tossed, there is no real reason to perform the heavy work of making stack traces nices until we actually want them to be nice such as when they are being printed.
11:35
I need to implement invoking the special native methods which do special things. These are internal VM functions, but they can easily be detected and handled accordingly to just return a value directly.
11:47
It would be easier to work with long, but I will go for int instead with high and low. This is because some systems might not have native 64-bit stuff so that means worse off systems do not need 64-bit math to work with stack traces which will slow it down.
13:12
ClassName is going to need to handle primitive types, so I have a solution
for that. If the name of the class is the primitive type name in the language
then it will be treated as one.
16:23
Okay I think the debug messages take up too much space!
16:39
Okay the byte code is being parsed wrong, this is our byte code:
67: iinc 6, 1
70: iinc 7, 1
73: goto 51
And this is how my VM decodes it:
@67#IINC~:[6, 1]
@69#ACONST_NULL~:[]
@70#IINC~:[7, 1]
@72#ACONST_NULL~:[]
@73#GOTO:[->@51]
So, this means IINC is being read improperly, it has a bad size but it does not break as much because ACONST_NULL is a single byte instruction.
17:46
Organization is doing wonders for getting stuff done, I like it.
17:56
I need to make arrays more optimized because right now everything is stored as an object which will mean an insane amount of memory is required.