ByteCode Format

The bytecode is currently structured to contain the following rules:

• file ::= stmts
• stmts ::= stmt stmts | Empty
• stmt ::= bytecode args
• args ::= arg args | Empty
• arg ::= constant | register
• bytecode ::= [5 byte little-endian] prefix byte 0x00, rest is value
• register ::= [5 byte little-endian] prefix byte 0x01, rest is value
• constant ::= [6 byte little-endian] prefix byte 0x02, next byte is constant type (see below), rest is value

As the current implementation only supports integers and floating point numbers, the constant types are:

• 0x00 = integer
• 0x01 = floating point

Byte-codes themselves are structured as such:

[2 byte code value][2 byte group value]

An example of a file may be:

00 0200 0100 01 01000000 02 00 01000000 = Total width 16 bytes

Reading as:

ByteCode 2, Group 1, first arg is register index 1, second arg is constant (int) value 1.

Under the current implementation, we just added 1 to register number 1, and stored the result back in register number 1. In pseudo-assembly:

add reg1, 1

Currently this is all the byte-code has, format-wise. Watch this space for revisions to include function blocks, string support etc.