Implementation: Compile to bytecode

[skx]

This was a quick hack which explains why the Run method walks the tokens directly from the lexer. However doing this is bad for two reasons:

• We don't spot syntax-errors until we run the script.
• We've got a performance hit if the same script is run multiple times.

Solution? Convert from the lexer-tokens to bytecode. This is trivial since we only have three main operations:

• Call a function, with no arguments.
• Run a comparison; return a value if it matches (true/false).
• Return a value (true/false).