[martin-hughes]

Add support for the Match expression.

This supports only integers in the "operand" positions, and only compares against actual integers in the packages. This seems to match a strict reading of the ACPI spec, where Operand is defined to be an Integer. I'm told (but haven't checked) that these are the limitations imposed by the Windows NT interpreter as well.

This PR works by adding a new argument resolution behaviour purely for the Match opcodes - they're a bit odd because they're raw (unprefixed) bytes in the middle of the expression, and two of the opcodes overlap with others (MTR and MEQ overlaps with Zeros and Ones, respectively)

I've added a basic test to check that the newly parsed opcodes can't be used outside the Match context, plus plenty of other tests that the implemented behaviour matches (ho ho) the behaviour I've written above. The ASL match.asl script has also been checked against uACPI and ACPICA which give the same result.

I haven't squashed the commit history and there's a few low-value commits, so if you're happy to merge then feel free to squash them @IsaacWoods

Isaac I'm pretty happy this is good to go - but I've put you as the reviewer in case you want a closer look

[IsaacWoods]

Great, thanks for working on this Martin! An elegant approach to this and nice tests :)