Make Location encoding algorithm consistent with other implementations.

[zrax]

Although DS doesn't actually encode any locations with negative prefixes anywhere outside of the tests, the computation was incorrect. This now matches the encodings generated by plAgeDescription::CalcPageLocation in H-uru/Plasma.

Yes, I'm aware that libhsplasma's computation is even more broken than this... I'll probably address that one next.

[dpogue]

@zrax Just checking because of the unexpected discoveries in the libHSPlasma PR when comparing with real game data, are there any changes needed here or is this good?

[zrax]

@zrax Just checking because of the unexpected discoveries in the libHSPlasma PR when comparing with real game data, are there any changes needed here or is this good?

No, this is based on the same faulty implementation I referenced for the libHSPlasma fix... This needs to be updated too.

EDIT: Actually, the original code produced correct results; this whole PR is based on my faulty read of the Plasma code :(. I'll just update it to be consistent with the change to libHSPlasma, and keep the Catch update.