Fix plLocation encoding/decoding to match Plasma, and add tests.

[zrax]

No description provided.

[dgelessus]

I looked into this a couple of years ago when implementing sequence number parsing in my NAGUS project, and also came to the conclusion that libHSPlasma was wrong :D If you want to compare notes (only for the MOUL format): these are my test cases (all taken from real ages currently in MOULa) and here's the text explanation of how I understand the encoding.

At first glance, your code looks right to me. I'll try to take a detailed look at the logic soon.

[zrax]

Good suggestion about using real PRP locations as test values... Despite modeling the tests after what Plasma's plAgeDescription::CalcPageLocation() does, this would have been even more wrong than before. Clearly, that function is not used in the way I think it is with negative prefixes.

[dpogue]

LGTM 👍

I guess Plasma itself is now the thing lacking unit tests for plLocation compatibility...