Armadillo compression algorithms

[gigaman]

Does anybody know what older versions of Armadillo used as their compression algorithm?
I know newer versions (3.10+) use either ZLIB, or just a simple RLE-like packing ("method 0").
Older versions (e.g. Armadillo 2.x), however, use 3 different methods (according to the compression setting you select in the program) - the same "method 0", and more complex methods 1 and 2 (especially method 2 is quite a lot of code). Does anybody know what they are? Is it some common compression algorithm, with source code available maybe?

Thanks!

[Squidge]

I've never done any reversing of armadillo's compression algorithms myself, I've always just got the armadillo rtl to do all the unpacking for me, and not worried about it. I suspect most other people have gone down the same path.

Is there some reason your looking for this info, rather than just using the built in functions?

[gigaman]

Well, yes, I'd like to have a portable source that unpacks the streams (compilable on non-Intel platforms, with thorough bounds checks, etc.)