Various fixes (Beta 1.8-1.8.9 spawnpoint bug, preprocesser #define guards, etc.)

[Nel-S]

• The Beta 1.8 - 1.8.9 spawn algorithms use (0,0) as their first-stage fallback point, not (8,8). (This can be seen by comparing the actual spawnpoint of 21945110879732238 in 1.8.9 Large Biomes (via its level.dat file) vs. 1.9 Large Biomes).
• If keywords like STRUCT or UNREACHABLE were already defined (for example, if a user's program includes a different library with that keyword followed by this one, which happened to me), the compilation encounters an error. I added #ifdef/#undef guards to undefine the previous keywords and so avoid the collisions.
• The Readme's description of spawnpoints was outdated, so I fixed those. I also clarified the makefile instructions, added some basic instructions for Windows users, and added information about make clean for anyone who encounters errors (like I did the first time I tried to use that).

These, meanwhile, will probably be the most controversial changes:

• The /locate command returns strongholds' coordinates with a chunk offset of (0,0), not (4,4). Since that's how most players will likely retrieve/check that structure's coordinates, I changed their returned positions to model that.
• The 1.18+ spawn algorithm also begins checking for topsolid non-waterlogged blocks from chunk offset (0,0), so on average a (0,0) chunk offset is a far better estimate of the world's most likely spawnpoint compared to an (8,8) chunk offset. Therefore I also changed that for estimateSpawn(); getSpawn() is not affected by the change.