Right now, I've switched from using Linked Lists for storing Lump and Wad Information to using Vectors. Programs that use the Linked Lists for doing data-manipulation will still function, though - I've left the LL code in. Everything will be better performance-wise for version 1.1.
Besides, it's still not too hard to pick up errors, to me. I know my code like the back of my hand. If something goes wrong, I usually find it.
And in my defense, when you're doing things like converting bytes to integers and vice-versa throughout the entire library, there is bound to be repeats in the code used to do such conversions.