I started writing a bytecode library but you all know how long I stick to a project before deleting it
It will be based on the mocha decompiler which largely still works, the only problems I have seen so far are the synchronised blocks and string append replacer. Mocha is nice apart from a few small things, such as it being old and not abstracting constant push instructions nicely (uses string values or constant pool indices)
I plan to rewrite it from the ground up incorporating extra features including but not unlimited to; generation of bytecode as well as source from IR code, more suitable abstractions, an application/library representation encapsulating type hierarchies and ideally some sort of pattern matching/transforming script system
I will also keep the constantpool because it simplifies type validation and will contribute to an application wide symbol table, I will not however provide an interface for searching it because that is fucking retarded
[center][size=99pt]ಠ_ಠ[/size][/center]