Handle multiple compile units in DWARF data

[Abyss-W4tcher]

Hi,

This PR adds support to detect different compile units (languages) in a DWARF data. The goal is to prevent name and type confusion accross multiple languages exposing concurrent symbols and types. When detected, everything is typed accordingly to a language, even base types, to avoid potentially dangerous shortcuts and assumptions.

To illustrate with a Linux kernel 6.5+, here are the resulting ISF base types (vmlinux provided in #57) :

['_Bool', '__int128', '__int128 unsigned', 'char', 'double', 'float', 'int', 'long int', 'long long int', 'long long unsigned int', 'long unsigned int', 'pointer', 'rust.!', 'rust.()', 'rust.__ARRAY_SIZE_TYPE__', 'rust.bool', 'rust.f32', 'rust.f64', 'rust.i128', 'rust.i16', 'rust.i32', 'rust.i64', 'rust.i8', 'rust.isize', 'rust.pointer', 'rust.u128', 'rust.u16', 'rust.u32', 'rust.u64', 'rust.u8', 'rust.usize', 'short int', 'short unsigned int', 'signed char', 'ssizetype', 'unsigned char', 'unsigned int', 'void']

I attached two ISF to this PR, to check out the differences before and after rust detection. Directly referring to #57, allows to see two distinct fs_struct and rust.fs_struct structs.

compile_units_handling.zipv2

Fixes #57, fixes #63