Find all dynamic libraries needed by ELF or MachO executables recursively

find_dynamic_deps.py

#!/usr/bin/env python3
import fnmatch
import platform
import sys
from pathlib import Path
from pprint import pprint

from elftools.elf.dynamic import DynamicSection
from elftools.elf.elffile import ELFFile
from macholib.MachO import MachO

system_libs = [
    "libstdc++*",
    "libc++*",
    "libgcc*",
    "libm",
    "libc",
    "libpthread",
    "libdl",
    "librt",
    "libutil",
    "libresolv",
    "libnsl",
    "libcrypt",
    "libX*",
    "libGL*",
    "ld-linux-*",
    "libcuda*",
    "libcublas*",
    "libnv*",
]


def get_needed_so(file_path):
    with open(file_path, "rb") as f:
        elf = ELFFile(f)
        for section in elf.iter_sections():
            if isinstance(section, DynamicSection):
                for tag in section.iter_tags():
                    if tag.entry.d_tag == "DT_NEEDED":
                        yield tag.needed


def get_needed_dylib(file_path):
    macho = MachO(file_path)
    for header in macho.headers:
        for command in header.commands:
            cmd = command[0]
            if cmd.cmd == "LC_LOAD_DYLIB":
                dylib_name = command[2].decode("utf-8").strip("\x00")
                yield dylib_name


if platform.system() == "Darwin":
    get_lib_needed = get_needed_dylib
else:
    get_lib_needed = get_needed_so


def get_needed_recursive(executables, search_paths):
    def _add_needed(name, path):
        if name in needed:
            return
        needed[name] = path
        for dep_file in get_lib_needed(path):
            if dep_file.startswith("/"):
                raise ValueError(f"Absolute path {dep_file} in {path}")
            for search_path in search_paths:
                dep_path = Path(search_path, dep_file)
                if dep_path.exists():
                    _add_needed(dep_file, dep_path)
                    break
            else:
                missing_name = dep_file.split(".")[0]
                if not any(fnmatch.fnmatch(missing_name, pattern) for pattern in system_libs):
                    print(f"Warning: {dep_file} required by {name} not found")

    needed = {}
    for executable in executables:
        executable = Path(executable)
        _add_needed(executable.name, executable)

    return needed


def main(bin_dir, search_paths):
    bin_dir = Path(bin_dir)
    executables = sorted(
        p for p in bin_dir.iterdir() if p.is_file() and p.stat().st_mode & 0o111 and not p.suffix
    )
    needed = get_needed_recursive(executables, search_paths)
    pprint(needed)


if __name__ == "__main__":
    bin_dir = sys.argv[1]
    search_paths = sys.argv[1:]
    main(bin_dir, search_paths)