#!/usr/bin/env python3

# Dynamic linking consistency checker
# hacked together by _that@xda, 2014-12 to 2022-03

import os
import string
import argparse


def ldcheck(files, libpath, printalldefined, printresolved, demangle):
    nmopts = "-C" if demangle else ""
    libs = files
    libswithpath = []

    # find all dependencies
    for filename in libs:
        if not os.path.isfile(filename) and libpath:
            filename = findlibraryinpath(filename, libpath)
        libswithpath.append(filename)
        with os.popen("readelf -d {0} | grep '\(NEEDED\)' | sed -r 's/.*\[(.*)\]/\\1/'".format(filename)) as f:
            for line in f:
                dependency = line.rstrip()
                if dependency not in libs:
                    libs.append(dependency)

    print("libs:", libswithpath)
    libsused = set((libswithpath[0],))

    # read all defined symbols
    syms = {}
    for filename in libswithpath:
        for sym in readsymbols(filename, nmopts):
            if sym["type"] != "U":
                # TODO: support symbols defined by multiple libs properly, and weak symbols
                # TODO: more proper handling of symbol versions, see https://maskray.me/blog/2020-11-26-all-about-symbol-versioning
                if "@@" in sym["name"]:
                    syms[sym["name"].replace("@@", "@")] = {"type": sym["type"], "file": filename}
                    syms[sym["name"].split("@@")[0]] = {"type": sym["type"], "file": filename}
                else:
                    syms[sym["name"]] = {"type": sym["type"], "file": filename}

    # resolve unresolved
    for filename in libswithpath:
        for sym in readsymbols(filename, nmopts):
            if sym["type"] == "U":
                resolved = syms.get(sym["name"])
                if resolved:
                    if filename in libsused:
                        libsused.add(resolved["file"])
                    if printresolved:
                        print("{0:25} {1:25} {2} {3}".format(filename, resolved["file"], resolved["type"], sym["name"]))
                else:
                    print("{0:25} {1:25} U {2}".format(filename, "UNRESOLVED #####", sym["name"]))
            else:
                if printalldefined:
                    print("{0:25}                  {1}".format(filename, sym["line"]))

    unused = set(libswithpath) - libsused
    if unused:
        print("unused:", unused)


def findlibraryinpath(filename, searchpath):
    for d in searchpath.split(os.pathsep):
        filepath = os.path.join(d, filename)
        if os.path.isfile(filepath):
            return filepath
    return filename


"""
output format of nm:
00003004 A __bss_start
         U __cxa_atexit
"""
def readsymbol(line):
    s = line if line[0] == " " else line.lstrip(string.hexdigits)
    s = s.lstrip()
    return {"type": s[0], "name": s[2:], "line": line}


def readsymbols(filename, nmopts=""):
    with os.popen("nm -D {0} {1}".format(nmopts, filename)) as f:
        for line in f:
            yield readsymbol(line.rstrip())

if __name__ == "__main__":
    parser = argparse.ArgumentParser(description='Check dynamic linkage consistency by matching defined vs unresolved symbols.')
    parser.add_argument('files', metavar='FILE', nargs='+', help="a dynamically linked executable or library.")
    parser.add_argument('-p', '--path', help="Search path for libraries (use like LD_LIBRARY_PATH)")
    parser.add_argument('-r', '--resolved', action='store_true', help="Print resolved symbols. By default only unresolved symbols are printed.")
    parser.add_argument('-a', '--alldefined', action='store_true', help="Print all defined symbols")
    parser.add_argument('-d', '--demangle', action='store_true', help="Demangle C++ names")

    args = parser.parse_args()
    ldcheck(args.files, args.path, args.alldefined, args.resolved, args.demangle)