java_class_parser.py

import binascii
import struct


CONSTANT_UTF8 = 1
CONSTANT_INTEGER = 3
CONSTANT_FLOAT = 4
CONSTANT_LONG = 5
CONSTANT_CLASS = 7
CONSTANT_DOUBLE = 6
CONSTANT_STRING = 8
CONSTANT_FIELDREF = 9
CONSTANT_METHODREF = 10
CONSTANT_INTERFACEMETHODREF = 11
CONSTANT_NAMEANDTYPE = 12
CONSTANT_METHODHANDLE = 15
CONSTANT_METHODTYPE = 16
CONSTANT_INVOKEDYNAMIC = 18

def print_byte(b):
    hex = str(binascii.hexlify(b), 'ascii')
    print(':'.join(hex[i:i+2] for i in range(0, len(hex), 2)))

def read_short(f):
    return struct.unpack('>H', f.read(2))[0]

def read_long(f):
    return struct.unpack('>I', f.read(4))[0]

def print_class_acc(acc):
    ACC_PUBLIC = 0x0001
    ACC_FINAL = 0x0010
    ACC_SUPER = 0x0020
    ACC_INTERFACE = 0x0200
    ACC_ABSTRACT = 0x0400
    ACC_SYNTHETIC = 0x1000
    ACC_ANNOTATION = 0x2000
    ACC_ENUM = 0x4000

    if acc & ACC_PUBLIC:
        print('PUBLIC', end=' ')
    if acc & ACC_FINAL:
        print('FINAL', end=' ')
    if acc & ACC_SUPER:
        print('SUPER', end=' ')
    if acc & ACC_INTERFACE:
        print('INTERFACE', end=' ')
    if acc & ACC_ABSTRACT:
        print('ABSTRACT', end=' ')
    if acc & ACC_SYNTHETIC:
        print('SYNTHETIC', end=' ')
    if acc & ACC_ANNOTATION:
        print('ANNOTATION', end=' ')
    if acc & ACC_ENUM:
        print('ENUM', end=' ')

def print_method_acc(acc):
    ACC_PUBLIC = 0x0001
    ACC_PRIVATE = 0x0002
    ACC_PROTECTED = 0x0004
    ACC_STATIC = 0x0008
    ACC_FINAL = 0x0010
    ACC_SYNCHRONIZED = 0x0020
    ACC_BRIDGE = 0x0040
    ACC_VARARGS = 0x0080
    ACC_NATIVE = 0x0100
    ACC_ABSTRACT = 0x0400
    ACC_STRICT = 0x0800
    ACC_SYNTHETIC = 0x1000

    if acc & ACC_PUBLIC:
        print('PUBLIC', end=' ')
    if acc & ACC_PRIVATE:
        print('PRIVATE', end=' ')
    if acc & ACC_PROTECTED:
        print('PROTECTED', end=' ')
    if acc & ACC_STATIC:
        print('STATIC', end=' ')
    if acc & ACC_FINAL:
        print('FINAL', end=' ')
    if acc & ACC_SYNCHRONIZED:
        print('SYNCHRONIZED', end=' ')
    if acc & ACC_BRIDGE:
        print('BRIDGE', end=' ')
    if acc & ACC_VARARGS:
        print('VARARGS', end=' ')
    if acc & ACC_NATIVE:
        print('NATIVE', end=' ')
    if acc & ACC_ABSTRACT:
        print('ABSTRACT', end=' ')
    if acc & ACC_STRICT:
        print('STRICT', end=' ')
    if acc & ACC_SYNTHETIC:
        print('SYNTHETIC', end=' ')

def parse_attr(f):
    attr_count = read_short(f)
    print('attr count', attr_count)
    for j in range(attr_count):
        name = constants[read_short(f)]['contents'].decode('utf-8')
        print('     name', name)
        length = read_long(f)
        print('     length', length)
        if name == 'Code':
            f.read(4)  # max_stack + max_locals
            code_length = read_long(f)
            code = f.read(code_length)
            print('     code', end=' ')
            print_byte(code)
            print('     execption ', read_short(f))
            parse_attr(f)
        else:
            f.read(length)


f = open('Simple.class', 'rb')

magic = f.read(4)
minor_ver = read_short(f)
major_ver = read_short(f)
constant_pool = read_short(f)
constants = [None] * constant_pool
for i in range(1, constant_pool):
    t = struct.unpack('>B', f.read(1))[0]
    if t == CONSTANT_CLASS:
        name = read_short(f)
        constants[i] = {
            'type': 'class ref',
            'name': name,
        }
    elif t == CONSTANT_UTF8:
        size = read_short(f)
        constants[i] = {
            'type': 'literal',
            'contents': f.read(size),
        }
    elif t == CONSTANT_STRING:
        constants[i] = {
            'type': 'string',
            'value': read_short(f),
        }
    elif t == CONSTANT_METHODREF:
        constants[i] = {
            'type': 'method',
            'class': read_short(f),
            'name_and_type': read_short(f),
        }
    elif t == CONSTANT_FIELDREF:
        constants[i] = {
            'type': 'field',
            'class': read_short(f),
            'name_and_type': read_short(f),
        }
    elif t == CONSTANT_NAMEANDTYPE:
        constants[i] = {
            'type': 'name_and_type',
            'name': read_short(f),
            '_type': read_short(f),
        }
    else:
        print(t)

# [print(i, x) for i, x in enumerate(constants)]

acc = read_short(f)
print_class_acc(acc)

this = read_short(f)
print('Class', constants[constants[this]['name']]['contents'])
extends = read_short(f)
print('Super Class', constants[constants[extends]['name']]['contents'])
print('=====================')
print('interface count', read_short(f))
print('field count', read_short(f))
print('\nMethods: ')
print('--------------------\n')
for i in range(read_short(f)):
    print_method_acc(read_short(f))
    print(constants[read_short(f)]['contents'])
    print('     descriptor', constants[read_short(f)]['contents'])
    parse_attr(f)
    print('\n')
parse_attr(f)