给定一个 python .pyc 文件，是否有一个工具可以让我查看字节码？

Question

CPython 解释器会自动将 Python 模块编译成 .pyc 文件。包含字节码的 .pyc 文件是二进制格式（编组代码？）。是否有 GUI（或命令行）工具可以让我查看字节码？

Answer 1

有一个名为 PyChrisanthemum 的可视化 Python 反汇编程序。

如 OP 已经发现的那样，您可以使用模块 dis (python 2.7.3, python 3.2.3) 以命令行方式执行此操作。

Answer 2

每个 *.pyc 文件都是一个包含以下内容的二进制文件：

• 一个四字节的幻数 - 它只是随着编组代码的每次更改而更改的字节；
• 四字节修改时间戳 - 是生成 .pyc 的源文件的 Unix 修改时间戳，以便在源更改时重新编译；
• 从 Python3.3+ 版本开始，接下来的四个字节是将源文件的大小编码为 long 的字段；
• 编组的代码对象。

为什么不直接使用 CPython 的内置功能来完成这项任务？

---

一个文件view_pyc_file.py

import platform
import time
import sys
import binascii
import marshal
import dis
import struct


def view_pyc_file(path):
    """Read and display a content of the Python`s bytecode in a pyc-file."""

    file = open(path, 'rb')

    magic = file.read(4)
    timestamp = file.read(4)
    size = None

    if sys.version_info.major == 3 and sys.version_info.minor >= 3:
        size = file.read(4)
        size = struct.unpack('I', size)[0]

    code = marshal.load(file)

    magic = binascii.hexlify(magic).decode('utf-8')
    timestamp = time.asctime(time.localtime(struct.unpack('I', b'D\xa5\xc2X')[0]))

    dis.disassemble(code)

    print('-' * 80)
    print(
        'Python version: {}\nMagic code: {}\nTimestamp: {}\nSize: {}'
        .format(platform.python_version(), magic, timestamp, size)
    )

    file.close()


if __name__ == '__main__':
    view_pyc_file(sys.argv[1])

用下一个 CPython 版本测试：

• 2.7.9
• 3.4.2
• 3.5.2

---

演示

文件内容main.py

$ cat main.py
print("Never give up")

通过python2.7创建和读取pyc文件

setivolkylany$~/Downloads/temp/temp$ python2.7 -m py_compile main.py 
setivolkylany$~/Downloads/temp/temp$ python2.7 view_pyc_file.py ./main.pyc
  1           0 LOAD_CONST               0 ('Never give up')
              3 PRINT_ITEM          
              4 PRINT_NEWLINE       
              5 LOAD_CONST               1 (None)
              8 RETURN_VALUE        
--------------------------------------------------------------------------------
Python version: 2.7.9
Magic code: 03f30d0a
Timestamp: Fri Mar 10 15:08:20 2017
Size: None

通过python3.4创建和读取pyc文件

setivolkylany$~/Downloads/temp/temp$ python3.4 -m py_compile main.py 
setivolkylany$~/Downloads/temp/temp$ python3.4 view_pyc_file.py __pycache__/main.cpython-34.pyc 
  1           0 LOAD_NAME                0 (print)
              3 LOAD_CONST               0 ('Never give up')
              6 CALL_FUNCTION            1 (1 positional, 0 keyword pair)
              9 POP_TOP
             10 LOAD_CONST               1 (None)
             13 RETURN_VALUE
--------------------------------------------------------------------------------
Python version: 3.4.2
Magic code: ee0c0d0a
Timestamp: Fri Mar 10 15:08:20 2017
Size: 23

通过python3.5创建和读取pyc文件

setivolkylany$~/Downloads/temp/temp$ python3.5 -m py_compile main.py 
setivolkylany$~/Downloads/temp/temp$ python3.5 view_pyc_file.py __pycache__/main.cpython-35.pyc 
  1           0 LOAD_NAME                0 (print)
              3 LOAD_CONST               0 ('Never give up')
              6 CALL_FUNCTION            1 (1 positional, 0 keyword pair)
              9 POP_TOP
             10 LOAD_CONST               1 (None)
             13 RETURN_VALUE
--------------------------------------------------------------------------------
Python version: 3.5.2
Magic code: 160d0d0a
Timestamp: Fri Mar 10 15:08:20 2017
Size: 23

---

基于：

• How can I understand a .pyc file content
• https://gist.github.com/anonymous/35c08092a6eb70cdd723
• https://nedbatchelder.com/blog/200804/the_structure_of_pyc_files.html

Answer 3

根据@Apteryx 关于 PEP 的注释扩展代码形式 @PADYMKO：

def view_pyc_file(path):
    """Read and display a content of the Python`s bytecode in a pyc-file."""

    file = open(path, 'rb')

    magic = file.read(4)
    bit_field = None
    timestamp = None
    hashstr = None
    size = None

    if sys.version_info.major == 3 and sys.version_info.minor >=7:
        bit_field = int.from_bytes(file.read(4), byteorder=sys.byteorder)
        if 1 & bit_field == 1:
            hashstr = file.read(8)
        else:
            timestamp = file.read(4)
            size = file.read(4)
            size = struct.unpack('I', size)[0]
    elif sys.version_info.major == 3 and sys.version_info.minor >= 3:
        timestamp = file.read(4)
        size = file.read(4)
        size = struct.unpack('I', size)[0]
    else:
        timestamp = file.read(4)

    code = marshal.load(file)

    magic = binascii.hexlify(magic).decode('utf-8')
    timestamp = time.asctime(time.localtime(struct.unpack('I', b'D\xa5\xc2X')[0]))

    dis.disassemble(code)

    print('-' * 80)
    print(
        'Python version: {}\nMagic code: {}\nTimestamp: {}\nSize: {}\nHash: {}\nBitfield: {}'
        .format(platform.python_version(), magic, timestamp, size, hashstr, bit_field)
    )

    file.close()