26

Are there any ways to decompile a dll and/or a .pyd file in order to extract source code written in Python?

Thanks in advance

Improve this question
3
  • 4
    is not same question(dll not a .pyc file ) ! My idea No can't ! Maybe convert to assembly code but how to follow data/function tree ? Commented Feb 24, 2016 at 14:23
  • 2
    @Torxed this is indeed not the same question, pyd files are produced by Cython. Unlike pyc/bytecode, the pyd files are binaries that do not contain the original python source. Although it does not protect from reverse-engineering, it does make it a lot harder, and prevents someone from obtaining the exact original sources of your code. Commented Mar 28, 2018 at 15:15
  • stackoverflow.com/questions/5287253/… Commented Mar 11, 2024 at 10:07

2 Answers 2

Reset to default
25

I assume the .pyd/.dll files were created in Cython, not Python?

Anyway, generally it's not possible, unless there's a decompiler designed specifically for the language the file was originally compiled from. And while I know about C, C++, Delphi, .NET and some other decompilers, I've yet to hear about Cython decompiler.

Of course, what Cython does is convert your Python[esque] code into C code first, which means you might have more luck finding a C decompiler and then divining the original Python code based on the decompiled C code. At the very least, this way you'll be dealing with translation from one (relatively) high-level language to another.

Worst-case scenario, you'll have to use a disassembler. However, recreating Python code from disassembler's output isn't going to be easy (pretty similar to divining the biological functions of a brain from chemical formulas of proteins that make up it's cells).

You might look at this question on ideas and suggestions regarding various decompilers and disassemblers, and proceed your investigation from there.

Improve this answer
Sign up to request clarification or add additional context in comments.

4 Comments

Thanks for the helpful reply, I'll make sure to do some extra research.
@YoussefImam I don't agree with this answer, see my answer stackoverflow.com/a/41075212/1422096
@Basj Reading your updated answer, it seems that .pyd files compiled by Cython do not have source code embedded into them after all. If you could show it to be otherwise, I would be really interested, but until then, I'll stand by my answer.
@Basj, is it possible to reverse engineer the C-code then, there is this software that I need to reverse engineer, it is written in Python and has .pyd files in it.
2

I don't agree with the accepted answer, it seems that yes, the content of the source code is accessible even in a .pyd.

Let's see for example what happens if an error arrives:

1) Create this file:

whathappenswhenerror.pyx

A = 6 print 'hello' print A print 1/0 # this will generate an error

2) Compile it with python setup.py build:

setup.py

from distutils.core import setup from Cython.Build import cythonize setup(ext_modules = cythonize("whathappenswhenerror.pyx"), include_dirs=[])

3) Now import the .pyd file in a standard python file:

testwhathappenswhenerror.py

import whathappenswhenerror

4) Let's run it with python testwhathappenswhenerror.py. Here is the output:

hello 6 Traceback (most recent call last): File "D:\testwhathappenswhenerror.py", line 1, in <module> import whathappenswhenerror File "whathappenswhenerror.pyx", line 4, in init whathappenswhenerror (whathappenswhenerror.c:824) print 1/0 # this will generate an error ZeroDivisionError: integer division or modulo by zero

As you can see the line of code print 1/0 # this will generate an error that was in the .pyx source code is displayed! Even the comment is displayed!

4 bis) If I delete (or move somewhere else) the original .pyx file before step 3), then the original code print 1/0 # this will generate an error is no longer displayed: 

hello 6 Traceback (most recent call last): File "D:\testwhathappenswhenerror.py", line 1, in <module> import whathappenswhenerror File "whathappenswhenerror.pyx", line 4, in init whathappenswhenerror (whathappenswhenerror.c:824) ZeroDivisionError: integer division or modulo by zero

But does this mean it's not included in the .pyd? I'm not sure.

Improve this answer

11 Comments

Moving the pyx shows that the Traceback uses some sort of link from the pyd (so in Linux) to lines in the pyx. The code is not in the pyd.
@Basj, if the pyx file is in same location and when you do import whathappenswhenerror, how do you know it's importing .pyd and not pyx? Probably it's importing pyx and hence the code display on error.
You can tell that it is using the pyx file by looking at the traceback (File "whathappenswhenerror.pyx", line 4)
Precisely, this answer is not correct. It seems, it's importing from .pyx
|

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.