Is it possible to get python bytecode without using co_code?

Question

Say I am in the python interpreter and define a function as follows:

def h(a):
  return a

If I want to look at the bytecode (not a disassembly using dis), I can typically use h.func_code.co_code. Is there any other way to look at the bytecode?

For example, in the interpreter, if I just type h without making it a function call, I get the address of the function. Can I use that address to get the bytecode? Is there some other way?

Some additional info from the comments:

The app is written in python and packed using something like Py2App, cx_freeze, or py2exe. I played some tricks on the executable and now, when launched, the executable dumps me to a python command line. From there, I manually typed my function h into the interepeter.

Other attributes of h.func_code are present such as co_varnames, co_argcount, etc, but co_code is not.

If I type in h.func_code.co_code into the interpreter, I get 'code' object has no attribute 'co_code'.

UPDATE: From the comments again. As far as I can tell the opcodes have been remapped for the python interpreter that was shipped with the app.

Why isn't h.func_code.co_code good enough? Using the memory address is not going to be any easier. — Martijn Pieters
– Martijn Pieters, Commented Sep 20, 2012 at 13:40
Because for some reason it isn't there. I'm learning about python reverse engineering and it seems there were some protections added to the piece of software. One of which seems to have made co_code unavailable. — mikeazo
– mikeazo, Commented Sep 20, 2012 at 13:42
For starters, do you know it's written in Python? What's it's type? (Also, dir can be overriden by the object, don't bet your farm on its output.) — user395760
– user395760, Commented Sep 20, 2012 at 13:49
@mikeazo that doesn't look like a version string. What's the full version string? — ecatmur
– ecatmur, Commented Sep 20, 2012 at 14:13
It is theoretically possible to remove the co_code attribute from the python interpreter. That would require a custom interpreter build. You could still just load all the python code into your own interpreter (just unzip the zipfiles in the Contents/Resources/ directory and put those on your path. — Martijn Pieters
– Martijn Pieters, Commented Sep 20, 2012 at 14:25

Martijn Pieters · Accepted Answer · 2012-09-20 14:28:21Z

5

If you have defined your own def h(): pass dummy function, and that function does not have a .func_code.co_code value, then most likely the included python interpreter is a custom compiled version where the co_code slot has been cleared.

The work-around is simple, copy the Contents/Resources zipfiles (in all subdirs) elsewhere, add those to your PYTHONPATH and import the code into your own interpreter.

answered Sep 20, 2012 at 14:28

Martijn Pieters

1.1m326 gold badges4.2k silver badges3.4k bronze badges

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

Collectives™ on Stack Overflow

Is it possible to get python bytecode without using co_code?

1 Answer 1

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

Comments

Your Answer

Sign up or log in

Post as a guest

Related