Skip to main content

High level python bytecode manipulation

Project description

PyBytecodeManipulator

A high level cross-version python bytecode manipulation library build ontop of 'dis' and 'inspect'

Supports code inlining, branch removing and arbitrary code injection into existing functions.

WARNING: using bytecode manipulation on a so low level as we do can break the python runtime at any point without a warning. We circumvent a lot of safety checks normally done.

WARNING: We cannot make sure that everything works as it should, expect broken code at runtime!

Supported python versions:

  • 3.8 and below: unsupported; will produce syntax errors due to using new features
  • 3.9 (WIP)
  • 3.10 (main development)
  • 3.11[a4] (forward porting; WIP)

Other version may work, but due to possible internal changes, we do not recommend using them together with this library!

Why are there so many print()-s?

Due to the breaking nature of anything this code touches, and the absents of any traces in the function itself, we decided to add a lot of "debug" statements indicating mostly the who-has-done-what-to-which-method for the runtime. This makes debugging broken code easier, as it is more clear what happened to each transformed function.

If you want them removed, create your own Fork of this and remove them, on your own risk.

We may use in the future the logging library, so you can disable our logger instance, but we are currently in an inter-stage of the code, so other stuff has priority.

Compatibility with other libraries

Debugging your injections

There is the possibility to "debug" functions using the execution emulator. It will be able to give you more information about a crash than the python core interpreter, but will be a lot slower than it.

It comes also with the possibility to run your bytecode in another interpreter version, so you can experiment with some stuff. In theory, it is also possible to run in python versions not supported by the bytecode manipulation system, but it is not recommended.

TransformationHandler() takes as an arg debug_code and debug_further_calls for activating it for all accessed methods.

BytecodePatchHelper() contains a method enable_verbose_exceptions() for activating it on that exact method.

Examples

Replacing a code constant (in the whole function body):

from bytecodemanipulation.Transformers import TransformationHandler

handler = TransformationHandler()


def test():
  return 0


handler.makeFunctionArrival("test", test)

# Replaces the constant '0' from the return with a '1'
handler.replace_method_constant("test", 0, 1)

handler.applyTransforms()

assert test() == 1

Code Formatting

We use the python formatting library "black" on our code

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

bytecodemanipulation-0.1.0.tar.gz (3.5 kB view hashes)

Uploaded Source

Built Distribution

bytecodemanipulation-0.1.0-py3-none-any.whl (3.5 kB view hashes)

Uploaded Python 3

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page