objectified/python-debug repository overview

Debugging Python applications using gdb inside a virtualenv is a bit of a pain (especially on e.g. OS X), so I solved that using Docker.

This image allows you to debug Python applications using gdb's Python support. In order to use it, cd to your project directory, and start a container using the following command:

docker run -ti --privileged --cap-add=SYS_PTRACE \
    --rm -v $PWD:/project \
    objectified/python-debug

From there on, you can pip install your requirements (pip is available inside the container, you can do pip install -r requirements.txt) and start your application either through gdb or as a separate process to which you let gdb connect.

For example, do this inside the container you just started:

$ cd /project
$ pip install -r requirements.txt
$ python your-longrunning-app.py

Now get another shell on the container:

$ docker exec -ti <your container id> /bin/bash

Changedir to /project, find the pid of the Python process you started earlier, and connect to it using gdb:

$ cd /project
$ gdb python <your pid>

From then on, you can do things like list threads, py-list (to see the current stack position), py-locals (to see variables in the current scope) and so on.

More info here:

https://wiki.python.org/moin/DebuggingWithGdb⁠