DEBUGPY-RUN

This utility facilitates debugging a Python command line program using Python extension in Visual Studio Code.

The Python debugger in VS Code is superb. However debugging a command line program which takes arguments is a little awkward to invoke. The official instructions require you to edit the command line arguments in your launch.json configuration which is cumbersome to do when you want to change arguments for each run, particularly because the arguments have to be quoted within a JSON data structure. This question on stackoverflow describes the problem, but there is no adequate solution. This utility provides a solution, follow the procedure to use it below.

If you have the VS Code Python extension installed then the full debugpy debugger is already bundled with it. You open a terminal window and run this utility to invoke your program with arguments. The utility finds the path where debugpy is installed and then runs it for the program and arguments you specify, in listen mode. Connect to it from within VS Code using the Python Remote Attach debug configuration (using the default host and port settings). You can control+c and then re-run the command with changed arguments using your shell history and command line editing facilities, for each debug run. You can also run debugpy-run remotely, with debugpy explicitly installed for this case, to debug from VS Code to a remote machine over a network.

This utility was developed on Arch Linux but should work on all Linux systems where VS Code is installed with the Python extension. The latest version and documentation is available at https://github.com/bulletmark/debugpy-run.

Installation

Arch users can install debugpy-run from the AUR.

Python 3.6 or later is required. Also, the Python packaging module is required. Note debugpy-run is on PyPI so just ensure that pipx is installed then type the following:

$ pipx install debugpy-run

To upgrade:

$ pipx upgrade debugpy-run

To remove:

$ pipx uninstall debugpy-run

Procedure to Use

1. Open VS Code for the directory where your command line program is located.

2. Ensure you have added a Debugging Configuration in your launch.json. Specify Remote Attach and just accept the default arguments (i.e. host = localhost, port = 5678). You only have to do this once for each project.

3. Open a terminal (either within VS Code, or external) and type:

   $ debugpy-run my-program -- --myargs
   

   Now debugpy-run will start the debugpy debugger for your program, output a message, and then wait to be connected by VS Code.

   ⚠️ As seen in the above example, you should specify -- after your program/module name so that debugpy-run knows where it's own options end, and the target options start.

   ⚠️ You must run my-program from same work directory as VS Code is using otherwise VS Code will not correlate your source file path with the code being run by debugpy. I.e. if the program is located at src/my-program with respect to the work directory then you must run it as debugpy-run src/my-program -- --myargs

4. In VS Code, start debugging, e.g. set a breakpoint then start the Remote Attach debug session.

5. At any point you can control+c the terminal command and restart it with new command line arguments (e.g. using the convenience of your shell history and editing commands) and then restart the debug session in VS Code.

Remote Debugging On Another Host

The debugpy-run utility first looks to find the debugpy package in your local ~/.vscode/extensions directory. If it fails to find that then debugpy-run next tries to import debugpy globally. This is is done so you can install both debugpy-run and debugpy on a remote headless server (e.g. where VS Code is not installed) and then debug a program on that server from VS Code on your laptop/PC remotely over the network.

So for example, I may have a program which runs on a server which want to debug from VS Code on my laptop. I first make sure I install the necessary software on the server (you can also do this in the programs virtual environment of course):

$ pipx install debugpy
$ pipx install debugpy-run

The start my program on the server using the debugger:

$ debugpy-run -p :5678 my-program -- --myargs

NOTE: We need to explicitly specify the :port for this case so that the port is opened on the external network interface so we can connect to it from another machine. By default, debugpy-run/debugpy otherwise only accept local connections.

Then I go back to my laptop, ensure I have set up Remote Attach debugging configured with host = my-server and port = 5678, then start debugging.

Of course, you could start debugpy directly yourself on the server but the debugpy-run wrapper is more convenient to use and makes the usage consistent with the familiar way you start debugpy-run on your laptop/PC.

Debugging A Program Running As Root

Another application of debugpy-run is that you can, as your normal user, easily use VS Code to debug a program you run as root. E.g. run a program using sudo:

$ sudo debugpy-run my-program -- --myargs

Now you can just Remote Attach to it in VS Code as your normal user.

Usage

Type debugpy-run -h to view the usage summary:

usage: debugpy-run [-h] [--listen | -C] [-W] [-p PORT] [-E] [-A] [-r]
                      [--log-to PATH | --log-to-stderr]
                      [-m MODULE | -c CODE | --pid PID] [-V]
                      [program] ...

Finds the "debugpy" package within your VSCode Python extension and then runs
it for "remote attach" debugging of the program/module you specify. If not
found in extensions, or bundled with this app, then tries to run the
global/venv installed "debugpy".

positional arguments:
  program               python program to execute and debug
  args                  remaining arguments to debug

options:
  -h, --help            show this help message and exit
  --listen              listen on given port, default=True
  -C, --connect         connect to given port rather than listen
  -W, --no-wait         do not wait on listen for client, start immediately
  -p PORT, --port PORT  [host:]port to use, default=5678
  -E, --no-extension    don't use the debugpy bundled in the extension
  -A, --no-app          don't use the debugpy bundled in this app
  -r, --run-on-error    re-run program/module even on error
  --log-to PATH         log to given path
  --log-to-stderr       log to stderr
  -m MODULE, --module MODULE
                        python module to execute and debug
  -c CODE, --code CODE  python code to execute and debug
  --pid PID             python pid to attach and debug
  -V, --version         output debugpy path and version

License

Copyright (C) 2021 Mark Blakeney. This program is distributed under the terms of the GNU General Public License. This program is free software: you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation, either version 3 of the License, or any later version. This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License at http://www.gnu.org/licenses/ for more details.