Debug-custom-Pylint-check-tests.md

Table of contents

• Introduction
• Run a single checker
  • Print statements
  • The Python debugger

Introduction

When trying to debug the tests for a custom Pylint checker, remember that the tests are just like any other backend test. You may find our [[guide to debugging backend tests|Debug-backend-tests]] useful. That said, there are some helpful tips for working with Pylint checkers in particular. We describe them below.

Run a single checker

Debugging a Pylint checker's tests often comes down to debugging the checker itself. To see how your checker behaves on some code, put that code into test.py. Then assuming that your checker has message with symbol my-checker-message, you can run your checker on test.py like this:

$ PYTHONPATH=scripts/linters ../oppia_tools/pylint-2.8.3/bin/pylint --load-plugins=pylint_extensions --disable=all --enable=my-checker-message test.py

--------------------------------------------------------------------
Your code has been rated at 10.00/10 (previous run: 10.00/10, +0.00)

(These instructions come from the Pylint documentation.) Note that the Pylint version number (2.8.3 in this case) may vary.

In this case, the checker found no issues with our code in test.py, but your checker might flag problems with the code in test.py. In either case, you should consider whether your checker behaved as expected. Did the checker flag any code that isn't actually a problem? Did the checker miss any problematic code? If the checker behaved unexpectedly, you should debug further. You may find two other debugging strategies useful: print statements and the Python debugger.

Print statements

You can print out debugging information from your checker code using print() statements. The output will appear when you run the checker as shown above. For example, if we added a print('DEBUGGING') statement to our checker, we might see:

$ PYTHONPATH=scripts/linters ../oppia_tools/pylint-2.8.3/bin/pylint --load-plugins=pylint_extensions --disable=all --enable=my-checker-message test.py
DEBUGGING

--------------------------------------------------------------------
Your code has been rated at 10.00/10 (previous run: 10.00/10, +0.00)

The Python debugger

Just like when debugging backend tests, you can use the Python debugger, pdb, with Pylint checkers. You just need to add this code to your checker:

import pdb; pdb.set_trace()

If you aren't familiar with pdb, take a look at our [[guide to debugging backend tests|Debug-backend-tests]].