Fix inconsistent argument exit code when argparse exit with its own error code

[dmrlawson]

Type of Changes

	Type
✓	🐛 Bug fix

Description

We were bitten by this causing us to think that pylint was completing successfully when in fact a typo in a configuration file was causing argparse to exit with exit code 2. This didn't happen in older versions of pylint for whatever reason. The documentation for pylint suggests that this means "error message issued" but that pylint executed properly when this is not the case.

It feels a bit clumsy to just add in try..except, but on the other hand it's perhaps unexpected that argparse will raise a SystemExit. I'm open to any suggestions to improve this.

[Pierre-Sassoulas]

Nice catch, we indeed missed that argparse could exit with its own exit code and the inconsistency with pylint's error codes. I think we should use -1 because it's the only way our bit encoded error code can be extended to 64 if required. Thoughts @DanielNoord ?

[dmrlawson]

Nice catch, we indeed missed that argparse could exit with its own exit code and the inconsistency with pylint's error codes. I think we should use -1 because it's the only way our bit encoded error code can be extended to 64 if required. Thoughts @DanielNoord ?

sys.exit(32) appears throughout the code. The "usage error" exit code could be factored out to be a constant and then changed but it's a bit more than I was intending here.

$ grep -r 'sys.exit(32)'
pylint/pyreverse/utils.py:        sys.exit(32)
pylint/pyreverse/utils.py:        sys.exit(32)
pylint/lint/run.py:            sys.exit(32)
pylint/lint/run.py:            sys.exit(32)
pylint/lint/run.py:            sys.exit(32)
pylint/lint/run.py:                sys.exit(32)
pylint/config/arguments_manager.py:            sys.exit(32)
pylint/config/config_initialization.py:        sys.exit(32)
pylint/config/config_initialization.py:            sys.exit(32)

[Pierre-Sassoulas]

You're right let's keep it as 32, I wasn't up to date on this aspect of the code sorry.

[Pierre-Sassoulas]

Could you fix the tests (there's some other test that is somewhat fragile you'll need to add the new file to it) and add a changelog entry please ?

[Pierre-Sassoulas]

Great job 👌

[coveralls]

Pull Request Test Coverage Report for Build 3699121501

• 8 of 8 (100.0%) changed or added relevant lines in 2 files are covered.
• No unchanged relevant lines lost coverage.
• Overall coverage increased (+0.001%) to 95.476%

---

Totals
Change from base Build 3698112388:	0.001%
Covered Lines:	17663
Relevant Lines:	18500

---

💛 - Coveralls

[github-actions]

🤖 According to the primer, this change has no effect on the checked open source code. 🤖🎉

This comment was generated for commit 244e1cf