chore: improve checkstyle double equals test regex & types

[hyperupcall]

Summary

General fixes for checkstyle.py, including a false positive fix for the no-test-double-equals rule.

[jthegedus]

If this change fixes a false positive, why does the lint script not show so in the PR checks?

[hyperupcall]

@jthegedus If I am understanding you correctly, the PR checks don't change because no existing code in main is matched by the old regular expression. Code from #1456 does match though (in particular, the [ ${line[0]} == 'usage: '* ]).

[jthegedus]

I thought the Workflows executed from the forked code, not what was on the target branch. https://docs.github.com/en/actions/managing-workflow-runs/approving-workflow-runs-from-public-forks

Which would lead me to think that this change hasn't matched any previous false positives.

[hyperupcall]

@jthegedus I don't think we are in disagreement - this change doesn't match any previous false positives, but it will address potential future potential ones ones. Specifically, the false positives are shown in the CI errors of #1456, which yes are executed from the forked code. I think we agree that these changes will fix those false positives?

Maybe it's weird that I introduced this fix separately instead of bundling it together with #1456? I'm not sure

[Stratus3D]

@hyperupcall I think I understand. As you say the false positive wasn't present in the codebase (since the wrong regex in this checkstyle.py script would have prevented it from ever being added).

@jthegedus you can validate this by reverting the change to line 125/old line 118 in checkstyle.py and running the file's tests like this: ./scripts/checkstyle.py --internal-test-regex. It'll print this:

no-test-double-equals: Failed negative test:
=> [[ "${lines[0]}" == 'usage: '* ]]

[Stratus3D]

Code seems fine but it's not clear to me why this check is needed at all.

The description states: Although it is valid Bash, it reduces consistency and copy-paste-ability. Is that all there is too it?

[hyperupcall]

@Stratus3D yes, that's it!

The check was originally added to help address #1313. Maybe the error message should change to be more descriptive?

[jthegedus]

Much better error msg, thanks!