More concise 'fail-under' >= comparisons

[jtkmckenna]

Type of Changes

	Type
✓	🐛 Bug fix

Description

Change the handling of 'fail-under' from a float to a string. When fail-under's comparison is made, we can round the score to the same precision as the fail-under.

Example, if a project reports a score of 8.79, to fix all pylint warnings is a large task, so we add a CI task that ensures the score only increases on each pull request.

If I use:

pylint iterate --fail-under 8.79

The CI task will run and return a non-zero exit code. This is because the score is more precisely '8.78563061304178', and the comparison

score_value >= linter.config.fail_under

is False. The documentation accurately describes the >= operator, but users can be left confused why the --fail-under only behaves some of the time (ie when the score rounds down in the reported score).
The help string for --fail-under:

Specify a score threshold under which the program will exit with error. (default: 10)

Here in this pull request, I set a local variable with a score rounded to the same number of decimal places as the user provided. If the user provides an integer (or if the default value is set), I do not round the score

[Pierre-Sassoulas]

Thank you for opening a merge request and documenting your proposed changes. I think keeping a float and failing a score of 8.789 when the fail under option is 8.79 is what is expected. If I used --fail-under 9 and the score was 8.6, I would expect pylint to fail.