Extend library_call_linter for character.only usage

[MichaelChirico]

Part of #884

Here I think the main question is whether this might be subsumed into library_call_linter().

[codecov-commenter]

Codecov Report

All modified and coverable lines are covered by tests ✅

Comparison is base (0dcd049) 99.64% compared to head (8a172a8) 99.64%.

❗ Current head 8a172a8 differs from pull request most recent head b95004b. Consider uploading reports for the commit b95004b to get more accurate results

Additional details and impacted files

@@           Coverage Diff           @@
##             main    #2279   +/-   ##
=======================================
  Coverage   99.64%   99.64%           
=======================================
  Files         115      115           
  Lines        5285     5331   +46     
=======================================
+ Hits         5266     5312   +46     
  Misses         19       19           

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[MichaelChirico]

NB: holding off on adding examples until we decide whether this should be subsumed into library_call_linter().

[AshesITR]

I'm in favor of subsuming - the linter name contains library that way. Someone unfamiliar with character.only will probably wonder what character_only_linter() does.

[MichaelChirico]

I'm in favor of subsuming - the linter name contains library that way. Someone unfamiliar with character.only will probably wonder what character_only_linter() does.

OK, done. Follow-up: Should either behavior (existing top-of-script enforcement vs. new character.only rules) be made optional with a parameter?

[AshesITR]

For now, I wouldn't create an option unless we get a feature request. The intention of the two lints appears to be the same - getting a clean library() call list at the top of scripts.