Feature Request: `swiftlint --fix-errors-only` (or, make `--fix` apply to errors only, and require `--fix --strict` for warnings)

Feature Request

• I searched for existing GitHub issues

swiftlint --fix-errors-only

upon entering a new code base into which i would like to inject swiftlint, there are already 700+ warnings that is not yet ready to agree upon. this is not the first time this has happened. we would like to iterate over the warnings, leaving many of them in place to remind us of things that we want to address … but we also want to use the --fix feature to be able to auto-correct those issues we've agreed should be errors.

frankly, i think --fix should apply to errors only, and if one wants to run swiftlint --fix --strict, that would take care of treating warnings as errors. but i understand that this may cause the existing behavior to change for many who rely on warnings when performing auto-correct.

thus, --fix-errors-only would be a way to address this while maintaining backward compatibility with those who may already rely on this.

Start at SwiftLint's command-line handling for --fix and --strict, then trace how warnings and errors are selected for autocorrection. Define the new flag or revised combination so errors can be fixed without changing existing --fix behavior, and verify that warning handling matches the chosen compatibility approach.