Don't count identifier expressions as variable usages in StrictUnusedVariable check

[IlanaRadinsky]

Before this PR

If you have a class like this

public class Test {
    private final String myField;

    public Test(String myField) {
        this.myField = myField;
    }
}

then the StrictUsedVariable check fails with

error: [StrictUnusedVariable] The field 'myField' is never read. Intentional occurrences are acknowledged by renaming the unused variable with a leading underscore. '_myField', for example.
    private final String myField;
                         ^

But if I apply the suggested fix

public class Test {
    private final String _myField;

    public Test(String myField) {
        this._myField = myField;
    }
}

I also get a StrictUnusedVariable error

error: [StrictUnusedVariable] The field '_myField' is read but has 'StrictUnusedVariable' suppressed because of its name.
    private final String _myField;
                         ^

See a1b70a5 https://app.circleci.com/pipelines/github/palantir/gradle-baseline/4165/workflows/756c72ee-8297-4da5-837a-40a3b010ad09/jobs/34598 for example of failing test before change.

After this PR

==COMMIT_MSG==
Don't count identifier expressions as variable usages in StrictUnusedVariable check
==COMMIT_MSG==

Possible downsides?

[changelog-app]

Generate changelog in changelog/@unreleased

Type

• Feature
• Improvement
• Fix
• Break
• Deprecation
• Manual task
• Migration

Description

Don't count identifier expressions as variable usages in StrictUnusedVariable check

Check the box to generate changelog(s)

• Generate changelog entry

[carterkozak]

Hmm, in this case I think we'd prefer to suggest removing the variable and assignment entirely rather than allowing underscores. That can be a bit tricky to auto-suggest, though.

[IlanaRadinsky]

@carterkozak Would it be preferable to update the error message without the auto-suggest, or just to close this PR?

[carterkozak]

Yep, ideally we would fail due to unused fields, but we wouldn't suggest renaming anything.

[IlanaRadinsky]

Tracking in #2361 since I haven't had a chance to come back to this yet.