Code Style: enforce named arguments for literals

[alrz]

Note: there is already a refactoring for adding named arguments.

[jcouv]

📝 I expect we'll add support for non-trailing named arguments in 15.5. Then enforcing that literals be named will make more sense (be less annoying).

[jcouv]

The non-trailing named arguments feature is merged to master, so this can go ahead.

[dpoeschl]

@jcouv What's the status of this one? I thought you had it almost working.

[jcouv]

One issue remains in the PR (how the new codefix interacts with the existing refactoring). I should be able to merge by next week (in time for 15.5 preview 2).

[jcouv]

Quick update: from discussions following the issue I mentioned above, that PR is snowballed. The existing refactoring would need to be removed and the code migrated into the new codefix.

[jcouv]

I've dropped my PR with some rationale. If anyone is interested, feel free to pick up this issue.

[alrz]

running such a fixer on the Roslyn codebase was actually rather annoying (consider all the Assert methods, with literals for the expected values).

probably it's best to limit this to invocations with >2 arguments.

[CyrusNajmabadi]

Unlikely to do this given lack of feedback.