[Wildcard Variables] Analyzer Implementation

[kallentu]

This meta issue tracks all the analyzer implementation work items. Referenced from: https://github.com/dart-lang/sdk/blob/main/pkg/analyzer/doc/process/new_language_feature.md

Work Items

• AST enhancements (AstBuilder)
  • For both locals and parameters, distinguish between wildcard and regular declarations. Keep element for future use in areas like renaming.
• [Wildcard Variables] Resolution of wildcard import directives #56245
  • May need some changing depending on [Wildcard Variables][spec] _ in imports language#3799
• Element model
• Type system updates
• Summary support
  • Make sure still works as intended.
• Resolution
  • ResolutionVisitor (resolve types)
  • ScopeResolverVisitor (resolve simple identifiers by scope)
    • Don't put wildcard variables in scope.
  • ResolverVisitor (type-based resolution)
• [Wildcard Variables] Verify that constant evaluation works as expected #56244
  • Verify with some tests that constants still work as expected with wildcards.
• Index and search
• Warnings (annotation-based, unused*, strict-mode-based, a few others)
  • ☂️ [wildcards] new and updated diagnostics / lints #56595
  • [Wildcard Variables] implement THIS_ACCESS_FROM_FIELD_INITIALIZER #56442
  • [Wildcard Variables] private_optional_parameter support #56381
  • OverrideVerifier and InheritanceOverrideVerifier (report errors and warnings related to overrides)
    • Make sure we can override methods and use _ for unused parameters. 267b8ef
  • ErrorVerifier (report other errors and warnings)
    • [Wildcard Variables] support for CONFLICTING_TYPE_VARIABLE_AND_MEMBER #56397
    • [Wildcard Variables] Support for wildcard record type positional fields #56271
    • From #initializing-formals, “Even though the parameters no longer collide, it is still an error to have two initializing formals named _”
    • From #super-parameters, “The positional formal parameter super._ is still allowed in non-redirecting generative constructors. Such a parameter forwards the argument's value to the super constructor invocation.”
    • [wildcards] produce a warning on late wildcard variables. #55905
      • Discussion: [Wildcard Variables][spec] Late, wildcarded variables language#3798
    • [wildcards] report a warning on final wildcard variables #55920
    • [Wildcard Variables] Update DUPLICATE_DEFINTION reporting #56089
    • [Wildcard Variables] Update DUPLICATE_FIELD_FORMAL_PARAMETER reporting #56092
  • FfiVerifier (report errors and warnings related to FFI)
  • Unused elements warnings
    • UNUSED_ELEMENT
      • f517dd6
    • UNUSED_LOCAL_VARIABLE
      • [Wildcard Variables] UNUSED_LOCAL_VARIABLE support #55719
    • UNUSED_CATCH_CLAUSE
      • [Wildcard Variables] Update UNUSED_CATCH_CLAUSE reporting #55723
    • UNUSED_CATCH_STACK
      • [Wildcard Variables] Update UNUSED_CATCH_STACK reporting #55721
    • UNUSED_FIELD
      • [Wildcard Variables] Update UNUSED_FIELD reporting #55862
  • Dead code
    • local function declarations: a446cd9
  • @useResult verification should not warn on results assigned to wildcards: a86ec25
• ExitDetector
• NodeLintRegistry

[kallentu]

@pq I updated what needs to be done in the analyzer with a short note from the kickoff. If it's easier to track with singular issues, we can do that too.

I suspect this work can start soon. I'm currently slowly adding language tests to the repo and I usually use those as a good indicator of what's working and what isn't.

[pq]

@kallentu:

From #super-parameters, “An occurrence of super._ as a declaration of a formal parameter in a constructor is a compile-time error.”

This seems out of sync with the current spec. Could you take a look?