[kunalgoesbyken]

Summary

Kotlin calls were resolved by name against functions only, so a call on a constructor property, a parameter, a typed or constructed local, an object or companion, or an unqualified member of the same class produced no edge, while the Java twin of each resolved. #431 reports the constructor-property case; the same missing receiver typing breaks all of them.

Changes

• bindings.ts: a Kotlin case in defName (so binding scopes line up with extract.ts) and a handleKotlin that binds val/var constructor parameters as fields, function parameters, and typed or constructor-initialised properties and locals. A plain constructor parameter is not visible to the methods and binds nothing. An UpperCamelCase receiver with no binding is the type (Utils.helper(), Registry.create()), as in Swift.
• A name declared again with a type we cannot read hides the outer binding instead of inheriting it: a lambda parameter, a for variable, a destructured name, an untyped local or property, a function-typed parameter.
• extract.ts: this.repo.save() reads as a field access, this@Outer.save() as a call on Outer, and a lowercase bare call inside a type resolves to the member first and the top-level function otherwise, as Swift does. A bare call inside apply, run, with or a buildX lambda is not read as a member of the enclosing class, since its receiver is another object.

$ graft callers FooService.handle      # before
handle · method · FooService.kt:L2-L2
  no indexed callers

$ graft callers FooService.handle      # after
handle · method · FooService.kt:L2-L2
  calls ← trigger (Bar.kt:L2-L4)
      3: fooService.handle(42)
  calls ← run (Bar.kt:L7-L9)
      8: service.handle(1)

An extension function, a call on a chained result (items.first().load()) and an overload set that differs only by parameter types stay unresolved.

Testing

On Windows 11 with Node 26:

• test/graph-kotlin-members.test.ts (new, five tests): property, this. property, parameter, local, object, companion and implicit-this calls; an unknown member, an unbound receiver and a plain constructor parameter resolving to nothing; the shadowing cases; receiver lambdas; and this@Outer. Four of the five fail on main; the second only asserts that nothing resolves.
• test/graph*, blast*, ask*, generic*, depth*, resolve* and covers* suites: 677 tests, 1 fail. The one failure is discoverWorkspaceChildren follows a symlinked child repo, a symlink permission error on Windows that fails the same way on main.
• tsc --noEmit passes.

Closes #431

[trailhq-graft]

🌱 graft blast radius

Nothing outside this diff depends on it. 1 area changed; no indexed dependents at depth 2.
Tests: no test reaches Graph Bindings.
Tag: L4XB — Graph Bindings · Nguyen Thanh Dat — Graph Bindings

Who knows this code — 2 people across 1 area

Area	Who knows it
Graph Bindings · changed	L4XB — 1 commit, last today · Nguyen Thanh Dat — 1 commit, last today

Ownership is git history over each area's own files, weighted towards recent work (120-day half-life). Merge commits and bots are dropped, and you are dropped from your own PR. A name with no @ has no GitHub handle in its commit email — tag them by hand, or add a .mailmap entry. A suggestion from history, not a CODEOWNERS rule.

Test signal per changed area — 1 ✗

Reached = a node under a test path has a resolved edge into the changed symbol. It undercounts anything called indirectly — through a CLI, a spawned process or a dynamic import — so read a low ratio as “look here”, never as a coverage gate.

• ✗ Graph Bindings — 0 of 14 reached · no test file reaches it
  • not reached: set, shadow, lookup, defName, kotlinDefName, resolveRecvType, isClassNode, visit, …6 more

_{graft blast · refs/graft/base...HEAD · depth 2 · 3 changed files}

Open the interactive graph → — click an area to see the code that changed, and the line that reaches it.