[proggeramlug]

Summary

#3828 (Function.prototype.call/.apply/.bind on function values) was
fixed by #3846 (Function.prototype.call.bind(method) uncurry-this), which
landed on main shortly before this. I verified empirically: building main's
runtime (with #3846, no other changes) makes the #3828 hasown standalone repro
return true/false, and the gap test below is byte-identical to Node.

This PR therefore drops the (now-redundant) runtime changes and keeps only a
regression test.

Why a separate test

#3846's test (test_gap_3716_uncurry_this.ts) exercises the
Function.prototype.call.bind(method) idiom. This test pins the distinct, real-world
function-bind shape that #3828 calls out as the Express (#3527) blocker:

function bind(this, that) {
  var target = this;
  var args = Array.prototype.slice.call(arguments, 1);
  return function () {
    return target.apply(that, args.concat(Array.prototype.slice.call(arguments)));
  };
}
var hasOwn = bind.call(Function.prototype.call, Object.prototype.hasOwnProperty);
hasOwn({a:1}, "a");  // true

i.e. a captured Function.prototype.call value re-applied through a user
function-bind closure with target.apply(that, args.concat(...)) and runtime
(non-literal) args — not the call.bind(...) call form.

It also covers typeof fn.call/apply/bind, object-thisArg call/apply/bind,
the captured Function.prototype.call value, and Object.prototype.hasOwnProperty
as a value.

Validation

• Byte-identical to node --experimental-strip-types on both the default
  (auto-optimize) and PERRY_NO_AUTO_OPTIMIZE=1 paths, against a main-with-fix(runtime): Function.prototype.call.bind(method) uncurry-this (#3716) #3846
  runtime.

Note

The minimal repros (1)/(2) in #3828 use a string thisArg (f.apply("T", …));
the residual {}-vs-"T" difference there is a separate boxed-primitive
JSON.stringify bug, filed as #3857 — unrelated to call/apply/bind dispatch.

[proggeramlug]

Out-of-scope boxed-primitive JSON serialization noted above is filed as #3857.