🐛 false positive noInvalidConstructorSuper

[sonnyp]

I think it's a regression in 12.1

Environment information

Rome 12.1.0
https://github.com/sonnyp/Workbench/

{
  "formatter": {
    "enabled": true,
    "ignore": [
      "src/icon-development-kit.json",
      "icon-development-kit-www",
      "src/lib",
      "troll"
    ],
    "indentStyle": "space",
    "indentSize": 2
  },
  "linter": {
    "enabled": true,
    "ignore": [
      "src/icon-development-kit.json",
      "icon-development-kit-www",
      "src/lib",
      "troll"
    ],
    "rules": {
      "recommended": true,
      "style": {
        "useBlockStatements": "off",
        "noParameterAssign": "off"
      }
    }
  },
  "vcs": {
    "enabled": true,
    "clientKind": "git",
    "useIgnoreFile": true
  }
}

What happened?

The rule lint/correctness/noInvalidConstructorSuper triggers when extending an object property, for example:

// KO 
class CodeView extends Gtk.Widget {
  constructor() {
    console.log('cool');
    super(); // This class calls super(), but the class extends from a non-constructor.
  }
}

// OK (removed the dot in Gtk.Widget)
class CodeView extends GtkWidget {
  constructor() {
    console.log('cool');
    super();
  }
}

Expected result

Extending an object property shouldn't trigger noInvalidConstructorSuper

Code of Conduct

• I agree to follow Rome's Code of Conduct

[sonnyp]

Merci!