Files
code_of_conquest_dnd/client/tests
Phillip Tarrant c5f127e0f1 test(creation-screen): close 3 vacuous guards — point-buy display, signal wiring, Luck sweep
Adversarial re-review of 6589d7c found three assertions in the same species as
the seven it fixed: green against the bug they name.

- test_the_displayed_scores_are_the_draft_s_scores spent zero points, so
  final() == rolled() and a Box/Value bound to `rolled[stat]` instead of
  `final[stat]` would pass unnoticed. Now presses the Plus node to force a
  divergence, and checks the Box/Rolled sub-label too (previously unchecked).
- No test ever emitted a node's `pressed` signal — every interaction went
  through the _on_*_pressed handler directly, so _wire() itself, and every
  card-to-index binding, had zero coverage. Minus had no coverage of any
  kind. Helper and interaction tests now press the authored nodes; added
  dedicated press tests for race/calling/pool/bonus/minus, plus the
  Plus/Minus disabled-boundary bindings.
- The Luck sweep only ever checked the default binding (human + sellsword),
  covering 1 of the 28 race x calling combinations whose authored prose
  (fragment/blurb) actually renders on screen. Now sweeps all 28.

Plus three cheap adds: a degenerate-origin (§13) guard, orphan-authored-card
counts for the race/calling/bonus rows, and a non-empty assertion on the
origin-card RichTextLabel.

Each fix verified by deliberately reintroducing its named bug, confirming the
guard test goes red, then reverting — see
.superpowers/sdd/task-6-report-fix2.md. 291 -> 298 tests, all green.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-13 15:38:31 -05:00
..