playwright: fix wait_for_hover false-pass on missing widget + cover both #186 sync primitives

[borisbat]

Follow-up to the #186 Copilot review — a real correctness fix plus the regression coverage it asked for.

Fix — wait_for_hover(expected=false) false-pass (#186 review)

wait_for_hover(ident, expected=false) returned the first snapshot when the widget was missing or had no hover field: the predicate find_widget(...)?["hover"] ?? false == expected coalesces a null hover to false, so a typo'd / unregistered ident reads as a successful "not hovered" and masks a real sync failure.

Guard with widget_exists — the same load-bearing check widget_rendered already uses for the identical daslib/json ?[] quirk (find_widget(...) != null is unreliable there; only a non-empty kind proves a real entry). The widget must actually be present before its hover state is compared:

return wait_until_sec(app, timeout_sec) $(var snap) {
    return widget_exists(snap, widget_ident) && (find_widget(snap, widget_ident)?["hover"] ?? false) == expected
}

Same false-positive discipline as widget_rendered / wait_for_series_shown. Only expected=false on a missing ident changes behavior (now times out → null → loud failure, instead of an immediate false pass); no existing caller relied on the old path.

Coverage — tests/integration/test_recording_sync.das (#186 review)

No dasImgui-side test exercised either primitive #186 added. New test (one app spawn, color_button_hover feature — CBH_RED carries the per-widget IsItemHovered telemetry):

• wait_for_hover: registers hover after move_to onto the swatch; clears after moving away (expected=false on a present widget); does NOT false-pass a missing ident (the regression above).
• record_check: returns its ok argument, and accumulates without panicking on false — the accumulate-don't-panic discipline. (The teardown-panic surfacing is exercised transitively by every record_* driver in the suite, which all share the g_record_failures → with_recording_app teardown path.)

Verification (local)

• test_recording_sync green headless.
• Lint + format clean (pre-push verified).

🤖 Generated with Claude Code

[Copilot]

Pull request overview

This PR fixes a correctness edge case in imgui_playwright.wait_for_hover(expected=false) where a missing/misspelled widget ident could incorrectly “pass” as “not hovered”, and adds an integration test covering both wait_for_hover and record_check (introduced in #186).

Changes:

• Fix wait_for_hover(..., expected=false) to require widget_exists(...) before comparing hover state, preventing false positives on missing widgets.
• Add tests/integration/test_recording_sync.das to exercise wait_for_hover (hover set/clear + missing-ident regression) and record_check’s return behavior.

Reviewed changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated 1 comment.

File	Description
widgets/imgui_playwright.das	Adds a widget_exists guard inside wait_for_hover to avoid false-passing on missing widget idents.
tests/integration/test_recording_sync.das	New integration test covering wait_for_hover hover transitions and the missing-ident regression, plus basic record_check behavior.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[Copilot]

Pull request overview

Copilot reviewed 2 out of 2 changed files in this pull request and generated no new comments.