Conversation
The mock was a passive WS↔MQTT pipe, so every bridge/request/* from the app silently dropped (no matching bridge/response/*) and <device>/set never flowed back as a state update. That's the root cause of "can't change states" in the dev stack. - seeder.py: full rewrite as a stateful engine. Owns authoritative state for devices, groups, bridge info and health; merges /set payloads into state and republishes; answers every bridge/request/<path> with a matching bridge/response/<path> envelope; emits bridge/event for rename/leave/interview/permit_join; simulates OTA progress ticks. Header docstring enumerates what we faithfully simulate vs. stub. - fixtures.py: 34 devices (up from 9) + 3 groups, covering lights, plugs, sensors, motion/contact/leak/smoke/vibration, switches/dimmers/buttons, curtain, siren, knob, locks, outdoor sensor. Retains the 9 original friendly_names UI tests rely on; renames Living Room Air Purifier to Bathroom Fan to match DeviceDetailUITests expectations. - bridge.py: stop echoing client command topics (/set, /get, bridge/request/*) back to WS subscribers — real z2m never does, and the echo was being misrouted as a device-state update by the app. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
UI tests were hitting the bridge at ws://localhost:8080/api with no ?token= query string, so the bridge (AUTH_TOKEN enforced) closed every connection with 1008 Invalid token. Tests that 'passed' were asserting against the app's empty-state UI; any test that needed real data from the bridge (groups, device state, logs) failed. - UITestHelpers.launchForTesting now passes UI_TEST_Z2M_TOKEN. - AppEnvironment.start reads UI_TEST_Z2M_TOKEN and forwards it as the ConnectionConfig authToken. - bridge.py logs every WS connect path + reject reason so future auth issues are obvious in container logs. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
- Add XCUIApplication.revealSearchField() helper for views that use .searchToolbarBehavior(.minimize); scrolling does not reveal the field. - DeviceListUITests/DeviceDetailUITests: use the helper instead of the old swipeDown approach. - GroupsUITests: AddGroupSheet and AddSceneSheet dismiss via drag indicator (they have no Cancel button by design). Update the three tests that assumed a Cancel button to dismiss the sheet the way the app actually wants users to. Each touched test now has a Behavior: comment describing what it is meant to verify. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Swipe-actions tests used `app.cells.firstMatch` which could land on a section header (DeviceListView groups by category by default) and produce no swipe response. Target a known device by name instead. The default XCUIElement.swipeLeft() is too short on iOS 26 to reveal trailing swipe actions when `allowsFullSwipe: false`. Added swipeLeftFar() helper that does a far-travel press-drag gesture. RenameDeviceSheet and RemoveDeviceSheet dismiss via the drag indicator (no Cancel button by design) — update tests to dismiss via swipeDown. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Nested NavigationStacks (LogsView/MainSettingsView wrap themselves in their own NavigationStack inside the SettingsView push) mean the assertion 'navigationBars.element(boundBy: 1) exists' never holds — both titles render into the same visible nav bar. - Settings tests: assert on navigationBars[title] for each pane (Server, General, MQTT, Adapter, Home Assistant, Logs). - testGeneralSettingsApplyAndCancel: Apply is always rendered in the confirmationAction slot but starts DISABLED; Cancel only appears once changes exist. Assert the correct enabled/existence state. - testGeneralSettingsHasLogLevelPicker: match the staticText row label instead of the enclosing cell. - seeder.py: emit a debug bridge-log line on every drift tick so the Bridge Log view has content when tests arrive (the startup log line is non-retained and races the app connect). - SEED_INTERVAL default lowered to 10s so activity-log state diffs arrive within test timeouts. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
- Connection editor: SettingsTextField does not set an accessibility identifier — locate the Host and Token rows by their enclosing cell containing the label StaticText, then type into the nested text field or secure field. Supply the mock bridge's auth token so the Connect flow reaches Main tab bar. - Protocol picker test: accept any of StaticText "Protocol", Button "Protocol", or a SegmentedControl — iOS 26 renders .automatic in a Form as a menu button, not a Picker. - PermitJoin: duration options are inside a Menu-style Picker, not always visible. Assert the Duration section header + Preset label are present. Dismiss via repeated app swipeDown so medium-detent sheets collapse cleanly. - Logs entry detail test falls back to Activity (drift-produced rows) and asserts the Logs back button appears after push. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Two tests expected behaviour that pre-dates commit e631ccf (the "device type specific configuration" boilerplate filter) and a paragraphFallback: false decision in makePairingGuide. - testNormalizesNestedPairingFromNotes: paragraphs in a Pairing subsection become the summary, not imitation steps. Only real stepList blocks populate primarySteps (avoids double-display). Added Behavior: comments spelling this out. - testCollectsOptionsAndPreservesResidualBlocks: switched the residual paragraph to a non-boilerplate sentence so miscSections count stays 1. - Added testOptionsBoilerplateIsStripped to explicitly cover the filter. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
New tests covering the recent Home redesign and OTA-aware notification coalescing: - HomeLayoutStoreTests: first-launch default (Groups hidden), hide/show persistence, IndexSet move, move-before drop semantics. - HomeSnapshotTests: coordinator exclusion from topology counts, LQI averaging, low-battery / weak-signal counts, PAN ID formatting, permit-join countdown derivation, devicesWithUpdates counter. - NotificationCoalescingTests: OTA occurrences track distinct device entries; displaying(occurrence:) swaps subtitle+device but keeps the aggregated count and log-ID list. - HomeUITests: Recent Events "Show All" opens Logs; Mesh card chevron pushes MeshDetailView. DeviceDocNormalizerTests: replace the skip-on-missing-local-corpus test with a curated set of inline markdown samples covering lights, sensors, covers, switches, buttons, and locks. Runs in every environment. Every added test carries a `Behavior:` comment. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
- Default setUp timeouts: waitForMainTab 30s → 15s; initial cell waits 20s → 10s; Connection flow 30s → 15s; Logs setup 30s → 10s. - Remove the stray Thread.sleep in testClearLogsEmptiesList. - CLAUDE.md (and AGENTS.md mirror): document the 2-worker parallel test invocation (~40% speedup, 1618s → 968s on a 2024 Mac) and the updated engine behaviour (SEED_INTERVAL=10s, stateful request handler, OTA progress ticks). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
This was referenced May 2, 2026
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
No description provided.