aglinxinyuan opened a new issue, #7504:
URL: https://github.com/apache/texera/issues/7504

   ### Task Summary
   
   Two files in `agent-service` are well covered by line count and thinly 
covered in substance:
   
   | File | Lines | Missed | Reported |
   |---|---:|---:|---:|
   | `src/agent/util/context-utils.ts` | 212 | 47 | 77.8% |
   | `src/agent/tools/workflow-execution-tools.ts` | 434 | 38 | 91.2% |
   
   The percentages understate the gap, because `bun` credits an entire function 
body once it is
   entered. Whole blocks inside `jsonToTableFormat` and 
`executeOperatorAndFormat` — the
   `__row_index__` gap row, the `NULL`/`null`/`undefined`/object cell mapping, 
the leading-tab header,
   the warnings line, the `totalRowCount` fallback — were already counted as 
covered while nothing
   asserted them.
   
   That matters because these two files decide what the model sees. 
`context-utils.ts` serializes the
   DAG into the prompt: topological order, port ordinals, schemas, redaction of 
properties when an
   operator has errored. `workflow-execution-tools.ts` turns an execution 
result back into the table
   the model reads. A defect in either is invisible in the UI and shows up only 
as the agent quietly
   reasoning over the wrong thing.
   
   Note for whoever picks this up: measure with `bun test --coverage 
--coverage-reporter=text`, which
   prints exact uncovered line numbers, then **mutate** — the percentage will 
not tell you which of
   those lines are actually pinned.
   
   ### Task Type
   
   - [ ] Refactor / Cleanup
   - [ ] DevOps / Deployment / CI
   - [x] Testing / QA
   - [ ] Documentation
   - [ ] Performance
   - [ ] Other
   


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]

Reply via email to