[coverage] Conformance findings: STATEMENT-023,STATEMENT-025

The issue describes two conformance test failures related to query timeout handling in the Thrift layer. Start by examining the failing tests in the coverage PR (linked). Look for the Thrift client code that sets TExecuteStatementReq.queryTimeout. For STATEMENT-025, validate that negative timeouts are rejected. For STATEMENT-023, ensure the timeout is enforced during query execution, not just set. The tests provide SQL to reproduce and expected YAML output.

Summary

Surfaced by the multi-language coverage fan-out while conformance-testing these SPEC-IDs against databricks/databricks-sql-nodejs. Each finding is committed as an expected-failure (xfail) test in the coverage PR — the test asserts the CORRECT (post-fix) behavior and stays red until THIS driver (databricks/databricks-sql-nodejs) is fixed, then flips green as a tripwire.

Findings

• STATEMENT-025 [thrift]: a negative queryTimeout is silently accepted and forwarded verbatim as TExecuteStatementReq.queryTimeout with no range validation, so the statement runs unbounded instead of the invalid value being reported to the caller
  • failing test: STATEMENT-025 — queryTimeout=0 is unlimited and a negative queryTimeout is rejected (see the coverage PR diff under tests/)
• STATEMENT-023 [thrift]: queryTimeout is accepted but never enforced at execution on Thrift — the driver polls to the query's natural completion (23.9s against a 5s deadline) instead of aborting at ~the deadline, the same accepted-but-never-applied defect class as databricks-odbc#136
  • failing test: STATEMENT-023 — queryTimeout aborts a long query at ~the deadline and leaves the session usable (see the coverage PR diff under tests/)

Reproduce & Expected

STATEMENT-023 — Validates that a configured client-side query timeout is ENFORCED at execution time -- not merely stored and echoed back by the getter (STATEMENT-021 asserts only that some timeout error surfaces;…

Reproduce:

SELECT COUNT(*) FROM range(0, 40000000) a JOIN range(0, 40000000) b ON a.id % 5000 = b.id % 5000 -- timeout nonce <unique-per-run>
SELECT 1

Expected (per the shared spec):

• completes without an exception
• result has exactly 1 row(s)
• full assertion contract:

result:
- label: timed_out
  error:
    contains:
    - timeout
    - timed out
    - cancel
- label: timed_out
  sql_state: HYT00
- label: timed_out
  elapsed_seconds_range:
    min: 4
    max: 25
- label: reused_statement
  no_exception: true
- label: reused_statement
  row_count: 1

STATEMENT-025 — Validates the BOUNDARY VALUES of the driver's client-side query-timeout surface -- the same surface STATEMENT-023 exercises with an ordinary finite deadline.

Reproduce:

SELECT 1 AS test_value
SELECT 1 AS test_value

Expected (per the shared spec):

• completes without an exception
• result has exactly 1 row(s)
• col 0, row 0 == 1 (type Int32)
• full assertion contract:

result:
- label: unlimited
  no_exception: true
- label: unlimited
  row_count: 1
- label: unlimited
  column:
    index: 0
    row: 0
    type: Int32
    equals: 1
- label: invalid
  exception_thrown: true

Context

• The behavior was first fixed in a DIFFERENT driver — reference PR: https://github.com/databricks/databricks-sql-go/pull/484 — which seeded the shared language-neutral spec. This issue tracks the same conformance gap in databricks/databricks-sql-nodejs; the reference PR is for cross-referencing the intended behavior, NOT a change to this repo.
• Coverage PR carrying the reproducing xfail test(s): https://github.com/databricks/databricks-driver-test/pull/1672