[coverage] Conformance findings: CLOUDFETCH-017

[peco-engineer-bot]

Summary

Surfaced by the multi-language coverage fan-out while conformance-testing these SPEC-IDs against adbc-drivers/databricks. 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 (adbc-drivers/databricks) is fixed, then flips green as a tripwire.

Findings

• CLOUDFETCH-017 [thrift, sea]: A non-numeric adbc.databricks.cloudfetch.prefetch_count (the CloudFetch in-memory chunk bound) throws ArgumentException instead of falling back to the default bound with a diagnostic, so a tuning typo fails the query rather than degrading gracefully
  • failing test: CloudFetch_InMemoryChunkBoundConfiguration (see the coverage PR diff under tests/)
• CLOUDFETCH-017: A non-numeric adbc.databricks.cloudfetch.prefetch_count (the CloudFetch in-memory chunk bound) throws ArgumentException on both Thrift and SEA instead of falling back to the default bound with a diagnostic, so a tuning typo fails the query rather than degrading gracefully

Reproduce & Expected

CLOUDFETCH-017 — A driver-level bound on how many decompressed CloudFetch chunks are held in memory at once (which is what caps peak RSS on a large result set) must behave as pure tuning: 1.

Reproduce:

SELECT id FROM range(0, 5000000)

Expected (per the shared spec):

• result has exactly 5000000 row(s)
• result has exactly 5000000 row(s)
• result has exactly 5000000 row(s)
• result has exactly 5000000 row(s)
• completes without an exception
• completes without an exception
• [thrift] OpenSession request configuration.cloudfetch_max_chunks_in_memory is absent
• [sea] CreateSession request session_confs.cloudfetch_max_chunks_in_memory is absent

Context

• The behavior was first fixed in a DIFFERENT driver — reference PR: https://github.com/databricks/databricks-odbc/pull/181 — which seeded the shared language-neutral spec. This issue tracks the same conformance gap in adbc-drivers/databricks; 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/1177