Repository navigation
feat(cross-surface): stage tpch and tpcds gates, record CI budget - #2428
Conversation
Wire tpch and tpcds STAGED_GATES on the Q-prefix maps. tpch runs unseeded-vs-unseeded at SF=0.01 (41/44 cells pass; fixed-stream needs seeded DF overrides first). tpcds full-99 staged at SF=0.01 (~13s wall); only Q38/Q88 pass, so the 10-15 blocking subset is deferred to the promotion item. Module docstring states the staged status and honest oracle limits. No GATES promotion, no CI wiring.
Codex Review SummaryThis comment shows the latest Codex review activity on this pull request.
ℹ️ About Codex in GitHubYour team has set up Codex to review pull requests in this repo. Reviews are triggered when you
Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings. |
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 0ccf5c47a3
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
Reuse _assemble_simple_duckdb_cell for the TPC-H, TPC-DS, and TPC-H Skew builders so the duplicate-code delta gate stays flat. The TPC-H unseeded-defaults rationale and scale seam stay in place. Record the three thin per-benchmark delegations in the scoped duplicate ignore list.
Port the expression Q15 relative-tolerance max filter and both-family Q17 empty-set NULL preservation. The tpch staged gate now runs green (44/44 cells, 0 divergent, 0 vacuous) at SF=0.01 unseeded.
Take the staged-small builder fixes (seeded SQL windows with DF parse-back, active-table loading, TPC-H Skew scale seam) and keep the TPC-H scale seam. Render the TPC-DS reference in the DuckDB dialect so Q90's alias parses. Replace the work-item handoff comment with the durable blocking-subset rationale. Regenerate the oracle coverage map.
|
Maintainer review: addressed all five threads in 3200d44. Adopted the staged-small builder fixes (seeded SQL windows with DF parse-back, active-table loading, skew scale seam) and kept the TPC-H Q11 scale seam; TPC-DS reference now renders with dialect=duckdb (Q90 alias parses); the gate comment states the durable blocking-subset rationale (only Q38/Q88 pass today); coverage map regenerated (tpch/tpcds staged). Threads resolved. CI re-running; enqueuing after green. |
…y oracle Staging tpch/tpcds gates exposed a generator blind spot: provenance rendered only when cross-surface was primary, so tpch/tpcds rows showed no provenance despite registered gates. Read the axis live from gate metadata for every registered gate and pin the updated contract.
|
Status note (takeover sweep): the medium-test failure on this branch is a stale cross-surface applicability pin. This branch predates the staged tpch/tpcds gate registration landing via the current queue batch (2440 merged; 2427/2394/2441/2443 queued). After that batch drains, rebase this branch onto develop and run make cross-surface-applicability-report to refresh the artifact and pins, then re-push. No code fix needed. |
Develop already stages the TPC-H, TPC-DS, NYC Taxi, TSBS, and TPC-H Skew gates with their parameter and NULL fixes, so take its versions. The remaining change discloses surface provenance for every registered gate and regenerates the oracle coverage map.
Wire tpch and tpcds STAGED_GATES on the Q-prefix maps. tpch runs
unseeded-vs-unseeded at SF=0.01 (41/44 cells pass; fixed-stream needs
seeded DF overrides first). tpcds full-99 staged at SF=0.01 (~13s
wall); only Q38/Q88 pass, so the 10-15 blocking subset is deferred
to the promotion item. Module docstring states the staged status
and honest oracle limits. No GATES promotion, no CI wiring.