aminghadersohi commented on PR #43943: URL: https://github.com/apache/superset/pull/43943#issuecomment-5591302722
Addressed all five items in [review 5140601116](https://github.com/apache/superset/pull/43943#pullrequestreview-5140601116) in `307f84d5bc234f62acdfcc02b049f34bd6891097`. 1. **Raw Gauge data:** removed renderability normalization from both get-data query paths. Saved, saved-with-form-data-key, and unsaved FastMCP requests retain mixed/all-non-finite groups in JSON/CSV/XLSX. Regressions assert returned rows, source `rowcount` distinct from `row_count`, and ordinary completeness semantics. Unsaved exports honor their format; the XLSX fallback also preserves non-finite groups. Preview/compile-only filtering and all-invalid errors remain covered by real generate/update compilation tests. This supersedes the previous comment's get-data normalization disposition. 2. **Temporal sentinels:** normalize empty/whitespace and `No filter` before comparing both operands. Schema and public FastMCP cases pass; missing/non-string comparators and genuinely conflicting restrictions still reject. 3. **Default-CI facet regression:** non-skipped tests assert x/y origins equal child-unit centers. Deliberately restoring outer width/height signals fails that test without `NODE_PATH`. Actual Vega scenegraph checks also pass. 4. **All-zero bounds:** pin ASCII `Range: 0 to 1` and Vega metadata min/max to 0/1; explicit degenerate bounds still reject. 5. **Needle length:** derive 68.5px from the 55–82px band midpoint; rendered endpoints must reach within that band. No new items were rejected. Earlier [22-suggestion dispositions](https://github.com/apache/superset/pull/43943#issuecomment-5574824746) and [human-review follow-ups](https://github.com/apache/superset/pull/43943#issuecomment-5575896351) remain applicable except the explicitly superseded raw-data behavior and needle length. All four earlier inline threads remain resolved; this review has no new inline thread to resolve. Validation: **1,810 affected Python tests**, **21 frontend Gauge tests**, **12 rendered specs**, full branch-applicable/staged pre-commit (including mypy), compileall, and optimized-mode renderability checks passed. Recorded fail-before cases cover raw-data/export and sentinel/needle fixes. **All 13 required checks passed on this exact head**, including [unit tests](https://github.com/apache/superset/actions/runs/34267917703/job/102221920856), [PostgreSQL integration](https://github.com/apache/superset/actions/runs/34267917682/job/102209195432), [Playwright](https://github.com/apache/superset/actions/runs/34267917673/job/102208404642), [pre-commit](https://github.com/apache/superset/actions/runs/34267917689/job/102201873695). The unit job initially exceeded its existing 30-minute limit after all 14,776 main-suite tests passed; one rerun completed successfully, including the remaining coverage batches. No tests, coverage thresholds, or timeouts were changed. The prior non-required dependency-check failure occurred while setup-uv fetched Astral's version manifest, before dependency checks. One failed-job rerun succeeded; the new head's [check-python-deps](https://github.com/apache/superset/actions/runs/34267917686/job/102201873517) also passed. No code failure was masked. PR testing/coordination sections are updated. No additional review is requested. -- 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] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
