fitzee commented on PR #44114: URL: https://github.com/apache/superset/pull/44114#issuecomment-5617319440
Addressed the latest review in `1d5e341711`: - Fixed the near-deadline regression: if the available timeout is no longer sufficient to complete the 500 ms dwell, capture proceeds using the already-satisfied readiness gate instead of starting a deterministically impossible wait. - Centralized the standard and tiled stable-readiness wait in one helper. - Added executable Node coverage for initial state, the 500 ms boundary, ready → loading → ready reset, and token reset. - Added chart-container predicate-selection coverage. - Renamed the wiring test so it no longer claims to execute the dwell. - Replaced brittle call-count/slicing assertions with predicate-based filtering. - Replaced hardcoded `500` assertions with `REPORT_CAPTURE_READINESS_STABILITY_MS`. - Added the `UPDATING.md` note. - Removed the unrelated setup-uv pin because it still performs remote manifest resolution on cache misses. I retained the per-tile/per-attempt dwell intentionally: scrolling changes which virtualized holders are active, and a blank retry follows a compositor repaint, so readiness from a previous tile or attempt is not transferable. The new insufficient-budget path prevents that safety check from consuming a capture that could still complete. Validation: repository-pinned Ruff formatting and lint pass. The JavaScript state-machine test executes directly when Node is available. The full local pytest invocation was blocked by the available shell image carrying an incompatible dependency set for this OSS head (`flask-caching` constructor mismatch), so CI remains the authoritative full-suite run. -- 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]
