github-actions[bot] opened a new pull request, #75593: URL: https://github.com/apache/airflow/pull/75593
* UI: Fix visual and console-error bugs in Dag list and details These are pre-existing bugs, all found while testing an unrelated change and all in the same corner of the UI, so they travel together rather than as a handful of one-line pull requests. Several survived review because nothing asserted on them: the Gantt axis only collided its React keys on runs of roughly twenty seconds or less, the warnings marker only stretched on a Dag that actually had a warning, and the HITL tab lookup only logged its 404 on a task that was not human-in-the-loop. Each now has a test that fails without its fix. The Dag status markers are consolidated because #74516 moved every other Dag status onto the shared StatusChip and left this one behind. That left two markers that looked identical, sat in different containers, and meant different things, with no single place to see everything wrong with a Dag. Co-Authored-By: Claude <[email protected]> * UI: Address review on the Dag status chip and warning translations The merged status chip was introducing the same class of console error the rest of the change removes. Rich detail was rendered inside a paragraph, which cannot hold the blocks the import-error view brings, so the browser closed the paragraph early and the file row inherited an error colour and type scale meant only for the traceback. Tests now fail on React's markup warnings instead of letting them scroll past in CI output, which is what let this through: the test that opens that very modal was passing while logging two nesting errors. Three assertions could not fail as written. One searched for a nested button only inside the inner button, so restoring the old wrapper would have gone unnoticed; one asked for a chip named after a deactivation that the import-error branch can never produce; one opened the modal for a failed request without checking the failure reached it. Warning types reached the heading as the backend's own lowercase enum values, which would have shown untranslated English in every locale, so they are translated now rather than passed through. The chip and its modal each walked the severity order themselves and would have disagreed the moment one changed. Co-Authored-By: Claude <[email protected]> --------- (cherry picked from commit da4fafc5ca22b25b670cf8d16daeac3ad4575c87) Co-authored-by: Brent Bovenzi <[email protected]> Co-authored-by: Claude <[email protected]> -- 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]
