Games screen test can read a row before its stats land #266

Closed
opened 2026-10-28 00:50:57 -05:00 by ethan-hann · 0 comments
Owner

GamesScreenViewModelTests.The_rows_show_as_soon_as_the_profile_list_arrives_and_fill_in_when_the_stats_land failed once in Linux CI on PR 265 (head 55fa9dcf, 2026-10-28 05:23 UTC). The same commit passed the Windows gate, the Linux VM gate and the Windows CI job, so this is a race in the test or the rig, not a product change.

Failure: after release.SetResult() and UntilAsync(() => !rig.Screen.IsLoading), the version column read ["1.1.0.4", "-"]: the second row had not filled in yet when IsLoading turned false. Assert.Equal() Failure: Collections differ at index 1, Expected "1.1.0.4", Actual "-".

The test dates from c037f829 (2026-10-21); the 0.9.21 batch changed the rigs' event delivery and dispatcher (D70), which likely widened the window. Find whether IsLoading can turn false before every row's stats are applied (a product ordering bug the player could see as a row stuck on a dash) or whether the test should wait for the rows themselves, fix that, and look for the same wait-on-a-flag-then-read-rows pattern in the other Games screen tests (tests-meaningful-and-fast.md).

This failure held back the 0.9.21 release: PR 265 (version bump + an earlier race fix in FakeEventDeliveryTests) is open with Linux CI red.

GamesScreenViewModelTests.The_rows_show_as_soon_as_the_profile_list_arrives_and_fill_in_when_the_stats_land failed once in Linux CI on PR 265 (head 55fa9dcf, 2026-10-28 05:23 UTC). The same commit passed the Windows gate, the Linux VM gate and the Windows CI job, so this is a race in the test or the rig, not a product change. Failure: after release.SetResult() and UntilAsync(() => !rig.Screen.IsLoading), the version column read ["1.1.0.4", "-"]: the second row had not filled in yet when IsLoading turned false. Assert.Equal() Failure: Collections differ at index 1, Expected "1.1.0.4", Actual "-". The test dates from c037f829 (2026-10-21); the 0.9.21 batch changed the rigs' event delivery and dispatcher (D70), which likely widened the window. Find whether IsLoading can turn false before every row's stats are applied (a product ordering bug the player could see as a row stuck on a dash) or whether the test should wait for the rows themselves, fix that, and look for the same wait-on-a-flag-then-read-rows pattern in the other Games screen tests (tests-meaningful-and-fast.md). This failure held back the 0.9.21 release: PR 265 (version bump + an earlier race fix in FakeEventDeliveryTests) is open with Linux CI red.
ethan-hann added this to the Code Cleanup milestone 2026-10-28 00:50:57 -05:00
Sign in to join this conversation.
No milestone
No assignees
1 participant
Notifications
Due date
The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set

Reference
ethan-hann/SavePoint#266
No description provided.