[GH] (pekko/fix-metrics-resizer-flaky-java25): Workflow run "Pull Requests" is working again!

[GitBox <gi...@apache.org>]

The GitHub Actions job "Pull Requests" on pekko.git/fix-metrics-resizer-flaky-java25 has succeeded.
Run started by GitHub user He-Pin (triggered by He-Pin).

Head commit for run:
409fde6c0bb5355a26c90c3b8455afd9b20598f7 / He-Pin <he...@gmail.com>
test: fix flaky 'stop an underutilizationStreak when fully utilized' on JDK 25

Motivation:
The test failed intermittently on JDK 25 CI:
  Some(UnderUtilizationStreak(..., 1)) was not empty (line 155)

Same race condition previously addressed in #2769 for the sibling test:
`sendToAll(await = true)` only guarantees each routee has *started*
processing (the first latch counted down), not that the resizer's later
inspection of `cell.currentMessage` will observe a non-null reference.
Under JDK 25's scheduling there is a brief window where currentMessage
can appear null while the actor transitions, causing `utilized` to be
computed as 1 instead of 2 and the underutilization streak to persist.

Modification:
After `sendToAll(await = true)`, enqueue one extra message per routee
with `await = false` so the mailbox is non-empty even if currentMessage
is momentarily indeterminate. Mirrors the fix already applied to
"update the underutilizationStreak highestUtilization if current
utilization is higher".

Result:
The test deterministically observes utilized == 2 on JDK 25 and the
underutilization streak is cleared as expected.

Report URL: https://github.com/apache/pekko/actions/runs/24881572793

With regards,
GitHub Actions via GitBox


---------------------------------------------------------------------
To unsubscribe, e-mail: notifications-unsubscribe@pekko.apache.org
For additional commands, e-mail: notifications-help@pekko.apache.org