Quarantine POST_MultipleRequests_PooledStreamAndHeaders

[Vinoth2562000]

Quarantine POST_MultipleRequests_PooledStreamAndHeaders

Description

This unquarantines the test, which failed intermittently on CI due to a coarse 100ms poll for the "StreamPooled" log.

Root cause

The test polled for the "StreamPooled" log with a 100ms initial delay. On CI, the second request's AcceptAsync sometimes ran on a freshly allocated HttpRequestHeaders before the first stream was pushed to the pool — so Assert.Same failed (same value, different reference).

Fix

Replace the polling WaitForLogAsync with an event-driven wait on TestSink.MessageLogged:

• Subscribe before the first request → log cannot be missed
• TCS fires the instant the log is written (inside _poolLock, right after StreamPool.Push)
• try/finally ensures the handler is unsubscribed on test failure

Assert.Same is preserved — the test still detects cache misses (its original purpose).

Screenshot

Fixes #52573

[dotnet-policy-service]

Thanks for your PR, @Vinoth2562000. Someone from the team will get assigned to your PR shortly and we'll get it reviewed.

[Copilot]

Pull request overview

This PR unquarantines POST_MultipleRequests_PooledStreamAndHeaders by replacing a polling-based log wait with an event-driven wait tied to TestSink.MessageLogged, aiming to eliminate intermittent CI failures while preserving the Assert.Same cache-miss detection.

Changes:

• Removed the [QuarantinedTest] attribute for POST_MultipleRequests_PooledStreamAndHeaders.
• Replaced WaitForLogAsync polling for the QUIC "StreamPooled" log with an event-driven TaskCompletionSource that completes when the log is emitted.
• Added try/finally cleanup to unsubscribe the MessageLogged handler.