Fix double-dilated timeout in ExpectMsgAsync<T>(Func<T, IActorRef, bool>, ...)

[Aaronontheweb]

Problem

TestKitBase_Expect.cs:207 resolves its timeout as:

timeout: RemainingOrDilated(RemainingOrDilated(timeout))

applying akka.test.timefactor twice. Every sibling overload applies it once.

Two consequences:

• With an explicit duration and timefactor = 3, a 5s wait becomes 45s (5 → 15 → 45).
• With null it is worse than a simple over-scale: the inner call resolves to RemainingOrDefault — the time actually remaining in the enclosing Within — and the outer call then dilates that remainder, yielding a bound larger than the budget it was derived from.

Found while tracking down MNTR flakes caused by the same class of mistake in test code (pre-dilating a value handed to a TestKit method that dilates internally).

Fix

Apply RemainingOrDilated once, matching every other overload.

Blast radius

This overload — ExpectMsgAsync<T>(Func<T, IActorRef, bool> isMessageAndSender, ...) — has no callers anywhere in this repository, so no existing test was relying on the inflated bound. It affects downstream users of Akka.TestKit only, for whom this restores the documented timefactor semantics.

Note the effective timeout gets shorter for anyone using this overload with a large timefactor, which is the correct behaviour but is a behavioural change worth calling out in review.

Verification

Akka.TestKit.Tests 320/321 (1 pre-existing skip) and Akka.TestKit.Xunit2.Tests 4/4 pass.

[Aaronontheweb]

LGTM