SelfAwareJoinDecider: remove infinite timeout on http lookup

[pjfanning]

Motivation

SelfAwareJoinDecider.selfContactPoint awaited ClusterBootstrap(system).selfContactPoint with Duration.Inf. The comment justified that with "the future has a timeout", which is only true once ClusterBootstrap.start() has run: ensureSelfContactPoint() schedules the timer that fails the promise, and it is called from start() alone. Bind the bootstrap routes manually and the promise is never completed by anything, so the await never returns.

Two further problems came out of review:

• The wait and the timer that makes it safe are coupled — the wait is only correct while it outlasts the timer — but they were separate literals in separate files, with nothing expressing the relationship. Raising the timer would silently make the decider time out first, replacing the deliberate "'Bootstrap.selfContactPoint' was NOT set, but is required..." error with a bare TimeoutException from somewhere else.
• Caching it as a lazy val does not fix the repeated-blocking case, because a lazy val re-runs its initialiser after a throw. In exactly the scenario this PR exists for, every canJoinSelf call would block a dispatcher thread for the full timeout again — Duration.Inf parked one thread once, this parks one per probe.

Modification

• Move the timer's duration to ClusterBootstrap.SelfContactPointTimeout and derive the decider's wait from it, behind a protected def selfContactPointTimeout that a test can override. The effective value is unchanged.
• Cache the resolved contact point in an AtomicReference. A failure is deliberately not cached, so a contact point set later is still picked up, but the blocking wait is paid at most once: reaching the timeout means nothing will complete the promise, so subsequent callers check it without blocking and fail fast until it does complete.
• Resolve against the promise directly instead of mapping it first, so the already-completed case needs no dispatcher hop and Future#value is meaningful.
• canJoinSelf reuses the self it already resolved instead of asking again.

Result

The two timeouts cannot drift apart. An unset contact point costs one blocking wait rather than one per probe, and is still picked up if it arrives late.

Tests

• sbt "management-cluster-bootstrap/test" — 54 succeeded, 0 failed (48 before this PR).
• New SelfContactPointResolutionSpec covers resolution, caching, the timeout, the block-once behaviour, setting a contact point after a timeout, and the ordering between the two timeouts.
• Directional: with the block-once guard removed so every call waits, "block for the timeout only once, then fail fast" fails with 505796658 nanoseconds was not less than 250 milliseconds.
• sbt "management-cluster-bootstrap/mimaReportBinaryIssues" — success.
• sbt "management-cluster-bootstrap/scalafmtCheck" "management-cluster-bootstrap/Test/scalafmtCheck" and +headerCheckAll — clean.

References

None - found by review of SelfAwareJoinDecider