KAFKA-20497: Add TimeOrderedCachingWindowStore#readOnly

[nicktelford]

The cache holds uncommitted writes that must not be visible under
READ_COMMITTED, so that isolation level bypasses the cache entirely and
delegates straight to the inner store's readOnly view. READ_UNCOMMITTED
requires a merged view through a ReadOnlyView. The ReadOnlyView
delegates to the existing fetchInternal/fetchKeyRange/fetchAllInternal
helpers and converts Instant arguments to epoch-milliseconds before
calling them; the cache key schema operates in longs throughout, so
converting early avoids repeated Instant-to-long conversions in iterator
hot paths.

KAFKA-20497

Reviewers: Bill Bejeck bbejeck@apache.org

[nicktelford]

@bbejeck

[github-actions]

A label of 'needs-attention' was automatically added to this PR in order to raise the
attention of the committers. Once this issue has been triaged, the triage label
should be removed to prevent this automation from happening again.

[bbejeck]

Thanks @nicktelford - overall looks good but there are a couple of failing tests

1. org.apache.kafka.streams.state.internals.TimeOrderedCachingPersistentWindowStoreTest#shouldThrowOnNullInstantInViewFetch
2. org.apache.kafka.streams.state.internals.TimeOrderedCachingPersistentWindowStoreTest#shouldNotThrowInvalidBackwardRangeExceptionWithNegativeFromKey

[nicktelford]

Thanks @nicktelford - overall looks good but there are a couple of failing tests

1. org.apache.kafka.streams.state.internals.TimeOrderedCachingPersistentWindowStoreTest#shouldThrowOnNullInstantInViewFetch
2. org.apache.kafka.streams.state.internals.TimeOrderedCachingPersistentWindowStoreTest#shouldNotThrowInvalidBackwardRangeExceptionWithNegativeFromKey

@bbejeck Rebased on trunk; these test failures should now be resolved.

[bbejeck]

@nicktelford - needs ./gradlew spotlessApply

[bbejeck]

Thanks @nicktelford - looks good overall but there are some test failures which I think can be resolved by updating the log statement on line 526 is in TimeOrderedCachingWindowStore

[nicktelford]

@bbejeck I think I fixed this the other day but forgot to push the new commit 🤦‍♂️
Also rebased on trunk.

[bbejeck]

Thanks @nicktelford - still getting the same test failures.