release-22.1: changefeedccl: Correctly handle dropped UDTs

[miretskiy]

Backport 1/1 commits from #107923.

/cc @cockroachdb/release

---

Schema feed is responsible for the determination
of correct schema descriptor corresponding to an
event occuring at certain time.

Due to the need to use low level API to retrieve
historical descriptor versions, it is possible that the schema feed would observe descriptors that were dropped/deleted/truncated.

A long outstanding bug in schema feed would
incorrectly check for low level value being 'nil' in order to determine if the descriptor was dropped.
This check is incorrect since the iterator may return a non-nil empty byte array reprsenting a tombstone.
A non-nil, but empty KV value would lead schema feed to incorrectly attempt to deserialize protocol message, resulting in a value type is not BYTES: UNKNOWN being returned.

This PR fixes the above issue, and adds a test
to verify correct behavior.

Fixes https://github.com/cockroachlabs/support/issues/2408

Release note (enterprise change): Fix a rare changefeed issue which is triggered when the parent database or types are dropped, and instead of exiting with a descriptive error message, the changefeed would observe an opaque error instead (value type is not BYTES: UNKNOWN)

Release justification: bug fix -- fix a long outstanding bug in changefeed.

[blathers-crl]

Thanks for opening a backport.

Please check the backport criteria before merging:

• Patches should only be created for serious issues or test-only changes.
• Patches should not break backwards-compatibility.
• Patches should change as little code as possible.
• Patches should not change on-disk formats or node communication protocols.
• Patches should not add new functionality.
• Patches must not add, edit, or otherwise modify cluster versions; or add version gates.

If some of the basic criteria cannot be satisfied, ensure that the exceptional criteria are satisfied within.

• There is a high priority need for the functionality that cannot wait until the next release and is difficult to address in another way.
• The new functionality is additive-only and only runs for clusters which have specifically “opted in” to it (e.g. by a cluster setting).
• New code is protected by a conditional check that is trivial to verify and ensures that it only runs for opt-in clusters.
• The PM and TL on the team that owns the changed code have signed off that the change obeys the above rules.

Add a brief release justification to the body of your PR to justify this backport.

Some other things to consider:

• What did we do to ensure that a user that doesn’t know & care about this backport, has no idea that it happened?
• Will this work in a cluster of mixed patch versions? Did we test that?
• If a user upgrades a patch version, uses this feature, and then downgrades, what happens?

[miretskiy]

Almost clean backport; had to fix up the test for 22.1 and add main_test.

[cockroach-teamcity]

This change is 

[fqazi]

Reviewed 4 of 4 files at r1, all commit messages.
Reviewable status: complete! 0 of 0 LGTMs obtained (waiting on @jayshrivastava)