Update `->Timestamp` to include `report-timezone`

It's possible that database connections checked back into the pool
could still have session timezones associated with them. This can
cause the database timezone tests to failure. This will reset the
connection pool before the test runs to ensure that doesn't happen.

The goal is to make it easier to write tests with different data and
report timezone scenarios easier. The tests already have found issues
with our current code and their handling of these timezone
conversions. These tests exposed a bug in MySQL date support which is
also fixed in this commit.

Fixes #6554, fixes #6994, fixes #7393

This commit adds a new `*report-timezone*` dynamic var that should get
bound before dates get parsed (i.e. in a query pipeline
middleware). This allows an easier API `(->Timestamp my-date-string)`
while still ensuring the dates are interpretted correctly.

This is a step toward centralization of our timezone and date
conversion logic. As of this commit there are still separate/similar
logic for handling reporting timezone etc in Pulses, drivers etc.

Fixes #7584, fixes #6402

Fix Mongo CI test failure

Enable Eastwood unused namespaces linter

Update driver deps [ci drivers]

Drop support for Java 7 

Sync metrics prototype

This test predates the move of min/max values of a field to
fingerprints. Rmoving this test as it's covered over there and doesn't
provide a lot of value in the fields sync code.

This commit adds a new "sync summary" message to sync/analyze/field
values scanning. It include information about how much data was
updated, begin/end timestamps around each phase etc.

Merge release 0.29.x

Update suggestions for adding external deps w/ Java 9

Do not use absolute URLs in illustrations

Don't overwrite PKs and FKs when infering special_type

Type inference: add partial matches for fields named *_country and *_state

Fix rule loading

Remove unnecessary use of flex-full on new query options

Separate SparkSQL dependencies [ci drivers]