Destroy open connections when resetting app DB; add `ReentrantReadWriteLock` around app DB `getConnection` (#21664)

* Destroy open connections when resetting app DB

* Use no-op dummy app DB

* Ok, don't shut down the app DB because it breaks in-memory app DBs

* Stop/start task scheduler

* Test fix ?

* Another test fix 

* Simplified impl

* Revert task changes

* Fix connection-pool destroying code

* Use ReentrantReadWriteLock for simpler implementation

* More doc

* indentation only

* Fix creation check for mysql

* Test that all persisting dbs can persist

* Erran tap> (whoops)

* `POST /api.ee/serialization/serialize/data-model` endpoint

* Sort namespaces

* Don't use `tru` for endpoint params validation

* Fix/refactor QB dirty results selector

* Don't clobber query results & last run card when canceling model changes

* Always rerun updated questions if they have a dirty query

* Add action that only reruns dirty question queries

* Stop rerunning model queries in a few scenarios where they don't need to run

* Fix e2e tests

* Add e2e tests to ensure reruns when necessary

* add instructions to convert key to der

* omit paragraph re: sslmode

* Implement advanced date/time/zone manipulation, part 1

Incorporate new functions into MBQL and add tests:
 - get-year
 - get-quarter
 - get-month
 - get-day
 - get-day-of-week
 - get-hour
 - get-minute
 - get-second

* Fix BigQuery implementations to call extract

Mark as not supported in legacy driver

* Add date extraction fns for Postgres

* Disable in MongoDB (for now at least)

Disable in BigQuery (legacy driver)

Add implementations for presto-jdbc

* Misc cleanup from Jeff's PR

* Update Jeff's implementation of bigquery-cloud-sqk

* Reorganized tests

* Mongo

* Oracle

* Sqlserver

* Sqlite

* Add casting supports for presto

* Remove Jeff's implementation of presto-jdbc because  its parent is
sql-jdbc

* Update presto-jdbc tests to use the same catalog for all datasets

* Add date extraction functions to the expression editor (#25382)

* make sure the semantic type of aggregated columns are integer

* no recursive call in annotate for date-extract func

* get-unit -> temporal-extract(column, unit)

* desguar nested datetime extraction too

* Changing graph viz settings toggles to inline, updating radio buttons to
segmentedControl inputs

* PR Feedback

* Adding missing placeholders to various viz settings

* Removing translation from $

* PR Feedback

* Remove Stacking option when only 1 stackable series is present

* Updating isValid property for stackable.stack_type

* Fixing Tests

* Upgrade Cypress to `v10.9.0`

* Migrate Cypress plugins to the new config

* Replace deprecated `ignoreTestFiles` option

* Migrate `Cypress.json` to the new config

* Extend the config

* Upgrade `cypress-grep` library

* Tag tests that rely on Webmail SMTP server

* Tag tests that rely on external QA databases

* Tag tests that rely on docker LDAP server

* Update the dev documentation

* Fix typo

* stop infinite error reload

* add inline forms for actions

* better style names

* simplify action click drill

* better validator type

* Refactor static waterfall chart to use a single entrypoint

* Consolidate static watetrfall chart

* Update waterfall chart text size to match combo-chart

* Move `<Values />` from XYChart/ to be under components/

Since it'll be used by waterfall chart

* Fix `<Values />` key error (forgot to add key attribute)

* Show data point values on waterfall chart

* Fix failed BE tests

* Fix test -> wrong `waterfall-type` used. And test both types

The waterfall component was changed to take a type argument, instead of having 2 functions for nearly identical
output. The failing test just passed the wrong string to the key argument.

* Change waterfall keys from strings to Clojure keys

* Remove redundant test. Categorical also handled in `waterfall-test`

* Fix clj-kondo lint error

* Cleanup conditional statements

* Miscellaneous code cleanup

* Update waterfall test to render data point values

* Fix waterfall chart data point values not center in the bar

* Make some API for static waterfall chart with total saner

Co-authored-by: Adam James <adam.vermeer2@gmail.com>

* Switch to humane-are

* Fix some busted `are`s

* Remove unused namespaces

* Implement DB specific execution permissions

Separate execution permissions from data permissions as a new dimension
like application features. Create a new endpoint for getting and setting
these permissions.

* Substitute persisted queries in parameter card references

```sql
select o.total, o.quantity, model_p.category, model_p.title
from orders o
left join {{#14-pg-products}} model_p -- reference to products model
on o.product_id = model_p.id
limit 4
```

Testing:

I'm unhappy with how verbose the test is. At some point we're going to
need a better, standardized way to test persistence. But the test makes
a persisted table of the test-data.categories
table (test/metabase/test/data/dataset_definitions/test-data.edn). It
runs some sql to update the values in the table to turn `"Winery"` ->
`"Winery from cached table"`, joins the two tables together to have both
at once.

There's a lot of moving pieces and the setup is a bit verbose. There's
also no obvious place where these types of tests should go. We'll need
to consolidate them in the future (and extend them).

* persisted macro

* docstring for `persisted-info-native-query`

* alignments and small nits

* Reuse persistence macro

* clj-kondo cleanups

* remove clj-kondo change before rebase