* Add Exasol as a partner driver

* Add Exasol partner driver on docs

* Remove Exasol from Community drivers

* Add toast message after moving questions or models

* Add test

* Address feedback: extract component into its own file

* Remove undo action on question move because it doesn't work in all cases

* Fix failed test

nodata user can't turn a question into a model

* Remove commented code that could confuses more people

Pursuant to reopening of #22732.

Proximate cause is occasional Java bigints getting in where the types assume it'll only be Clojure bigints. On the current reproduction I have it fixes the bug but I am still determining ultimate cause and if there are other phenomena. There was also differential behavior of backport vs. on-master implementation which I still don't understand.

This one needs some sync to have gone through, or else it will prehistoric empetuses towards decelization

* fix missing enter full-screen button

* fix dashboard background trimmed in the dark mode

* update specs

Serialization of Databases, Tables, Fields

This brought a few core changes:
- Add `serdes-entity-id` to abstract the field used for the ID
- Pass the options to `extract-one` so it can eg. do encryption things.
- Handle dates in YAML storage and ingestion
- `:serdes/meta` now holds the entire hierarchy, not just the leaf model+ID pair.

There's an open problem here about the right way to handle secrets like
a database's password. Do we assume both sides have the same
`MB_ENCRYPTION_SECRET_KEY`? Provide a serdes-specific password the user
just made up, and every secret gets decrypted with the source key, encrypted with
the serdes key, stored, decrypted with the serdes key, and encrypted with
the destination key?

* Remove anchor time setting

* Remove anchor time setting

* Update setting

* Allow styling `SettingSelect` component

* Add `ModelCachingScheduleWidget`

* Remove `PersistedModelRefreshIntervalWidget`

* Pass `disabled` prop to expression input

* Install `cron-expression-validator`

* Wrap cron validator to support i18n

* Fix React hooks style

* Handle "custom" select value

* Clean cron input when selecting built-in schedule

* Connect custom cron input

* Update API endpoint to change schedule

* wip

* Handle the "year" CRON component

* Extract components, make explainer text live

* Add default custom schedule value

* Add helpful docs

* Don't allow configuring CRON seconds part

* Fix error message

* Tweak popover spacing

* Workaround validation lib issue

* Run prettier

* Replace interval-hours and anchor-time with cron

Removed the two settings and replaced with a single cron schedule setting

Renamed the /set-interval endpoint to /set-refresh-schedule

* Ignore "year" part in schedule endpoint

* Fix variable

* Use a temp scheduler and initialize the refresh job

Running into errors when updating the triggers for each database's
refresh job because the "job" itself didn't exist. Reminder of what's
going on here: There's a single refresh job. It has a trigger for each
database. So updating the trigger would fail since it doesn't exist
since there was no job to hold the triggers.

This error is quite clear in the tests run locally:

```
ERROR in metabase.api.persist-test/set-refresh-schedule-test (util.clj:421)
Uncaught exception, not in assertion.

        clojure.lang.ExceptionInfo: Error in with-temporary-setting-values: Couldn't store trigger 'DEFAULT.metabase.task.PersistenceRefresh.database.trigger.1' for 'DEFAULT.metabase.task.PersistenceRefresh.job' job:The job (DEFAULT.metabase.task.PersistenceRefresh.job) referenced by the trigger does not exist.
    location: metabase.public-settings/persisted-models-enabled
     setting: "persisted-models-enabled"
       value: true
```

But this logging is absent from the logging in Github annoyingly:

```
FAIL in metabase.api.persist-test/set-refresh-schedule-test (persist_test.clj:26)
Setting new cron schedule reschedules refresh tasks
Setting :persisted-models-enabled = true
expected: "0 0 0/12 * * ? *"
  actual: (nil)
```
Whic doesn't give us the error, just the test result.

* update to newer API `set-interval` -> `set-refresh-schedule`

```
;; old
{:hours 4}

;; new
{:cron "0 0 0/1 * * ? *"}
```

Co-authored-by: Case Nelson <case@metabase.com>
Co-authored-by: dan sutton <dan@dpsutton.com>

* Wait for the connected question to fully load

* Set default values for source dashboard filters

This step alone cuts the test execution time in half!

* Visit dashboards directly instead of hitting the `back` button

* Simplify function name

* Tidy up comments

* Wait for the target dashboard card to load

* Make setup function more explicit and more descriptive

* Make sure results loaded in a public source dashboard before testing click behavior

* update colors
* align border colors
* change border-dark hex value

Co-authored-by: Maz Ameli <maz@metabase.com>

* Assert upgrade URL

* Fix test id selector after refactoring

* Fix public sharing E2E tests

* Add test for updated public sharing copy

* Make upgrading URL link text consistent

* Update embedding copy

* Fix array key error when using jt

* Make E2E assertion more explicit

* Reference reproduced issue #23501

* Add repro for #20133

* Fix PK and FK indicators were gone

* Fix `SortIcon` opacity

* Fix spacing between column name and sort icon

When `(premium-features/is-hosted?)` is true (IE, in the cloud) we
display some information about whitelisted ip addresses on the database
admin screen.

> If your database is behind a firewall, you may need to allow connections from our Metabase Cloud
> IP addresses: 18.207.81.126 - 3.211.20.157 - 50.17.234.169

A previous change interpreted this text as markdown so now we can
include the link here.

One problem is that this markdown is a translated string. So we're gonna
need this markdown duplicated into all of our other languages. And those
will point to english documentation. Not sure if there's a great
solution there.

For testing:

If you want to see this in the app, edit public settings cloud-gateway-ips:

```clojure
(defsetting cloud-gateway-ips
  (deferred-tru "Metabase Cloud gateway IP addresses, to configure connections to DBs behind firewalls")
  :visibility :public
  :type       :json
  :setter     :none
  :getter     (fn []
                (when #_(premium-features/is-hosted?) true
                  (fetch-cloud-gateway-ips-fn))))
```

And then go to admin > databases and edit a db or start to create a new one

* LDAP now also properly allows `nil` in names.

There were still places where "Unknown" was substituted in to name keys, which we don't need to do anymore as 'nil' is
valid for first/last names of users.

* Remove unused namespaces/refers

* Adjust LDAP so that even nil name values update the Metabase user

Also changed a test in both the OSS and EE LDAP tests to make sure this is indeed true that 'nil' values are correctly
set in the app-db when LDAP does not send `givenName` and/or `sn`.

Write `storage.yaml` and `ingest.yaml` to serialize all the way to YAML files and back.

Lots of generative testing to check it's isomorphic.

* Refactor admin part of question moderation E2E test

* Remove redundant test

* Do not use hard coded user names

* add plain date type for date picker
* better field handling