* Druid Docker image [WIP]

* More Druid Docker stuff [WIP]

* Test fixes 

* Remove maxRowsQueuedForOrdering from query, don't think we need it

* Force test Druid host -> localhost for now

* Update sync tests for updated test data

* Revert unused changes

* Allow multiple `--spec` arguments in Cypress `run` command (#15917)

* Split e2e CI groups (#15924)

* Add option to choose between source-folder and test-files in CCI

* Rearrange Cypress files

* Split e2e workflows

* Split native queries into a separate CI group

* Improve Cypress runner command line config (#15854)

* Rename Cypress CI groups to `e2e-tests` (#15877)

* Remove the `node` parameter from Cypress CI job matrix (#15878)

* Merge EE Cypress tests with the OSS ones (#15887)

* Reorganize test folder: Move e2e support files into a separate folder (#15899)

* Refine CI checksum calculation - extract e2e files (#15900)

* Run tests on Java 8 again (fix #15707); run tests against Java 16

* Switch from Nashorn to GraalVM JS engine

* Remove unused var

* Add SMTP server for testing

* Try running maildev in the default executor for Cypress

* Add first email-based Cypress test

* Try different ports

* Stop using ENVs for maildev

* Use camelCase for `const`

Co-authored-by: Luis Paolini <paoliniluis@gmail.com>

* Bumped Lein 2.8 with 2.9.5 in CircleCI image
Simplify executors (all use the unified image now)

* Pinning down to specific container

* [ci nocache]

* Fix SAML redirect to /collection/root/

* Test util HTTP client should automatically URL-encode query parameters.

* Simplify HTTP client

* Minor tweaks

* Remove *url-encode-query-parameters?*

* experimental CI fix: cache build script deps

* Fix typo

* Fix Circle config again

* Fetch BE deps if deps.edn files change

* Use CI executor for be-deps, so we have the Clojure CLI

Fix Oracle SSL tests

Define new test for Oracle SSL connectivity, in oracle_test.clj, similar to how things work in mysql_test.clj

Add new test util macro, with-env-keys-renamed-by, to support running tests with environ keys temporarily renamed

Using new test macro from both MySQL and Oracle SSL connectivity tests

Removing now unneeded be-tests-oracle-ssl-ee CircleCI job

Removing now unneeded test-selector parameter for test-driver orb in CircleCI config.yml

Updating JVM_OPTS to use a trust store that starts with cacerts and adds the RDS root CA, rather than one only
containing the RDS root CA

Adding ssl-cert config field to MySQL DB details map, to hold the server cert chain in PEM format (similar to what is done in MongoDB driver)

Updating MySQL driver init to map :ssl-cert into :serverSslCert for the JDBC url, when ssl is in use and cert is provided (the MariaDB driver we are using accepts PEM format certificates inline directly for the param value, so no need to shepherd into a temp file)

Adding new test to mysql_test.clj to run a single test while connecting via SSL with PEM cert

Update CircleCI config:
 - use extra-env to set all the MySQL SSL instance DB related vars (for an RDS instance, currently)
 - adding the rds-combined-ca-bundle.pem certificate to resources/certificates
 - loading that cert bundle from resources directory via env var

Adding to/fixing assertion in connection-spec-test for :ssl

Port the metabase.mbql utility namespaces to ./shared/ so they can be used on both the frontend and backend.

* Port code for building i18n resources to Clojure build scripts; remove dep on gettext

* Use Metabase classloader

* Test fixes 



* Bump adoptopenjdk version

* Delete gettext dependency from Dockerfile

* Update developers-guide.md

* Address PR feedback

* Fix frontend singular msgstr format

Co-authored-by: Luis Paolini <paoliniluis@gmail.com>

It's possible for the scheduler to get in a weird state if the sync
fails while it executes.

This change makes it *only* recreate a job/task if the schedule has
changed or if it is missing. Previously, clearing the state at every
start had bad effects if the JVM had terminated during the sync.

Adds a vector of Exception classes that signal a "fatal" exception
during sync for a specific database. If these exceptions occur, the sync
for that database stops and will pick up next time.

This will have to be expanded per driver, but I don't see a way around
that, as each driver will have its own, unique way of failing.

metabase/metabase#14817

Update Oracle driver to have the `ssl` (boolean) option, with the expectation that any keystore/truststore params have been set as JVM parameters

Updating CircleCI config to allow for the following parameters to the `test-driver` orb:
- `extra-env`: a string that will be prepended to the `lein test` invocation, in order to set environment variables specifically for that test run.  Expected to look like `VAR_NAME=a VAR_NAME_2=b`. Defaults to blank (i.e. nothing).
- `test-selector`: a string that represents the Leiningen test [selectors](https://cljdoc.org/d/leiningen/leiningen/2.9.5/api/leiningen.test#test). Defaults to `:all` (i.e., run all tests).

Adding RDS root CA truststore file to our resources so that CircleCI (and anyone else, really) can refer to it

Defining new `be-tests-oracle-ssl-ee test-driver` task, and is identical to it except for the env vars required to connect via SSL instead (by specifying the RDS root CA trust store).  It only runs a few Oracle driver specific tests, not the full suite, since we are really only concerned with connectivity here.

Updating Oracle driver docs to discuss relevant JVM params and configuration

* Shared CLJ/CLJS lib (PoC/WIP)

* PoC 2.0

* Fixes 

* More test fixes 

* Bump shadow-cljs version

* Fix more stuff

* Need to ^:export the exports

* CI fixes 

* Add eslintignore

* Ignore cljs files for FE code coverage

* Try prefixing CLJS -> JS import with goog:

* Revert indentation change

* No goog:

* Add .prettierignore

* Use advanced build for now for JS tests unit we can figure out how to make it work

* Change string-operations-from-subquery to use test-data instead of sample-data, because the latter currently is incompatible with Redshift (see #14784)

* Add MSSQL_MEMORY_LIMIT_MB: 1024 to CircleCI config (cherry pick of https://github.com/metabase/metabase/commit/4a714f3e21fb06fdb73b11fa7812f417b38a069e)

1. Rename optimize-datetime-filters middleware -> optimize-temporal-filters (it's more accurate, because this also
optimizes date or time filter clauses)

2. optimize-temporal-filters middleware now optimizes relative-datetime clauses (which represent a moment in time
relative to when the query is ran, e.g. "last month") in addition to absolute-datetime clauses (which represent an
absolute moment in time, e.g. 2021-02-15T14:40:00-08:00) . This middleware rewrites queries so we filter against
specific temporal ranges without casting the column itself, meaning we can leverage indexes on that column. See #11837
for more details

3. Added new validate-temporal-bucketing middleware that throws an Exception if you try to do something that makes no
sense, e.g. bucket a DATE field by :time or a TIME field by :month. This is a better situation then running the query
and waiting for the DB to complain. (In practice, I don't think the FE client would let you generate a query like this
in the first place)

4. Fix random test failures for MySQL in task-history-cleanup-test

* Use dashboard count in search scoring

[Fixes #14945]

* Fix sorting priority bug

(thanks, Dan!)

* Show (hidden) scores in search results for debugging purposes

* Bump memory limit for SQL Server on CI

c.f. https://metaboat.slack.com/archives/C5XHN8GLW/p1614272426000400

Ports changes introduced in #14942 over to master branch
This is a cherry-pick of 70ad97a4.

* Enable running a single spec in Cypress

* Rename flag name to run Cypress files from a specific folder

* Quick sync

* Try reducing -Xmx for :ci profile a bit to prevent random test failures

* Update quick sync parameter name

* Update docstring for dataset

* Fix sync-database! schema

* Fix schema again

* Test fix 

* Rework tests in api.gtap-test

* Return actual error message when trying to save an invalid sandbox query

* Minor refactor

* GTAP modal should display errors

* Fix flow failure

* Test fix

* Another flow fix 



* Longer timeout for the test that keeps failing

* Return error message as :message

* New cache key for drivers/uberjar so Oracle/Vertica errors stop popping up

* Revert cache key rotation

* #14612 Repro: Sandboxing limitations should have meaningful UI error message/feedback (#14690)

* Fix random Cypress test failures when fingerprinting fails?

* Bump no-output-timeouts to 15m because CircleCI is extra slow lately

Co-authored-by: Nemanja Glumac <31325167+nemanjaglumac@users.noreply.github.com>

* remove projectId

* turn off --record

* remove parallel/group

* Separate Oracle/Vertica plugin manifests for EE driver versions

* Bump test HTTP client timeout on CircleCI to fix random failures

* include edition in cypress group

Co-authored-by: Cam Saul <github@camsaul.com>

* Use GitHub Actions to run front-end linters

* Add docs/, needed for lint-docs-links

* No need to honor ci skip since the execution is path-conditional already

* Remove front-end linters from Circle CI

* package.lock, in case any npm/run scripts changed

Co-authored-by: Cam Saul <1455846+camsaul@users.noreply.github.com>

* Split metabase.db into connection, connection-pool-setup, env, setup, and util

* Remove old logging string

* Test fixes 

* Test fix 

* Deprecate parse-connection-string

* Remove application-db-mock-id

* Full support for MB_DB_CONNECTION_URI for application DB without parsing

* Remove *allow-potentailly-unsafe-connections*

* Test fixes 

* Test fix 

* Test fixes 

* Refactor load/dump commands

* New metabase.cmd.copy namespace

* More tweaks etc.

* 

*  2000

* Test fix 

* 

* Make linter happy

* Test fix

* Fix typo

* Use correct skip key for backend tests

This completes one step towards removing Enzyme.
The complete list of all tests is at #13657.

* Run each test in isolation

* Convert `EditUserModal.integ.spec.js` to Cypress

* Convert `PeopleListingApp.integ.spec.js` to Cypress

* Convert `GroupDetailApp.integ.spec.js` to Cypress

* Delete `GroupsListingApp.integ.spec.js`

It was already covered in Cypress.

* Delete `NewUserModal.integ.spec.js`

It was already covered in Cypress.

* Convert `UserActivationModal.integ.spec.js`

* Convert `UserPasswordResetModal.integ.spec.js`

* Convert `UserSettings.integ.spec.js`

* Remove integration tests

* Remove integration tests from CI

* Fix linter errors

* Delete `AuditTable.integ.spec.jsx`

* Clean up `people.cy.spec.js`

* Update assertion that makes sure XHR didn't happen

* Remove redundant command from PR template

Note: `yarn-lint` already runs `yarn flow`.
https://github.com/metabase/metabase/pull/14174#discussion_r549790198

* Remove explicit mention of timezones in `developers-guide.md`

* add step to validate Liquibase migrations to CI

* addColumn should only allow one column.

* Only allow one column added per addColumn change

* Add strict-mode for change sets

* Check that migrations are in order/IDs are unique

* Fix CI

* Tests for the linter!!

* Require remarks when adding new Tables

* Fix CI