[ci skip]

@jaril informed me that Replay.io is missing metadata
from Metabase E2E tests. Apparently trying to access
system environment variable from Cypress support file
didn't work.

The solution for this is to prefix the environment variable
with `CYPRESS_` and then to read it using `Cypress.env()`

This is what Cypress documentation suggests:
https://docs.cypress.io/guides/guides/environment-variables#Option-3-CYPRESS_

[ci skip]

* Make custom `button` command chainable

* Chain `.button` directly

No need to use `.within()` anymore.
We can chain the button command directly because it now receives
the previously yielded subject.

* Fix collections bulk actions test

* remove /reference pages + reducers

* Remove redundant metadata selectors

* Remove redundant metadata actions

* Remove `QueryButton` component

* Remove `List` component

* Remove dead link

* Move `NoDatabasesEmptyState` to `NewModelOptions`

* remove obvious /reference specific e2e tests and ask a bunch of questions [ci skip]

* Remove data reference section from e2e test suite [ci skip]

* Remove segments data reference e2e tests [ci skip]

* Remove reference section from the browse page test [ci skip]

* Remove the last few mentions of the reference page in e2e tests [ci skip]

* Fix sso button lint error

---------

Co-authored-by: Anton Kulyk <kuliks.anton@gmail.com>
Co-authored-by: Nemanja Glumac <31325167+nemanjaglumac@users.noreply.github.com>
Co-authored-by: Alexander Polyankin <alexander.polyankin@metabase.com>

[ci skip]

* Make E2E custom icon command chainable

* Refactor E2E helpers

We can now chain the `.icon()` command on the previous subject.

* Chain `.icon()` on prev subject in all tests

* Fix moderation-question test

* Add `dimensionOptionsForValidation`

* Add `getMLv1CompatibleDimension` dimension method

* Use `getMLv1CompatibleDimension`

* Fix `getMLv1CompatibleDimension` usage

* Add tests for `getMLv1CompatibleDimension`

The underlying issue was fixed, but we are not sure exactly when.
This reproduction should prevent future regressions.

* Apply inclusion and exclusion patterns to `driver/syncable-schemas` for sql-jdbc drivers

* Fix kondo

* Fix postgres test

* Remove unnecessary test DB names

* Update 15946 e2e test
- modal() expects ModalContent to be present in DOM, but NativeQuerymodal uses ModalBody directly
- add extra check for run button to be enabled

* Ensure collection is present in dataset query during query conversion

* Unskip e2e test for #15946

* Handle string tableId

Fixes #29935.

Part of the issue here was incorrect hand-rolling of the second
argument to `:case`, the default value. In the failing test case given
in #29935, it was written as `{:default 0}`. But that's accidentally
transferring for legacy MBQL works, where the default is an option and
the options come last rather than first in a clause.

This also fixes `lib.common/->op-arg` to make sure it recurses
properly. It wasn't recursing into lists or MBQL clauses, which left
unresolved `(fn [query stage] ...)` functions in eg. `:case`
alternatives, since those are nested inside lists.

* add migration for typed collections

* add hydration of collections to card+dashboard api

- adds it for:
  - GET /api/question/:id
  - GET /api/dashboard/:id

* fix test that checks for collection shape

* get tests working with hydrated collection

* typo fixes

* fix more tests

- PUT requests don't hydrate :collection

* update tests to fix param_values {} vs nil inconsistencies

Part of #31157.

* fix dashboard save failure after deleting tabs

* fix undo regression and remove unused code in delete reducer case

https://github.blog/changelog/2022-10-11-github-actions-deprecating-save-state-and-set-output-commands/

* Use faster inserts in MySQL

* Add local_infile check

* PR feedback

* Fix boolean handling for MySQL

* Fix tests; fix MySQL fallback if local_infile is turned off

* Turn local_infile on for MySQL tests

* Another way of doing MySQL local_infile :/

* sanitize-file-name was a misstep. Not needed for security and could cause problems

* Mistaken docstring

This was causing ambiguous references, since `order-by` clauses
reference columns from previous stages by name.

Fixes #30401.

* Make stage index required in MLv2 methods

* Add wrappers for `appendStage` and `dropStage`

* Fix notebook stage index management

* Fix `QuestionRowCount`

* Fix `StructuredQuery` limit method

* Clean redundant stages in notebook

This is because `source-field` was missing from a field dimension.

* Add initial filterable columns implemenation

* Implement filterable-columns

Change filters to return internal clauses rather than external operators

* Add comments