* Add dashboard parameter mapping for date range

* Fix unit tests

* Add e2e tests

* Update logic to use date range only for specific units

* Add Mantine `Popover`

* Fix spacing

* Deprecate previous popovers

* Update isNavigationAllowed logic to prevent new model creation

* Add simple model creation test

* Remove identifier

* Update comment

* Add a test case for saving a new model

* Avoid using fetchMock directly in QueryBuilder unit tests

* Remove setupCollectionsEndpointsByIds in favor of setupCollectionByIdEndpoint

* Add native question test base

* Update isNavigationAllowed tests

* Make isNavigationAllowed tests pass

* Update shouldShowUnsavedChangesWarningForSqlQuery logic to account for new questions

* Add a test case to run new native question

* Update mock to include individual collection GET endpoint

* Add DataSourceSelectors to MockNativeQueryEditor

* Update test beforeunload event to reflect new behavior

* Use setupCollectionByIdEndpoint instead of modifying setupCollectionsEndpoints

* Update failing test

* Use serializeCardForUrl

* Add a test for native query editing

* Do not trigger warning when leaving empty question

* Add beforeunload tests for creating new questions

* Remove redundant div

* Make assertions consistent

* Use single resetAllMocks()

* Fix post-rebase conflict

* Simplify getShouldShowUnsavedChangesWarning's logic

* Add isModifiedFromNotebook to store

* Show unsaved changes warning when structured question has been modified via notebook editor

* Use explicit return type, as it's not obvious that this function needs to be a Promise

* Remove unused selector dependency

* Update mocks with new attribute

* Add a test for editing notebook questions

* Extract waitForSaveNewQuestionToBeEnabled out of triggerNotebookQueryChange

* Introduce waitForSaveModelToBeEnabled & waitForSaveModelToBeDisabled

* Add a test case

* Remove redundant navigation

* Update isNavigationAllowed logic and tests

* Refactor tests and add more test cases

* Add more test cases

* Shorten code

* Add a test for saving the question

* Refactor isNavigationAllowed

* Add a test for not showing modal

* Add a test for non-notebook changes

* Organize code

* Use existing setUIControls

* Add extra assertions for post-save navigation to check if question is properly mark as non-dirty

* Add extra assertions for post-save navigation to check if question is properly mark as non-dirty

* Add a post-save nagivation test

* Add logic to reset UI controls isModifiedFromNotebook state after saving the question

* Rename to allowedPathnames

* Add explanatory comment

* Update failing e2e test

* Fix #32963 e2e repro failing

* Allow `/model/${slug}` to make models-query-editor e2e test pass

* Update isNavigationAllowed logic to make e2e tests pass

* Update isNavigationAllowed tests

* Add 1 more model location

* Revert identifier name change

* Use optional chaining operator for brevity

* Fix typo

* Rename notebook to structured

* Rename getQuestionLocations to getNativeQuestionLocations and don't use getNativeQuestionLocations in getStructuredQuestionLocations

Moves `migrate-click-through` and `migrate-remove-admin-from-group-mapping-if-needed` to custom migrations, so that all migrations are handled by liquibase.

Resolves #32611, #32612

Co-authored-by: John Swanson <john.d.swanson@gmail.com>

* remove redundant cy.wait

* try again to fix test failure in CI that doesn't repro locally

We want to add some tests to verify that this (slightly hidden) feature continues working.

Fixes: https://github.com/metabase/metabase/issues/21163

* Add `:added` metadata to driver multimethods

Commit which introduced defmulti was determined
with `git log -L<line>,+1:<file>. Then first version it appeared in
was extracted from `git tag --contains <sha>`.

* Add driver methods changelog test

* Add `:changelog-test/ignore` to various defmultis

Multiple defmultis added after 42 incl. are missing entries from
changelog. Marking them this way makes it possible to add those
entries retrospectively without having failing tests, one by one.

* no splitting when there is only 1 metric

* checking for unique cards, not series

* beefy e2e test

* Conditionally start Docker containers for E2E tests

* Check out before initializing Docker containers

* Use official action to log in to DockerHub

* Evaluate conditional expressions

* Try shell conditionals

* Fix errors due to unset vars

* Run either Mongo or Pg+MySQL

* Remove `branch` from version.properties file

* Remove branch from logs

* Remove `branch` from `application-name`

The `application-name` string is needed for Google drivers.
See: https://github.com/metabase/metabase/issues/2627

https://www.notion.so/metabase/Filter-clause-text-descriptions-556639c83bd24e3f9a306b68d36e8a1f



Co-authored-by: metamben <103100869+metamben@users.noreply.github.com>
Co-authored-by: Anton Kulyk <kuliks.anton@gmail.com>

* Update migration and add test

Sets semantic_type to null for metabase_field if value was type/Number,
to avoid log warnings.

* Remove test for trivial migration as suggested

As migration tests are costly and change is trvial, it is better
to test small migration as one in question manually. Hence removing
the test.

Enable column-filter, disable distribution and summarize-column drills for aggregated columns (#34681)

* Update isNavigationAllowed logic to prevent new model creation

* Add simple model creation test

* Remove identifier

* Update comment

* Add a test case for saving a new model

* Avoid using fetchMock directly in QueryBuilder unit tests

* Remove setupCollectionsEndpointsByIds in favor of setupCollectionByIdEndpoint

* Add native question test base

* Update isNavigationAllowed tests

* Make isNavigationAllowed tests pass

* Update shouldShowUnsavedChangesWarningForSqlQuery logic to account for new questions

* Add a test case to run new native question

* Update mock to include individual collection GET endpoint

* Add DataSourceSelectors to MockNativeQueryEditor

* Update test beforeunload event to reflect new behavior

* Use setupCollectionByIdEndpoint instead of modifying setupCollectionsEndpoints

* Update failing test

* Use serializeCardForUrl

* Add a test for native query editing

* Do not trigger warning when leaving empty question

* Add beforeunload tests for creating new questions

* Remove redundant div

* Make assertions consistent

* Use single resetAllMocks()

* Fix post-rebase conflict

* Work around Cypress not handling beforeunload-related prompts

* Our app registers beforeunload event listener e.g. when editing a native SQL question.
* Cypress does not automatically close the browser prompt and does not allow manually
* interacting with it (unlike with window.confirm). The test will hang forever with
* the prompt displayed and will eventually time out. We need to work around this by
* monkey-patching window.addEventListener to ignore beforeunload event handlers.
*
* @see https://github.com/cypress-io/cypress/issues/2118

* Make `Chromatic` workflow 3rd-party friendly

* Remove unused output

* Unify the use of `GITHUB_TOKEN`

* Update label check

* Fix wrong token reference

[ci skip]

This was causing scheduled runs to fail due to a
missing closing quotes in a string.

[ci skip]