* lint: add stylelint support for emotion components

* drop comment

* fix pattern order

* Use one version of typescript

* use :focus not :outline

* Use the logical CSS property "inset-inline-start:" (not "inline-start:" which doesn't exist) as replacement for the physical css property "left:"

* Use mb-color-focus for focus color

* Drop dead code

---------

Co-authored-by: Raphael Krut-Landau <raphael.kl@gmail.com>

* fix: limit xrays on agg models to source fields

Fixes #43793

1. `table-like?` was checking aggregations but not breakouts. This meant
   that legacy metric definitions would try to apply to the second stage
   which is generally illegal and would cause things like implicit joins
   to come in where they make no sense.
2. Trying to bin a `count` metric-like model would give a bin span of
   0.0 which would blow up those questions.
3. Trying to self reference breakout fields, in the precense of a
   question's  aggregation meant that most fields would not be exposed to
   the next stage and should be ignored. (similar to 1.)

* Fix tests

* Show alert sending time in alert form

* F

* Add csv-stringify package

* Split values at comma

WIP: Parse values using json

* Add label to ParameterValue and add ParameterValue to value config

* Use custom label in FieldValuesWidget

* Allow passing displayValue from outside of RemappedValue

* Allow passing displayValue in ParameterFieldWidgetValue

* Get custom value label in FormatterParameterValue

* Allow rendering custom displayValue in FieldValuesWidget

* Add custom ParameterValue types

* Fix ParameterValuePicker type

* Fix jest import for csv-stringify

* Add missing RTKQ exports for paramters

* Use direct implementation of useGetParameterValues

* Fix ValueSourceModal tests

* Add tests for value parsing and stringifying

* Use custom label in list too

* Render the custom label in a parameter value

* Setup parameter mocks in embed test

* Add info about splitting values with comma

* Add model hint when user is using labels

* Add test for cusom labels on dashboards

* Handle label-value pairs in test

* Test custom labels in parameter search input

* Use the label to parse the value for text inputs

* Fix comma's to commas

Co-authored-by: Kamil Mielnik <kamil@kamilmielnik.com>

* Use strong instead of styling link

* Use getInput instead of ignoring eslint rule

* Use arrow function

* Add newlines between tests

* Use uppercase for test value

* Name value

* Use getLabel and getValue helpers to avoid indexing confusion

* Pull out confusing condition

* Move out getLabel helper

* Be more explicit about non-existent config values

Co-authored-by: Uladzimir Havenchyk <125459446+uladzimirdev@users.noreply.github.com>

* Move setupParameterValuesEndpoints to setup

* Add loading state

* Do not use alert to implement info box

* Fix formatting

* Remove undefined from getValue signature

* Use getLearnUrl

* Remove dependency on column

* Make custom field labels work in the absence of fields

* Support custom values for non-field parameters

* Add test for native question parameters

* Add newline in test

* Add test case for when show-metabase-links is false

* Just fall back to rendering value when fetching label, or in case it failed

* Add test for loading state of FormattedParameterValue

* Do not fetch custom label for FormattedParameterValue

* Add table with ip addresses

* Add test for IPAddress filter with custom labels

* Add test for custom labels on type/Quantity fields

* Remove getInput helper in test

* Remove unused hasCustomLabel value

* Split up into common and premium tests

* Make fieldValues a map for performance

---------

Co-authored-by: Kamil Mielnik <kamil@kamilmielnik.com>
Co-authored-by: Uladzimir Havenchyk <125459446+uladzimirdev@users.noreply.github.com>

* fix: param-fields when native model is aggregated

Fixes #42829

The param-field lookup required result_metadata to match
filterable-columns in order to find parameter targets. This adds a
fallback where if the param field is missing from result_metadata,
then we look at filterable-columns.

* Unskip test

Per discussion in Slack, I made a mistake with the migration IDs as I wanted to backport to 50 but used 51 IDs.
This PR uses 50 IDs and places the migrations in the correct place.

Not sure if it's mergable in this state because Stats has run with the PR containing the incorrect migration IDs.

Refine the reproduction for " Entity picker surfaces recent questions that are not valid joins" (#44974) (#45091)

* Refine the reproduction for 44794

* Move skip to top-level again

* Remove superflouous click

* Verify wether or not the recents tab exists and is open

* Make the test description more descriptive

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

* Move question details to a variable

---------

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

This works on master, but this reproduction will prevent #38265 from
reoccurring.

E2E repro for "Editing Column description of Model is confusing if it contains aggregations" (#45113)

* Fix checks and tests

* Fix checks and tests

Reproduce #27643: Field filter linked to Boolean column causes an error when used in Embedding (#45038)

* Add a repro for #27643

* Clarify the race condition behavior in the comment

* User Parameter Values Are Unique Per Dashboard

Fixes: #43001

Might be related: #44858

User Parameter Values previously stored just the `parameter-id` and `user-id` as data to look up the last used
parameter value. This isn't sufficient as the parameter id is not guaranteed to be unique, in particular this is true
when a dashboard is duplicated: the parameters are copied to the new dashboard without changing the parameter id at
all.

This means that we need to also store the dashboard-id in the user_parameter_value table and use that to update/get
the last used value.

The migration removes existing entries to the user_parameter_value table because I want a non-nullable constraint on
the dashboard_id column, and existing entries will not have a value. The only way to backfill those values would be to
look through every dashboard containing parameters, and for every parameter check for a matching ID. Even if you can
do this, there's no way to disambiguate if the same paramter exists on 2 dashboards anyway, so one of them will be
wrong. I think it's not worth the trouble considering that removing the entries in this table doesn't break anything;
it just becomes a mild inconvenience that a user has to select a filter value again (since the dashboard will use the
default value).

* alter test to check for uniqueness across dashboard

This test makes 2 dashboards with parameters of the same ID and asserts that changing the value on dashboard-a does
not change the value on dashboard-b.

* adjust migration to pass linter rules

* remove the extra rollback on migration

* Update src/metabase/models/user_parameter_value.clj

Co-authored-by: bryan <bryan.maass@gmail.com>

* Update src/metabase/models/user_parameter_value.clj

Co-authored-by: bryan <bryan.maass@gmail.com>

* adjust parameters in test so we don't get logged warnings

* put update!/insert! in a transaction to avoid any race conditions

---------

Co-authored-by: bryan <bryan.maass@gmail.com>

* Support compiling SQL with inlined parameters directly

* Test fix

* Make `format-honeysql` a multimethod

* Add test for #32543

* Test also fixes #44915

* Add explicit test for #44915

* Better test for #44915

* More test fixes 

* Fix Kondo error

* Optimize: compute chart title href only on mouseenter and focus

* Improve comment

* Improve comment

* Format code

* Update existing tests

* Extract helpers

* Update unit tests

* Make getHref argument-less

* include dark background to PNG exports when night theme is enabled

* fix css

* override border for png downloads

`Lib.diagnoseExpression` is stricter about this than the QP or FE
expression parser, making it reject expressions that tried to compare
datetimes to strings like `2024-07-02 14:34:17Z`.

Those used to work in 49.x, but a new `Lib.diagnoseExpression` call
while editing expressions caused it to regress.

Fixes #44916.

**MetadataProvider cache**

The reusable MetadataProviders is a generally useful thing - within some
dynamic scope like an API request, a `clojure.core.cache` instance can
be provided as `lib.metadata.jvm/*metadata-provider-cache*`, and that
will be used to cache the provider for each `database-id`.

Reusing those providers can significantly reduce appdb traffic,
re-fetching the same metadata in different `MetadataProvider`
instances.

**Dashboard loads**

The FE can supply `?dashboard_load_id=uuid_string` on requests for a
dashboard load:

- `/api/dashboard/:id`
- `/api/dashboard/:id/query_metadata`
- `/api/dashboard/:id/dashcard/:dashcard_id/card/:card_id/query`

That provides a good cache key across that burst of N+2 API requests
that are part of a single load. Currently it is used for two things:

- Caching `get-dashboard` (used by the first two requests)
- Selecting the `clojure.core.cache` for `*metadata-provider-cache*`

Both use a TTL of 10 seconds, which is generally enough for the first
screenful of a dashboard load, if not the whole thing. (Browsers will
only issue a handful of requests at once, so a few long-running queries
might delay later parts of the dashboard from loading until after the
TTL expires.)

* Fix #39771

* Add repro for #39771

* Add explanatory comment

* Fix z-index

* Use Mantine Tooltip in Ellipsified to fix z-index issues

* Revert "Fix z-index"

This reverts commit 951e3dca86be73ba2123930fbfaaae45c5fa1f0e.

* Revert "Fix #39771"

This reverts commit 4165da5022d566253e875938a84bd1f137c92357.

* Update repro to new tooltip

* Update assertions