* initial migrations

* autofill permission_id when adding new sandbox

* use new table name in backend code

* add test for migration

* fix migration lint errors

* WIP fix tests

* small cleanup

* use CONCAT to fix migration on mysql

* break out fk constraint into a separate migration

* address feedback

* fix GET api/session endpoint for settings managers

* fix test

update specs and ensure users need to wait until stats are loaded

`constantly` evaluates its argument at compile time so when we do
`(constantly (deferred-tru))` the user's locale is undefined.

Change to `(fn [] (deferred-tru))` so the error message can be returned
in user's locale.

Parameter with source is a card:
- deduplicates values
- remove null/empty values
- properly search number

* Switch to the new Hawk test runner

* Fix some stuff

* Fix everything

* Commentz

* Bump init timeout?

* Fix Kondo config

* Remove unused

* Bump test runner version to get `:exclude-tags` support

Removed pulses from search API and corresponding tests. There _may_ be cruft tied to pulse search elsewhere because we can't blindly remove pulses yet (need to provide legacy support at the model level), but I think I got it all.

Fixing check such that a create-table check failure shows correctly instead of showing as a can-read failure.

* Remove unused code

* Add `<Tab.Link />` component

* Allow to customize `TabList's` styled components

* Add `tab` arg to model detail URL helper

* Add `:tab` routes for model detail page

* Use link tabs in model detail page

* Open actions tab directly in e2e tests

* Add safety checks for tabs

* Don't nest `TabLink` under `Tab`

* Use explicit tab routes

* Use `IndexRedirect`

* Don't use `Array.prototype.at`

Fails with Node 14.x :(

* remove database only after additional confirmations

* specs

* review

* fix typos

* Require rendering routes manually

* Remove `Link` mocks

* Return `history` from `renderWithProviders`

* Fix `react-router-redux` doesn't work in Jest

* Fix `ViewHeader` tests

* Don't pass not used prop

* Avoid `@ts-ignore`

* Add migration to remove defaults

* Add rollback

* Fix migration

* Fix migration

* Fix migrations

* whitespace

* Add test defaults for DashboardCard

* Remove defaults from create-dashboard-card!

* Fix test

* Fix tests

* Tighten API endpoints

* Fix endpoints

* Formatting

* Fix tests

* Fix tests

* Fix tests

* Fix tests

* Fix test

* Fix e2e tests

* Update e2e tests

* Fix e2e tests

* Bump athena driver to 2.0.35

* Include aws core for auth

* convert SelectList to typescript

* add SelectList unit tests

* add SelectList storybook file

* update tests

* add PivotTable unit tests

* obey the linter

* measure leftHeader cell content

* test cell data width detection

* convert PivotTable to typescript

* remove unused hasCustomColors

* User Resizable Pivot Table Columns (#27692)

* move column widths to state

* add drag resizing

* only get initial header widths on row change

* better column change handling

* return if there are no valid rowIndexes

* address review comments

* add drag resizing

* address review comments

* persist column width settings

* update unit tests

* fix rebase

* e2e test column resize and persistence

* keep column width state locally unless changed by the user

this handles a couple situations:
 (1) where an existing pivot table is opened and is magically "changed" because we auto-measured header widths and loaded that data into visualization_settings, and
 (2) where, like pinned cards in dashboards, there is no updatable state for visualization settings, we need to be able to do measurements and update it within the component and can't rely on mutable state of a parent component

* subscription permissions refactor (https://github.com/metabase/metabase/security/advisories/GHSA-ch8f-hhq9-7gv5)

* move recipient filtering to API-level (https://github.com/metabase/metabase/security/advisories/GHSA-492f-qxr3-9rrv)

* Script to copy athena jdbc jar to our maven repo

Since amazon is still not putting their athena jdbc on to maven, the
easiest thing for us to do is setup a maven-y s3 repo and copy their jar
into it.

This script allows you to run:
`./bin/athena.bb <version> [--deploy]`

For example:
`./bin/athena.bb 2.0.35 --deploy`

Will download the 2.0.35 jar from amazon's release bucket, create and
update the necessary maven repo files, then upload all those files to
our maven repo.

It uses the jar build WITHOUT the AWS SDK.

You must have `bb` (babashka), `aws` (awscli), and `shasum` commands available.
You must have aws credentials available, e.g. through `aws configure`

* Add comment explaining script to script

* Use iterate instead of loop for find-pred

Co-authored-by: metamben <103100869+metamben@users.noreply.github.com>

* Update bin/athena.bb

Co-authored-by: metamben <103100869+metamben@users.noreply.github.com>

* Update bin/athena.bb

Co-authored-by: metamben <103100869+metamben@users.noreply.github.com>

---------

Co-authored-by: metamben <103100869+metamben@users.noreply.github.com>

We made a mistake on the Clojure CLI version

Replaced, as needed, instances of str/lower|upper-case with the -en version. Also updated rules in our kondo config.

* Add `UserInfo` type

* Use `UserInfo` for existing types

* Add creator fields to action types

* Show action creator in action list

* Add actions.creator_id

* Add creator_id and hydrated creator to actions

* Punt on locking down time types

* Workaround for DB migration issue

* Move hydration down the pipe

* HydratedUser -> DefaultUser