Co-authored-by: Cal Herries <39073188+calherries@users.noreply.github.com>

Previously, JS files isn't picked up by VS Code, so they'll have limited
intellisence. Also, Jest and Cypress has conflicting names e.g.
`describe` or `expect`. So, this commit attempt to resolve that as well.

* Ignore JS tests when doing CodeQL scans

* Do not run CodeQL scans unless needed

Adds --include-field-values flag to serdes v2 export command

* Use GitHub basic runners for `druid` and `sparksql` tests

* Reduce the `drivers` timeout to 45 mintues

* sort stacked tooltip rows

* update separator color

Adds error reporting and --abort-on-error flag to v2 import command

* Fix wrong aggregation field's `base_type`

It should be the same as the underlying field, not a hard-coded number
typ (`type/Integer` or `type/Float`)

* Add unit tests

* WIP action data type tests

* DRY up the actions tests

* lots of type tests

* make the tests work

* reorganize tests

* reduce some flakes

* updates after rebase

* Add /api/logs/query_execution endpoint for downloading logs.

Intended for admins to use only via api to manually save logs if they want. Paired with auto-deleting the
query_execution after N days (30 by default), this should help admins keep the query_execution table light.

* Add docstrings

* Add a test for the logs/query_execution endpoint

This test checks that the correct rows are returned, that is, those within the specified number of days.

* Tests should pass with h2 now.

* Move logs endpoint to Enterprise, allowed only with `:advanced-config`

* Forgot to remove api.logs ns require from OSS routes

* Wrap logs endpoint in proper premium feature

* Test will filter further to only look for THIS run's temp rows

* Alter query to give results for a month specified as YYYY-MM

This lets admins do paginated log downloads, which is good if the query_execution table is quite large (very likely,
at least initially).

* Add sanity check test. Address other review feedback

* Release script fix

* Add test 

* Metabase lib PoC RFC [ci skip]

* Cleanup [ci skip]

* `:lib/uuid` overhaul [ci skip]

* More tests [ci skip]

* Basic join support [ci skip]

* Simplify a few things, more tests, etc. [ci skip]

* Pipeline queries FTW [ci skip]

* Metabase® Lib™ v2

* Fix cljs test assertions

* No bees please

* Update docstring

* Mega simplification

* Fix Kondo warnings

* Address PR feedback

* Fix Malli util for Cljc

* Oops, `instrument!` should be public

* Cloverage do not instrument JS util namespace

* Fix Cloverage again

* Don't count MySQL or Postgres namespaces for Coverage

* Address PR feedback

* Move dispatch value keyword function to util

* Revert acccidental require

* Test fix

* Fix Cljs =?

* Fix totally unrelated broken links

* Build Script overhaul

* Remove stray deps.edn

* Include :ci profile for build and release tests

* Fix release script

* Update markdown dox

* Update another dox

* Tweakz

* Don't pin new version of data.xml to core project, just for build scripts

* Ignore unrecognized options

* Fix i18n/enumerate

* expose number of active users as a setting

The new approach uses a dev-mode-only CLJC reader conditional
`#?(:cljs-dev ...)` to make `metabase.util.devtools` empty
except in dev builds.

The previous approach for this relied on dead code elimination.
But DCE relies on tracing function calls and `:require`s from the
`:entries` list in shadow-cljs.edn. Top-level namespace code is
handled differently; in particular a namespace listed in `:entries`
like `metabase.util.devtools` will always have its top-level code
in the build.