* feat(richtext-lexical): 'bottom' position value for plugins
* feat: TestRecorderFeature
* chore: restructuring to seed and clear db before each test
* chore: make sure all tests pass
* chore: make sure indexes are created in seed.ts - this fixes one erroring test
* chore: speed up test runs through db snapshots
* chore: support drizzle when resetting db
* chore: simplify seeding process, by moving boilerplate db reset / snapshot logic into a wrapper function
* chore: add new seeding process to admin test suite
* chore(deps): upgrade jest and playwright
* chore: make sure mongoose-specific tests are not skipped
* chore: fix point test, which was depending on another test (that's bad!)
* chore: fix incorrect import
* chore: remove unnecessary comments
* chore: clearly label lexicalE2E test file as todo
* chore: simplify seed logic
* chore: move versions test suite to new seed system
* test: saving multiple versions with unique fields
* fix: saving multiple versions with unique fields
Version schemas were getting produced with uniqueness constraints which
caused updates to likely fail because each successive version would
potentially reuse unchanged unique key-values (particularly ident keys
like slug) from previous versions.
This was due to `mongoose.buildSchema` not respecting buildSchemaOptions.disableUnique.
This regression was introduced with the fix at commit c175476e.
* test: eslint fix
Co-authored-by: Dan Ribbens <dan.ribbens@gmail.com>
