ref https://linear.app/tryghost/issue/ONC-323
When the store gets into a bad state for new posts that causes saves to fail we can detect that by looking at the `model.isNew` property. Currently our best approach to fix this state is to restart the app.
- added a `didTransition()` hook to our `lexical-edit.new` route
- detects the bad state, logs the error, and triggers a browser refresh
- logs with a `recreatedPostIsGood` property that will let us know if we could instead just try recreating the post and avoiding a full refresh (so far we have no reproduction case so we need to learn what we can)
- added `sinon-chai` dependency for better assertions on spies/stubs
- added `sentry-testkit` dependency so we can test our Sentry integration calls
- we can't use sinon for these calls because of the way Sentry's es6 imports work
- extracted our full Sentry config object generation to a util function so it can be re-used in unit tests
- updated our integrations list to disable the default `dedupe` integration because it can cause very unexpected/difficult to debug test failures when you're asserting using `sentry-testkit`
closes https://linear.app/tryghost/issue/PLG-176
The editor files were previously stubbed for testing because we didn't have a way to load the externally-hosted files. This made testing slow and difficult because the only way to test the Admin integration was via Ghost's e2e browser tests.
- unstubbed the editor globals so `fetchKoenigLexical()` actually tries to import the external assets
- updated `ember-cli-build` to copy the Koenig UMD file over to the assets directory in development/test builds
- updated `environment.js` to set the required filename for the default asset import to successfully hit the test environment hosted files
- updated lexical editor acceptance tests to demonstrate the editor loads successfully for new and existing posts
Promoted our beta editor to the default editor. Keep an eye on (or subscribe to) https://ghost.org/changelog/ for release announcements with full details.
- moved the beta editor (Lexical-based editor) to the default editor; all pages and posts will now use it
- all mobiledoc (previous editor) posts will remain mobiledoc until opened in the editor at which point will be converted to Lexical on the fly and open in the new editor
no issue
- improves test run times by splitting the tests into multiple groups and running each in their own browser instance
- settled on 2 browser instances as that seems to best fit within Travis' memory and CPU constraints
- updated ember-cli-code-coverage config to work with parallel builds
no issue
- upgrade to latest `ember-source` and related dependencies including `ember-cli`
- upgrade to latest `ember-mocha` and modern ember testing setup
- https://github.com/emberjs/rfcs/blob/master/text/0268-acceptance-testing-refactor.md
- switch from using global acceptance test helpers and `native-dom-helpers` to using the new `ember-test-helpers` methods
- use [`chai-dom`](https://github.com/nathanboktae/chai-dom) assertions where in some places (still a lot of places in the tests that could use these)
- pin `ember-in-viewport` to 3.0.x to work around incompatibilities between different versions used in `ember-light-table`, `ember-infinity`, and `ember-sticky-element`
- incompatibilities manifested as "Invalid value used as weak map key" errors thrown when using `ember-light-table` (subscribers screen)
- pin `ember-power-datepicker` to unreleased version that contains a move from global acceptance test helpers to modern test helpers
no issue
- switch `jscs` and `jshint` inline config to `eslint` config
- fix eslint errors, predominantly in tests where the config now the main app config more closely
no issue
- increases individual test timeout from 5sec to 15sec
It seems Travis is occasionally _very_ slow and will timeout on heavier acceptance tests resulting in random failures. Hopefully this will reduce the number of random test failures we see.
no issue
- increases the timeout because acceptance tests sometimes hit the limit which then caused a knock-on effect through other tests
- fix settings/navigation acceptance tests where it picked up previous test failure error messages
- fix user model unit tests so that `expect` is always called _after_ the runloop has finished handling property updates
