refs https://github.com/TryGhost/Toolbox/issues/202
- during DB init, we have to create all the tables
- right now we loop over all tables and call the `createTable` command
- this command checks if the table exists and if not, creates the table
- this works fine but it means we query the database for every table
- in MySQL, we query the information_schema table, which we've seen
issues with before because it doesn't have indexes
- the smarter thing to do here is to get all the tables that already exist,
remove them from the list, and just straight up create them without
further checks
- this entire thing should be protected by the migration lock so we
shouldn't encounter issues from multiple processes initializing the DB
and tables existing after the initial check
- this commit also removes the check from `createTable` because this isn't
really needed. We should be using the migration utils, which do
check for existing tables. I've added a note to the function and
audited anywhere we still call the function
- this commit removes (- 49 tables + 1 initial check) 48 queries from
the initial DB init
refs https://github.com/TryGhost/Toolbox/issues/214
- Disconnected Stripe state is now a special case after the defaul test configuration was changed to have fake Stripe keys included mimicing configured Stripe
- The disconnected state is now achieved through an API call instead of mocking internal services from the test level. This seems like a cleaner approach, although still a bit cumbersome
refs https://github.com/TryGhost/Toolbox/issues/214
- These variables need to be present in the configuration during the Ghost's boot time initialization, which caused a need to remember to mock settings cache before agent initialization.
- By moving the values into default settings config it removes the need to do any work during test environment setup. Yey!
- We should put default values in to test-specific settings-defaults.json In similar situations. Specifically, when we find a need to mock settings cache to be able to start Ghost instance in a certain state
refs https://github.com/TryGhost/Toolbox/issues/214
- Having this config for test environment allows to pre-populate default settings values in the settings table
- Right now the default-settings.json is an exact copy of the original "/data/schema/default-settings/default-settings.json". Having a starter file as an exact copy, allows to track the differences between environments as they are introduced easier
refs https://github.com/TryGhost/Toolbox/issues/214
- The values configuration for the settings table need to become configurable to be able to run our test environment with a pre-defined set of configurations (e.g Stripe-related values).
- This change makes it possible to define the default settings file location (currently a JSON)
- A new key is now exposed through the "paths.defaultSettings" key in settings, which can be overloaded for the needs of the environment
refs https://github.com/TryGhost/Toolbox/issues/214
- The `defaultSettings` path name in the config (one pointing to routes yaml file) creates confusion with the `defaultSettings` which populate defaults for in the database settings table.
- Furthermore, the name collision creates a problem when trying to make database default settings dynamic - being able to load them from configurable file path.
- Rename makes "routing" explicit to avoid ambiguity and free up the name for the database defaults
- The value seems to be safe to be renamed as all keys used in `overrides.json` are taking priority - the name "defaultRouteSettings" hasn't surfaced at any point in the git history
refs https://github.com/TryGhost/Toolbox/issues/214
- TestAgent was used to initialize both Admin & Member API agents, which is somewhat confusing because Member API does not have the same "loginAs" functionality like Admin API does
- Having distinct agents for each API makes the class API cleaner with possibility to extract common functionality even further
- Fixed test fixtures so that members with subscriptions also have products/tiers
- Fixed test fixtures so that default&free tiers can be updated for tests
- Added tests for the signin functionality and welcome page redirects
- Extended `setupStripe` to setup other Members settings - this needs some more
thought around how we proceed
refs https://github.com/TryGhost/Team/issues/1355
- Uses the models to query the database for testing the event presence.
- Prototyped an util to help with event assertion and correct error messages and lower lines of codes.
- Given there are side-effects between tests (adding or removing members), the event count depends on the previous tests, this isn't optimal.
- the previous logic only allowed one flag to be mocked at a time because it kept recalling sinon.stub
- now it's possible to mock multiple flags with different settings as we always just add to the same stub
refs https://github.com/TryGhost/Toolbox/issues/163
- as of Node 15, unhandled rejections will exit the process so if
Ghost is running on Node 15+ and encounters one, it will kill Ghost
- if Sentry is enabled, it will add a handler for the event that will
send it to Sentry but the logging is sent to stdout/stderr, which means
we lose it in Ghost logs
- this commit adds a process handler for the `unhandledRejection` event
which will log the reason to Ghost logs and prevent Ghost from
exiting
refs https://github.com/TryGhost/Toolbox/issues/214
- After calling `DELETE /settings/stripe/connect/` Admin API endpoint, stripe did not fully disconnect causing other Member API endpoints to behave as if Stripe was still configured
no issue
- The test was failing because the match for a the data has changed, which was returned in a short format YYYY-MM-DD.
- Added a new matcher for short date format -anyShortDate. Can be used for match dynamic dates in short format
- Ghost themes are allowed to define the API version they want to use, but this concept has never really worked correctly
- We have a theory that it doesn't ever do quite what people want, and so always loading latest would not really break anything - this commit tests that theory
- We're pinning to canary, as we're aiming to get rid of the concept of versions altogether
- I could have done return config.get('api:versions:default');, but this actually returns v4 (although the same as canary, conceptually different)
and it also seemed like an unnecessary level of indirection. This change should be easy to understand and to revert if we are wrong
refs https://github.com/TryGhost/Toolbox/issues/214
- Having to fill out `process.env` variables in tests is a frustrating developer experience. Test environment should be configured with smart defaults, so the developer writing test cases would modify variables like this only is special test cases.
https://github.com/TryGhost/Team/issues/1374
When we receive a web hook to update payment details for a customer
which is not associated with a Member, instead of proceeding and erroring
we now return early and respond to Stripe with a 200. A test has been
added to ensure this
refs https://github.com/TryGhost/Toolbox/issues/207
- we want to start writing some E2E tests that involve automated
pointing and clicking around the frontend of Ghost to test that
members of Ghost sites can still do what we expect
- we've decided to look in to Playwright for this - it looks __really__
nice
- this is a VERY basic first test - it'll check for a 200 on the
homepage of whatever we provide as the TEST_URL env variable, or
default to a (manually-run) Ghost instance on port 2368
- also adds a `yarn test:browser` command to run the tests using the
Playwright CLI, and a sample GitHub Action workflow which we can
manually run with a site URL
- there's a lot more to add here in terms of test framework but this
gets us started
refs https://github.com/TryGhost/Toolbox/issues/214
- Calling `getMembersAPIAgent` and `getAdminAPIAgent` separately was booting Ghost twice, which caused a significant performance degradation.
- Additionally, having two calls was slightly ugly and having once utility function that delivers multiple agents at once feels like more readable syntax
refs https://github.com/TryGhost/Toolbox/issues/210
- After updating to the newest `@tryghost/express-test`, it's dependency `@tryghost/jest-snapshot` was missmatched with the one used in Ghost itself. This caused multiple instances of SnapshotManager class in the system, which is meant to be a singleton!
- The issue was causing following error to appear: "IncorrectUsageError: Unable to run snapshot tests, current test was not configured"
refs https://github.com/TryGhost/Toolbox/issues/210
- Having assertions executed in the declaration order using superagent has lead to many frustrating developer experiences. For example when status code assertion fails there was usually no additional context to investigate the failure further. In this situation having a "response body" assertion executed usually would give many more clues on the actual problem at hand.
- The order of the assertions has been changed from the order in which they were declared to the following one based on assertion type:
1. `exect`
2. `expectHeader`
3. `expectStatus`
- For more on this topic check the documentation: cfd6e88fe2/packages/express-test (assertion-execution-order)
- for some reason, this test seems to be failing now we've pulled it out
of the general CI
- it makes sense when the repo is clean, because the html files don't
exist, but I don't understand how they were working before... 🤔
- anyway, we should be overriding the path to the test fixtures admin
view files here
- this fixes the unit tests
refs https://github.com/TryGhost/Toolbox/issues/208
- we currently run all test commands separately in CI and this can make
it harder to replicate any issues we see in CI because we have to
manually go and copy the order of the tests
- it's also nicer if there's just one command that runs so we can make
changes globally to CI
- this commit adds a test:ci yarn command which will run the tests in
order of speed/importance, with the -b AKA fail-fast flag so we
don't have to wait for all tests to run before finding out about
issues
- this also cleans up running unit tests in the DB matrix
refs https://github.com/TryGhost/Toolbox/issues/208
- unit tests (shouldn't) be database dependent so they don't need to be
included in the DB matrix
- this prevents an extra 6 executions of unit tests, which isn't a big
deal given they only take 10s to run, but semantically they belong
outside so this commit does that
- also updates the canary step to rely on this extra test run
