refs https://github.com/TryGhost/Team/issues/1723
- Added count.replies to comments
- Added replies endpoint
- Limited returned replies to 3.
- Replaced likes_count with count.likes in comments
- Instead of fetching all the likes of a comment to determine the total count, we'll now use count.likes
- Instead of fetching all the likes of a comment to determine whether a member liked a comment, we'll now use count.liked (which returns the amount of likes of the current member, being 0 or 1). This is mapped to `liked` to make it more natural to work with.
The `members.test.snap` file changed because we no longer include `liked: false` if we didn't fetch the liked relation. And in the comments events of the activity feed the liked property is therefore removed.
These changes requires an update to the `bookshelf-include-count` plugin:
- Updated to also work for nested relations
- This moves the count queries from the `bookshelf-include-count` plugin to the `countRelations` method of each model.
- Updated to keep the counts after saving a model (crud.edit didn't return the counts before)
- we shouldn't need individual LICENSE files because these packages
won't be published, so the top-level one applies
- also cleaned up README files to remove mentions of Lerna monorepos and
install instructions
refs https://github.com/TryGhost/Toolbox/issues/354
- set packages to `private: true`
- removed repository link - these packages won't be published so this
link won't be seen anywhere
- removed `publishConfig`
- As of Ghost 5.0 we only use the V2 version of jsonErrorRenderer
- Removed the old one, and renamed the V2 to not have a suffix any more
- Added 100% coverage to tests whilst here
- trying to call new RequestNotAcceptableError with a variable triggers a lint warning in newer versions of eslint-plugin-ghost
- this workaround is worth it for the safety of not allowing single strings to be passed in!
- this middleware block is used in v4 but not in v5
- we want to remove it and then rename handleJSONResponseV2 so that we have one single consistent error handling block
refs https://github.com/TryGhost/Toolbox/issues/292
- There's a need to distinguish different types of RequestNotAcceptableError erros by their code. The code is also having an instructional name to give it more explicit utility (nice clue for a developer seeing the error)
refs https://github.com/TryGhost/Toolbox/issues/280
- When an outdated client receives a 404 as a response there's no clear way to act on it. Plain 404 says nothing about need to update.
- In such cases the resourceNotFound handler should return a 406 error indicating the Ghost instance needs an update.
refs https://github.com/TryGhost/Toolbox/issues/280
- When an outdated client receives a 404 as a response there's no clear way to act on it. Plain 404 says nothing about need to update.
- In such cases the resourceNotFound handler should return a 406 error indicating the client need to update.
no issue
- When adding more methods under test to this file the require statement will becom difficult to read in one long line. Multiline statement is way nicer to track in git history too
- in the event we get an unknown error bubble up, we don't handle the
templating on the error name
- `@tryghost/tpl` throws an error because we pass an undefined string:
`Cannot read properties of undefined (reading 'replace')`
- this commit adds handling to fallback to a different user message in
that event so we don't cause a 500 error
no issue
Change to error handling caused all theme errors to be reported in Sentry, this fix (and a respective fix in Ghost) allows the error to be prepared for sentry before replacing the stack
refs: https://github.com/TryGhost/Team/issues/1369
If we prepare the error for users to view before using Sentry, then the error passed to Sentry will have the stack trace removed for production environments.
@tryghost/errors@1.2.5 also made it so that the error is not mutated, but cloned and a new one is returned.
refs https://github.com/TryGhost/Toolbox/issues/203
- without `--all`, c8 will ignore files that aren't covered in tests, so
they won't pull the test coverage down
- this means we have artificially high coverage scores
- this commit adds `--all` where previously missing
