no issue
We changed `reschedule` event to trigger adapter's `unschedule` and `schedule` methods since we now generate separate tokens(urls) for consistency as two different url(token) is needed to complete the reschedule functionality.
no issue
The default scheduling generates a known, independent URL for publishing a resource. In case of resource being rescheduled or unscheduled, the adapter expects the the same URL to remove/update existing jobs. The URL includes a JWT token for API auth which is calculated from post model and appended to URL.
There was a bug in token generation which meant If we go to update or delete the job i.e. unschedule a post then a new token is used which means the existing scheduled job cannot be removed. This PR:
- removes issued at (`iat`) timestamp from token generation which lead to a different token being generated for same payload
- Fixes timestamp being used for URL calculation from resource model
* Updated scheduler to use v2 API by default
* Updated scheduling for post/page resource types
* Extended base method to take options param with token and jwt options
* Updated token expiration to 6 hours after publish/blog start time to allow retries
closes#10060
- Implemented scheduling for posts and pages
- Added cache invalidation when scheduling
- Refactored admin token eneration function to accept existing key as parameter in tests
- Added Ghost Scheduler Integration fixture
- Added fixture for permissions for post publish action
- Migrated getScheduled method to v2
- Did not add support for 'from' and 'to' parameters as they were not used by DefaultScheduler
- This method needs rethinking in a long run as it's an ugly hack and should rather become proper endpoint that returns JSON data instead of models
- Removed unused auth middleware from v2 routes
- Added internal scheduler role
- Implemetnted transactions in v2 frame
- This takes into account scenario mentioned in c93f03b87e
- Specifically:
>if two queries happening in a transaction we have to signalise
knex/mysql that we select for an update
otherwise the following case happens:
you fetch posts for an update
a user requests comes in and updates the post (e.g. sets title to "X")
you update the fetched posts, title would get overriden to the old one
closes#10773
- The refactoring is a substitute for `urlService.utils` used previously throughout the codebase and now extracted into the separate module in Ghost-SDK
- Added url-utils stubbing utility for test suites
- Some tests had to be refactored to avoid double mocks (when url's are being reset inside of rested 'describe' groups)
no issue
- case: restart Ghost and while having a scheduled post
- caused by 4acc375fb6 (diff-4726ce3c4d18d41afad4b46cb0aa7dd3)
- the bug exists since 2.12
- Bookshelf added support (or better said fixed a bug) for accessing previous attributes
- `object.updated('published_at')` always returned "undefined", because the self-implementation < 2.12 only remembered previous attributes after update (see https://github.com/TryGhost/Ghost/blob/2.11.0/core/server/models/base/index.js#L234)
- but `object.previous('published_at')` returns the current value (object.get('published_at') === object.previous('published_at') -> and that's why rescheduling on bootstrap never worked
- might fix https://forum.ghost.org/t/scheduled-posts-never-publish/6873/10
- reduced timeouts on scheduling unit tests
refs #9866
- if we start with v2 controllers, the code base should not require specific api controllers
- because e.g. `require('../api/posts')` will no longer exist
- if you require the api folder, you will get the latest available version by default e.g. `require('../api').posts`
- this branch does not touch the test env (!)
refs #9178
- we have to take care that we don't end up in circular dependencies
- e.g. API requires UrlService and UrlService needs to require the API (for requesting data)
- update the references
- we would like to get rid of the utils folder, this is/was the most complicated change
closes#8568
- use our `urlJoin` util to concatenate the URL (not the query part of it, as this is not supported in `urlJoin`) and to prevent possible missing or double slashes, as `config.apiUrl` could be with or without trailing slash
