Repos are generally better since otherwise we end up duplicating
things between front-end and API.
Types moved to by CONST values within a class for better visibilty
of usage and listing of types.
- Tweaked how books are passed on update to prevent unassignment if
parameter is not provided.
- Added books to validation so they show in docs.
- Added request/response examples.
- Added tests to cover.
- Added child book info to shelf info.
Review of #1908
- Moved list to the left of the page to align with other navigational
items.
- Hid list of no shelves, to help hide shelf references if not in use.
- Tweaked test to ensure it wasn't finding shelf name in breadcrumb
rather than list being tested.
* Started mass-refactoring of the current entity repos
* Rewrote book tree logic
- Now does two simple queries instead of one really complex one.
- Extracted logic into its own class.
- Remove model-level akward union field listing.
- Logic now more readable than being large separate query and
compilation functions.
* Extracted and split book sort logic
* Finished up Book controller/repo organisation
* Refactored bookshelves controllers and repo parts
* Fixed issues found via phpunit
* Refactored Chapter controller
* Updated Chapter export controller
* Started Page controller/repo refactor
* Refactored another chunk of PageController
* Completed initial pagecontroller refactor pass
* Fixed tests and continued reduction of old repos
* Removed old page remove and further reduced entity repo
* Removed old entity repo, split out page controller
* Ran phpcbf and split out some page content methods
* Tidied up some EntityProvider elements
* Fixed issued caused by viewservice change
- Created BookChild class to share some page/chapter logic.
- Gave entities the power to generate their own permissions and slugs.
- Moved bits out of BaseController constructor since it was overly
sticky.
- Moved slug generation logic into its own class.
- Created a facade for permissions due to high use.
- Fixed failing test issues from last commits
- Removes 'on*' attributes from elements.
- Also updated script logic to remove scripts instead of escaping.
- All JS injection removal now uses DomDocument + xpath parsing.
Avoids possible permission issues where included content shown in search or preview
where the user would not normally have permission to view the included content.
Closes#1178