Still uses gravatar as a default.
Updated URL placeholders to follow LDAP format.
Potential breaking config change: `GRAVATAR=false` replaced by `AVATAR_URL=false`
Builds upon #1111
Extended the base Laravel translation system to
allow a locale to be based upon another.
Also adds functionality to take base & fallback locales into account when fetching
an array of translations.
Related to work done in #1159
For #939
- Adds way to check if current language is RTL via config system.
- Made TinyMCE default direction be based on current language text
direction.
- Fixed bullet points to be RTL compatible.
- Set page content body to have direction based on content.
- Added {$service}_AUTO_REGISTER and {$service}_AUTO_CONFIRM_EMAIL env
options for each social auth system.
- Auto-register will allow registration from login, even if registration
is disabled.
- Auto-confirm-email indicates trust and will mark new registrants as
'email_confirmed' and skip 'confirmation email' flow.
- Also added covering tests.
- Updated homepage selection UI to be more scalable
- Cleaned homepage selection logic in code
- Added seed test data for bookshelves
- Added bookshelves to permission system
LDAP group sync was trying to find users based on the external_auth_id
which is not garunteed to match the username entered so somtimes
the search for a user would fail.
This passes the username to the group sync.
Picked up by @yoyokko in #959.
Fixes permissions with test to cover in the event a page is created,
with permission, in a chapter but the user does not have permission to
see the parent book.
Fixes#912
404 handling now not a hack-around and uses Laravel 'fallback' routes
instead. Prevents errors with the session when you have mulitple errors
on a page where a post/put/delete is made.
Deleted old static file.
Default output depends on app-public setting.
Otherwise can be overidden in `.env` file via `ALLOW_ROBOTS`
Otherwise view file can be customized.
Fixes#779
Added per-entity weighting changes.
Now Books score higher than chapters which score higher than pages.
Reduced queries required on search by only searching once but at a
higher count to see if there's another page.
