Superdesk-core

Fixed

- Report original exception details to sentry.
- Ensure username is unique.
- Remove the link to other stories.
- Keep renditions until item expiry.
- Fix renditions if published it can't be removed.
- Activate token should be active for 7 days.
- Fix bad argument handling in data upgrades.
- Fix image cropping if mimetype is missing for some rendition.
- Fix `place` filter condition.
- Reset `event_id` for unlinked stories.
- Fix resending of stories when `NO_TAKES` is on.
- Fix image cropping for renditions defined via vocabulary.
- Fix dateline is required for picture item kill.
- Fix push to send all renditions of package items and embedded items.
- Fix reuters content formatting.
- Fix `item:expired` notification to send item id.
- Fix rss guid handling when it's not permalink.
- Fix email subject with utf8.
- Stop retrying enqueue on error.
- Make word counts consistent with UI.

Added

- Add mark stories for desks.
- Add endpoint to generate content api tokens for subscribers.
- Activity reports generating and management.
- Add audit collection gc.
- Add EFE specific parser.
- Password-less authentication via xmpp.
- Document authentication.
- New operators for filter conditions.
- Store ingested data that could not be parsed.
- Add `CELERY_PREFIX` config option.
- Add `evolvedfrom` to ninjs output.
- Copy metadata field from parent to associated items.
- Add archived collection curation functions.
- Add group label to macros.
- Add filter to block api content.

Changed

- Refactor ingest providers registry.
- Refactor search providers registry.
- Add associations to rewrites and takes.
- Update celery to 4.0.
- Make single-line view a user preference.

Fixed

- Update sign off if mapped field is updated on user.
- Fix editing content profiles with custom fields.
- Fix publishing with content profiles.

Added

- Implement email publishing to multiple users via bcc.

Changed

- Bypass content validation using profile for auto published content.

Fixed

- Clean up old item locks.
- Mark search provider as used when an item is fetched.
- Fix issues with dateline validation for content profile.

Fixed

- Push notification when dictionary is created/updated.
- Fix `backend_meta` service on python3.4.
- Parse byline from rss.
- Fix slow query in duplicate item.
- In app factory load `CORE_APPS` before `INSTALLED_APPS`.
- Clear SMS flag for new takes and updates.
- Load init files from core if missing in data folder.
- Fix xml file ingest with BOM or non-unicode encoding.

Changed

- Set default content expiry to `0`.
- Use `CONTENTAPI_URL` setting for content API.
- Use projections in search results.
- App init command won't override modified data. Use `-f` to keep old behaviour.
- Make `byline` empty on kill.
- Remove `By` prefix from `byline` in nitf output.
- Initial [Content API Refactoring](http://superdesk.readthedocs.io/en/latest/contentapi.html).

Added

- Allow duplication of items in text archive.
- Add `Place` to criteria fields for content filters.
- Add `PUBLISHED_CONTENT_EXPIRY_MINUTES` for published items.
- Add `firtscreated` field to superdesk ninjs output.
- Add `desk:preferred` user preference.
- Add `suggestions` resource for related items suggestion.

Fixed

- Fix comments notifications for users with `.` in username.

Fixed

- Fix parsing of non-unicode nitf.
- Fix ingestion with refresh rate < 1m.