- Engine will no longer error on importing a zip with multiple manifest files. This behavior matches more what our Cloud site does. It will choose which manifest to use based on a prescribed order of importance: cmi5 > SCORM (1.2/2004) > AICC > TinCan > SCORM 1.1
- Resolved some minor discrepancies between .NET and Java endpoint results
- Add ability to specify the 'tracking' parameter on the registration launchLink endpoint in the API
- Engine will no longer try to fetch metadata for URLs from the example domains (example.com, example.net, etc.), URLs that include ":scormdriver:", and URLs without a top level domain
- For Concurrent Launch, converted HTML output strings to translated strings
- Made xAPIPipe processing smarter with URLs in a slightly incorrect format
- Updated redaction logic in Java to help with connection string debugging
- When importing an AICC course by passing a url to the manifest file, we will no longer strip parameters from the manifest url, and will pass them along on all four file requests.
- Removed limited use of MD5 hashing to make Engine FIPS compliant
- Added a setting to allow configuring system-level configuration settings through the REST API using a virtual "system tenant name" [see ApiConfigurationSystemTenantName]
- Adds missing cloudfront DLL to .NET release.
- Fixed malformed policy error with Cloudfront cookies created with the .NET version of Engine
- Fix bug with Cloudfront auth cookies for AICC content
- Updated some of our URL encoding to fix a bug with package names that have special charcters
- Fixed a potential caching bug with tenant-scoped configurations for customers running without a system database
- Fix potential arithmetic overflow error when importing AICC course with very large max_time_allowed value (.NET Engine only)
- Fix for launching courses on Oracle that have an empty item_identifier
- Fixed a bug with how some 3rd party AICC data is handled in .NET environments
- Fixes the schema audit of the upgrade tool when run on PostgreSQL 10.