Release v1.11.0, December 2021

This is identical to v1.10.5. This is tagged again so that it matches the A-plus version 1.11.

Release notes:
https://apluslms.github.io/releases/v1_11.html

Release v1.12.0, December 2021, Release Candidate 1

Major changes:

• Remove the gitmanager. The new Git manager is a separate service.
• Add new API for setting up and updating courses. The new Git manager needs it.
• Add new JWT authentication and authorization to the HTTP requests between A-plus and MOOC-Grader. It is also possible to log into the grader web site with JWT tokens (but usually using the web UI is not necessary). JWT tokens make the system more secure. IP address-based firewalls are not the only authorization mechanism any longer.
• Support multiple mounts in exercise grading containers. There is a new field mounts in addition to the old mount in the container field of the exercise config.yaml.
• The script for starting asynchronous grading (containers) has been converted to a Python module from a shell script. The module to use may be changed in the settings as before.

Release notes:
https://apluslms.github.io/releases/v1_12.html

Release v1.10.5, December 2021

Bug fixes:

• Improve error handling particularly in parsing course configurations.
  For example, the MOOC-Grader server used to crash if a course had syntax errors in its YAML files. Now, the server does not respond with HTTP 500 Internal server error and instead, more descriptive error messages are returned.

Release v1.10.4, November 2021

Bug fix:

• Course configurations cache: detect changes in the included base configuration (YAML) templates in exercises and update the cache. In the past, the cache did not load changes in the exercise if the changes were made in the base template without modifying the exercise config.yaml file itself. The base template is defined with the include field.

Release v1.10.3, October 2021

Bug fixes:

• Fix the model solutions of radio button questions in questionnaires when the choices included both correct and neutral choices. The model solution used to show validation errors and if the question included feedback, the server crashed.

Release v1.10.2, September 2021

Fix a crash when the request does not have any lang parameter.

Release v1.10.1, September 2021

Bug fix:

• Use only short language codes "en" even if
  the lang parameter of the grader protocol includes
  a long language code "en-gb".
  The old code fails to load English course contents
  that are defined with the key "en" when the grader
  uses the long key "en-gb".

Release v1.10.0, September 2021

Release notes: https://apluslms.github.io/releases/v1_10.html

Release v1.10.0, August 2021, Release Candidate 1

v1.10.0rc1

Release v1.10.0, August 2021, Release Candidate 1

v1.9.1, August 2021

Release v1.9.1, August 2021

Changes:
* Add a new configuration field checkbox_feedback to
  questionnaire fields (questions) with type: checkbox.
  If this field is set, the feedback/hints for selected options
  is rendered under the checkbox instead of the list
  under the question after all checkboxes.
  By default, all feedback is rendered under the question.
  This reverts the change from v1.8 (summer 2020) that
  always forced the feedback to be rendered under the checkbox.
  (Though, inverse feedback when the option is not selected
  is always rendered under the question.)
* Show hints/feedback in questionnaire model solutions.
* Show hints/feedback when the correct answer is revealed
  in questionnaires after the last submission attempt
  (when reveal_model_at_max_submissions true).
* Add a new environment variable ENV_SETTINGS_PREFIX for defining
  a prefix for other environment variables that are used
  to update Django app settings.
  Remove the DISABLE_LOAD_ENVIRONMENT_SETTINGS variable.