We have been bucketing all users into the relative dates experiment
since May 18, 2020. We no longer need to keep this as an
ExperimentWaffleFlag and can convert to a CourseWaffleFlag (so it
continues to support exemptions).
* Commits:
feat: Implement POST endpoint for discussions API
feat: Implement update helper for discussions API
feat: Implement to_internal_value helper for discussion API
feat: Add lti_configuration to discussions API payload
refactor: Base discussions serializer off of ModelSerializer
refactor: Remove unused helper from discussions API
fix: Use the correct name for serializer to_internal_value
The /jump_to/ LMS endpoint is used in a number of places
to direct users to courseware. It currently only redirects to
Legacy courseware URLs, which then conditionally may
redirect to the Learning MFE.
Two issues with this:
1. Performance Impact: In most cases, going to Legacy first
is just an extra redirect.
2. Confusion for Privileged Users: Neither course nor global
staff are auto-redirected from the Legacy experience to the
MFE. Thus, these priviliged users confusingly never see the
MFE by default; they must always manually click into it.
This commit makes it so that /jump_to/ directs
users to whatever the default courseware experience is
for them. For staff of courses active in the new experience,
this will impact (at a minimum) the "View Live"
links in Studio, all links on the old and new LMS
course outline, and the "Resume" links on the course
dashboard. Learners should see no difference other than
a performance improvement when following courseware links
from the LMS.
This also adds an optional 'experience=[legacy|new]'
query param to /jump_to/, allowing us to specifically
generate Legacy courseware URLs for the
"View in Legacy Experience" tool.
TNL-7796
This:
1. Introduces a new override using the `pluggable_override` decorator.
It is now possible to specify a custom way of getting XBlock's icon
by defining `GET_UNIT_ICON_IMPL` in settings.
2. Introduces a way to add custom `XBLOCK_MIXINS` by defining
`XBLOCK_EXTRA_MIXINS` in settings. This allows, e.g. to add
new fields to XBlocks.
* MST-734 Fix production issue on Learner Onboarding Status Panel
Fix the prod issue where learning sequence service object missing the needed get_user_course_outline service API
The user web API call currently returns 500
This was initially introduced as a temporary flag to be able to get more
information. But if we get this kind of issue again, we'll need
something like this logging to determine the source of the session
collision. Rather than removing the code and adding it back in later,
convert this temporary switch into an opt-in setting that can be used
again in the future.
BREAKING_CHANGE: 'safe_session.log_request_user_changes' switch no
longer exists and is replaced with the 'LOG_REQUEST_USER_CHANGES' django
setting which defaults to 'False'
Previously they also had to have a valid JWT cookie which led to a weird
corner case where a user was logged in but still showed the login form
resulting in some confusion and odd behavior.
This change gives precedence to the session token to determine whether
or not someone is logged into the LMS but ensures that if you go through
the login flow, you refresh your JWT cookies. This should not cause any
breakage for MFE flows that might redirect to the LMS login page since
the JWT would get refreshed if it's out of date but the session is
valid.
Six frames was not enough because for DRF views the request gets wrapped
in a proxy object and so we need more of the stack to see what part of
the code we're in that actually invokes the use change.
Exposed the Date header on the outline api so clients can accurately compute times relative to the dates returned by the API; this was previously done with the course API (#26979)
Browser time is notoriously unreliable for this, especially for a Learner-facing countdown call-to-action based on the access expiration date. (REV-2126)
Using the Date header for this allows the client to make use of information that is already sent, does not require additional calls nor modifying the API, and could be generalized to more or all our APIs without modifying them.
