If the first date the user selects is invalid it is not set on the
model. This is because jQuery datepicker's getDate returns the
current date when an invalid value is put in the field and the
model's initial value is also the current date. So always update
the date.
TNL-4366
Changes in studio to allow the hide_after_due setting to be utilized.
Includes:
-python changes to contentstore, where the data is stored.
-refactoring of timed-examination-preference-editor.underscore, to add this
setting and make the editor more accessible.
-javascript changes to link the new setting to the correct data model.
-sass updates to fix a11y issues on the editor modal.
-addition of post-due visibility information to course outline in studio.
-new tests: python, js, acceptance, and a11y
TNL-4115. Previously, course updates (which are intended to be posted with
dates, for sorting in the LMS) could be authored in studio with a valid
date, nothing, or a random string as the "date" for the update. As there
is no validation for this in studio, everything succeeded with no warning.
However, the LMS has problems parsing some of these values, and barfs when
loaded by learners.
The fix does two big things:
- gracefully handles invalid dates in LMS. These updates are now treated as
having a date of today, for sorting purposes.
- turns on validation in studio. Now, it is not only impossible to enter
invalid dates in studio, but notifications will draw the course author's
eye if any invalid updates were previously saved.
Test additions for this commit:
Adds:
- unit test for LMS parsing
- Jasmine test to confirm invalid dates cannot be set by the user
-also adds event to setAndValidate instead of using a global object
- fix for lettuce test
-It is no longer valid to enter the string "January 1, 2013" as this test
had been doing. Keyed-in entries must use MM/DD/YY format.
- adaptation asides to be imported from the XML
- updating SplitMongo to handle XBlockAsides (CRUD operations)
- updating Studio to handle XBlockAsides handler calls
- updating xblock/core.js to properly init XBlockAsides JavaScript
Make base.html Mako template safe by default by:
1. Add page-level default of html escaping
2. Fix escaping of all variables in base.html
3. Fix escaping of all dependent underscore templates
Also includes additional best practices for certificates and
textbooks JavaScript/Underscore in order to complete that work.
TNL-3425
The function initializeXBlock() expects a DOM element, and is passed one in most
cases. However, when adding a new XBlock component in Studio, the function is
passed a jQuery object, which ends up being forwarded to the actual
initialisation function of the XBlock.
It turns out that loading JavaScript with $.getScript
causes Chrome to treat the file as an XHR request
and not JS. I've switched over to using RequireJS
to load the URL which already has the ability to
dynamically load files and have the browser
recognize them.
added unit tests and fixed JS tests
added bok choy tests
fixed broken botchoy tests
fixed course discovery broken test
removed monkey patch at class level
changes after feedback from matte
Changed 'forceRoundTime' to 'false' in BaseDateEditor
Updated bok choy acceptance test 'studio.test_studio_outline.py:EditingSectionsTest.test_can_edit_subsections' to check times aren't rounded to the nearest half hour after saving
Also added myself to AUTHORS
