add DB migration and fix earlier mistakes in student migration history
add tests and fix bugs that came out of those unit tests
remove unused import
pep8/pylint
address some PR feedback
fix tests
fix broken test
try to mock time
use freeze-gun to overload the system time to simulate the future
ConfigurationModels can be managed using the admin site. They are
append-only, and track the user who is making the change, and the time
that the change was made. The configuration is stored in the database,
and cached for performance.
[LMS-1220]
To mark a language as dark-launched, add it to the DARK_LANGUAGES
django conf setting. To activate a dark-launched language, set he
query parameter `preview-lang` to the language code on any url.
[LMS-2045]
[LMS-2077]
[LMS-2076]
configure middleware
add test for session inactive timeouts
add Studio inactive session timeout test
change login method used
add create_test_account to test
make sure the expected redirect URL is right
fix indenting problem
fix doc string since we moved from minutes to seconds
use utility methods rather than calling another test to set up and activate an account
clean up code violations
respond to PR feedback
use optional params to make code cleaner
pylint fix on test files
fix errorenous logic when running a microsite that could reside in a hosting environment with a marketing site in front of it
pep8/pylint fixes
address PR feedback, remove underscore from test hostname
more pep8/pylint cleanup. Skip test_microsites test, it works on localdev, not on Jenkins. Need to talk with QA team
manually add Ned's single-to-double quote fix
change aws.py runtimes so that the microsite_dir that is read from configuration is changed to a python path
Conflicts:
lms/templates/help_modal.html
Refactored stub services for style and DRY
Added unit tests for stub implementations
Updated acceptance tests that depend on stubs.
Updated Studio acceptance tests to use YouTube stub server; fixed failing tests in devstack.
I suspect there are very few circumstances that an enrollment event is emitted from the CMS, but if it ever were to happen, it would bomb out because the "crum" middleware is not used in that app. We are hoping to replace this request caching hack in the near future, but for now it is the best option we could come up with.
I noticed the stack appear when I created a new course in my dev environment. Presumably the course author is automatically enrolled in the course, which, causes the tracking code to bomb out. Luckily we catch the error and just log it, but we really should be capturing this type of enrollment event.
@jzoldak showed me that setting `USE_I18N = DEBUG` didn't actually accomplish
what I had hoped it would -- changing DEBUG didn't also change USE_I18N.
This pull request accomplishes what I was trying to accomplish, without being
quite so clever about it.
