Prevent user sign up on LOD imports #11827
Conversation
Use dataset flag to infer if we should allow sign up or not.
github-actions
bot
added
DEV: backend
Build Artifacts
|
|
LGTM - learners on a LOD can no longer sign up. Tested on Android and Ubuntu with imported, created and migrated learners plus regression tested. |
👏🏽
| @@ -19,6 +19,7 @@ | |||
| from kolibri.core.content.constants.schema_versions import MIN_CONTENT_SCHEMA_VERSION | |||
| from kolibri.utils.android import ANDROID_PLATFORM_SYSTEM_VALUE | |||
| from kolibri.utils.android import on_android | |||
| from kolibri.utils.lru_cache import lru_cache | |||
There was a problem hiding this comment.
let's remember to remove this file in 0.17, after stopping python 2.7 support
| @@ -486,6 +487,7 @@ def get_device_info(version=DEVICE_INFO_VERSION): | |||
| return info | |||
|
|
|||
|
|
|||
| @lru_cache() | |||
There was a problem hiding this comment.
Using lru_cache without a max_size does not use lru capabilities, it's a normal cache without memory limit. This is not a problem in this case because not many dataset_ids will be used to call this function. Maybe using a @cache decorator would result in a clearer code for future devs.
There was a problem hiding this comment.
Ah, I was misled by our backport, which has a default max_size parameter: https://github.com/learningequality/kolibri/blob/develop/kolibri/utils/lru_cache.py#L54 - we should ensure other uses of lru_cache in the codebase add a max_size parameter as well, as none of them do.
There was a problem hiding this comment.
Looks like the functools version has a default maxsize too? https://docs.python.org/3/library/functools.html#functools.lru_cache
Summary
References
Fixes #11817
Reviewer guidance
Do a learner only import from a facility that allows learners to sign up.
Ensure that you cannot create an account.
Testing checklist
PR process
Reviewer checklist
yarnandpip)