This fixes TNL-7973 [1].
Background
----------
> An alert was sent out because a malformed capa problem caused block
> transformers to fail to run.
> This specific error/alert was triggered when an optionresponse problem
> lacked text. The authoring-based fix is to edit the offending
> option-response problem to remove the empty option.
> Expected behavior: We should be able to keep the error local to the
> ProblemBlock and not blow up the whole course publish block
> transformer collection process.
> This is a really easy error for authors to make, and the consequences
> to them (one problem doesn’t work) are disproportionate to the
> consequences [...] (alerting triggered).
- [1] https://openedx.atlassian.net/browse/TNL-7973
Reverts #26731
The code in common/lib/capa/capa/safe_exec needs
to remain Python 3.5-compatible, since edx-sandbox
(ie codejail) is still running Python 3.5.
django-not-configured is an error raised by pylint (with
the pylint-django plugin) when it's not correctly configured.
We should not be applying lint amnesty for such a violation.
The limits can be defined in settings.CODE_JAIL['limits_overrides'],
which is a dict mapping context ids (generally, course run keys)
to overrides to be applied to settings.CODE_JAIL['limits'].
This will allow us to temporarily alter the codejail limits
for certain contexts in order to enable, for example,
a certain course run's tasks to run longer while a large exam's
grades are recomputed.
TNL-7649
* Workaround the error in find_question_label.
This is a temporary log, which will give us information and unblock the course team.
PROD-1326
* raise exception after logging it.
See the commit that added round_away_from_zero for details.
IN the case of dashboard_data.py:
We don't first cast to decimal here because in python 2 this didn't do
anything. The python 2 rounding method would just cast things to float
before doing the rounding anyway so it just wastes cpu cycles. In
python 3 this causes some typing issues so just removing it.
We did this already for the 'get_html' function for a capa problem but
there wasn't a test to ensure this works correctly for the
get_question_answer function that powers the show answer button.
I ran into it while doing a quick smoke test on the python 3 sandbox.
