If a 0/0 is achieved, via an error and it happens to be any one of the non late scores it will have a score of "Infinity", which is a String and prevents the score of the Submission from being updated. It doesn't matter how late the "Infinity" shows up, as long as it's before the due date. Since the builds will be ordered by score and Strings are ordered before Floats.
The NAN issue that was "resolved" in firstdraft/appdev_template#113 was to address specifically "Nan %" being displayed on assignments that had no specs. The change only modified the display score, not that the actual "score" part of the Build's test_output column. That change is no longer relevant since grades now conditionally hides builds and stuff if there are no tests.
This change makes the "score" zero, if there are no tests and displays a message that there is an error in the tests..
I suddenly stopped being able to replicate this error— but you can check the Submission#"faa0bd24-624f-43af-8597-24ad06399967" for the Infinity nonsense.
Related to https://github.com/firstdraft/grades/issues/444
If a 0/0 is achieved, via an error and it happens to be any one of the non late scores it will have a
score
of"Infinity"
, which is a String and prevents the score of the Submission from being updated. It doesn't matter how late the "Infinity" shows up, as long as it's before the due date. Since the builds will be ordered by score and Strings are ordered before Floats.The
NAN
issue that was "resolved" in firstdraft/appdev_template#113 was to address specifically "Nan %" being displayed on assignments that had no specs. The change only modified the display score, not that the actual "score" part of the Build'stest_output
column. That change is no longer relevant since grades now conditionally hides builds and stuff if there are no tests.This change makes the
"score"
zero, if there are no tests and displays a message that there is an error in the tests..Since
rails grade
is just runningIf you run that
rspec
command in a project has an error that prevents the tests from fully running it produces something like:This change fixes it so that it does
Do this
I suddenly stopped being able to replicate this error— but you can check the Submission#"faa0bd24-624f-43af-8597-24ad06399967" for the Infinity nonsense.