Judging runs shown in pending state after doing a repeated rejudging

[eldering]

Description of the problem

Judging runs stay showing as pending, while they should be marked as unevaluated.

Your environment

Current main, actually reproduced on wfbaku branch.

Steps to reproduce

1. Submit a program that fails on all test cases and have lazy evaluation enabled.
2. Rejudge it with a number of times to repeat, e.g. 5 times.
3. Let it get rejudged.
4. Observe that for all rejudgings except the last one, there are judging runs that show as pending, while they should just be marked as unevaluated (with a ?) due to lazy evaluation.