@filmor have you tried just changing the runtime and leaving the tests as-is for a first pass? I looked through the runtime and didn't see a mistake so I think I would start by ruling out problems with the test updating (which can easily be done in a followon branch)

Not needed. The failures before were due to one place where I forgot to replace text_type by str and instead removed it entirely, so instead repr was called which was buggy at the time. I fixed that (and a few other missing replacements of long and a missing import of operator) and now tests are passing. I'm not a big fan of doing multiple PRs for the same "change".

Codecov Report

Merging #1158 into master will decrease coverage by 0.27%.
The diff coverage is 100.00%.

@@            Coverage Diff             @@
##           master    #1158      +/-   ##
==========================================
- Coverage   86.53%   86.25%   -0.28%     
==========================================
  Files           1        1              
  Lines         297      291       -6     
==========================================
- Hits          257      251       -6     
  Misses         40       40              

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 140a9b8...cc4211c. Read the comment docs.

Also, if you don't increase the coverage back to the target, or lower the target a bit, it is going to complain on every subsequent PR.

The coverage is always against master, it will not repeatedly complain.