[BUG] Fix bug with merge tasks that allows for tasks larger than max size allowed

[samster25]

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Comparison is base (195174c) 85.55% compared to head (e1106d6) 85.50%.
Report is 1 commits behind head on main.

❗ Current head e1106d6 differs from pull request most recent head fc9b5be. Consider uploading reports for the commit fc9b5be to get more accurate results

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #1882      +/-   ##
==========================================
- Coverage   85.55%   85.50%   -0.05%     
==========================================
  Files          55       55              
  Lines        6194     6194              
==========================================
- Hits         5299     5296       -3     
- Misses        895      898       +3     

see 2 files with indirect coverage changes

[jaychia]

Mostly LGTM, had some questions about the logic of self.accumulator_ready()

[jaychia]

Why is a None accumulator considered ready?

In the main loop logic, ready accumulators are naively yielded. Should the accumulator_ready function instead yield true only for these:

1. It is Some, and has no size (we won't be merging this anyways, might as well say its ready to yield)
2. It is Some, has a size, and the size is smaller than min

[samster25]

done