fix: ensure that we can dask.compute multiple collections simultaneously

[douglasdavis]

Fix #299

@lgray your reproducer in #299 runs for me with this PR!

I'm surprised at how small of a code change this is w.r.t. how gnarly of a problem it seemed when I first started digging into upstream dask code 😛

[lgray]

make the reproducer a test?

[lgray]

If we can merge this and cut a new release there's a few things I need for some tests on a batch queue.

[codecov-commenter]

Codecov Report

Merging #312 (e153158) into main (c00b6d3) will increase coverage by 0.00%.
The diff coverage is 100.00%.

❗ Current head e153158 differs from pull request most recent head ee2c45c. Consider uploading reports for the commit ee2c45c to get more accurate results

❗ Your organization is not using the GitHub App Integration. As a result you may experience degraded service beginning May 15th. Please install the Github App Integration for your organization. Read more.

@@           Coverage Diff           @@
##             main     #312   +/-   ##
=======================================
  Coverage   92.31%   92.31%           
=======================================
  Files          21       21           
  Lines        2562     2564    +2     
=======================================
+ Hits         2365     2367    +2     
  Misses        197      197           

Impacted Files	Coverage Δ
src/dask_awkward/lib/optimize.py	93.99% <100.00%> (+0.05%)	⬆️

📣 We’re building smart automated test selection to slash your CI/CD build times. Learn more

[douglasdavis]

make the reproducer a test?

for some reason uproot.dask http (in pytest) takes quite a while when I try to run it locally, I had started putting together something using existing fixtures. It mimics the reproducer and then some. when CI is green I'll merge and cut a release