test: simplify memory tests

[wjones127]

Replace the memory integration tests that used spans with one that tracks general statistics in an custom allocator. Advantages:

• Guaranteed not to miss any allocations. Previously due to issues with spans not propagated across all Tokio tasks, we would miss allocations.
• Works on multiple platforms

Disadvantages:

• Running with plain cargo test will have failures
• Still can't run in unit tests.

As a side note, I did look into exposing this in Python, so we could use it in Python tests. However, I found it wasn't useful in many cases because it only tracked allocations from the Lance binary. This meant it missed allocations made by PyArrow, for example, so the insertion tests couldn't be written in Python. If we want to do allocation tracking in Python, we will probably have to do something that uses the LD_PRELOAD trick.

[codecov-commenter]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 81.94%. Comparing base (4d2bcb9) to head (902d9f6).
⚠️ Report is 2 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #5214      +/-   ##
==========================================
- Coverage   82.02%   81.94%   -0.08%     
==========================================
  Files         342      342              
  Lines      141495   141516      +21     
  Branches   141495   141516      +21     
==========================================
- Hits       116059   115969      -90     
- Misses      21595    21677      +82     
- Partials     3841     3870      +29     

Flag	Coverage Δ
unittests	81.94% <ø> (-0.08%)	⬇️

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.