WIP: Fix writing compound dtype datasets with a single field

[rly]

Summary

• Fix compound dtype detection in __list_fill__ to handle single-field compound dtypes (e.g., HERD's keys table with dtype [('key', 'O')])
• The old check len(dtype) > 1 missed single-field compound dtypes, causing get_data_shape to return a 2-D shape that mismatched the 1-D maxshape, producing an h5py error
• Use dtype.names is not None to correctly detect all compound dtypes

Fixes #1419

Test plan

• test_write_dataset_list_single_field_compound_datatype — unit test for writing a single-field compound dtype dataset
• test_write_herd_as_child — integration test writing/reading HERD as a child group of a container (mirrors the NWB use case)
• Existing compound dtype and HERD tests continue to pass

🤖 Generated with Claude Code

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 93.06%. Comparing base (0897b93) to head (790f249).

Additional details and impacted files

@@            Coverage Diff             @@
##              dev    #1420      +/-   ##
==========================================
+ Coverage   92.85%   93.06%   +0.21%     
==========================================
  Files          41       41              
  Lines        9989     9989              
  Branches     2054     2054              
==========================================
+ Hits         9275     9296      +21     
+ Misses        436      417      -19     
+ Partials      278      276       -2     

☔ 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.