Clean up stale comments and in-place uncertainty ops in Combiner

[mwcraig]

#978 is merged; this is now rebased onto main and stands alone as a single commit.

Small cleanup of the Combiner.median_combine / average_combine / sum_combine methods:

• Rewrite the stale comment claiming the median uncertainty "still uses numpy" (the code calls uncertainty_func, not np.ma.median).
• Replace the three stale # Convert uncertainty to plain numpy array (#351) comments (the code converts to the combiner's array namespace via xp.asarray).
• Drop a redundant second uncertainty = xp.asarray(uncertainty) in median_combine.
• Rewrite the in-place uncertainty /= ... / uncertainty *= ... as rebinding so the methods do not rely on mutable arrays (immutable backends such as jax).

No behaviour change; no new tests. The diff is CHANGES.rst (+3) and ccdproc/combiner.py (+21/-23).

Test matrix

Measured on this branch after the rebase. Every figure is identical to main at 034c77e, which is the point of a no-behaviour-change cleanup.

backend	pytest ccdproc	vs main
numpy	410 passed, 5 skipped	unchanged
jax (JAX_ENABLE_X64=True)	396 passed, 10 skipped, 2 xfailed, 7 xpassed	unchanged
dask	399 passed, 16 skipped	unchanged
dask + escape baseline enforced	399 passed, 16 skipped	unchanged
array-api-strict, test_combiner.py	68 failed, 23 passed, 1 xfailed	unchanged

The strict test_combiner.py failures are pre-existing and tracked in #971 — device and np.asarray boundary issues, unrelated to this cleanup.

An earlier revision of this description quoted a 393 / 379 / 383 matrix and "74 failed, 13 passed" under strict. Those predate both #985 (which added coverage upload from the jax and dask jobs) and the rebase onto the reviewed #978, and are superseded by the table above.

🤖 Generated with Claude Code

https://claude.ai/code/session_01DTN9DnPnLKK2u7knnMJ2gA

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 96.31%. Comparing base (e29ca51) to head (c37c0de).
⚠️ Report is 1 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #980      +/-   ##
==========================================
- Coverage   96.32%   96.31%   -0.01%     
==========================================
  Files           9        9              
  Lines        1685     1684       -1     
==========================================
- Hits         1623     1622       -1     
  Misses         62       62              

Flag	Coverage Δ
dask	95.41% <100.00%> (-0.01%)	⬇️
jax	95.53% <100.00%> (-0.01%)	⬇️
numpy	96.19% <100.00%> (-0.01%)	⬇️

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

☔ View full report in Codecov by Harness.
📢 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.