fix(rust): Handle empty lists in list.agg with scalar aggregations

[Zakir032002]

Fixes #26237

When list.agg() is used with scalar aggregations (such as first()) on columns
containing a mix of nulls and empty lists, Polars may panic with an assertion
failure in deposit().

This occurs because scalar aggregations on empty lists can produce no output,
while deposit() expects one value per valid group based on the validity bitmap,
leading to a length mismatch.

This PR pads the aggregation result with null values when necessary so the number
of values matches the number of valid groups before calling deposit().

The fix is applied to functions:

• evaluate_on_list_chunked
• evaluate_on_array_chunked

AI disclosure

• I used AI assistance to help reason about the root cause .
• I have personally reviewed all code changes and confirm they are correct,
  relevant, and consistent with the existing codebase.

[codecov]

Codecov Report

❌ Patch coverage is 42.85714% with 8 lines in your changes missing coverage. Please review.
✅ Project coverage is 81.01%. Comparing base (c1c140f) to head (1dacac6).
⚠️ Report is 3 commits behind head on main.

Files with missing lines	Patch %	Lines
crates/polars-expr/src/expressions/eval.rs	42.85%	8 Missing ⚠️

Additional details and impacted files

@@           Coverage Diff           @@
##             main   #26424   +/-   ##
=======================================
  Coverage   81.01%   81.01%           
=======================================
  Files        1782     1782           
  Lines      243119   243130   +11     
  Branches     3078     3078           
=======================================
+ Hits       196954   196972   +18     
+ Misses      45362    45355    -7     
  Partials      803      803           

☔ 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.

[orlp]

I don't think this is the right solution, to me the problem should be fixed elsewhere. I don't see why 'some aggregations may produce fewer values' should be acceptable, and should be hidden by just extending with nulls.