Fix custom field serializer inside a flattened field

[gaoflow]

Fixes #453.

A custom field serializer on a field of a flattened dataclass crashed serialization:

@serialize
class Child:
    value: int = field(serializer=str)

@serialize
class Parent:
    child: Child = field(flatten=True)

to_dict(Parent(child=Child(value=3)))  # SerdeError: name 'value_serializer' is not defined

Cause

A flattened dataclass field is rendered inline in the enclosing class's generated serialize function, so it emits references like value_serializer(...). But the loop that registers custom field serializers (and skip_if predicates) in the generated globals only iterated the top-level fields — the flattened child's serializer was never put in scope.

Fix

Recurse into flattened (and optional-flattened) dataclass fields when collecting field serializers / skip_if predicates, so they are available in the enclosing scope. Handles nested flattens too.

Tests

Added test_flatten_custom_field_serializer (basic + nested flatten; fails without the patch). Full suite green (5169 passed). Deserialization was already unaffected (a flattened child is deserialized via its own scoped function).

Thanks @AustinScola for the report and minimal reproducer.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 90.98%. Comparing base (f09d503) to head (5f62546).
⚠️ Report is 75 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #756      +/-   ##
==========================================
+ Coverage   90.83%   90.98%   +0.15%     
==========================================
  Files          13       13              
  Lines        2040     2352     +312     
  Branches      370      462      +92     
==========================================
+ Hits         1853     2140     +287     
- Misses        124      140      +16     
- Partials       63       72       +9     

☔ 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.