Refactor JUnitReportEngine into focused partial-class files aligned with TrxReportEngine layout

[Copilot]

JUnitReportEngine.cs had grown into a single large unit (~800+ LOC), making navigation and targeted reviews expensive. This PR restructures it into focused partial files, matching the established TrxReportEngine organization while preserving behavior and API shape.

• Refactor scope

  • Split JUnitReportEngine into concern-specific partial files:
    • JUnitReportEngine.cs (orchestration + retry/write flow entry points)
    • JUnitReportEngine.XmlWriting.cs (JUnit XML serialization)
    • JUnitReportEngine.SuiteBuilding.cs (result → suite/testcase model shaping)
    • JUnitReportEngine.FileNaming.cs (default name/template resolution/sanitization)
    • JUnitReportEngine.Models.cs (internal nested DTOs: SuiteSet, Suite, TestCase)
  • Kept internal sealed visibility and class semantics unchanged.

• Behavioral invariants preserved

  • No public/internal API changes.
  • No logic changes to:
    • report generation pipeline (GenerateReportAsync / retry behavior),
    • suite aggregation and duplicate testcase disambiguation,
    • XML element/attribute emission order and escaping,
    • filename template resolution and invalid/reserved-name handling.

• Codebase consistency

  • Naming and file boundaries now mirror the sibling TRX engine pattern, reducing cognitive switching cost across report providers.

// Before: one monolithic file
internal sealed class JUnitReportEngine { /* orchestration + xml + model-building + file naming + models */ }

// After: same type split by concern
internal sealed partial class JUnitReportEngine { /* orchestration */ }
internal sealed partial class JUnitReportEngine { /* xml writing */ }
internal sealed partial class JUnitReportEngine { /* suite building */ }
internal sealed partial class JUnitReportEngine { /* file naming */ }
internal sealed partial class JUnitReportEngine { /* nested models */ }

[Copilot]

Copilot wasn't able to review any files in this pull request.