Wrap failures in the woven final-field writer like the runtime accessor

[evanchooly]

Fixes #4355

The __writeXxx method that AddFieldAccessorMethods#writeFinalField weaves for final fields called Field.set without handling its checked IllegalAccessException, so a failure surfaced as an undeclared checked exception.

The Field.set call now sits in a trying(...) block that catches Exception and rethrows new RuntimeException("Failed to set final field '<name>'", e). This is the same type, message and catch scope as the runtime path in NestmateAccessorGenerator, so both tiers fail the same way. The catch covers all Exceptions, not only IllegalAccessException, to match the runtime tier, and every failure names the field. The Field lookup and caching stay outside the try.

Test: TestGeneration#testGeneratorFinalFieldWriteFailureIsWrapped puts a Field that was never made accessible into the woven class's __fieldCode cache. That makes Field.set throw IllegalAccessException through the AOT accessor. The test then checks the exception type, message and cause. With the fix reverted, the test fails because the raw IllegalAccessException reaches the caller.

Results: core 1300 tests (16 skipped), 0 failures under both critter and reflection. critter-maven verify passes (6 unit tests and 3 invoker tests).

Merge order

This PR is one of a batch that came out of the critter review issues (#4352–#4359). Suggested merge order:

1. Run each JDK test leg on its own JDK #4361, Add a non-blocking CI leg that denies final field mutation (JEP 500) #4360 (CI only, independent)
2. Don't re-set properties already passed to an entity's constructor #4362, then Test record entities with both mappers #4366 (Test record entities with both mappers #4366 depends on Don't re-set properties already passed to an entity's constructor #4362)
3. Wrap failures in the woven final-field writer like the runtime accessor #4364, Give mapper copies with different mapping settings their own runtime models #4363, Name the cause when critter runtime generation can't access an entity #4365 (+ Document critter runtime generation's class loader requirement morphia-docs#20), Critter clean-ups: CritterClassLoader.findClass ordering and stale docs #4367 (independent, any order)
4. Strengthen critter mapper tests: tier assertions and weak-cache release #4368 last (conflicts with Give mapper copies with different mapping settings their own runtime models #4363 and Name the cause when critter runtime generation can't access an entity #4365 in TestCritterMapper.java; rebase after they merge)

This PR: no dependencies; merges cleanly with the rest of the batch.

[Copilot]

Copilot review overview

🟢 Approval recommended

The focused change aligns exception behavior and includes targeted regression coverage, with no unresolved findings.

Review effort: Balanced
Findings: None

What changed in this PR

Aligns woven final-field write failures with runtime accessor behavior, addressing #4355.

Changes:

• Wraps Field.set exceptions in a field-specific RuntimeException, preserving the cause.
• Adds a regression test checking the exception type, message, and cause.

File	Description
core/​src/​test/​java/​dev/​morphia/​critter/​parser/​generator/​TestGeneration.java	Tests wrapping of a forced final-field write failure.
core/​src/​main/​java/​dev/​morphia/​critter/​parser/​generator/​AddFieldAccessorMethods.java	Adds exception wrapping while leaving field lookup and caching unchanged.

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.