Preserve backtrace through JsError → JsValue → JsError round-trip

[AlbertMarashi]

Summary

Closes #4608
Supersedes #4607

• Errors caught by internal exception handlers (e.g. async module evaluation) now preserve their backtrace through the JsError → JsValue → JsError round-trip via promise rejection
• Stores the backtrace on the Error object in JsError::into_opaque (both Native and Opaque variants) and recovers it in JsError::from_opaque
• Moves backtrace capture in handle_error before the exception handler check, so catchable errors also get a backtrace before being intercepted by internal handlers
• Adds regression tests

Details

Previously, the backtrace was only captured for uncatchable errors (inside if !err.is_catchable()) and for catchable errors that reached handle_throw(). Errors intercepted by internal handlers (e.g. the module async wrapper) were stored as pending_exception before either path could capture a backtrace. When these errors were later converted to JsValue for promise rejection, the backtrace was lost entirely.

This PR:

1. Captures the backtrace unconditionally in handle_error, before the exception handler check
2. Stores the backtrace on the Error object during into_opaque (for both Repr::Native and Repr::Opaque errors) so it survives conversion to JsValue
3. Recovers the backtrace in from_opaque by reading it back from the Error object

This approach was suggested by @jedel1043 in #4607 — extending the Error type with a backtrace field rather than injecting positions, which would be incorrect with nested try blocks.

Test plan

• test_call_error_preserves_backtrace — calls undefined() at module top level, verifies backtrace with exact string match
• test_nested_call_error_preserves_backtrace — nested foo → baz → import.meta.non_existent(), verifies full multi-frame backtrace
• test_explicit_throw_preserves_backtrace — explicit throw new Error("test") inside a function, verifies backtrace frames are present

Made with Cursor

[AlbertMarashi]

@jedel1043 requesting review

[github-actions]

Test262 conformance changes

Test result	main count	PR count	difference
Total	52,862	52,862	0
Passed	49,471	49,471	0
Ignored	2,249	2,249	0
Failed	1,142	1,142	0
Panics	0	0	0
Conformance	93.59%	93.59%	0.00%

[randomboi404]

I recommend you to fix the issues in your code and make it pass the CI otherwise it can't be merged.