[Calls] Execute bounded module initialization consistently in symbolic and native preparation #487

Part of #483.

Observed boundary

Priority: P0. 93 module-initialization first rejections; prepared Perm blocked later.

93 is the number of first failures, not evidence that all 93 are harmless literal initializers. JacoDB #378 is a completed frontend foundation, not completion of Calls module replay/setup.

Implementation scope

Replace the dependency-initializer whitelist with a checked supported initialization contract shared by source preparation, frontend lowering, symbolic state setup and fresh native replay. Start with pure scalar/array/object literals and source-attested constant dependencies, then define supported deterministic effects. Preserve evaluation order, module-instance sharing within a scenario and reset between samples. Keep unsupported environment-dependent effects explicit.

Acceptance criteria

• Perm.ts with const zeroCache: number[][] = [] reaches preparation and a real scenario with matching initialized state.
• Arrays/objects are constructed with original identity/aliasing; accepting source syntax alone is insufficient.
• Supported initializers run once per module instance, in the same order as native execution; errors and unsupported effects are retained.
• The fixed target-class universe excludes unrelated module statements, while any target-class lifecycle statements use the separately checked lifecycle contract.

Existing work to reuse or coordinate

• [TS] Execute export-default expressions during module initialization jacodb#378

Evidence and delivery rules

Use the frozen v7 200-class ledger and the source-linked reproducer; do not overwrite sealed experiment output. Reuse existing loaders, domain projection, source instrumentation, candidate/replay contracts and bounded observations. Keep original class members and independently frozen domain policies. Run focused source/native checks for the changed contract and leave broad suites to CI; unsupported and newly exposed later gates remain recorded.

Local evidence: iccq2026-paper/experiments/class-coverage-round-robin-20261002/real-class-campaign-20261014/{original/selection-ledger-v7.json,V7_INFRASTRUCTURE_FRONTIER.md} and real-class-machine-diagnostic-20261014/revised-protocol-v8/README.md.