Preserve JSON input object order

[dnagoda]

📚 Stack (bottom → top)

1. Add batch mode for JSONL function runs #595 Add batch mode for processing multiple inputs via JSONL
2. Preserve JSON input object order #590 Preserve JSON input object order
3. Refine batch failure handling #591 Refine batch failure handling
4. Reuse the compiled provider across runs #592 Reuse the compiled provider across runs
5. Compute humanized bytes lazily #593 Compute humanized bytes lazily
6. Add minimal batch output with --batch-full-output opt-in #594 Add minimal batch output with --batch-full-output opt-in

Stacked PRs — review bottom-up.

---

Why

function-runner parses each JSON input into serde_json::Value and
reserializes it to bytes before handing it to the Wasm module. With
serde_json's default map, object keys are reordered lexicographically. JS
functions that rely on Object.keys() ordering can then behave differently under
replay than in production, breaking parity.

What

• Enable serde_json's preserve_order feature so input object key order is
  retained end-to-end through the runner.
• Add a regression test covering nested, metafield-like message objects to lock
  in the ordering guarantee.

This brings function-rerunner inputs closer to real production bytes.

Testing

• cargo test json_input_preserves_object_key_order_in_raw_bytes
• cargo build --release
• cargo test
• Re-ran the 17 discount localized-message mismatch rows with the fixed release
  runner → full semantic matches.