Qualify auto-generated sub-orchestration IDs across continue_as_new generations

[Copilot]

Summary

Auto-generated sub-orchestration IDs were based only on event_id; after continue_as_new, event_id restarts and children could reuse prior-generation IDs, stalling parent waits against already-completed child instances.
This change makes auto IDs generation-qualified for execution N>1 while keeping execution 1 format unchanged for compatibility.

• Runtime ID derivation

  • Updated replay action resolution so pending sub-orchestration IDs become:
    • execution 1: sub::{event_id}
    • execution N>1: sub::{execution_id}.{event_id}
  • Keeps replay determinism and removes cross-generation collisions.

• Replay-engine regression coverage

  • Added a focused test validating execution 2 emits sub::2.{event_id} for auto-generated child IDs.

• Docs

  • Updated sub-orchestration behavior docs and replay-engine test spec to reflect generation-qualified auto IDs.

let final_instance = if instance.starts_with(SUB_ORCH_PENDING_PREFIX) {
    if execution_id == INITIAL_EXECUTION_ID {
        format!("{SUB_ORCH_AUTO_PREFIX}{event_id}")
    } else {
        format!("{SUB_ORCH_AUTO_PREFIX}{execution_id}.{event_id}")
    }
} else {
    instance
};

Checklist

• Tests added/updated
• Regular tests pass locally (cargo nt)
• Doc tests pass locally (cargo test --doc --all-features)
• Docs updated
  • Existing docs updated where behavior changed
  • New doc added under docs/ if introducing a new area
  • Linked from docs/README.md

Links / Design notes

• Code pointers:
  • src/runtime/replay_engine.rs (auto ID resolution now includes execution_id)
  • tests/replay_engine/sub_orchestration.rs (new generation-specific ID test)
  • docs/sub-orchestrations.md
  • docs/replay-engine-test-spec.md