at-object-delete — object delete (Node proxy)
Status: measured-inconclusive (node-proxy)
Claim
On this Node/V8 microbench, delete on a null-prototype object vs Map.delete was ~1× — inconclusive as a speed gate.
How measured
-
Harness:
examples/bench/evidence_at001_at003.js -
runtime: node-proxy
-
Raw lines:
at-object-delete.jsonl
Runtime (this evidence file)
| Field | Value |
|---|---|
| When | 2026-10-22 Asia/Shanghai (JSONL ts is UTC) |
| runtime | node-proxy |
| node | v20.19.2 |
| v8 | 11.3.244.8-node.38 |
Median before (object_delete) | ~1.341 ms |
Median after (map_delete_free) | ~1.323 ms |
| ratio before/after | ~1.01× |
Lint mapping
arktsperf-lint at-object-delete: severity stays package default (see evidence-policy). JSONL is Node proxy, not device.
Limits
- Do not treat as proof of speedup. Shape/IC effects need a different harness.
- Default stays off.
Landed 2026-10-22
Deep-dive: original harness timed fill + delete together (fill dominates → ~1×). Isolated delete-only on Node is modest (~1.7× obj vs Map). Still node-proxy; keep off. No speed escalate.