refactor(noq-proto)!: Remove identity_hash from public API

[Frando]

Description

Based on #643

Removes identity_hash from the public API of noq_proto.

Breaking Changes

• changed: noq_proto::PathId no longer implementsidentity_hash::IdentityHashable

Notes & open questions

Not sure if it's worth it?

Change checklist

• Self-review.
• All breaking changes documented.

[github-actions]

Performance Comparison Report

e87b1e3a859b249f44202ccc9853470dc361263b - artifacts

No results available

---

1e533fb14fc82166fb2e0c48a34369eed42773b2 - artifacts

No results available

---

33ed91ca2a087c8b359d3ceb396c1c16f25728f1 - artifacts

Raw Benchmarks (localhost)

Scenario	noq	upstream	Delta	CPU (avg/max)
large-single	5916.4 Mbps	7845.8 Mbps	-24.6%	97.7% / 98.9%
medium-concurrent	5508.6 Mbps	7736.5 Mbps	-28.8%	96.3% / 97.8%
medium-single	4122.3 Mbps	4744.2 Mbps	-13.1%	95.9% / 98.3%
small-concurrent	3870.3 Mbps	5317.5 Mbps	-27.2%	97.4% / 99.6%
small-single	3576.0 Mbps	4857.6 Mbps	-26.4%	96.4% / 98.6%

Netsim Benchmarks (network simulation)

Condition	noq	upstream	Delta
ideal	3079.3 Mbps	3911.2 Mbps	-21.3%
lan	782.5 Mbps	810.3 Mbps	-3.4%
lossy	69.8 Mbps	55.9 Mbps	+25.0%
wan	83.8 Mbps	83.8 Mbps	~0%

Summary

noq is 23.6% slower on average

---

86939ff6747963f1d067111ee85c491e7e52f356 - artifacts

Raw Benchmarks (localhost)

Scenario	noq	upstream	Delta	CPU (avg/max)
large-single	5808.2 Mbps	7962.8 Mbps	-27.1%	97.0% / 98.4%
medium-concurrent	5822.0 Mbps	7861.9 Mbps	-25.9%	96.4% / 97.9%
medium-single	3545.6 Mbps	4561.7 Mbps	-22.3%	95.8% / 98.3%
small-concurrent	3893.9 Mbps	5462.1 Mbps	-28.7%	97.6% / 99.7%
small-single	3639.9 Mbps	4721.1 Mbps	-22.9%	96.4% / 98.6%

Netsim Benchmarks (network simulation)

Condition	noq	upstream	Delta
ideal	3100.8 Mbps	4079.1 Mbps	-24.0%
lan	782.4 Mbps	810.4 Mbps	-3.5%
lossy	55.9 Mbps	69.8 Mbps	-20.0%
wan	83.8 Mbps	83.8 Mbps	~0%

Summary

noq is 24.9% slower on average

[github-actions]

Documentation for this PR has been generated and is available at: https://n0-computer.github.io/noq/pr/646/docs/noq/

Last updated: 2026-05-22T07:21:57Z

[matheus23]

I'm not sure. identity-hash has had only exactly one release that is three years old.
It has zero dependencies and otherwise only depends on stable rust std traits.

It only uses core imports (not std) and is no_std-compatible.

I really see no reason to not depend on it in this niche way. Even if we end up removing IntMap, we can keep the identity hash impl for PathId without almost any added cost.

[flub]

Meh, I'm really conflicted. I'm tempted to lean to +1. Mostly I think rust isn't great here. OTOH the newtype is essentially free. But I can see this gets annoying if we have to do it for lots of reason.

I think I'd rather do this now so we have it as a pattern and then see how it plays out and if we want to expose it anyway.

[matheus23]

As @flub says, let's see how it plays out and play it safe.