Duplicated crates: Tracking Issue

[macladson]

Description

This is a tracking issue for crates in Lighthouse where we have to compile multiple versions. I'll try to keep this updated as things change.

Worst 10 duplicate crates by compile time

Using cargo build --release --timings I extracted the following data:

Rank	Crate	Version	Compile Time	Wasted Time	Dependants
1	rustls	v0.22.4	12.4s	12.4s	tokio-rustls
		v0.23.35			futures-rustls, hyper-rustls, libp2p-quic, libp2p-tls, quinn, quinn-proto, reqwest, tokio-rustls
2	h2	v0.3.27	7.4s	7.4s	hyper
		v0.4.12			hyper
3	hyper	v0.14.32	6.9s	6.9s	warp
		v1.8.1			beacon_node, hyper-rustls, hyper-timeout, hyper-tls, hyper-util, igd-next, mockito, reqwest, tonic, validator_client
4	darling_core	v0.13.4	2.0s	4.0s	darling, darling_macro
		v0.20.11			darling, darling_macro
		v0.21.3			darling, darling_macro
5	syn	v1.0.109	3.8s	3.8s	compare_fields_derive, context_deserialize_derive, darling_core, darling_macro, match-lookup, metastruct_macro, neli-proc-macros, strum_macros, test_random_derive
		v2.0.110			alloy-rlp-derive, alloy-sol-macro, async-trait, auto_impl, clap_derive, darling_core, darling_macro, derive_more, ethereum_ssz_derive, futures-macro, libp2p-swarm-derive, prost-derive, serde_derive, strum_macros, superstruct, thiserror-impl, tokio-macros, tracing-attributes, tree_hash_derive, (+30 more)
6	rustix	v0.36.17	2.7s	2.7s	procfs
		v1.1.2			tempfile, terminal_size
7	itertools	v0.10.5	1.3s	2.6s	beacon_chain, beacon_node_fallback, beacon_processor, lighthouse_network, metastruct_macro, network, operation_pool, state_processing, store, types
		v0.13.0			ethereum_ssz, milhouse, superstruct
		v0.14.0			alloy-rpc-types-eth, compare_fields, eip4844, ekzg-single-open, prost-derive, ssz_types
8	rustls-webpki	v0.102.8	2.0s	2.0s	rustls
		v0.103.8			libp2p-tls, rustls
9	strum_macros	v0.24.3	1.4s	1.4s	strum
		v0.27.2			strum
10	http	v0.2.12	1.4s	1.4s	h2, headers, headers-core, http-body, hyper, warp
		v1.3.1			alloy-json-rpc, attohttpc, h2, http-body, hyper, hyper-rustls, hyper-util, igd-next, mockito, reqwest, tonic, tower-http, (+3 more)

[macladson]

Removing warp would yield very large gains here as it is soley responsible for the duplicated rustls, h2 and hyper crates.
Upgrading darling and syn might also be possible since many of the older v1.0 syn versions come from our own crates.
itertools also looks like another good option as both older versions come solely from crates under our control.