perf: zero-alloc Jid dedup and session lock helper by jlucaso1 · Pull Request #500 · oxidezap/whatsapp-rust

jlucaso1 · 2026-04-07T15:41:26Z

Summary

Zero-alloc Jid dedup

Replace HashSet + retain + .clone() with in-place sort_unstable_by + dedup_by
Eliminates N String allocations per group send (user + server cloned into HashSet)
New helpers in wacore::types::jid: sort_dedup_by_user() and sort_dedup_by_device()
Order change is safe: downstream resolve_devices is a batch query, order-independent

Session lock helper

Extract session_mutexes_for() on Client, replacing duplicated 8-line lock acquisition pattern in send.rs and signal.rs

Misc

Remove unnecessary to_jid.clone() in group stanza builder (last use of owned value)

Test plan

cargo fmt --all
cargo clippy --all --tests — zero warnings
cargo test --workspace --exclude e2e-tests — 1119 tests pass

Summary by CodeRabbit

Refactor
- Streamlined session lock acquisition mechanism for device encryption handling, reducing code complexity.
- Consolidated deduplication utilities for improved maintainability and consistency across session management operations.

- Add sort_dedup_by_user() and sort_dedup_by_device() to wacore JidExt, replacing HashSet+retain+clone with in-place sort+dedup (0 allocations vs N String clones per group send) - Add session_mutexes_for() helper on Client, deduplicating the 8-line multi-lock acquisition pattern from send.rs and signal.rs - Remove unnecessary to_jid.clone() in group stanza builder (last use)

coderabbitai · 2026-04-07T15:41:40Z

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: Organization UI

Review profile: ASSERTIVE

Plan: Pro

Run ID: 3412eaed-474e-44f1-a875-0c051f2065df

📥 Commits

Reviewing files that changed from the base of the PR and between fe12a54 and 3861a55.

📒 Files selected for processing (4)

src/features/signal.rs
src/send.rs
wacore/src/send.rs
wacore/src/types/jid.rs

📝 Walkthrough

Walkthrough

The PR refactors session lock and deduplication logic across multiple modules by introducing helper methods: session_mutexes_for centralizes per-device session mutex collection, while sort_dedup_by_user and sort_dedup_by_device replace ad-hoc deduplication patterns. This simplifies repetitive inline implementations previously scattered across the codebase.

Changes

Cohort / File(s)	Summary
Session Lock Acquisition Refactoring `src/features/signal.rs`, `src/send.rs`	Replaced manual per-key session lock loops with delegated calls to new `session_mutexes_for` helper method; added `pub(crate) async fn session_mutexes_for` to `Client` that collects and returns mutexes for a slice of device-encryption JIDs in order.
JID Utility Functions `wacore/src/types/jid.rs`	Added two new exported sorting-and-deduplication utilities: `sort_dedup_by_user` (dedup by user/server pair) and `sort_dedup_by_device` (dedup by user/server/agent/device tuple); updated doc comments for `write_protocol_address_to` and `cmp_for_lock_order`.
Deduplication Pattern Consolidation `wacore/src/send.rs`	Replaced ad-hoc `HashSet`-based deduplication of JID lists with calls to new utility functions `sort_dedup_by_user` and `sort_dedup_by_device`; removed unnecessary clone in stanza node construction.

Estimated code review effort

🎯 3 (Moderate) | ⏱️ ~22 minutes

Possibly related PRs

perf: single-allocation to_protocol_address_string() for session lock keys #380: Refactors the same session-lock key derivation in the send path by replacing per-key string construction with an efficient single-step approach.
feat: add Signal protocol feature API and consolidate internal helpers #474: Introduces centralized Client::session_lock_for and updates session-locking logic across the codebase, overlapping with this PR's session mutex acquisition refactoring.
fix: per-device session locks in DM send path (#462) #463: Directly modifies the DM send path's per-device session lock acquisition, introducing helpers to collect per-device lock mutexes and replace inline per-key locking loops.

Poem

🐰 A helper hops in, gathering locks with care,
No more string-building scattered here and there,
Sort and dedup now dance as one refrain,
Lock-key collection flows through a cleaner lane! 🔐✨

🚥 Pre-merge checks | ✅ 3

✅ Passed checks (3 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title 'perf: zero-alloc Jid dedup and session lock helper' directly summarizes the main changes: eliminating allocations in Jid deduplication and introducing a session lock helper method.
Docstring Coverage	✅ Passed	Docstring coverage is 100.00% which is sufficient. The required threshold is 80.00%.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

✨ Finishing Touches

📝 Generate docstrings

Create stacked PR
Commit on current branch

🧪 Generate unit tests (beta)

Create PR with unit tests
Commit unit tests in branch perf/zero-alloc-dedup-and-session-lock-helper

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

jlucaso1 · 2026-04-07T15:42:36Z

@claude review

claude · 2026-04-07T15:42:59Z

Claude finished @jlucaso1's task in 52s —— View job

Code Review: perf: zero-alloc Jid dedup and session lock helper

Read changed files and PR diff
Review zero-alloc Jid dedup helpers
Review session lock helper extraction
Review misc changes
Post review feedback

Clean PR — good dedup optimization and nice extraction of the duplicated lock pattern. A few observations:

`sort_dedup_by_user` / `sort_dedup_by_device` — looks correct

The sort_unstable_by + dedup_by approach is a solid zero-alloc replacement for the HashSet + retain + .clone() pattern. The behavioral change from unordered-retain to sorted-dedup is safe since resolve_devices is a batch query (order-independent), as noted in the PR description.

One nit: sort_dedup_by_user (wacore/src/types/jid.rs:35) sorts by user then server, but cmp_for_lock_order (line 27) sorts by server then user. This is fine since they serve different purposes (dedup vs lock ordering), but worth being aware of if anyone later tries to unify them.

`dedup_by` argument order

dedup_by passes (a, b) where a is the later element and b is the earlier one (consecutive duplicates, a is removed). The equality predicates here are symmetric so this is correct, but it's a common footgun worth noting — if these ever became asymmetric (e.g., merging fields from a into b), the argument order matters.

`session_mutexes_for` — good extraction

The helper at src/send.rs:1496-1507 cleanly deduplicates the 8-line pattern from both send.rs and signal.rs. The buf reuse across iterations avoids per-iteration allocation, which is nice.

Minor: The method assumes jids are already sorted (from build_session_lock_keys). This implicit precondition could silently break if someone calls session_mutexes_for with unsorted jids. Consider either:

A debug assertion: debug_assert!(jids.windows(2).all(|w| cmp_for_lock_order(&w[0], &w[1]).is_le()))
Or documenting the precondition in the doc comment (e.g., /// Jids must be pre-sorted via build_session_lock_keys``)

This is low-risk since both call sites already go through build_session_lock_keys, but a debug_assert would be a nice safety net.

`to_jid.clone()` removal

wacore/src/send.rs:1210 — removing the unnecessary .clone() on the last use of to_jid is correct. Good catch.

Unused `HashSet` import?

wacore/src/send.rs:15 still imports HashSet, which is still used elsewhere in the file (lines 1245, 1246, 1323), so the import is correctly retained.

Overall

LGTM — the changes are well-scoped, the allocation savings are real (N clones per group send eliminated), and the session lock helper reduces duplication nicely. The only suggestion is the optional debug_assert for the sort precondition on session_mutexes_for.

github-actions · 2026-04-07T15:46:32Z

Bencher Report

Branch	perf/zero-alloc-dedup-and-session-lock-helper
Testbed	ubuntu-latest

🚨 1 Alert

Benchmark	Measure Units	View	Benchmark Result (Result Δ%)	Upper Boundary (Limit %)
libsignal_benchmark::session_optimization_group::bench_decrypt_with_previous_session previous_session:setup_with_archived_sessions()	Instructions instructions x 1e3	📈 plot 🚷 threshold 🚨 alert (🔔)	47.03 x 1e3 (+7.45%) Baseline: 43.77 x 1e3	45.96 x 1e3 (102.33%)

Click to view all benchmark results

Benchmark	Instructions	Benchmark Result instructions (Result Δ%)	Upper Boundary instructions (Limit %)
binary_benchmark::attr_parser_group::bench_attr_parser attr_lookup:setup_attr_marshaled()	📈 view plot 🚷 view threshold	6,199.00 (-2.56%) Baseline: 6,361.59	6,679.67 (92.80%)
binary_benchmark::child_iteration_group::bench_get_children_by_tag	📈 view plot 🚷 view threshold	523,708.00 (-23.75%) Baseline: 686,844.88	721,187.13 (72.62%)
binary_benchmark::jid_optimization_group::bench_jid_to_owned_access jid_access:setup_jid_heavy_marshaled()	📈 view plot 🚷 view threshold	20,910.00 (-4.59%) Baseline: 21,915.47	23,011.24 (90.87%)
binary_benchmark::marshal_group::bench_marshal_allocating	📈 view plot 🚷 view threshold	98,703.00 (-10.91%) Baseline: 110,787.77	116,327.15 (84.85%)
binary_benchmark::marshal_group::bench_marshal_auto_allocating	📈 view plot 🚷 view threshold	98,731.00 (-7.91%) Baseline: 107,212.34	112,572.96 (87.70%)
binary_benchmark::marshal_group::bench_marshal_auto_huge_bytes_allocating	📈 view plot 🚷 view threshold	533,017.00 (-0.08%) Baseline: 533,418.00	560,088.90 (95.17%)
binary_benchmark::marshal_group::bench_marshal_auto_long_string	📈 view plot 🚷 view threshold	15,955.00 (-3.36%) Baseline: 16,509.80	17,335.29 (92.04%)
binary_benchmark::marshal_group::bench_marshal_auto_many_children_allocating	📈 view plot 🚷 view threshold	14,813,622.00 (-5.97%) Baseline: 15,754,914.66	16,542,660.39 (89.55%)
binary_benchmark::marshal_group::bench_marshal_exact_allocating	📈 view plot 🚷 view threshold	118,631.00 (-17.17%) Baseline: 143,230.65	150,392.18 (78.88%)
binary_benchmark::marshal_group::bench_marshal_exact_huge_bytes_allocating	📈 view plot 🚷 view threshold	534,447.00 (-0.07%) Baseline: 534,840.60	561,582.63 (95.17%)
binary_benchmark::marshal_group::bench_marshal_exact_long_string	📈 view plot 🚷 view threshold	18,004.00 (-3.00%) Baseline: 18,560.67	19,488.70 (92.38%)
binary_benchmark::marshal_group::bench_marshal_exact_many_children_allocating	📈 view plot 🚷 view threshold	28,200,726.00 (-17.88%) Baseline: 34,342,321.52	36,059,437.60 (78.21%)
binary_benchmark::marshal_group::bench_marshal_huge_bytes_allocating	📈 view plot 🚷 view threshold	533,456.00 (-0.08%) Baseline: 533,857.00	560,549.85 (95.17%)
binary_benchmark::marshal_group::bench_marshal_long_string	📈 view plot 🚷 view threshold	15,928.00 (-5.76%) Baseline: 16,901.34	17,746.41 (89.75%)
binary_benchmark::marshal_group::bench_marshal_many_children_allocating	📈 view plot 🚷 view threshold	14,815,122.00 (-5.97%) Baseline: 15,756,097.29	16,543,902.16 (89.55%)
binary_benchmark::marshal_group::bench_marshal_reusing_buffer	📈 view plot 🚷 view threshold	108,446.00 (-8.71%) Baseline: 118,788.14	124,727.55 (86.95%)
binary_benchmark::marshal_group::bench_marshal_reusing_buffer_vec_writer	📈 view plot 🚷 view threshold	98,803.00 (-7.91%) Baseline: 107,284.34	112,648.56 (87.71%)
binary_benchmark::roundtrip_group::bench_roundtrip large:setup_large_marshaled()	📈 view plot 🚷 view threshold	91,558.00 (-3.58%) Baseline: 94,955.65	99,703.43 (91.83%)
binary_benchmark::roundtrip_group::bench_roundtrip small:setup_small_marshaled()	📈 view plot 🚷 view threshold	7,431.00 (-1.84%) Baseline: 7,570.00	7,948.50 (93.49%)
binary_benchmark::roundtrip_group::bench_roundtrip_auto large:setup_large_marshaled()	📈 view plot 🚷 view threshold	91,589.00 (-0.85%) Baseline: 92,377.83	96,996.72 (94.42%)
binary_benchmark::roundtrip_group::bench_roundtrip_auto small:setup_small_marshaled()	📈 view plot 🚷 view threshold	7,454.00 (+0.98%) Baseline: 7,381.93	7,751.03 (96.17%)
binary_benchmark::roundtrip_group::bench_roundtrip_exact large:setup_large_marshaled()	📈 view plot 🚷 view threshold	107,134.00 (-0.94%) Baseline: 108,151.35	113,558.92 (94.34%)
binary_benchmark::roundtrip_group::bench_roundtrip_exact small:setup_small_marshaled()	📈 view plot 🚷 view threshold	8,966.00 (+0.81%) Baseline: 8,893.93	9,338.63 (96.01%)
binary_benchmark::unmarshal_group::bench_unmarshal large:setup_large_marshaled()	📈 view plot 🚷 view threshold	41,989.00 (-5.57%) Baseline: 44,466.08	46,689.38 (89.93%)
binary_benchmark::unmarshal_group::bench_unmarshal small:setup_small_marshaled()	📈 view plot 🚷 view threshold	2,716.00 (-1.47%) Baseline: 2,756.54	2,894.37 (93.84%)
binary_benchmark::unpack_group::bench_unpack_compressed	📈 view plot 🚷 view threshold	556,090.00 (-0.01%) Baseline: 556,137.16	583,944.01 (95.23%)
binary_benchmark::unpack_group::bench_unpack_uncompressed	📈 view plot 🚷 view threshold	773.00 (+0.10%) Baseline: 772.20	810.81 (95.34%)
libsignal_benchmark::conversation_group::bench_full_dm_conversation full:setup_conversation_data()	📈 view plot 🚷 view threshold	27,434,615.00 (-0.93%) Baseline: 27,693,099.09	29,077,754.04 (94.35%)
libsignal_benchmark::dm_group::bench_dm_decrypt_first_message decrypt_prekey:setup_dm_with_first_message()	📈 view plot 🚷 view threshold	5,512,366.00 (-0.61%) Baseline: 5,546,120.80	5,823,426.84 (94.66%)
libsignal_benchmark::dm_group::bench_dm_encrypt_first_message first_msg:setup_dm_session()	📈 view plot 🚷 view threshold	162,318.00 (-8.28%) Baseline: 176,970.04	185,818.54 (87.35%)
libsignal_benchmark::dm_group::bench_dm_encrypt_subsequent_message subsequent:setup_established_dm_session()	📈 view plot 🚷 view threshold	163,047.00 (-8.26%) Baseline: 177,727.87	186,614.26 (87.37%)
libsignal_benchmark::dm_group::bench_dm_session_establishment setup:setup_dm_users()	📈 view plot 🚷 view threshold	17,232,286.00 (-0.27%) Baseline: 17,279,475.04	18,143,448.79 (94.98%)
libsignal_benchmark::group_messaging_group::bench_group_create_distribution_message create:setup_group_sender()	📈 view plot 🚷 view threshold	298,493.00 (+0.48%) Baseline: 297,055.10	311,907.85 (95.70%)
libsignal_benchmark::group_messaging_group::bench_group_decrypt_message decrypt:setup_group_with_encrypted_message()	📈 view plot 🚷 view threshold	12,622,794.00 (+0.22%) Baseline: 12,595,526.63	13,225,302.96 (95.44%)
libsignal_benchmark::group_messaging_group::bench_group_encrypt_message encrypt:setup_group_with_distribution()	📈 view plot 🚷 view threshold	713,228.00 (-0.56%) Baseline: 717,228.94	753,090.38 (94.71%)
libsignal_benchmark::session_optimization_group::bench_decrypt_with_previous_session previous_session:setup_with_archived_sessions()	📈 view plot 🚷 view threshold 🚨 view alert (🔔)	47,033.00 (+7.45%) Baseline: 43,773.92	45,962.61 (102.33%)
libsignal_benchmark::session_optimization_group::bench_message_key_eviction eviction:setup_message_key_eviction()	📈 view plot 🚷 view threshold	14,247,117.00 (-8.38%) Baseline: 15,550,867.85	16,328,411.24 (87.25%)
libsignal_benchmark::session_optimization_group::bench_out_of_order_decryption out_of_order:setup_out_of_order_messages()	📈 view plot 🚷 view threshold	5,116,228.00 (-6.34%) Baseline: 5,462,485.05	5,735,609.30 (89.20%)
libsignal_benchmark::session_optimization_group::bench_promote_matching_session promote:setup_promote_matching_session()	📈 view plot 🚷 view threshold	298,844.00 (-59.55%) Baseline: 738,709.26	775,644.72 (38.53%)
libsignal_benchmark::signature_group::bench_key_generation keygen	📈 view plot 🚷 view threshold	2,830,495.00 (+0.15%) Baseline: 2,826,196.61	2,967,506.44 (95.38%)
libsignal_benchmark::signature_group::bench_signature_creation sign:setup_keypair_with_message()	📈 view plot 🚷 view threshold	3,467,011.00 (-0.05%) Baseline: 3,468,670.99	3,642,104.54 (95.19%)
libsignal_benchmark::signature_group::bench_signature_verification verify:setup_keypair_with_message()	📈 view plot 🚷 view threshold	126,598,813.00 (+0.98%) Baseline: 125,370,956.33	131,639,504.15 (96.17%)
reporting_token_benchmark::content_extraction_group::bench_content_extraction extended:setup_extended_message()	📈 view plot 🚷 view threshold	11,851.00 (+0.07%) Baseline: 11,842.39	12,434.51 (95.31%)
reporting_token_benchmark::content_extraction_group::bench_content_extraction simple:setup_simple_message()	📈 view plot 🚷 view threshold	3,879.00 (+0.84%) Baseline: 3,846.88	4,039.22 (96.03%)
reporting_token_benchmark::full_generation_group::bench_full_token_generation extended:setup_full_gen_extended()	📈 view plot 🚷 view threshold	77,193.00 (-11.84%) Baseline: 87,556.11	91,933.91 (83.97%)
reporting_token_benchmark::full_generation_group::bench_full_token_generation simple:setup_full_gen_simple()	📈 view plot 🚷 view threshold	69,103.00 (-13.17%) Baseline: 79,581.49	83,560.57 (82.70%)
reporting_token_benchmark::key_derivation_group::bench_key_derivation	📈 view plot 🚷 view threshold	43,398.00 (-14.53%) Baseline: 50,773.33	53,312.00 (81.40%)
reporting_token_benchmark::message_encoding_group::bench_message_encoding extended:setup_extended_message()	📈 view plot 🚷 view threshold	5,939.00 (+2.33%) Baseline: 5,803.66	6,093.84 (97.46%)
reporting_token_benchmark::message_encoding_group::bench_message_encoding simple:setup_simple_message()	📈 view plot 🚷 view threshold	2,214.00 (+3.17%) Baseline: 2,145.90	2,253.19 (98.26%)
reporting_token_benchmark::token_calculation_group::bench_token_calculation	📈 view plot 🚷 view threshold	19,365.00 (-11.58%) Baseline: 21,900.26	22,995.27 (84.21%)

🐰 View full continuous benchmarking report in Bencher

jlucaso1 merged commit b56be2e into main Apr 7, 2026
8 checks passed

jlucaso1 deleted the perf/zero-alloc-dedup-and-session-lock-helper branch April 7, 2026 15:50

This was referenced Apr 12, 2026

perf: single-buffer ProtocolAddress + reusable hot-loop address construction #518

Merged

perf: 11 allocation trims from hot-path audit #570

Merged

coderabbitai Bot mentioned this pull request Jul 29, 2026

perf(jid): skip the agent normalisation when the raw agents already match #1183

Merged

Navigation Menu

Uh oh!

Uh oh!

perf: zero-alloc Jid dedup and session lock helper - #500

perf: zero-alloc Jid dedup and session lock helper#500
jlucaso1 merged 1 commit into
mainfrom
perf/zero-alloc-dedup-and-session-lock-helper

jlucaso1 commented Apr 7, 2026 •

edited by coderabbitai Bot

Loading

Uh oh!

coderabbitai Bot commented Apr 7, 2026 •

edited

Loading

Walkthrough

Changes

Estimated code review effort

Possibly related PRs

Poem

Uh oh!

jlucaso1 commented Apr 7, 2026

Uh oh!

claude Bot commented Apr 7, 2026 •

edited

Loading

Uh oh!

github-actions Bot commented Apr 7, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

1 participant

Uh oh!

Uh oh!

Conversation

jlucaso1 commented Apr 7, 2026 • edited by coderabbitai Bot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Summary

Zero-alloc Jid dedup

Session lock helper

Misc

Test plan

Summary by CodeRabbit

Uh oh!

coderabbitai Bot commented Apr 7, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Walkthrough

Changes

Estimated code review effort

Possibly related PRs

Poem

Uh oh!

jlucaso1 commented Apr 7, 2026

Uh oh!

claude Bot commented Apr 7, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Code Review: perf: zero-alloc Jid dedup and session lock helper

sort_dedup_by_user / sort_dedup_by_device — looks correct

dedup_by argument order

session_mutexes_for — good extraction

to_jid.clone() removal

Unused HashSet import?

Overall

Uh oh!

github-actions Bot commented Apr 7, 2026

Bencher Report

🚨 1 Alert

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

1 participant

jlucaso1 commented Apr 7, 2026 •

edited by coderabbitai Bot

Loading

coderabbitai Bot commented Apr 7, 2026 •

edited

Loading

claude Bot commented Apr 7, 2026 •

edited

Loading

`sort_dedup_by_user` / `sort_dedup_by_device` — looks correct

`dedup_by` argument order

`session_mutexes_for` — good extraction

`to_jid.clone()` removal

Unused `HashSet` import?