Update Rust crate rand to 0.10

[renovate]

This PR contains the following updates:

Package	Type	Update	Change
rand (source)	dependencies	minor	0.8 → 0.10
rand (source)	dev-dependencies	minor	0.9 → 0.10
rand (source)	dependencies	minor	0.9 → 0.10

---

Release Notes

rust-random/rand (rand)

v0.10.2

Compare Source

Fixes

• Fix possible memory safety violation due to deserialization of UniformChar from bad source (#​1790)

Changes

• Document required output order of fn partial_shuffle and apply #[must_use] (#​1769)
• Avoid usage of unsafe in contexts where non-local memory corruption could invalidate contract (#​1791)

v0.10.1

Compare Source

This release includes a fix for a soundness bug; see #​1763.

Changes

• Document panic behavior of make_rng and add #[track_caller] (#​1761)
• Deprecate feature log (#​1763)

v0.10.0

Compare Source

Changes

• The dependency on rand_chacha has been replaced with a dependency on chacha20. This changes the implementation behind StdRng, but the output remains the same. There may be some API breakage when using the ChaCha-types directly as these are now the ones in chacha20 instead of rand_chacha (#​1642).
• Rename fns IndexedRandom::choose_multiple -> sample, choose_multiple_array -> sample_array, choose_multiple_weighted -> sample_weighted, struct SliceChooseIter -> IndexedSamples and fns IteratorRandom::choose_multiple -> sample, choose_multiple_fill -> sample_fill (#​1632)
• Use Edition 2024 and MSRV 1.85 (#​1653)
• Let Fill be implemented for element types, not sliceable types (#​1652)
• Fix OsError::raw_os_error on UEFI targets by returning Option<usize> (#​1665)
• Replace fn TryRngCore::read_adapter(..) -> RngReadAdapter with simpler struct RngReader (#​1669)
• Remove fns SeedableRng::from_os_rng, try_from_os_rng (#​1674)
• Remove Clone support for StdRng, ReseedingRng (#​1677)
• Use postcard instead of bincode to test the serde feature (#​1693)
• Avoid excessive allocation in IteratorRandom::sample when amount is much larger than iterator size (#​1695)
• Rename os_rng -> sys_rng, OsRng -> SysRng, OsError -> SysError (#​1697)
• Rename Rng -> RngExt as upstream rand_core has renamed RngCore -> Rng (#​1717)

Additions

• Add fns IndexedRandom::choose_iter, choose_weighted_iter (#​1632)
• Pub export Xoshiro128PlusPlus, Xoshiro256PlusPlus prngs (#​1649)
• Pub export ChaCha8Rng, ChaCha12Rng, ChaCha20Rng behind chacha feature (#​1659)
• Fn rand::make_rng() -> R where R: SeedableRng (#​1734)

Removals

• Removed ReseedingRng (#​1722)
• Removed unused feature "nightly" (#​1732)
• Removed feature small_rng (#​1732)

v0.9.5

Compare Source

What's Changed

• rand v0.9.5: backport #​1790 by @​dhardy in #​1803

Full Changelog: rust-random/rand@0.9.4...0.9.5

v0.9.4

Compare Source

Fixes

• Fix doc build (#​1766)

Full Changelog: rust-random/rand@0.9.3...0.9.4

v0.9.3

Compare Source

v0.9.2

Compare Source

Deprecated

• Deprecate rand::rngs::mock module and StepRng generator (#​1634)

Additions

• Enable WeightedIndex<usize> (de)serialization (#​1646)

v0.9.1

Compare Source

Security and unsafe

• Revise "not a crypto library" policy again (#​1565)
• Remove zerocopy dependency from rand (#​1579)

Fixes

• Fix feature simd_support for recent nightly rust (#​1586)

Changes

• Allow fn rand::seq::index::sample_weighted and fn IndexedRandom::choose_multiple_weighted to return fewer than amount results (#​1623), reverting an undocumented change (#​1382) to the previous release.

Additions

• Add rand::distr::Alphabetic distribution. (#​1587)
• Re-export rand_core (#​1604)

v0.9.0

Compare Source

Security and unsafe

• Policy: "rand is not a crypto library" (#​1514)
• Remove fork-protection from ReseedingRng and ThreadRng. Instead, it is recommended to call ThreadRng::reseed on fork. (#​1379)
• Use zerocopy to replace some unsafe code (#​1349, #​1393, #​1446, #​1502)

Dependencies

• Bump the MSRV to 1.63.0 (#​1207, #​1246, #​1269, #​1341, #​1416, #​1536); note that 1.60.0 may work for dependents when using --ignore-rust-version
• Update to rand_core v0.9.0 (#​1558)

Features

• Support std feature without getrandom or rand_chacha (#​1354)
• Enable feature small_rng by default (#​1455)
• Remove implicit feature rand_chacha; use std_rng instead. (#​1473)
• Rename feature serde1 to serde (#​1477)
• Rename feature getrandom to os_rng (#​1537)
• Add feature thread_rng (#​1547)

API changes: rand_core traits

• Add fn RngCore::read_adapter implementing std::io::Read (#​1267)
• Add trait CryptoBlockRng: BlockRngCore; make trait CryptoRng: RngCore (#​1273)
• Add traits TryRngCore, TryCryptoRng (#​1424, #​1499)
• Rename fn SeedableRng::from_rng -> try_from_rng and add infallible variant fn from_rng (#​1424)
• Rename fn SeedableRng::from_entropy -> from_os_rng and add fallible variant fn try_from_os_rng (#​1424)
• Add bounds Clone and AsRef to associated type SeedableRng::Seed (#​1491)

API changes: Rng trait and top-level fns

• Rename fn rand::thread_rng() to rand::rng() and remove from the prelude (#​1506)
• Remove fn rand::random() from the prelude (#​1506)
• Add top-level fns random_iter, random_range, random_bool, random_ratio, fill (#​1488)
• Re-introduce fn Rng::gen_iter as random_iter (#​1305, #​1500)
• Rename fn Rng::gen to random to avoid conflict with the new gen keyword in Rust 2024 (#​1438)
• Rename fns Rng::gen_range to random_range, gen_bool to random_bool, gen_ratio to random_ratio (#​1505)
• Annotate panicking methods with #[track_caller] (#​1442, #​1447)

API changes: RNGs

• Fix <SmallRng as SeedableRng>::Seed size to 256 bits (#​1455)
• Remove first parameter (rng) of ReseedingRng::new (#​1533)

API changes: Sequences

• Split trait SliceRandom into IndexedRandom, IndexedMutRandom, SliceRandom (#​1382)
• Add IndexedRandom::choose_multiple_array, index::sample_array (#​1453, #​1469)

API changes: Distributions: renames

• Rename module rand::distributions to rand::distr (#​1470)
• Rename distribution Standard to StandardUniform (#​1526)
• Move distr::Slice -> distr::slice::Choose, distr::EmptySlice -> distr::slice::Empty (#​1548)
• Rename trait distr::DistString -> distr::SampleString (#​1548)
• Rename distr::DistIter -> distr::Iter, distr::DistMap -> distr::Map (#​1548)

API changes: Distributions

• Relax Sized bound on Distribution<T> for &D (#​1278)
• Remove impl of Distribution<Option<T>> for StandardUniform (#​1526)
• Let distribution StandardUniform support all NonZero* types (#​1332)
• Fns {Uniform, UniformSampler}::{new, new_inclusive} return a Result (instead of potentially panicking) (#​1229)
• Distribution Uniform implements TryFrom instead of From for ranges (#​1229)
• Add UniformUsize (#​1487)
• Remove support for generating isize and usize values with StandardUniform, Uniform (except via UniformUsize) and Fill and usage as a WeightedAliasIndex weight (#​1487)
• Add impl DistString for distributions Slice<char> and Uniform<char> (#​1315)
• Add fn Slice::num_choices (#​1402)
• Add fn p() for distribution Bernoulli to access probability (#​1481)

API changes: Weighted distributions

• Add pub module rand::distr::weighted, moving WeightedIndex there (#​1548)
• Add trait weighted::Weight, allowing WeightedIndex to trap overflow (#​1353)
• Add fns weight, weights, total_weight to distribution WeightedIndex (#​1420)
• Rename enum WeightedError to weighted::Error, revising variants (#​1382) and mark as #[non_exhaustive] (#​1480)

API changes: SIMD

• Switch to std::simd, expand SIMD & docs (#​1239)

Reproducibility-breaking changes

• Make ReseedingRng::reseed discard remaining data from the last block generated (#​1379)
• Change fn SmallRng::seed_from_u64 implementation (#​1203)
• Allow UniformFloat::new samples and UniformFloat::sample_single to yield high (#​1462)
• Fix portability of distribution Slice (#​1469)
• Make Uniform for usize portable via UniformUsize (#​1487)
• Fix IndexdRandom::choose_multiple_weighted for very small seeds and optimize for large input length / low memory (#​1530)

Reproducibility-breaking optimisations

• Optimize fn sample_floyd, affecting output of rand::seq::index::sample and rand::seq::SliceRandom::choose_multiple (#​1277)
• New, faster algorithms for IteratorRandom::choose and choose_stable (#​1268)
• New, faster algorithms for SliceRandom::shuffle and partial_shuffle (#​1272)
• Optimize distribution Uniform: use Canon's method (single sampling) / Lemire's method (distribution sampling) for faster sampling (breaks value stability; #​1287)
• Optimize fn sample_single_inclusive for floats (+~20% perf) (#​1289)

Other optimisations

• Improve SmallRng initialization performance (#​1482)
• Optimise SIMD widening multiply (#​1247)

Other

• Add Cargo.lock.msrv file (#​1275)
• Reformat with rustfmt and enforce (#​1448)
• Apply Clippy suggestions and enforce (#​1448, #​1474)
• Move all benchmarks to new benches crate (#​1329, #​1439) and migrate to Criterion (#​1490)

Documentation

• Improve ThreadRng related docs (#​1257)
• Docs: enable experimental --generate-link-to-definition feature (#​1327)
• Better doc of crate features, use doc_auto_cfg (#​1411, #​1450)

v0.8.8

Compare Source

What's Changed

• v0.8: fix no-std + serde1 build by @​dhardy in #​1825

Full Changelog: rust-random/rand@0.8.7...0.8.8

v0.8.7

Compare Source

What's Changed

• rand v0.8.7: backport #​1790 by @​dhardy in #​1804

Full Changelog: rust-random/rand@0.8.6...0.8.7

v0.8.6

Compare Source

What's Changed

This release back-ports a fix from v0.10. See also #​1763.

Changes

• Deprecate feature log (#​1772)

• Drop the experimental simd_support feature.

New Contributors

• @​nwalfield made their first contribution in #​1772

Full Changelog: rust-random/rand@0.8.5...0.8.6

---

Configuration

📅 Schedule: (UTC)

• Branch creation
  • At any time (no schedule defined)
• Automerge
  • At any time (no schedule defined)

🚦 Automerge: Disabled by config. Please merge this manually once you are satisfied.

♻ Rebasing: Whenever PR becomes conflicted, or you tick the rebase/retry checkbox.

🔕 Ignore: Close this PR and you won't be reminded about these updates again.

---

• If you want to rebase/retry this PR, check this box

---

This PR was generated by Mend Renovate. View the repository job log.

[greptile-apps]

PR author is in the excluded authors list.

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 223b6b6a25

ℹ️ About Codex in GitHub

Codex has been enabled to automatically review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

When you sign up for Codex through ChatGPT, Codex can also answer questions or update the PR, like "@codex address that feedback".

[chatgpt-codex-connector]

Preserve seed replay for the simulation harnesses

Upgrading the simulation crates from rand 0.9 to 0.10 changes the sampler behind calls like random_range/random_bool, so a ChaCha8Rng seeded with an old failing seed will no longer drive the same operation stream. That breaks the reproducibility contract documented in crates/gossip-coordination/src/sim/mod.rs:31-34 and the seed-replay workflow relied on by crates/gossip-coordination/src/sim/harness_tests.rs:577-579 and crates/gossip-coordination/src/sim/mega_sim_tests.rs:495-521: any seed captured from CI or a proptest-regressions file on the 0.9 line stops reproducing the original failure after this bump.

Useful? React with 👍 / 👎.

[chatgpt-codex-connector]

Keep bootstrap CI generation on a value-stable RNG

The eval harness uses StdRng::seed_from_u64 in tools/eval-harness/src/metrics.rs:268-283,543 specifically so bootstrap confidence intervals are deterministic and comparable in CI. Moving this dependency to rand 0.10 weakens that guarantee because StdRng on the 0.10 line is explicitly non-portable, so the same BootstrapConfig { seed: 42 } can start producing different intervals after a routine dependency refresh. That will make report diffs noisy and undermine the “same input, same CI” behavior this module documents.

Useful? React with 👍 / 👎.

[kilo-code-bot]

CRITICAL: The RNG stack is only half-upgraded here

crates/gossip-coordination/src/sim/mod.rs and the other simulation modules still use the 0.9 rand API with rand_chacha::ChaCha8Rng. Moving only rand to 0.10 is not source-compatible: 0.10 changed the trait surface, while rand_chacha 0.9 still sits on the older rand_core line. As written, the test-support targets stop compiling until the RNG crates and imports are updated together.

[kilo-code-bot]

CRITICAL: This leaves the crate on incompatible rand APIs

crates/gossip-coordination-etcd/src/sim_etcd_kv.rs and crates/gossip-coordination-etcd/tests/concurrent_cas_contention.rs still import rand::{Rng, SeedableRng} and drive ChaCha8Rng directly. Bumping only rand to 0.10 is not source-compatible here: the 0.10 traits no longer match rand_chacha 0.9, so these simulator/test targets stop compiling until the RNG versions and imports move in lockstep.

[kilo-code-bot]

CRITICAL: The simulator still depends on the pre-0.10 RNG API

crates/gossip-persistence-inmemory/src/sim/mod.rs and crates/gossip-persistence-inmemory/tests/differential_oracle.rs still use the 0.9 rand API with rand_chacha::ChaCha8Rng. Updating only rand to 0.10 leaves those call sites on incompatible traits, so the test-support code stops compiling until the RNG stack is upgraded together.

[kilo-code-bot]

Code Review Summary

Status: 4 Issues Found | Recommendation: Address before merge

Overview

Severity	Count
CRITICAL	4
WARNING	0
SUGGESTION	0

Fix these issues in Kilo Cloud

Issue Details (click to expand)

CRITICAL

File	Line	Issue
crates/gossip-coordination/Cargo.toml	21	rand 0.10 is still paired with rand_chacha 0.9, leaving simulation and test-support code on incompatible rand_core trait versions.
crates/gossip-coordination-etcd/Cargo.toml	32	The etcd simulator/tests still drive ChaCha8Rng through rand traits, so this partial RNG upgrade leaves those targets on incompatible APIs.
crates/gossip-persistence-inmemory/Cargo.toml	20	The in-memory simulator/test code still uses rand_chacha 0.9, so the rand 0.10 bump breaks the seeded RNG trait stack.
tools/eval-harness/Cargo.toml	17	The harness source still calls the pre-0.9 rng.gen_range(...) API, so moving this direct dependency to rand 0.10 leaves the standalone tool source-incompatible.

Files Reviewed (6 files)

• Cargo.lock - generated lockfile, dependency graph reviewed
• crates/gossip-coordination-etcd/Cargo.toml - 1 issue
• crates/gossip-coordination/Cargo.toml - 1 issue
• crates/gossip-persistence-inmemory/Cargo.toml - 1 issue
• tools/eval-harness/Cargo.lock - generated lockfile, dependency graph reviewed
• tools/eval-harness/Cargo.toml - 1 issue

Previous Review Summary (commit 6741630)

Current summary above is authoritative. Previous snapshots are kept for context only.

Previous review (commit 6741630)

Status: 4 Issues Found | Recommendation: Address before merge

Overview

Severity	Count
CRITICAL	4
WARNING	0
SUGGESTION	0

Fix these issues in Kilo Cloud

Issue Details (click to expand)

CRITICAL

File	Line	Issue
crates/gossip-coordination/Cargo.toml	21	rand 0.10 is still paired with rand_chacha 0.9, which leaves the simulation/test-support targets on split rand_core versions.
crates/gossip-coordination-etcd/Cargo.toml	32	This crate still drives ChaCha8Rng through rand traits, so bumping only rand to 0.10 leaves the etcd simulator/tests on incompatible RNG APIs.
crates/gossip-persistence-inmemory/Cargo.toml	20	The in-memory persistence simulator still uses rand_chacha 0.9, so this partial RNG upgrade keeps the test code on incompatible trait versions.

Other Observations (not in diff)

Issues found in unchanged code that cannot receive inline comments:

File	Line	Issue
tools/eval-harness/src/metrics.rs	555	The bootstrap sampling loop still calls rng.gen_range(...), so the eval harness source remains on the pre-0.9 rand API after this dependency bump.

Files Reviewed (6 files)

• Cargo.lock - generated lockfile, no direct review findings
• crates/gossip-coordination-etcd/Cargo.toml - 1 issue
• crates/gossip-coordination/Cargo.toml - 1 issue
• crates/gossip-persistence-inmemory/Cargo.toml - 1 issue
• tools/eval-harness/Cargo.lock - generated lockfile, no direct review findings
• tools/eval-harness/Cargo.toml - 1 related unchanged-code issue in tools/eval-harness/src/metrics.rs

---

_{Reviewed by gpt-5.5-2026-04-23 · Input: 119.8K · Output: 20.6K · Cached: 645.1K}