Revise index / reset / skip methods by dhardy · Pull Request #44 · rust-random/rand_core

dhardy · 2026-01-12T07:44:30Z

Added a CHANGELOG.md entry

Summary

Make fn index private; add fn word_offset as a pub replacement.

Remove fn reset; rename fn generate_and_set to reset_and_skip.

Motivation

The old fn index was easy to mis-use: RustCrypto/stream-ciphers#438 (comment). I don't think this was the first time.

We also don't need two "reset" methods; it should be equivalent to call reset_and_skip(0).

@nstilt1 please review.

dhardy · 2026-01-13T13:31:15Z

From the other PR and patch I take it that @nstilt1 finds this acceptable.

… change (#500) # Changes * Update `rand_core` to a preview from rust-random/rand_core#44 * I need to change the `Cargo.toml` once that PR makes it through * Update `set_stream()` to reset the `block_pos` + `index` to 0. Before, calling `set_stream()` would basically behave like this: ```rust fn set_stream(stream_id: S) { self.stream = S.into(); if self.index != 64 { // generate_and_set is called, resulting in self.block_pos += 4; self.index = self.index; // no change to the index, but change to the block_pos } } ``` * * I had a test that had repeated `set_stream()` calls and the only way to replicate it was to manually find the `word_pos` that resulted from each call. We can remove this test or at the very least reduce the test, since I had to add more assertions to figure out what was going on in the previous implementation. * Update `get_block_pos()` to return the current `block_pos` instead of returning `block_pos + 4` due to `generate()` calls incrementing the pos by 4. * This is not a breaking change since `get_block_pos()` was not implemented for `rand_chacha`, but it is a breaking change for anyone who was using `get_block_pos()` from the prerelease versions. * Update docs to explain that `set_stream()` nullifies any calls made to `set_word_pos()` or `set_block_pos()`. --------- Co-authored-by: Arthur Gautier <arthur.gautier@arista.com>

dhardy · 2026-01-12T07:44:30Z

Added a CHANGELOG.md entry

Summary

Make fn index private; add fn word_offset as a pub replacement.

Remove fn reset; rename fn generate_and_set to reset_and_skip.

Motivation

The old fn index was easy to mis-use: RustCrypto/stream-ciphers#438 (comment). I don't think this was the first time.

We also don't need two "reset" methods; it should be equivalent to call reset_and_skip(0).

@nstilt1 please review.

dhardy · 2026-01-13T13:31:15Z

From the other PR and patch I take it that @nstilt1 finds this acceptable.

… change (#500) # Changes * Update `rand_core` to a preview from rust-random/rand_core#44 * I need to change the `Cargo.toml` once that PR makes it through * Update `set_stream()` to reset the `block_pos` + `index` to 0. Before, calling `set_stream()` would basically behave like this: ```rust fn set_stream(stream_id: S) { self.stream = S.into(); if self.index != 64 { // generate_and_set is called, resulting in self.block_pos += 4; self.index = self.index; // no change to the index, but change to the block_pos } } ``` * * I had a test that had repeated `set_stream()` calls and the only way to replicate it was to manually find the `word_pos` that resulted from each call. We can remove this test or at the very least reduce the test, since I had to add more assertions to figure out what was going on in the previous implementation. * Update `get_block_pos()` to return the current `block_pos` instead of returning `block_pos + 4` due to `generate()` calls incrementing the pos by 4. * This is not a breaking change since `get_block_pos()` was not implemented for `rand_chacha`, but it is a breaking change for anyone who was using `get_block_pos()` from the prerelease versions. * Update docs to explain that `set_stream()` nullifies any calls made to `set_word_pos()` or `set_block_pos()`. --------- Co-authored-by: Arthur Gautier <arthur.gautier@arista.com>

dhardy · 2026-01-12T07:44:30Z

Added a CHANGELOG.md entry

Summary

Make fn index private; add fn word_offset as a pub replacement.

Remove fn reset; rename fn generate_and_set to reset_and_skip.

Motivation

The old fn index was easy to mis-use: RustCrypto/stream-ciphers#438 (comment). I don't think this was the first time.

We also don't need two "reset" methods; it should be equivalent to call reset_and_skip(0).

@nstilt1 please review.

dhardy · 2026-01-13T13:31:15Z

From the other PR and patch I take it that @nstilt1 finds this acceptable.

… change (#500) # Changes * Update `rand_core` to a preview from rust-random/rand_core#44 * I need to change the `Cargo.toml` once that PR makes it through * Update `set_stream()` to reset the `block_pos` + `index` to 0. Before, calling `set_stream()` would basically behave like this: ```rust fn set_stream(stream_id: S) { self.stream = S.into(); if self.index != 64 { // generate_and_set is called, resulting in self.block_pos += 4; self.index = self.index; // no change to the index, but change to the block_pos } } ``` * * I had a test that had repeated `set_stream()` calls and the only way to replicate it was to manually find the `word_pos` that resulted from each call. We can remove this test or at the very least reduce the test, since I had to add more assertions to figure out what was going on in the previous implementation. * Update `get_block_pos()` to return the current `block_pos` instead of returning `block_pos + 4` due to `generate()` calls incrementing the pos by 4. * This is not a breaking change since `get_block_pos()` was not implemented for `rand_chacha`, but it is a breaking change for anyone who was using `get_block_pos()` from the prerelease versions. * Update docs to explain that `set_stream()` nullifies any calls made to `set_word_pos()` or `set_block_pos()`. --------- Co-authored-by: Arthur Gautier <arthur.gautier@arista.com>

dhardy · 2026-01-12T07:44:30Z

Added a CHANGELOG.md entry

Summary

Make fn index private; add fn word_offset as a pub replacement.

Remove fn reset; rename fn generate_and_set to reset_and_skip.

Motivation

The old fn index was easy to mis-use: RustCrypto/stream-ciphers#438 (comment). I don't think this was the first time.

We also don't need two "reset" methods; it should be equivalent to call reset_and_skip(0).

@nstilt1 please review.

dhardy · 2026-01-13T13:31:15Z

From the other PR and patch I take it that @nstilt1 finds this acceptable.

… change (#500) # Changes * Update `rand_core` to a preview from rust-random/rand_core#44 * I need to change the `Cargo.toml` once that PR makes it through * Update `set_stream()` to reset the `block_pos` + `index` to 0. Before, calling `set_stream()` would basically behave like this: ```rust fn set_stream(stream_id: S) { self.stream = S.into(); if self.index != 64 { // generate_and_set is called, resulting in self.block_pos += 4; self.index = self.index; // no change to the index, but change to the block_pos } } ``` * * I had a test that had repeated `set_stream()` calls and the only way to replicate it was to manually find the `word_pos` that resulted from each call. We can remove this test or at the very least reduce the test, since I had to add more assertions to figure out what was going on in the previous implementation. * Update `get_block_pos()` to return the current `block_pos` instead of returning `block_pos + 4` due to `generate()` calls incrementing the pos by 4. * This is not a breaking change since `get_block_pos()` was not implemented for `rand_chacha`, but it is a breaking change for anyone who was using `get_block_pos()` from the prerelease versions. * Update docs to explain that `set_stream()` nullifies any calls made to `set_word_pos()` or `set_block_pos()`. --------- Co-authored-by: Arthur Gautier <arthur.gautier@arista.com>

dhardy · 2026-01-12T07:44:30Z

Added a CHANGELOG.md entry

Summary

Make fn index private; add fn word_offset as a pub replacement.

Remove fn reset; rename fn generate_and_set to reset_and_skip.

Motivation

The old fn index was easy to mis-use: RustCrypto/stream-ciphers#438 (comment). I don't think this was the first time.

We also don't need two "reset" methods; it should be equivalent to call reset_and_skip(0).

@nstilt1 please review.

dhardy · 2026-01-13T13:31:15Z

From the other PR and patch I take it that @nstilt1 finds this acceptable.

… change (#500) # Changes * Update `rand_core` to a preview from rust-random/rand_core#44 * I need to change the `Cargo.toml` once that PR makes it through * Update `set_stream()` to reset the `block_pos` + `index` to 0. Before, calling `set_stream()` would basically behave like this: ```rust fn set_stream(stream_id: S) { self.stream = S.into(); if self.index != 64 { // generate_and_set is called, resulting in self.block_pos += 4; self.index = self.index; // no change to the index, but change to the block_pos } } ``` * * I had a test that had repeated `set_stream()` calls and the only way to replicate it was to manually find the `word_pos` that resulted from each call. We can remove this test or at the very least reduce the test, since I had to add more assertions to figure out what was going on in the previous implementation. * Update `get_block_pos()` to return the current `block_pos` instead of returning `block_pos + 4` due to `generate()` calls incrementing the pos by 4. * This is not a breaking change since `get_block_pos()` was not implemented for `rand_chacha`, but it is a breaking change for anyone who was using `get_block_pos()` from the prerelease versions. * Update docs to explain that `set_stream()` nullifies any calls made to `set_word_pos()` or `set_block_pos()`. --------- Co-authored-by: Arthur Gautier <arthur.gautier@arista.com>

dhardy · 2026-01-12T07:44:30Z

Added a CHANGELOG.md entry

Summary

Make fn index private; add fn word_offset as a pub replacement.

Remove fn reset; rename fn generate_and_set to reset_and_skip.

Motivation

The old fn index was easy to mis-use: RustCrypto/stream-ciphers#438 (comment). I don't think this was the first time.

We also don't need two "reset" methods; it should be equivalent to call reset_and_skip(0).

@nstilt1 please review.

dhardy · 2026-01-13T13:31:15Z

From the other PR and patch I take it that @nstilt1 finds this acceptable.

… change (#500) # Changes * Update `rand_core` to a preview from rust-random/rand_core#44 * I need to change the `Cargo.toml` once that PR makes it through * Update `set_stream()` to reset the `block_pos` + `index` to 0. Before, calling `set_stream()` would basically behave like this: ```rust fn set_stream(stream_id: S) { self.stream = S.into(); if self.index != 64 { // generate_and_set is called, resulting in self.block_pos += 4; self.index = self.index; // no change to the index, but change to the block_pos } } ``` * * I had a test that had repeated `set_stream()` calls and the only way to replicate it was to manually find the `word_pos` that resulted from each call. We can remove this test or at the very least reduce the test, since I had to add more assertions to figure out what was going on in the previous implementation. * Update `get_block_pos()` to return the current `block_pos` instead of returning `block_pos + 4` due to `generate()` calls incrementing the pos by 4. * This is not a breaking change since `get_block_pos()` was not implemented for `rand_chacha`, but it is a breaking change for anyone who was using `get_block_pos()` from the prerelease versions. * Update docs to explain that `set_stream()` nullifies any calls made to `set_word_pos()` or `set_block_pos()`. --------- Co-authored-by: Arthur Gautier <arthur.gautier@arista.com>

dhardy · 2026-01-12T07:44:30Z

Added a CHANGELOG.md entry

Summary

Make fn index private; add fn word_offset as a pub replacement.

Remove fn reset; rename fn generate_and_set to reset_and_skip.

Motivation

The old fn index was easy to mis-use: RustCrypto/stream-ciphers#438 (comment). I don't think this was the first time.

We also don't need two "reset" methods; it should be equivalent to call reset_and_skip(0).

@nstilt1 please review.

dhardy · 2026-01-13T13:31:15Z

From the other PR and patch I take it that @nstilt1 finds this acceptable.

… change (#500) # Changes * Update `rand_core` to a preview from rust-random/rand_core#44 * I need to change the `Cargo.toml` once that PR makes it through * Update `set_stream()` to reset the `block_pos` + `index` to 0. Before, calling `set_stream()` would basically behave like this: ```rust fn set_stream(stream_id: S) { self.stream = S.into(); if self.index != 64 { // generate_and_set is called, resulting in self.block_pos += 4; self.index = self.index; // no change to the index, but change to the block_pos } } ``` * * I had a test that had repeated `set_stream()` calls and the only way to replicate it was to manually find the `word_pos` that resulted from each call. We can remove this test or at the very least reduce the test, since I had to add more assertions to figure out what was going on in the previous implementation. * Update `get_block_pos()` to return the current `block_pos` instead of returning `block_pos + 4` due to `generate()` calls incrementing the pos by 4. * This is not a breaking change since `get_block_pos()` was not implemented for `rand_chacha`, but it is a breaking change for anyone who was using `get_block_pos()` from the prerelease versions. * Update docs to explain that `set_stream()` nullifies any calls made to `set_word_pos()` or `set_block_pos()`. --------- Co-authored-by: Arthur Gautier <arthur.gautier@arista.com>

dhardy · 2026-01-12T07:44:30Z

Added a CHANGELOG.md entry

Summary

Make fn index private; add fn word_offset as a pub replacement.

Remove fn reset; rename fn generate_and_set to reset_and_skip.

Motivation

The old fn index was easy to mis-use: RustCrypto/stream-ciphers#438 (comment). I don't think this was the first time.

We also don't need two "reset" methods; it should be equivalent to call reset_and_skip(0).

@nstilt1 please review.

dhardy · 2026-01-13T13:31:15Z

From the other PR and patch I take it that @nstilt1 finds this acceptable.

… change (#500) # Changes * Update `rand_core` to a preview from rust-random/rand_core#44 * I need to change the `Cargo.toml` once that PR makes it through * Update `set_stream()` to reset the `block_pos` + `index` to 0. Before, calling `set_stream()` would basically behave like this: ```rust fn set_stream(stream_id: S) { self.stream = S.into(); if self.index != 64 { // generate_and_set is called, resulting in self.block_pos += 4; self.index = self.index; // no change to the index, but change to the block_pos } } ``` * * I had a test that had repeated `set_stream()` calls and the only way to replicate it was to manually find the `word_pos` that resulted from each call. We can remove this test or at the very least reduce the test, since I had to add more assertions to figure out what was going on in the previous implementation. * Update `get_block_pos()` to return the current `block_pos` instead of returning `block_pos + 4` due to `generate()` calls incrementing the pos by 4. * This is not a breaking change since `get_block_pos()` was not implemented for `rand_chacha`, but it is a breaking change for anyone who was using `get_block_pos()` from the prerelease versions. * Update docs to explain that `set_stream()` nullifies any calls made to `set_word_pos()` or `set_block_pos()`. --------- Co-authored-by: Arthur Gautier <arthur.gautier@arista.com>

dhardy requested a review from newpavlov January 12, 2026 07:44

Revise index / reset / skip methods
d072c70

dhardy force-pushed the push-xzmlvzwurnrl branch from 20e67fa to d072c70Compare January 12, 2026 07:46

nstilt1 mentioned this pull request Jan 12, 2026
chacha20: rand_core v0.10.0-rc.4 + set_stream()/get_block_pos() change RustCrypto/stream-ciphers#500
Merged

changed word_offset() to use an immutable reference to self
8935760

dhardy merged commit eca8fcc into masterJan 13, 2026
13 checks passed

dhardy deleted the push-xzmlvzwurnrl branch January 13, 2026 13:31

dhardy mentioned this pull request Jan 19, 2026
Update rand_core to 0.10.0-rc-4 rust-random/rand#1712
Merged
1 task

dhardy mentioned this pull request Jan 23, 2026
Tracker: rand_core v0.10 #13
Closed
17 tasks

takumi-earth pushed a commit to earthlings-dev/rand_core that referenced this pull request Jan 27, 2026
Revise index / reset / skip methods (rust-random#44)
3a1a11f

vxfemboy mentioned this pull request Jul 27, 2026
[PR] Update Rust crate rand_core to 0.10.0 vxfemboy/purrcrypt#20
Closed
1 task