Let `RngCore: TryRngCore<Error = Infallible>` by dhardy · Pull Request #45 · rust-random/rand_core

dhardy · 2026-01-12T11:57:45Z

Added a CHANGELOG.md entry

Summary

Revise the core traits according to @newpavlov's suggestion and similar to #41 but keeping the infallible methods.

No trait renames are done in this PR. Also not sure if we need any.

Details

TryRngCore is implemented for any R: DerefMut where Target: TryRngCore.
RngCore: TryRngCore<Error = Infallible> with automatic implementation over TryRngCore
UnwrapMut is removed
CryptoRng = RngCore + TryCryptoRng or as close as possible without sum traits; it is kept since I believe some people need dyn CryptoRng
utils::{next_u64_via_u32, next_word_via_fill} now use closures instead of R: RngCore

Questions

Should utils fns support potentially-fallible uses (try_next_u64_via_u32 takes FnMut() -> Result<u32; E> for generic E)? I think unnecessary.

dhardy · 2026-01-13T09:17:00Z

Updated. Some things need further discussion (see above).

This fixes changes made in #2195 This is due to the refactor made in rand_core in rust-random/rand_core#45 which dropped the "trait dependency" between CryptoRng and RngCore

tarcieri · 2026-01-19T22:33:59Z

We noticed something a little weird here:

So:

pubtraitCryptoRng:TryCryptoRng<Error = Infallible>{}

impl<R:TryRngCore<Error = Infallible>>RngCoreforR{ ...

I would expect that means any type which impls CryptoRng also impls RngCore, but in RustCrypto/traits#2197 we had to add explicit RngCore bounds (or reborrow as in RustCrypto/traits#2195)

newpavlov · 2026-01-19T22:35:16Z

I guess we should revert my suggestion above and add a test for this.

tarcieri · 2026-01-19T22:50:37Z

I don't get it... I tried to make a contrived version in the playground and it... works?

https://play.rust-lang.org/?version=stable&mode=debug&edition=2024&gist=be7a3fdc3b7bc93fb22c1529d99d4f03

tarcieri · 2026-01-19T22:55:48Z

Aha, the ?Sized bound breaks the blanket impl of RngCore for TryRngCore.

I think adding ?Sized to the bounds of the blanket impl should fix the problem?

Updated contrived example:

https://play.rust-lang.org/?version=stable&mode=debug&edition=2024&gist=a352bef7b0f1c475ed0715326782d5dc

Adds `?Sized` to the bounds of the blanket impls, so unsized types can also be used by way of the blanket impl. See also: #45

tarcieri · 2026-01-19T23:03:46Z

Opened a PR to add ?Sized to the bounds of the blanket impl: #51

baloo · 2026-01-20T05:40:12Z

- /// Wrap RNG with the [`UnwrapMut`] wrapper.
- fn unwrap_mut(&mut self) -> UnwrapMut<'_, Self> {
- UnwrapMut(self)
- }


Dropping UnwrapMut causes me trouble in dhkem. UnwrapErr requires R to be Sized which may not be possible if the Rng comes from a trait.

see RustCrypto/KEMs#182

I only see failures "failed to load source for dependency rand_core". Can you point me to a job failing because of UnwrapErr?
Note that UnwrapErr itself does not require the Sized bound and its only field is public, so you should be able to create it using UnwrapErr(rng).

Hm, we already have test which test the ?Sized case, see: https://github.com/rust-random/rand_core/blob/master/src/lib.rs#L620-L623

@baloo(&mut rng).unwrap_err() should be equivalent to the old rng.unwrap_mut().
(We could perhaps add unwrap_mut(&mut self) -> UnwrapErr<&mut Self>, but it shouldn't be necessary.)

Maybe we should just remove the unwrap_err method in favor of wrapping in UnwrapErr?

Yeah, UnwrapErr(rng) worked, I was convinced I tried it out yesterday night, but I guess not. Thanks a lot!

tarcieri · 2026-01-23T14:42:08Z

So, with a RngCore: TryRngCore<Error = Infallible> bound there are two ways to write the blanket impl:

You can blanket impl RngCore for R: TryRngCore<Error = Infallible>
Or you can blanket impl TryRngCore<Error = Infallible> for types which impl RngCore

Either way the RngCore: TryRngCore<Error = Infallible> bound is met.

I think #2 feels a little more straightforward to me? Types which don't need fallibility impl the infallible trait and get the fallible trait for free. Maybe instead of this:

impl<R>RngCoreforRwhereR:TryRngCore<Error = Infallible> + ?Sized,{

...it could be:

impl<R>TryRngCoreforRwhereR:RngCore{typeError = Infallible;[...]

As far as I can tell the two approaches are duals of each other and should achieve the same set of impls.

newpavlov · 2026-01-23T14:45:14Z

Previously, we were using the second option but it was causing issues with overlapping blanket impls. Also, I think Rng: TryRng<Error = !> is the correct approach to this problem. Ideally, it would be nice if we were able to write implementation of Rng which would desugar to TryRng<Error = !>, but I am not sure how such language feature could be implemented.

tarcieri · 2026-01-23T14:50:12Z

Can you give a concrete example of the problem with overlapping blanket impls?

Rng: TryRng<Error = !>

Is that not just RngCore: TryRngCore<Error = Infallible> with different trait names and syntax for Infallible?

newpavlov · 2026-01-23T14:54:14Z

We were unable to add DerefMut blanket impl for Try(Crypto)Rng. One of consequences of it is that we needed both UnwrapErr and UnwrapMut. It was mentioned in the original proposal.

Is that not just RngCore: TryRngCore<Error = Infallible> with different trait names and syntax for Infallible?

Yes, I just wrote it in a shorter fashion.

tarcieri · 2026-01-23T15:03:08Z

Aah OK, I see it now

dhardy · 2026-01-23T16:46:32Z

Is that not just RngCore: TryRngCore<Error = Infallible> with different trait names and syntax for Infallible?

Effectively RngCore is a trait alias (see #41) with extra methods.

We could move those extra methods to an extension trait (add trait RngExt here, then let rand::RngExt: rand_core::RngExt), though it seems we don't have much justification for the extra trait. (See also this comment.)

tarcieri · 2026-01-23T16:47:56Z

Yeah it seems fine given what you're trying to do

…ndom#45)

dhardy added 10 commits January 12, 2026 08:14

impl TryRngCore for R: DerefMut where Target: TryRngCore
752fdbf

Remove UnwrapMut, moving fn re to UnwrapErr
87c0cd8

impl TryRngCore for UnwrapErr
25c4f5d

Let RngCore: TryRngCore
ea88cf9

Fix tests
c7fd184

Doc: RngCore and TryRngCore traits
a2a35c3

Doc: crypto traits
20003bd

Doc: block module
4a467b5

Doc: utils module
635c64e

Let next_u64_via_u32 and next_word_via_fill use closures
ae26d43

dhardy mentioned this pull request Jan 12, 2026
Merge RngCore, TryRngCore#41
Closed
3 tasks

dhardy added 2 commits January 12, 2026 12:18

Shut Clippy up
b0fabb0

CHAEGLOG
b339a5c

newpavlov reviewed Jan 12, 2026
View reviewed changes

Comment threadCHANGELOG.md Outdated
Comment threadsrc/utils.rs Outdated
Comment threadsrc/lib.rs
Comment threadsrc/lib.rs Outdated
Comment threadsrc/lib.rs Outdated
Comment threadsrc/lib.rs

dhardy added 2 commits January 13, 2026 09:14

Doc: review updates
269a63e

utils: use TryRngCore / Result
061219d

newpavlov reviewed Jan 14, 2026
View reviewed changes

Comment threadCHANGELOG.md Outdated

newpavlov reviewed Jan 14, 2026
View reviewed changes

Comment threadsrc/lib.rs Outdated

dhardy changed the title ~~Let RngCore: TryRngCore~~Let RngCore: TryRngCore<Error = Infallible>Jan 14, 2026

Review updates
c675388

newpavlov reviewed Jan 14, 2026
View reviewed changes

Comment threadsrc/lib.rs Outdated

newpavlov approved these changes Jan 14, 2026
View reviewed changes

dhardy added 2 commits January 14, 2026 18:28

Doc tweak for CryptoRng
d6ea763

Merge CHANGELOG
7fbc8a7

dhardy merged commit 48d9605 into masterJan 14, 2026
13 checks passed

dhardy deleted the push-zqznkttsmlvp branch January 14, 2026 18:37

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

baloo mentioned this pull request Jan 19, 2026
elliptic-curve: fixup rand_core API use RustCrypto/traits#2197
Merged

tarcieri added a commit that referenced this pull request Jan 19, 2026
Support ?Sized types in Rng*: TryRng* blanket impls
c14a054
Adds `?Sized` to the bounds of the blanket impls, so unsized types can also be used by way of the blanket impl. See also: #45

tarcieri mentioned this pull request Jan 19, 2026
Support ?Sized types in Rng*: TryRng* blanket impls #51
Merged

baloo reviewed Jan 20, 2026
View reviewed changes

This was referenced Jan 23, 2026
Reorganize changelog for v0.10.0 #59
Merged
Tracker: rand_core v0.10 #13
Closed

takumi-earth pushed a commit to earthlings-dev/rand_core that referenced this pull request Jan 27, 2026
Let RngCore be a sub-trait of TryRngCore<Error = Infallible> (rust-ra…
2a0b42e
…ndom#45)

Liam-DeVoe mentioned this pull request Mar 31, 2026
Upgrade to rand 0.10 hegeldev/hegel-rust#161
Merged

BemusedVermin mentioned this pull request Apr 27, 2026
chore(deps): bump rand_core 0.9→0.10 and rand_xoshiro 0.7→0.8 BemusedVermin/evolution-simulation#230
Merged
3 tasks

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

Uh oh!

Uh oh!

Conversation

dhardy commented Jan 12, 2026• edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Summary

Details

Questions

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

dhardy commented Jan 13, 2026

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

tarcieri commented Jan 19, 2026

Uh oh!

newpavlov commented Jan 19, 2026

Uh oh!

tarcieri commented Jan 19, 2026

Uh oh!

tarcieri commented Jan 19, 2026• edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

tarcieri commented Jan 19, 2026

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

tarcieri commented Jan 23, 2026

Uh oh!

newpavlov commented Jan 23, 2026• edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

tarcieri commented Jan 23, 2026

Uh oh!

newpavlov commented Jan 23, 2026• edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

tarcieri commented Jan 23, 2026

Uh oh!

dhardy commented Jan 23, 2026

Uh oh!

tarcieri commented Jan 23, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

dhardy commented Jan 12, 2026•
edited
Loading

tarcieri commented Jan 19, 2026•
edited
Loading

newpavlov commented Jan 23, 2026•
edited
Loading

newpavlov commented Jan 23, 2026•
edited
Loading