-
Notifications
You must be signed in to change notification settings - Fork 92
revealing current seed as a public read-only property #44
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
base: master
Are you sure you want to change the base?
Conversation
@@ -31,6 +31,9 @@ public protocol Seedable { | |||
/// The seed type. | |||
associatedtype Seed | |||
|
|||
/// Current seed value. | |||
var seed: Seed { get } |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Requiring Seedable
to have a seed
property is a breaking change since types outside of this project can conform to this protocol and would have to update their code accordingly.
public typealias Seed = (UInt64, UInt64) | ||
|
||
/// The seed value. | ||
public var seed: Seed { |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
It would be better if _state
was made public and mutable as state
instead. See my comment about Seedable
.
public typealias Seed = (UInt32, UInt32, UInt32, UInt32) | ||
|
||
/// The seed value. | ||
public var seed: Seed { |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
x
, y
, z
, and y
can probably be made into a public mutable state
variable which can be exposed instead.
@@ -38,6 +37,11 @@ public struct XorshiftStar: RandomBytesGenerator, Seedable, SeedableFromRandomGe | |||
UInt64, UInt64, UInt64, UInt64 | |||
) | |||
|
|||
/// The seed value. | |||
public var seed: Seed { |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Similar to with Xoroshiro
, should probably change _state
to be public and mutable instead.
/// The seed type. | ||
public typealias Seed = [UInt32] | ||
|
||
/// The seed value | ||
public var seed: Seed { |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Unless ChaCha(seed: rng.seed).seed == rng.seed
, this should be removed since it may not be valid to consider this as a seed. Please add a test to ensure this invariant is met.
Side note: this makes me wish Swift had more of Rust's abilities regarding protocols/traits. How I'd implement pub trait Seedable<S> {
fn from_seed(seed: S) -> Self where Self: Sized;
fn reseed(&mut self, seed: S);
} Then there would be no need for impl<S: IntoIterator<Item=u64>> Seedable<S> for ChaCha {
default fn from_seed(seed: S) -> ChaCha {
for s in seed.into_iter() {
/* ... */
}
}
default fn reseed(&mut self, seed: S) {
for s in seed.into_iter() {
/* ... */
}
}
}
impl<S> Seedable<S> for ChaCha
where
S: for<'a> IntoIterator<Item=&'a u64>
{
default fn from_seed(seed: S) -> ChaCha {
// Dereference the integer
let iter = seed.map(|&s| s);
ChaCha::from_seed(iter)
}
default fn reseed(&mut self, seed: S) {
// Dereference the integer
let iter = seed.map(|&s| s);
self.reseed(iter)
}
} Specialization would allow for: impl Seedable<u64> for ChaCha {
fn from_seed(seed: u64) -> ChaCha {
ChaCha::from_seed(&[seed])
}
fn reseed(&mut self, seed: u64) {
self.reseed(&[seed])
}
} Edit: apparently because of problems with specialization, this implementation wouldn't be currently possible in Rust either 😢 https://play.rust-lang.org/?gist=409dba7bc6162833285c8437815bc178&version=nightly |
Sometimes it might be helpful to be able to log seed that was used to create generator.
So we can reproduce same random sequence using that seed we know.
For example it can be logged for test suit session whose data is generated randomly.
And if some tests fail, we can fix them and re-run test suit using that seed to verify it on the same 'random' data.