Move ISAAC generators to their own crate #551
Conversation
|
I think the ISAAC generators should be removed from the table as well. Other than that, it looks good to me. |
|
I'd like to have more comprehensive overviews of available PRNGs. These ISAAC generators are not quite CSPRNGs but that was where they fit best. Okay, I admit, part of this is just not wanting to throw away all the work @pitdicker did on these implementations. 😄 |
rand_isaac/README.md
Outdated
|
|
||
| ## Crate Features | ||
|
|
||
| `rand_isaac` supports `no_std` and `alloc`-only configurations, as well as full |
There was a problem hiding this comment.
This crate always works with no_std, and only has the serde1 feature?
There was a problem hiding this comment.
Whoops, I forgot to update this bit. 👍
There was a problem hiding this comment.
Okay, I admit, part of this is just not wanting to throw away all the work @pitdicker did on these implementations. smile
Git history preserves everything 😄. Still, having a good implementation of ISAAC in the ecosystem seems like good to have.
Until you purge it! But discoverability and ease-of-use also matter I think. |
Part of #431
cargo doc --no-deps --all --all-featuresseems to fail in my tests unless I remove the--all-featuresflag. Not sure what's wrong here.I left the PRNG comparisons in the
prngmodule doc for now. We may decide to move that underdoc/later.I also left the benchmarks as-is; this allows them to be compared across crates. I'm not sure whether to keep
rand_isaaca dev-dependency just for this; perhaps eventually we should move the benchmarks.For now the lib depends on
rand_isaacanyway, just for the deprecation warnings (existing ones, and new ones for the names in theprngmodule).