Tom Kaitchuck

Thanks for putting this together. Would you like to collaborate working in a single branch? I don't have such a CPU to test on, which has held me back on...

@Jefffrey Are you able to repro #195 ? If so can you take a look at #197

Fixed in 0.1.17

@decathorpe This is very confusing. Is it the case that this fails on `s390x` ``` fn test_or() { assert_eq!(((u64::MAX as u128) > 64, u64::MAX as u128); } ``` If so,...

I don't understand how the error you have above for operations::test::test_add_length could be generated if the assertion in the `main` you wrote passes.

@jamessan Ok good. That change is not actually necessary for the algorithm to be correct. We can make the test less sensitive by having it work with `[u64; 2]` and...

If because this is solely a DoS attack vulnerability and not a soundness issue it does not meet the criteria for inclusion, then feel free the close this.

@dtolnay Good point. In testing I found that if I modified your example to have two layers of wrapping methods each annotated with `#[inline(never)]` that problem goes away. (Perhaps the...

> This is because we expose raw hashes as part of the RawEntry and RawTable APIs, where the user will pass hashes calculated using the standard method of Hash::hash That...

On thought: The trait I am using do do the dispatch is here: https://github.com/tkaitchuck/aHash/blob/master/src/lib.rs#L120 It is defined for all hasher types so that it can work seamlessly with other hashers....