Gerd Zellweger

1 core, 4 generator, 20M events, things are fine: | name | ('elapsed', 'mean') | ('elapsed', 'std') | ('elapsed', 'min') | ('elapsed', 'max') | |-------:|----------------------:|---------------------:|---------------------:|---------------------:| | 0 | 7.70756 |...

seems like problems start with any amount of cores >1: 2 cores, 4 generators: | name | ('elapsed', 'mean') | ('elapsed', 'std') | ('elapsed', 'min') | ('elapsed', 'max') | |-------:|----------------------:|---------------------:|---------------------:|---------------------:|...

Thanks, any reason why you'd need a non-static lifetime in the new_from_slice? e.g., would this be simpler: `pub fn new_from_slice(slice: &'static [T])`

Sounds good to me, while we're waiting for v2, I will work on adding support for Substr + tests.

I added tests and implement traits for Substr too.

hi, is there anything I can do to move this PR forward?

> This breaks library upgraders - to avoid this you'd need to keep the default rkyv feature and point it to rkyv-32. To do this you'll also need to use...

@idelvall since we added this, we haven't spotted any new errors of this form. However, of course the CI cache utility is much lower now.

Might be able to put it behind `extra_traits` until rustc has "a better fix" for this.

Hm sorry about that, the cr4 function uses `from_bits_truncate` from bitflags https://github.com/gz/rust-x86/blob/ae3306a372c82a92b2e0f7ca81c6664455625c7f/src/controlregs.rs#L152 to make sure the value only sets the bits that are known to the Cr4 bitflag struct. This...