Eliza Weisman

it's worth noting that the type was previously named `CausalCell`, but was renamed to `UnsafeCell` in #111. here's Carl's rationale at the time: https://github.com/tokio-rs/loom/pull/111#discussion_r391882249

Try increasing the value of `max_branches` by a bit?

In this particular case, it's somewhat important that `loom`'s lazy static be used in the test, as using a "real" lazy static results in non-deterministic behavior, hitting an internal loom...

I'd be happy to help out with this, btw.

Because Loom's `UnsafeCell` contains additional data for tracking its state in the model, I'm not sure if it can be used with `MaybeUninit` in the same way as the std...

It would be really nice to have a fix for the double panic...it's definitely annoying when tests fail that way, as it's hard to get useful information out of them....

this is extremely cool!

:wave: hi @bdonlan, sorry this went so long without a review --- are you still interested in working on this? i'd love to give this branch a review if you're...

Yeah, we should probably put a link in the README/lib.rs docs that links to the docs added in #253.

@Restioson > Would a PR to implement this be accepted? Yes, that would be great! It should probably be simulated using `Atomic::unsync_load`: https://github.com/tokio-rs/loom/blob/73e6c14da41978e7a20e1993c2554a16a3dcb597/src/rt/atomic.rs#L200-L216 We might want to add `get_mut` methods...