Kunal Sareen

> If "their own API" is in Rust, they can use Option in their API, too. That's not FFI then. It's just Rust, which is perfectly fine. Unfortunately, most modern...

This is slightly off-topic, but perhaps we want to add functions like `fn is_null() -> bool` and `fn store_null()` to the `Edge`/`Slot` trait to aid debugging and help implement reference...

You need to still expose `flush_mutator` (or whatever we end up calling it) in the case where the `ScanStackRoots` work-packet is not created by the binding for VMs where it...

I think it's probably related to https://github.com/iv-org/invidious/pull/2895 as mentioned in my issue here: https://github.com/iv-org/invidious/issues/3104#issuecomment-1133956440.

Ah right. Thank you for the context! The system is a bit counterintuitive to me, so if possible is it agreeable add a config setting that controls this (with default...

This feature already exists as `ObjectReference.is_movable`: https://github.com/mmtk/mmtk-core/blob/1b9cfe40a8bd34fd0c7cc3681a4fe5e7cdb7e167/src/util/address.rs#L568-L571 If you want it to be specific to current GC, you could edit the `is_movable` function itself instead of creating a new function....

I mean technically `is_movable` is already incorrect since it does not take pinning bits/(transitively) pinned roots into account. I think this whole thing needs to be rethought.

Yes. I meant that a binding can't even rely on `is_movable` for accurate information since it basically only checks policy/build flags. I guess it depends on how fine-grained we want...

Hmm. `{Gen,Sticky}Immix` are still really bad in comparison to `Immix` for lusearch. That doesn't make any sense. Looking back at my old results, it seems like yeah, that was the...

The performance looks good (better than master for most things) but still not much better than Immix which is odd. But this PR does not need to solve all the...