Gábor Lehel

> Also, we have no other choice if we want to remain backwards compaible. Safe const code can already create references to fn. (Hmm. I thought an analogous case would...

A simple and boring option is `IsGraph` (which of course doesn't address the "what if it isn't" issue from @snowleopard's previous comments).

@jeehoonkang Very interesting, thanks for sharing! Two questions occurred to me: 1. At one point you write: > We compare the performance of PEBR with that of the NR (noreclamation)...

Yeah, figuring out how to abstract over `out` and `move` is much less obvious than `mut` and shared. Possibly we would have to take a capability-based perspective on it (see...

As I wrote [here](https://github.com/rust-lang/rust/issues/34511#issuecomment-373516961) I think this use case would be better addressed by something modelled after how closures work.

Cross-pasting from the duplicate issue I opened: > There is some previous discussion about HKTs, type lambdas, type inference, and so on in the comments on the [default type parameters...

This is besides the main point, but is there some reason this is not a suitable alternative to recursive typedefs? ```rust struct Instruction { fun: fn(&mut Stack, &[Instruction]) } ```

One useful piece of data would be what proportion of `catch_unwind` calls are/aren't using `AssertUnwindSafe`. The places where it's *not* being used are where the mechanism is providing value, by...

@LegionMammal978 > Why should this logical implication hold? As I understand it, `Send` and `Sync` mean that an object can be sent or shared across threads without any of its...

> This reasoning works for atomics, but I don't think it follows for `Mutex`es. I'm not sure if you were just pointing out a technical detail, or if you intended...