Joe Eli McIlvain

Here's a more minimal repro which segfaults without using partial application (it uses an explicit `class val` to hold the actor) and a bit less of the bouncing around between...

I have [some ideas about what is going on here](https://ponylang.zulipchat.com/#narrow/stream/190365-runtime/topic/pony_traceunknown.20shallow.20traces.20immutables) but need to discuss further with Sean.

To summarize what we discussed in [the Zulip thread](https://ponylang.zulipchat.com/login/#narrow/stream/190365-runtime/topic/pony_traceunknown.20shallow.20traces.20immutables): The current Pony runtime has a correctness bug due to what is usually a valid optimization, but in this case is...

In the sync call today, Jason says he has a fix in mind: Basically, ban generics whose instantiations result in double-ephemeral in the public signature.

It seems overly restrictive to prevent any consuming of both `x` and `x2` during the `match` block, because there are some examples of safe code that does that. Wouldn't it...

> Could you link some of the samples where the consumes you've mentioned are used? I didn't necessarily have specific samples in mind - I was just speaking hypothetically about...

Yeah, I'm pretty convinced at this point that we won't be able to get around this until we stop treating a `match` clause as a new `let` binding... Unless we...

Here's an example of iso invalidation that Jason and I were talking about on the sync call: ```pony class UsesNoConsume[T] // (for the case where T is an iso) var...

Note that instead of assigning the `value = None` result to a variable `t2`, you could also pass it as an argument to a function that does some unknown things...

In the sync call today, I said it's my position that we basically have two options here: - :one: "Reinvent" the `iso` and aliasing part of the type system so...