Noa

Ah ok, there were parallel `make` builds occuring, and when I gave each job a different `XARGO_HOME` that resolved the issue. Maybe `flock` is broken? Edit: actually I have no...

There will likely be an `io::ErrorKind::NotSupported` added in Rust 1.52 (rust-lang/rust#78880), I think that be a good mapping for `nix::Error::UnsupportedOperation` if `From for io::Error` were ever to be added again.

Also, I'd like to make a case for why that impl should be added again: `nix::Error` is basically a subset of `io::Error`. While `io::Error` can hold an arbitrary user error...

I've got a wip branch changing around half so far of the uses of RawFd to the io-safe versions.

> Note: This API exists so that it's possible to e.g. clone the underlying Rc. Finding this late, but if the `IntoRaw`-trait-based approach was going to be taken, there could...

@Shadow53 I believe `Rejection::find` requires `'static` because `::downcast()` does, otherwise, would `&'a T` and `&'b T` be the same types or different types? I do have my own issue with...

This might count as a breaking change, but what if there was a `#( #var )?` optionality specifier? Basically the same as `#()*` but expecting the variable to be an...

I'll check it out, thanks! I might fork it and experiment with a method on Cc to decrement refcount with a custom drop function, since we're running into UB with...

Oh, I was hoping to use it as an IndexMap; ModuleType/ComponentType also have `imports`/`exports` public as a IndexMap. Maybe if the `Vec` was wrapped in a `ImportType` struct or something?

Or an `ImportsMap` that works similar to a multimap or something, since that seems to be what it's functioning as.