Techcable

Yeah I figured it might be a good way to start contributing to the code base. I wanna help make this run faster :)

[syn](https://github.com/dtolnay/syn) works very well.

I think classes should be value-based by default and should have to explicitly request reference schematics (forcing heap-allocation). If this is done, the compiler is free to stack-allocate and copy...

You only need to do the optimization for small immutable objects, which are the most common.

> The problem is that the [jedi xontrib](https://github.com/xonsh/xonsh/blob/0ba5bec86f79a9a1d517862e1c98e3979898e291/xontrib/jedi.py) is not aware of any xonsh syntax, but treats the input as python code. > During parsing, while creating the AST, the...

> @Techcable are you working on this or should I give it a try? Feel free to give it a try! I haven't even started yet :)

Personally, I find the Python PEG grammar easier to read than the ply grammar xonsh uses currently. Maybe you could just copy the Python PEG grammar (maybe using pgen2), and...

Maybe use the [`archery`](https://docs.rs/archery/) crate? That is what [`rpds`](https://docs.rs/rpds) uses.

I still have this exact same problem on stable, 1.55 and latest nightly 😦 I am building on alpine in a VM.

On my machine, running `cargo miri test -p without-alloc` on the `master` branch (commit 3bd012468b8305f14f2) fails with the same message as it does for the PR: Backtrace for `cargo miri...