niner

> Apparently in RakuAST with 6.d, there is no issue: but maybe there should be, as `m1` is a submethod. But `m1` is not in any way involved in constructing...

Thanks for the report! Makes perfect sense the way you describe it. Since you already seem to have done a lot of the investigation and know where in the code...

@timo Looks like you have done quite a bit of work in this area. Looks to me like the problem is well defined and the idea for how to solve...

That would make it impossible to look at the references file.

Sigils are shortcuts for type constraints (and default values). `@` means `Positional` (default value `Array`). `Int @` means `Positional[Int]` (default value `Array[Int]`). They are roughly equivalent. Thus `sub foo(Int @a)`...

I don't like this approach very much. It pushes more code into the VM which is the wrong direction. The VM already does too much. Every complex operation pushed into...

> That would be nice indeed. However, as far as I understand it, the JSON standard is a finite standard that does not allow any non-standard graphemes. Thus, it can...

> The performance benefit is indeed gone. But what is the point of changing it back just for the fact that we can? What happened to "if it ain't broke,...

> I thought it was ok to use internal interfaces inside the core? I was talking about JSON::Fast

This does another file system access though for the stat() call. Is that really faster than resizing the buffer as needed?