Andreas Rumpf

Does this do the tokenization properly? A good test case is system.nim after line 145: ```nim proc new*[T](a: var ref T, finalizer: proc (x: ref T) {.nimcall.}) {. magic: "NewFinalize",...

> Quick Q: Do you expect the whole string "pass a finalizer...", including "{.nimcall.}", to be of a uniform color Yes, this one. There is not much complexity going on...

I'm keeping it open as it's so interesting but be aware we don't "officially" support TCC for reasons like this.

Afaik it's caused by nimlangserver using the lockfile feature. And for a reproducible build using a specific Nim version makes sense. That said I think nimlangserver shouldn't use a lockfile.

The code is ugly.

Why would I care how CPython does things... We need to live with this code for good, so clean it up so that the control flow doesn't look like it...

> Pointing out which part requires amend would be helpful. Well almost every line is a what-the-heck for me...

Sorry, but after looking at it again, I still think it's of unacceptable quality. Feel free to open a new PR that implements the feature without looking at how Python...

Still an issue.

Yeah, but I wonder why overloading resolution picks up ``consts``. :-)