Procedure call scopes

[keenanwoodall]

Issue: #4917

Managing "scopes" when declaring ui layouts can get a bit hairy.
Given an api where you can open() and close() ui elements, you have two options:

1. Just open() and close()

open("Root")
  open("Sidebar"); close()
  open("Editor")
    open("Tabs"); close()
    open("Text"); close()
  close()
close()

2. Use @(deferred_out)

@(deferred_out=close)
element :: proc () {
  open()
}

...

element("Root")
{
  { element("Sidebar") }
  element("Editor")
  {
    { element("Tabs") }
    { element("Text") }
  }
}

I don't think anything is inherently wrong with either, but it can start to feel a bit verbose.

Currently, opening a scope immediately after a function call is not legal syntax

// error: call does not return a value and cannot be used as a value
element() {
}

Describe the solution you'd like
Allow a scope to be opened after a function call when the proc has a @(deferred_x) attribute, or if necessary a new attribute, say @(scoped_x=)
This would behave like defer, but would defer to the end of the scope associated with the function call, calling the deferred proc immediately if no scope is present.
This would allow the UI in my example scenario to be defined like so:

element("Root") {
  element("Sidebar")
  element("Editor") {
    element("Tabs")
    element("Text")
  }
}

This is very much just a "nice to have", but I totally understand if it would conflict with Odin's principles and introduce too much "magic"/ambiguity. Maybe there's a better way to go about it, I just wanted to throw out the idea as it would make writing ui of in this way feel a bit smoother and remove some of the "noise"