Discuss code splicing

[Bert-Proesmans]

Hi, Youtube suggested me your older video about literate programming. (The blogpost link in the description doesn't work anymore, btw)

I was wondering if you considered implementing something alike code splicing?
Now, splicing can happen in two ways;

1. Leave markers in code to be replaced by the next code block
2. In the next code block, target a specific point of syntax in previous code block to insert new code after/before

Both approaches come with challenges; what to splice into, where to splice into
And, of course, how to do this in a minimalistic way. Preferable programming language agnostic?

Do you happen to have thought about another approach? Thoughts about more/different challenges?
Below I wrote naïve examples, those still need a lot of thought about syntax, comment-syntax compatibility, type-safety, splice scope, single-or multiple replacements, should the tool be language agnostic working as a text tool or syntax/semantic model aware...

I feel like with the Rust compiler for type inference, LSP for easing on pesky syntax requirements, and the current level of care for error messages, we are at a point in time where literate programming could be elevated to a next level.
EDIT; Also, we now have solid syntax parsers that support multiple programming languages at once.

---

Example, approach 1, see below. Note that the square brackets are for demonstration purposes and not part of the literate document;

[Descriptive text]
... We iterate over our list of numbers, and print them in the end

[Code]

let numbers = [-1,5,8];
for number in numbers.into_iter()<!-- Process numbers --!> {
  println!{"{:?}, number);
}

[Descriptive text]
... And for that reason we need the natural square root of these numbers

[Code]

<!-- include -->
use std::f64;
<!-- include -->

<!-- START Process numbers --!>
.filter(|n| *n > 0)
.map(|n| (*n as f64).sqrt())
<!-- END Process numbers --!>

[Code result]

use std::f64;

let numbers = [-1,5,8];
for number in numbers.into_iter()
.filter(|n| *n > 0)
.map(|n| (*n as f64).sqrt())
{
  println!{"{:?}, number);
}

---

Example, approach 2, see below. Note that the square brackets are for demonstration purposes and not part of the literate document;

[Descriptive text]
... We iterate over our list of numbers, and print them in the end

[Code]

let numbers = [-1,5,8];
for number in numbers.into_iter() {
  println!{"{:?}, number);
}

[Descriptive text]
... And for that reason we need the natural square root of these numbers

[Code]

<!-- @after include -->
use std::f64;
<!-- @after include -->

<!-- @after fn_call into_iter--!>
.filter(|n| *n > 0)
.map(|n| (*n as f64).sqrt())
<!--  @after fn_call into_iter --!>

[Code result]

use std::f64;

let numbers = [-1,5,8];
for number in numbers.into_iter()
.filter(|n| *n > 0)
.map(|n| (*n as f64).sqrt())
{
  println!{"{:?}, number);
}