feat!: use jinja for templating

[seblyng]

AI-assisted🤖

Seems to work but I want to look through the code myself as well.
Should also decide if we want to do this since it will be a breaking change.

I am perfectly fine with this since it is probably only us who are using this. We can also just implement something super simple or just ask AI to convert all the old potential http files to use the new syntax. I would expect that to be very easy

[seblyng]

This was a clippy lint if you wondered

1. clippy: large size difference between variants
   for further information visit https://rust-lang.github.io/rust-clippy/rust-1.91.0/index.html#large_enum_variant
   `#[warn(clippy::large_enum_variant)]` on by default [large_enum_variant]
   output.rs:61:5: the largest variant contains at least 328 bytes
   output.rs:60:5: the second-largest variant contains at least 24 bytes
   output.rs:61:13: consider boxing the large fields or introducing indirection in some other way to reduce the total size of the enum: `Box<HttpRequestInfo>`

[seblyng]

We now have something very powerful like jinja (if we merge this), and we can potentially avoid some hardcoded restrictions we have. For example, we have a requirement now for a list to be like this:

[[label]]
name = "foo"
value = "foo"

[[label]]
name = "bar"
value = "bar"

So name and value keys.

I propose that we at least keep it like this for this PR to make the scope a bit smaller, and then potentially do another breaking change if we want.

We can avoid to "map" it internally (like we do in for example the list_option_to_string and other places) and instead just do it like this in the template:

["{{ label | select_multiple | map(attribute="value") | join('", "') }}"]

I haven't thought about potential pitfalls with this though