Description
My use case with HTMX is that a response may either include the full page (extending base.html) or just a single block content depending on a request header. Right now, using Minijinja, this is implemented centrally in a custom middleware which decides what to render.
I'd love to switch to Rinja, but right now this is not possible without having two versions of every struct, modifying every endpoint to either return the full- or the partial template struct, with the latter one annotated with #[attribute(block = "content")].
Would you consider extending the Template derive macro to be able to generate methods on the struct which allow rendering each block individually? I'm thinking of something like:
{% block title %}Hello {{title}}{% endblock %}
{% block content %}...{% endblock %}#[derive(Template)]
#[attribute(file = "foo.html", generate_block_methods = "true")]
struct Foo {
title: String
}
// Generated
impl Template for Foo {..}
impl Foo {
fn render_block_title(&self) -> Result<String> { ... }
fn render_block_content(&self) -> Result<String> { ... }
}The methods could even be used in the existing render methods (possibly annotated #[inline]) to avoid generating redundant code.
Another approach could be extending Template and DynTemplate with a fn render_block(&self, block: &str) -> ... method.