Easier way to render blocks without code duplication

[Giwayume]

If I have a template that uses lots of data from its struct, and I want to render individual blocks from the template (which also use the same data), there ends up being a lot of code duplication.

#[derive(Template)]
#[template(path = "my_template.html")]
pub struct MyTemplate<'a> {
    a: &'str,
    b: &'str,
    c: &'str,
    d: &'str,
    e: &'str,
    f: &'str,
    g: &'str,
}

#[derive(Template)]
#[template(path = "my_template.html", block = "sidebar")]
pub struct MySidebarTemplate<'a> {
    a: &'str,
    b: &'str,
    c: &'str,
    d: &'str,
    e: &'str,
    f: &'str,
    g: &'str,
}

MyTemplate { ... }.render()...
MySidebarTemplate { ... }.render()...

This is just a simple example, in a real application it's so much worse.

It would be nice if I could just take MyTemplate and tell it to render a specific block. Like:

MyTemplate { ... }.render_block("sidebar")

Conceptually, this could mean "render the full template, but discard everything from the HTML response that is not a part of the block", in terms of implementation. It may be a little more wasteful on CPU but it will save memory and clean up the code drastically.

Is something like this possible?

[Kijewski]

A similar idea was asked in askama-rs/askama#328. Maybe you can answer if you think my idea is worth exploring?

[Kijewski]

Implemented in askama-rs/askama#337. Will be contained in the next askama release.

[Giwayume]

I don't see rinja in the dependency tree of the askama main branch. Is that some upcoming dependency swap?

[GuillaumeGomez]

askama next release will be current rinja main branch.