benchmark_template.rs

benchmark_template.rs

use criterion::{black_box, criterion_group, criterion_main, Criterion};
use bartholomew::content::Content;
use std::path::PathBuf;


fn doit(n: u64) -> Result<(), anyhow::Error> {

    let  doc: String = r#"title = "The Goals of Bartholomew"
    description = "We have plans. Big plans. Actually, they're small plans."
    date = "2021-12-23T17:05:19Z"
    tags = ["bartholomew"]
    [extra]
    author = "Matt Butcher"
    author_page = "/author/butcher"
    ---


    Bartholomew is intended to be three things:

    1. A simple-to-use CMS-like system that feels like Jekyll or other static site generators
    2. A flexible system that is cheap to run, but easy to extend. (Cloud is expensive!)
    3. An exhibition of the value of WebAssembly on the server-side

    Static site generators are great for many things. But they are also limiting. To keep
    a site feeling fresh, you have to keep building and publishing the site because all
    elements are rendered at build time. If you want to write dynamic elements of any sort,
    you have to resort to JavaScript.

    But at the same time, these static site generators are super easy to use. Content is
    just written in plain text, and is rendered into HTML. Menus and navigation are
    built up automatically. And dealing with images and other files is as easy as dropping
    a file in a directory. We love that experience. So we tried to combine the best of the
    static site generator with a server-side technology.

    {{ alert "warning" "Bartholomew is a work in progress" }}"#.to_string();


    let script_path = PathBuf::from("./shortcodes/");

    let  contents: Content = doc.parse()?;
    let mut c = Content::new(contents.head, contents.body, script_path);
        let _out = c.render_markdown();

    Ok(())
}

fn criterion_benchmark(c: &mut Criterion) {
    c.bench_function("render content template", |b| b.iter(|| doit(black_box(1))));
}

criterion_group!(benches, criterion_benchmark);
criterion_main!(benches);

render content template time:   [459.67 us 460.79 us 461.96 us]                                    
                        change: [-1.8546% -0.9871% +0.0853%] (p = 0.04 < 0.05)
                        Change within noise threshold.
Found 3 outliers among 100 measurements (3.00%)
  1 (1.00%) high mild
  2 (2.00%) high severe

using two Content structs because the script path is different inside WASI.