Development

Watch mode

Call website.watch(data) instead of website.build(data) to start a file-watching loop. Hauchiwa runs an initial build, then re-runs affected tasks whenever a watched file changes.

match args.mode {
    Mode::Build => website.build(data)?,
    Mode::Watch => website.watch(data)?,
}

Each loader and task declares which file patterns it watches. When a file changes, only the tasks that depend on that file - directly or transitively - are re-run. Everything else is served from cache.

Watch mode also starts a WebSocket server on a free port. Inject the live-reload script into your HTML to have the browser refresh automatically after each rebuild:

config.task().merge(|ctx, deps| {
    let refresh = ctx.env.get_refresh_script();  // Some(...) in watch mode, None otherwise
    let script_tag = refresh.unwrap_or_default();
    // include script_tag in your <head>
    Ok(output)
});

The server feature flag must be enabled for the HTTP development server. The live feature flag must be enabled for the WebSocket live-reload.

Logging

Enable the logging feature and call hauchiwa::init_logging() at the start of main to get structured log output with ANSI colours, uptime timestamps, and progress bars for parallel tasks:

[dependencies]
hauchiwa = { version = "*", features = ["logging"] }

fn main() -> anyhow::Result<()> {
    hauchiwa::init_logging()?;

    let mut site = Blueprint::<()>::new()
        // ...
        .finish();

    site.build(())?;
    Ok(())
}

Without this feature, Hauchiwa emits tracing events but does not install a subscriber - you can bring your own if you have an existing logging setup.

Diagnostics

website.build() returns a Diagnostics value containing per-task execution times. Two built-in renderers are available for visualising where time is spent.

Mermaid diagram

render_mermaid returns a Mermaid graph string with nodes colour-coded by duration (green = fast, yellow = moderate, red = slow, blue = cached):

let diagnostics = website.build(data)?;
println!("{}", diagnostics.render_mermaid(&website));

Paste the output into mermaid.live to see the graph.

Waterfall chart

render_waterfall returns an SVG timeline showing tasks laid out in parallel lanes with duration labels - useful for spotting bottlenecks in large graphs: