Duster.js - Watch & precompile dust.js templates

A simple Node script duster.js to watch a directory of .dust templates and compile them into .js files which can be included into an HTML file.

Why duster.js? Autocompile your templates to use dust.js in the browser

The dust.js documentation does not mentioned a clear way to work with dust templates in a purely client-side approach, instead focusing on server-side node.js applications.

For my backbone.js app, the only option was to include the dust-full.js file and compile the templates on each browser load. The file is much larger than the normal dust-core.js and this approach provides no extra value over other templating solutions (performance, browser caching, external file management etc).

So I wrote a script to pre-compile dust.js files whenever they are modified in a folder.

Install

npm install -g dusterjs

Usage

The most simple usage: compile a bunch of templates (the views directory) into a single JS file (templates.js). Only files with the .dust extension will be compiled.

$ duster views templates.js

To watch the input directories for changes and recompile, use the --watch (-w) option:

$ duster -w views templates.js

To generate one .js file per dust template, use the --no-concat option (for example, views/home.dust would become views-js/home.js and views/shared/nav.dust would become views-js/shared/nav.js):

$ duster --no-concat views views-js

A complete list of options:

  --verbose, -v   verbose mode
  --watch, -w     watch input directory(s) for changes
  --concat, -c    concatenate all compiled templates into one javascript file (turn off with --no-concat)  [string]  [default: true]
  --minify, -m    minify all the compiled templates (turn off with --no-minify)                            [default: true]
  --interval, -i  set the polling interval (in milliseconds)                                               [default: 100]
  --help          show usage information and exit
  --version       show program version and exit

As a library

Duster.js is now also available as a library.

duster.compileFile (fileName, callback)

Passes to callback the JavaScript resulting from compiling the dust template.

duster.compileFileSync (fileName)

Returns the JavaScript resulting from compiling the dust template.

duster.minify (js)

Returns a minified version of js (a piece of JavaScript code). Uses uglifyjs.

duster.compileAll (inputs, output, options)

Compiles all .dust templates in the array of directory names inputs, and writes them to output (which is either a file name or directory name, depending on whether or not options.concat is truthy). If options.minify is truthy, the javascript will be compressed. If options.concat is truthy, the javascript will all be written to one file (specified by output), otherwise each .dust template will get its own .js file in the directory specified by output.

If an error occurs, an array of strings is thrown. If no errors occur, an object is returned with the following properties:

{
    bytesRead: 54983, // The total number of JavaScript bytes produced by compiling the dust templates
    bytesWritten: 46563 // The total number of JavaScript bytes after compressing
}

duster.watch (inputs, output, [options], [callback])

Watches each of the directories in inputs recursively for changes. If any files or directories are added, changed or removed, the output is recompiled. options is passed to compileAll. callback should be a function taking two arguments; the first argument is the error (if an error occurred), the second argument is the result of compileAll.

duster.watch(["views", "../views"], "js/templates.js", {}, function (err, results) {
    if (err) 
        return console.error("** Error:", err);
    console.log("Templates updated at", new Date().toLocaleTimeString());
});

TODO

It would be more efficient when using --watch and --no-concat to only compile the one file that has changed (currently, it recompiles them all).
It would also be more efficient to do this in general with --no-concat, by looking at timestamps.

More information

Linkedin wrote a dust.js tutorial: https://github.com/linkedin/dustjs/wiki/Dust-Tutorial

by Dan McGrady http://dmix.ca

lee-houghton/dusterjs