anseki/skeleton-loader

Loader module for webpack to execute your custom procedure. It works as your custom loader.

skeleton-loader

Loader module for webpack to execute your custom procedure. It works as your custom loader.

By default, skeleton-loader only outputs the input content. When you specify a function, skeleton-loader executes your function with the input content, and outputs its result. The function does something, it might edit the content, it might parse the content and indicate something in a console, it might do anything else.

That is, you can specify a function in webpack configuration instead of writing new custom loader.

skeleton-loader is useful when:

• You couldn't find a loader you want.
• You don't want to write a special loader for your project.
• You want to add something to the result of another loader.
• You want to do additional editing.
• etc.

For example:

// webpack.config.js
module.exports = {
  entry: './app.js',
  output: {
    filename: 'bundle.js'
  },
  module: {
    rules: [{
      test: /\.js$/,
      loader: 'skeleton-loader',
      options: {
        procedure: function(content) {
          // Change the input content, and output it.
          return (content + '').replace(/foo/g, 'bar');
        }
      }
    }]
  }
};

// webpack.config.js
// ...
test: /\.html$/,
// ...
// skeleton-loader options
options: {
  procedure: function(content) {
    // Remove all elements for testing from HTML.
    return (content + '').replace(/<div class="test">[^]*?<\/div>/g, '');
  },
  toCode: true
}

// webpack.config.js
// ...
test: /\.json$/,
// ...
// skeleton-loader options
options: {
  procedure: function(content) {
    var appConfig = JSON.parse(content);
    // Check and change JSON.
    console.log(appConfig.foo);
    appConfig.bar = 'PUBLISH';
    return appConfig;
  },
  toCode: true
}

// webpack.config.js
// ...
// skeleton-loader options
options: {
  // Asynchronous mode
  procedure: function(content, options, callback) {
    setTimeout(function() {
      callback(null, 'Edited: ' + content);
    }, 5000);
  }
}

Installation

npm install --save-dev skeleton-loader

Usage

Documentation:

• Loaders
• Using loaders (for webpack v1)

Options

You can specify options via query parameters or an options (or skeletonLoader for webpack v1) object in webpack configuration.

procedure

Type: function
Default: undefined

A function to do something with the input content. The result of the procedure is output.
The following arguments are passed to the procedure:

• content
  The content of the resource file as string, or something that is passed from previous loader. That is, if another loader is chained in loaders list, the content that is passed from that loader might not be string.
• options
  Reference to current options. This might contain either or both of sourceMap and meta if those are passed from previous loader. Also, it might contain options.resourceOptions.
• callback
  A callback function for asynchronous mode. If the procedure doesn't receive the callback, the loader works in synchronous mode.

In the procedure function, this refers to the loader context. It has resourcePath, query, etc. See: https://webpack.js.org/api/loaders/#the-loader-context

The result of the procedure can be any type such as string, Object, null, undefined, etc.
For example:

// app.js
var config = require('config.json');

// webpack.config.js
// ...
// skeleton-loader options
options: {
  procedure: function(config) {
    if (initialize) {
      return; // make config be undefined
    }
    return process.env.NODE_ENV === 'production' ? config : {name: 'DUMMY'}; // data for test
  }
}

In synchronous mode, the procedure has to return the content. The content is output as JavaScript code, or passed to next loader if it is chained.

For example:

// webpack.config.js
// ...
// skeleton-loader options
options: {
  procedure: function(content, options) {

    // Do something with content.
    console.log('Size: ' + content.length);
    content = (content + '').replace(/foo/g, 'bar'); // content might be not string.

    // Check the resource file by using context.
    if (this.resourcePath === '/abc/resource.js') {

      // Change current option.
      options.toCode = true;
    }

    // Return the content to output.
    return content;
  }
}

If the procedure receives the callback, the loader works in asynchronous mode. To return either or both of SourceMap and meta data, it must be asynchronous mode.
In asynchronous mode, the procedure has to call the callback when it finished.

The callback accepts the following arguments:

• error
  An error object, when your procedure failed.
• content
  The content that is output as JavaScript code, or passed to next loader if it is chained. This can be any type such as string, Object, null, undefined, etc.
• sourceMap
  An optional value SourceMap as JavaScript object that is output, or passed to next loader if it is chained.
• meta
  An optional value that can be anything and is output, or passed to next loader if it is chained.

For example:

// webpack.config.js
// ...
// skeleton-loader options
options: {
  procedure: function(content, options, callback) { // Switches to asynchronous mode
    // Do something asynchronously.
    require('fs').readFile('data.txt', function(error, data) {
      if (error) {
        // Failed
        callback(error);
      } else {
        // Done
        callback(null, data + content);
      }
    });
  }
}

options.resourceOptions

The options argument has resourceOptions property if a query string is specified with the resource file, and it is an object that is parsed query string.
This is useful for specifying additional parameters when importing the resource files. For example, you can specify the behavior with resource files.

var
  all = require('file.html'),
  noHead = require('file.html?removeHead=yes'),;

// webpack.config.js
// ...
// skeleton-loader options
options: {
  procedure: function(content, options) {
    if (options.resourceOptions && options.resourceOptions.removeHead) {
      content = content.replace(/<head[^]*?<\/head>/, ''); // Remove <head>
    }
    return content;
  }
}

The query string is parsed in the same way as loader-utils.

toCode

Type: boolean
Default: false

When the content is not JavaScript code (e.g. HTML, CSS, JSON, etc.), a loader that is specified as a final loader has to convert the content to JavaScript code and output it to allow another code to import the content.
If true is specified for toCode option, the content is converted to JavaScript code.
If the loader is specified as not a final loader, this option is ignored (i.e. the content is not converted, and it is passed to next loader).

For example:

// webpack.config.js
module.exports = {
  // ...
  module: {
    rules: [
      // HTML code is converted to JavaScript string.
      // It works same as raw-loader.
      {test: /\.html$/, loader: 'skeleton-loader?toCode=true'},

      // JSON data is converted to JavaScript object.
      // It works same as json-loader.
      {
        test: /\.json$/,
        loader: 'skeleton-loader',
        options: {
          procedure: function(content) { return JSON.parse(content); },
          toCode: true
        }
      }
    ]
  }
};

// app.js
var html = require('file.html');
element.innerHTML = html;

var obj = require('file.json');
console.log(obj.array1[3]);

cacheable

Type: boolean
Default: true

Make the result cacheable.
A cacheable loader must have a deterministic result, when inputs and dependencies haven't changed. This means the loader shouldn't have other dependencies than specified with context.addDependency.
Note that the default value is true.