Skip to content

Latest commit

 

History

History
180 lines (120 loc) · 7.7 KB

CONTRIBUTING-CODE.adoc

File metadata and controls

180 lines (120 loc) · 7.7 KB

Contributing Code

This guide will give you all the necessary information you need to become a successful Asciidoctor.js developer and code contributor!

Table of Contents

Introduction

Asciidoctor.js is direct port of Asciidoctor from Ruby to JavaScript using Opal, a Ruby-to-JavaScript transpiler. The project includes a Rake build script that executes the Opal compiler on the Asciidoctor source code to produce the Asciidoctor.js scripts. A npm build is used to assemble and prepare the distribution files, using the Rake build underneath.

Opal parses the Ruby code and any required libraries, then rewrites the code into JavaScript under the Opal namespace. The resulting JavaScript can be executed within any JavaScript runtime environment (such as a browser).

To interact with the generated code, you either invoke the JavaScript APIs directly, or you can invoke native JavaScript objects from within the Ruby code prior to compilation.

Setup

To build Asciidoctor.js, you’ll need some tools:

  • Node.js >= 0.12 and npm (we recommend nvm to manage multiple active Node.js versions)

  • Bower

  • Rake and Bundle

Start by cloning the source from GitHub:

$ git clone --recursive git://github.com/asciidoctor/asciidoctor.js

Next, switch to the asciidoctor.js directory and run npm’s install command followed by Bower’s install command:

$ cd asciidoctor.js
  npm install
  bower install

You’re now ready to build Asciidoctor.js.

Tip
 Opal.js, the Ruby runtime in JavaScript is available in bower_components/opal/opal/current/opal.min.js

Building

This section describes how to build Asciidoctor.js from source, execute the tests and try out the examples.

Building Asciidoctor.js

To build Asciidoctor.js, simply run the npm build script from the root of the project:

$ npm run build
Note
 The build task will make some minor code changes on the asciidoctor submodule. As you may know String are immutable in JavaScript, so we need to replace gsub! and sub! methods. These changes are made at build time to keep the Ruby code as fast as possible.

This command produces some files in the build directory:

build/

  • asciidoctor.js — includes core and extensions

  • asciidoctor-core.js — no extensions API

  • asciidoctor-extensions.js — extensions API only

  • asciidoctor-docbook.js — DocBook (4.5 and 5) converters

  • asciidoctor-all.js — core, extensions API and Opal
    (DocBook isn’t the main target of webapp. For this reason, we choose to keep it separate.)

build/npm/

  • asciidoctor-core.js — no extensions API, will automatically load the DocBook converters

  • asciidoctor-extensions.js — extensions API only

  • asciidoctor-docbook.js — DocBook (4.5 and 5) converters

Each file has a minified (.min.js) version.

You’ll see these scripts in action when you run the examples, described next.

Building and running the examples

To build the examples, simply run the npm examples task from the root of the project:

$ npm run examples

This command produces another JavaScript file in the build directory, asciidoctor_example.js. This script includes:

  • a string that contains an AsciiDoc source document

  • a call to the Asciidoctor API to render the content of that string to HTML

  • an event listener that inserts the generated HTML into the page

All the JavaScript in that file was generated from a Ruby script by Opal.

Point your browser at build/asciidoctor_example.html. You should see the Asciidoctor.js README. The content on the page was rendered from AsciiDoc by Asciidoctor.js when you loaded the page!

Building and running benchmark

Currently, you can run benchmark against 3 JavaScript environment:

  • node

  • phantomjs

  • jjs

The following command will run benchmark against Node.js:

$ npm run benchmark node

Once build/benchmark is initialized, you can run benchmark again (without building the project):

$ node build/benchmark/run.js

Debugging

Compiling a Ruby application to JavaScript and getting it to run is a process of eliminating fatal errors. When the JavaScript fails, the message isn’t always clear or even close to where things went wrong. The key to working through these failures is to use the browser’s JavaScript console.

Chrome / Chromium

Chrome (and Chromium) has a very intuitive JavaScript console. To open it, press Ctrl+Shift+J or right-click on the page, select Inspect Element from the context menu and click the Console tab.

When an error occurs in the JavaScript, Chrome will print the error message to the console. The error message is interactive. Click on the arrow at the start of the line to expand the call trace, as shown here:

error in chrome console

When you identify the entry you want to inspect, click the link to the source location. If you want to inspect the state, add a breakpoint and refresh the page.

Chrome tends to cache the JavaScript files too aggressively when running local scripts. Make a habit of holding down Ctrl when you click refresh to force Chrome to reload the JavaScript.

Another option is to start Chrome with the application cache disabled.

$ chrome --disable-application-cache

Firefox

Firefox also has a JavaScript console. To open it, press Ctrl+Shift+J or right-click on the page, select Inspect Element from the context menu and click the Web Console tab.

When an error occurs in the JavaScript, Firefox will print the error message to the console. Unlike Chrome, the error message is not interactive. Its power, instead, lies under the hood.

To see the call trace when an exception occurs, you need to configure the Debugger to pause on an exception. Click the Debugger tab, click the configuration gear icon in the upper right corner of that tab and click Pause on exceptions. Refresh the page and you’ll notice that the debugger has paused at the location in the source where the exception is thrown.

error in javascript debugger

The call trace is displayed as breadcrumb navigation, which you can use to jump through the stack. You can inspect the state at any location by looking through the panels on the right.

Modifications to Asciidoctor (for compatibility)

Compiling Asciidoctor to JavaScript currently requires some changes in Asciidoctor. The goal is to eventually eliminate all of these differences so that Asciidoctor can be compiled to JavaScript as is.

Here’s a list of some of the changes that are currently needed:

  • Named posix groups in regular expressions are replaced with their ASCII equivalent

    • JavaScript doesn’t support named posix groups, such as [[:alpha:]])

  • A shim library is needed to implement missing classes in Opal, such as File and Dir

  • All mutable String operations have been replaced with assignments (this is done at build time)

    • JavaScript doesn’t support mutable strings

  • $~[0] used in place of $& and $~[n] in place of $n after running a regular expression (where n is 1, 2, 3…​)

  • Opal doesn’t recognize modifiers on a regular expression (e.g., multiline)

  • Optional, non-matching capture groups resolve to empty string in gsub block in Firefox (see http://www.bennadel.com/blog/1916-different-browsers-use-different-non-matching-captured-regex-pattern-values.htm)

  • Assignments without a matching value are set to empty string instead of nil (in the following example, b is set to empty string)

    a, b = "value".split ',', 2