pulsar/docs/creating-a-package.md
2013-10-16 17:05:26 -07:00

21 KiB

Authoring Packages

Packages are at the core of Atom. Nearly everything outside of the main editor is handled by a package. That includes "core" pieces like the file tree, status bar and more.

A package can contain a variety of different resource types to change Atom's behavior. The basic package layout is as follows (though not every package will have all of these directories):

my-package/
  grammars/
  keymaps/
  lib/
  menus/
  spec/
  snippets/
  stylesheets/
  index.coffee
  package.json

package.json

Similar to npm packages, Atom packages can contain a package.json file in their top-level directory. This file contains metadata about the package, such as the path to its "main" module, library dependencies, and manifests specifying the order in which its resources should be loaded.

In addition to the regular npm package.json keys available, Atom package.json files have their own additions.

  • main (Required): the path to the CoffeeScript file that's the entry point to your package
  • stylesheets (Optional): an Array of Strings identifying the order of the stylesheets your package needs to load. If not specified, stylesheets in the stylesheets directory are added alphabetically.
  • keymaps(Optional): an Array of Strings identifying the order of the key mappings your package needs to load. If not specified, mappings in the keymaps directory are added alphabetically.
  • menus(Optional): an Array of Strings identifying the order of the menu mappings your package needs to load. If not specified, mappings in the keymap directory are added alphabetically.
  • snippets (Optional): an Array of Strings identifying the order of the snippets your package needs to load. If not specified, snippets in the snippets directory are added alphabetically.
  • activationEvents (Optional): an Array of Strings identifying events that trigger your package's activation. You can delay the loading of your package until one of these events is trigged.

Source Code

If you want to extend Atom's behavior, your package should contain a single top-level module, which you export from index.coffee (or whichever file is indicated by the main key in your package.json file). The remainder of your code should be placed in the lib directory, and required from your top-level file.

Your package's top-level module is a singleton object that manages the lifecycle of your extensions to Atom. Even if your package creates ten different views and appends them to different parts of the DOM, it's all managed from your top-level object.

Your package's top-level module should implement the following methods:

  • activate(rootView, state): This required method is called when your package is loaded. It is always passed the window's global rootView, and is sometimes passed state data if the window has been reloaded and your module implements the serialize method. Use this to do initialization work when your package is started (like setting up DOM elements or binding events).

  • serialize(): This optional method is called when the window is shutting down, allowing you to return JSON to represent the state of your component. When the window is later restored, the data you returned is passed to your module's activate method so you can restore your view to where the user left off.

  • deactivate(): This optional method is called when the window is shutting down. If your package is watching any files or holding external resources in any other way, release them here. If you're just subscribing to things on window, you don't need to worry because that's getting torn down anyway.

Simple Package Code

my-package/
  package.json # optional
  index.coffee
  lib/
    my-package.coffee

index.coffee:

module.exports = require "./lib/my-package"

my-package/my-package.coffee:

module.exports =
  activate: (rootView, state) -> # ...
  deactivate: -> # ...
  serialize: -> # ...

Beyond this simple contract, your package has access to Atom's API. Be aware that since we are early in development, APIs are subject to change and we have not yet established clear boundaries between what is public and what is private. Also, please collaborate with us if you need an API that doesn't exist. Our goal is to build out Atom's API organically based on the needs of package authors like you.

Check out wrap-guide for a simple example of Atom's package API in action.

Stylesheets

Stylesheets for your package should be placed in the stylesheets directory. Any stylesheets in this directory will be loaded and attached to the DOM when your package is activated. Stylesheets can be written as CSS or LESS (but LESS is recommended).

Ideally you will not need much in the way of styling. We've provided a standard set of components. You can view all components by using the command palette (cmd-p) and searching for "styleguide" or just cmd+ctrl+G.

If you do need styling, we try to keep only structural styles in the package stylesheets. Colors and sizing should be taken from the active theme's ui-variables.less. If you follow this guideline, your package will look good out of the box with any theme!

An optional stylesheets array in your package.json can list the stylesheets by name to specify a loading order; otherwise, stylesheets are loaded alphabetically.

Keymaps

'.tree-view-scroller':
  'ctrl-V': 'changer:magic'

It's recommended that you provide key bindings for commonly used actions for your extension, especially if you're also adding a new command.

Keymaps are placed in the keymaps subdirectory. By default, all keymaps are loaded in alphabetical order. An optional keymaps array in your package.json can specify which keymaps to load and in what order.

See the main keymaps documentation for more detailed information on how keymaps work.

Menus

Menus are placed in the menus subdirectory. By default, all menus are loaded in alphabetical order. An optional menus array in your package.json can specify which menus to load and in what order.

Application Menu

'menu': [
  {
    'label': 'Packages'
    'submenu': [
      {
        'label': 'My Package'
        'submenu': [
          {
            'label': 'Toggle'
            'command': 'my-package:toggle'
          }
        ]
      }
    ]
  }
]

It's recommended that you create an application menu item for common actions with your package that aren't tied to a specific element.

To add your own item to the application menu simply create a top level menu key in any menu configuration file in menus (since the above is CSON it should end with .cson)

The menu templates you specify are merged with all other templates provided by other packages in the order which they were loaded.

Context Menu

'context-menu':
  '.tree-view':
    'Add file': 'tree-view:add-file'
  '#root-view':
    'Inspect Element': 'core:inspect'

It's recommended to specify a context menu item for commands that are linked to specific parts of the interface, like adding a file in the tree-view.

To add your own item to the application menu simply create a top level context-menu key in any menu configuration file in menus (since the above is CSON it should end with .cson)

Context menus are created by determining which element was selected and then adding all of the menu items whose selectors match that element (in the order which they were loaded). The process is then repeated for the elements until reaching the top of the DOM tree.

In the example above, the Add file item will only appear when the focused item or one of its parents has the tree-view class applied to it.

Snippets

An extension can supply language snippets in the snippets directory which allows the user to enter repetitive text quickly.

".source.coffee .specs":
  "Expect":
    prefix: "ex"
    body: "expect($1).to$2"
  "Describe":
    prefix: "de"
    body: """
      describe "${1:description}", ->
        ${2:body}
    """

A snippets file contains scope selectors at its top level (.source.coffee .spec). Each scope selector contains a hash of snippets keyed by their name (Expect, Describe). Each snippet also specifies a prefix and a body key. The prefix represents the first few letters to type before hitting the tab key to autocomplete. The body defines the autofilled text. You can use placeholders like $1, $2, to indicate regions in the body the user can navigate to every time they hit tab.

All files in the directory are automatically loaded, unless the package.json supplies a snippets key. As with all scoped items, snippets loaded later take precedence over earlier snippets when two snippets match a scope with the same specificity.

Language Grammars

If you're developing a new language grammar, you'll want to place your file in the grammars directory. Each grammar is a pairing of two keys, match and captures. match is a regular expression identifying the pattern to highlight, while captures is an object representing what to do with each matching group.

For example:

{
  'match': '(?:^|\\s)(__[^_]+__)'
  'captures':
    '1': 'name': 'markup.bold.gfm'
}

This indicates that the first matching capture ((__[^_]+__)) should have the markup.bold.gfm token applied to it.

To capture a single group, simply use the name key instead:

{
  'match': '^#{1,6}\\s+.+$'
  'name': 'markup.heading.gfm'
}

This indicates that Markdown header lines (#, ##, ###) should be applied with the markup.heading.gfm token.

More information about the significance of these tokens can be found in section 12.4 of the TextMate Manual.

Your grammar should also include a filetypes array, which is a list of file extensions your grammar supports:

'fileTypes': [
  'markdown'
  'md'
  'mkd'
  'mkdown'
  'ron'
]

Bundle External Resources

It's common to ship external resources like images and fonts in the package, to make it easy to reference the resources in HTML or CSS, you can use the atom protocol URLs to load resources in the package.

The URLs should be in the format of atom://package-name/relative-path-to-package-of-resource, for example, the atom://image-view/images/transparent-background.gif would be equivablent to ~/.atom/packages/image-view/images/transparent-background.gif.

You can also use the atom protocol URLs in themes.

Writing Tests

Your package should have tests, and if they're placed in the spec directory, they can be run by Atom.

Under the hood, Jasmine is being used to execute the tests, so you can assume that any DSL available there is available to your package as well.

FIXME: Explain the following

  • jasmine
  • jasmine-focused
  • spec/fixtures and global.project
  • setTimeout
  • whatever else is different in spec-helper

Running tests

Once you've got your test suite written, the recommended way to run it is apm test. apm test prints its output to the console and returns the proper status code depending on whether tests passed or failed.

Publishing

Atom bundles a command line utility called apm which can be used to publish Atom packages to the public registry.

Once your package is written and ready for distribution you can run the following to publish your package:

cd my-package
apm publish minor

This will update your package.json to have a new minor version, commit the change, create a new Git tag, and then upload the package to the registry.

Run apm help publish to see all the available options and apm help to see all the other available commands.

Full Example

Let's take a look at creating our first package.

To get started hit cmd-p, and start typing "Package Generator." to generate the package. Once you select the package generator command, it'll ask you for a name for your new package. Let's call ours changer.

Now, changer is going to have a default set of folders and files created for us. Hit cmd-r to reload Atom, then hit cmd-p and start typing "Changer." You'll see a new Changer:Toggle command which, if selected, pops up a new message. So far, so good!

In order to demonstrate the capabilities of Atom and its API, our Changer plugin is going to do two things:

  1. It'll show only modified files in the file tree
  2. It'll append a new pane to the editor with some information about the modified files

Let's get started!

Changing Keybindings and Commands

Since Changer is primarily concerned with the file tree, let's write a key binding that works only when the tree is focused. Instead of using the default toggle, our keybinding executes a new command called magic.

keymaps/changer.cson can easily become this:

'.tree-view-scroller':
  'ctrl-V': 'changer:magic'

Notice that the keybinding is called ctrl-V--that's actually ctrl-shift-v. You can use capital letters to denote using shift for your binding.

.tree-view-scroller represents the parent container for the tree view. Keybindings only work within the context of where they're entered. For example, hitting ctrl-V anywhere other than tree won't do anything. You can map to body if you want to scope to anywhere in Atom, or just .editor for the editor portion.

To bind keybindings to a command, we'll use the rootView.command method. This takes a command name and executes a function in the code. For example:

rootView.command "changer:magic", => @magic()

It's common practice to namespace your commands with your package name, and separate it with a colon (:). Rename the existing toggle method to magic to get the binding to work.

Reload the editor, click on the tree, hit your keybinding, and...nothing happens! What the heck?!

Open up the package.json file, and notice the key that says activationEvents. Basically, this tells Atom to not load a package until it hears a certain event. Let's change the event to changer:magic and reload the editor.

Hitting the key binding on the tree now works!

Working with Styles

The next step is to hide elements in the tree that aren't modified. To do that, we'll first try and get a list of files that have not changed.

All packages are able to use jQuery in their code. In fact, we have a list of some of the bundled libraries Atom provides by default.

Let's bring in jQuery:

{$} = require 'atom'

Now, we can query the tree to get us a list of every file that wasn't modified:

magic: ->
  $('ol.entries li').each (i, el) ->
    if !$(el).hasClass("modified")
      console.log el

You can access the dev console by hitting alt-cmd-i. When we execute the changer:magic command, the browser console lists the items that are not being modified. Let's add a class to each of these elements called hide-me:

magic: ->
  $('ol.entries li').each (i, el) ->
    if !$(el).hasClass("modified")
      $(el).addClass("hide-me")

With our newly added class, we can manipulate the visibility of the elements with a simple stylesheet. Open up changer.css in the stylesheets directory, and add a single entry:

ol.entries .hide-me {
  display: none;
}

Refresh atom, and run the changer command. You'll see all the non-changed files disappear from the tree. There are a number of ways you can get the list back; let's just naively iterate over the same elements and remove the class:

magic: ->
  $('ol.entries li').each (i, el) ->
    if !$(el).hasClass("modified")
      if !$(el).hasClass("hide-me")
        $(el).addClass("hide-me")
      else
        $(el).removeClass("hide-me")

Creating a New Pane

The next goal of this package is to append a pane to the Atom editor that lists some information about the modified files.

To do that, we're going to first create a new class method called content. Every package that extends from the View class can provide an optional class method called content. The content method constructs the DOM that your package uses as its UI. The principals of content are built entirely on SpacePen, which we'll touch upon only briefly here.

Our display will simply be an unordered list of the file names, and their modified times. Let's start by carving out a div to hold the filenames:

@content: ->
  @div class: 'modified-files-container', =>
    @ul class: 'modified-files-list', outlet: 'modifiedFilesList', =>
      @li 'Test'
      @li 'Test2'

You can add any HTML5 attribute you like. outlet names the variable your package can uses to manipulate the element directly. The fat pipe (=>) indicates that the next set are nested children.

We'll add one more line to magic to make this pane appear:

rootView.vertical.append(this)

If you hit the key command, you'll see a box appear right underneath the editor. Success!

Before we populate this, let's apply some logic to toggle the pane off and on, just like we did with the tree view:

# toggles the pane
if @hasParent()
  rootView.vertical.children().last().remove()
else
  rootView.vertical.append(this)

There are about a hundred different ways to toggle a pane on and off, and this might not be the most efficient one. If you know your package needs to be toggled on and off more freely, it might be better to draw the interface during the initialization, then immediately call hide() on the element to remove it from the view. You can then swap between show() and hide(), and instead of forcing Atom to add and remove the element as we're doing here, it'll just set a CSS property to control your package's visibility.

You might have noticed that our two li elements aren't showing up. Let's set a color on them so that they pop. Open up changer.css and add this CSS:

ul.modified-files-list {
  color: white;
}

Refresh Atom, hit the key combo, and see your brilliantly white test list.

Calling Node.js Code

Since Atom is built on top of Node.js, you can call any of its libraries, including other modules that your package requires.

We'll iterate through our resulting tree, and construct the path to our modified file based on its depth in the tree:

path = require 'path'

# ...

modifiedFiles = []
# for each single entry...
$('ol.entries li.file.modified span.name').each (i, el) ->
  filePath = []
  # ...grab its name...
  filePath.unshift($(el).text())

  # ... then find its parent directories, and grab their names
  parents = $(el).parents('.directory.modified')
  parents.each (i, el) ->
    filePath.unshift($(el).find('div.header span.name').eq(0).text())

  modifiedFilePath = path.join(project.rootDirectory.path, filePath.join(path.sep))
  modifiedFiles.push modifiedFilePath

modifiedFiles is an array containing a list of our modified files. We're also using the node.js path library to get the proper directory separator for our system.

Let's remove the two @li elements we added in @content, so that we can populate our modifiedFilesList with real information. We'll do that by iterating over modifiedFiles, accessing a file's last modified time, and appending it to modifiedFilesList:

# toggles the pane
if @hasParent()
  rootView.vertical.children().last().remove()
else
  for file in modifiedFiles
    stat = fs.lstatSync(file)
    mtime = stat.mtime
    @modifiedFilesList.append("<li>#{file} - Modified at #{mtime}")
  rootView.vertical.append(this)

When you toggle the modified files list, your pane is now populated with the filenames and modified times of files in your project. You might notice that subsequent calls to this command reduplicate information. We could provide an elegant way of rechecking files already in the list, but for this demonstration, we'll just clear the modifiedFilesList each time it's closed:

# toggles the pane
if @hasParent()
  @modifiedFilesList.empty()
  rootView.vertical.children().last().remove()
else
  for file in modifiedFiles
    stat = fs.lstatSync(file)
    mtime = stat.mtime
    @modifiedFilesList.append("<li>#{file} - Modified at #{mtime}")
  rootView.vertical.append(this)

Included Libraries

FIXME: Describe `require 'atom'

In addition to core node.js modules, all packages can require the following popular libraries into their packages:

Additional libraries can be found by browsing Atom's node_modules folder.