Note: I forked the original page.js because I needed IE9 support. This repo is just a fork with the merged PR #288 necessary for fixing IE9. Once the PR is merged, I'll probably delete this repo. But until then, feel free to use it if you need the support.

Isn't it fun having to polyfill and fix bugs for IE and Safari? ;)

Don't forget to also add the html5-history-api from npm and load it before your router.

---

Tiny Express-inspired client-side router.


page('/', index)
page('/user/:user', show)
page('/user/:user/edit', edit)
page('/user/:user/album', album)
page('/user/:user/album/sort', sort)
page('*', notfound)
page()

Installation

There are multiple ways to install page.js. With package managers:

$ npm install page # for browserify
$ component install visionmedia/page.js
$ bower install visionmedia/page.js

Running examples

To run examples do the following to install dev dependencies and run the example server:

$ git clone git://github.com/visionmedia/page.js
$ cd page.js
$ npm install
$ node examples
$ open http://localhost:4000

Currently we have examples for:

• basic minimal application showing basic routing
• notfound similar to basic with single-page 404 support
• album showing pagination and external links
• profile simple user profiles
• query-string shows how you can integrate plugins using the router
• state illustrates how the history state may be used to cache data
• server illustrates how to use the dispatch option to server initial content
• chrome Google Chrome style administration interface
• transitions Shows off a simple technique for adding transitions between "pages"
• partials using hogan.js to render mustache partials client side

NOTE: keep in mind these examples do not use jQuery or similar, so portions of the examples may be relatively verbose, though they're not directly related to page.js in any way.

API

page(path, callback[, callback ...])

Defines a route mapping path to the given callback(s). Each callback is invoked with two arguments, context and next. Much like Express invoking next will call the next registered callback with the given path.

page('/', user.list)
page('/user/:id', user.load, user.show)
page('/user/:id/edit', user.load, user.edit)
page('*', notfound)

Under certain conditions, links will be disregarded and will not be dispatched, such as:

• Links that are not of the same origin
• Links with the download attribute
• Links with the target attribute
• Links with the rel="external" attribute

page(callback)

This is equivalent to page('*', callback) for generic "middleware".

page(path)

Navigate to the given path.

$('.view').click(function(e){
  page('/user/12')
  e.preventDefault()
})

page(fromPath, toPath)

Setup redirect from one path to another.

page.redirect(fromPath, toPath)

Identical to page(fromPath, toPath)

page.redirect(path)

Calling page.redirect with only a string as the first parameter redirects to another route. Waits for the current route to push state and after replaces it with the new one leaving the browser history clean.

page('/default', function(){
  // some logic to decide which route to redirect to
  if(admin) {
    page.redirect('/admin');
  } else {
    page.redirect('/guest');
  }
});

page('/default');

page.show(path)

Identical to page(path) above.

page([options])

Register page's popstate / click bindings. If you're doing selective binding you'll like want to pass { click: false } to specify this yourself. The following options are available:

• click bind to click events [true]
• popstate bind to popstate [true]
• dispatch perform initial dispatch [true]
• hashbang add #! before urls [false]
• decodeURLComponents remove URL encoding from path components (query string, pathname, hash) [true]

If you wish to load serve initial content from the server you likely will want to set dispatch to false.

page.start([options])

Identical to page([options]) above.

page.stop()

Unbind both the popstate and click handlers.

page.base([path])

Get or set the base path. For example if page.js is operating within /blog/* set the base path to "/blog".

page.exit(path, callback[, callback ...])

Defines an exit route mapping path to the given callback(s).

Exit routes are called when a page changes, using the context from the previous change. For example:

page('/sidebar', function(ctx, next) {
  sidebar.open = true
  next()
})

page.exit('/sidebar', function(ctx, next) {
  sidebar.open = false
  next()
})

page.exit(callback)

Equivalent to page.exit('*', callback).

Context

Routes are passed Context objects, these may be used to share state, for example ctx.user =, as well as the history "state" ctx.state that the pushState API provides.

Context#save()

Saves the context using replaceState(). For example this is useful for caching HTML or other resources that were loaded for when a user presses "back".

Context#handled

If true, marks the context as handled to prevent default 404 behaviour. For example this is useful for the routes with interminate quantity of the callbacks.

Context#canonicalPath

Pathname including the "base" (if any) and query string "/admin/login?foo=bar".

Context#path

Pathname and query string "/login?foo=bar".

Context#querystring

Query string void of leading ? such as "foo=bar", defaults to "".

Context#pathname

The pathname void of query string "/login".

Context#state

The pushState state object.

Context#title

The pushState title.

Routing

The router uses the same string-to-regexp conversion that Express does, so things like ":id", ":id?", and "*" work as you might expect.

Another aspect that is much like Express is the ability to pass multiple callbacks. You can use this to your advantage to flatten nested callbacks, or simply to abstract components.

Separating concerns

For example suppose you have a route to edit users, and a route to view users. In both cases you need to load the user. One way to achieve this is with several callbacks as shown here:

page('/user/:user', load, show)
page('/user/:user/edit', load, edit)

Using the * character we can alter this to match all routes prefixed with "/user" to achieve the same result:

page('/user/*', load)
page('/user/:user', show)
page('/user/:user/edit', edit)

Likewise * can be used as catch-alls after all routes acting as a 404 handler, before all routes, in-between and so on. For example:

page('/user/:user', load, show)
page('*', function(){
  $('body').text('Not found!')
})

Default 404 behaviour

By default when a route is not matched, page.js invokes page.stop() to unbind itself, and proceed with redirecting to the location requested. This means you may use page.js with a multi-page application without explicitly binding to certain links.

Working with parameters and contexts

Much like request and response objects are passed around in Express, page.js has a single "Context" object. Using the previous examples of load and show for a user, we can assign arbitrary properties to ctx to maintain state between callbacks.

To build a load function that will load the user for subsequent routes you'll need to access the ":id" passed. You can do this with ctx.params.NAME much like Express:

function load(ctx, next){
  var id = ctx.params.id
}

Then perform some kind of action against the server, assigning the user to ctx.user for other routes to utilize. next() is then invoked to pass control to the following matching route in sequence, if any.

function load(ctx, next){
  var id = ctx.params.id
  $.getJSON('/user/' + id + '.json', function(user){
    ctx.user = user
    next()
  })
}

The "show" function might look something like this, however you may render templates or do anything you want. Note that here next() is not invoked, because this is considered the "end point", and no routes will be matched until another link is clicked or page(path) is called.

function show(ctx){
  $('body')
    .empty()
    .append('<h1>' + ctx.user.name + '<h1>');
}

Finally using them like so:

page('/user/:id', load, show)

Working with state

When working with the pushState API, and page.js you may optionally provide state objects available when the user navigates the history.

For example if you had a photo application and you performed a relatively expensive search to populate a list of images, normally when a user clicks "back" in the browser the route would be invoked and the query would be made yet-again.

An example implemenation might look as follows:

function show(ctx){
  $.getJSON('/photos', function(images){
    displayImages(images)
  })
}

You may utilize the history's state object to cache this result, or any other values you wish. This makes it possible to completely omit the query when a user presses back, providing a much nicer experience.

function show(ctx){
  if (ctx.state.images) {
    displayImages(ctx.state.images)
  } else {
    $.getJSON('/photos', function(images){
      ctx.state.images = images
      ctx.save()
      displayImages(images)
    })
  }
}

NOTE: ctx.save() must be used if the state changes after the first tick (xhr, setTimeout, etc), otherwise it is optional and the state will be saved after dispatching.

Matching paths

Here are some examples of what's possible with the string to RegExp conversion.

Match an explicit path:

page('/about', callback)

Match with required parameter accessed via ctx.params.name:

page('/user/:name', callback)

Match with several params, for example /user/tj/edit or /user/tj/view.

page('/user/:name/:operation', callback)

Match with one optional and one required, now /user/tj will match the same route as /user/tj/show etc:

page('/user/:name/:operation?', callback)

Use the wildcard char * to match across segments, available via ctx.params[N] where N is the index of * since you may use several. For example the following will match /user/12/edit, /user/12/albums/2/admin and so on.

page('/user/*', loadUser)

Named wildcard accessed, for example /file/javascripts/jquery.js would provide "/javascripts/jquery.js" as ctx.params.file:

page('/file/:file(*)', loadUser)

And of course RegExp literals, where the capture groups are available via ctx.params[N] where N is the index of the capture group.

page(/^\/commits\/(\d+)\.\.(\d+)/, loadUser)

Plugins

An example plugin examples/query-string/query.js demonstrates how to make plugins. It will provide a parsed ctx.query object derived from node-querystring.

Usage by using "*" to match any path in order to parse the query-string:

page('*', parse)
page('/', show)
page()

function parse(ctx, next) {
  ctx.query = qs.parse(location.search.slice(1));
  next();
}

function show(ctx) {
  if (Object.keys(ctx.query).length) {
    document
      .querySelector('pre')
      .textContent = JSON.stringify(ctx.query, null, 2);
  }
}

Available plugins

• querystring: provides a parsed ctx.query object derived from node-querystring.
• body-parser: provides a req.body object for routes derived from body-parser.
• express-mapper: provides a direct imitation of the Express API so you can share controller code on the client and the server with your Express application without modification.

Please submit pull requests to add more to this list.

Running tests

In the console:

$ npm install
$ npm test

In the browser:

$ npm install
$ npm run serve
$ open http://localhost:3000/

Support in IE8+

If you want the router to work in older version of Internet Explorer that don't support pushState, you can use the HTML5-History-API polyfill:

  npm install html5-history-api

How to use a Polyfill together with router (OPTIONAL):

If your web app is located within a nested basepath, you will need to specify the basepath for the HTML5-History-API polyfill. Before calling page.base() use: history.redirect([prefixType], [basepath]) - Translation link if required.

• prefixType: [string|null] - Substitute the string after the anchor (#) by default "/".
• basepath: [string|null] - Set the base path. See page.base() by default "/". (Note: Slash after pathname required)

Pull Requests

• Break commits into a single objective.
• An objective should be a chunk of code that is related but requires explaination.
• Commits should be in the form of what-it-is: how-it-does-it and or why-it's-needed or what-it-is for trivial changes
• Pull requests and commits should be a guide to the code.

License

(The MIT License)

Copyright (c) 2012 TJ Holowaychuk <[email protected]>

Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the 'Software'), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED 'AS IS', WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.