pushState + ajax = pjax
Clone or download
Pull request Compare This branch is 31 commits ahead of defunkt:master.
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Failed to load latest commit information.
script
test
vendor
.eslintrc
.gitignore
.travis.yml
CONTRIBUTING.md
Gemfile
Gemfile.lock
LICENSE
README.md
bower.json
jquery.pjax.js
package.json

README.md

pjax = pushState + ajax

pjax is a jQuery plugin that uses ajax and pushState to deliver a fast browsing experience with real permalinks, page titles, and a working back button.

pjax works by fetching HTML from your server via ajax and replacing the content of a container element on your page with the loaded HTML. It then updates the current URL in the browser using pushState. This results in faster page navigation for two reasons:

  • No page resources (JS, CSS) get re-executed or re-applied;
  • If the server is configured for pjax, it can render only partial page contents and thus avoid the potentially costly full layout render.

Status of this project

jquery-pjax is largely unmaintained at this point. It might continue to receive important bug fixes, but its feature set is frozen and it's unlikely that it will get new features or enhancements.

Installation

pjax depends on jQuery 1.8 or higher.

npm

$ npm install jquery-pjax

standalone script

Download and include jquery.pjax.js in your web page:

curl -LO https://raw.github.com/defunkt/jquery-pjax/master/jquery.pjax.js

Usage

$.fn.pjax

The simplest and most common use of pjax looks like this:

$(document).pjax('a', '#pjax-container')

This will enable pjax on all links on the page and designate the container as #pjax-container.

If you are migrating an existing site, you probably don't want to enable pjax everywhere just yet. Instead of using a global selector like a, try annotating pjaxable links with data-pjax, then use 'a[data-pjax]' as your selector. Or, try this selector that matches any <a data-pjax href=> links inside a <div data-pjax> container:

$(document).pjax('[data-pjax] a, a[data-pjax]', '#pjax-container')

Server-side configuration

Ideally, your server should detect pjax requests by looking at the special X-PJAX HTTP header, and render only the HTML meant to replace the contents of the container element (#pjax-container in our example) without the rest of the page layout. Here is an example of how this might be done in Ruby on Rails:

def index
  if request.headers['X-PJAX']
    render :layout => false
  end
end

If you'd like a more automatic solution than pjax for Rails check out Turbolinks.

Check if there is a pjax plugin for your favorite server framework.

Also check out RailsCasts #294: Playing with PJAX.

Arguments

The synopsis for the $.fn.pjax function is:

$(document).pjax(selector, [container], options)
  1. selector is a string to be used for click event delegation.
  2. container is a string selector that uniquely identifies the pjax container.
  3. options is an object with keys described below.
pjax options
key default description
timeout 650 ajax timeout in milliseconds after which a full refresh is forced
push true use pushState to add a browser history entry upon navigation
replace false replace URL without adding browser history entry
maxCacheLength 20 maximum cache size for previous container contents
version a string or function returning the current pjax version
scrollTo 0 vertical position to scroll to after navigation. To avoid changing scroll position, pass false.
type "GET" see $.ajax
dataType "html" see $.ajax
container CSS selector for the element where content should be replaced
url link.href a string or function that returns the URL for the ajax request
target link eventually the relatedTarget value for pjax events
fragment CSS selector for the fragment to extract from ajax response

You can change the defaults globally by writing to the $.pjax.defaults object:

$.pjax.defaults.timeout = 1200

$.pjax.click

This is a lower level function used by $.fn.pjax itself. It allows you to get a little more control over the pjax event handling.

This example uses the current click context to set an ancestor element as the container:

if ($.support.pjax) {
  $(document).on('click', 'a[data-pjax]', function(event) {
    var container = $(this).closest('[data-pjax-container]')
    var containerSelector = '#' + container.id
    $.pjax.click(event, {container: containerSelector})
  })
}

NOTE Use the explicit $.support.pjax guard. We aren't using $.fn.pjax so we should avoid binding this event handler unless the browser is actually going to use pjax.

$.pjax.submit

Submits a form via pjax.

$(document).on('submit', 'form[data-pjax]', function(event) {
  $.pjax.submit(event, '#pjax-container')
})

$.pjax.reload

Initiates a request for the current URL to the server using pjax mechanism and replaces the container with the response. Does not add a browser history entry.

$.pjax.reload('#pjax-container', options)

$.pjax

Manual pjax invocation. Used mainly when you want to start a pjax request in a handler that didn't originate from a click. If you can get access to a click event, consider $.pjax.click(event) instead.

function applyFilters() {
  var url = urlForFilters()
  $.pjax({url: url, container: '#pjax-container'})
}

Events

All pjax events except pjax:click & pjax:clicked are fired from the pjax container element.

event cancel arguments notes
event lifecycle upon following a pjaxed link
pjax:click ✔︎ options fires from a link that got activated; cancel to prevent pjax
pjax:beforeSend ✔︎ xhr, options can set XHR headers
pjax:start xhr, options
pjax:send xhr, options
pjax:clicked options fires after pjax has started from a link that got clicked
pjax:beforeReplace contents, options before replacing HTML with content loaded from the server
pjax:success data, status, xhr, options after replacing HTML content loaded from the server
pjax:timeout ✔︎ xhr, options fires after options.timeout; will hard refresh unless canceled
pjax:error ✔︎ xhr, textStatus, error, options on ajax error; will hard refresh unless canceled
pjax:complete xhr, textStatus, options always fires after ajax, regardless of result
pjax:end xhr, options
event lifecycle on browser Back/Forward navigation
pjax:popstate event direction property: "back"/"forward"
pjax:start null, options before replacing content
pjax:beforeReplace contents, options right before replacing HTML with content from cache
pjax:end null, options after replacing content

pjax:send & pjax:complete are a good pair of events to use if you are implementing a loading indicator. They'll only be triggered if an actual XHR request is made, not if the content is loaded from cache:

$(document).on('pjax:send', function() {
  $('#loading').show()
})
$(document).on('pjax:complete', function() {
  $('#loading').hide()
})

An example of canceling a pjax:timeout event would be to disable the fallback timeout behavior if a spinner is being shown:

$(document).on('pjax:timeout', function(event) {
  // Prevent default timeout redirection behavior
  event.preventDefault()
})

Advanced configuration

Reinitializing plugins/widget on new page content

The whole point of pjax is that it fetches and inserts new content without refreshing the page. However, other jQuery plugins or libraries that are set to react on page loaded event (such as DOMContentLoaded) will not pick up on these changes. Therefore, it's usually a good idea to configure these plugins to reinitialize in the scope of the updated page content. This can be done like so:

$(document).on('ready pjax:end', function(event) {
  $(event.target).initializeMyPlugin()
})

This will make $.fn.initializeMyPlugin() be called at the document level on normal page load, and on the container level after any pjax navigation (either after clicking on a link or going Back in the browser).

Response types that force a reload

By default, pjax will force a full reload of the page if it receives one of the following responses from the server:

  • Page content that includes <html> when fragment selector wasn't explicitly configured. Pjax presumes that the server's response hasn't been properly configured for pjax. If fragment pjax option is given, pjax will extract the content based on that selector.

  • Page content that is blank. Pjax assumes that the server is unable to deliver proper pjax contents.

  • HTTP response code that is 4xx or 5xx, indicating some server error.

Affecting the browser URL

If the server needs to affect the URL which will appear in the browser URL after pjax navigation (like HTTP redirects work for normal requests), it can set the X-PJAX-URL header:

def index
  request.headers['X-PJAX-URL'] = "http://example.com/hello"
end

An X-PJAX request header is set to differentiate a pjax request from normal XHR requests. In this case, if the request is pjax, we skip the layout html and just render the inner contents of the container.

Check if your favorite server framework supports pjax here: https://gist.github.com/4283721

Server side JS variables

Sometimes you might need to send some data from server along with HTML contents. This can be done by sending X-PJAX-Vars response header which contains JSON encoded values. These values are saved with state as $.pjax.state.vars. You can access them with options.vars, no matter whether it is a fresh request or popstate.

require 'json'
if request.headers['X-PJAX']
  my_vars = {:user => "in"}
  response.headers['X-PJAX-Vars'] = JSON.generate(my_vars)
end

Layout Reloading

Layouts can be forced to do a hard reload when assets or html changes.

First set the initial layout version in your header with a custom meta tag.

<meta http-equiv="x-pjax-version" content="v123">

Then from the server side, set the X-PJAX-Version header to the same.

if request.headers['X-PJAX']
  response.headers['X-PJAX-Version'] = "v123"
end

Deploying a deploy, bumping the version constant to force clients to do a full reload the next request getting the new layout and assets.

Client redirect

Using a Location: url header would redirect just the pjax request to a new location and load contents from there. Instead use X-PJAX-Location response header to instruct pjax to make a full page redirect.

require 'json'
if request.headers['X-PJAX']
  response.headers['X-PJAX-Location'] = "https://some.new.location"
else
  response.headers['Location'] = "https://some.new.location"
end

You can handle the redirect with options.redirect_handler(new_location, options, xhr) method (for ex. to display a message and a loader).

<title>, <meta> and <link> tags

HTML response could also inclulde some meta information of the requested page. This meta can be sent with the contents, enclosed in <head> tag, as in a common HTML document.

<link rel="stylesheet" /> is ignored. Load all your stylesheets at page load or dynamically using JS.

<meta class="pjax" /> and <link class="pjax" /> tags are automatically removed at pjax request. This is intended to mark metas that belong to on specific page only.

Legacy API

Pre 1.0 versions used an older style syntax that was analogous to the now deprecated $.fn.live api. The current api is based off $.fn.on.

$('a[data-pjax]').pjax('#pjax-container')

Expanded to

$('a[data-pjax]').live('click', function(event) {
  $.pjax.click(event, '#pjax-container')
})

The new api

$(document).pjax('a[data-pjax]', '#pjax-container')

Which is roughly the same as

$(document).on('click', 'a[data-pjax]', function(event) {
  $.pjax.click(event, '#pjax-container')
})

NOTE The new api gives you control over the delegated element container. $.fn.live always bound to document. This is what you still want to do most of the time.

Contributing

$ git clone https://github.com/defunkt/jquery-pjax.git
$ cd jquery-pjax/

To run the test suite locally, start up the Sinatra test application.

$ bundle install
$ bundle exec ruby test/app.rb
== Sinatra/1.4.5 has taken the stage on 4567 for development with backup from WEBrick

# in another tab:
$ open http://localhost:4567/