Express State

Share configuration and state data of an Express app with the client-side via JavaScript.

Overview

Goals

Express State is designed to make it easy to share configuration and state data from the server to the client. It can be used to share any data that needs to be available to the client-side JavaScript code of the an app: e.g., the current user, a CSRF token, model data, routes, etc.

Progressively enhanced Web apps can be built by rendering an app's initial state on the server and using Express State as the conduit through which the server passes data and control over to the client-side JavaScript code.

How It Works

Configuration and state data are exposed to client-side JavaScript via two methods: app.expose() and res.expose(), both of which make the data available on a special state "locals" object for views/templates to serialize and embed into HTML pages.

When views/templates embed this exposed data into an HTML page, it is serialized as literal JavaScript. The JavaScript serialization format is limited to expressions that initialize namespaces and the exposed data assigned to those namespaces, which is a superset of JSON that includes regular expressions and functions.

Features

Express State was written because of the shortcomings of express-expose. The following is a list of features highlighting differences when compared with express-expose:

• An efficient and powerful serialization format: Literal JavaScript is used to namespace exposed data that is a superset of JSON and includes regular expressions and functions. This avoids the cost of allocating and parsing large JSON strings on the client and enables things like sharing routes defined as regular expressions with a client-side URL router.

• Smart namespacing: A root namespace can be set via an app's state namespace setting and it will be prepended to namespaces passed to expose() unless they already contain it or they start with "window.". The "global" on to which the namespaces are created can also be controlled.

• Precise data value overrides: Sub-values within exposed objects can be easily overridden without clobbering the entire object. Request scoped values can even override data exposed at the app's scope.

• Lazy serialization: Exposed data objects are stored by reference, making them "live" and allowing their values to be updated even after the object has been exposed. Only the namespaces and data that are still reachable after the series of expose() calls will be serialized. Serialization can happen at anytime, on demand, by calling the toString() method on state "locals" objects.

When data is not going to change the {cache: true} option can be set to eagerly serialize exposed objects, making repeated toString() calls more efficient.

• Explicit extension of each Express app: Express State's functionality has to be explicitly added to an Express app via the exported extend() function. This prevents problems in complex apps where multiple versions of Express and/or multiple Express apps are used.

Installation

Install using npm:

$ npm install express-state

Usage

Extending an Express App

To use Express State with an Express app, the app must first be extended. Use the extend() method that Express State exports:

var express  = require('express'),
    expstate = require('express-state'),

    app = express();

expstate.extend(app);

Once extended, the app will have the app.expose() method, and response objects will have the res.expose() method.

Note: It's perfectly fine for the same Express app to be extended more than once; after the first time the app is extended, the subsequent extend() calls will be noops.

Exposing Data

Data can be exposed at two different scopes: the app's scope, and a request/response's scope via app.expose() and res.expose() respectively.

Express State uses Express's built-in "locals" system. When data is exposed at the app's scope, a special app.locals.state object is created and used as the backing store for all app.expose() calls. Express also merges app.locals with res.locals to create the context object in which views/templates are rendered. This means that, by default, data exposed at the app's scope will also be present when rendering views/templates for all requests.

Express State sets up a similar relationship using prototypal inheritance where res.locals.state inherits from app.locals.state. This means data exposed at the request scope will also contain exposed data from the app's scope. If values for the same namespace are exposed at both scopes, the request/response scope takes precedence and shadows the value at the app's scope.

Exposing App-Scoped Data

When data that needs to be exposed to the client-side JavaScript code is not request-specific and should be available to all requests, it should be exposed at the app's scope using app.expose().

The following example exposes a Flickr API key required by Flickr to identify requests:

app.expose({
    api_key: '02348notreal2394879137872358bla'
}, 'MY_APP.Flickr');

The client-side JavaScript code can now look up the Flickr API key at MY_APP.Flickr.api_key when it needs to make a request to Flickr's API.

Exposing Request-Scoped Data

When data that needs to be exposed to the client-side JavaScript is request-specific, it should be exposed at the request/response's scope using res.expose().

The following example shows how to create a middleware function to expose the current person's Cross Site Request Forgery (CSRF) token—this is a best practice where the CSRF is used to validate HTTP requests that mutate state:

// Add Express' packaged `cookieParser()`, `session()`, and `csrf()` middleware.
app.use(express.cookieParser());
app.use(express.session({secret: 'something secure, not this!'}));
app.use(express.csrf());

// Create a middleware function that will expose the CSRF token for the current
// request only.
app.use(function (req, res, next) {
    res.expose(req.session._csrf, 'MY_APP.CSRF_TOKEN');
    next();
});

The client-side JavaScript code can now add the X-CSRF-Token HTTP header with the value at MY_APP.CSRF_TOKEN to all XHRs it makes to the server.

Increase Performance of Unchanging or Static Data

It's common to expose app-scoped data which will not change during the lifecycle of the Express app instance. To improve per-request performance, this unchanging/static data can be eagerly serialized and cached by setting the {cache: true} option:

var CONFIG = {
    hostname : 'example.com',
    someOther: 'constant value'
};

app.expose(CONFIG, 'MY_APP.config', {cache: true});

Setting this option allows Express State to optimize the serialization process by keeping the serialized value around and re-using it every time the toString() method is invoked (which happens for every request.)

Note: When a large amount of data needs to be exposed to the client-side, it is recommended to come up with a strategy where all data which is common to most/every request be exposed at the app-scope with the {cache: true} option set.

Untrusted User Input

Always escape untrusted user input to protected against XSS attacks!

Express State provides a mechanism to expose configuration and state data as first-party JavaScript, which means any untrusted user input should be properly escaped based on the [OWASP HTML escaping recommendations][OWASP].

Express State will automatically encode any <, >, / characters within string values of exposed data to their Unicode counterparts during serialization. This provides a basic level of protection against XSS attacks by not allowing the "</script><script>" character sequence within an exposed string value to be interpreted and cause the browser prematurely close a script element and reopen a new one.

Even with the basic XSS protection Express State provides, it's still important to always escape untrusted user input.

Exposing Functions

Express State allows for functions to be serialized and sent to the browser, but this has a few limitations and practical constraints:

• A TypeError will be thrown if a native built-in function is being serialized, like the Number constructor. Native built-ins should be called in wrapper functions, which can be serialized.

• Functions should only be exposed if they are dependency free and monadic in nature. The original scope in which a function defined is not guaranteed to be present in the client-side environment. If a function references variables or has other dependencies outside its scope, it's likely not to work properly.

• Application code should not be sent to the browser by exposing it via Express State. That would be a misuse of this library and it's recommended that client-side code be organized into serve-able files or modules allowing the browser to download the code via standard <script src=""> elements or a script loader.

Setting a Root Namespace

A common practice is to set a root namespace for an app so all of its exposed data is contained under one global variable in the client-side JavaScript code. A root namespace can be setup for an app using the state namespace setting:

app.set('state namespace', 'MY_APP');

Now anytime data is exposed, the root namespace will be prepended unless it already exists in the namespace passed into the expose() call or the passed-in namespace starts with "window.".

With the above "MY_APP" root namespace, the following are all equivalent and result in MY_APP.foo === 123 in the client-side JavaScript:

// These all have the same result on the client: `MY_APP.foo === 123`
app.expose(123, 'foo');
app.expose(123, 'MY_APP.foo');
app.expose(123, 'window.MY_APP.foo');

Setting a root namespace helps keep code DRY and configurable at the app level while having the "window." escape hatch for data that needs to be exposed at a specific namespace on the client.

Overriding Exposed Values

Objects that are exposed through either expose() method are stored by reference, and serialization is done lazily (unless the {cache: true} option was set). This means the objects are still "live" after they've been exposed. An object can be exposed early during the life cycle of a request and updated up until the response is sent.

The following is a contrived example, but shows how values can be overridden at any time and at any scope:

app.expose({root: '/'}, 'url');

app.use(function (req, res, next) {
    res.expose(req.path, 'url.path');
    res.expose(req.query, 'url.query');
    next();
});

On the client, the resulting url object would look like the following for a request to the URL "/foo?bar=baz":

{ root: '/',
  path: '/foo',
  query: { bar: 'baz' } }

Notice how exposing values at the url.path and url.query namespaces did not clobber the original url object exposed at the app's scope.

However, previously exposed data can be completely clobbered by simply exposing a new value at the same namespace. When this happens, Express State is smart enough to know it can release its references to the previous value objects and not waste CPU and bytes serializing them.

Serialization

Express State serializes exposed data to literal executable JavaScript. The JavaScript produced during serialization is limited to expressions that initialize namespaces and the exposed data assigned to those namespaces, which is a superset of JSON that includes regular expressions and functions.

JavaScript, as the serialization format, is more powerful and efficient than JSON. It avoids the cost of allocating and parsing large JSON strings on the client and enables things like sharing routes defined as regular expressions with a client-side URL router.

The special app.locals.state and res.locals.state objects contain a custom toString() method implementation, which serializes the objects to JavaScript that is human readable and can be embedded inside a <script> element in an HTML page.

The following example shows a series of expose() calls and the resulting output from serialization:

app.expose({bar: 'bar'}, 'foo');
app.expose(/baz/, 'foo.baz');
app.expose(function () { return 'bla'; }, 'a.very.big.ns');

// Serialize `app.locals.state` and log the result.
console.log(app.locals.state.toString());

The output of the console.log() call would be:

(function (root) {
/* -- Data -- */
root.foo = {"bar":"bar"};
root.foo || (root.foo = {});
root.foo.baz = /baz/;
root.a || (root.a = {});
root.a.very || (root.a.very = {});
root.a.very.big || (root.a.very.big = {});
root.a.very.big.ns = function () { return 'bla'; };
}(this));

Note: A TypeError will be thrown if a native built-in function is being serialized, like the Number constructor. Native built-ins should be called in wrapper functions