http-enhanced

http-enhanced

A drop-in replacement (add-on, really) for Node.js standard library's http module.

It adds a few methods to the http.ServerResponse and http.IncomingMessage prototypes. If you don't call any of the methods, nothing changes. You'll just have some extra methods lying around that never did anything.

Install

From the npmjs registry:

npm install --save http-enhanced

Usage

Most often, you'll create a server like this:

var http = require('http')
http.createServer(function (req, res) {

  // req (request) is an instance of http.IncomingMessage
  var url = req.url

  // res (response) is an instance of http.ServerResponse
  res.writeHead(200, {'Content-Type': 'text/html'})
  res.write('Hello, world!')
  res.end()

}).listen(80)

A simple change will let you use some shortcuts:

var http = require('http-enhanced')
http.createServer(function (req, res) {

  // save incoming data to req.data and wait until the request ends,
  //   or callback immediately (setImmediate) if it already has
  req.readToEnd('utf8', function(err, data) {
    var reversed = data.split('').reverse().join('')

    // 1. set HTTP status code to 200,
    // 2. set the Content-Type header to 'application/json',
    // 3. and stringify the given object, all in one go:
    res.json({success: true, message: reversed})
  })

}).listen(80)

Request

request.readToEnd([encoding], [callback])

• encoding String | null If specified, callback gets a String instead of a Buffer.
• callback Function | null Call some function(err, data) when the request ends.
  • If only one argument is given, it must be the callback.

Read the request into a buffer until the end event is triggered, which will trigger callback(Error || null, Buffer || null). This uses "new-style" streams, listening for readable events and calling read(), coercing to a Buffer when needed.

This function can be called multiple times, with or without the callback.

A popular use case might be if you want to upload a file and do a lot of I/O independently but at the same time. You might call req.readToEnd() with no arguments at the beginning of your request handler, and then set off your expensive I/O calls.

You can get back a string if you specify the encoding, e.g., req.readToEnd('utf8', function(err, string) { ... }). This is exactly equivalent to calling:

req.readToEnd(function(err, buffer) {
  var string = buffer.toString(encoding)
  ...
})

If the request has already ended, any captured buffer will be immediately returned, via setImmediate (which replaced process.nextTick in node v0.10). This might occur if you start listening for data at some point, in which case the request is flipped to "old-style" streams, and end might occur before you listen for it.

For that reason, and that calling req.read() from multiple listeners could produce problems, you should not use either of these:

req.setEncoding('utf8') // no!
req.on('data', function(chunk) { ... }) // robot, NO!

So you shouldn't call req.readToEnd() (without a callback) in your pipeline unless you're going to call it again with a callback, later.

request.readData(callback)

Wraps req.readToEnd() and uses the request's Content-Type header to determine whether to parse the request as JSON or a form querystring.

callback = function(error, data) { ... }

• application/json: Returns result of JSON.parse. Interprets empty application/json requests as null, instead of throwing an Error, since JSON.parse('') will raise a SyntaxError normally. If JSON.parse(body) throws an error due to invalid JSON, calls back with the error.
• application/x-www-form-urlencoded: Returns result of querystring.parse.
• otherwise, returns the same thing as readToEnd, a Buffer.

Does not work for uploads (use something like formidable).

Returns the parsed querystring for GET requests.

Response

response.writeEnd(data)

• data String String to write to response

The standard http built-in response.end(data) is supposed to write the data and then end the response. From the docs:

If data is specified, it is equivalent to calling request.write(data, encoding) followed by request.end().

But sometimes it doesn't, and writeEnd makes sure that's what it really does (minus the optional encoding).

res.writeEnd('Hello world')

response.writeAll(statusCode, contentType, data)

• statusCode Number Three-digit HTTP status code
• contentType String MIME type
• data String String to write to response

Roll writeHead(statusCode, contentType) and writeEnd(data) all into one:

res.writeAll(200, 'text/xml', '<root>Haha, not really.</root>')

response.json(object)

• object Object JSON-stringifiable object

Write response to end with Content-Type: application/json and HTTP status code 200, encoding the object with JSON.stringify.

res.json({success: true, message: "Hello world!"})

If JSON.stringify throws an error trying to encode your object (e.g., if it has circular references), it will fall back to util.inspect with the options: {showHidden: true, depth: null}.

response.html(data)

• data String HTML to write to response.

Set status code to 200 and Content-Type to text/html.

res.html('<p><i>Hello</i> world!.</p>')

response.text(data)

• data String Plain text to write to response.

Set status code to 200 and Content-Type to text/plain.

res.text('Hello world.')

response.die([error])

• error String | Error Will call error.toString().

Set status code to 500 (if it's currently 200) and Content-Type to text/plain, using the string representation of the given error, prepended with the label "Failure: ", as the response body (or just "Failure" if no error is provided). If you want to use a 4xx or 5xx status code other than 500, call, e.g., .status(418) before calling .die().

res.die('Goodbye, cruel world.')

response.redirect(location)

• location String (a URL)

Set status code to given status code (302 by default) and the Location header to the given string. Also writes the text, "Redirecting to: /index?error=404" (or whatever url you use).

res.redirect('/index?error=404')

To use a different 3xx status code, set it before calling redirect.

res.status(303).redirect('/login')

License

Copyright 2013-2014 Christopher Brown. MIT Licensed.