Chrome Web Server - an HTTP web server for Chrome (chrome.sockets)

====

Basic usage:

var app = new chrome.WebApplication(options)

options: object, with keys

• handlers: array of handlers,
• renderIndex: boolean (whether to render index.html if in directory)
• port: int (port to listen on)

Handlers var handlers = [ ['/favicon.ico',FavIconHandler], ['/stream.',StreamHandler], ['/static/(.)',StaticHandler], ['.*', DefaultHandler] ]

handlers is an array of 2 element arrays where the first item is a regular expression for the URL and the second is the handler class, which should extend BaseHandler

    function StaticHandler() {
        this.disk = null
        chrome.runtime.getPackageDirectoryEntry( function(entry) { this.disk = entry }.bind(this) )
        BaseHandler.prototype.constructor.call(this)
    }
    var FavIconHandlerprototype = {
        get: function(path) {
            // USE HTML5 filesystem operations to read file
            
        },
        onReadFile: function(evt) {
            if (evt.error) {
                this.write('disk access error, perhaps restart JSTorrent')
            } else {
                this.write(evt)
            }
        }
    }
    _.extend(StaticHandler.prototype,
             StaticHandlerprototype,
             BaseHandler.prototype
            )

todo: create small example pages

====

Now updated to use the new chrome.sockets API! (old version used the now deprecated chrome.socket)

====

Get it in the chrome web store: https://chrome.google.com/webstore/detail/web-server-for-chrome/ofhbbkphhbklhfoeikjpcbhemlocgigb

The default behavior right now is very simple. You choose a directory to serve static content. It is now able to stream large files and handle range requests. It also sets mime types correctly.

Here is an example project based on it: https://chrome.google.com/webstore/detail/flv-player/dhogabmliblgpadclikpkjfnnipeebjm

====

MIT license

I wrote this because the example app provided by google would lock and hang and had all sorts of nasty race conditions. Plus it would not stream large files or do range requests, HEAD requests, etc, etc.

The design of this is inspired heavily by to the Python Tornado Web library. In this as well as that, you create an "app" which registers handlers. Then under the hood it will accept connections, create an HTTPConnection object, and that has an associated IOStream object which handles the nonblocking read/write events for you.