tlrobinson / jack

[UNMAINTAINED] A port of Rack for the JavaScript JSGI web server interface

JSGI & Jack

JSGI is a web server interface specification for JavaScript, inspired by Ruby's Rack (http://rack.rubyforge.org/) and Python's WSGI (http://www.wsgi.org/). It provides a common API for connecting JavaScript frameworks and applications to webservers.

Jack is a collection of JSGI compatible handlers (connect web servers to JavaScript web application/frameworks), middleware (intercept and manipulate requests to add functionality), and other utilities (to help build middleware, frameworks, and applications).

Homepage:

• http://jackjs.org/
• http://narwhaljs.org/

Source & Download:

• http://github.com/tlrobinson/jack/
• http://github.com/tlrobinson/narwhal/

Mailing list:

• http://groups.google.com/group/narwhaljs

IRC:

• #narwhal on irc.freenode.net

JSGI Specification

View the JSGI specification.

Example JSGI Application

function(env) {
    return {
        status : 200,
        headers : {"Content-Type":"text/plain"},
        body : ["Hello world!"]
    };
}

JSGI vs. Rack

JSGI applications are simply functions, rather than objects that respond to the "call" method. The body must have a forEach method which yields objects which have a toByteString method, as opposed to Strings.

JSGI vs. WSGI

WSGI uses a start_response function to set the HTTP status code and headers, rather than returning them in an array. JSGI is similar to WSGI 2.0: http://www.wsgi.org/wsgi/WSGI_2.0.

Contributors

• Tom Robinson
• George Moschovitis
• Kris Kowal
• Neville Burnell
• Isaac Z. Schlueter
• Jan Varwig
• Irakli Gozalishvili
• Kris Zyp
• Kevin Dangoor
• Antti Holvikari
• Tim Schaub

Acknowledgments

This software was influenced by Rack, written by Christian Neukirchen.

http://rack.rubyforge.org/

License

Copyright (c) 2009 Thomas Robinson <280north.com>

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 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.

jack's Issues

session improvements

• explicit session deserialization
• only serialize dirty sessions.

tests don't run

I think you mean module.id instead of require.id in the tests, no?

I can't get tests to run without:
http://github.com/tschaub/jack/commit/d93b2a7d690e7c7fec0a1d97853c0de656bc01b3

Pull Request: add response validation and update jack tests

I made a series of commits, first to add validation to responses to make sure they are JSGI compliant, and secondly to fix the currently failing jack testcases.

http://github.com/paulbaumgart/jack/commits/error-handling

Reviews appreciated.

Pull Request: update detection for whether jackup is being run as a standalone program

This just implements the more correct version that doesn't require type coercion:
http://github.com/paulbaumgart/jack/commit/55697bf8a45089d9adeb2a3e167b1e0874377b04

Pull Request: The Response.deleteCookie function not functioning correctly.

Changing { expires: 0 } to { expires: new Date(0) } fixed the problem.

"jackup" not picking up packages in the local packages directory

To reproduce on a debian box:

Create a new narwhal package, f.ex: tusk init my-project

Create a local packages directory inside my-project and create a new
package inside that packages directory, f.ex:

mkdir packages && cd packages && tusk init my-local-package

Add lib/foo-module.js file inside my-local-package/lib directory.

Now add a test.js file inside my-project/lib directory and require
foo-module defined in my-local-package.

If you invoke test.js using narwhal then everything works fine. If
instead you invoke test.js using jackup, the system is unable to load
the foo-module.

-s server doesn't work

line 115 of jackup should read var server = options.server

Node.js

Hows support for node?

Pull Request: fix for exceptions being thrown by the SIMPLE webserver if parts of the response are undefined

Currently, if either the headers or the body of the response are not defined, various exceptions get thrown. http://github.com/paulbaumgart/jack/commit/671e21b2c0bfb5d13a7a08518e15b9105e4231fb adds a few extra checks to prevent that.

I haven't had a chance to look at the other handlers to see if they have the same behavior.

Jack: ERROR: ReferenceError: "file" is not defined.

I'm getting this error in Jack's file.js module. To reproduce:

1. Start the example site. jackup jack/examples/example.js
2. Create an "example.txt" file at jack/examples/example.txt
3. Navigate the browser to http://localhost:8080/example.txt

Browser shows blank page, log in the terminal shows ERROR: ReferenceError: "file" is not defined.

Seems to be related to File = require("file") on line 3 of jack/lib/jack/file.js

use require("uri").parse instead of require("uri").URI.parse

See tschaub/jack@ee2a7a7cf7ec29d807fcfa6881c80103cbb4b2ac

parseMultipart should use binary.js

It should provide a way to access the uploaded data as a ByteString/Array in addition to the tempfile method.

That would be extremely useful in AppEngine deployments that do not provide write access to the file system and thus the tempfile solution in inapropriate

Response.redirect

Deprecate in favor of the Redirect middleware?

Non-form-encoded bodies should not be parsed as though they are form encoded.

When a non-form-encoded content-type is present, the body in Request should not be parsed as though it is form-encoded.

encode multipartFormData hasOwnProperty guard

need to check in encodeMultipart to accomodate arrays that have had functions added to their prototype

s/platform/engine/g

Since http://github.com/tlrobinson/narwhal/commit/a18eaf5eb105fb5a93957e2950962a6aa3decbd7, jackup doesn't run. Replacing system.platform with system.engine looks to fix things.

request helpers (pull-request)

remote.ip
remote.acceptEncoding

Simple Server should be able to be stopped via a signal

Shouldn't have to kill the process to stop it. It should be able to be stopped by a signal on unix or by a command sent to a port other than the web listening port on windows. I'm not so sure that's how it should be done on windows, any ideas about that?

Response.redirect throws exception

pull-request

patch:
http://github.com/gmosx/jack/commit/a05b47bf961f05ee8510c1357d6c5fa97f69d836

Jackup does not recognize windows absolute paths to configuration module

if the path to the configuration module is specified via an absolute path such as c:\myapp.js, jackup considers this to be a relative path and prepends File.cwd() to it:

if (config.charAt(0) !== "/")
        config = File.join(File.cwd(), config);

Upgrade to JSGI 0.3

This covers the pull request for upgrading to JSGI 0.3 when we are ready.
http://github.com/kriszyp/jack/

Pull Request: add objective-j, cappuccino, and plist mime-types to mime.js

http://github.com/paulbaumgart/jack/commit/464890d63f1e4852aa28cd70e2083b1d2fd62ee9

MethodOverride

MethodOverride calls request.POST() that deserializes multipart data to tempfiles (even if not required). It should manually parse the _method param or use GET data?

Response Constructor and ContentLength middleware set inappropriate headers for 304 responses

Content-Type should not be set if the status passed into the constructor is a 304.

ContentLength Middleware should not add a Content-Length to a 304 response.

Here is a commit fixing these issues:
http://github.com/nrstott/jack/commit/70604c18b569fd644dcf9b730a6846fb47a630d9

BUG, Parse Mulitpart does not properly handle non-file sections with values that have newlines

Here is the fix: http://github.com/nrstott/jack/commit/18e73ba3e5fc976bc3252978395783a6bc89e2f0

I also greatly simplified the algorithm while I was at it.

This bug would occur when you had a text area with
"line
break"
in it. The parse multipart would set the value to be just "line"