eclipse-vertx / vert.x Goto Github PK

Complete this module.

Consider adding methods for getting, setting and appending primitive types.

Test remains to be done.

Implement this.

Initially use new Java 7 async file IO.

Produce a vert.x core library as part of the distribution.

This can be used embedded in any Java application.

Complete this.

Remaining to do:

1. HTTPS
2. Cookies
3. Pipelining test currently commented out
4. Some missing tests, including: chaining methods, multiple connections, pause/resume
5. Write some examples including: Websockets, sendfile, POST/PUT example

Hi running tests (with gradle or with ant) fails at org.vertx.tests.core.file.FileSystemTest
getting null as message for all fails. I am on ubuntu 11.10, using openjdk 1.7

Copy of the gradle report:
org.vertx.tests.core.file.FileSystemTest
testMkdirWithParentsFails
testChmod
testMkdirCreateParents
testMkdirCreateParentsWithPerms
testCopyEmptyDir
testMkdirSimple

On this page: http://vertx.io/wsver-0002/ruby_web_tutorial.html in step 4 ("Get the persistor up and running") there is the following line which isn't grammatically correct and a bit confusing.

Vert.x ships with an out of the box bus module (busmod) called MongoPersistor. This is a busmod is component which communicates with other components on the vert.x event bus by exchanging JSON messages.

I think it should be more like the following?

Vert.x ships with an out-of-the-box bus module (busmod) called MongoPersistor.  A busmod is a component which communicates with other components on the vert.x event bus by exchanging JSON messages.

Timer functionality

Module for spawning processes and capturing their input output streams etc

Implement this.

Complete the Net module.

Remaining to do:

1. SSL support
2. Sendfile

Complete this.

Consider renaming when() and provide other ways of composing. Update methods in other modules to return Completions as appropriate

Due to an error in the way handlers are applied in DefaultSockJSBridge, it means that only websockets will work for all sockJS transport.

This is because the bridge is applying a catch all handler for all requests which wipes out any non websockets handlers applied by the sock js server.

This effects 1.0.0.beta2

hybi-10 is now supported by IE, Thunderbird and Chrome.

We need to update our implementation to hybi-10 so it works (the current impl now won't work with the latest version of Chrome)

Someone has made a pull request in Netty here:

netty/netty#26

We should take a look and merge it into our forked Netty.

For interfacing with JMS providers

Pingers so cluster connections are properly closed when nodes disappear from network

Due to the overzealous nature of IntelliJ refactoring, the string 'body' was renamed to 'payload' in multiple places in the project, including some of the static html used by SockJS.

This may cause SockJS to fail to work with all transports.

This effects vertx-1.0.0.beta2

Around the area of nodes joining and leaving the cluster

For integrating with Twitter/Facebook/Google+

For reading/writing from console

Complete these with associated tests. We should accept any value that's a primitive type, String, or implements the Threadsafe interface

Java examples ant build script in the 1.0.0.beta.1 distribution is broken.

To fix, change:

to

Modify build.xml so vertx.bat is in the bin directory of the distribution along with vertx

When two or more event loops attempt to send to the new server concurrently and the connection has not yet been setup

blah blah

We should write a super lightweight framework in the style of sinatra/express that can be layered on top of core HTTP.

We need to create a package manager for node.x

We should also host a public repository (e..g on Amazon) where people can upload and download modules.

All modules other than core should be managed via the package manager.

If you are running vert.x in clustered mode and do not specify -cluster-host, it will default to 0.0.0.0 which means if you have multiple nodex running on the network they won't see each other (since they try to connect to the others at 0.0.0.0).

Workaround is to always specify -cluster-host when running clustered.

AMQP 0.9.1 client.

Asynchronous connection pool will be needed for:

Net, HTTP, AMQP, Redis etc

Remove buffer factory methods etc

FYI:

http://purplefox.github.com/vert.x/ -- returns a 404

Is there another location / site with documentation?