The environment that I chose to run it after some searching and recommendations is:

1. Open Fire – This is really the chat server using XMPP (jabber)
2. StropheJS – This is the XMPP JS library that I choose to use.

Setting up Open Fire

Setting up Open Fire was relatively easy. I basically followed the steps from the documentation and help from Malcom .

However 2 things complicate matters.

1. I am running Open Fire behind a firewall so needed to set up a Reverse Proxy in order to configure it.
2. VJ’s maybe running behind a firewall too. So Open Fire needs to be able to communicate on port 80 as well. Luckily Open Fire supports “http-bind” so I just need to set it up.

Reverse Proxy for Open Fire and HTTP-BIND

In order to configure Open Fire, I need to access the port 9090. But I can’t change the firewall. So instead, set up a Virtual Host to (Reverse) proxy it. This is also where I administer my Open Fire server.

/etc/httpd.conf


<VirtualHost openfire.mydomain.com>
ServerName openfire.mydomain.com
ServerAlias openfire.mydomain.com
ProxyRequests Off
ProxyPass / http://127.0.0.1:9090
ProxyPassReverse / http://127.0.0.1:9090
</VirtualHost>

Now I can configure Open Fire by going to http://openfire.mydomain.com/ without the 9090 port reversed proxied. example

Important (for HTTP-BIND)

Configure the Server/Domain Name to be Another sub-domain – instead of openfire.mydomain.com, use chat.mydomain.com because

1. openfire.mydomain. com is reverse proxied for administration only and chat.mydomain.com is the server/domain of the XMPP server that clients will connect to.
2. So chat.mydomain.com needs 5222 open for chat clients talking XMPP. But because I am behind a firewall and chat clients may also be behind a firewall, we cannot use 5222 to communicate.
3. Luckily Open Fire has HTTP-BIND, that means it can talk HTTP that wraps the XMPP. But HTTP-BIND runs on port 7070.
4. Since I cannot open 7070 in the firewall. I need to Reverse Proxy this as well so that clients can use a normal looking URL to talk HTTP(XMPP) to.

/etc/httpd.conf

<VirtualHost chat.mydomain.com>
ServerName chat.mydomain.com
ServerAlias chat.mydomain.com
ProxyRequests Off
ProxyPass /http-bind http://127.0.0.1:7070/http-bind/
ProxyPassReverse /http-bind http://127.0.0.1:7070/http-bind/
</VirtualHost>

SparkWeb

I need also a web chat client for JANE, the “helpdesk” agent. so I installed Spark Web. Installing it is straight forward too. download it and put it in the server.

<VirtualHost chat.mydomain.com>
...
<Directory "/usr/local/sparkweb">
AllowOverride All
Order deny,allow
</Directory>
</VirtualHost>


and edit the index.html


function jive_sparkweb_getConfig()
{
return {
server: "localhost",
connectionType: "http",
port: "80"
};
}

Now Start Coding!!

I use these for development.

1. Eclipse IDE
2. Aptana Studio

With the help of the samples from Aptana, I was able to get started with a chrome window for my chat widget. However, the main task for me here is to integrate with the Strophe JS.

Integrate with Strophe.js

Basically I modified echobot.js since it already provides the Login and onMessage Handler.

First include all the necessary Strophe .js files.

....
<script language='javascript' type='text/javascript' src='lib/strophejs/strophe.js'></script>
...

BOSH SERVICE (credit to Malcom)

specify the BOSH Service url

var BOSH_SERVICE = "http://chat.mydomain.com/http-bind";

Login

The login function is straight forward, except I am fixing the domain so that the client (user) need not remember it.

var DOMAIN = "chat.mydomain.com";

….

connection.connect($(‘#jid’).get(0).value + “@” + DOMAIN, $(‘#pass’).get(0).value, onConnect);

….

Send Message

• This function just sends the message, here I pass in the elemid which is a textarea.
• and JANE is my “helpdesk” so this widget only can only chat with her.
• sends a .c(“body”).t(text) instead of cnode(body)
• clears the textarea

var JANE = "Jane@chat.mydomain.com";

function sendMessage(elemid){
if (connection.connected && connection.authenticated) {
var text = $('#' + elemid).get(0).value;
if (text.length > 0) {
var from = Strophe.getNodeFromJid(connection.jid);
var to = JANE;
var reply = $msg({
to: JANE,
from: connection.jid,
type: "chat"
}).c("body").t(text);

connection.send(reply.tree());
log(text, "from");


$('#' + elemid).get(0).value = "";
}
}
else {
log("You have to log in before you can ask Jane");
}
}

Receive Message

• This function is a registered handler when a message arrives

function onMessage(msg){
var to = msg.getAttribute('to');
var from = msg.getAttribute('from');
var type = msg.getAttribute('type');
var elems = msg.getElementsByTagName('body');

if (type == "chat" && elems.length > 0) {
var body = elems[0];
log(Strophe.getText(body), "to");
}


return true;
}

> defaults write com.apple.dock persistent-others -array-add '{ "tile-data" = { "list-type" = 1; }; "tile-type" = "recents-tile"; }'

> killall Dock


if you’ve mis-typed and the Dock would no start again you can remove this domain(com.apple.dock) and key(persistent-others) by using

> defaults delete com.apple.dock persistent-others

and redo the above…. you can also use

> defaults read com.apple.dock persistent-others

to see the current values

defaults man page and more excellent tips from MacOSXTips.co.uk

It’s really just 3 steps :

1. mvn release:prepare -DdryRun=true
2. mvn release:clean release:prepare
3. mvn release:perform

But here are some things I learnt though…

1. Upon release:prepare errors, use release:rollback before release:clean (wipe out all the pom.* release.properties that’s needed for release plugin)
2. Undoing subversion changes using  svn merge -r 999:998 http://subversion/project/dir to revert back to the last working version. This is because release:prepare errors might have occurred after copy checked into subversion.

And finally, some pre-requisites :

Working copy must be SNAPSHOT

Dependencies must NOT be SNAPSHOT – to overcome this I had to  download and install a ‘special version’ into your own (local) maven repository :

mvn:deploy

will deploy to my scm definition in my pom.xml

<scm>
<connection>scm:svn:http://subversion/project/dir</connection>
<developerConnection>scm:svn:http://subversion/project/dir</developerConnection>
<url>http://subversion/project/dir</url>
</scm>

and reference to it in your repository

<repository>
<id>vocanic-m2</id>
<url>http://subversion/m2/</url>
<snapshots>
<enabled>true</enabled>
</snapshots>
</repository>

Mac OS subversion 1.4.4 – use this version because on Leopard, subversion 1.5.1 has a problem and somehow can’t create tags on subversion causing the release:prepare to check in the release copy but cannot move/create a tag copy – that’s where I learn (2) above – see here also

FireFox will cache your form values when you refresh a page…. I had this problem and was causing my animation to go haywire…. if you need to turn this auto caching off you can do :


<form autocomplete="off" /> or <input ... autocomplete="off" />


# The loopback network interface
auto lo
iface lo inet loopback

## Network interface(s)
## You should (un)comment and edit according to your needs.

# The primary network interface (dhcp)
#auto eth0
#iface eth0 inet dhcp

# The primary network interface (static IP)
auto eth0
iface eth0 inet static
address 192.168.1.45
netmask 255.255.255.0
#gateway 192.168.1.1
gateway 192.168.1.254
broadcast 192.168.1.255

\u@\h\


• user @ hostname

\[\e[1;4;33m\]

• enclosing “\[" and "\]” – to /not/ confuse the screen mode so that auto wrapping is preserved.
• “\e” – is the escape, you could use “33″ too
• “1;4;33″ – “1″ is for “bold”, “4″ is for “underline”, and “33″ is for “yellow”, you can add background colour by adding the ascii code for background colour – just separate these numbers with “;”

\[\e[0m\]

• 0 resets to normal display

\W

• working directory