Archive for June, 2009

LiveCycle Data Services (LCDS) – Quick Start

Posted in Flex/AIR, LiveCycle Data Services with tags , , , , on June 16, 2009 by devgirl

With the recent work I’ve been doing to add LiveCycle Data Services samples to the Flex Data Access category for Tour de Flex, I felt the need to add some additional helper information to assist people in getting started using it quickly. I’m including some basic information on how to configure the client (using Flex 3) and server side, and provide links to more detailed information as needed. This post is not intended to explain all things Data Services, but I am hoping to provide some overview information for those just beginning to take a look at it, as well as some helpful info and links that might be needed to get on board fast!

So… you want to use Data Services, where do you begin? First I suggest you think about and try to answer the following questions:

  • What kind of data access do I need or what does the service need to do?
  • What platform/language do I need to use (Java, .NET, Cold Fusion etc)?
  • Will there be a database needed, and if so how complex will the table structure be?
  • Will there be any other layers involved, such as Hibernate for instance?

LiveCycle Data Services (LCDS) – Feature Overview

First, a quick overview of the different features provided by LCDS…

RPC Services
Making calls to a remote service or operation is super easy using LCDS. There are various ways to achieve this depending on protocol you want to use etc… they include an HTTP call (using the HTTPService class),  a SOAP call (using the WebService class) or using AMF via Remoting (using the RemoteObject class). I strongly encourage using Remoting where possible due to the significant gains in response time. James Ward has a fantastic and powerful example of different response times for different types of invocations here. The best way to use this app is to select the ‘Census – Guide Me’ option from the drop down list and go through each type of call to get the full effect of response times.

Data Management
If you need to persist and manage data between clients, LiveCycleData Service’s Data Management is definitely the way to go. Using the DataService object, you are immediately given amazing power over your data, including CRUD operations (Create, Read, Update, Delete) with conflict management handling options, auto-commit of data, persisting to a local cache when offline (in an AIR app), and so much more. And what’s even better is the little bit of code needed to make all of this happen! It is still totally amazing to me how powerful yet easy it can be compared to anything I’ve done in the past. The Tour de Flex samples for Data Management are the best resource to check out to get started in this area quickly.

Messaging
The most common use of messaging is in a one-to-many relationship, when there is information that should be broadcast real-time to many at once. This is often called publish-subscribe messaging, or topic-based messaging. The topic is the shared item that a ‘producer’ (the one sending the message) and the ‘consumer’ (the one receiving the message) both subscribe to. A general but explanatory example often used for publish-subscribe messaging is that of subscribing to a stock ticker topic that is broadcasting all of the stock prices real-time to its subscribers. Another type of messaging supported by LCDS is point-to-point messaging, which involves a one-to-one relationship. All of this is described in-depth in the LCDS Developer documentation for those looking for more information. Adobe offers the open-source BlazeDS that offers most of the same Messaging features as well. More info on BlazeDS is given later in this post.

What next?

So once you decide what you need to do and which LCDS features you need to use, I recommend going to the Flex Data Access category in Tour de Flex and locating a sample for the specific area of LCDS you’re planning on implementing. If your solution involves  server side code, for instance if you’re doing Remoting or Data Management, the following steps detail what you should do next to configure the server side:

Server-side Configuration Steps
1.    Create your project in your favorite IDE (I use Eclipse) and be sure to include the following libraries (found in the LCDS or BlazeDS install packages):

a.   acrobat-core.jar
b.   acrobat-core-charset.jar
c.   backport-util-concurrent.jar
d.   cfdataservicesadapter.jar – (Cold Fusion only)
e.   cfgatewayadapter.jar (Cold Fusion only)
f.    commons-codec-1.3.jar
g.   commons-httpclient-3.0.1.jar
h.   commons-logging.jar
i.    concurrent.jar
j.    flex-acrobat
k.   flex-messaging-common.jar
l.    flex-messaging-core.jar
m.  flex-messaging-data.jar
n.   flex-messaging-data-req.jar
o.   flex-messaging-opt.jar
p.   flex-messaging-proxy.jar
q.   flex-messaging-remoting.jar

2.    Create a destination in the appropriate XML configuration file (remoting-config.xml, data-management-config.xml etc.)  The next section has more information about configuration files and LCDS.

3.    Point the destination to the appropriate server side ‘assembler’, (if using Data Management) or class you are using to perform the server side RPC operations and set required parameters on that destination.

Note:  An assembler is a general term for the server side code that will be used to process your interactions with your data. There are a few different types of assemblers built into LCES, or you can extend their AbstractAssembler class to create your own custom class. Examples of each are part of the recent additions to Tour de Flex, and more information about the Java classes used can be found here

4.    Copy your compiled class files for your service or assembler into the WEB-INF/classes directory. Ensure any other configuration files needed for things like a Hibernate integration etc are added to their appropriate server location as well. Note: Tour de Flex has more information regarding configuring LCDS with Hibernate.

For a detailed explanation of the server side configuration needed for your web application that uses LCDS, see here.

Client-side Configuration Steps
1.    Create your Flex project. There are actually a couple ways you can go about doing this. The conventional way will include your required libraries in the build path and set up your compiler options as needed (adding the –services option) so you are ready to use the various LCDS components in your application. However, I personally find the non-conventional way along with using runtime configuration of channels easier and feel that I have more control over it that way.

a.    Conventional way: this link sums it up
b.    Non-conventional way: create a basic Flex project (no server specified) and add the necessary SWC’s to the included libraries for the build path (fds.swc, playerfds.swc and the locale folder found in your webapps WEB-INF/flex/libs folder). Note: these SWC’s are also shipped with LCDS

As I said above, using this option requires you to use runtime channel definitions and set them on your LCDS components, but also offers you full control over which channel is used when, and allows you to easily switch where one is pointing (for instance localhost versus another server destination). Also important to note, configuring this way does not require the services configuration to be compiled into your SWF. An example of configuring channels at runtime is shown below, and used in all of the Tour de Flex Data Access samples:

    <mx:ChannelSet id="channelSet">
         <mx:RTMPChannel id="rtmpChannel" url="rtmp://tourdeflex.adobe.com:2037"/>
    </mx:ChannelSet>
        <mx:DataService id="ds" destination="hibernate-contact" autoCommit="false" channelSet="{channelSet}" fault="onFault(event)"/>

Config file C-R-A-Z-Y! Why are there so many configuration files and what are they used for?

When I first started using LCDS, it seemed a little confusing to me that there were so many different XML files for configuring things, I was just accustomed to having to edit one main config file for many of the things I had worked on, so to have to touch all the different ones for the different services seemed a bit overwhelming at first.

Now that I have worked with it for awhile, I realize there really is one main configuration file called services-config.xml, and it just includes the rest of the config files for the specific services (remoting-config.xml, data-management-config.xml, messaging-config.xml etc).  The services-config file defines channels, logging and security settings, and then has the includes for the rest of the configuration files.  You would typically not specify application level destinations here, but would put them in the appropriate *-config.xml file based on the guidelines below:

remoting-config.xml – set up your destinations used for remoting here (ie: RemoteObject from your Flex client)

data-management-config.xml – set up your destinations for Data Management applications here (ie: those referred to from the DataService object in your Flex client).

messaging-config.xml – set up your destinations for messaging applications here (ie: those referred to from  Producer/Consumer objects from your Flex client).

proxy-config.xml – defines Proxy Service destinations for working with web services and HTTP services (REST services).

Channel Confusion! Which, when and why?

Channels are used in LCDS to provide the transport of your data from end to end over a specified protocol (HTTP, AMF, RTMP etc).  They specify an endpoint for your data to be sent to as well as handle how it should get there. Channels are defined in the services-config.xml file, and there are many different types that can be defined for them. It can be very confusing to try to read the documentation given and figure out what you actually need. Damon Cooper has a fantastic and detailed summary of the different channels depending on what you might need located here.

BlazeDS

BlazeDS is the remoting and HTTP-based messaging technology Adobe is contributing to the community under LGPL v3. BlazeDS offers publication of the Action Message Format (AMF3) binary data protocol specification. Certified builds, warranty protection and enterprise support subscriptions are available. A comparison of the features included in LCDS versus BlazeDS can be found here. There’s also a full test drive available in BlazeDS you can use to check out the included features.

References

The LCDS Gurus

Christophe Coenraets – site includes TONS of cool samples using more advanced LCDS features and information. Christophe has a way of explaining the more complex features in a manner that seems much less intimidating and always supplies source code so you can see exactly what he did to make something happen.
Sujit Reddy G – again lots of great info on LCDS, BlazeDS and more, but some particularly helpful info on using Flash Builder 4 with Data Access.
Dr Flex – a very cleverly written site packed with some great and useful information on a variety of topics.
Cornel Creanga –  former J2EE Software Developer turned Adobe Evangelist with a lot of useful information and samples to check out.
Greg Wilson – great blogger and Adobe Evangelist with a bunch of LCDS, LiveCycle and Tour de Flex sample information.
LCDS Livedocs/Developer’s Guide – in-depth information related to Data Services, includes all the details for the subject matter covered in this post.

LiveCycle Data Services – included samples – included in LCDS is a great set of samples that show off the different features. Some of these were already the basis of the samples added recently to Tour de Flex, but it is in you best interest to start by installing LCDS, running the samples locally to see how they work, and then checking out the included source code. For instance, when running locally your path to the LCDS samples webapp might be: http://localhost:8080/lcds-samples/ depending on your configuration. Don’t miss the ‘Take the Test Drive’ option as well… that is actually the best place to start as some of the other samples on the main samples page become a bit more advanced.

Advertisements

Launch an AIR application from within another AIR application

Posted in Adobe AIR, Flex/AIR with tags , , on June 8, 2009 by devgirl

Have you ever needed to launch another AIR application from within an AIR application? I have found myself wondering how this would be done exactly, and noticed that Rich Tretola just did a great blog post this morning on two ways to do it, including source code. He presented the solutions at 360 | Flex recently and was kind enough to share it on his blog. He shows one solution using an undocumented AIR class called ProductManager, and another of how to load from the Browser API air.swf in the event that you are not sure the other AIR app is installed on the user’s machine. Check it out if you find yourself in need of this functionality in the future!