read more...

read more...

This is something that hasn't been made clear enough in the documentation, so I thought it would be a good idea to give a short overview.

Configuration

First, set these configuration parameters to false, their purpose should be self-explanatory:

<param name="ELEMENT_AUTO_RELOAD">false</param>
<param name="SITE_AUTO_RELOAD">false</param>
<param name="TEMPLATE_AUTO_RELOAD">false</param>
<param name="L10N_RESOURCEBUNDLE_AUTO_RELOAD">false</param>

By disabling all these parameters, RIFE will not automatically check for changes in these file types and your application will run at full-speed.

Note that when you set up you repository configuration participant like this:

<participant param="XmlSelectorHostname">ParticipantConfig</participant>

You can create configuration files that are specific for each host (development, production, ...). In each of these files you can then set the appropriate configuration parameters and still include a common configuration file that contains all the application settings.

This site for example has the following configuration files:

src/
    rep/
        config-common.xml
        config-dawn_uwyn_com.xml
        config-elm_uwyn_office.xml
        config-ferrari_uwyn_office.xml
        config-pine_uwyn_office.xml

The host dawn.uwyn.com is the production server and all the others are development machines. They all include config-common.xml.

Packaging

Next, you should pre-compile all your templates during the packaging of your web application. This is handled by the com.uwyn.rife.template.TemplateDeployer class. You will typically do this in your Ant build process, this is an example of a task that does just that:

<target name="precompile">
    <echo message="Precompiling html templates:"/>
    <java classname="com.uwyn.rife.template.TemplateDeployer"
            failonerror="true"
            fork="true">
        <arg value="-verbose"/>
        <arg value="-t"/>
        <arg value="enginehtml"/>
        <arg value="-d"/>
        <arg value="${build.classes}"/>
        <arg value="-encoding"/>
        <arg value="UTF-8"/>
        <arg value="${templates.dir}"/>
        <classpath refid="application.classpath"/>
        <classpath refid="lib.classpath"/>
    </java>
</target>

You will need to do this for each template type that you use. After this process, your templates will be regular Java classes and will not have to be parsed during production anymore. Your application will thus respond instantly once it has been started up.

In two weeks I will be giving a talk about RIFE at Fosdem in Brussels, Belgium. The final date and time are set, it will be on Sunday the 26th at 10h00 in the morning.

Given the usual Saturday night drinking and pub discussions, the audience will probably be slim but that's ok. I'm going to test-drive a completely new presentation I've been writing for TSSJS 2006. This talk will give me the opportunity to fine-tune it and see what works and what doesn't.

My RIFE session proposal has been accepted as a BOF for JavaOne 2006. I'm very glad that it also entitles me to a full complimentary pass to the conference, since otherwise I wouldn't have been able to afford attending (airfare, hotel and entrance totaled to more than €4000).

If you can't make it to JavaOne, check out our conference agenda for other RIFE sessions, straight from the source.

These are the details of my submission, and what you can expect:

Cutting-edge productivity with RIFE

About RIFE

RIFE is a full-stack, open-source Java web application framework, offering fast results with the promise of maintainability and code clarity. While many are attracted to RIFE for the productivity gains it offers, the framework actively tries to solve difficult problems with creative solutions that combine power and flexibility with comfortability. RIFE has pioneered many novel ideas in Java web application development, including: pure Java web continuations, run-time POJO-driven CRUD generation, scalable component oriented architecture, state handling without sessions, focus on REST as an application API, bi-directional logicless template engine, and the integration of a content management framework.

About this session

This session demonstrates how RIFE lets you cut corners when there's no need for a complex setup. You'll see how trivial CRUD functionalities are instantly available and how they can completely be customized without the need of restarting your servlet container. The rich POJO-driven meta-data system will be demonstrated as well as how it propagates and drives most of the framework's modules. We'll also develop a component from scratch and introduce the web engine, the template engine, the persistence layer and the content management framework.

Pre-requisits

Attendees should have experience with dynamic web application development and have at least an intermediate level of Java programming knowledge.