On 29/06/2011 00:34, Reinhard Pötz wrote:
On 06/28/2011 10:29 AM, Francesco Chicchiriccò wrote:
Gents,
Cocoon 3 is using Cocoon Spring Configurator 2.1[1]; by default this is
triggered by an empty
<configurator:settings/>
in main applicationContext.xml.
In my project I need to:
1. always reload sitemap.xmap from the filesystem during development
Am I right that you use C3 within a Wicket application and can't use
the "default" reloading classloader stuff based on the Cocoon Maven
Plugin (mvn cocoon:prepare)?
Actually not: I gave up in using the cocoon-sample-wicket-webapp as a
base for my application because of the missing "servlet:/" feature.
I need the reloading feature that a block has (thanks to the Cocoon
Maven Plugin) but I also want to build a webapp that can be overlayed.
If I understood well, the cocoon-sample / cocoon-sample-webapp pair does
not meet these needs barely because the cocoon-sample-webapp is an empty
shell for cocoon-sample. Hence, if you overlay a WAR produced by
cocoon-sample-webapp, you can't handle into the sitemap stuff.
This is the reason why I am building my project (a Cocoon3-based
framework for delivering content managed by the enterprise Hippo CMS -
http://forge.onehippo.org/gf/project/hct/) as a maven WAR artifact with
more or less the same "developer-friendly" features that a block has.
Basically, my META-INF/cocoon/spring/servlet-service.xml declares the
XMLSitemapServlet as the following:
<bean name="org.onehippo.forge.hct.service"
class="org.apache.cocoon.servlet.XMLSitemapServlet">
<servlet:context mount-path="" context-path="classpath:/COB-INF/"/>
</bean>
2. catch the real cause of a ProcessingException (I was inspired by [2])
yes, makes sense.
Hence I set something like as
<configurator:settings>
<configurator:property name="org.apache.cocoon.reloading.sitemap"
value="true"/>
<configurator:property
name="org.apache.cocoon.exception.ProcessingException.unroll"
value="true"/>
</configurator:settings>
Then, I modified [3] and [4] in order to handle this kind of
configuration: the patch is attached to this e-mail.
Do you think this could be useful in general? If so I would commit
gladly.
If my "C3 within Wicket" assumption is right, I'm fine with your
suggestions.
What do you think, then?
Thanks.
--
Francesco Chicchiriccò
Apache Cocoon Committer and PMC Member
http://people.apache.org/~ilgrosso/
