[richfaces-issues] [JBoss JIRA] (RF-5460) Strange a4j resource prefix with url-based servlet mappings

Strange a4j resource prefix with url-based servlet mappings ----------------------------------------------------------- Key: RF-5460 URL: https://issues.jboss.org/browse/RF-5460 Project: RichFaces Issue Type: Bug Security Level: Public(Everyone can see) Components: core Affects Versions: 3.2.2 Environment: JSF RI 1.2.11, RF 3.2.2SR1, Facelets 1.1.15b1, Glassfish v2. Reporter: Fab Mars Priority: Minor Labels: a4j, mapping, prefix, servlet My web.xml has several servlet mappings. <servlet-mapping> <servlet-name>Faces Servlet</servlet-name> <url-pattern>*.jsf</url-pattern> </servlet-mapping> <servlet-mapping> <servlet-name>Faces Servlet</servlet-name> <url-pattern>/somepath1/*</url-pattern> </servlet-mapping> <servlet-mapping> <servlet-name>Faces Servlet</servlet-name> <url-pattern>/somepath999/*</url-pattern> </servlet-mapping> I'm doing this because I'm using some phaselistener that preproceses some things and also enables short-urls. Now, any page has a4j resources written like this: <script type="text/javascript" src="/MyContext/somepath999/a4j/g/3_2_2.SR1org/richfaces/renderkit/html/scripts/skinning.js"> That is all a4j resources are now comprising the LAST servlet-mapping, just between the context name and the /a4j/ path part. I tried to workaround it, by putting the *.jsf mapping as the last in the web.xml but to no avail: the last servlet mapping path was always used. To reproduce, just download the demo application: http://download.jboss.com/jboss-richfaces/richfaces-demo-3.2.2.SR1-jee5.war Edit web.xml and add to the existing *.jsf mapping the following: <servlet-mapping> <servlet-name>Faces Servlet</servlet-name> <url-pattern>/yeswecan/*</url-pattern> </servlet-mapping> Then deploy and look at the source of any page: <img class="rich-spacer " height="1" id="j_id258" src="/richfaces/yeswecan/a4j/g/3_2_2.SR1images/spacer.gif" width="1" style="height:10px;" /> I'm not sure this is a bug or a feature of some sort. The code obviously takes ANY valid servlet mapping, which is fair enough. But maybe it should just check whether there is an extension-based mapping amongst them, and if yes, use it. Worse, this issue is also probably (conceptually) linked to https://jira.jboss.org/jira/browse/RF-3586 Anyway, you're free to requalify this issue if you want. I'm reporting what I found and I'll find a workaround in the meantime. Thanks.