Re: valid XML checking

[Sachin Patel]

David, you bring up a good point on wether or not to save the converted version.  Currently the serialized version is fully qualified but I think we could provide a way for the editors to create and open up a converted in-memory only version of the plan and the serialized version would always remain unqualified.  The thing I'm not quite sure about is how to convert from the in-memory copy back to unqualified form.   - sachin On Mar 9, 2006, at 7:02 PM, Sachin Patel wrote: I'm already doing this on importing applications. Non-qualified plans can be imported, and I run the conversion to make it a valid doc so that the EMF models can process it.  Once its processed and the namespaces have been entered from that point forward you have to use qualified elements if you're modifying the resource via the Text/XML editor otherwise you invaliding the document and thus the editors can be opened against it.   I may be able to put in a hook to always run SchemaConverstionUtils on opening the editor which would resolve the problem.  Go ahead and open a defect.  But if you put unqualified elements and run the WTP xml validator its going to fail since technically it is an invalid document.  I could add a ResourceChangeListener or Builder as well which would help as well. But the "real solution" will be for users not have to modify the xml at all and there needs to be a complete set of deployment plan editors.  I've been very behind on this effort, what would help would be if someone could mock up how they wanted things to be layed out in the editors and then I could get it done much quicker. - sachin On Mar 9, 2006, at 2:59 PM, David Jencks wrote: The builders all run plans through SchemaConversionUtils to install the correct namespaces.  Is there some way to get eclipse to do that? I don't know if it would be a good idea to then save the converted version. thanks david jencks On Mar 9, 2006, at 10:53 AM, Lin Sun wrote: Hi there,   I’ve spent quite some time trying to get the MDB sample (http://opensource2.atlassian.com/confluence/oss/display/GERONIMO/JBoss+to+Geronimo+-+EJB-MDB+Migration) deployed using the eclipse plugin.   I was able to get it running eventually but I had to spend extra time in modifying the deployment plan to bypass the valid XML checking.   The extra editing is are not needed if I use the command line deployer or admin console.   The problem is that the plugin checks the deployment plan very strictly.  I have to use fully qualified elements, otherwise the plans are not treated as valid xmls, and the model editor will not be able to parse the resource and throws exception in the eclipse log.   For example:  the following is not treated as valid Geronimo deployment plan in eclipse environment.   But thecommand line deployer or admin console likes it:     <ejb-ref>     <ref-name>ejb/CustomerHome</ref-name>     <target-name>       geronimo.server:EJBModule=MDBDemo,J2EEApplication=null,J2EEServer=geronimo,j2eeType=EntityBean,name=CustomerEJB    </target-name>   </ejb-ref>   I had to change it to the following just for Eclipse environment.  (xmlns:nam=http://geronimo.apache.org/xml/ns/naming-1.0 is specified earlier in the plan)     <nam:ejb-ref>     <nam:ref-name>ejb/CustomerHome</nam:ref-name>     <nam:target-name>       geronimo.server:EJBModule=MDBDemo,J2EEApplication=null,J2EEServer=geronimo,j2eeType=EntityBean,name=CustomerEJB     </nam:target-name>   </nam:ejb-ref>   Ideally, we don’t want users to perform such extra efforts while porting their Geronimo specific plans in eclipse.   I am not sure what can be done in the plugin code to perform a loose XML checking… it seems to use the WTP/pre-requisites’ code to perform the XML checking.   Thoughts?   Thanks,   Lin