Yep
2010/10/6 Blake Sullivan <blake.sullivan@oracle.com>On 10/6/10 12:01 PM, Leonardo Uribe wrote:
2010/10/6 Blake Sullivan <blake.sullivan@oracle.com>
On 10/6/10 11:35 AM, Leonardo Uribe wrote:HiNo. While the VDL may have a preferred extension, the authors should be free to route whatever content they want to the VDL. For example, in the particular case of running JSPX content through Facelets, .jspx would not be a preferred mapping for Facelets.
In theory, shouldn't the vdl be responsible to indicate which prefix/suffix pattern is used to identify if a physical viewId should be handled by an specific vdl?.
Ok, I see, that is for backward compatibility with apps that uses JSPX content for jsp. But in such case, which one is responsible to define such mapping?. Note right now it is not possible to create another vdl without break RestoreView algorithm. In theory, the ViewDeclarationLanguageFactory should do it (I did some attempts on this issue and finally I added some methods there).
If a particular JAR contains .jspx content that it wants to run through the facelets engine in a compatibility mode, that JAR would contain the mapping in its faces-config.xml and an application using that JAR wouldn't have to concern itself with that implementation detail.
But note this implementation detail is made explicit on the JAR, but the implementation proposed is just a workaround for facelets vdl. Let's take a look at how it is described on faces-config.xml
<faces-config-extension>
<facelets-processing>
<file-extension>.jspx</file-extension>
<process-as>jspx</process-as>From my point of view it says: "...for facelets vdl, if .jspx extension is found on the physical resource file, process it as some specific 'mode', and if .view.xml extension is found on the physical resource file, process it as xml...".
</facelets-processing>
<facelets-processing>
<file-extension>.view.xml</file-extension>
<process-as>xml</process-as>
</facelets-processing>
</faces-config-extension>
Since this is a facelets-specific extension, facelets can do whatever it wants with it
Does that suggest a vdl can process a file in different "modes"? Maybe yes, there could be some difference about how should be processed from facelets vld.
Are you talking about needing a custom way of coughing up resources, or that the scheme for finding resources needs to know about the VDL in some way? In the JAR case, it doesn't care about the VDL at all, it is simply loading the requested resources out of the JAR regardless of extension.
Now suppose the .jspx files are in some directory inside the JAR. How facelets can found them? ... through a custom javax.faces.view.facelets.ResourceResolver implementation ... Can we configure this artifact from faces-config.xml?.... no .... If we configure ResourceResolver interface that will work? .... no because the algorithm on RestoreView is very specific on that field and does not take into account ResourceResolver to find resources.
I agree that the resource loading scheme in JSF is weak and can be improved and that can be done with or without this enhancement. We could also later generalize the registration for other VDLs, but the fact that we would need to support options would make this much more complicated, so I question whether the effort and complexity would be worth it without further examples.
Things are different if we take as target files inside META-INF/resources (composite components). No ResourceResolver instance is involved and there will be no problem from RestoreView algorithm (because those files will not be used for create views).
In few words, my concern with this one is if this is being solved, why don't do it in a more standard way for all vdl and solve all those holes in the spec once for all?
Leonardo Uribe