The problem with the mechanism introduced in JSF-2.0 is to rely on Expression Language evaluation for the dynamic parts. But that doesn't perform well so the JSF containers serve it with a cache header. Please checkout Jakobs enhanced resource handler [1] (especially the Requirements Wiki page) and based on it the myfaces-commons-resourcehandler [2] . It's pluggable. The issue should be fixed in JSF-2.2 (or latest 2.3, not sure if Ed has it on the schedule already).
LieGrue, strub [1] http://www.jakobk.com/2011/11/bachelor-thesis-about-relative-resource-handler [2] https://cwiki.apache.org/MYFACES/myfaces-commons-resourcehandler.html ----- Original Message ----- > From: "Carlson, John W." <carlso...@llnl.gov> > To: MyFaces Discussion <users@myfaces.apache.org>; Mark Struberg > <strub...@yahoo.de> > Cc: Java Programmers List <j...@llnl.gov> > Sent: Thursday, February 16, 2012 10:36 PM > Subject: RE: CODI 1.0.1 NPE in JSF ResourceHandler when evaluating > ValueExpression > > Uh, are you guys saying the FacesContext isn't available in a > ResourceHandler? If it's not available, this is a major pain for people > with dynamic resource generation. Essentially, one would have to pass all > the > necessary information through the URL, or through a non-JSF session. > > I realize that static resource generation is more desirable, and that we're > moving towards HTML5 with dynamic stuff happening on the browser. > > If this is not fixed with JSF2, then I would say my statement about JSF not > being suitable for a readonly/downloads application is still valid. >