...source requests to the web root.
[ARQ-818] Expand WebArchiveClassLoader implementation to redirect all…
… resource requests to the web root.
I'd prefer to have some kind of fallback implemented as well - classloader should also try to find original resource name if the lookup of adjusted name fails.
Are you sure? What root do you have in mind? I don't see anything like this in servlet 3.0 and Java EE 6 spec.
By the way how about WEB-INF/lib lookup?
10.7.2 Web Application Class Loader: "The class loader that a container uses to load a servlet in a WAR must allow the developer to load any resources contained in library JARs within the WAR following normal Java SE semantics using getResource."
To get at your primary question regarding the 'fallback":
Let's invert this and ask another way: When should CL.getResource() requests relative to the root, and not "WEB-INF/classes", ever be allowed? We should be as strict as the spec allows.
pushed upstream, 278e7d4