I get the rationale behind the implementation details here -- what I don't have insight into are what extensions are needed that can't be achieved within the existing spec/extension mechanism. Thx for cleaning up the imports.
On Tue, Sep 22, 2009 at 3:07 PM, Paul Lindner <[email protected]> wrote: > On Tue, Sep 22, 2009 at 2:59 PM, <[email protected]> wrote: > > > One nit, but the majority concern I have is in the overall result of > > this approach. Seems to me this encourages ad hoc modification of the > > spec, leading to inconsistencies and incompatibility between gadgets. > > Why not use the existing <Requires>/<Optional> syntax? It's limited, but > > typically can get the job done, using a known pattern. > > > > I agree in principal, however in practice we've seen the need at the > implementer level to add extensions in this area. Henning has a patch that > implements a stax-based parser that allowed for namespaced extensions. I > didn't have time to test and merge that patch so I went ahead and added in > the functionality you see. > > > > > http://codereview.appspot.com/121064/diff/1/2 > > File > > > java/gadgets/src/main/java/org/apache/shindig/gadgets/spec/ModulePrefs.java > > (right): > > > > http://codereview.appspot.com/121064/diff/1/2#newcode26 > > Line 26: import org.w3c.dom.*; > > please expand these wildcards, verbose as it may be. > > > > > Silly IDE.. I'll fix it.. > > > > http://codereview.appspot.com/121064 > > >
