On Wed, 28 May 2003, Fernando Martins wrote:
>
>
> On Wed, 28 May 2003, Jason Carreira wrote:
>
> > >
> > > Doesn't it allready does this with the exception of
> > > extracting servletPath according to the spec? If this is a
> > > problem of ambiguous spec then maybe we should determine what
> > > the majority of containers are doing, and also submit this
> > > ambiguity to Sun. I thought Jetty was famous for keeping
> > > strict to spec. So if in Jetty url-pattern * doesn't work,
> > > maybe it's not in the spec. Then if /* returns paths as
> > > described above, it's not too much work to adapt
> > > ServletDispatcher to it (to spec?). Should I really need a
> > > custom dispatcher? I'm not implemeting nothing new, and it's
> > > not a question of making it work in 2 containers, but just to
> > > make it work according to spec, which in the end means what
> > > should be in url-pattern.
> > >
> > > Thanks
> > > Fernando
> > >
> >
> >
> > ServletDispatcher is based on finding the action name by using "." as
> > the delimter:
> >
> > // Get action name ("Foo.action" -> "Foo" action)
> > int beginIdx = name.lastIndexOf("/");
> > int endIdx = name.lastIndexOf(".");
> >
> > return name.substring(((beginIdx == -1) ? 0 : (beginIdx + 1)),
> > (endIdx == -1) ? name.length() : endIdx);
> >
> > If we refactored out a parent class, we could have 2 subclasses, one of
> > which finds the Action name from the "." and the other of which uses
> > everything after the last "/" as the action name.
> >
> > Jason
>
> Yes but if "." doesn't exist it allready uses everything after the last
> "/" without throwing Exceptions, so in fact it's allready doing both
> options. The problem is how to get the name to parse from the request, and
> that should be done according to spec.
>
> I only see reason for 2 classes if there is really some conflicting
> behaviour that will not be possible to have transparently in
> the same class.
>
> Even if it will be really necessary 2 classes, you still have to figure
> out what should go in url-pattern for it to work.
>
> Fernando
Replying to my own post...
I was thinking that being ServletDispatcher in the heart of request
processing, it should be as fast as possiblee. So maybe the reason
for 2 subclasses is not so bad in regard for performance.
Less "if's" :)
Fernando
-------------------------------------------------------
This SF.net email is sponsored by: ObjectStore.
If flattening out C++ or Java code to make your application fit in a
relational database is painful, don't do it! Check out ObjectStore.
Now part of Progress Software. http://www.objectstore.net/sourceforge
_______________________________________________
Opensymphony-webwork mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/opensymphony-webwork
