On 11/26/08, Felix Meschberger <[EMAIL PROTECTED]> wrote:
> Hi all,
>
>  After the big reorganization of servlet resolution in April of this year
>  and some more experiences with using Sling, I propose the extension of
>  our servlet/script resolution process.
>
>  Actually it is a single change only: Consider request selectors for all
>  requests. Currently request selectors are only considered for GET or
>  HEAD requests and ignored for any other requests.
>
>  Over time use cases arose, which would be easier implementable if we
>  would support using selectors to influence servlet/script resolution for
>  all methods.
>
>  Example: Say you are building an application providing a configurable
>  input form system. The most important configuration would be the action
>  upon form submission. Submitted forms are sent back to the resource
>  providing the form. The actual action would be encoded as a request
>  selector:
>
>  For example we would like to be able to send a form based mail message.
>  We would have script "/apps/myapp/form/mail/POST.esp", which handles the
>  "myapp/form" resource type and is called using the "mail" selector. The
>  form itself together with some nice GUI surroundings etc. would be
>  stored in the /content/feedback resource.
>
>  Now updating the /content/feedback resource should be handled by Sling's
>  default POST servlet. Thus POSTing to /content/feedback would update the
>  content. Submitting the form would POST to /content/feedback.mail.html
>  and should thus not store any form input but trigger the mail/POST.esp
>  script.
>
>  I have prepared a prototype implementation of this new mechanism in my
>  sandbox at [1].
>
>  A nice side effect of this prototype is, that it is much simpler codewise.
>
>  WDYT?
+1 to allow selector mapping for all methods.

regards, toby

>
>  Regards
>  Felix
>
>  [1]
>  
> http://svn.apache.org/repos/asf/incubator/sling/whiteboard/fmeschbe/servlet_resolver
>

Reply via email to