On Fri, 13 Apr 2001, Klaus Sonnenleiter wrote:
> Craig,
>
> since I'm not really familiar with what the standard says, I can't comment
> on that. But I can only tell you what I observed in other HTTP servers and
> it appears that most convert a %3F into a question mark some time before
> sending the request to the classloader or to the filter that looks for the
> file. My current problem is actually limited to a specific area and I think
> taking a calculated risk and deviating slightly from the standard (if
> that's what it is), would not be the worst of all options.
>
Remy has pointed out the RFCs for interpreting URIs.
> However, I would be interested in finding out what exactly the desired
> behavior for the final version of Tomcat 4.0 is.
The controlling authority for Tomcat 4.0 behavior is the Servlet 2.3 and
JSP 1.2 specifications (both in "Proposed Final Draft" status) at:
http://java.sun.com/products/servlet/download.html
http://java.sun.com/products/jsp/download.html
Both specs contain references to other specs they rely on as
well. Basically, the spec defines what a web application developer can
assume is portable across servlet containers.
There are lots of features in Tomcat 4.0 that go beyond the spec, which
are basically defined by the code contributions that are made. But the
specifics of the behavior you are talking about here seems to be pretty
clearly defined in the relevant specifications.
> Speaking of which, is
> there a target release date yet?
>
It doesn't make much sense to have Tomcat 4.0 go final until the specs do
(there have in fact been some minor changes -- and some clarifications are
still being discussed in the JSR-053 expert group that produced these
specs even as I type). Therefore, we'll continue to release "betas", with
increasing functionality and performance along with bug fixes, until then.
> Klaus Sonnenleiter
>
Craig