On 05/06/2012 20:57, Konstantin Kolinko wrote:
> 2012/6/5 <[email protected]>:
>> Author: markt
>> Date: Tue Jun 5 18:17:58 2012
>> New Revision: 1346510
>> ---
>> tomcat/trunk/test/org/apache/tomcat/util/http/mapper/TestMapperContextRoot.java
>> (added)
>> +++
>> tomcat/trunk/test/org/apache/tomcat/util/http/mapper/TestMapperContextRoot.java
>> Tue Jun 5 18:17:58 2012
>> @@ -0,0 +1,79 @@
>> + private static class Bug53356Servlet extends HttpServlet {
>> +
>> + private static final long serialVersionUID = 1L;
>> +
>> + @Override
>> + protected void doGet(HttpServletRequest req, HttpServletResponse
>> resp)
>> + throws ServletException, IOException {
>> + // Confirm behaviour as per Servler 12.2
>> + boolean pass = "/".equals(req.getPathInfo());
>> + if (pass) {
>> + pass = (req.getServletPath() == null);
>> + }
>> + if (pass) {
>> + pass = (req.getContextPath() == null);
>> + }
>> +
>
> Looking into Servlet spec 3.0 rev.a chapter 12.2.
>
> It says that contextPath and servletPath are empty string (""), but
> the code above expects nulls.
I'll get that fixed. I mis-read the spec.
> Regarding contextPath being "":
> It is surely strange, as
>
> 1. it contradicts other places like 3.5 (definition of "Context Path"
> and famous equation of requestURI = contextPath + servletPath +
> pathInfo ).
>
> 2. There is also ServletContext.getContextPath() that provides the
> correct information.
>
> 3. Though looking at Javadoc [3], it says that
> HttpServletRequest.getContextPath() may be different with
> ServletContext.getContextPath(), but it does not mention this special
> servlet mapping as a case.
>
> [3]
> http://docs.oracle.com/javaee/6/api/javax/servlet/http/HttpServletRequest.html#getContextPath%28%29
>
> I wonder if it was ever raised with the EG.
Not that I am aware of.
> It is worth mentioning the chapter number (12.2) in the comment in Mapper.java
>
> I suspect that this feature is not covered by official tests, because
> otherwise we would have had to deal with it earlier.
Correct. I'm not surprised.
> So maybe we should respect ch.3.5 here, instead of 12.2?
Until the EG says otherwise, I think we should go with 12.2
Mark
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
