Greg Wilkins
Greg Wilkins
> @gregw, this means that I can in Jetty reject `Content-Length` before the client is sending the body? Yes - and I think in Undertow also. > All of this...
I would like there to be a vibrant third party market of pluggable authentication and authorization modules.... but therejust isn't and JASPIC has not made one grow. I think implementations...
Here is another example that illustrates this question: ```xml Servlet0 org.example.ZeroServlet Servlet0 /foo/* X org.example.XFilter X Servlet0 *.txt /foo/* ``` Should a request to `/foo/bar.txt` invoke filter `X` once, twice...
See also #318 for a related clarification
I don't think that a container should rely on the application to not swallow the exception. But there are some questions that would be good to clarify: + if the...
@grgrzybek Great Work!!! > I hope this summary will be a good inspiration for improving the specification ;). Well at the very least it is inspiration for improving Jetty :)...
> What would be the problem specifying that access to "/c1" should _always_ be redirected to "/c1/" and then normal servlet lookup will take place? I think something should be...
Currently the jetty implementation interprets both `close()` and `complete()` method calls as a write operations, thus it cannot be legally called if there is a pending write operation. Specifically this...
@stuartwdouglas I think undertows is also a plausible interpretation of the spec - thus why we really need to narrow it down to one behaviour and to well describe it....