php-java-bridge-users Mailing List for PHP/Java Bridge

The problem is that jetty doesn't flush the response buffer. In pseudo code:

protected void doPut (HttpServletRequest req, HttpServletResponse res) {

   res.getOutputStream().write(1);
   res.setContentLength(1);

   try {res.flushBuffer(); } catch (Exception t) {t.printStackTrace(); }
   try {resOut.close(); } catch (Exception t) {t.printStackTrace(t);}
   // response must be flushed right here, not at the end of the
service() routine

   Thread.wait(15000);

 }


> Connection:keep-alive

It's a HTTP/1.0 request. If jetty sets keep-alive for a HTTP/1.0
request, we've found another bug, I think.


BTW: Jetty contains another serious bug:

  requestDispatcher.include(req, res);

If req is an instance of HttpServletRequestWrapper, jetty tries to
find the original request in the thread context. This fails if the
current thread is not a servlet thread. You can verify this bug by
clicking on the jsr223.jsp example of the JavaBridge.war; jetty
crashes with a NullPointerException.


Regards,
Jost Bökemeier



On 3/8/10, php...@li...
<php...@li...> wrote:
> I filled a bug with Eclipse and I was asked:
>
> "Greg Wilkins 2010-03-05 02:27:12 EST
> Can you give a bit more information about the request that is not being
> flushed.
>
> Does the response have Connection:keep-alive set and a Content-Length?"
>
> Do you know the answer? I'll update the bug report.
>
> Thanks,
>
> McKinley
>
> On Fri, Feb 19, 2010 at 5:52 AM,
> <php...@li...> wrote:
>> This is indeed a Jetty 7 bug. Jetty doesn't flush the output buffer
>> "res.flushBuffer()" for HTTP/1.0 connections, so the code runs into a
>> wait timeout. I won't change the bridge code to work around this bug.
>> But Jetty 7.0  supports HTTP/1.1 chunked connections:
>>
>> Please add
>>
>> <context-param>
>>           <param-name>promiscuous</param-name>
>>           <param-value>true</param-value>
>> </context-param>
>>
>> after the </listener> declration of your WEB-INF/web.xml to work around
>> this Jetty 7 bug.
>>
>>
>> Regards,
>> Jost Bökemeier
>>
>
> ------------------------------------------------------------------------------
> Download Intel&#174; Parallel Studio Eval
> Try the new software tools for yourself. Speed compiling, find bugs
> proactively, and fine-tune applications for parallel performance.
> See why Intel Parallel Studio got high marks during beta.
> http://p.sf.net/sfu/intel-sw-dev
> _______________________________________________
> php-java-bridge-users mailing list
> php...@li...
> https://lists.sourceforge.net/lists/listinfo/php-java-bridge-users
>

I filled a bug with Eclipse and I was asked:

"Greg Wilkins 2010-03-05 02:27:12 EST
Can you give a bit more information about the request that is not being
flushed.

Does the response have Connection:keep-alive set and a Content-Length?"

Do you know the answer? I'll update the bug report.

Thanks,

McKinley

On Fri, Feb 19, 2010 at 5:52 AM,
<php...@li...> wrote:
> This is indeed a Jetty 7 bug. Jetty doesn't flush the output buffer
> "res.flushBuffer()" for HTTP/1.0 connections, so the code runs into a
> wait timeout. I won't change the bridge code to work around this bug.
> But Jetty 7.0  supports HTTP/1.1 chunked connections:
>
> Please add
>
> <context-param>
>           <param-name>promiscuous</param-name>
>           <param-value>true</param-value>
> </context-param>
>
> after the </listener> declration of your WEB-INF/web.xml to work around
> this Jetty 7 bug.
>
>
> Regards,
> Jost Bökemeier
>