Just FYI .. lets see if this is something we can expect to be fixed in
the JDK, and understand any reasons behind the current behavior
http://bugs.sun.com/bugdatabase/view_bug.do?bug_id=7045148
asankha
On 05/13/2011 03:18 PM, Oleg Kalnichevski wrote:
On Fri, 2011-05-13 at 14:00 +0530, Asankha C. Perera wrote:
Hi Oleg
Creating a native HttpCore test case will take some time as I will be
traveling from tomorrow .. meanwhile I came across a somewhat hacky
solution to the issue and wanted to check how you feel about it.. I do
believe the real bug is in Java, and I guess we should file a bug :)
The solution would be to throw an IllegalStateException instead of a
ConnectionClosed, and requesting our user to handle the case by way of
Javadocs. I know this is not the most elegant way to solve this.. but a
Java fix could take a long time.. What do you think?
thanks
asankha
Asankha
I think I have a slightly less hacky work-around for the problem. How
about creating a custom version of ContentDecoderChannel that catches
ConnectionClosedException and re-throws it as a runtime exception of
some sort? This would enable your code to handle truncated messages and
would not require the hack to be applied to HttpCore? Would that work
for you?
I also believe we should release HttpCore 4.1.1 without a fix for the
problem and take more time to work out a proper solution in the 4.2
branch. Would that be acceptable?
Oleg
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
--
Asankha C. Perera
AdroitLogic, http://adroitlogic.org
http://esbmagic.blogspot.com
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
