FWIW- Looks like that last NPE will occur if the ByteBufferWrapper() I create to write to the response is "dispose()"d prematurely. I'm trying to do micro optimizations on internal buffer handling to eek out every last drop of performance.
jb
On Jul 20, 2011, at 11:15 AM, Oleksiy Stashok wrote:
> On 07/20/2011 06:10 PM, Jon Brisbin wrote:
>> On Jul 20, 2011, at 11:03 AM, Oleksiy Stashok wrote:
>>
>>> Which Grizzly version are you using? Is it one taken from the trunk or 2.1.1?
>> 2.1.1
> Please try to checkout Grizzly 2.0 code from trunk (version 2.1.2-SNAPSHOT) and test it?
>
> Thanks.
>
> WBR,
> Alexey.
>
>
>>> Again would be easier, if I'd be able to reproduce this on a simple test.
>> Yeah, I know that would be a lot simpler. Unfortunately I'm trying to keep up with another group that's doing something similar and I don't have a lot of time to extract the appropriate bits of this application and put them into separate test.
>>
>> I was kinda hoping someone might recognize the error and say "adjust this knob and you'll be all set". :)
>>
>> I'm using HttpServer.createSimpleServer() and registering a single HttpHandler implementation that accepts HTTP uploads and streams them to a FileChannel object using the NIOInputStream to read from the client, if it makes any difference.
>>
>> If I have time, I'll try and break this stuff out into a test case...
>>
>> jb
>>
>>
>>> Thanks.
>>>
>>> WBR,
>>> Alexey.
>>>
>>>> Thanks!
>>>>
>>>> Jon Brisbin
>>>>
>>>> http://jbrisbin.com
>>>> Twitter: @j_brisbin
>>>>
>>>>
>>
>> Thanks!
>>
>> Jon Brisbin
>>
>> http://jbrisbin.com
>> Twitter: @j_brisbin
>>
>>
>
Thanks!
Jon Brisbin
http://jbrisbin.com
Twitter: @j_brisbin