I am sorry. I would like to fix typo. :(
>But, I *couldn't* commit test codes for maven because user-generated
>thrift codes were needed for testing(any idea?).
Regards,
Bongjae Chang
On 11/28/11 11:17 PM, "Bongjae Chang" <bongjae.chang_at_gmail.com> wrote:
>Hi Alexey,
>
>I fixed bugs as your advice and committed them to extras/thrift.
>
>I edited 3 points from original sources which I attached.
>
>1. Processing more than 1 thrift packet in ThriftFrameFilter
>2. Adding input buffer dispose() in ThriftServerFilter
>3. Removing redundant input buffer copy in ThriftClientFilter
>
>Commit revision is "f1d895bf".
>
>But, I could commit test codes for maven because user-generated thrift
>codes were needed for testing(any idea?).
>
>So you can test it in only original attached files.
>
>If there are any issues or more suggestions about this, please let me
>know.
>
>Thanks!
>
>Regards,
>Bongjae Chang
>
>
>On 11/28/11 7:29 PM, "Oleksiy Stashok" <oleksiy.stashok_at_oracle.com> wrote:
>
>>Hi Bongjae,
>>
>>> I implemented thrift server and client modules which based on Grizzly
>>>v2.
>>thank you very much for your contribution!
>>
>>>
>>> Thrift site is http://thrift.apache.org .
>>>
>>> And Netty is already supporting thrift server and client modules.
>>>
>>> Netty tool site is https://github.com/cgbystrom/netty-tools
>>>
>>> So I tried to support them using Grizzly for performance improvement.
>>>
>>> I attached sources and test codes. Grizzly-thrift codes are located at
>>> modules/thrift/src/main and please read the README file. :)
>>> Could you please review them?
>>Can I suggest you move this code to Grizzly git repository under extras
>>directory for now, so it will simplify development for you and reviewing
>>for folks who's interested.
>>The code looks fine in general.
>>The most important thing, which is probably missed (if I'm not missing
>>anything), in the ThriftFrameFilter there is no logic, which handles
>>situation when input Buffer contains more than 1 thrift packet (you can
>>see the GIOP sample) to see how we propose to handle this (splitting
>>buffer etc).
>>
>>Hope this will help.
>>
>>Thanks.
>>
>>WBR,
>>Alexey.