directory-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Trustin Lee <trus...@gmail.com>
Subject Re: [mina] Bug in VmPipeFilter
Date Thu, 24 Nov 2005 09:43:07 GMT
Hi Niklas,

2005/11/24, Niklas Therning <niklas@trillian.se>:
>
> While writing a test for traffic control on VmPipe transports I may have
> stumbled across a bug in VmPipeFilter. Basically I have an echo server
> and a client which repeatedly sends a single byte to the server. When I
> run this test I get an exception:
>
> Exception in thread "IoThreadPool-1" java.lang.IllegalStateException:
> Already released buffer.  You released the buffer too many times.
>     at
> org.apache.mina.common.ByteBuffer$DefaultByteBuffer.release(
> ByteBuffer.java:593)
>     at
> org.apache.mina.util.ByteBufferUtil.releaseIfPossible(ByteBufferUtil.java
> :43)
>     at
> org.apache.mina.filter.ThreadPoolFilter.processEvent(ThreadPoolFilter.java
> :616)
>     at
> org.apache.mina.filter.ThreadPoolFilter$Worker.processEvents(
> ThreadPoolFilter.java:372)
>     at
> org.apache.mina.filter.ThreadPoolFilter$Worker.run(ThreadPoolFilter.java
> :326)
>
> I think this is because the ThreadPoolFilter releases buffers right
> after it has called messageReceived/messageSent. Since VmPipeFilter
> sends the same message object (i.e. the buffer) to these two methods the
> ThreadPoolFilter will already have released the buffer when VmPipeFilter
> calls nextFilter.messageReceived.
>
> I have attached a simple application which demonstrates the bug. I've
> also attached a fix to VmPipeFilter which also adds correct byte counts
> if messages are buffers. Let me know if the fix looks ok and I will
> check it in.


We could simply acquire the buffer once more?  WDYT?

You've got an access to the repository, so you can directly check it in. :)

Cheers,
Trustin
--
what we call human nature is actually human habit
--
http://gleamynode.net/

Mime
View raw message