What is the difference to use CumulativeProtocolDecoder than to use
DemuxingProtocolCodec?
2007/5/12, angel figueroa <[EMAIL PROTECTED]>:
Hi,
the example that you mention is for the decoder. But what about the
Encoder.?
Thanks in advance for all the information that you are providing.
Regards,
Angel Figueroa
Maarten Bosteels-4 wrote:
>
> Hi Angel,
>
> You should not use the TextLineCodecFactory for reading binary data.
> From the example data, I assume your messages have header with a fixed
> length of 2 bytes
> which should be interpreted as the number of remaining bytes, right ?
>
> I have been working on a tutorial on writing your own
> ProtocolDecoder/Encoder, unfortunately I don't have the code right here.
> but I tried to reproduce it off the top of my head.
> I believe it does what you want, but I have not tested it :-)
>
> The decoder itself is stateless: it stores the conversational state in
the
> IoSession.
> This means you can use the same decoder instance for all your sessions.
>
> import org.apache.mina.filter.codec.CumulativeProtocolDecoder;
> import org.apache.mina.filter.codec.ProtocolDecoderOutput;
> import org.apache.mina.common.IoSession;
> import org.apache.mina.common.ByteBuffer;
>
> public class ByteArrayProtocolDecoder extends CumulativeProtocolDecoder
{
>
> private static class DecoderState {
> /** whether we have already read the fixed-length header */
> boolean headerRead = false;
> /** the length of the payload */
> int length;
> }
>
> /** the header has a fixed length of 2 bytes */
> private final static int HEADER_LENGTH = 2;
>
> private final static String DECODER_STATE_KEY =
> ByteArrayProtocolDecoder.class.getName() + ".DECODER_STATE";
>
> protected boolean doDecode(IoSession session, ByteBuffer in,
> ProtocolDecoderOutput out) throws Exception {
> DecoderState state = (DecoderState) session.getAttribute
> (DECODER_STATE_KEY);
> if (state == null) {
> state = new DecoderState();
> session.setAttribute(DECODER_STATE_KEY, state);
> }
> if (!state.headerRead) {
> // we have not yet read the header => check if have enough bytes
to
> read the fixed-length header
> if (in.remaining() >= HEADER_LENGTH) {
> state.length = in.getShort();
> state.headerRead = true;
> } else {
> // not enough bytes to decode a message, MINA will call us again
> when there is more data available
> return false;
> }
> }
> if (state.headerRead) {
> // we have already read the lengt header, check if all the data is
> available
> if (in.remaining() >= state.length) {
> // ok, message complete
> byte[] bytes = new byte[state.length];
> in.get(bytes);
>
> // this will cause IoHandler.messageReceived() to be called with
a
> byte[] as the message
> out.write(bytes);
>
> // remove the decoder state to be ready for the next message
> session.removeAttribute(DECODER_STATE_KEY);
> return true;
> }
> // not enough bytes available
> return false;
> }
> return false;
> }
> }
>
> Hope that helps,
> Maarten
>
> On 5/11/07, angel figueroa <[EMAIL PROTECTED]> wrote:
>>
>>
>> Which protocol Filter can be use for receiving Binary Data?
>> The first two byte is a binary number. I am using TextLineCodecFactory
>> but
>> i
>> thing it does not call the messageReceived.
>>
>> Any Help would be appreciated?
>>
>> Example
>> Associated Data - 70 bytes
>> 000: 0044 4953 4F30 3036 3030 3030 3430 3038
>> .DISO00600004008
>> 008: 3030 3832 3230 3030 3030 3030 3030 3030
>> 0082200000000000
>> 010: 3030 3034 3030 3030 3030 3030 3030 3030
>> 0004000000000000
>> 018: 3030 3035 3130 3230 3237 3433 3030 3030
>> 0005102027430000
>> 020: 3335 3030 3103 35001.
>>
>>
>> --
>> View this message in context:
>>
http://www.nabble.com/ProtocolCodecFilter-for-Binary-Data-tf3729533.html#a10439119
>> Sent from the mina dev mailing list archive at Nabble.com.
>>
>>
>
>
--
View this message in context:
http://www.nabble.com/ProtocolCodecFilter-for-Binary-Data-tf3729533.html#a10440123
Sent from the mina dev mailing list archive at Nabble.com.
