Its a method that allows you to see who the real owner of the memory your object referring to is. Packet state, header and buffer's get peered to each other it is very possible to loose track of who the owner is.

This debug method JMemory.toDebugString() prints out the following information:

JMemory@b3f004a: size=1506, owner=Pcap.class
JMemory@b3f004a: size=1506, owner=Pcap.class
JMemory@b3f004a: size=886, owner=Pcap.class
JMemory@af7afb0: size=1506, owner=nio.JMemoryPool$Block.class
JMemory@b3f004a: size=86, owner=Pcap.class
JMemory@b3f004a: size=886, owner=Pcap.class
JMemory@b3f004a: size=86, owner=Pcap.class
JMemory@af7e788: size=1506, owner=nio.JMemoryPool$Block.class
JMemory@af7ed7a: size=886, owner=nio.JMemoryPool$Block.class
JMemory@af7f100: size=86, owner=nio.JMemoryPool$Block.class
JMemory@af7f166: size=1506, owner=nio.JMemoryPool$Block.class
JMemory@af7f758: size=1506, owner=nio.JMemoryPool$Block.class
JMemory@af7ffc0: size=886, owner=nio.JMemoryPool$Block.class
JMemory@af80346: size=86, owner=nio.JMemoryPool$Block.class
[truncated....]

This output is from iterating over a capture file and all its packets. The JMemory.toDebugString() was called in 2 places. First on the original JBuffer that supplied with JBufferHander.nextPacket method and the second time after a new PcapPacket was created with the libpcap buffer copied into it.

You can see 2 owners. Pcap.class and JMemoryPool$Block.class. Pcap class represents memory owned and allocated by libpcap library itself. While the Block.class is the owner using jNetPcap's default allocation scheme while allocates memory in chunks and hands it out for packet copies.

I've done the Http header and its one of those Mime based headers that can have many dozen's of different types of fields. It seems unreasonable to have the header author define everyone of those fields as a separate accessor method. We need a better header container for those type of headers (http, SIP, Mail message, etc..)

The existing Http definition (found under tests directory), already keeps track of all those headers in a Map<String, String>. There is an accessor method to retrieve the map and then using normal map methods you can check for each field. I think that is the right approach since there are so many possible fields, and most of them are almost always not there. The JHeader.decodeHeader method properly sets the offset and length properties for each Http field, thus it is simply a matter of a normal numerical sort to get the present fields to display properly in a formatter.

A single AbstractMimeHeader baseclass can take care of all of those protocols at once, including http.

One currently unresolved issue with Http is the multi-part contents. I've been thinking about this quiet a bit, and it doesn't make any sense to implement multiparts as hardcoded embeded sub-headers, although that is how they appear. What I'm leaning towards is a new type of sub-header that can act as as a proxy (a surrogate) and place actual standalone headers as sub-headers. This will allow for example Http header to be completely standalone as a header after http and as an embeded sub-header for Http header's with multiparts.

New text handling in JBuffer seems to be working well for Http so that part is done and working.

Just some of the things I'm working on and thinking about.

hello everyone.
i have a problem in sending the packet that i have overwritten.

the code is below

pcap.dispatch(1000, new PcapHandler() {	
  public void nextPacket(String user, long seconds, int useconds,
      int caplen, int len, ByteBuffer buffer) {

    buffer.put(getByteMAC(MAC_NAS));//newdes
    buffer.put(getByteMAC(MAC_LAPTOP_WIRE));//new source

    for(int  jj=0 ;jj<5;jj++){
      System.out.println("jj is "+jj);
      pcap.sendPacket(buffer.array());
    }
  }
}

the console output is like this
jj is 0 (there is no loop in there)
but when i didnt send the packet, the loop worked well
jj is 0, jj is 1 ....

note: i have checked that i could send using pcap.sendPacket(buffer.array()), with arp packet that i've created and overwritten it.

anyone know why i couldnt send the packet?

Best Regards

Timmy

Hi *,

I wrote a custom header (ieee 802.11 radiotap) which worked fine until rc3.

i did change all the annotations. its compiling, even decoding all those optional subheaders (writing correct values in optionsBitmap, optionsLengths and optionsOffsets).

but with rc3 i'm getting a nullpointer-exception

Opening: /home/[... snip ...]/captures/radiotap_IEEE80211_LLC_IP4_TCP
Protocol for this DLT [(127) which is (IEEE802_11_RADIO)] --> [org.jnetpcap.packet.header.IEEE802dot11_RADIOTAP]
Reading 1 packet...
Packet caplen=104 wirelen=104
id=1 name=TSFT is_present=0 present_bitmap=0x482E optionsBitmap=0x0 pad=0 offset of next field=8 length=0  
[... snip ...]
id=14 name=DB_ANTENNA_NOISE is_present=0 present_bitmap=0x482E optionsBitmap=0x105C pad=0 offset of next field=16 length=0  

Exception in thread "main" java.lang.NullPointerException
	at org.jnetpcap.packet.structure.JField.getLength(JField.java:238)
	at org.jnetpcap.packet.format.JFormatter.stylizeBitField(JFormatter.java:508)
	at org.jnetpcap.packet.format.JFormatter.stylizeSingleLine(JFormatter.java:609)
	at org.jnetpcap.packet.format.TextFormatter.fieldBefore(TextFormatter.java:94)
	at org.jnetpcap.packet.format.JFormatter.format(JFormatter.java:274)
	at org.jnetpcap.packet.format.JFormatter.format(JFormatter.java:278)
	at org.jnetpcap.packet.format.JFormatter.format(JFormatter.java:304)
	at org.jnetpcap.packet.format.JFormatter.format(JFormatter.java:254)
	at org.jnetpcap.packet.format.JFormatter.format(JFormatter.java:347)
	at org.jnetpcap.packet.format.JFormatter.format(JFormatter.java:316)
	at org.jnetpcap.packet.JPacket.toString(JPacket.java:694)
	at de.fuberlin.mi.cst.WiFiClipboard.WiFiClipboard.nextPacket(WiFiClipboard.java:39)
	at org.jnetpcap.Pcap.loop(Native Method)
	at org.jnetpcap.Pcap.loop(Pcap.java:1947)
	at de.fuberlin.mi.cst.WiFiClipboard.WiFiClipboard.testRadiotap(WiFiClipboard.java:19)