Wireshark-dev: Re: [Wireshark-dev] new packet-usb-audio dissector?

From: Jaap Keuter <jaap.keuter@xxxxxxxxx>
Date: Sun, 09 Dec 2007 12:44:09 +0100

Hi,

There is little difference in having your dissector in either a plugin or build-in. On the Win32 platform you have limited function access through what's listed in libwireshark.def, but that is mostly complete.

The big plus is that your rebuild cycle for a plugin is much shorter than for a build in one. So starting of with a plugin makes initial development simpler, while deployment of the finished code may well be done as build-in dissector.

The agentx plugin is the current example, since the h323 dissector was recently moved to be build-in.

Thanx,
Jaap

Németh Márton wrote:

Hi,

I would like to implement the USB Audio Subclass
( http://www.usb.org/developers/devclass_docs/audio10.pdf ) dissector for
Wireshark, so I started to read the documentation. In "Wireshark Developer's
Guide" ( http://www.wireshark.org/docs/wsdg_html/#ChDissectSetup ) it was
recommended to start a new dissector to be a "plugin". Is this the recommended
format or should I start writing the new dissector into
wireshark/epan/packet-usb-audio.c? What are the drawbacks and gains writing a
plugin vs. writing a built-in dissector?

I also found the following sentece:

"You can find a good example for these files in the h223 plugin directory."

However, the h223 packet dissector is located at wireshark/epan/dissectors/packet-h223.c,
so I assume it is not (or no longer) a plugin.

In the wireshark/doc/README.plugins (from current SVN), I found that I should take
the wireshark/plugins/agentx. So where should I start?

	Márton Németh