ANNOUNCEMENT: Live Wireshark University & Allegro Packets online APAC Wireshark Training Session
July 17th, 2024 | 10:00am-11:55am SGT (UTC+8) | Online

Ethereal-dev: Re: [ethereal-dev] Help needed to add new file format decoding capabi lities to

Note: This archive is from the project's previous web site, ethereal.com. This list is no longer active.

From: Guy Harris <gharris@xxxxxxxxx>
Date: Thu, 17 Jun 2004 01:37:13 -0700
On Thu, Jun 17, 2004 at 10:26:26AM +0200, Yann Samama wrote:
> I want to add to Ethereal the ability to read new type of files that are
> generated by some commercial sniffers.
> As I read through the different files in the wiretap sub-directory, it
> raised several issues :
> 
> 1/ Is it possible to have a "read" function that can work on non-contiguous
> blocks of data?

"Non-contiguous" in what sense?

> 2/ Is it possible to have the data-link layer type to be defined on a per
> packet basis, rather than by a per file basis?

Yes.  See WTAP_ENCAP_PER_PACKET.

> 3/ What are the differences between the "read" and the "seek and read"
> functions?

The "read" routine is expected to scan forward to the next packet
record, and read it, advancing "wth->data_offset" as appropriate, and
setting "*data_offset" to a value that will let the "seek and read"
routine re-read that packet record.  It should use "wth->fh" when
reading.  It fills in a buffer, and possibly a pseudo-header, associated
with the "wtap" structure.

The "seek and read" routine takes an offset value as an argument, as
well as:

	a pointer into a buffer into which to read the packet data;

	a pointer to a "wtap_pseudo_header" union into which to put
	pseudo-header data, if any;

	the number of bytes of packet data to read;

and reads the packet that started at the specified offset (the offset
doesn't *have* to be a byte offset in the file, but it has to be
something sufficient to let the "seek and read" routine find the record;
except in some of the text file formats, it's an offset in the file).