On Oct 28, 2012, at 4:08 PM, Jakub Zawadzki <darkjames...@darkjames.pl> wrote:

> if user description:
>  - not set than: format("%s (%s)", friendly_name, interface_description)
>  - set: only user descr.
> 
> Anyway such code should imho go to winpcap.

Which code?  The code to get the friendly name, or that code and the code to 
get the user description, or that plus the code that chooses the format in 
which to display that information?

Currently, the pcap API has no notion of separate "friendly name" and "vendor 
description" for an adapter; arguably, it should.  pcap_findalldevs() would 
need a data structure change to support that; personally, I'd prefer adding a 
new API that supplies pcap-NG Interface Description Block-style information, as

        1) that makes it easier to map to an IDB

and

        2) that makes it extensible.

Libpcap/WinPcap also has no configuration files it reads or writes, so it has 
no place to store user descriptions.  I'm less sure it should have that.

I'd prefer leaving "how to display it" up to the application (tcpdump, *shark, 
etc.).

And, yes, ultimately libpcap/WinPcap should supply a friendly name if available 
and a vendor description if available; however, libpcap major releases don't 
come out very often (once a year or so), and WinPcap is even slower (the latest 
release is over two years old, and is based on libpcap 1.1.0, not on the latest 
1.3.0 release, which means it can't even read pcap-ng files), so, just as 
Wireshark does its own enumeration of network interfaces if libpcap is so old 
that it lacks pcap_findalldevs(), it can do its own friendly-name fetching if 
libpcap/WinPcap doesn't yet have a call that supplies the friendly name.
___________________________________________________________________________
Sent via:    Wireshark-dev mailing list <wireshark-dev@wireshark.org>
Archives:    http://www.wireshark.org/lists/wireshark-dev
Unsubscribe: https://wireshark.org/mailman/options/wireshark-dev
             mailto:wireshark-dev-requ...@wireshark.org?subject=unsubscribe

Reply via email to