Re: [Wireshark-commits] rev 53090: /trunk/ /trunk/asn1/acse/: acse.cnf packet-acse-template.c /trunk/asn1/cmip/: cmip.cnf packet-cmip-template.c /trunk/asn1/disp/: packet-disp-template.c /trunk/epan/dissectors/: ...

[Guy Harris <guy () alum mit edu>]

On Jan 10, 2014, at 10:59 AM, Stig Bjørlykke <stig () bjorlykke org> wrote:

> Updating the RTSE dissector to a new-style was done by returning
> tvb_length(tvb), which in this case is always 0.  Returning 0 from a
> new-style dissector means this package was not for us, which is wrong
> in this case.

That's why I never converted all dissectors to new-style dissectors - having the return value both be an "is this a 
packet for my protocol?" indicator *and* a "length dissected in the tvbuff handed to me" indicator doesn't always work.

Something better is needed.  Perhaps handing a ptvcursor to the dissector, using the ptvcursor to handle the "how much 
was dissected?" part, and having the dissector return a gboolean indicating whether the packet was accepted or not 
would be better.  (It would also let heuristic dissectors use the same signature.)
___________________________________________________________________________
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-request () wireshark org?subject=unsubscribe