Re: speed optimizations

["Didier" <dgautheron () magic fr>]

Hi,

On Thu, 31 Oct 2013 22:20:50 +0100, Anders Broman wrote
> > I have a very rough first patch from our version working with current svn.
> Sounds interesting, why not attach it to a bug report even in it's 
> rough state? I had some vauge idea to use the "ignore packet flag" 

cf https://bugs.wireshark.org/bugzilla/show_bug.cgi?id=9366
need a lot of cleaning but it works for some meaning of work...

> > call_dissector(next_dissector)
>
> Why do we need to know if a dissector has sub dissectors? What if we 
> kept a Gslist of all protocol handles in a
> frame, much like the "protocols string" but always populated on the 
> first pass one could then match (some) filters
> against that list before deciding if the frame should be dissected 
> or not.
IIRC the list is too big but it's a long time ago and it could be ok now. 
> > When compiling wireshark it's also extract which dissectors don't call
> > subdissectors or only call them via wireshark API
>
> If this is similar to a previous patch it requres changes to many 
> dissectors.
They was a lot of thing in this patch :)

Regards
Didier
___________________________________________________________________________
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