Wireshark-bugs: [Wireshark-bugs] [Bug 8787] 9P dissector - compute fids on first visit

Date: Thu, 01 Aug 2013 15:01:22 +0000

Comment # 22 on bug 8787 from
(In reply to comment #20)
> (In reply to comment #18)
> > Yes, tcp_dissect_pdus() is probably what you want to use to handle, well,
> > TCP being TCP.
> 
> Do you want it in this patch or can it wait for another ticket?

New functionality can definintely wait...

> > Another thing which has changed since this started: the ep_ and se_
> > allocators have been deprecated.  Ideally those allocations would be
> > replaced with wmem (see doc/README.wmem).
> 
> wmem looks nice! It doesn't seem such a hard change, but I'd rather have it
> in a different patch as this one already is quite big, and I'm more of the
> "incremental" type :P (It'd go with the tcp_dissect_pdus one that should be
> quite small)

It would be better to do the wmem-ification now as it should be pretty small,
but I'm OK with doing it later.

> Also, I suppose this means I probably should replace g_malloc with wmem with
> the epan scope? does tvb allocated data gets allocated with this as well,
> and should use wmem_free instead of g_free? (e.g. tvb_get_string returns a
> g_malloc'd string at the moment, I suppose there would be two functions
> laying around while the transition is being done?)

g_malloc->wmem is, I think, less important than ep_/se_->wmem.  And until APIs
like tvb_get_string() return wmem there's not much for you to do anyway...


You are receiving this mail because:
  • You are watching all bug changes.