Wireshark-bugs: [Wireshark-bugs] [Bug 11322] Regression: TCP reassembly can fail DISSECTOR_ASSER

Date: Tue, 30 Jun 2015 19:56:04 +0000

Comment # 6 on bug 11322 from
(In reply to Hadriel Kaplan from comment #5)
> (In reply to Pascal Quantin from comment #3)
> > As explained in the packet-tcp.c comments, this sanity check was added on
> > purpose so as to detect faulty dissectors rejecting packets while still
> > changing the desegment related variables while they should not. Could you
> > double check what is done by your Lua script? 
> 
> Unfortunately, it's not "my script" - it's just one I happen to have found
> that exercises the TCP-reassembly stuff. (found on
> https://github.com/diegonc/packet-bnetp)
> 
> I only mention that it's not mine because there could be any number of such
> Lua scripts out there that do something similar - since it worked in
> previous releases.

That's why this kind of change is done in major releases and are not backported
in earlier branches: to avoid breaking compatibility. Having custom scripts /
plugins working between major releases is something that we have never
guaranteed (some Lua APIs have changed also if I'm not mistaken).

> 
> 
> Right I figured, but I don't see why it's an error.  I'll comment on that
> separately.

As explained in comment #4, this kind of checks was already done for heuristic
sub dissectors. I just ensured that it is done uniformly whatever the sub
dissector type (by port, conversation or heuristic).


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