[Guy Harris <guy@xxxxxxxxxx>]

> > It also kind of bugs me that the
> > smb dissector can append text to the info field after the subdissector
> > has handled the packet.
> 
> Yes, that's definitely wrong.

...but if we don't call a subdissector, it's OK.

I don't know whether, for SMBs that *don't* encapsulate other protocols,
we should move the "Error: XXX" stuff before any stuff added by the
dissector for that particular SMB; that's a side-effect of your patch.

If not, we could, instead, have "dissect_smb_command()" return a Boolean
value indicating whether a subdissector for another protocol was called,
and append the error indication only if it wasn't called.