On Tue, Oct 28, 2003 at 06:18:03AM -0800, Dinesh G Dutt wrote: > Guy Harris writes: > > if you have the frame length, either the ability to set the reported > > length of a tvbuff to a *larger* value than it currently has should be > > added, and you should use that, or you should construct a new > > "top-level" tvbuff with the specified reported length; > What would happen if reassembly is attempted for such frames ?
It should fail, as reassembly without all the frame data won't work (it'd leave holes in the reassembled packet), so code that does reassembly should check for that (it's done in at least some places in reassemble.c, and done by at least some callers of that code as well).
