[Guy Harris <guy@xxxxxxxxxx>]

> I wonder, is there ever a need to create a new dfilter and apply it
> to an already-created proto_tree, causing this scheme not to work?
> It probably doesn't matter, since re-creating the proto_tree is not
> too time-consuming.

Given that we don't store protocol trees, that's likely to happen only
if you create a new dfilter and apply it to a currently-displayed
protocol tree.

But, if it ever happens, it could be done by making a pass over the
existing protocol tree and, for each node:

	checking if it's in the list of interesting fields for the
	dfilter;

	if so, adding the item to the list attached to the dfilter's
	entry for that hf_ value.