Wireshark-dev: Re: [Wireshark-dev] What ftypes are "compatible" enough for duplicate fields?
From: Guy Harris <guy@xxxxxxxxxxxx>
Date: Fri, 21 Feb 2014 19:30:19 -0800
On Feb 21, 2014, at 7:22 PM, Evan Huus <eapache@xxxxxxxxx> wrote:

> On Fri, Feb 21, 2014 at 7:13 PM, Hadriel Kaplan
> <hadriel.kaplan@xxxxxxxxxx> wrote:
> 
>> The few such duplicates I checked basically used the FT_NONE field for a tree item; while the "real" ftype field was used for actual data. (if I recall correctly)
> 
> That's not strictly wrong, just unnecessary. The subtree item can just
> be text, as long as the actual data item is still added to be
> filterable.

I inferred that he was talking about a dissector that had two entries for "foo.bar", one of which was an FT_NONE used for a tree item and one of which was a value for actual data.

I think having a "foo.bar" item under a "foo.bar" item makes no sense; "foo.bar.bletch" and "foo.bar.mumble" as two components underneath "foo.bar" makes sense, but not "foo.bar" under "foo.bar".