Re: Info on Bug #1154

On Sun, 2005-01-30 at 20:56 +0100, Henrik Nordstrom wrote:
> On Sun, 30 Jan 2005, Serassio Guido wrote:
>
> > Looking to connStateCreate() code, I need only a detail about the 3 state
> > logic of disable-pmtu-discovery:
> >
> > always ==> disable PMTU for any call to connStateCreate()
> > off ==> never disable PMTU
> > intercepted ==> disable PMTU only when port->transparent is defined
> >
> > It's right ?
>
> Almost.
>
> intercepted ==> disable PMTU only when the connection is set in
> transparent mode. See the code relating to the clientNatLookup call.
>
> For clarity I think transparent should be used instead of intercepted.

transparent has very specific meaning in rfc2616, I really don't think
we should refer to transparent for anything other than that.

We already have idiomatic use of transparent as:
* NTLM
* interception

Using it in code or documentation, other than to resolve the ambiguity,
is bad IMO.

Rob

-- 
GPG key available at: <http://www.robertcollins.net/keys.txt>.