[
Lists Home |
Date Index |
Thread Index
]
Miles Sabin wrote:
> Bullard, Claude L (Len) wrote,
>
>>All the "always resolves" means is that if the http: endowed URI is
>>handed to a resolver, it will resolve. What gets returned may be
>>caca, but it will resolve.
>
>
> What? Even http://this.host.does.not.exist/ ?
>
> You won't even get a 404 from that. Do you count TCP-level failures as
> "resolving"?
The thing is, unless you have knowledge embedded in the client of the
resolver there is no way to tell if a TCP or DNS level failure is
transient or final, without multiple retries. Even then, it may not be
correct.
For instance, http://www.foo.com/ may not resolve because my DNS server
is unreachable or because the DNS server hasn't been configured
properly. Now, that's fine if I know that "www.foo.com" or "foo.com", or
".com" `should' exist. However, if I don't have such a priori
information, I can't tell the difference between http://www.foo.com/ and
http://www.foo.blah/ when they produce the same error.
--
Mike
|
