I hate the new mail system

[Tor Arntsen]

On 8 March 2017 at 19:46, Pete Turnbull <pete at dunnington.plus.com> wrote:

> Notice that - unlike normal whois servers - this one apparently requires
> some other stuff, possibly including the text "whois", as part of the query.
> That may explain why a normal whois client gets an error, because the
> standard way to make a query is simply to send the string to query (eg,
> "uni-stuttgart.de" or "dunnington.cx"):
>
> $ whois -h whois.denic.de uni-stuttgart.de
> % Error: 55000000007 Request not clearly specified
>
> OK, so back to telnet, and try it with the syntax DENIC claims to want:
>
> $ telnet whois.denic.de 43
> Trying 81.91.170.6...
> Connected to whois.denic.de.
> Escape character is '^]'.
> -T dn uni-stuttgart.de
> [lots of output]
> $
>
> Aha!  That works.  But I can't replicate it with most whois clients.
>
> However, it /does/ work with the jwhois client, which some linux systems
> have, and which seems to have some special method to deal with DENIC. It
> doesn't work with the RIPE client - despite DENIC recommending that - unless
> you construct a rather odd-looking query by adding at least "--T dn":
[..]
> Pete Turnbull

I did an strace and I can confirm that the Linux 'whois' client that I
used from those various sites sends '-T dn' (or actually -T dn,ace)

   write(3, "-T dn,ace uni-stuttgart.de\r\n", 28) = 28

I can't see where this whois originates from, it has version number
'5.2.<something>'. Its man page refers to RFC 3912, but RFC 3912 says
nothing about -T.  RFC 3912's single example wouldn't have worked in
this case. So I wonder what replaced RFC 3912, and why there's a
mismatch between documentation and functionality.