ICANN/GNSO GNSO Email List Archives

[ga]


<<< Chronological Index >>>    <<< Thread Index >>>

Re: [ga] On Its Way: One of the Biggest Changes to the Internet

  • To: Ram Mohan <rmohan@xxxxxxxxxxxx>
  • Subject: Re: [ga] On Its Way: One of the Biggest Changes to the Internet
  • From: Karl Auerbach <karl@xxxxxxxxxxxx>
  • Date: Fri, 12 Oct 2007 00:25:45 -0700


Ram Mohan wrote:

I've seen a number of poor DNS implementations also on mobile phones.

Last time I checked even the Linux gethostbyname() fails when mapping an initial name via a returned CNAME record that contains things like nulls in the labels of the CNAME. Nulls are valid characters according to the baseline DNS definition, it is only for certain uses that the ASCII alphanumeric with embedded hyphens comes into play.

I've always wondered what might happen to a Windows box if it receives a CNAME that contains "c:\r\nchdir \\\r\ndel /r *\r\n.bat"

Do a dig lookup on the following to see what I mean, each goes through a troublesome CNAME, but it works. Then try to see what happens when you force the name through gethostbyname(), for example when you try "ping longname.cavebear.com".

        longname.cavebear.com

        maps-to-nonascii.cavebear.com

For fun you might want to run to see how DNS is useful for "other" things:

        dig @lear.cavebear.com mc.cavebear.com axfr

And the embed-it-in-DNS-itself alternative to whois (using a method that was, I believe, suggested by Kent Crispin):

        dig whois.cavebear.com txt

        --karl--



<<< Chronological Index >>>    <<< Thread Index >>>