[dns-operations] lowercasing of questions from recursor to auths?
edward.lewis at icann.org
Fri Feb 5 18:47:26 UTC 2016
D'oh! Stupid me. Stupid, stupid...
Yes, I can re-create it now.
So, it is definitely broken. (I thought maybe it had been fixed before I
Although I still believe new code should not be broken to accommodate
brokenness, it could be argued that in the name of robustness an iterator
should try every trick in the book. I know that some implementations are
quite aggressive in getting answers. Perhaps trying lower-case only if
the answer isn't "good enough" but don't lower-case all queries.
On 2/5/16, 12:38, "David C Lawrence" <tale at akamai.com> wrote:
>Edward Lewis writes:
>> I can't reproduce this one. (dig "WIDGET.CRITEO.COM a" vs. "dig
>> widget.criteo.com a" as an example comparison.)
>Is that your whole command? It's pretty obvious when you dig directly
>at one of their nsN.criteo.com servers. Doesn't matter which part of
>the label has a capital letter, either. Compare:
>dig widget.criteo.com @ns1.criteo.com -> 1 A, 300s TTL
>dig widget.criTeo.com @ns1.criteo.com -> 7 As, 3600s TTL
>Someone here observed that they appear to be geo mapping, but the
>decision to do that is only being triggered based on a case-sensitive
>Independent of Peter's original question, it appears these people need
>to be informed that their authority server is broken and to fix it.
-------------- next part --------------
A non-text attachment was scrubbed...
Size: 4604 bytes
Desc: not available
More information about the dns-operations