wrong ans. name

[Andrew Brown]

>...
>which, i am almost certain, stems from the fact that m has multiple
>cnames.  i get, for example:
>
>;; ANSWER SECTION:
>m.doubleclick.net.      5M IN CNAME     exnjld1avip.doubleclick.net.
>m.doubleclick.net.      5M IN CNAME     ancald2avip.doubleclick.net.
>m.doubleclick.net.      5M IN CNAME     ancald1avip.doubleclick.net.
>m.doubleclick.net.      5M IN CNAME     anvald2avip.doubleclick.net.
>m.doubleclick.net.      5M IN CNAME     anvald1avip.doubleclick.net.
>m.doubleclick.net.      5M IN CNAME     uucald2avip.doubleclick.net.
>m.doubleclick.net.      5M IN CNAME     uucald1avip.doubleclick.net.
>m.doubleclick.net.      5M IN CNAME     uuvald2avip.doubleclick.net.
>m.doubleclick.net.      5M IN CNAME     uuvald1avip.doubleclick.net.
>m.doubleclick.net.      5M IN CNAME     exnjld4avip.doubleclick.net.
>m.doubleclick.net.      5M IN CNAME     exnjld1bvip.doubleclick.net.
>exnjld1avip.doubleclick.net.  1H IN A  209.67.38.35

it seems to me that bind, in the case of a multiple-cname answer,
should be inserting the A record that corresponds to the data of the
last CNAME in the list (in order to satisfy gethostby*() at the very
least).  it seems to be to be serving up the CNAMEs in proper
round-robin fashion, but only inserting the first A record it finds
for the given CNAMEs.  at least, i'm always getting the same A record
in the answer.

otoh, inserting all of them would probably result in a completely
different failure mode (eg, packet truncated most of the time).

-- 
|-----< "CODE WARRIOR" >-----|
codewarrior at daemon.org             * "ah!  i see you have the internet
twofsonet at graffiti.com (Andrew Brown)                that goes *ping*!"
andrew at crossbar.com       * "information is power -- share the wealth."