Anybody care to comment on the abysmal behaviour of communicator
with respect to DNS lookups? It looks like they have made some
seriously bad choices in how to use the res_*() library to resolve
requested URLs. It sometimes takes me 4-5 repeat attempts to get
the damn thing to resolve, pages which force a new client to pop
up generally fail first time. It also seems to cache its DNS lookup
state with some 'interesting' outcomes for temp failed DNS resolution...
Netscapes FAQ attempts to suggest its down to bad local DNS installation
but I would like to canvas opinion on this before assuming bind is at
fault. I've tested this with forwarding requests in resolve.conf and also
calling a local bind 8.1.1 daemon on the box. No difference.
I did try searching the archives. This issue doesn't seem to have come up in
recent months.
cheers
-George
Received on Wed Dec 24 1997 - 00:37:32 NZDT