work properly with the multi interface when the user has provided the
fd_sets!

we need to point to user-provided fd_sets

from malloc()

before transfers, when doing multiple ones on the same handle.

even when the realloc() actually gets a new memory block

would reset the timeout for each incoming data, which would make veeery
slow responses be allowed to take even more time since the timeout would only
be reached if the time between two received data chunks was longer than the
set timeout value...

needed

used for "exported" globals.

rule: "Curl_" prefix for library-wide private symbols, "curl_" is for
exported symbols.

needs to be more portable and in the test server we can skip the extra
safety

any name resolved data in any curl handle struct. That way, we won't mind
if the cache entries are pruned for the next time we need them. We'll just
resolve them again instead.

This changes the Curl_resolv() proto. It modifies the SessionHandle struct
but perhaps most importantly, it'll make the internals somewhat dependent
on the DNS cache not being disabled as that will cripple operations somewhat.
Especially for persistant connections.

from the host names before comparing

now have --interface working properly