On Sat, 15 Oct 2011, Yang Tse wrote:

> I don't remember if the reason for not doing so was that some stacks failed
> later on when addressing '::1' or something. Maybe someone has more fresh
> info about this.

I think I let it remain like that as I've kind of assumed that A) a sensible
libc implementation would never cause an actual DNS resolve of such an address
anyway since we don't ask for a canonical name and B) there could be other
magic involved that would allow a function to be clever and return info we
don't have. Not that I can think of any particular right now...

Andy, are you saying that your test target makes an actual DNS resolve of an
IP-only string? And if so, what exactly does it try to resolve and what is the
result?

-- 
  / daniel.haxx.se
-------------------------------------------------------------------
List admin: http://cool.haxx.se/list/listinfo/curl-library
Etiquette:  http://curl.haxx.se/mail/etiquette.html