On Tue, Apr 27, 2010 at 18:59:16 +0200, Kurt Roeckx wrote: > I didn't look at the source, but if it's already using > getaddrinfo() and going over all the addresses it returned, > I don't see why it should be broken with either value of > the option. > > So I can only assume that it doesn't go over all the > addresses returned by getaddrinfo(), and that that is > the real issue. > Whatever "the real issue" means. It assumes that bind() on in6addr_any allows receiving ipv4 packets, as the spec says it does. I know it could use multiple sockets, but that would be a less trivial patch, and wouldn't make any actual difference afaict. > I also think that applications should set it to 1, so I would > disagree with that patch. And it would still have a problem > on an OS that doesn't allow you to set that to 0. > I couldn't care less about openbsd. Cheers, Julien
Attachment:
signature.asc
Description: Digital signature