> First - I do think the PREFER_LARGEST_SCOPE is needed for a large collection > of applications. (With all the noise about site-local in this group, I > wonder if a REQUIRE_GLOBAL_SCOPE wouldn't be a good idea...)
Or make PREFER_LARGEST_SCOPE the default? But that isn't the subject of this draft ... > However, it does not make the hard things possible. > > As an application programmer, I'm not sure this will really give me enough > control when/if I need it. What I *think* I would like to see, in addition to > this sockopt, is an interface similar to getaddrinfo - but for address pairs, > not just for destination addresses. Something that lets me specify both the > source and the destination characteristics that I care about - and have it > return a linked list of all the src/dst address combinations that are > possible given those constraints. Sorted in the same order that the system > would have used to choose. This would of course include flags such as > described in this draft. (I'm not convinced this much control is really > needed yet...) I don't quite understand this. You can only express properties about the local addresses (you can't tell from the DNS whether the address you looked up has any particular property). Having said that, being able to do all of address selection in getaddrinfo and have it return a list of src/dst pairs would make certain things simpler. But I think that is orthogonal to adding the preferences and/or requirements on local address types. Erik -------------------------------------------------------------------- IETF IPng Working Group Mailing List IPng Home Page: http://playground.sun.com/ipng FTP archive: ftp://playground.sun.com/pub/ipng Direct all administrative requests to [EMAIL PROTECTED] --------------------------------------------------------------------
