On Sun, 24 Jun 2001, Francis Dupont wrote:
>    switches are necessary in many cases (especially when
>    you want the same code to work on a traditional ipv4 system that uses
>    gethostby{name,addr} and on a new ipv6-compliant system which uses
>    get{addr,name}info).
>
>    > => I have ported many applications to IPv6 on BSD systems without
>    > the problems you seemed to have encountered...
>
>    it depends from the application and the policy of the programmer.
>    i prefer not to modify existing ipv4 code (for compatibility with older
>    systems), so i generally use A LOT of switches. i also try to be careful
>    when handling common ipv6/ipv4 cases, and this make my work a little
>    harder.
>
> => I have the opposite policy so results are different. My target is
> integrated systems where IPv6 is *not* optional, no more than IP is!

That is a longer term goal.  Currently that only works in closed
environments.

>From my viewpoint, the applications wrt. ipv6 support can be split into a
few categories:

 1) no ipv6 support
 2) ipv6 support togglable at compile time; if you use the app on
ipv4-only system, it won't even run (or ipv4 features don't work properly)
[for example xinetd]
 3) ipv6 support togglable at compile time; if you use the app on
ipv4-only system, ipv4 features work
 4) ipv6 support detected/configured at runtime

I don't consider applications to be integrated properly unless 4) applies.
I get the impression you are doing mostly 2).  As ipv4 isn't going away
any time soon, except in some closed environments like 3G mobile phones,
you have to be able to deal with all kinds of situations.

As said, I don't think this works yet in real life (consider: OS
distributions), and is counter-productive for the spread of IPv6 as
distributions cannot ship ipv6 ready apps just in case unless they're 3)
or 4) (because that'd hurt the operation of production software).

-- 
Pekka Savola                 "Tell me of difficulties surmounted,
Netcore Oy                   not those you stumble over and fall"
Systems. Networks. Security.  -- Robert Jordan: A Crown of Swords

--------------------------------------------------------------------
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]
--------------------------------------------------------------------

Reply via email to