On Tue, Feb 21, 2006 at 10:20:53PM +0100, Ignatios Souvatzis wrote:
>On Tue, Feb 21, 2006 at 01:11:52PM -0800, jonathan@dsg.stanford.edu wrote:
>> 
>> In message <20060221210414.GA27831@beverly.kleinbus.org>,
>
>> I think the real ``bug'' is forcing IPv6 down people's throats
>> whethher they want it or no, which leads to nobody bothering to
>> test IPv6-less configs, which leads to bugs (or more accurately,
>> buggy error messgaes) like the one George encountered.
>
>I agree here; but note that we surely also have a lot of code that
>breaks or at least gives similar error messages if we disable IPv4.


I never tried to run without ipv4, but I think that is a valid
concern.

The reason I run NetBSD (among other reasons) is the ability
to setup a clean deployment, minimalistic and absent of gory
hacks. To that end, if someone wants a keyboard/display/disk only
install, it should be supported because that alone should work
before other drivers and protocols are added. Of course I mean
as a target goal, philosophically, not a required endpoint; and
I expect developer motivation to arise from their own issues.

There probably would be a lot of unnecessary or vague warnings,
brokenness and similar for disabling IPv4; but that doesn't mean
they shouldn't be corrected, only that for IPv4 dependencies are
much less likely to become noticed because that's such a standard
protocol.

For me, when I looked at my kernel conf (I knew an option there
caused the warning), IPv6 would have been one of my last guesses.
If you google for "socket: Protocol not supported" you'll find it
comes up a lot on different OSs, for different software, and is
often unanswered. -- in this case, a more useful warning from ssh
could have explained my issue, in the time it took to read it.

(adding to my list of patches to write--now saved to disk *sigh*)

// George


-- 
George Georgalis, systems architect, administrator <IXOYE><
http://galis.org/ cell:646-331-2027 mailto:george@galis.org