svn commit: r351423 - in head: . sbin/ping6 sbin/ping6/tests

[Bjoern A. Zeeb]

On 26 Aug 2019, at 16:31, John Baldwin wrote:

> On 8/26/19 1:59 AM, Bjoern A. Zeeb wrote:
>> In other notes (and I keep saying that), I can see a world when ping
>> doesn’t exist anymore as IPv4 doesn’t exist anymore (I partially
>> already live in that world).  The fact that people still do not 
>> prepare
>> themselves for this time is a bit strange to me as by the time 
>> FreeBSD
>> 14 is still in support this IPv6-only world might very well happen 
>> for a
>> majority of people.  And FreeBSD 14-CURRENT really is only a year 
>> away
>> now.   So breaking what’s been good for almost 20 years now for a 
>> few
>> more years doesn’t really seem to be worth to me.
>
> Eh, I think having 'ping' around on even IPv6 systems is sensible.
> ping is not inherently version-specific in name, only ping6 is.  
> Having
> ping not include ipv4 bits for WITHOUT_INET=yes is fine, but I think 
> not
> having ping as a command is just nonsense.  The fact that we have 
> ping6
> instead of ping -6 (compared to say, traceroute, ssh, etc. which all
> have unified commands) is just a user-interface bug we are stuck
> maintaining compatibility for, not a goal to shoot for.

Yes, I think we agreed with that and the wording we are using for the 
current (and future) FreeBSD situation just differed.

To rephrase:  I think it is a good idea (especially given the startup 
script use it) to have a command to send ICMP echo requests for a 
supported protocol family.

I also think it would be a good idea to preserve the legacy of ping6(8) 
which behaves exactly the same as the current ping6(8) even if the code 
is shared with ping(8) and installed as a hardlink or similar.

/bz