Please test fping 3.16-rc2 (fping6 is no more!)

[David Schweikert]

Hi,

Thanks Axel for testing 3.16-rc1!

I have now published fping 3.16-rc2 with a very big change: the fping
and fping6 binaries are now united into a single binary, as modern
IPv6-enabled tools should be. There are now additional options '-4' and
'-6' to force IPv4-only or IPv6-only mode.

Here are the release notes:

* INCOMPATIBILITY WARNING:
fping and fping6 are now unified into one binary. This means that for
example doing 'fping www.google.com' is going to ping the IPv6 IP of
www.google.com on IPv6-enabled hosts. If you need exact compatibility with
old versions, you can:
- compile fping with --disable-ipv6 (or use a wrapper, and call 'fping -4')
- compile fping with --enable-ipv6 and rename it to fping6 (same as 'fping -6')

* (feature) Unified 'fping' and 'fping6' into one binary (#80)
* (feature) New option '-4' to force IPv4
* (feature) New option '-6' to force IPv6
* (feature) Support kernel-timestamping of received packets (#46)
* (feature) Simplify restrictions: only -i >= 1 and -p >= 10 are enforced now
* (feature) --enable-ipv6 is now default (you can use --disable-ipv6 to disable IPv6 support)
* (bugfix) Fix option -m to return all IPs of a hostname
* (bugfix) Fix option -H (ttl) for IPv6
* (bugfix) Fix option -M (don't fragment) for IPv6
* (bugfix) Fix option -O (ToS) for IPv6
* (bugfix) Fix compatibility issue with AIX (#69, @blentzgh)
* (bugfix) Fix option -q not suppressing some ICMP error messages (#83)
* (bugfix) Fix option -M expecting an argument, when it shouldn't
* (bugfix) Fix minor issues found by Coverity Scan

Please test and let me know if you notice anything.

Cheers
David

[Axel Beckert]

Hi,

On Tue, Jan 31, 2017 at 02:04:39PM +0100, David Schweikert wrote:
> Thanks Axel for testing 3.16-rc1!

I'll probably upload 3.16-rc2 to Debian Experimental in the next few
days, too.

> I have now published fping 3.16-rc2 with a very big change: the fping
> and fping6 binaries are now united into a single binary, as modern
> IPv6-enabled tools should be. There are now additional options '-4' and
> '-6' to force IPv4-only or IPv6-only mode.

Nice!

Haven't tried it yet, so a few curious questions:

1) Can fping now ping mixed IPv4 and IPv6 addresses in the same
process?

If so, I have a feature request on that topic (because the same issue
already annoys me with oping :-):

Can you add an option (e.g. -0 or -5) which pings both, A and AAAA
record (as if both IP addresses were given) if a host has both, A and
AAAA record? Or maybe that should be even the default instead of IPv6?

Example:

→ host www.ethz.ch
www.ethz.ch. has IPv4 address 129.132.19.216
www.ethz.ch. has IPv6 address 2001:67c:10ec:4380::216
→ fping -5 www.ethz.ch
www.ethz.ch/129.132.19.216 is alive
www.ethz.ch/2001:67c:10ec:4380::216 is alive

On the other hand that would probably imply that if an host returns
more than one A or AAAA record, both should be pinged:

→ host foobar.weebly.com
foobar.weebly.com. is an alias for pages-wildcard.weebly.com.
pages-wildcard.weebly.com. has IPv4 address 199.34.228.53
pages-wildcard.weebly.com. has IPv4 address 199.34.228.54

Currently, fping (3.15) works as follows:

→ fping -A foobar.weebly.com
199.34.228.53 is alive

But what I'd find helpful is, if it would reply like this:

→ fping -A foobar.weebly.com
199.34.228.53 is alive
199.34.228.54 is alive

> If you need exact compatibility with old versions, you can:
> - compile fping with --disable-ipv6 (or use a wrapper, and call 'fping -4')
> - compile fping with --enable-ipv6 and rename it to fping6 (same as 'fping -6')

2) So "--enable-ipv6" implies "--disable-ipv4"? If so, that option
should probably be renamed to "--only-ipv6" or similar to be
unambiguous.

Kind regards, Axel
--
/~\ Plain Text Ribbon Campaign | Axel Beckert
\ / Say No to HTML in E-Mail and News | a...@deuxchevaux.org (Mail)
X See http://www.nonhtmlmail.org/campaign.html | a...@noone.org (Mail+Jabber)
/ \ I love long mails: http://email.is-not-s.ms/ | http://abe.noone.org/ (Web)

[David Schweikert]

Hi Axel,

On Tue, Jan 31, 2017 at 3:49 PM, Axel Beckert <a...@deuxchevaux.org> wrote:

I'll probably upload 3.16-rc2 to Debian Experimental in the next few
days, too.

Cool, thanks!

Haven't tried it yet, so a few curious questions:

1) Can fping now ping mixed IPv4 and IPv6 addresses in the same
   process?

Yes:

  $ fping ::1 127.0.0.1

  ::1 is alive

  127.0.0.1 is alive

 

If so, I have a feature request on that topic (because the same issue
already annoys me with oping :-):

Can you add an option (e.g. -0 or -5) which pings both, A and AAAA
record (as if both IP addresses were given) if a host has both, A and
AAAA record? Or maybe that should be even the default instead of IPv6?

You mean like this?

$ fping -A -m www.google.com

2a00:1450:4001:816::2004 is alive

216.58.210.4 is alive

(www.ethz.ch is not pingable for me, btw. :))

Also, as you assumed:

$ fping -A -m foobar.weebly.com

199.34.228.54 is alive

199.34.228.53 is alive

This has always been what '-m' was for, but it seems that I broke it a while

ago (and fixed it now)...

It might be useful to have an option to print the original name that was given

in addition to the numerical name. For example like this:

foobar.weebly.com (199.34.228.54) is alive

foobar.weebly.com (199.34.228.53) is alive

There is an option '-n' that you can combine with '-A', but it does a reverse-DNS

lookup, so you might not get the same name, and later you don't know what address

was for what host.

>     If you need exact compatibility with old versions, you can:
>     - compile fping with --disable-ipv6 (or use a wrapper, and call 'fping -4')
>     - compile fping with --enable-ipv6 and rename it to fping6 (same as 'fping -6')

2) So "--enable-ipv6" implies "--disable-ipv4"? If so, that option
   should probably be renamed to "--only-ipv6" or similar to be
   unambiguous.

No, I probably should reformulate that text... When you call the fping binary

'fping6' (with a sym-link, for example), it is going to implicitely set the

'-6' option at runtime. Bad idea?

 

Cheers

David

--
You received this message because you are subscribed to the Google Groups "fping-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to fping-users+unsubscribe@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Axel Beckert]

Hi David,

On Tue, Jan 31, 2017 at 06:20:06PM +0100, David Schweikert wrote:
> Haven't tried it yet, so a few curious questions:
> >
> > 1) Can fping now ping mixed IPv4 and IPv6 addresses in the same
> > process?
>
> Yes:
>
> $ fping ::1 127.0.0.1
> ::1 is alive
> 127.0.0.1 is alive

Cool, thanks!

> > If so, I have a feature request on that topic (because the same issue
> > already annoys me with oping :-):
> >
> > Can you add an option (e.g. -0 or -5) which pings both, A and AAAA
> > record (as if both IP addresses were given) if a host has both, A and
> > AAAA record? Or maybe that should be even the default instead of IPv6?
>
> You mean like this?
>
> $ fping -A -m www.google.com
> 2a00:1450:4001:816::2004 is alive
> 216.58.210.4 is alive

Yes!

> (www.ethz.ch is not pingable for me, btw. :))

*sigh*

> Also, as you assumed:
>
> $ fping -A -m foobar.weebly.com
> 199.34.228.54 is alive
> 199.34.228.53 is alive
>
> This has always been what '-m' was for,

I must admit, I was not aware of that option. And its description in
the (3.15) man page seems suboptimal as I probably would have overseen
it even when I looked for it.

But as I see, you even already fixed the documentation to be way more
clear, too. Yay! :-)

> but it seems that I broke it a while ago (and fixed it now)...

*snicker*

> It might be useful to have an option to print the original name that was
> given in addition to the numerical name. For example like this:
>
> foobar.weebly.com (199.34.228.54) is alive
> foobar.weebly.com (199.34.228.53) is alive
>
> There is an option '-n' that you can combine with '-A', but it does
> a reverse-DNS lookup, so you might not get the same name, and later
> you don't know what address was for what host.

That's not that bad IMHO, but yes, might be confusing.

> > If you need exact compatibility with old versions, you can:
> > > - compile fping with --disable-ipv6 (or use a wrapper, and call
> > 'fping -4')
> > > - compile fping with --enable-ipv6 and rename it to fping6 (same as
> > 'fping -6')
> >
> > 2) So "--enable-ipv6" implies "--disable-ipv4"? If so, that option
> > should probably be renamed to "--only-ipv6" or similar to be
> > unambiguous.
>
> No, I probably should reformulate that text... When you call the
> fping binary 'fping6' (with a sym-link, for example), it is going to
> implicitely set the '-6' option at runtime. Bad idea?

No, that's fine as it is what users would expect. But my comment was
about the compile time (configure) options. (Maybe adding the same for
"fping4" if IPv6 is set as default might be a nice addition.)

[David Schweikert]

On Tue, Jan 31, 2017 at 19:13:29 +0100, Axel Beckert wrote:
> > > 2) So "--enable-ipv6" implies "--disable-ipv4"? If so, that option
> > > should probably be renamed to "--only-ipv6" or similar to be
> > > unambiguous.
> >
> > No, I probably should reformulate that text... When you call the
> > fping binary 'fping6' (with a sym-link, for example), it is going to
> > implicitely set the '-6' option at runtime. Bad idea?
>
> No, that's fine as it is what users would expect. But my comment was
> about the compile time (configure) options. (Maybe adding the same for
> "fping4" if IPv6 is set as default might be a nice addition.)

--enable-ipv6 does not imply --disable-ipv4. You can't disable IPv4 at
compile time.

The idea with 'fping6' is to maintain compatibility for systems that
already use fping6. New systems should use preferably 'fping -4' or
'fping -6' if forcing IPv4/IPv6 is needed.

Cheers
David