RFC: discard ping replies with rtt > timeout (-t) ?

[David Schweikert]

Hi,

Currently, fping behaves a bit strangely regarding timeouts, when in
"count" or "loop" mode. Take this example:

fping -c 3 -p 5000 host

fping will behave as follows:

1) send ping and wait 5000 ms
2) send ping and wait 5000 ms
3) send ping and wait 1000 ms (default -t timeout)

If the host is 1300 ms away, this will result in two pings being
received, and one lost. That's really unexpected, I would say.

This issue was already reported in github issue #32
(https://github.com/schweikert/fping/issues/32), but I am not
sure about how to fix it.

One proposal would be to change the timeout behavior as follows
(for fping >= 4.0):

- When in loop or count mode, set the default timeout (-t) value,
based on the value of -p, instead of 1000 ms.

In the example above, it means that "-t 5000" would be implicit,
but you could override it, if you want.

- Enforce the -t value by discarding replies with rtts larger
than the timeout.

Let me know what you think! You can either post your comment in the
github issue, or reply here.

Cheers
David