On Thu, Apr 12, 2007 at 10:49:30AM +0200, Ulrich Eckhardt wrote: > > Resolving this bug, as I don't believe any aspect of ping's behavior > > should change. > > I wouldn't simply close this. The point is that the behaviour of ping is > inconsistent, albeit in the context of me doing something > dubious/stupid/useless. If I ping a broadcast address, ping _must_ expect > multiple answers, because it is by definition a one-to-many ping. Therefore, > waiting must not terminate after the first answer, too, because it is > not 'the answer' but just 'an answer'.
Remember that when you first submitted this bug (I know it was a while
ago), you were using -c1. Quoting the man page description of -c:
-c count
Stop after sending count ECHO_REQUEST packets. With deadline
option, ping waits for count ECHO_REPLY packets, until the
timeout expires.
ping behaves exactly as documented, and entirely consistently, in this
case. What behavior do you expect?
> In that context, a proper error-message would also be welcome. Even when
> broadcasting, ping still claims duplicate answers, as if those were
> unexpected. Yes, it does give a warning that I'm pinging a broadcast address
> (heck, I explicitly requested!) but not why that might be a bad idea.
Would you prefer that the "(DUP!)" note be supressed when -b is
provided?
> At the very least, this inconsistency should be documented in the manpage (I
Which inconsistency are you referring to?
noah
signature.asc
Description: Digital signature
