12/22 Why does last output show most connecting from IP instead of hostname?
\_ because FreeBSD is stupid. -tom
\_ because tom is stupid. -FreeBSD
\_ wait a minute. FreeBSD can't write.
\_ How is it stupid, tom? There's only so much room to store
or display hostnames, I'd imagine. If the hostname is longer
than that when reversed, would you rather get
"cx425.sanjose.a" or a real IP address that you could resolve
yourself?
\- tom lacks fu to "last | ip2hostname" --psb
\_ It's stupid because every other system in the world uses
hostname there and truncates it to the length of the field.
Intelligent systems which want the full length use an
extended utmp/wtmp. And if you're going to use IP's, it's
the height of stupidity to only use them some of the time.
-tom
\_ Reason #1 "it's stupid because other people truncate
so you have incomplete and invalid hostnames" Uh..huh
Reason #2: it's stupid because... it's the height of
stupidity. Good, tom.
\_ How about Reason #1, if you want to get the normal
output of last now you have to do hundreds of name
lookups. Typical BSD arrogance, "who cares what
behavior people need or expect." -tom
\_ OK, tom: would you rather have chopped off,
indecipherable hostnames (which is what _every
one_ of those ip addrs would be) or ip addresses
that you _could_ look up?
\- if you need to do lookups, you do them,
if you dont need to do them, you dont.
you are right this does make life harder
for low-fu people and might not be a
good decision if your goal is to
maximize the number of people using your
OS. but there are other goals --psb
\_ I would rather have "last" be what it
always has been, and provide an additional
option for new functionality, instead of
changing the default behavior which has
existed for n years (n > 20). In terms of
functionality, I find soda's "last" to be
far less useful than a typical "last",
klee ttyAm 128.32.191.92 Fri Dec 22 19:45 - 19:46 (00:00)
because the information I'm looking for
is usually not a specific hostname. -tom
\_ So back to my question, why are some with IP and some hostname?
What causes it to be diff? thank you.
\_ Picking a few lines at random, we have:
mchowla ttyAz 209.131.52.33 Fri Dec 22 19:47 - 19:50 (00:02)
samli ttyAm <DEAD>charon.sun.com<DEAD> Fri Dec 22 19:47 - 20:02 (00:15)
as you can see, if the hostname were much longer than
"charon.sun.com", it wouldn't fit and you'd have to chop it
off.. let's say it was "sjsu3.sj.ca.ibx4.colo37.cnw4.cnw.net"
you'd see something like "sjsu3.sj.ca.ibx4" as the hostname,
which completely useless. Better to give the IP, where you could
at least run something like "last | ip2hostname" and have all
of the hostnames (if poorly formatted)
\_ so basically, it's inconsistent behavior. Why not all IPs?
And why a 16 char limit? IP is 15 char max(for now).
\- do you feel stuff reporting in 512b
blocks and have done so for n>20yrs
should comtinue to do so? --psb
\_ I dunno, my feeling is that it is
better to fix things than to have
fundamentally broken stuff lay around
forever along with alternate "fixed"
ways of doing things. (Ever tried to
program the Windows API? It fucking
sucks, and this is exactly why.) -blojo
\_ Hear hear. It's not like this
is even some programming interface.
How many critical cross platform
apps do you know of that depend on
the formatting of last?
\- tom, there are times where change is merited.
the chown user.group was changed to user:group ...
i bitched about that till i realized it made sense.
"the rest of the world" doesnt do everthing right the
first time ... so its a good thing they dont have their
head up their ass and can look around and learn. --psb
\_ Tom said it's stupid so it's stupid. Why do you all waste his
precious netrek time asking why? He said so and that should be good
enough for the likes of you!
\_ this thread is priceless. it brings a tear to my eye.
\- "i am tom, hear me roar" |
http://csua.com/feed/