|
|
| Author |
Message |
| 25 new of 70 responses total. |
devnull
|
|
response 25 of 70:
| 
May 22 04:27 UTC 1999 |
Re #3: Really? I didn't know grex had 90GB of disk space.
(I recently saw the maintainers of news.mit.edu saying that they want their
new machine to have 90GB...)
|
dang
|
|
response 26 of 70:
|
May 22 19:42 UTC 1999 |
Then I'm behind the times. I don't do much with usenet.
|
other
|
|
response 27 of 70:
|
May 27 01:31 UTC 1999 |
i'm greatly in favor of drew's suggestion in resp:24 above. and tin as
well...
|
fungster
|
|
response 28 of 70:
|
May 30 06:39 UTC 1999 |
How about allowing access on port 79 to members so that they can use
their own Usenet service? Dejanews (oops, "Deja.Com") is notroiously
crappy on lynx, and is even more so with its news interface. And,
reference.com (my favorite site) doesn't seem to be keeping up.
In addition, usenetchannel.com went out of business a couple of
years ago.
For that matter, how about aliasing trn, tin, and rn to one of the
many READ ONLY news services out there, like news.ripco.com,
news.netcom.ca, or www.talkway.com? (www.talkway.com, despite its
name, offers both web-based reading and posting on the http
protocol (port 80) and reading only on port 79. I'm sure they wouldn't
mind if Grex linked off this site.) Indeed, with the overhead of the
huge table enhanced files from services like deja.com, this would
actually be more beneficial to Grex's link, and use up less
bandwidth.
Also, you might want to talk to jared and see if he will "share"
nether.net's news.nether.net (actually iagnet.net news feed)
service. This may be doubtful, though, because the license may
only cover one block of IP numbers.
Here's a link to legally accessible (as opposed to news servers
left on because the admin forgot or was too stupid to turn off
global accessing), public, read only news servers:
http://www.gj.net/~bhkraft/
|
scg
|
|
response 29 of 70:
|
May 30 17:28 UTC 1999 |
Members have access to all TCP and UDP ports, so port 79 is there.
|
fungster
|
|
response 30 of 70:
|
May 30 22:54 UTC 1999 |
I meant port 119 (NNTP). Why not allow access to port 119 on www.talkway.com
for everyone? (79 is finger.)
|
scg
|
|
response 31 of 70:
|
May 30 23:16 UTC 1999 |
You're right. I should have read more carefully. Anyhow, as I said, members
have access to all TCP and UDP ports. 119 would be in cluded in all, just
as 79 would.
|
mdw
|
|
response 32 of 70:
|
May 31 01:45 UTC 1999 |
If it helps any, port 131072 isn't included in "all ports".
|
dang
|
|
response 33 of 70:
|
May 31 18:40 UTC 1999 |
Why not?
|
mdw
|
|
response 34 of 70:
|
Jun 1 01:49 UTC 1999 |
Think about it.
|
dang
|
|
response 35 of 70:
|
Jun 1 15:24 UTC 1999 |
Well, the only reason I can think of is that it's above 65,536. If
that's the case, then I was confused because I thought it was unique for
some reason.
|
mdw
|
|
response 36 of 70:
|
Jun 2 06:52 UTC 1999 |
What degree program are you in? Try port 131095.
|
janc
|
|
response 37 of 70:
|
Jun 2 14:19 UTC 1999 |
I have no idea what Marcus is talking about either.
|
rtg
|
|
response 38 of 70:
|
Jun 2 16:29 UTC 1999 |
It appears that 'port number' is a 16-bit integer. When I connect to grex
at port numbers higher than 65,536, I get a service that appears to be a
16-bit truncation of the number I requested.
|
scg
|
|
response 39 of 70:
|
Jun 2 17:07 UTC 1999 |
131015 appears to be telnet, aka 23.
|
mdw
|
|
response 40 of 70:
|
Jun 3 00:23 UTC 1999 |
Jeeze, and here I thought I was surrounded by computer people.
|
scg
|
|
response 41 of 70:
|
Jun 3 00:36 UTC 1999 |
Er, I should have said 131095, not 131015, since that was the number being
discussed in Marcus's most recent response. 131095 is telnet, aka port 23.
I'm guessing this has something to do with 65536 * 2 + 23 = 131095. Going
by that system, 65536 * 2 = 131072, which presumably means that 131072 = 0,
which is an invalid port. That all seems straight forward enough, but is
there any practical application for which this would be useful information,
other than confusing other Coop participants?
|
mdw
|
|
response 42 of 70:
|
Jun 3 04:19 UTC 1999 |
Absolutely none.
|
dang
|
|
response 43 of 70:
|
Jun 4 21:41 UTC 1999 |
I don't see how I should have known that it wrapped. If I were writing
telnet, I would return an invalid port error, and not truncate it.
|
mdw
|
|
response 44 of 70:
|
Jun 5 05:17 UTC 1999 |
Ah, but would you be surprised to see something like "sin.sin_port =
atoi(cp)"? A few newer C compilers might complain about the truncation,
but most won't.
Anyways, now you've learned where "telnet 7731424898 131015" will take
you.
|
remmers
|
|
response 45 of 70:
|
Jun 8 18:06 UTC 1999 |
This 16-bit restriction on port numbers discriminates against systems
which wish to offer more than 65536 services and should therefore be
abolished.
|
other
|
|
response 46 of 70:
|
Jun 8 19:23 UTC 1999 |
gosh, yes! every user everywhere should have their own personal port on every
machine they connect with. with only 65536 ports available, that kind of
essential progress will be stifled and all of society as we know it will
crumble and decay!
|
hhsrat
|
|
response 47 of 70:
|
Jun 10 02:00 UTC 1999 |
Are there more that 65536 possible services that a system could offer?
|
remmers
|
|
response 48 of 70:
|
Jun 10 11:30 UTC 1999 |
Of course! But to save space, I won't list them all here.
|
ryan
|
|
response 49 of 70:
|
Jun 10 13:56 UTC 1999 |
This response has been erased.
|