Hi Folks, LimeWire has recently been including a new header during HTTP exchanges: X-Create-Time. This creation time is designed to create a network-wide...
... Why don't you use the Last-Modified header instead? The definition of the header allows to return something else than "date of the last write access". ... ...
Thanks for the comments, Christian. ... We are using a new X-Create-Time header because it is not intended to be the last modified time. It is intended to be...
Interesting. What are the anticipated uses of this new header? -Justin ... -- Justin Chapweske - Founder, Onion Networks http://onionnetworks.com/ ...
Justin Chapweske
justin@...
Jun 1, 2004 8:16 pm
20057
... Uhh, no, Dave. The G2 (or, "MP") design is sound. The G1 out-of-band query design is flawed. Why? The G2 design recognizes that UDP packets can be sent by...
Correct me if I'm wrong, but UDP packets are always received in whole. Therefore, it is trivial to look at byte 16 and determine if the type matches a known...
... Mike brought up some good points. Me and Greg had an almost identical conversation regarding different protocols multiplexed onto the same UDP listening...
... Yes, the points are indeed valid. Changing the entire protocol for some small (easily overcome) issues is a bit drastic, though. ... Any design (even with...
... I agree -- I didn't say it was impossible to identify a G1 packet, just that it is more difficult than it needs to be, and requires going deeper into the...
Mike, Admittedly, much of my earlier gripe isn't specific to MP Udp data, depite me naming that traffic explicitly. Most of the rant, even, isn't *exclusively*...
Since I've been getting GDF digests I sometimes miss the brunt of interesting discussions :) . I want to make one point about the design of Gnutella's OOB...
Has Limewire considered placing their handshaking test that Christopher Rohrs coded in the GDF files section? I know I'm about 3 years late on this, but the...
On Thu, 2004-06-03 at 15:04, Susheel wrote: ... Is there some way of providing feedback to the LimeWire users that have the buggy clients installed to upgrade?...
Most users would be confused by the presence of a status code (or even the reason phrase). It's not a huge issue, as LimeWire nodes will be able to parse the...
For a bit now, I have been receiving a significant number of pongs from hosts in Asia, mainly China, who are listening on port 2500, and running Gnucleus...
... Technically, the "X-Node:" header with a double colon is not malformed in HTTP (there's no requirment that the colon be followed by a space, so the second...
Asked it before, but how to test a new client... And also, how come I never seem to connect to a limewire or bearshare client... Jeroen ... From: Sam Berlin...
Hi, QRT take care to don't loose our bandwith in that way... And don't forget, some people talk both language. Frederic ... now, I have been receiving a...
Hi, I see GWC requests from LimeWire 4.0.5 where "?urlfile=1" is already appended to the URL - encoded, of course. Please, fix your code, you're trashing...
Don't say "fix our code" if it is stupid gwebcaches. You can say "program defensively" or "clean up stupid gwebcaches". Thanks -greg ... From: Christian Biere...
Hi folks, I'm currently working on the local host cache implementation of gtk-gnutella and I wonder what a reasonalable default number of hosts to keep around...
Richard Eckart
wyldfire@...
Jun 6, 2004 4:45 pm
20081
We keep around 400 hosts and our best source is pongs with long uptime information. However, everyone should implement that uptime feature and reset it if the...
You're making a false assumption here: in the /8 CIDR blocks you have NO block allocated only to CHINA, but to APNIC which subdelegates some parts of these...
