Slyck.com
 
Slyck Chatbox - And More

K-Lite is Dieing..

Discussions about the FastTrack network and any clients that connect to it
Forum rules
PLEASE READ BEFORE POSTING: Slyck Forum Rules

Postby fire01 » Fri Jul 16, 2004 9:11 am

You're wrong ashton i wasn't using kceasy i was using kazaa lite k++ v 2.4.3
fire01
 
Posts: 18
Joined: Tue Jun 15, 2004 12:50 am

Postby ashton » Fri Jul 16, 2004 1:04 pm

Ooble wrote:It should be easy to fix - just add "unsigned" to the variable - but then I'm not exactly experienced with this sort of thing. 'Course, this only fixes the problem if it only happens on giFT.
In gift-fasttrack the stats are unsigned integers already.
Last edited by ashton on Fri Jul 16, 2004 1:08 pm, edited 1 time in total.
"There is a very thin line between genius and madness"
-Julian
User avatar
ashton
 
Posts: 290
Joined: Thu Nov 20, 2003 5:22 pm
Location: New York City

Postby ashton » Fri Jul 16, 2004 1:08 pm

fire01 wrote:You're wrong ashton i wasn't using kceasy i was using kazaa lite k++ v 2.4.3
Which is actually KMD 2.0.2, I get this issue with gift-fasttrack and when I check the supernode it is always a newer node(2.6.x) and as I mentioned gift-fasttrack does use unsigned int's for each stat field variable. I am wondering if they have recently made changes to the stats code in 2.6.4??
"There is a very thin line between genius and madness"
-Julian
User avatar
ashton
 
Posts: 290
Joined: Thu Nov 20, 2003 5:22 pm
Location: New York City

Postby Ooble » Fri Jul 16, 2004 7:06 pm

ashton wrote:In gift-fasttrack the stats are unsigned integers already.

In which case it's getting an incorrect number from the network. Not sure what you can do about this, except add in code to fix it.
[code:1]char str[20];
scanf("%19s", &str);
if (!strcmpi(str, "Hello, world!"))
printf("Hello, programmer.\n");[/code:1]
User avatar
Ooble
 
Posts: 2372
Joined: Sat Dec 06, 2003 9:33 am
Location: Wouldn't you like to know, SlyckTom? Just so you can sell my address to the RIAA? No chance.

Postby ashton » Sat Jul 17, 2004 2:16 am

Ooble wrote:
ashton wrote:In gift-fasttrack the stats are unsigned integers already.

In which case it's getting an incorrect number from the network. Not sure what you can do about this, except add in code to fix it.
This is why i am wondering if the 2.6.4 update of KaZaA may have changed the stats provider or mucked it up, cause the packet data once decrypted is a negative, this happens rarely and seems only when connected to these 2.4.6 supernodes.
"There is a very thin line between genius and madness"
-Julian
User avatar
ashton
 
Posts: 290
Joined: Thu Nov 20, 2003 5:22 pm
Location: New York City

Postby tm, » Sat Jul 17, 2004 2:51 am

ashton wrote:
tm wrote:
ashton wrote:the function disable_networklicense has been removed from the FastTrack stack because the recording industry would be able to proove they could shut FT down

@Ashton,
Do you know when this happened - that Sharman no longer gave itself the ability to shut down any FT client? Does this mean that older versions with the "feature" can still be remotely disabled? And were you able to recreate Sharman's code that would enable you (or anyone else) to remotely disable an old FT client?

This certainly throws a monkeywrench into Sharman's legal strategy of claiming that they had no direct control over the network.
Pre 1.7.X clients contain that function. This call is injected into the root Supernodes, I have a list somewhere of them. This is what used to broadcast the "Ugrade" packet that contained the payload of features and also a tailing and encrypted section that could disable a client license, a la Morpheus, so they could target any client. There really is no way to recreate this as the functionality has most likely been removed from the root Supernodes, iMesh no longer has to pay for their FT stack because A. The contract has expired from the "Consumer Empowerment" firm, a company that Joltid made up to license out the stacks which no longer exists and B. The function has been removed as i stated from the FT stack.

Grokster was still using the v.1.6 stack until just a few months ago (and maybe some users never upgraded?) so it's funny that they won against the RIAA in court based on the argument that there was no way to control users - all the while Grokster users could have been remotely shut down on a whim.

I never saw much said about those root supernodes - other than their existence. I wonder if anyone has ever come close to hacking into one of them or breaking the encryption?

With KCeasy around today, it's a wonder why there is still a lot of effort to create a v2.6 K-lite client. Maybe old habits are hard to break.
tm,
 
Posts: 1192
Joined: Wed Apr 21, 2004 12:23 am

K-Lite is Dieing..

Postby SlyckChuck » Sat Jul 17, 2004 2:56 am

With kazaa nothing is beyond mucking up. The app is dying just like the headline says. Too many fake files and too few who share. Let's do ourselves a favor and let it die away.... :idea:
The 80's Still Rule
User avatar
SlyckChuck
 
Posts: 7025
Joined: Sun Jan 18, 2004 12:57 pm
Location: On Earth

Postby Ooble » Sat Jul 17, 2004 5:08 am

Let's not - how else are we supposed to protect ourselves from the RIAA?
[code:1]char str[20];
scanf("%19s", &str);
if (!strcmpi(str, "Hello, world!"))
printf("Hello, programmer.\n");[/code:1]
User avatar
Ooble
 
Posts: 2372
Joined: Sat Dec 06, 2003 9:33 am
Location: Wouldn't you like to know, SlyckTom? Just so you can sell my address to the RIAA? No chance.

Postby hafnium » Sat Jul 17, 2004 5:15 am

Ok, I am going to repeat myself and I suggest those who do not believe me look up how two's complement works. There is no problem with the stats generated by the network. Only the code which displays them on screen is broken as it assumes the numbers are signed when they are actually not.

Oh and those root supernodes haven't been existing for a long time. The remote update code on the other hand is still in all versions of Kazaa. You can inject and update from anywhere if you have the right key...
hafnium
 
Posts: 10
Joined: Sun Jun 27, 2004 12:02 pm

Postby ashton » Sat Jul 17, 2004 5:47 am

hafnium wrote:Ok, I am going to repeat myself and I suggest those who do not believe me look up how two's complement works. There is no problem with the stats generated by the network. Only the code which displays them on screen is broken as it assumes the numbers are signed when they are actually not.

Oh and those root supernodes haven't been existing for a long time. The remote update code on the other hand is still in all versions of Kazaa. You can inject and update from anywhere if you have the right key...
Would you stop saying things that are totally incorrect. My Mac application XFactor was showing negative files and the packets i sniffed and decrypted contained a negative integer coming from the supernode, this started happening with KMD 2.6.4. Also no the remote upgrade code is not in kazaa 2.6, if you think it is show me the code and offset. :wink:
"There is a very thin line between genius and madness"
-Julian
User avatar
ashton
 
Posts: 290
Joined: Thu Nov 20, 2003 5:22 pm
Location: New York City

Postby hafnium » Sat Jul 17, 2004 6:12 am

ashton wrote:
hafnium wrote:Ok, I am going to repeat myself and I suggest those who do not believe me look up how two's complement works. There is no problem with the stats generated by the network. Only the code which displays them on screen is broken as it assumes the numbers are signed when they are actually not.

Oh and those root supernodes haven't been existing for a long time. The remote update code on the other hand is still in all versions of Kazaa. You can inject and update from anywhere if you have the right key...
Would you stop saying things that are totally incorrect. My Mac application XFactor was showing negative files and the packets i sniffed and decrypted contained a negative integer coming from the supernode, this started happening with KMD 2.6.4.

See, that's exactly what I meant about you needing to look up two's complement. Whether the number is negative or not cannot be deduced from the sniffed packets. It is solely a matter of interpretation. I think we agree that there is actually a postive number of nodes on the network. It thus follows that you and the display code interpret the numbers incorrectly.

As for the auto update code. I don't care enough to look up where exactly it is. I just don't want people to be misinformed.
hafnium
 
Posts: 10
Joined: Sun Jun 27, 2004 12:02 pm

Postby ashton » Sat Jul 17, 2004 11:28 am

As for the auto update code. I don't care enough to look up where exactly it is. I just don't want people to be misinformed.
Your misinforming people here bby saying this function does indeed exist in in KMD 2.6 when in fact it does not, if you 'think' it does then some factual information showing this would be wise. Thanks
"There is a very thin line between genius and madness"
-Julian
User avatar
ashton
 
Posts: 290
Joined: Thu Nov 20, 2003 5:22 pm
Location: New York City

Postby lovedrug » Sat Jul 17, 2004 9:21 pm

it wont die :roll:
User avatar
lovedrug
 
Posts: 55
Joined: Fri Jul 16, 2004 7:42 am

Postby jd » Sun Jul 18, 2004 4:55 pm

There's no remote update code in KMD since many versions ago, prolly since 2.0.0 maybe earlier.

And about that 32-bit integer: since we know that there aren't 2^31 nodes or even users, we know that something's wrong with that number.
jd
 
Posts: 11
Joined: Wed Mar 10, 2004 2:51 pm

Postby jd » Sun Jul 18, 2004 5:00 pm

ashton wrote:Cause they can't unpack 2.6. :lol:


KMD26 is using a homemade protection so it's very easy to unpack. With source code to klr.exe already available I'm surprised no-one even attempted reading the source code to see exactly what needs to be done to unpack it properly. Well, their loss I guess.
jd
 
Posts: 11
Joined: Wed Mar 10, 2004 2:51 pm

Postby hafnium » Sun Jul 18, 2004 5:05 pm

jd wrote:There's no remote update code in KMD since many versions ago, prolly since 2.0.0 maybe earlier.


I suppose you guys have done an exhaustive search of the entire code base to prove that?

Hint: The exact same key which is used for the update verification is also part of the protocol encryption. If it was missing nothing would work.

jd wrote:And about that 32-bit integer: since we know that there aren't 2^31 nodes or even users, we know that something's wrong with that number.


We are talking about the number of files, not users.
hafnium
 
Posts: 10
Joined: Sun Jun 27, 2004 12:02 pm

Postby ashton » Sun Jul 18, 2004 5:52 pm

hafnium wrote:I suppose you guys have done an exhaustive search of the entire code base to prove that?
You are correct.
"There is a very thin line between genius and madness"
-Julian
User avatar
ashton
 
Posts: 290
Joined: Thu Nov 20, 2003 5:22 pm
Location: New York City

Postby HEAT84 » Mon Jul 19, 2004 2:05 am

Somebody in this thread said that KCeasy was immune to this problem. Well apparently not because I was having to disconnect and reconnect several times with KCEasy befor it showed 2 million users.

And how come everyone ignored the post that had a link to a fix in it? Does it not work?
User avatar
HEAT84
Official Slyck Conspiracy Nut
 
Posts: 1270
Joined: Mon Jan 27, 2003 8:41 am
Location: Neither Here Nor There

Postby ashton » Mon Jul 19, 2004 2:39 am

HEAT84 wrote:Somebody in this thread said that KCeasy was immune to this problem. Well apparently not because I was having to disconnect and reconnect several times with KCEasy befor it showed 2 million users.

I may have mentioned something, XFactor is immune to this problem, not sure what about kceasy as i don't often use it.
HEAT84 wrote:And how come everyone ignored the post that had a link to a fix in it? Does it not work?

It's not a "fix" it is a temporary patch, it does solve the problem permanently.
"There is a very thin line between genius and madness"
-Julian
User avatar
ashton
 
Posts: 290
Joined: Thu Nov 20, 2003 5:22 pm
Location: New York City

Postby tm, » Mon Jul 19, 2004 4:21 am

HEAT84 wrote:Somebody in this thread said that KCeasy was immune to this problem. Well apparently not because I was having to disconnect and reconnect several times with KCEasy befor it showed 2 million users.

And how come everyone ignored the post that had a link to a fix in it? Does it not work?

It was pointed out earlier in the thread that cross-links do occur frequently between the 2 sub-networks. That posted "fix" was just a list of working nodes. As I discovered earlier, node-shopping is really just a bandaid, not a cure. The only feasible permanent solution is to drop K-lite (until they update to 2.6 core), and switch to using KCeasy with the latest FT plugin.

I just checked over at the Imesh forums, and users are reporting the same problem of varying user population - which will be "fixed" in the v.5 Imesh by simply dropping the user statistics.

I wonder though, is there a way to connect to both the old Klite/Imesh network and the new Fasttrack 2.6 network at the same time?
tm,
 
Posts: 1192
Joined: Wed Apr 21, 2004 12:23 am

Postby tomy » Mon Jul 19, 2004 6:58 am

tm wrote:I wonder though, is there a way to connect to both the old Klite/Imesh network and the new Fasttrack 2.6 network at the same time?

Right now im using old k-lite k++ and i get over 2 million users. I think im connected to both the old Klite/Imesh network and the new Fasttrack 2.6 network at the same time.
tomy
 
Posts: 158
Joined: Tue Mar 23, 2004 2:55 pm

Postby BigWillyStyle42 » Mon Jul 19, 2004 7:24 pm

I have to agree with hafnium on this one. That's a display bug nothing else. Let's pretend for one second that the number of files is stored in a 8 bit number to simply this example. From the packet you get the following 8 bits:
Code: Select all
11111111


How many files are on the network? 255 or -1? Well obviously since it makes no sense for the number of files and users to take on negative values there are 255 files on the network. That's a display bug.
BigWillyStyle42
 
Posts: 798
Joined: Mon Sep 29, 2003 8:22 pm
Location: Rockville, MD

Postby ashton » Mon Jul 19, 2004 9:30 pm

BigWillyStyle42 wrote:I have to agree with hafnium on this one. That's a display bug nothing else. Let's pretend for one second that the number of files is stored in a 8 bit number to simply this example. From the packet you get the following 8 bits:
Code: Select all
11111111


How many files are on the network? 255 or -1? Well obviously since it makes no sense for the number of files and users to take on negative values there are 255 files on the network. That's a display bug.

Answer me this, the negative value has never been seen before Sharman released 2.6.4 a few weeks back, now every client receives this negative value from supernodes(from time to time) even third-party clients like xfactor and kceasy, it's a new bug introduced from kazaa 2.6.4, call it a "display" bug or whatever, but the fact remains it was introduced into the network with the last KMD.
"There is a very thin line between genius and madness"
-Julian
User avatar
ashton
 
Posts: 290
Joined: Thu Nov 20, 2003 5:22 pm
Location: New York City

Postby HEAT84 » Tue Jul 20, 2004 2:54 am

Now when I connect it only says 55 users. Are things getting worse?
User avatar
HEAT84
Official Slyck Conspiracy Nut
 
Posts: 1270
Joined: Mon Jan 27, 2003 8:41 am
Location: Neither Here Nor There

Postby SlyckTom » Tue Jul 20, 2004 4:25 pm

I started seeing this before I went on vacation but didnt think anything of it. I also use Kazaa lite and saw 95,000 -odd users.

So its seems Sharman is finally, after all this time, segmenting the KL crowd out of FastTrack...I read through the entire thread but probably missed this point...but why does the KL client connect to the 2.0 million portion of the network mostly but not at other times? What is causing this? (perhaps time for a story on this...)
SlyckTom
 
Posts: 5713
Joined: Fri Jul 26, 2002 7:22 pm
Location: New York City

PreviousNext

Return to FastTrack/Kazaa

Who is online

Users browsing this forum: No registered users and 1 guest

© 2001-2008 Slyck.com