Quote:
Originally posted by Joakim Agren
And since the nodes that you are connected to constantly changes they probably do not stay intouch with an individual servent for a long enough time to get all the QueryHits they need.
|
I don't think the network really is that volatile. And even if it was, the spynode wouldn't have to gather all that information in one session since your client has an unique servent id, that will allow to identify it, even after its IP has changed. They could track you for weeks, before they finally think they've gathered enough information.
Quote:
|
That is not entirely true.It depends on what security features you mean but just a new countermeasurment function to deny connections to hostile nodes is still very doable for the LimeWire team to impliment whitout beeing close source.
|
You cannot identify hostile hosts. The spynode could use any user agent string without anyone ever recognizing the spynode. It's not a crime to spoof a user agent string.
Quote:
|
And the QueryHIt messages are not proof of any file transfers.
|
Advertising a huge index of stolen copyrighted material is proof enough for most ISPs to send their users a notice when the RIAA tells them to, and the ISP can shut you down, even it is not entirely proven that you actually uploaded those files. And if they did so, what are you going to do? Sue them? (You'd better not if you don't want your computer confiscated by the local police.)
Quote:
|
I think that you underestimate the US court system.
|
And I think you should face reality. No judge would honestly believe you if you said you were just advertising this huge index of stolen copyrighted material to steal more copyrighted material and you weren't actually uploading anything. This is not only about proof. If a bomb explodes in your local mall and you happened to have bought the ingredients for that bomb, you go to jail, no matter what you did.
There are so many people that were found guilty although they were truly innocent and you think you would get away with saying:
"Hey, I weren't actually uploading anything." That's pathetic.
Quote:
|
To get some real hard evidence they have to prove that I do really share files and hence they have to first make a query and in the search results that comes up they have to Browse My shared directory to see if I share alot of files or if I am a just a Small sharer. And then start to download files from me and aslo save the result from the Browsing of my host.
|
You simply don't get it, do you? THEY DON'T HAVE TO BROWSE YOUR F U C K I N G HOST, TO GET AN ALMOST COMPLETE LIST OF YOUR SHARES!
Quote:
|
That is not correct!.I have set the amount of files that someone have to have in their shared directory to 20 in order for them to download anything from me. LimeWire takes the information from the Pong messages that contain data about the IP,portnumber,number of files and the total size of all those files. If a host send out a number below 20 files in their Pong response then that node will be refused and my node upon a request will not send out the http string data to that node that he needs to make an http connection and upload from me. And yes ofcourse the Anti Freeloader only works for Gnutella connections and not for uploads but since he cannot make a http connection to me then he cannot make uploads.
|
That's it. Okay, I really don't get it. Are you dumb or something? Somebody tells you that freeloader blocking does not work as you expected and what do you do? YOU SIMPLY MAKE SOMETHING UP, AND SAY THAT'S WHY IT WORKS AS I THOUGHT! You are unbelievable. I have the source code on my hard drive. I read the source code. I edited the source code and never it occurred to me, that LimeWire would ever reject a HTTP connection without even sending an error code. Not to mention that the algorithm you described wouldn't even work, since you don't necessarily receive many pongs at all and identifying a client according to their IP number is anything but safe.
The freeloader algorithm is in StandardMessageRouter.java in the method handlePingReplyForMe() and in HTTPUploader.java in the readHeader() method and all it does is killing uploads to browsers (identified by their user string) and killing gnutella connections to hosts not sharing files.