I just installed the newest 0.9.6c, now it can finish probing my firewall within 1-2 minute. But I guess it is because this time I install by the .exe(win32 installation file at Sourceforge.net), instead of the java web start autoinstaller at the official site. May be the java web start autoinstaller are not good for Windows XP, but I don't know why.
Besides, I found that if install Ants to location other than the default one(C:/program file/...), the "Chunk" folder will not be create, the Chunk location in setting will be the root dir of ANts. Is it a bugs? :goodjob
Sorry, my newly installed Ants is being fcuked again now :hole
It was ok during the first install. But later when I changed to install back to its default location(sure I had uninstalled the previous copy), the chunk folder was not created and the chunk location in setting still stick at root dir of ANts. Also, I can't connect to irc anymore, just like those previous version. It keep trying connect to irc then disconnect and cycle again. Firewall probing is endless. :sw
Can't help you with the Chunks problem. But the IRC channel seems not to be very reliable in general. Sometimes it works, sometimes not.
The firewall probing is initiated by the IRC chatroom so I would expect that not to work when IRC doesn't.
As long as you get some peers, you should be fine. If you wan to get peers without the IRC channel, make sure that you have the Gwebcache options ticked on the settings tab.
"Publish IP on Gwebcache" and "Search peers on Gwebcache"
There's also an "unofficial" IRC channel that is available on the accessirc.net server. It's not widely known or used so you may find yourself the only one there.
Hope that helps.
Doubt is not a pleasant condition, but certainty is absurd. - Voltaire.
Yes, you should make sure you are running the latest version. 096b was apparently pretty well broken according to Gwren's release comment on Sourceforge.
Sure it is 0.9.6c man... I got the install file a few hours ago at http://sourceforge.net/projects/antsp2p/, stated 0.9.6c right?
Can it be wrong?
It could also be that not many users have updated to 096c yet, assuming 096c nodes are not compatible with previous versions. A lot of Ants users are in Europe and most of us are at work or school at the moment.
Doubt is not a pleasant condition, but certainty is absurd. - Voltaire.
I found that the sequence like that: when a newly installed Ants is open, the Chunk folder still not exist. After sometime of operation(1-2 minutes), I can see the Chunk folder created with some index folder also(but the chunk location in setting is still saying C:/Program Files/ANts/, not C:/Program Files/ANts/Chunk/)... Then unless the irc connected, the firewall probing will continuous in red. I guess the irc cause an immediate failure to Ants functioning. Why not Ants creators setup a reliable private irc for the purpose?
Lack of IRC does not cause Ants to stop functioning. Your node should remember nodes it's connected to in the past and try them. You can also add node IPs directly if you now someone you trust, or you can use the Gwebcaches, or Jabber buddies if you use the Jeti plugin.
The firewall probing has no effect on the actual functionality of Ants, it's purely to indicate firewall configuration issues visually. If you have connections, and you are proxying data, there is no problem and the firewall probing stuff can be ignored.
Doubt is not a pleasant condition, but certainty is absurd. - Voltaire.
9.6c seems to be running fine. I installed with Java Web Start. If you have problems remove it and reinstall. I have had trouble with IRC window occasional not opening, that is a simple mater of restarting Ants
Assuming you ip address is correct and ports open...
When you get message that it is probing, go to the IRC list and find you name. Right click on your name and hit reprobe. When i do this mesage immidiatly turns blue and tells me that i am set correctly. Down the road Gwren should put a reprobe button righr on the connection panel for people to find.
"With 2000 years of examples behind us we have no excuse when fighting, for not fighting well."
T. E. Lawrence
http://mysite.verizon.net/tsafa1/swordreview.htm
http://mysite.verizon.net/tsafa1/workout.htm
ABC - you sound very positive and happy with ANts - I am not at home so it is hard to say yet, I had to upgrade just before leaving for work - so I hope I find the same thing when I get home. What are you finding with searches - in my opinion this still needed fixing - but it may be done by now.
Best regards
Chris
I don't like the idea of PUSH based replication. Wouldn't that make it possible for a hacked node to overwhelm it's neighbours and pollute the DHT by sending out huge amounts of nonsense results?
Doubt is not a pleasant condition, but certainty is absurd. - Voltaire.
Maybe the query interval and result number could be configurable depending on node "load". Gwren has made much of Ants' efficiency wrt. this, so I wouldn't want to interrupt up a node that is busy up/downloading/proxying with a search.
Perhaps part of a node's hash table (chosen at random) could be sent along with file data it uploads?
It would be outside the layer of encryption around the file data, but inside the point-to-point encryption.
Am I right in thinking that it would then be available to every node in the chain it passes along?
Doubt is not a pleasant condition, but certainty is absurd. - Voltaire.
Me? A coder? No. I'm just throwing ideas around. If we can come up with something that sounds good, then I'll email Gwren with the thread URL.
I would keep it as simple as possible. Using resources and methods already in place.
To keep it PULL based, a downloader could automatically request some of an uploaders hash table along with the file data it really wants.
The uploader adds the hash table info to it's first few packets and it all gets wrapped up in the EE encyption, safe from prying eyes along the way.
That would tend to give downloaders information about files from uploaders sharing the files they want. This would tend to give better search results for downloaders:
eg:
Someone likes to download files of type X.
I share lots of stuff of type X.
If they once find something of mine and download it, they will get file information back from me about more stuff from my share, which will tend to be stuff they want.
In order to spread the information widely, there would have to be a random element. The uploader could add some file hash data from it's remote shared index as well as or instead of it's own file info, on a random basis.
I would weight the choice of what to send based on what filetype is being downloaded, what the uploader has already advertised this session/day/week/whatever, and maybe on other factors yet to be determined.
What does everyone think?
Doubt is not a pleasant condition, but certainty is absurd. - Voltaire.
Email sent.Originally Posted by ABC_thellookoflove
There is no way most people could run an Emule server on their PC, let alone one programmed in JAVA!
Doubt is not a pleasant condition, but certainty is absurd. - Voltaire.
Posting Permissions
Reply With Quote
Bookmarks