Timestamps are in GMT/BST.
[0:02] <Rust> nextgens, yes.. the jolly rogers coockbook is prolly only a tiny bit better than the anarchist one..
[0:09] * MikeW (i=Mike@) Quit ()
[0:10] <MineHaunter> I'm running a test node compiled for win32 with gcj and it seems ok
[0:10] <MineHaunter> it's compiled from the svn jars
[0:14] * OctobersDark (n=October@) has left #freenet
[0:22] * Xosy (n=newbie@) has joined #freenet
[0:23] <Xosy> was checking on freenetproject.org and dont see this simple question in the FAQ
[0:23] <Xosy> i wanted to know what license freenet is released under
[0:23] <Xosy> is it GPU license, so that anyone can modify source and release new version?
[0:23] <MineHaunter> I guess it's GPL
[0:26] <MineHaunter> yes, it's gpl, there is a copy of the license in the freenet folder
[0:31] * Apophis2 (n=Apophis@) has joined #freenet
[0:32] * Xosy (n=newbie@) Quit ()
[0:36] * Tritman (i=Tritman@) Quit (Remote closed the connection)
[0:41] * Tritman (i=Tritman@) has joined #freenet
[0:50] * hjubal (n=hjubal@) Quit ("Leaving")
[0:53] * JasonMog8D (n=foo@) Quit (Read error: 54 (Connection reset by peer))
[0:54] * ronja13 (n=Miranda@) Quit ("Miranda IM! Smaller, Faster, Easier. http://miranda-im.org")
[1:05] * Zothar (n=Zothar@) Quit ("Chatzilla 0.9.74 [Firefox 1.5.0.8/2006102516]")
[1:05] <_ph00> <Rust> "the jolly rogers coockbook is prolly only a tiny bit better than the anarchist one." <== both suck. if you want to blow up stuff, I would suggest some miltary handbook
[1:06] <_ph00> I mean, those "cookbooks" use teaspoons intead of grams FFS!
[1:06] <_ph00> (not even coounting that half the 'recepes' are plain wrong
[1:08] <Rust> that's what i said, i think
[1:17] <_ph00> you didn't, actually. but from your msg one can figure out that you also think they are both low quality
[1:17] <Rust> well, that's what i meant.. i'm just to tired to write sensibly
[1:21] <_ph00> k
[1:22] <_ph00> it's scary...
[1:22] <_ph00> people read that kind of stuff while actually having in mind to experiment with that
[1:23] <_ph00> I would *never* try to mix explosives following a recepe that tells me to "add two teaspoons of this and one tablespoon of that"
[1:24] <_ph00> btw, if you search some torrent sites, you probably will find some real military manuals, including all that sabotaging stuff, emergency weapons etc
[1:24] * Iceman_B (n=Ice@) Quit ("Fate-chan is hotter than Nanoha-chan ^_^")
[1:25] <Rust> oh yes
[1:25] <Rust> that stuff is pretty easy to come by actually, if you look hard enough, and have a clue
[1:25] <Rust> i just never really bothered making a bomb in my backyard..
[1:26] <_ph00> well...
[1:26] <_ph00> you know the saying, hope for the best, prepare for the worst
[1:27] <_ph00> I see a world going every day more towards religious fanatism and fascist govenments
[1:27] <_ph00> so, maybe, knowing how to blow up stuff can be useful in the future (hopefully not)
[1:27] <_ph00> [I never tried to make a bomb in my backyard either, anyways]
[1:29] <Rust> that's why i like the idea of having stuff on freenet, but alas
[1:29] <Rust> anyways, it's frustrating me that my node doesn't connect to new pers
[1:29] <Rust> peers
[1:29] <_ph00> it doesn't?
[1:29] <_ph00> that's weird
[1:30] <_ph00> you mean you can connect to your peers but you can't add any new ones?
[1:30] <Rust> i think most of the new peers are shown as "never connected"
[1:30] <Rust> and that's what they get from me
[1:31] <_ph00> hm
[1:31] <_ph00> and you are sure that they added you, right?
[1:31] <_ph00> what you mean "you think"? are they shown as 'never connected' or not?
[1:32] <Rust> yes they added me
[1:32] <Rust> and i think that most
[1:32] <Rust> might be all, but might not be
[1:32] <Rust> i'm really too tired and busy to notice
[1:40] <Rust> and now one of the never connected turned busy
[1:40] <_ph00> tired busy frustrated and pissed off, not the best condition to figure out what's going wrong
[1:41] <_ph00> but
[1:41] * harry123 (n=egsavage@) has joined #freenet
[1:41] <Rust> and now it turned connected
[1:41] <_ph00> one of the never connected turning busy, mean that at least they found each other
[1:41] <_ph00> see?
[1:41] <_ph00> maybe it just takes time
[1:41] <Rust> htta's what i think now
[1:41] <_ph00> what kind of connection have you got?
[1:41] <Rust> i wonder if this has been happening before
[1:42] <Rust> an 8mb adsl, methinks
[1:42] <_ph00> do you have firewall, router? did you open your freenet port? (etc)
[1:42] <Rust> no.. still haven't gotten around to it
[1:42] <_ph00> 8Mb, down, and how much up? (probably 1)
[1:42] <Rust> and i did think that it might cause this
[1:42] <Rust> (no - didn't open the port, i should)
[1:43] <_ph00> actually, it does work anyways, I am behind a NAT where I can't port forward, and freenet works, but if you open the port it's easier for the node to connect to peers
[1:43] * JasonMog8D (n=foo@) has joined #freenet
[1:43] <_ph00> yes you should
[1:43] <_ph00> you know what port to open?
[1:44] <Rust> i'm sure i can find it on the site
[1:44] <Rust> the problem is that i don't know the router password
[1:47] * Tritman (i=Tritman@) Quit (Remote closed the connection)
[1:48] * Tritman (i=Tritman@) has joined #freenet
[1:50] <_ph00> <Rust> the problem is that i don't know the router password <== did you try the default password?
[1:50] <_ph00> there must be a list...
[1:50] <Rust> yes
[1:50] <Rust> i did
[1:50] <_ph00> ah OK
[1:50] <Rust> even the technician pass
[1:50] <Rust> this is just bizarre
[1:50] <_ph00> and that didn't work...
[1:51] <_ph00> who set that up?
[1:51] <_ph00> the router
[1:51] <Rust> i have no idea
[1:51] <Rust> it's not my house reall
[1:51] <_ph00> ?!?
[1:51] <_ph00> aaahhh...
[1:51] <Rust> what i should do is call the internet provider and yell at them
[1:51] <Rust> but i'm lazy
[1:51] <_ph00> you broke in someone's house and you're using their PC
[1:51] <Rust> lol no
[1:52] <Rust> i'm staying with some people
[1:52] <Rust> for some months
[1:52] <_ph00> huh? how can you yell at the ISP if you don't know the password to your router?
[1:52] <Rust> so they'll tell me
[1:52] <_ph00> I know, it's not yours
[1:52] <_ph00> who? ISP?
[1:52] <Rust> i'm pretty sure they've set up the network here
[1:52] <Rust> yeah
[1:52] <_ph00> your friends could tell you, probably, not ISP
[1:53] <Rust> they're technoidiots
[1:53] <_ph00> hm
[1:53] <_ph00> call them and ask who set up the stuff
[1:53] <Rust> that's what i'll do
[1:53] <Rust> i'm just lazy
[1:53] * JasonMog8D (n=foo@) Quit (Client Quit)
[1:53] <_ph00> that's what I do: set up home computers, networks etc for computerdiots
[1:53] <_ph00> fix pc's etc
[1:54] <_ph00> not much work, not much money, but NO BOSSES
[1:54] <_ph00> yay
[1:54] <Rust> heh
[1:54] <Rust> i do that for free
[1:54] <Rust> i guess you're smarter than i am
[1:54] <Rust> (:
[2:05] <_ph00> you do that for free, and you have to work
[2:05] <_ph00> I do only that
[2:05] <_ph00> like living without the need for a job
[2:05] <_ph00> (you don't get rich that way, only survive... as a poor guy)
[2:06] <Rust> good enough simetimes
[2:06] <_ph00> but that's good enough for me, I theink my freedom is worth more than money (and I don't have a wife who whines about getting "real jobs")
[2:07] <_ph00> I work some 15-20 hors a week max
[2:07] <_ph00> no early awakenings, no slips & ties, no show-up every day... just keep my cell phone on, and wait for calls
[2:07] <_ph00> mon-fri 9-5
[2:08] <_ph00> you call me on a weekend, I tell you to fuck off
[2:08] <_ph00> btw, here's gone another saturday night on IRC
[2:08] <_ph00> [I guess that's one of the signs that you may be a geek]
[2:09] <Rust> yes..
[2:09] <Rust> doing what you do for a living is one too..
[2:11] <_ph00> hm... not really
[2:11] <_ph00> most geeks I know go for "real" jobs
[2:12] <Rust> and having two seperate conversations with the same person on two different channels is also a sign
[2:12] <_ph00> heh
[2:12] <_ph00> that's an almost secure sign of being a geek
[2:14] * phrosty (n=phrosty@) Quit (Read error: 131 (Connection reset by peer))
[2:14] * murdockl (i=murdockl@) Quit (Read error: 110 (Connection timed out))
[2:14] * phrosty (n=phrosty@) has joined #freenet
[2:19] * harry123 (n=egsavage@) has left #freenet
[2:20] * Rust (n=for@) Quit ("Cats are like furry dilettanti, or the reverse?")
[2:22] * auc1 (n=chatzill@) has left #freenet
[2:24] * PhrostByte (n=phrosty@) has joined #freenet
[2:27] * phrosty (n=phrosty@) Quit (Read error: 131 (Connection reset by peer))
[3:07] * Iceman_B (n=Ice@) has joined #freenet
[3:21] * drxvii (n=drxvii@) has joined #freenet
[3:43] * drxvii (n=drxvii@) Quit ()
[3:44] * drxvii (n=drxvii@) has joined #freenet
[3:44] * strautc (n=strautc@) has left #freenet
[5:06] * _ph00 (n=z@) Quit ("Leaving")
[5:08] * drxvii (n=drxvii@) Quit ()
[5:09] * NoJohnDoe (n=chatzill@) has joined #freenet
[5:09] <NoJohnDoe> sup freenet
[5:10] <NoJohnDoe> happens that someone can give me this damn noderef?
[5:10] <NoJohnDoe> any1 here anyway?
[5:13] <fridim> Hi
[5:14] <NoJohnDoe> can you read me?
[5:15] <fridim> I've updrage to stable 996. It's restarted the node, but the Fproxy is still not on. The Hard drive is working. It seems like it reconstructs the database or something. Is it normal ? It started few minutes ago
[5:15] <fridim> NoJohnDoe, #freenet-refs
[5:17] <NoJohnDoe> I actually am trying, fridim - seems that nobody can read me there. Is it why I have not regstered my Nick yet?
[5:25] * NoJohnDoe (n=chatzill@) Quit ("Chatzilla 0.9.75 [Firefox 1.5/2005111116]")
[5:36] * Fu2 (n=eatme@) has joined #freenet
[5:36] * Fu2 (n=eatme@) has left #freenet
[6:33] * JustMe (i=JustMe_@) Quit ()
[6:38] * Q-collective (n=Q-collec@) Quit (Remote closed the connection)
[6:52] * Q-collective (n=Q-collec@) has joined #freenet
[7:48] * TheSeeker (i=Fridlekh@) Quit (Read error: 110 (Connection timed out))
[7:48] * drxvii (n=drxvii@) has joined #freenet
[7:50] * drxvii (n=drxvii@) has left #freenet
[8:09] * _ph00 (n=z@) has joined #freenet
[8:10] * _ph00 (n=z@) Quit (Client Quit)
[8:14] * _ph00 (n=z@) has joined #freenet
[8:23] * zaox (n=Administ@) has joined #freenet
[8:40] * whiterabbit (n=Whiterab@) has joined #freenet
[8:52] <nextgens> fridim> yes, it's normal
[9:10] * sanity (n=ian@) Quit ()
[9:28] * TheSeeker (i=Fridlekh@) has joined #freenet
[9:41] <nextgens> ok, there's a bug on the /queue page
[9:41] <nextgens> it sends FCP identifiers through the form
[9:42] <nextgens> wich works ... but not if things have been HTMLEncoded in the meantime
[9:42] * nextgens wonders if that's a HTMLNode bug or something else
[10:43] * the-moog (n=the-moog@) has joined #freenet
[10:51] * RustY (n=for@) has joined #freenet
[10:55] <RustY> so _ph00, wanna give me some tech advice?
[10:55] <_ph00> *me*
[10:55] <_ph00> ?
[10:55] <RustY> sure
[10:55] <RustY> well, anyone really
[10:55] <RustY> but i see you're here
[10:55] <_ph00> there at least one senior dev on line right noe
[10:56] <RustY> on no, not freenet advice
[10:56] <RustY> general pc advice
[10:56] <_ph00> but of course, I'll try
[10:56] <_ph00> (I'm not such a PC guru...)
[10:56] <RustY> well, i bought this new deskstar 80gb drive, right?
[10:56] <RustY> i plugged it in
[10:56] <_ph00> anyways, what's the problem?
[10:56] <RustY> master on the same cable as the cd
[10:57] <RustY> set the cd to slave
[10:57] <_ph00> secondary master?
[10:57] <RustY> bios recognizes it
[10:57] <RustY> yup
[10:57] <_ph00> is that IDE?
[10:57] <RustY> even win2k sees it
[10:57] <RustY> but it won't show up in my computer
[10:57] <_ph00> huh?
[10:57] <RustY> pretty sure it is
[10:57] <RustY> *scratches head*
[10:57] <RustY> i'll check, one second
[10:57] <_ph00> how can you say that win sees it then?
[10:58] <RustY> i see it in device manager, but i don't see it in my computer
[10:58] <_ph00> no real need, if it connects it's the right sort
[10:58] <_ph00> hm
[10:58] <_ph00> look, I never had such a problem
[10:58] <RustY> have you got any idea tho?
[10:59] <_ph00> but if I had, the first thing I would do, I would boot a LiveCD to see if the HD will work under a different OS
[10:59] <RustY> hmm.. i should try that, yes
[11:00] <RustY> it just puzzles me that it shows up under device manager, but doesn't work
[11:00] <RustY> :\
[11:00] <_ph00> another thing you can try, is uninstalling it i your devixce manager, then reboot,
[11:00] <RustY> i'll try that to
[11:00] <RustY> some more googling might give some results
[11:00] <MineHaunter> RustY: probably your hd is not partitioned and formatted
[11:00] <_ph00> another thing yet, would be plugging it in another PC and see if it works
[11:01] <_ph00> no partition
[11:01] <_ph00> right
[11:01] <RustY> MineHaunter, i just found a post about this drive having this problem, heh
[11:01] <_ph00> I thounght if it was unformatted, windows should see it anyway, and when you click on it it should prompt you to format
[11:02] <_ph00> but if there are no partitions on it, that's different: widows wouldn't show it at all in My Computer
[11:02] <RustY> yes.. i feel silly now
[11:02] <RustY> thanks to both of you, i'll do that now
[11:03] <_ph00> 'this drive having this problem' like it's a know problem of that particular model?
[11:03] <_ph00> if nothing works, we'll try something else
[11:03] <RustY> well, they don't partition it in the factory.. and i've found like 5 posts about people doing it now
[11:03] <_ph00> ah, so 'that's the problem
[11:03] <_ph00> no partition
[11:03] <RustY> yup
[11:04] <RustY> god bless partition magic
[11:04] <_ph00> ok then, so you have a chance to partition it the way you want to
[11:04] <_ph00> for instance, you may want to install a nice win-lin dual boot
[11:04] <_ph00> hehe
[11:04] <_ph00> I'm for free software, but I love partition magic
[11:05] <_ph00> (anyways, you can do the same thing with most LiveCDs)
[11:05] * AnyQ22 (n=admin@) Quit (Read error: 110 (Connection timed out))
[11:05] <_ph00> but it's just a bit less easy than using partition magic (which is way too difficult for a lot of people)
[11:05] <RustY> well, i think i'll use this new one for data to be shared by both the lin and the win, music and such
[11:06] <_ph00> you have partition magioc installed?
[11:06] <RustY> so i guess it'll have to be fat32
[11:06] <RustY> i do
[11:06] <_ph00> I never installed it, I run off the bootable CD only
[11:06] <_ph00> no why
[11:06] <_ph00> why fat32
[11:06] <_ph00> for linux and windows both?
[11:06] <_ph00> you can do that ext2
[11:07] <_ph00> there are ext2 drivers for windows
[11:07] <RustY> hmmmm
[11:07] <RustY> how well would it work?
[11:07] <_ph00> so you can have only one small ntfs partition for windows, and all the rest on ext2
[11:07] <_ph00> I never got a problem with that
[11:07] <RustY> last time i tried to read a linux partition from windows i couldn't get it to do it
[11:08] <RustY> tho i think it was ext3
[11:08] <_ph00> ext3 is basically ext2 plus a journalling system
[11:08] <_ph00> I run ext2
[11:09] <_ph00> probably, the wext2 driver for windows will read ext3 too, buty I'm not sure
[11:09] <_ph00> I do know that it works fine on ext2
[11:09] <RustY> hmm
[11:09] <RustY> thanks
[11:09] <RustY> i shall google it
[11:10] <_ph00> yes
[11:10] <_ph00> google windows ext2 drivers, you should get to a file less than 2 MB in size
[11:11] <_ph00> you run it and it will mke your ext3 partitions visible and writable under windows
[11:11] <_ph00> ext2
[11:11] <RustY> hmm
[11:12] <_ph00> I told yofine for me
[11:12] <_ph00> and with no 4GB max file size or other fat32 lame stuff
[11:13] * Apophis2_ (n=Apophis@) has joined #freenet
[11:13] <_ph00> I diob't know whether read/write is faster on fat32 or on ext2 + windows, tho
[11:13] <_ph00> but I guess it can't be slower than fat32
[11:14] * fridim (n=fridim@) Quit (""Putain ils sont o?? ?" -- le petit poucet")
[11:15] * fridim (n=fridim@) has joined #freenet
[11:16] <RustY> hah
[11:16] <RustY> yea, i guess you're right
[11:20] * Apophis2 (n=Apophis@) Quit (Read error: 145 (Connection timed out))
[11:20] <RustY> heh
[11:20] <RustY> apparently there's even support for reiserfs
[11:28] <_ph00> oh yeah?
[11:28] <_ph00> I didn't even know that
[11:28] <_ph00> but I didn't look
[11:28] <_ph00> (as I run ext2...)
[11:28] <_ph00> I'm the last one doing that, probably
[11:29] <RustY> well, the project has been discontinued, and i can't be bothered to look for another one - so ext3 it is
[11:30] * Apophis2 (n=Apophis@) has joined #freenet
[11:30] <_ph00> ext3?
[11:30] <_ph00> why not ext2
[11:30] <_ph00> that will give you more disk space
[11:30] <_ph00> (but no journalling)
[11:33] <RustY> i can afford to lose a couple gb for journalling..
[11:37] <nextgens> would you move onto an other chan. please ?
[11:38] <RustY> *shuts up*
[11:38] <RustY> sure.
[11:39] <RustY> i do have a freenet problem
[11:39] <RustY> my node won't start.
[11:39] <nextgens> then this channel is adequate for that problem :)
[11:39] <RustY> (:
[11:40] <RustY> i turned it off last night, before shutting down the pc
[11:40] <RustY> started the pc today (figured out the drive problem), and tried to start freenet
[11:40] <nextgens> stick your wrapper.log onto a pastebin
[11:40] <RustY> and i just gets stuck on the restarting node page
[11:40] <RustY> sex
[11:40] <RustY> sec
[11:42] <RustY> http://dark-code.bulix.org/z5r8om-23351
[11:44] <nextgens> INFO | jvm 3 | 2006/11/13 11:31:38 | Node initialization completed.
[11:44] <nextgens> well, according to the log, your node has started up
[11:44] <RustY> i saw that, yes
[11:44] <nextgens> but is outdated
[11:44] <nextgens> latest is 996
[11:44] <RustY> i'll update, but it does'n't connect to anyone
[11:44] <RustY> doesn't
[11:44] <nextgens> are you accessing fproxy from an other computer on the lan ?
[11:45] <RustY> no
[11:45] <nextgens> hmm
[11:45] <nextgens> but do you have access to fproxy ?
[11:45] <nextgens> [11:40] < RustY> | and i just gets stuck on the restarting node page
[11:45] * Apophis2_ (n=Apophis@) Quit (Read error: 110 (Connection timed out))
[11:45] <nextgens> isn't it a browser caching problem ?
[11:46] <nextgens> have you tried ctrl+r or whatever refresh is on your browser ?
[11:46] <RustY> it should redirect me, shouldn't it?
[11:46] <nextgens> it shouldn't
[11:46] <RustY> oh
[11:46] <RustY> anyway
[11:46] <RustY> i did refresh it myself
[11:46] <RustY> and it still doesn't connect to anyone
[11:46] <nextgens> try to update first
[11:46] <RustY> hmm k
[11:47] <nextgens> even though 994 should connect to any node
[11:47] <nextgens> should be able to connect to any I ought to have said
[11:52] <RustY> well, i ran the update script
[11:52] <RustY> but now when i use the "start the node" shortcut it says that it's locked
[11:52] <RustY> and doesn't start it at all
[11:53] <nextgens> isn't it migrating the datastore ?
[11:53] <Caco_Patane> http://againstred.blogspot.com/2006/11/only-paranoid-survive-freenet-project.html
[11:54] * hopeatikari (i=95416dc@) has joined #freenet
[11:54] * Iceman_B^ (n=Ice@) has joined #freenet
[11:54] * Caco_Patane (n=caco@) Quit ("Leaving")
[11:55] * Iceman_B (n=Ice@) Quit (Read error: 104 (Connection reset by peer))
[11:56] <nextgens> it's mostly a cut & paste of the what-is page, isn't it ?
[11:58] <RustY> what do you mean nextgens?
[11:58] <RustY> it closed the update window, and that's it
[11:58] <nextgens> RustY> isn't the node doing many access on your hard drive ?
[11:58] <nextgens> es
[11:58] <nextgens> RustY> what does your wrapper.log say ?
[12:00] <RustY> http://dark-code.bulix.org/p2erqw-23353
[12:01] <nextgens> your node hasn't started at all
[12:01] <nextgens> accoding to timestamps, it has been shut down 10 mins ago
[12:01] <RustY> that's my problem exactly
[12:02] <RustY> it doesn't start
[12:02] <nextgens> ok
[12:02] <nextgens> get a command prompt
[12:02] <nextgens> and issue :
[12:02] <nextgens> net start freenet-darknet-8888
[12:02] <nextgens> or whatever the service name is
[12:03] <RustY> The service database is locked.
[12:03] <RustY> More help is available by typing NET HELPMSG 2180.
[12:03] <RustY> that's my problem
[12:03] <nextgens> hmmmkay
[12:03] <nextgens> the only solution I know to that specific problem is to restart
[12:04] <RustY> that's what i've been planning
[12:04] <RustY> thanks man
[12:04] <RustY> bye
[12:04] * RustY (n=for@) Quit ("Success always occurs in private and failure in full view.")
[12:07] * Apophis2 (n=Apophis@) Quit ("Verlassend")
[12:07] * Apophis2 (n=Apophis@) has joined #freenet
[12:15] * toad_ (n=toad@) has joined #freenet
[12:15] * ChanServ sets mode +o toad_
[12:37] <nextgens> hey toad_
[12:37] <nextgens> toad_> no major issue have been reported on 995 yet
[12:37] <nextgens> :)
[12:48] <_ph00> mostly beacuse ppl upgraded staight to 996 without even running 995 once
[12:51] <nextgens> 996 I meant :)
[12:51] <_ph00> ayways, I don't know what you guys have done (and even if you told me I wouldn't understand) but I noticed that peers connect more easily now. As I already said a bunch of times, I'm behind a NAt where I cant port forward. up to 994, I always had some disconnected peers, often while I knwy that they were up, they had some problems connecting. Well, not any more, I have all my peers connected mosty of the time now
[13:01] * Iceman_B^ (n=Ice@) Quit (Read error: 60 (Operation timed out))
[13:02] * Iceman_B (n=Ice@) has joined #freenet
[13:12] <toad_> nextgens: there's a guy on frost who had a problem with it
[13:13] <toad_> _ph00: interesting, maybe some of the USK changes fixed ARKs?
[13:13] * toad_ has 16 connected peers now, which isn't exceptional
[13:14] <toad_> Sorry guys but this is a really bad one. Files that I could download the same day before changing to this version I am lucky to get 50% of them now.
[13:14] <toad_> Inserts are *SLOW* but I guess that has not changed too much.
[13:14] <toad_> I am guessing the conversion process borked my datastore.
[13:14] <_ph00> the main difference is that you are reachable from the internet, while I'm not, I'm behind NAT with no forwarded ports
[13:14] <nextgens> toad_> I've experienced a really lower memory footprint on both of my nodes
[13:15] <nextgens> toad_> my "main" node fits in less than 50M
[13:15] <toad_> _ph00: fair point; so anyone who i'd need to look up an ARK for would connect to me anyway because they'd send me packets
[13:15] <nextgens> with ~20 jobs on the global queue
[13:15] <nextgens> whereas it used to be almost twice that amount
[13:15] <toad_> nextgens: not the case here, it's showing 100-160M
[13:16] <toad_> goes up to 160 then GCs down to 100
[13:16] <nextgens> # Used Java memory: 52.4 MiB
[13:16] <nextgens> # Allocated Java memory: 78.4 MiB
[13:16] <nextgens> # Maximum Java memory: 127 MiB
[13:16] <toad_> <li>Used Java memory:?141 MiB</li>
[13:16] <toad_> <li>Allocated Java memory:?222 MiB</li>
[13:16] <toad_> <li>Maximum Java memory:?247 MiB</li>
[13:16] <nextgens> it used to allocate 127M
[13:16] <toad_> i have a 256 limit though
[13:16] <toad_> i could reduce it and see what happens
[13:16] <_ph00> I have 2 disconnected peers out of 22 total, which is good, compared to what I used to get. Some peer couldn't connect though they were on line, which seems not to be happening any more.
[13:17] * nextgens has 256M of physical mem ;)
[13:17] <_ph00> but maybe it's just a coincidence
[13:17] <_ph00> (let's see... if it stays thios way, it's not)
[13:17] <toad_> _ph00: tell me if it does
[13:17] <_ph00> k
[13:17] <nextgens> _ph00> the USK fixes have probably helped arks
[13:17] <toad_> nextgens: any ideas about the guy on frost i pasted from above?
[13:17] <nextgens> yep
[13:17] <toad_> i really am not convinced that i messed up the datastore
[13:17] <nextgens> your migration code is not dumbass proff
[13:17] <nextgens> proof
[13:18] * nextgens bets he restarted the node straight away during the migration process
[13:18] <toad_> "Files that I could download the same day before changing to this version I am lucky to get 50% of them now."
[13:18] <toad_> that's the suspicious thing
[13:18] <nextgens> as it was using more cpu than it used to
[13:18] <toad_> nextgens: or the wrapper did? :)
[13:18] <nextgens> maybe :)
[13:18] <nextgens> :$
[13:20] <toad_> I can see that it might have broken the store, but I doubt it
[13:20] <toad_> alternatively it might have broken USKs, but I'm reasonably sure I fixed that
[13:21] <nextgens> it performs really better for me
[13:21] <nextgens> now indexes are loading almost fully
[13:22] <toad_> better than 992?
[13:22] <nextgens> toad_> and what about the number of threads ?
[13:23] <toad_> 994 had some serious USK bugs
[13:23] <nextgens> here they are less threads in use as well
[13:23] <nextgens> ~80 ATM
[13:23] <toad_> nextgens: was running 204 when i pasted the memory stats
[13:23] <toad_> <div class="infobox-content">
[13:23] <toad_> <ul><li>Used Java memory:?28.5 MiB</li>
[13:23] <toad_> <li>Allocated Java memory:?58.0 MiB</li>
[13:23] <toad_> <li>Maximum Java memory:?185 MiB</li>
[13:23] <toad_> <li>Available CPUs:?1</li>
[13:23] <toad_> <li>Running threads:?219</li>
[13:23] <toad_> </ul></div>
[13:23] <toad_> that's with max=192
[13:23] <nextgens> well, it seems that at startup it uses less than 30M
[13:24] <nextgens> wich is good
[13:24] <nextgens> soon freenet is gonna use less ram than firefox :D
[13:26] <nextgens> does anyone know good documentation about bloom filters ? and more generally speaking probabilistic data structures ?
[13:26] * nextgens is studying to prepare google interviews
[13:27] <toad_> sorry ...
[13:27] * RustY (n=for@) has joined #freenet
[13:27] <toad_> i found some on the web when i googled for them though
[13:27] <toad_> maybe wikipedia, there are also some open course materials that i found
[13:27] <nextgens> I've seen the wikipedia pages
[13:27] <toad_> when you have time we need to finish dealing with freenet-ext.jar
[13:27] <nextgens> it's a bit light :/
[13:27] <toad_> well, bloom filters are really cool
[13:28] <toad_> and they're really simple too, once you grok the concept
[13:28] <nextgens> sure, unless you have to handle removal ;)
[13:28] <toad_> :)
[13:28] <nextgens> counting bloom filters are a bit trickyer
[13:28] <nextgens> and wikipedia only says "be carefull about special cases"
[13:28] <toad_> hash the value, divide the hash into N bits, keep a 2^N array of either bits or counters ...
[13:29] <toad_> it's really not hard
[13:29] <nextgens> and efficient :)
[13:29] <toad_> obviously if you want removal, you need counters, and if you have counters, they can still get maxed out
[13:29] * greycat (i=rfc1413@) has joined #freenet
[13:29] <toad_> so you either handle that with a flag (indicating that the bloom filter has problems), or have an external structure (another bloom filter perhaps? :) ) to track the maxed out bits
[13:30] <toad_> you could maybe do it like they do with hashtables
[13:30] <toad_> migrating gradually from a small one to a bigger one at each step
[13:30] <toad_> but as i've never actually implemented a bloom filter, ignore me ;)
[13:30] <nextgens> the problem beeing you loose time while doing it :/
[13:31] <nextgens> either you preallocate too much and loose time on each operation
[13:31] <nextgens> or not enough and you loose time when you need to iterate in order to grow up
[13:31] <nextgens> toad_> aren't we using a bloom filter on the DS ?
[13:31] <toad_> yeah, the best way most of the time is probably just to let a few elements max out, and reconstruct it occasionally :|
[13:31] <toad_> nextgens: nope
[13:31] <nextgens> huh ?
[13:31] <toad_> we probably should
[13:31] <nextgens> why not ?*
[13:31] <nextgens> indeed :)
[13:32] <toad_> well, BDB may use one internally
[13:32] <toad_> i doubt it though, it doesn't know we have a structure where the common case is lookup failure
[13:32] <toad_> if you want to implement one, go ahead :)
[13:32] <nextgens> I really thought we were using one ...
[13:33] <nextgens> weren't we on .5 ?
[13:33] <toad_> no, we planned to use one
[13:33] <toad_> i don't think we ever implemented it
[13:33] <CIA-12> toad * r10907 /trunk/freenet/src/freenet/clients/http/filter/ (CSSTokenizerFilter.java CSSTokenizerFilter.jflex): Another fix from UniquePerson, fixes comment encoding in CSS.
[13:33] <nextgens> ok :/
[13:33] <toad_> well you can check out NativeFSDatastore.java if you like :)
[13:33] <toad_> but it's a monster
[13:33] <nextgens> well, I happend it to my todo
[13:34] <nextgens> but after freenet-ext, the installer, sts ...
[13:34] <toad_> quite
[13:34] <toad_> i suggest you file a bug for it
[13:34] <nextgens> await the implementation in the next decade as you use to say ;)
[13:34] <nextgens> ok, will do
[13:38] <toad_> nextgens: does the current freenet-ext.jar work with x86-32?
[13:38] <toad_> nextgens: should we get people to upgrade to it / make the auto-updater update it?
[13:38] <toad_> nextgens: i mean does FEC work with it?
[13:40] <nextgens> fec works on #7 with x86-32
[13:40] <toad_> nextgens: would it be worth looking at native implementations of Rijndael eventually?
[13:40] <nextgens> it's probably a good idea
[13:40] <toad_> okay, and you've deployed fec #7?
[13:40] <nextgens> hmm
[13:40] <nextgens> what for ?
[13:41] <toad_> so it would be a good idea to fix the updater to update freenet-ext.jar as well, using the same trick we use on windows?
[13:41] * nextgens thinks that java's crypto code isn't that slow
[13:41] <toad_> well we use our own crypto code
[13:41] <toad_> because we don't want to make the user deal with the export crap
[13:41] <nextgens> hmm
[13:41] <toad_> which sadly does still exist
[13:41] <toad_> freenet:CHK@X~sw6twmj6s-dTwirMXtCxENtsQpEA8GajDkAPZyUHg,tBWeD~gg4gi-wKkkeADTh0s~j8a0HCHjLJRNI4HpF2U,AAEA--8?type=text/html
[13:41] <toad_> that's a profile from UniquePerson on fred running frost
[13:41] <nextgens> I suggest you spend some time on the auto-updater/update over mandatory
[13:42] <toad_> there may be significant optimisations possible on USM.waitFor and maybe MessageFilter.match
[13:42] <toad_> the rest is mostly crypto
[13:42] <toad_> nextgens: good idea
[13:42] <nextgens> and maybe before you do, you can up the requirements in NodeStarter
[13:43] <nextgens> there is a constant for freenet-ext.jar
[13:43] <nextgens> it will display a useralert if the user isn't up to date
[13:43] <toad_> hmmm
[13:44] <toad_> ok
[13:44] <toad_> so the first thing is refactoring in NodeUpdater and make freenet-ext.jar auto-update'able
[13:44] <toad_> can do that today
[13:44] <toad_> take a few hours
[13:45] <toad_> then probably it would be a good idea to get into binary blobs
[13:45] * bleearrgh (i=bleargh@) has joined #freenet
[13:45] * Zothar_Work (n=chatzill@) has joined #freenet
[13:45] <toad_> binary blobs -> instant broadcast of revocation -> update over mandatory
[13:45] <nextgens> ok
[13:45] <toad_> or something like that
[13:46] <toad_> there are bugs too of course, and there is the big item of multi-container manifests (and more control over manifests)
[13:46] <toad_> and maybe transparent container support at the same time
[13:46] * nextgens is against giving away more control over manifests
[13:46] <toad_> errr, more control over containers
[13:47] <toad_> what goes in what container is always going to be a matter of heuristics
[13:47] <toad_> we may as well allow clients to specify it
[13:47] <toad_> also, should we exclude the metadata from the container
[13:47] <toad_> ?
[13:48] <nextgens> what for would we?
[13:48] <toad_> there is significant feeling on frost that activelinks forcing you to load the whole container is bad
[13:48] <nextgens> i'm not convinced
[13:48] <nextgens> it helps spreading the data
[13:48] <toad_> also the metadata in the container is unlikely to be compressed more by being in the container, because it's a messy binary format that doesn't look like anything else
[13:48] <nextgens> and once the activelink has loaded, loading the main page is quick
[13:49] <nextgens> and it's good in terms for plausible deniability
[13:49] <toad_> yes, but that means loading the activelink (and therefore the index) takes ages
[13:49] <nextgens> you would have fetched the container and it will be in your store as soon you've seen the active link
[13:49] <toad_> and nothing which is activelinked from a major index site will ever fall out of freenet
[13:49] <nextgens> not if everyones does ... and the network works
[13:50] <nextgens> indeed
[13:50] <nextgens> that's the point :)
[13:50] <toad_> well that was half the point in 0.5, the other half being related to DBR availability (doesn't apply any more)
[13:51] <nextgens> maybe we should introduce some kind of DBRs back
[13:51] <toad_> well my preference is that indexes don't activelink to objectionable sites, but obviously that has a censorship effect and people may be unwilling to drop alinks which aren't really objectionable
[13:51] <toad_> nextgens: why? USKs work fine
[13:52] <nextgens> it was convenient to set up "meeting points"
[13:52] <toad_> nextgens: what about this: we have multiple containers. by default, the activelink and the HTML go into the first container, the images go into the second.
[13:52] <nextgens> well, arguably you can do the same with USKs
[13:52] <toad_> nextgens: meeting points?
[13:53] <toad_> nextgens: compromize solution, so the alink loads relatively fast, and the page loads quickly, but an unpopular site will have its images fall out :)
[13:53] <nextgens> that seems sensible
[13:53] <toad_> it seems a little bit *too* custom-built to drop unvisited porn sites, but it DOES look like a nice compromize to me
[13:53] <nextgens> but from an usability PoV, it's a bit shark
[13:53] <toad_> is it?
[13:54] <nextgens> hmm
[13:54] <toad_> the alternative is that the index loads the entire site - or at least the first 2MB container's full
[13:54] <nextgens> maybe that will help "cleaning" freenet from pedo galleries though
[13:54] <toad_> which means that ENTRY POINT for example will have to load hundreds of megs of sites
[13:54] <nextgens> so it's probably a good idea
[13:54] <toad_> nextgens: you do realise the convo will be posted on frost as evidence that we are plotting to scourge freenet clean of KP?
[13:55] * nextgens is not pro pedo.
[13:55] <toad_> well, i think there are practical reasons for a heuristic of that type - the first container includes the alink, the HTML, and the CSS
[13:55] <nextgens> and I have never hidden that fact
[13:55] <toad_> me neither, but freenet must not be designed for censorship
[13:55] <toad_> at the same time, dropping stuff that isn't accessed is a basic freenet value
[13:56] <nextgens> and I feel comfortable with that, even if anonymous morrons on frost will say I'm the root of Evil
[13:56] <toad_> having a 108x36 icon for a freesite on an index is a good idea; this can be accomplished by inserting it separately as a CHK
[13:56] <toad_> having it as part of the freesite is something we've done since 0.5; in 0.5 it just loaded the metadata initially, but then the whole container
[13:56] <toad_> but it was usually excluded from the container
[13:56] <toad_> to make it load faster
[13:57] * toad_ notes that indexes which want to preload the whole site should be able to do so by including relevant link rel=preload's
[13:57] <nextgens> it sounds like we are attempting to cheat in order to get better results at benchmarks
[13:57] <toad_> in 0.5 we also had to have it to show whether a DBR was retrievable; not a problem since USKs
[13:58] <toad_> nextgens: well maybe that's necessary; most of freenet isn't freesites, it's files posted on frost
[13:58] <nextgens> but that's true that using pre= it's fair
[13:58] <toad_> that was probably the justification people gave
[13:58] * EKB (i=EKB@) has joined #freenet
[13:58] <nextgens> so that we let index maintainers the choice of what they want to do
[13:59] <toad_> ok, i think i can defend the policy i've proposed anyway
[13:59] * nextgens agrees :p
[13:59] <toad_> but you can change it for each site if you don't like it; there's no way we can enforce it
[13:59] <nextgens> that makes a concensus anyway
[13:59] <nextgens> as the one who are missing are always wrong ;)
[13:59] <toad_> of course the index maintainers can enforce it
[14:00] <toad_> they can fetch the alink with a small max-size, and if it doesn't load, then reinsert the image as CHK :)
[14:00] <nextgens> I would say "not our business"
[14:00] <nextgens> of course index maintainers can even not link to a freesite
[14:00] <toad_> ok
[14:00] <toad_> or not activelink to it
[14:00] <toad_> yeah
[14:00] <toad_> EP alinks to porn sites, which personally I object to, but that's just me
[14:00] <toad_> Indicia doesn't, except in the porn section
[14:01] <toad_> (which personally I think is the best compromize)
[14:01] <toad_> for an index site
[14:01] <toad_> but that's again not something we can get involved in
[14:01] <toad_> ok
[14:01] <toad_> another point re containers
[14:01] <toad_> if a site is small, it should probably all go in one container
[14:01] <toad_> how small?
[14:02] <nextgens> fits in 32K ? :p
[14:02] * toad_ notes that this will cause certain sites he objects to to go in a single container and hence to be auto-loaded by their activelink; that's just life :(
[14:02] <toad_> nextgens: well, it can't be one key in any case
[14:02] <nextgens> how big is an average png using the standard activelink format ?
[14:02] <toad_> nextgens: it can be one key of metadata plus one key of data
[14:03] <toad_> blocks in 0.7 are either data or metadata, not both
[14:03] <nextgens> toad_> but that's also good for plausible deniability
[14:03] <toad_> nextgens: what is?
[14:03] <nextgens> "I've got it in my store cause I've seen the activelink"
[14:03] <toad_> yeah
[14:04] * toad_ should check that we always check the store for every block before starting to fetch any blocks from non-store
[14:04] <toad_> otherwise the fact that you have ALL the blocks could give you away
[14:04] <nextgens> -rw------- 1 nextgens nextgens 129K 2006-08-19 18:10 src/freenet/sites/flog/logo.png
[14:04] <nextgens> hmm
[14:04] <toad_> yikes, that's a big PNG
[14:04] <nextgens> -rw------- 1 nextgens nextgens 10K 2006-09-08 23:44 src/freenet/sites/flog/activelink.png
[14:04] <toad_> which one is it?
[14:04] <toad_> show me it on the web
[14:04] <nextgens> not the good one
[14:04] <nextgens> the activelink is 10K big
[14:05] <nextgens> anyway, it can stick into one block
[14:05] <toad_> my activelink is 4557 bytes
[14:05] <nextgens> meaning that 2 blocks is probably enough
[14:05] <toad_> well we're talking about the whole site here anyway
[14:05] <nextgens> your activelink is WAY to huge
[14:05] <toad_> http://192.168.1.7:8888/USK@J585KtAJ7UN2~4i17hf7C9XbufMnticJeUDYLcB0dvo,lxZhX2snsExxemocIlI~ZJRFVdVLBLIFZhqV3yswR9U,AQABAAE/toad/10/activelink.png
[14:05] <nextgens> it's f*cking up my flog's design ;)
[14:05] <toad_> it's square, but it's small! :)
[14:06] <toad_> so compress it :)
[14:06] * nextgens will set the size on next edition
[14:06] <toad_> or use one of the other frogs i've quoted on my site
[14:06] <toad_> e.g. the newsbyte one with me dribblng :|
[14:06] <nextgens> it will look awfull but will work
[14:06] * EKB (i=EKB@) has left #freenet
[14:06] <toad_> or one of the photos
[14:06] <toad_> okay
[14:06] <toad_> so how big can a site be to be included in one container?
[14:07] <toad_> it will have one block of metadata in any case
[14:07] <nextgens> anway, I suggest you improve the updater for now :)
[14:07] <toad_> 1+1 is one obvious, defensible answer
[14:07] <nextgens> we need to give time to thinking
[14:07] * toad_ thinks maybe 1+8 would be reasonable
[14:07] <toad_> well we can always change it later
[14:07] <nextgens> in order to produce a good solution
[14:07] <nextgens> but maybe that ought to take place on @devl
[14:07] <toad_> well maybe posting to the list is a good idea
[14:07] <nextgens> otherwise bback is gonna complain in six months ;)
[14:08] * toad_ will post to devl
[14:08] * toad_ hopes bback is on devl :)
[14:14] * toad_ will crosspost to frost
[14:20] <toad_> done
[14:23] * hopeatikari (i=95416dc@) Quit ("l")
[14:23] * nextgens doesn't think you should spend to much time there
[14:24] <nextgens> most of the thread end up in trolls
[14:24] <nextgens> ie: without any new valid point to be raised
[14:24] <nextgens> sometimes there is no argumentation at all
[14:24] <nextgens> the perfect troll
[14:27] <toad_> a lot of users don't go anywhere near the mailing lists
[14:28] <toad_> i get many bug reports from frost
[14:28] <toad_> i have to bridge the anonymous and visible communities...
[14:29] <toad_> nextgens: do you think the database is fast enough to use it for queued requests? if not for all queued requests, then for some of them?
[14:29] <nextgens> I dunno
[14:29] <nextgens> I haven't done any test with it
[14:29] <toad_> there are ways we could optimise it; we could have a stack of the next 100 requests to run for example, pulled ahead of time from the database, if latency was a problem
[14:30] <toad_> and fproxy requests etc stay in RAM
[14:30] <nextgens> hmm
[14:30] <nextgens> I'm not sure that fproxy requests are the worst
[14:30] <toad_> well, the point is, we could eliminate most of the memory usage of big queued requests
[14:31] <nextgens> that's like bloom filtering DS access ...
[14:31] <nextgens> it would be optimization
[14:31] <toad_> well, yeah
[14:31] <nextgens> probably usefull, but the hard part is to assign priorities
[14:31] <toad_> but it might significantly reduce memory requirements
[14:31] <toad_> and more importantly make them predictable
[14:31] <nextgens> the question is more whom are we targetting ?
[14:32] <toad_> no matter how many requests you have global-queued, your node won't need more than 128M
[14:32] <nextgens> to whom that feature would be usefull ?
[14:32] <nextgens> people inserting huge things
[14:32] <toad_> or requesting huge things
[14:33] <nextgens> true
[14:33] <toad_> or requesting lots of things
[14:33] * OctobersDark (n=October@) has joined #freenet
[14:33] <nextgens> I suggest you fill in a bug in the BTS :)
[14:33] <toad_> i don't see what the problem is with optimising freenet for filesharing; ian has decreed that filesharing isn't necessarily illegal files, and we had a student to do Thaw for our SoC
[14:33] <nextgens> and schedule it right after update over mandatory if load-balancing hasn't been sorted out in the meantime
[14:34] <toad_> just as long as it's optimised for everything else (fproxy) too :)
[14:34] <nextgens> toad_> don't shout it too loud, it's illegal in France you know :)
[14:34] <toad_> :|
[14:34] <toad_> nextgens: sharing legal files is legal in france
[14:35] <toad_> tools for sharing legal files that mainly get used for sharing illegal ones may be illegal
[14:35] <nextgens> not if using tools that might be used to perform copyright infringment
[14:35] <nextgens> ie: any tool
[14:35] <nextgens> including internet
[14:35] <toad_> tools which are manifestly intended to be used for copyright infringement was the wording iirc
[14:35] <toad_> so the internet is okay, because lots of legitimate business happens there too
[14:36] <nextgens> I'm not good enough in english to phrase it correctly :)
[14:36] <nextgens> and anyway ... it's the same in english all depends on how you define "manifestly intended"
[14:36] <toad_> hmmm
[14:37] <toad_> well that's what i heard
[14:37] <nextgens> and as it hasn't been defined ;)
[14:37] <toad_> it could mean anything, yeah
[14:37] <toad_> vague criminal offences are wonderful things
[14:37] <toad_> (illegal under human rights legislation probably, but wonderful things)
[14:38] <nextgens> convenient tools I would say
[14:38] * toad_ looking forward to IPRED2 being struck down on the grounds that patent infringement is far too vague and woolly to be part of criminal law
[14:38] <toad_> nextgens: indeed
[14:38] <toad_> judges will take the legislation and put precision into it, eventually
[14:39] <toad_> but it takes tme
[14:39] <toad_> time
[14:39] <toad_> (/me notes that IPRED2 hasn't passed yet, and may pass without the patent bit ... though Lord Sainsbury just retired, after telling me that the UK would oppose that ... )
[14:39] * timmy2chk (n=Vincent@) has joined #freenet
[14:40] <toad_> (I wonder if that means that his successor will take a harder line?)
[14:40] * toad_ sees ludicrously implausible conspiracy theories here...
[14:41] * timmy2chk (n=Vincent@) Quit (Client Quit)
[14:41] <toad_> bbiab
[14:41] * nextgens goes back to real work
[14:42] * timmy2chk (n=Violent@) has joined #freenet
[15:07] * _ph00 (n=z@) Quit ("Leaving")
[15:10] * mazzanet_ (n=mazzanet@) has joined #freenet
[15:25] * mazzanet (n=mazzanet@) Quit (Read error: 110 (Connection timed out))
[15:30] * peacenik (n=chatzill@) has joined #freenet
[15:34] * zaox (n=Administ@) Quit (Read error: 131 (Connection reset by peer))
[15:35] * _ph00 (n=z@) has joined #freenet
[15:35] * peacenik (n=chatzill@) Quit ("Chatzilla 0.9.76 [Firefox 1.5.0.8/2006102516]")
[15:36] * SecretPolice (n=anon@) has joined #freenet
[15:37] <_ph00> help!
[15:37] <_ph00> I'm trying to make a bunch of one-click buttons for freenet
[15:38] <_ph00> one for the node itself, one for thaw and one for frost
[15:38] <_ph00> no problem with first 2
[15:38] <_ph00> but I can't figure out what command should it run
[15:39] <anonymouse> see frost.bat (or .sh or whatever for your platform)
[15:39] <anonymouse> doesn't that contain the info you want?
[15:39] <_ph00> I tried sh /path/to/frost/frost.sh and it did start, but it tried to run in the wrong directory, probablt the freenet dir
[15:39] <anonymouse> you could cd to the right dir first
[15:40] <_ph00> I tried that, but I get "unknown command cd"
[15:40] <_ph00> when I do that on a terminal, I run one line with cd /path/to/frost && ./frost.sh &
[15:41] <_ph00> and then enter and exit
[15:41] <anonymouse> uhm
[15:41] <anonymouse> is the first line of your script a bang path to some sh?
[15:41] <_ph00> but when I tried to do the same using a panel button, I got that weird 'unknown command' output
[15:41] <anonymouse> IIRC, cd is just a built-in shell command
[15:42] <anonymouse> if your .sh script says #!/bin/sh or something at the beginning, then good
[15:42] <anonymouse> otherwise you may wish to add that
[15:44] <_ph00> I'm sorry, where were we (I had someone here who wouldn't stop talking)
[15:44] <anonymouse> you know what's great for that?
[15:44] <_ph00> where should I add that #! thing?
[15:44] <anonymouse> earplugs
[15:44] <_ph00> heh
[15:45] <anonymouse> well most easily, put it into the frost.sh
[15:45] <anonymouse> or see if it's there already
[15:45] <_ph00> I'd need some invisible ones
[15:45] <_ph00> hm
[15:45] <_ph00> nice
[15:45] * _ph00 checks
[15:46] <_ph00> no
[15:47] <_ph00> all frost.sh does is java -jar Thaw.jar
[15:47] <_ph00> plus a coupler of symbols...
[15:47] <_ph00> oh
[15:47] <_ph00> that one was there too
[15:47] <_ph00> #!/bin/sh
[15:47] <_ph00> and then java -jar etc
[15:47] <anonymouse> ok well
[15:48] <anonymouse> between the bang line and the java line, do your "cd" in there
[15:48] <_ph00> and "$@" after that, on the same line
[15:48] <_ph00> cd /path/to/frost ?
[15:48] <_ph00> hm kay
[15:48] <anonymouse> yeps
[15:49] <_ph00> then I could run directly from the panel button and the script will cd for me
[15:49] <_ph00> I get it
[15:49] <_ph00> (well, that wasn't so difficult....)
[15:49] <anonymouse> good
[15:50] <_ph00> do I need the last / after the path's last dir name?
[15:50] <anonymouse> i do not think it matters
[15:50] <_ph00> k
[15:51] <_ph00> OK, it's working. thx
[15:55] * railk (n=railk@) has joined #freenet
[15:59] <anonymouse> welcome
[15:59] * _ph00 (n=z@) Quit ("Leaving")
[16:05] * RustY (n=for@) Quit ()
[16:24] * freenut (n=harvardh@) has joined #freenet
[16:31] * sanity (n=ian@) has joined #freenet
[16:31] * ChanServ sets mode +o sanity
[16:40] <freenut> how do i run frost?cant get here.
[16:44] * SecretPolice (n=anon@) Quit ("Chatzilla 0.9.75 [Firefox 2.0/2006101023]")
[16:53] * railk (n=railk@) Quit ("Cya, wouldn't want ta be ya!")
[16:54] * _ph00 (n=z@) has joined #freenet
[16:57] * vfervers_ (n=vfervers@) has joined #freenet
[17:05] * richard223 (n=joe@) has joined #freenet
[17:06] <richard223> heh
[17:06] * Urs_ShPo (n=gaim@) has joined #freenet
[17:06] <richard223> 100 packets per second
[17:06] <richard223> my home firewall router can't handle that :(
[17:07] * ^wAc^ (i=none@) Quit ("?SysCurSion?")
[17:07] * MikeW (n=em@) has joined #freenet
[17:09] * maelztrom (i=nobody@) Quit (Remote closed the connection)
[17:18] * Wotan02 (n=Miranda@) has joined #freenet
[17:24] * sanity (n=ian@) Quit ()
[17:26] * freenut (n=harvardh@) has left #freenet
[17:31] * stpeter (n=inthelob@) has joined #freenet
[17:31] <stpeter> hi all
[17:32] <stpeter> can freenet run on jvm 1.4?
[17:34] <anonymouse> i don't even know the answer, but i know once i updated to 1.5_09 (from _06) it got a lot faster
[17:36] * sich (n=sich@) has joined #freenet
[17:43] * digger3 (n=digger3@) has joined #freenet
[17:43] <digger3> How long does it normally take for freenet to download a new version? (996)
[17:44] <timmy2chk> for me it took several hours
[17:44] <digger3> can I see the progress anywhere?
[17:44] <timmy2chk> dont know
[17:44] <timmy2chk> i do not think so
[17:46] <digger3> ok
[17:47] * digger3 (n=digger3@) has left #freenet
[17:48] * Urs_ShPo (n=gaim@) Quit ("Leaving.")
[17:50] <sich> you can use ./update.sh in your freenet directory
[17:58] <richard223> hmm... still getting nodeAveragePingTime too high
[17:58] <richard223> # Output Rate: 3.93 KiBps
[17:58] <richard223> # Input Rate: 2.39 KiBps
[17:59] <richard223> something is seriously broken at my site
[17:59] <richard223> nodeAveragePingTime: 6391ms
[18:00] <richard223> any ideas?
[18:09] * sanity (n=ian@) has joined #freenet
[18:09] * ChanServ sets mode +o sanity
[18:09] * NoJohnDoe (n=chatzill@) has joined #freenet
[18:14] * NoJohnDoe (n=chatzill@) Quit ("Chatzilla 0.9.75 [Firefox 1.5/2005111116]")
[18:18] * PhrostByte (n=phrosty@) Quit ("baseball is wrong: man with four balls cannot walk.")
[18:21] <TheSeeker> how many connections do you have richard?
[18:21] * Wotan02 (n=Miranda@) Quit ("Miranda IM! Smaller, Faster, Easier. http://miranda-im.org")
[18:26] <stpeter> the right main class is freenet.node.Node or freenet.node.NodeStarter?
[18:26] <nextgens> NodeStarter
[18:27] <richard223> 21/1/0/0/16 connections
[18:27] <richard223> but my bandwidth isn't saturated
[18:27] <nextgens> toad_> any progress on extending the updater to fetch freenet-ext yet ?
[18:27] <richard223> I have 24Kbps up
[18:27] <TheSeeker> I've never heard of anyone's bandwidth being saturated... no matter how many connections they have :P
[18:28] <richard223> I can saturate it with eMule or BitTorrent,
[18:28] <TheSeeker> load ballancing just doesn't work right now ...
[18:28] <TheSeeker> freenet isn't bittorrent or freemule :P
[18:28] <richard223> 0.5 used to saturate my output bandwidth
[18:29] <richard223> ping went from 50mS to 2000mS
[18:29] <TheSeeker> 0.5 had better load ballancing after being tweaked for years.
[18:29] <richard223> yeah, in 2004 when I first used freenet, it would rape my connection
[18:30] <nextgens> TheSeeker> more b/w usage doesn't mean better load-balancing
[18:30] <richard223> but I found out that the problem is not the KB/s but the packets per second
[18:30] <nextgens> TheSeeker> any progress with your gcj experiments ?
[18:30] <richard223> and freenet generates about 100 packets per second on my node
[18:30] <nextgens> TheSeeker> the new /contrib should be far easier to build now
[18:31] <richard223> that's what is giving me trouble. my firewall/router can't handle that many pps
[18:31] <nextgens> richard223> according how many peers you have that's not much
[18:31] <TheSeeker> nextgens: I haven't done any gcj experiments...
[18:31] * nextgens is confused
[18:31] <nextgens> I meant TheBishop_
[18:31] <TheSeeker> ah :)
[18:31] * timmy2chk (n=Violent@) Quit ("what remains of my violence")
[18:32] <nextgens> toad_> maybe we should have an option to let the user change the "packet coallescing delay"
[18:32] <richard223> if my brother on the other computer starts Azureus, my peers start reporting 60000mS delay
[18:32] <nextgens> toad_> so that people like richard223 could lower the amount of pps
[18:32] <sanity> current load balancing is definitely under-using available bandwidth
[18:32] <sanity> we are waiting on mrogers to implement his new token-based load balancing
[18:32] <richard223> I checked his config, he's not doing more than 5kb/s in either direction
[18:32] <TheSeeker> richard223: that's probably because your brother is using an applicatino that's using 100% of your upload bandwidth, and your router has no QOS
[18:33] <richard223> even if I ping something, like google, the ping value is OK. but freenet reports astronomical values
[18:33] <nextgens> maybe your isp is doing qos on udp flows
[18:33] <TheSeeker> "bittorrent detected... THROTTLE"
[18:33] <TheSeeker> ?
[18:33] <nextgens> nope
[18:34] <nextgens> most do to encourage people using their voip solutions
[18:34] <richard223> now that he closed azureus, my latencies are OK again, none of my peers backs me off
[18:34] <nextgens> they "whitelist" game servers in order to please gamers
[18:35] <richard223> haha no nextgens, not my ISP. it's part of the telco (so they don't like voip), and my latency is 50mS to the first router, and 200mS or more to anything else
[18:35] <nextgens> it doesn't mean anything
[18:35] <nextgens> :)
[18:36] <richard223> it's a monopoly
[18:36] <nextgens> most isps won't throwtle icmp packets
[18:36] <richard223> they don't give a shit anyway
[18:36] <nextgens> as "bad ping values" is bad for PR
[18:36] <nextgens> and noone is tunneling over icmp yet
[18:36] <richard223> my isp is famous for their bad ping values
[18:36] <anonymouse> that we know of =)
[18:36] <nextgens> freenet will ... soon ^^-^^ ;)
[18:36] <anonymouse> it's just an arms race
[18:37] <anonymouse> then they'll come up with some way to detect likely icmp tunneling and throttle it
[18:37] * nextgens gtg, bbiab
[18:37] <nextgens> sure, but it's tricky to do it well
[18:37] <nextgens> and most won't as each time you queue things, you increase the latency
[18:37] <richard223> AFAIK, my ISP doesn't throttle anything,
[18:38] <richard223> I get excellent values with BitTorrent and eMule
[18:38] <richard223> the problem is the lack of investment from them. they have about 600.000 adsl customers,
[18:38] <richard223> with the same equipment from 5 years ago
[18:39] <richard223> again, they are the monopoly so they don't give a shit
[18:40] <richard223> anyway I think the problem is my router because it gets really hot when I run something heavy
[18:40] <sanity> <flamebait>is there any point in gcj any more?</flamebait>
[18:40] <sanity> or kaffe for that matter?
[18:40] <toad_> hi
[18:40] <toad_> sanity: kaffe, no, once sun produces source code
[18:40] <toad_> sanity: gcj, definitely
[18:41] <sanity> i guess it is still useful to have compiled binaries
[18:41] <toad_> sanity: the ability to produce binaries is useful
[18:41] <toad_> right
[18:41] <toad_> also the ability to debug JVM bugs
[18:41] <sanity> but still, there is much less motivation for gcj now - if not in theory, definitely in practice
[18:41] <toad_> I'd like to be able to use generics etc in our code
[18:41] <sanity> toad_: any reason not to now?
[18:42] <stpeter> freenet compiles with gcj on win :)
[18:42] <toad_> sanity: maybe we should wait until sun delivers on its promise?
[18:42] <stpeter> and also seems to run fine
[18:42] <toad_> stpeter: really?
[18:42] <stpeter> yes
[18:42] <toad_> stpeter: i thought it had problems...? 0.5 or 0.7?
[18:42] <sanity> generics are nice, but the syntax can get clunky fast if you have lots of nested generic types
[18:42] <toad_> used correctly they can save a lot of space
[18:43] <toad_> sanity: nested?
[18:43] <sanity> toad: they have deilivered : http://download.java.net/.a2af6706b541c76a90f3597ee8e6a6eb/openjdk/jdk7/
[18:43] <toad_> for(object o : collection) is pretty cool too
[18:43] <sanity> toad: at least for jdk7
[18:43] <sanity> toad_: that is cool
[18:43] * sanity searches for an example of clunkiness..
[18:43] <stpeter> toad_: u need to remove the sleepycat/persist and je/latch/java5 classes from the ext jar
[18:43] <toad_> WOAH
[18:43] <toad_> that was fast
[18:44] <toad_> Java HotSpot? Virtual Machine source under GNU General Public License version 2.
[18:44] * toad_ wonders if it's "GPL 2" or "GPL 2 or later" ... looks like just GPL2
[18:44] <stpeter> toad_: and use some not already committed classpath paches
[18:44] <toad_> stpeter: cool, that's using mainline GCJ too?
[18:44] <stpeter> toad_: for example file.lock/unlock
[18:44] <sanity> here is a nice line of generics code:
[18:44] <sanity> EventSorter.Tuple<EventProcessorFactory, HashMap<ProfileInstance, EventProcessor>> t = Main.es.profileToIEPMMap
[18:44] <sanity> .get(new EventProfile(EventProfile.MEDIA_ID));
[18:45] <stpeter> toad_: with gcj 4.2
[18:45] <toad_> sanity: well, if GCJ has in fact caught up with us ...
[18:45] <stpeter> toad_: waht u mean with mainline?
[18:45] <toad_> stpeter: with a _released_ GCJ? woah
[18:45] <sanity> see how verbose it gets even with just nesting one parameterized type within another?
[18:45] <sanity> wait, it compiles with gcj, but does it run?
[18:45] <toad_> sanity: yeah, but if you're gonna do *that* you deserve it ...
[18:45] <stpeter> toad_: yes , seems to :)
[18:46] <stpeter> i'm using nativej.mtsystems.ch
[18:46] <sanity> toad: nothing wrong with nesting types
[18:46] <nextgens> toad_> sanity> you forgot the number of supported platforms
[18:46] <toad_> sanity: of course, but how would you simplify the above syntax?
[18:46] <stpeter> sanity: yes , seems to :)
[18:46] <nextgens> sun's jvm runs only on x86 and sparc ;)
[18:46] <sanity> toad_: the above syntax? i don't think you can
[18:46] <toad_> sanity: well, if freenet works with GCJ _now_ then i'm reluctant to break compatibility
[18:46] <sanity> toad_: within java's constraints
[18:46] <nextgens> I thought sanity would know with his ppc :p
[18:46] <sanity> toad_: ml's approach is better
[18:47] <sanity> nextgens: i have an intel mac :-)
[18:47] <nextgens> that's cheating :p
[18:47] <toad_> GCJX still isn't in mainline GCJ...
[18:48] <toad_> sanity: with the external TCPA chip :(
[18:48] <nextgens> and they are busy working on ecj atm
[18:48] <sanity> toad_: back to important stuff: any new developments with mrogers load balancing?
[18:48] <toad_> nextgens: well, is it realistic to think that GCJ will support generics soon?
[18:48] <sanity> toad_: there is a TCPA chip?
[18:48] <nextgens> dunno
[18:48] <toad_> sanity: yes
[18:48] <sanity> toad_: that sucks
[18:48] <nextgens> the right question is will classpath support them
[18:48] <sanity> toad_: still, to the best of my knowledge, i don't think apple use it
[18:48] <toad_> sanity: read the label before buying! they
