Timestamps are in GMT/BST.
[0:05] * TheSeeker (~Fridlekh@cpe-65-28-158-211.socal.rr.com) Quit (Read error: 60 (Operation timed out))
[0:35] * jayo (~jay@ool-18bf6dac.dyn.optonline.net) has joined #freenet
[0:37] * jayo is now known as jay
[0:54] <lostlogic_> 67 backed off
[0:54] <lostlogic_> all my routes are backed off
[0:54] * lostlogic_ is now known as lostlogic
[0:55] <salahx> heh
[0:57] <salahx> sounds liek nusines as usua aroudn here :)
[0:58] <lostlogic> :-(
[0:58] <lostlogic> freenet is drinking all my bandwidth and all I get out of it is RNFs
[0:58] <lostlogic> that's not nice.
[0:58] <lostlogic> *pouts*
[0:58] <salahx> what version are you using ?
[0:59] <lostlogic> 5087
[0:59] <salahx> well appearey the "Stable" etowrk needs to "Heal" from 508[5|6]
[1:00] <lostlogic> ahh, which would splain 5087 nodes that therefore actually work being overloaded for now?
[1:00] * lostlogic patients up for now.
[1:00] <salahx> Binggo
[1:01] <lostlogic> that is good explanation, I was able to download some flogs, sites, and splitfiles earlier before the network knew anything about me :-P
[1:01] <lostlogic> anywho, time for the sleeping. laters :)
[1:01] <salahx> I susepct they'll be 5088 soon witha LAStKnownGood of 5087 :)
[1:01] <salahx> ok!
[1:01] <lostlogic> yah
[1:06] * jay (~jay@ool-18bf6dac.dyn.optonline.net) Quit (".")
[1:21] * KenMan (~chaziller@pcp403292pcs.mntcrm01.md.comcast.net) has joined #freenet
[1:38] * Ash-Fox (Hal-9000@aae32.neoplus.adsl.tpnet.pl) Quit (Client Quit)
[1:41] * Ash-Fox (Hal-9000@aae32.neoplus.adsl.tpnet.pl) has joined #FreeNET
[1:43] * TheSeeker (~Fridlekh@cpe-65-28-158-211.socal.rr.com) has joined #Freenet
[1:53] <KenMan> word to the wise... if you ever should leverage a tazer against a computer, make sure you are *not* grounded first. Especially if you aim for the PSU :o
[1:53] <salahx> hehe
[1:54] <KenMan> anyway, i made some 'shocking' changes to my machine, perhaps it will operate better now...
[1:55] <KenMan> i found that a NIC was sharing an interrupt with the chipset or built-in sound . Also, I caught missing clock interrupts during one of those hiccups... and upgraded from 2.6.2 to 2.6.7
[1:56] <KenMan> maybe I hit something important. Time will tell ...
[2:01] * TheSeeker (~Fridlekh@cpe-65-28-158-211.socal.rr.com) Quit (Read error: 60 (Operation timed out))
[2:38] * tav|offline (~tav@espians.com) Quit (Read error: 104 (Connection reset by peer))
[2:39] * jabawok_w (~jabawok_w@203-59-60-38.perm.iinet.net.au) has joined #freenet
[2:39] * xena (~xena@espians.com) Quit (Client Quit)
[3:30] <iip_i2p> <Newsbyte> heal, my ass
[3:31] <iip_i2p> <Newsbyte> and what's a taer?
[3:31] <iip_i2p> <Newsbyte> and what's a tazer?
[3:32] * tyesinclair (~tyesincla@170.252.64.1) has joined #freenet
[3:32] * tyesinclair (~tyesincla@170.252.64.1) has left #freenet
[3:33] <iip_i2p> <Newsbyte> the closest I could find was; taser: A trademark used for a high-voltage stun gun.
[3:35] <iip_i2p> <Newsbyte> toad, you there? After a night running, I now have a whopping nine nodes/connections
[3:37] <salahx> Connections open (Inbound/Outbound/Limit) 74 (55/19/200)
[3:37] <salahx> Transfers active (Transmitting/Receiving) 16 (5/11)
[3:37] <salahx> Data waiting to be transferred None
[3:37] <salahx> Total amount of data transferred 546 MiB
[3:38] <salahx> Unstable 60172
[3:38] <salahx> I dunno why uinstable is working but stable isn't
[3:40] <salahx> hoy many non-5087 nodes do you have in your RT ?
[3:44] * TheSeeker (~Fridlekh@cpe-65-28-158-211.socal.rr.com) has joined #Freenet
[3:47] * Ash-Fox (Hal-9000@aae32.neoplus.adsl.tpnet.pl) Quit (Client Quit)
[3:49] * salahx (salahx@sc1-24.217.174.147.charter-stl.com) Quit ()
[4:06] * TheSeeker (~Fridlekh@cpe-65-28-158-211.socal.rr.com) Quit (Read error: 110 (Connection timed out))
[4:18] * oierw (~mathew@cpe-69-75-126-12.hawaii.rr.com) Quit ("hmm.")
[4:26] <iip_i2p> <Newsbyte> salah, are you asking me?
[4:37] * Lemoni (~Lemoni@munyoo.cs.mu.OZ.AU) has joined #freenet
[4:46] * kers (~kers@86.ppp144.rsd.worldonline.se) has joined #freenet
[4:46] * kers_ (~kers@86.ppp144.rsd.worldonline.se) has joined #freenet
[4:46] * kers (~kers@86.ppp144.rsd.worldonline.se) has left #freenet
[4:46] * kers_ is now known as kers
[4:47] * kers (~kers@86.ppp144.rsd.worldonline.se) Quit (Client Quit)
[4:47] * kers (~kers@86.ppp144.rsd.worldonline.se) has joined #freenet
[4:54] * oierw (mathew@cpe-69-75-126-12.hawaii.rr.com) has joined #freenet
[4:58] * duana (~duana@adsl-4-132.VIC.dft.com.au) has joined #freenet
[5:05] <duana> ok
[5:07] <duana> are there some decent specs of the FNP around anyone knows about? found one in the CVS but it was kinda incomplete and hard to u'stand
[5:09] <oierw> the only real fnp spec is the code
[5:09] <oierw> it's been changing quite a bit recently to get certain things working
[5:10] <duana> yeah i was afraid of that... thanks
[5:11] <oierw> it's not all *that* convuluted
[5:14] <duana> do u know if they're planning to make FCP encrypted?
[5:17] <duana> we want to make an applet that can insert messages onto freenet that can be used on machines that don't run nodes
[5:17] <oierw> why would they want to do that?
[5:18] <oierw> i mean, if you already are running rreenet on a trusted computer, you can just use ssh prot forwarding or something like that. it doesn't show up as freenet traffic and it's just as secure
[5:18] <oierw> also running an applet for fred would be... difficult
[5:18] <oierw> fcp itself isn't that heavy, but just the applet part is not fun
[5:19] <oierw> unless you run fred on the same computer as the webserver...
[5:19] <iip_i2p> <wart> Hiya. After upgrading to 5087 I'm getting OOM errors during seeding as my node starts up. Any advice?
[5:19] <duana> oierw: get back to u - dinner is ready :)
[5:20] <oierw> wart: there've been a lot of report of 5087 not playing nice
[5:20] <oierw> or maybe that was 5086
[5:20] <iip_i2p> <wart> both, as far as I can tell, although I few problems with 5086.
[5:24] <oierw> a merge from unstable somehow got a bunch of nodes with a crypto problem, they think
[5:25] <oierw> and they were going to blame the nbi stuff, but it wasn't even in stable eyt
[5:25] <oierw> so... i dunno
[5:25] <jabawok_w> anyone else got some early 5087 reports?
[5:25] <iip_i2p> <wart> Sounds like there are problems upon problems
[5:25] <jabawok_w> personally i found it uses slightly more mem - seems to be running ok though..
[5:26] <iip_i2p> <wart> When I can persuade it to start, I'll let you know...
[5:26] <jabawok_w> lol
[5:26] <jabawok_w> i did find freenet a little slow - maybe because so many nodes have been knocked offline by the startup prob
[5:26] <jabawok_w> (with 5087)
[5:27] <iip_i2p> <wart> So I'm not the only one experiencing this problem?
[5:27] <jabawok_w> still haven't been able to load YoYo completely - sits for about 10mins before giving up retrieving all the pictures
[5:29] <Fisu_> The mainpage pics in YoYo haven't changed in ages, they should be all over freenet
[5:29] <iip_i2p> <wart> Damn, 5085 was working so well, too.
[5:29] <jabawok_w> yep
[5:30] <duana> shdnt the pics be on your localnode and therefore come up straight away? i dont u'stand why they dont
[5:30] <jabawok_w> thats what i would have thought... got a 15G datastore here... only 70% full too
[5:31] <iip_i2p> <wart> You can adjust the datastore size simply by changing the storeSize value and rstarting the node, right?
[5:31] <jabawok_w> can anybody give me a basic explanation of "restarted" "rejected" and "backed off" on the RNF page?
[5:32] <jabawok_w> so i believe
[5:32] <jabawok_w> actually I understand "rejected" that makes sense.. what about the other two ^^
[5:38] * srob99 (~srob99@203-59-84-243.dyn.iinet.net.au) has joined #freenet
[5:39] * shendaras (shendaras@ip68-97-113-208.ok.ok.cox.net) Quit ("Boom")
[5:44] <jabawok_w> duana and iip_i2p and Fisu_ - YoYo working now... still getting a lot of RNFs on other pages though (but thats 'normal' ie hasn't really changed from 5085)..
[5:44] <jabawok_w> the YoYo thing may have been a connection limit in firefox (not sure if i've increased max connections on this one yet)
[5:44] * TheSeeker (~Fridlekh@cpe-65-28-158-211.socal.rr.com) has joined #Freenet
[5:47] <iip_i2p> <wart> Hmm, simple enough solution: increase the JVM's memory limit
[5:50] <duana> how do u ppl keep ur bandwidth in check? the few times ive run my node (im new to freenet) it chews up 500mb in 12 hours. i've tried using trickle and the setting in the config file to limit it, but it doesnt seem to work after doing that
[5:52] <jabawok_w> whats that a solution to?
[5:53] <oierw> it should be fairly decent now
[5:53] <oierw> at bandwidth limiting
[5:56] <jabawok_w> I use linux ingress policing for incoming and htb for outgoing traffic shaping - and set freenet's limits to slightly higher than the o/s forced limits..
[5:56] * TheSeeker (~Fridlekh@cpe-65-28-158-211.socal.rr.com) Quit (Read error: 60 (Operation timed out))
[5:57] <duana> ok thanks .. some things to try
[6:07] <iip_i2p> <wart> Anyone else getting excessively high loads with 5087?
[6:08] * nextime (~nextime@danex.i-m-c.it) Quit ("nextime has no reason")
[6:08] <iip_i2p> <wart> load average: 14.13, 9.10, 6.76
[6:08] <iip_i2p> <wart> Load: 120%
[6:11] * nextime (~nextime@danex.i-m-c.it) has joined #freenet
[6:13] <jabawok_w> yep.. due to more mem usage i think here
[6:13] <jabawok_w> only got 384M on freenet box - and its running mldonkey as well so.. 250 odd meg into swap...
[6:13] <jabawok_w> iip_i2p: post a few lines from "vmstat 5" (not the first line)
[6:13] <iip_i2p> <wart> This is rather frustrating
[6:13] <Iakin3> [11:37] <iip_i2p> <wart> Hiya. After upgrading to 5087 I'm getting OOM errors during seeding as my node starts up. Any advice?
[6:13] <Iakin3> wart, read the mailing list..
[6:14] <Iakin3> There has been other that has had the same problem
[6:14] <Iakin3> (or just increase your -Xmx)
[6:14] <iip_i2p> <wart> I have, I didn't see anything about rediculous CPU usage
[6:14] <iip_i2p> <wart> I'm no longer running out of memory
[6:16] <Iakin3> duana: You know that the bandwidth config items are specified in bytes, not bits, per second?
[6:16] <iip_i2p> <wart> jabawok, There is plenty of read activity
[6:17] <jabawok_w> what about si and so
[6:17] <duana> Iakin3 - good chance i did use bits per sec.. thanks ill check
[6:18] <iip_i2p> <wart> Aye, frequent swapping
[6:18] <iip_i2p> <wart> I'm getting something like two minimum request interval violations per second
[6:20] <duana> hmmm nope had it as bytes
[6:20] <Iakin3> Did you remember to remove the '%' comment sign
[6:20] <duana> yer
[6:21] <Iakin3> And it was outputBandwidthLimit that you set to something?
[6:21] <duana> yes
[6:21] <Iakin3> Hmm..
[6:21] * guido^pe (~unknown@dsl-213-023-238-173.arcor-ip.net) has joined #freenet
[6:21] <Iakin3> If you set it to something lower?
[6:22] <Iakin3> The limiting isn't really exact
[6:22] <iip_i2p> <Newsbyte> a lot of OOM is being reported with the 5087
[6:23] <duana> wd the number of connections have something to do with it?
[6:24] <duana> not if theyre sharing the same pipe i guess
[6:24] <iip_i2p> <Newsbyte> heh...well, *my* numder of connections is about 7
[6:24] <iip_i2p> <wart> Newsbyte, Yes, a few fixes have been posted
[6:25] <duana> i guess i'll keep playing around and in the meantime pay to get unlimited uploads :/
[6:25] <iip_i2p> <wart> Newsbyte, you have strange problems.
[6:25] <iip_i2p> <wart> ;)
[6:26] <iip_i2p> <Newsbyte> worth noting these strange problems only commenced with the 5085 build, though
[6:27] <Pascal> WTF: logLevel=Error 27155759104 Jul 28 05:47 freenet.log
[6:27] <iip_i2p> <wart> And that not many people are seeing what you're seeing. Besides, 5085 worked wonderfully for me. 5086 had some problems, but still worked.
[6:27] <jabawok_w> anybody know what this means? :The Insert Request failed.
[6:27] <jabawok_w> Reason: Request failed gracefully: freenet.client.WrongStateException: Wrong state: FAILED should be DONE: after waiting for next process
[6:27] <Pascal> uptime 1 day 9 hours 46 minutes
[6:28] <iip_i2p> <Newsbyte> I wouldn't know about that. When I reported that on the maillist, some said they experienced the same, so it can't be that freaky
[6:28] <Iakin3> Newsbyte, thats right.. any idea of the cause?
[6:28] <iip_i2p> <Newsbyte> though toad claimed it was the encryptionbug that caused it
[6:29] <iip_i2p> <Newsbyte> but if it were, then it should have been solved in the 5087
[6:30] <Iakin3> jabawok: It simply means that the insert failed
[6:30] <Iakin3> Newbyt, do you mean the large logs?
[6:30] <Iakin3> or the connectioncount?
[6:31] <iip_i2p> <Newsbyte> anyways...do any og you guys have a homepage where you could set the freenethelp site on (the html mirror, that is) for a few days?
[6:32] <iip_i2p> <wart> Can anyone tell me what this minimum request interval stuff is?
[6:32] <Iakin3> If you mean the latter... it wont do any good to your node if everyone else still has non-upgraded nodes
[6:32] <Iakin3> It is a mechanism for limiting the number of requests a node gets
[6:32] <iip_i2p> <Newsbyte> MRI...is explaned on the freenethelp.org wiki site :-)
[6:32] <Iakin3> Ahhh
[6:32] <iip_i2p> <wart> Thanks
[6:33] <iip_i2p> <wart> At least disabling rate limitig is keeping my load average under 6
[6:33] <iip_i2p> <Newsbyte> good, good...so, ermm...the homepage/webspace...anyone?
[6:36] <Iakin3> Can someone who experiences these new OOMs try to figure out wheter or not they are still there when not using the native big integer code
[6:36] <iip_i2p> <Newsbyte> n07 4ll 47 1ce !!
[6:36] <Iakin3> NewsByte, I could have hosted the pages if I wherent going for vacation in two days
[6:37] <Iakin3> (with the web server shut down)
[6:37] <iip_i2p> <Newsbyte> good!I only need a copple of days! :-)
[6:45] <Iakin3> Is that enough?
[6:45] <Iakin3> I am leaving on friday
[6:46] <wind789> are there any plans to add a feature where freenet will shut down without interupting transfers that are already in progress?
[6:52] * guido^pe (~unknown@dsl-213-023-238-173.arcor-ip.net) Quit (Read error: 54 (Connection reset by peer))
[6:53] * guido^pe (~unknown@dsl-213-023-243-072.arcor-ip.net) has joined #freenet
[7:08] * guido^pe (~unknown@dsl-213-023-243-072.arcor-ip.net) Quit ()
[7:09] * guido^pe (~unknown@dsl-213-023-243-072.arcor-ip.net) has joined #freenet
[7:12] <Fisu_> Transfers active (Transmitting/Receiving)410 (328/82)
[7:12] <Fisu_> 328 simultaneous outbound transfers on 128Kbyte connection?
[7:13] <Fisu_> Build: 5087
[7:14] <mikeDOTd> Fisu_: the default max connections is 128
[7:14] <toad_> <salahx> FReent uses "negative trust" i.e. in punshes bad nodes
[7:14] <toad_> <salahx> other system use "positive trust" i.e. reawrd good nodes
[7:14] <toad_> wrong
[7:15] <toad_> freenet uses positive trust
[7:15] <toad_> that's what bidi does
[7:15] <toad_> amongst other things
[7:15] <mikeDOTd> Fisu_: perhaps you have it changed in freenet.conf/freenet.ini?
[7:15] <Fisu_> mikeDOTd: Connections open (Inbound/Outbound/Limit)86 (52/34/200)
[7:16] <toad_> negative trust is worthless unless identity is a nonzero cost
[7:16] <toad_> usually it isn't on freenet
[7:17] <toad_> <wind789> I'm on a regular cable connection, I can upload up to about 60 kb/sec - you CAN?!
[7:17] <toad_> 60kbits/sec surely?
[7:17] <toad_> 60kB/sec is half a megabit!
[7:17] <Iakin3> [13:04] <wind789> are there any plans to add a feature where freenet will shut down without interupting transfers that are already in progress?
[7:18] <toad_> you can't get that here... not unless you pay loads...
[7:18] <Iakin3> Hmm.. care to clarify that?
[7:18] <Fisu_> toad_: The default ADSL here is 512/512
[7:18] <mikeDOTd> some cable modems in the US support rates that high
[7:19] <toad_> <wind789> my upload bandwidth is much higher than my download bandwdith - that's just wierd
[7:21] <toad_> <salahx> until it gets the request (which includes the exnrpytion key the node does not know), it has no idea - no, it never has any idea
[7:21] <toad_> only the client end of the node knows
[7:21] <toad_> if it gets a local request with the decrypt key
[7:24] <toad_> <duana> do u know if they're planning to make FCP encrypted?
[7:24] <toad_> <duana> we want to make an applet that can insert messages onto freenet that can be used on machines that don't run nodes
[7:24] <toad_> no, but you could use FNP if you want it encrypted
[7:24] <toad_> of course you'd have no plausible deniability as a transient
[7:26] <toad_> <jabawok_w> i did find freenet a little slow - maybe because so many nodes have been knocked offline by the startup prob
[7:26] <toad_> what startup prob?
[7:26] <toad_> Newsbyte had a problem but it was just that the node was still running?
[7:27] <Iakin> if test $1 -eq "dynamic"
[7:27] <Iakin> then
[7:27] <Iakin> LIBPATH="-L.libs"
[7:27] <Iakin> INCLUDELIBS="-lgmp"
[7:27] <Iakin> else
[7:27] <Iakin> STATICLIBS=".libs/libgmp.a"
[7:27] <Iakin> fi
[7:27] <Iakin> And someone help me getting that right?
[7:28] <Iakin> s/And/can
[7:29] <toad_> <jabawok_w> I use linux ingress policing for incoming and htb for outgoing traffic shaping - and set freenet's limits to slightly higher than the o/s forced limits.. - surely you want to set freenet's limits lower than the forced limits, to give it a chance with high level limiting?
[7:31] <toad_> <iip_i2p> <wart> At least disabling rate limitig is keeping my load average under 6 - EH?
[7:32] <toad_> Fisu: probably that is a reporting bug
[7:32] <toad_> also connections != transfers, you can have 100 transfers on 1 conn
[7:36] <guido^pe> I have problem starting Freenet on my FreeBSD machine. It'll start up and then consume cpu time without end until I stop it again but it will never show any lifesigns besides this
[7:36] <iip_i2p> <wart> toad, As I said ;) My load average hovered around 12 until I disabled limiting, at which time it stayed near 6. I recently reseeded and now its staying near 2.
[7:36] <guido^pe> (like port 8888 becoming available or a "store" directory appearing)
[7:36] <toad_> aaargg
[7:36] <iip_i2p> <wart> Coincidence, perhaps?
[7:37] <iip_i2p> <wart> Also, I'm getting fewer MRI violations since I reseeded (no longer one or two every second), but they're still relatively frequent
[7:37] <toad_> ugh
[7:38] <toad_> report on support:
[7:38] <toad_> RROR: The resource freenet/support/CPUInfo
[7:38] <toad_> mation/libjcpuid-x86-linux.so was not a valid library for this platform
[7:38] <toad_> java.lang.UnsatisfiedLinkError: /tmp/jcpuid15030lib.tmp: libgcc_s.so.1: cannot
[7:38] <toad_> pen shared object file: No such file or directory
[7:38] <toad_> at java.lang.ClassLoader$NativeLibrary.load(Native Method)
[7:38] <Iakin3> So?
[7:38] <toad_> :<
[7:38] <toad_> that should be ERROR
[7:38] <Iakin3> Read on
[7:38] <Iakin3> Ah
[7:39] <toad_> it needs to be statically linked...
[7:39] <Iakin3> Maybe he just copied part of the message?!
[7:39] <toad_> Iakin3: there's a stack trace
[7:39] <toad_> but it's not much use
[7:39] <toad_> we know what the problem is
[7:45] * TheSeeker (~Fridlekh@cpe-65-28-158-211.socal.rr.com) has joined #Freenet
[7:49] <Iakin3> Right..
[7:49] <Iakin3> So whats the problem?
[7:50] <Iakin3> Or.. rather, I know what the problem is
[7:50] <Iakin3> but why are you troubled by it?
[7:50] <toad_> Iakin3: the problem is that we don't statically link libjcpuid
[7:50] <guido^pe> toad_: could you spare some time to help me with my problem?
[7:50] <toad_> guido^pe: what problem?
[7:51] <guido^pe> the one I stated above. I cannot start Freenet on FreeBSD with Sun Java 1.4.2
[7:52] * tyesinclair (~tyesincla@170.252.64.1) has joined #freenet
[7:52] <toad_> Iakin3: are you dealing with the native issues at the moment?
[7:52] <toad_> what are you doing?
[7:52] <guido^pe> <guido^pe> I have problem starting Freenet on my FreeBSD machine. It'll start up and then consume cpu time without end until I stop it again but it will never show any lifesigns besides this
[7:52] <kers> guido^pe: i got exactly the same problem with 1.4.2 on osx - it's problems with that version (of java)
[7:52] <toad_> IMHO we need to do: 1) compile all included native libs statically
[7:53] <toad_> 2) include i586, k6, k62, and k63
[7:53] <toad_> Iakin: are you working on this?
[7:53] <guido^pe> kers: Does that mean I need to downgrade?
[7:53] <toad_> if not, I will, soon
[7:55] <kers> guido^pe: i think 5087 requires 1.4.x (correct me someone if i'm wrong) - so you would need to find a newer java build (wich i can't for osx, atleast) :/
[7:57] * tyesinclair (~tyesincla@170.252.64.1) has left #freenet
[7:58] <guido^pe> Well, theoretically I could try 1.4.1 or 1.5 beta, right?
[8:02] * toad_ makes jcpuid use -static-libgcc...
[8:04] <iip_i2p> <Newsbyte> good!I only need a copple of d
[8:08] <mikeDOTd> kers: osx 10.3 has Java(TM) 2 Runtime Environment, Standard Edition (build 1.4.2_03-117.1)
[8:10] <wind789> <toad_> <wind789> I'm on a regular cable connection, I can upload up to about 60 kb/sec - you CAN?! -Yes, with multiple connection I can get up to 60 killa Bytes, not bits freenet it using 61 kb/sec now
[8:10] * TheSeeker (~Fridlekh@cpe-65-28-158-211.socal.rr.com) Quit (Read error: 110 (Connection timed out))
[8:10] <toad_> multiple connection?
[8:11] <toad_> you mean you have several IPs?
[8:11] <wind789> <toad_> <wind789> my upload bandwidth is much higher than my download bandwdith - that's just wierd - I mean freeenet usualy uploads more stuff from my node than it downloads
[8:11] <wind789> no just one, I mean if I have multiple connecitons open
[8:11] <wind789> sending yo many people at once ect...
[8:11] <wind789> to*
[8:11] * nicktastic (~nicktasti@c-67-163-227-74.client.comcast.net) has joined #freenet
[8:13] <wind789> with 1 connection it only gets about 40 kb/sec max I think
[8:19] * mazzanet (~mazzanet@mazzanet.registered) Quit ("Read error: Connection reset by beer")
[8:19] * mazzanet (~mazzanet@mazzanet.registered) has joined #freenet
[8:21] <toad_> hmmm
[8:21] <toad_> how do i compile a .so and tell it to link statically to libc etc?
[8:25] <kers> mikeDOTd: does that work for you ?
[8:26] <nicktastic> toad_: As far as I know, you can't statically link a shared library. That defeats the purpose of it being a shared library.
[8:27] <toad_> nicktastic: I know we can, because we do
[8:27] <toad_> I want to know how to do that with libc
[8:28] <mikeDOTd> kers: i run my node on my linux box, but in the past i've been able to run a node under osx
[8:29] * srob99 (~srob99@203-59-84-243.dyn.iinet.net.au) Quit (zelazny.freenode.net irc.freenode.net)
[8:29] * jabawok_w (~jabawok_w@203-59-60-38.perm.iinet.net.au) Quit (zelazny.freenode.net irc.freenode.net)
[8:29] * lostlogic (~lostlogic@node-402418b2.mdw.onnet.us.uu.net) Quit (zelazny.freenode.net irc.freenode.net)
[8:29] * srob99 (~srob99@203-59-84-243.dyn.iinet.net.au) has joined #freenet
[8:29] * jabawok_w (~jabawok_w@203-59-60-38.perm.iinet.net.au) has joined #freenet
[8:29] * lostlogic (~lostlogic@node-402418b2.mdw.onnet.us.uu.net) has joined #freenet
[8:29] <toad_> i think i've fixed the problem with libgcc_s.so.1
[8:30] <nicktastic> toad_: Which libraries do you link in that manner? jbigi?
[8:30] <toad_> ah, no i haven't
[8:30] * toad_ does so...
[8:30] <toad_> nicktastic: jcpuid had dynamic libgcc_s.so.1
[8:31] <toad_> that was a major problem as it is heavily versioned
[8:31] <toad_> jbigi didn't
[8:31] <toad_> so should run on any compatible libc6 system... maybe
[8:33] <toad_> okay, new freenet-ext.jar uploaded
[8:37] <nicktastic> toad_: How did you instruct the linker to merge the contents of libgcc_s.so.1 into libjcpuid-x86-linux.so?
[8:37] <toad_> nicktastic: -static-libgcc
[8:38] <nicktastic> haha Very nice
[8:38] * nicktastic takes note
[8:40] * toad_ builds jbigi so's for pentium 1, MMX, K6, K6-2, K6-3
[8:40] * toad_ wonders if K6-2 is identical to K6-3 or they do some wierd cache size optimizations
[8:41] <toad_> bbl
[8:42] * toad_ also thinks we need i586 to fall back on for cyrix, via, centaur, etc
[8:42] * toad_ mailed iakin re that
[8:44] <toad_> actually some of them have mmx and 3dnow... hrrm
[8:44] * Hory (~Miranda@81.196.25.110) has joined #FreeNet
[8:44] * toad_ still thinks the solution with wierd x86 compatible chips is to just bench all compatible cores
[8:45] <toad_> i suppose we'd have to detect features first though
[8:52] * greycat (~wooledg@192.35.79.70) has joined #freenet
[8:53] * robilad is now known as robilad[afk]
[9:02] <hirvox> just allow the user to manually select the appropiate .so
[9:02] * duana (~duana@adsl-4-132.VIC.dft.com.au) has left #freenet
[9:04] <Iakin3> Hehe 'at least I think I have' :)
[9:04] <greycat> Ohhh, loooky, we're spiralling down into the RNF+pcaching eat-up-all-the-blocks-so-we-have-less-and-less-of-them insertion deathtrap!
[9:05] <toad_> Iakin3: hi
[9:05] <greycat> 2-3 attempts ago, I managed to insert 73/78 blocks. Then 72. Now 70.
[9:05] <Iakin3> Hup, well, yes I am working on it.. but not _that_ particular issue
[9:05] <toad_> hirvox: the problem is that on obscure platforms we don't know what's the right one and neither does the user
[9:05] <toad_> Iakin3: I'm building the pentium 1/mmx/k6/etc .so's
[9:05] <greycat> If this trend continues, I'll soon be inserting NEGATIVE blocks! That should be fun!
[9:05] <toad_> i will commit them when they are done
[9:06] <toad_> i've fixed the jcpuid stuff
[9:06] <toad_> hopefully we won't run into libc issues
[9:06] <Iakin3> I am adding those extra architectures, as well as freebsd special-casing and dynamic linking to gmp..
[9:06] <toad_> as i couldn't find how to link that statically
[9:06] <toad_> which archs?
[9:06] <toad_> non-x86 ones/
[9:06] <toad_> ?
[9:09] <Iakin3> [15:03] <toad_> i suppose we'd have to detect features first though
[9:09] <greycat> 71/78 this time.
[9:09] <Iakin3> That code is prepared for..
[9:10] <Iakin3> toad: No.. those that your are compiling for :)
[9:10] <toad_> Iakin3: am I tripping you up?
[9:10] <Iakin3> tripping me up?
[9:10] <toad_> are you still working on build.sh?
[9:10] <Iakin3> The was a phrase I dont know the meaning of
[9:10] <Iakin3> Aha..
[9:11] <Iakin3> Well, dont worry.. I will merge your shit
[9:11] <Iakin3> :)
[9:11] <toad_> i haven't modded it
[9:11] <Iakin3> ok
[9:11] <toad_> i just want to know if you're going to make any more changes
[9:11] <Iakin3> Well.. please stay out of the jbigi build scripts if possible
[9:11] <toad_> and if so, what changes?
[9:12] <Iakin3> Yea, well.. I am working on that 'use a preexistant gmp.so if it exists'
[9:12] <toad_> ahh cool (as a fallback)
[9:12] <Iakin3> So people can do their 'emerge gmp' etc
[9:13] <greycat> *cough*apt-get
[9:13] <toad_> do you have a linux machine to test it on?
[9:13] <Iakin3> Nope..
[9:13] <Iakin3> Or well.. if I want to
[9:13] <toad_> do you want a shell on one?
[9:13] <Iakin3> But I dont :)
[9:13] <hirvox> *cough*my rpms*cough* :)
[9:13] <toad_> brb
[9:13] <Iakin3> I am working in mingw, so the new build-scripts shouldn't be a problem to run in linux
[9:15] <hirvox> Iakin:I'm having trouble getting net.i2p.util.NativeBigInteger to detect jbigi and jcpuid in LD_LIBRARY_PATH
[9:17] <toad_> hmm
[9:17] <toad_> it's cvs add -b for binaries, right?
[9:21] <Iakin3> toad: how do I do an if `uname -sr` startswith MINGW then
[9:21] <Iakin3> in shell scripting?
[9:21] <toad_> ummm
[9:21] <toad_> i'm not sure :)
[9:21] <Iakin3> hirvox: What name have you given it?
[9:21] <Iakin3> Anyone else?
[9:21] <toad_> iirc start-freenet.sh has an example
[9:22] <hirvox> libjbigi-linux-none.so and libjcpuid-x86-linux.so
[9:22] <toad_> case `uname -sr` in
[9:22] <toad_> MINGW*) do something ;;
[9:22] <toad_> esac
[9:23] <toad_> is it cvs add -kb for binaries?
[9:24] <toad_> hmmm
[9:24] <toad_> http://elib.cs.berkeley.edu/admin/cvs/cvsaddko.html
[9:24] <toad_> is that still correct?
[9:25] <greycat> 71/78 blocks again.
[9:27] <Iakin3> New build scripts committed..
[9:27] <vsalento> cvs add -kb is correct
[9:27] <Iakin3> Please try and see if they work
[9:27] <Iakin3> (jbigi build scripts that is)
[9:27] <toad_> Iakin3: what changes?
[9:28] <Iakin3> k6 etc. added
[9:28] <toad_> hmmm
[9:28] <Iakin3> dynamic lib building added
[9:28] <Iakin3> experimental freebsd building added
[9:29] <greycat> 71/78 blocks once again (lots of CANCELLED this time).
[9:30] <toad_> Iakin3: you need -static-libgcc, don't you?
[9:30] <Iakin3> Noone has complained so far...
[9:31] <toad_> okay, you don't
[9:31] <toad_> good
[9:31] <Iakin3> They have only complained about jcpuid..
[9:31] <toad_> for some reason it doesn't use it
[9:31] <Iakin3> and said that it did load '..-none-so'
[9:31] <Iakin3> hirvox, what messages do you get?
[9:31] <toad_> okay, building...
[9:32] <Iakin3> It tries the jar first, then the LD_LIBRARY_PATH
[9:32] <toad_> bbiab
[9:32] <toad_> Iakin3: why do we not include pentium pro btw?
[9:33] <toad_> and what will it do with e.g. a cyrix 5x86?
[9:34] <Iakin3> Bah.. they are on their own
[9:35] <Iakin3> They can use the generic code or the dynamically linked one
[9:35] <Iakin3> Noone runs windows on a PPro nowadays
[9:35] <mikeDOTd> but they run on cyrix?
[9:35] <toad_> what about linux?
[9:35] <toad_> and yes, people do
[9:35] <toad_> for "firewall" machines etc
[9:36] <mikeDOTd> my current quad-cpu server is ppro based
[9:37] <Iakin3> Is it running windows?
[9:37] <mikeDOTd> freebsd
[9:37] <mikeDOTd> ah
[9:37] <toad_> how much did you get it for on ebay? :)
[9:38] <mikeDOTd> i see
[9:38] * lolo-laptop (~lostlogic@68.251.84.226) has joined #freenet
[9:38] <mikeDOTd> toad_: it was a donation, in exchange for hosting it on some pipe
[9:38] <greycat> 74/78 blocks this time.
[9:38] <toad_> Compiling C code...
[9:38] <toad_> ./build.sh: line 60: syntax error: unexpected end of file
[9:38] <mikeDOTd> i don't think it's reasonable to assume people won't run windows on a ppro system
[9:38] <toad_> Building jbigi lib that is statically linked to GMP
[9:38] <toad_> Compiling C code...
[9:38] <toad_> ./build.sh: line 60: syntax error: unexpected end of file
[9:39] <toad_> this is after it's done the athlon
[9:39] <toad_> so i'm cvs adding the other files
[9:39] * sanity|m (~symirc@future-is.orange.co.uk) has joined #freenet
[9:39] <Iakin3> toad, update from cvs again
[9:39] <Iakin3> I fixed it I thinkg
[9:39] <mikeDOTd> do you guys need a solaris build?
[9:39] <toad_> Iakin3: can I still commit the other files?
[9:39] <toad_> mikeDOTd: it'd be nice...
[9:39] <Iakin3> Sure
[9:39] <mikeDOTd> toad_: ok
[9:40] <toad_> I don't know how we'd use it though
[9:40] <toad_> we need detection code...
[9:40] <mikeDOTd> solaris has sysctl too
[9:40] * sanity|m (~symirc@future-is.orange.co.uk) Quit (Read error: 54 (Connection reset by peer))
[9:40] <mikeDOTd> if you're going that route
[9:41] <Iakin3> Toad, I am in the process of building the new windows libs too.. So, plz wait for them before updating freenet-ext.jar again
[9:41] <Iakin3> It will probably take a bunch of hours though before they are done
[9:41] <toad_> okay
[9:41] <toad_> note on security: if anyone sees anyone other than me, Iakin, Thelema, Hirvox, Sanity committing binaries, shoot to kill
[9:42] <toad_> why does it think i've modified libjbigi-linux-athlon.so etc?
[9:43] * toad_ supposes he probably has
[9:45] * TheSeeker (~Fridlekh@cpe-65-28-158-211.socal.rr.com) has joined #Freenet
[9:46] * toad_ wonders why we still have so few connections
[9:46] <toad_> i used to get 100+ on stable
[9:46] <toad_> now i have 75, and newsbyte has ~ 7
[9:49] * toad_ wonders why it is taking so long to commit these libs...
[9:49] <lolo-laptop> maybe freenet is shrinking?
[9:49] <toad_> ah, they're in
[9:49] <toad_> good
[9:49] <lolo-laptop> maybe nodes on pre 5087 are so backed up in connection openeres they are rejecting?
[9:49] * tyesinclair (~tyesincla@170.252.64.1) has joined #freenet
[9:49] * tyesinclair (~tyesincla@170.252.64.1) has left #freenet
[9:50] <toad_> heh
[9:51] <toad_> the ReactionOS freesite does not contain a downloadable reactionos from freenet
[9:51] <toad_> otoh the xbox stuff looks findable
[9:52] <greycat> Grrr! I made the mistake of trying to click the open connections page in a non-newborn mozilla....
[9:52] <greycat> BTW: 76/78 blocks this time.
[9:55] <toad_> greycat: http://127.0.0.1:8888/servlet/nodeinfo/networking/ocm?setMode=Connection
[9:55] <greycat> toad_: can I "save" that mode without editing the .conf file and restarting the node?
[9:55] <toad_> greycat: don't know
[9:56] <greycat> Connections open (Inbound/Outbound/Limit) 75 (33/42/200)
[9:56] <greycat> Transfers active (Transmitting/Receiving) 84 (29/55)
[9:56] <greycat> Data waiting to be transferred 376 Bytes
[9:56] <greycat> Total amount of data transferred 917 MiB
[9:56] <greycat> I normally have more connections than that.
[9:56] <toad_> bbiab
[9:56] <toad_> it might just be all the 5086/5085's
[9:56] <toad_> which have major connection problems
[9:56] <toad_> otoh it might not be...
[9:58] <toad_> bbiab
[10:04] <Iakin3> toad: Did you get the build to start?
[10:06] <greycat> 77/78 blocks.
[10:07] * TheSeeker (~Fridlekh@cpe-65-28-158-211.socal.rr.com) Quit (Read error: 110 (Connection timed out))
[10:07] * robilad[afk] is now known as robilad
[10:08] <Iakin3> Its going forwards again :)
[10:09] <greycat> 78/78 blocks, failed (RNF) on the mapfile key.
[10:13] <Pascal> toad: those test settings are working even better than the ones I had. Unfortunatly the log filled up my hdd
[10:15] <Pascal> toad: looks like the same "Interpolated crazy value" problem you had yesterday. Been fixed yet?
[10:18] * kers (~kers@86.ppp144.rsd.worldonline.se) Quit (Read error: 104 (Connection reset by peer))
[10:22] * kers (~kers@3.ppp144.rsd.worldonline.se) has joined #freenet
[10:24] * Ash-Fox (Hal-9000@aae32.neoplus.adsl.tpnet.pl) has joined #FreeNET
[10:24] * Pascal (Pascal@c-24-13-17-191.client.comcast.net) Quit ()
[10:25] * Pascal (Pascal@c-24-13-17-191.client.comcast.net) has joined #freenet
[10:35] <lolo-laptop> Pascal: yeah, that is fixed
[10:36] <Pascal> lolo-laptop: thanks
[10:39] <greycat> 78/78 blocks, failed (RNF) on the mapfile key again.
[10:42] <wind789> greycat:doing an upload test?
[10:42] <greycat> wind789: something like that.
[10:42] <wind789> stable or uinstable?
[10:43] <greycat> wind789: this is the SAME DAMNED FILE I've been trying to insert since yesterday.
[10:43] <greycat> stable, 5087
[10:43] <wind789> lol
[10:43] <wind789> got a chk? I can try to download if you'd like
[10:45] <greycat> 78/78+RNF again.
[10:45] <wind789> darn
[10:56] * Iakin3 noticed that he has a build #5093 running on one of his machines
[10:59] * thelema|aikido is now known as thelema
[11:00] <thelema> hi
[11:17] <thelema> why are we trying to fit a bunch of implementations of the native code into the same freenet-ext?
[11:17] <thelema> rather, why are we trying to fit them *all* in?
[11:18] <thelema> It seems a reasonable idea to have freenet-ext-x86
[11:18] <thelema> ,freenet-ext-mac
[11:18] <thelema> ,freenet-ext-wierdCPUs
[11:19] <thelema> three .jars will cover everyone, and we can still keep the x86 jar (the one that nearly everyone will use) smaller in size.
[11:19] <KenMan> yeah, that does seem more logical, but simple is good too.
[11:21] <KenMan> Okay, about 7 hours ago, my node served out 243 items in an hour. This is about 4 times normal. But requests are 5-10 times normal...
[11:22] <KenMan> upon further inspection, sentData is crap... queries per hour are too high...
[11:23] <toad_> <Pascal> toad: those test settings are working even better than the ones I had. Unfortunatly the log filled up my hdd
[11:23] <toad_> <Pascal> toad: looks like the same "Interpolated crazy value" problem you had yesterday. Been fixed yet?
[11:23] <toad_> huh?
[11:24] <toad_> well, the x86 jar will be quite large
[11:24] <greycat> another 78/78+RNF some time while I was /away
[11:24] <toad_> the others won't make much difference probably
[11:25] <toad_> KenMan: sentData vs sentDataNonInsert ?
[11:25] <toad_> what are the causes of sentData failure?
[11:26] <KenMan> yes, more than half apparently were inserts :o
[11:26] <greycat> My node's also passed 1 GB data sent, with an uptime ~15 hours.
[11:26] <toad_> well, inserts are supposed to get transfer failed more often...
[11:26] <toad_> because we start sending before we have a confirmed route...
[11:27] <KenMan> by design, or by coincidence ??!
[11:27] <KenMan> ahh
[11:28] <thelema> toad_: better than making one *huge* jar, and making the x86 people download mac and whatever else code.
[11:28] <toad_> thelema: if the mac and whatever else code is 500kB, and the x86 code is 2MB, what difference does it make?
[11:28] <toad_> if we can separate them automatically, then perhaps
[11:28] <toad_> but NOT if it makes the install more complex for the end-user
[11:29] <KenMan> Good thing I have broadband!! - Transfers active (Transmitting/Receiving) 253 (164/89)
[11:29] <toad_> KenMan: are the transmits real?
[11:29] <KenMan> how the hell would I know ??
[11:30] <KenMan> I can only guess yes, as they don't normally go this high...
[11:30] <greycat> Connections open (Inbound/Outbound/Limit) 72 (32/40/200)
[11:30] <greycat> Transfers active (Transmitting/Receiving) 155 (89/66)
[11:30] <greycat> Data waiting to be transferred 5,646 Bytes
[11:30] <greycat> Total amount of data transferred 1,030 MiB
[11:30] <toad_> hmmm
[11:30] <KenMan> well, at least we are back to xmits > receives
[11:30] <toad_> lets see if i can reproduce these "early RNF"s
[11:30] <thelema> toad_: it'll make a 25% difference in each filesize.
[11:31] <KenMan> btw, how are you guys managing to run .so's from inside of jar files ? or do you extract them first ??
[11:31] <thelema> toad_: native math isn't necessary to run freenet. it's a luxury, and it shouldn't be that difficult to download a special jar for mac users.
[11:32] <thelema> KenMan: the code extracts them first
[11:32] <toad_> thelema: it makes freenet usable on many platforms
[11:32] <thelema> KenMan: that was the issue someone just had with running code out of /tmp
[11:32] * KenMan screws his hair back into his head...
[11:33] <thelema> toad_: 99+% of which are x86. But in the case that someone wants to run fred on a ppc box or something wierder, they have to do a little extra work. big deal.
[11:34] <toad_> thelema: whereas an extra few hundred kilobytes takes how many seconds to download on a typical freenet link?
[11:34] <toad_> freenet-credible link?
[11:34] <toad_> they'll probably have to install the JVM anyway on windows
[11:35] <toad_> sh*t
[11:35] <toad_> grrr
[11:35] <Pascal> toad: the requestSendTimeCutoff and successfulSendTimeCutoff values you suggested the other day
[11:35] <toad_> it didn't compress the logs...
[11:35] <toad_> Pascal: hrrm?
[11:36] <Pascal> [10:41] <toad_> <Pascal> toad: those test settings are working even better than the ones I had. Unfortunatly the log filled up my hdd
[11:36] <toad_> ah
[11:36] <toad_> hmmm
[11:36] <Pascal> [12:42] * toad_ getting a LOT of Interpolated crazy value: -1998.766210074004
[11:37] <Pascal> but that was the bug I was running into
[11:37] <toad_> freenet is niced, therefore running compiles in the foreground...
[11:37] <KenMan> okay toad, here's my picture you requested -> http://mywebpages.comcast.net/jkcorson/FreeNet/5087c.png
[11:37] <toad_> Pascal: ahh
[11:37] <toad_> I thought the changes that caused that had been reverted?
[11:38] <Pascal> fixed now
[11:38] <toad_> KenMan: hrrm
[11:38] <Pascal> KenMan: where did that come from?
[11:38] <KenMan> I know the pale blue line shouldn't spike like that, it was just a peer starting up frost.
[11:38] <toad_> backed off doesn't follow requests so closely here...
[11:38] <KenMan> well, queries are through the roof...
[11:39] <toad_> hmmm
[11:39] <toad_> i think my stable node is stuck again
[11:39] <KenMan> Pascal - I've been charting interesting (and not so interesting) things with GnuPlot...
[11:40] <toad_> very long latency to access web interface
[11:40] <toad_> although it does get there in the end
[11:40] * KenMan believes he cured himself of all that ailed his machine...
[11:40] <toad_> ah, excellent:
[11:40] <toad_> Couldn't retrieve key: SSK@IWmzQkYKOlOefQoWbGCQSrkS3vsPAgM%2cNzRr-Pj88cVT0bN%7e1urLVg/FIND//
[11:40] <toad_> Hops To Live: 15
[11:40] <toad_> Error: Route Not Found
[11:40] <toad_> Attempts were made to contact 8 nodes.
[11:40] <toad_> * 0 were totally unreachable.
[11:40] <toad_> * 8 restarted.
[11:40] <toad_> * 0 cleanly rejected.
[11:40] <toad_> * 31 backed off.
[11:40] <toad_> we have 75 open connections!
[11:40] * Lemoni (~Lemoni@munyoo.cs.mu.OZ.AU) Quit ("Leaving")
[11:40] <toad_> ah, restarted. grrr.
[11:41] <toad_> it's credible because of the restarted's
[11:41] <greycat> I've been getting shitloads more RNFs than normal too.
[11:41] * toad_ supposes he should turn off the kenmanhack, then he might see some of the instarnf's
[11:41] <toad_> but i'd have to restart for that...
[11:41] * TLF (francisco@122.Red-81-40-116.pooles.rima-tde.net) has joined #freenet
[11:42] <toad_> Hops To Live: 15
[11:42] <toad_> Error: Route Not Found
[11:42] <toad_> Attempts were made to contact 4 nodes.
[11:42] <toad_> * 0 were totally unreachable.
[11:42] <toad_> * 4 restarted.
[11:42] <toad_> * 0 cleanly rejected.
[11:42] <toad_> * 35 backed off.
[11:43] * KenMan feels frustrated and cynical because too much is changing too fast ...
[11:44] <KenMan> of course, change _can_ be good...
[11:45] <thelema> KenMan: certain things were merged to stable a bit too fast...
[11:45] <thelema> toad_: do you think that 8 restarted is good?
[11:46] <toad_> thelema: no, but it could be simply caused by the cpu usage which is going nuts at the moment
[11:46] <toad_> no idle time at all in the last 5 minutes
[11:46] <KenMan> thelema: that's why I'm focused on stable now...
[11:47] * TheSeeker (~Fridlekh@cpe-65-28-158-211.socal.rr.com) has joined #Freenet
[11:47] <toad_> so how do i find out what the problem is?
[11:48] <thelema> toad_: you're saying the restarteds could be simply caused by CPU usage on your box?
[11:48] * jay (jcl@ool-18bf6dac.dyn.optonline.net) has joined #freenet
[11:48] <KenMan> here is my 2-cent analysis regarding qph - as mRIs are updated more frequently, the average (for routes) is much lower, thus reducing the backpressure for query rate control...
[11:48] <toad_> eh?
[11:48] <toad_> well the average will be lower
[11:49] <toad_> but only where it was too high before because nodes hadn't been updated
[11:49] <thelema> toad_: or are you blaming NBI's cpu usage reducing effect for creating restarteds?
[11:49] <toad_> where is the problem?
[11:49] <toad_> thelema: i'm blaming the cpu usage
[11:49] <KenMan> by having a lower average, my node is gobbling it up, spewing forth several times the normal amount of queries / unit of time...
[11:49] <toad_> i don't know what it's caused by
[11:49] <toad_> but i have many reports of it from users
[11:50] <thelema> many reports of way high CPU usage and memory use increase?
[11:50] <toad_> Checkpoint: Opening connection 9
[11:50] <toad_> freenet.Message: DataRequest 65
[11:50] <toad_> freenet.Message: QueryRejected 1
[11:50] <toad_> freenet.client.InternalClient$ClientMessageVector 1
[11:50] <toad_> freenet.interfaces.LocalNIOInterface$ConnectionShell 1
[11:50] <toad_> freenet.node.states.data.DataStateInitiator 24
[11:50] <toad_> freenet.node.states.data.TrailerWriteCallbackMessage:false:false 6
[11:50] <toad_> freenet.node.states.data.TrailerWriteCallbackMessage:true:true 7
[11:50] <toad_> freenet.node.states.request.RequestInitiator 3
[11:50] <toad_> freenet.node.states.request.SendFinished 3
[11:50] <toad_> okay, that's not good
[11:50] <toad_> but it looks like it's just log contention
[11:50] <toad_> if i turn off the log, i'll have no chance of finding bugs
[11:50] <toad_> grrrrrr
[11:51] <toad_> the only times i've managed to reproduce the 100% cpu usage problems have been when i've had some logging
[11:51] <KenMan> can you 'net log' by any chance ??
[11:51] <toad_> and when i do get it, guess what? it's the logging! :<
[11:51] <thelema> bbiab
[11:51] <toad_> KenMan: doesn't matter, logging produces significant CPU usage no matter where it gets written
[11:52] * toad_ needs $2K to spend on a dual processor system...
[11:52] <KenMan> maybe the efficiency boost of NBI has allowed your system to swamp your disk now...
[11:52] * KenMan pokes fun at the toad :)
[11:52] <toad_> KenMan: huh?
[11:52] <toad_> it's not a disk problem
[11:52] <toad_> it's a cpu problem
[11:53] <KenMan> why can logging take so much CPU ??
[11:53] <toad_> and it's NOT caused by connection openers
[11:53] <toad_> KenMan: because there's lots of it
[11:54] <KenMan> that's not a good answer. Lots of it should cause lots of disk, not lots of CPU...
[11:54] <toad_> hmmm
[11:54] <toad_> it was the gzip...
[11:54] <toad_> but that's gone
[11:54] <toad_> because there's lots of it
[11:54] <toad_> lots of string operations
[11:54] <toad_> etc etc
[11:54] <toad_> it takes up a lot of cpu if you have lots of it
[11:54] <toad_> even if one of it takes up very little cpu
[11:54] <KenMan> heh
[11:55] <toad_> it wasn't the gzip
[11:55] <toad_> it's still pegged
[11:55] <toad_> a lot of it seems to be system time
[11:55] <toad_> it alternates between ~
[11:55] <toad_> 20% and ~ 45%
[11:55] <toad_> 15% and 45%
[11:56] * Pascal (Pascal@c-24-13-17-191.client.comcast.net) Quit ()
[11:57] <toad_> i suppose it might be caused by the inter-thread stuff in the logging
[11:57] <toad_> but there really is no other choice
[11:57] <toad_> and i don't see why notifying on a lock would be expensive if nobody else is waiting
[11:58] <toad_> the high system usage suggests it might actually be caused by the disk writes...
[11:58] <KenMan> so you need to tune your selective logging a little better, perhaps ??
[12:01] <toad_> no
[12:01] <toad_> sometimes it seems to work reasonably
[12:02] * toad_ adds some buffering...
[12:03] <toad_> that might just be it
[12:03] <toad_> the writes may be too small
[12:03] <KenMan> wait, doesn't the system buffer for you ? especially at the disk-write level ??
[12:03] * toad_ added a 32kB buffer
[12:04] <toad_> well yeah but if you're sending 80 byte writes to the kernel, it's not very efficient
[12:04] <KenMan> :)
[12:04] <toad_> since it's written by a dedicated thread, and it's all timestamped, there's no reason not to buffer it
[12:05] * KenMan forecasts that the system will crash without getting that last crucial buffer to disk :p
[12:05] <toad_> freenet hardly ever crashes my PC
[12:05] <KenMan> i just slapped myself for being so pessimistic...
[12:08] <toad_> if this works... it'd be nice (although a little embarrassing, i'm sure there was some reason i took some buffering out)
[12:08] <KenMan> 12:19:19 Caught freenet.node.rt.EstimatorFormatException: No point 0 reading tcp/IPADDRESS ...
[12:08] <toad_> probably it wasn't buffered since before we made the separate thread
[12:08] <toad_> brb
[12:10] <toad_> so far so good...
[12:11] <KenMan> I really think I am cured. My use of the crummy VIA built-in sound was causing some clock interrupts to get lost !! So I shut it off and plugged in a real soundcard.
[12:11] <toad_> 50 running connectionopeners and 80% cpu idle...
[12:11] <toad_> KenMan: hmmm
[12:11] <toad_> I should try that sometime
[12:11] <toad_> i was using a good one until it broke
[12:11] <toad_> okay, this is very promising...
[12:11] <toad_> but perhaps there's a real cpu-gets-maxed-out-permanently bug...
[12:12] <toad_> definite improvement in initial responsiveness anyway
[12:12] <KenMan> geek-rock lyrics... "my CPU , it gets soo hot, it CAN'T cool down, baby"
[12:13] <KenMan> actually , I just substituted 'CPU' for 'motorcycle', and ripped off Joe Satriani
[12:15] <wind789> does freenet reserve a certain amount of bandwidth for message handling?
[12:15] <toad_> wind789: no
[12:16] <toad_> interesting
[12:16] <toad_> i get intervals, sometimes quite long, of no idle time...
[12:17] <toad_> hmmm
[12:17] <toad_> maybe it's not an interval...
[12:18] * TheSeeker (~Fridlekh@cpe-65-28-158-211.socal.rr.com) Quit (Read error: 110 (Connection timed out))
[12:18] <toad_> hmmm
[12:19] <toad_> this is rather unpromising...
[12:19] * Pascal (~Pascal666@m876936d0.tmodns.net) has joined #Freenet
[12:19] <toad_> hmmm
[12:19] <KenMan> now freenet requires an opteron 844 system in order to run ??
[12:19] <toad_> 850 :)
[12:19] * interrupt (~chatzilla@64.122.23.213) has joined #freenet
[12:19] <Pascal> [11:10] * toad_ needs $2K to spend on a dual processor system..
[12:19] <Pascal> if it would help debug fred I'll give you an account on mine
[12:20] <Pascal> but you'll have to walk me through it, i'm still new to linux
[12:20] <thelema> Pascal: useradd toad (as root)
[12:20] <toad_> i've been meaning to organize my various accounts and things, and use them more...
[12:20] <thelema> Pascal: passwd toad (to set his password)
[12:20] <KenMan> will you keep it up 24 hours a day ? and let toad use up all the CPU ??
[12:22] <Pascal> KenMan: I'm sure he can use his own box for most things
[12:22] <toad_> Pascal: :)
[12:22] <Pascal> toad_: would that help?
[12:22] <toad_> yeah
[12:22] <toad_> quite possibly
[12:22] <KenMan> :)
[12:23] <KenMan> you know, FreeNet should be able to rent a high-powered, high-netted server for small $$'s per month for Toad... no?
[12:24] * Pascal666 (~Pascal666@m067936d0.tmodns.net) has joined #Freenet
[12:24] <KenMan> I mean, easily under $100/mo should buy something pretty serious...
[12:24] <Pascal666> damn, hit wrong key :(
[12:24] <Pascal666> toad_: ok, I'll add you as soon as I get home
[12:25] <wind789> can sombody insert some junk? I wanna download somthing
[12:25] <wind789> prefereable stable
[12:26] <toad_> KenMan: can we?
[12:26] <toad_> i mean wwould it?
[12:26] * interrupt (~chatzilla@64.122.23.213) has left #freenet
[12:27] <toad_> hmmm
[12:27] <toad_> no idle over last 9*30 seconds...
[12:27] * thelema doesn't think a high powered high-netted server is going to really help much.
[12:27] <toad_> and not much for last 13*30 seconds
[12:28] <thelema> there's got to be an optimization in logging (probably not logging nearly as much) that'll help things.
[12:28] <toad_> thelema: it's just that if i have enough logging to debug stuff, it slows stuff down, often by too much
[12:30] * interrupt (~chatzilla@64.122.23.213) has joined #freenet
[12:31] <toad_> hmmm
[12:31] <toad_> these pauses are getting very annoying
[12:33] <toad_> KenMan seems to think they're caused by the built-in sound card on the mobo
[12:33] <toad_> the cheap sound card i have spare isn't good enough to play games with (drivers problem w/ winex)
[12:33] <toad_> but we can deal with that another day
[12:33] <KenMan> wind789: start looking for this - CHK@YfiV-hiTe7yXrJyVbt28RewR25cSAwI,AS4wghE6OR728xEfH5HMPA
[12:34] <toad_> KenMan: how did you fix it? just use a kernel without the driver for the onboard sound?
[12:34] <KenMan> yes, and turn off the sound in the BIOS
[12:35] <KenMan> cat /proc/interrupts - are any shared ??
[12:35] <Pascal666> is /. down?
[12:35] <KenMan> not for me
[12:35] <Pascal666> hmmm 503 Service Unavailable
[12:36] <KenMan> "Linux: SGI & NASA Plan 10240-Processor Altix Cluster"
[12:36] <iip_i2p> <Pseudonym> it was for me a little while ago
[12:36] * Pascal (~Pascal666@m876936d0.tmodns.net) Quit (Read error: 104 (Connection reset by peer))
[12:36] * Pascal666 is now known as Pascal
[12:37] * salahx (salahx@sc1-24.217.174.147.charter-stl.com) has joined #freenet
[12:37] <toad_> 9: 183414539 XT-PIC ehci_hcd, uhci_hcd, uhci_hcd, uhci_hcd, via82cxxx, eth0
[12:37] <KenMan> blehhhcchhh kill that ...
[12:37] <toad_> that has ethernet AND USB AND onboard sound
[12:37] <KenMan> if you don't use USB, remove those drivers too
[12:38] <toad_> I do use usb
[12:38] <toad_> mouse, keyboard
[12:38] <Pascal> hwo do you remove USB drivers? I've been tyring to figure that one out for some time now
[12:38] <toad_> I could put the keyboard on PS/2
[12:38] <toad_> i think the mouse port is broken
[12:38] <toad_> although perhaps i have a suitable serial port
[12:38] * toad_ would prefer to stay on USB though
[12:38] <KenMan> try poking about in the BIOS to get them separated... i compile a kernel with no USB support, and turn off the hardware in the BIOS
[12:38] <toad_> KenMan: is it just the sound card?
[12:39] <toad_> I mean, PCI ought to be abl to share interrupts without any major issues
[12:39] <KenMan> yeah, just get rid of sound support and cross you fingers
[12:39] <KenMan> it didn't do such a great job here...
[12:39] <toad_> KenMan: that's just because of the sound card though right?
[12:40] <KenMan> i don't really know yet. but i believe so.
[12:40] <toad_> okay
[12:40] <KenMan> (because of the builtin sounds card)
[12:40] * toad_ going to switch off pc to install ancient sound card whose drivers are incompatibl with winex
[12:41] <KenMan> just do without sounds for a day
[12:41] <KenMan> disable in BIOS and run with it
[12:41] <toad_> i'd have to reboot anyway
[12:41] <toad_> bye...
[12:41] * toad_ (toad@82-32-16-91.cable.ubr03.azte.blueyonder.co.uk) Quit ("using sirc version 2.211+KSIRC/1.3.10")
[12:43] <KenMan> i shot the sheriff, but I did not shoot his to-oad ...
[12:45] <Pascal> KenMan: :)
[12:45] * KenMan feels bad because he encouraged people to use their VIA builtin sound months back... :(
[12:47] <salahx> I ahve one of those, seems to work fine
[12:48] <KenMan> well, the two people I encouraged to do it were both reporting similar troubles last night...
[13:00] * TLF (francisco@122.Red-81-40-116.pooles.rima-tde.net) Quit ("http://www.vitaliano.esp.cc || Adi?s a todos y a todas y que les vaya bien.")
[13:00] * piranha (piranha@3ffe:b80:1ca1:0:0:0:deca:fbad) Quit (Read error: 60 (Operation timed out))
[13:00] * TLF (francisco@122.Red-81-40-116.pooles.rima-tde.net) has joined #freenet
[13:01] * piranha (piranha@3ffe:b80:1ca1:0:0:0:deca:fbad) has joined #freenet
[13:01] <hirvox> hrm, I'm seeing loads of these: 28.7.2004 20:15:13 (freenet.transport.WriteSelectorLoop, Network writing thread, NORMAL): fixKeys added sun.nio.ch.SelectionKeyImpl@25eaea
[13:02] <hirvox> about 10 per second
[13:03] <salahx> me too
[13:03] <salahx> Toad said it a logging bug
[13:05] <hirvox> hmm, would 23 per-cpu-type optimized .sos be enough for a x86 rpm?
[13:06] <Pascal> maybe we should have fred download the correct lib. freenet-ext.jar is getting too big.
[13:06] <hirvox> naah, just make more .jars. personally, I'd like to keep architecture-specific files at an arm's length from the node itself
[13:07] <salahx> i suggest tdynamic linking instead
[13:08] <hirvox> my libs _are_ dynamically linked
[13:08] <Pascal> keep current cpus (P4 and Althon?) in freenet-ext.jar and put the rest elsewhere
[13:09] <salahx> well most Linux distrobution are going to have libgmp avaiable (or irts easy enoguh to install), though it may not be optimized
[13:09] <salahx> Most Windows distributions won't thouigh
[13:09] <Pascal> we can simply add a note to the readme telling people to install libgmp
[13:10] <Pascal> basically make it a prerequisite
[13:10] <salahx> well it optioal, it'll work with out
[13:10] <hirvox> yeah, making it mandatory is silly
[13:10] <salahx> dfor Gento it woud lbe something like USE="gmp" emerge net-p2p/freenet
[13:11] <salahx> for other packagin system; howeverh tye do it is the packagers problem anyway
[13:11] * robilad is now known as robilad[afk]
[13:12] <hirvox> yeah, let the packagers worry about it
[13:21] * Pascal is now known as Pascal666
[13:28] <greycat> 77/78 blocks while at lunch.
[13:31] * Unrest (Unrest@ACAE9FB0.ipt.aol.com) has joined #freenet
[13:32] <Unrest> hello
[13:32] <Unrest> anyone alive in here?
[13:32] <Pascal666> always
[13:32] <salahx> yes
[13:32] <Unrest> ok
[13:33] <Unrest> everyone here from a freenet, or former freenet?:)
[13:33] <Unrest> one would assume.. but this is irc..:)
[13:34] <Unrest> pascal? :)
[13:34] <hirvox> what do you mean by former freenet?
[13:34] <Unrest> oh
[13:34] <Unrest> well
[13:35] <Unrest> Cleveland freenet vanished a few years ago
[13:35] <hirvox> this is www.freenetproject.org's channel
[13:35] <Unrest> not the same, eh?
[13:36] <hirvox> I've never even heard of Cleveland's freenet
[13:37] <Unrest> obviously not..lol.. Cleveland freenet was a unix script BBS.. based at Case Western Reserve University
[13:38] <Unrest> dial-up, mostly..
[13:38] <hirvox> this freenet is an anonymous p2p network
[13:38] <Unrest> but in those days, you could telnet.. anyone remember telnet ? :P
[13:39] <Unrest> the search continues for CFN people..:) lol
[13:41] <KenMan> Unrest: good luck !
[13:41] <Unrest> thanks..:)
[13:41] * Unrest (Unrest@ACAE9FB0.ipt.aol.com) has left #freenet
[13:42] <Pascal666> ever wonder about Ian's choice of name? With freenet6 and all the freenets out there....
[13:42] <KenMan> gosh, i hope toad is not still struggling with stupid hardware problems , that would suck !!
[13:42] <KenMan> I hate monkeying with hardware. It always takes so long to get things right...
[13:43] <KenMan> however, it was revealed to me that my system only has a few months to live, if I don't do something about the dust in there :(
[13:43] <Pascal666> KenMan: thing actually get to "right" for you? I usually get to good enough :(
[13:43] <KenMan> And I just vaccuumed in there a few months ago...
[13:43] * kers (~kers@3.ppp144.rsd.worldonline.se) Quit ("Leaving")
[13:44] <KenMan> to give you a sense of how bad it was, just 'poofing' with my mouth caused much of the dust to move around...
[13:45] <KenMan> but of course, the finest dust is still clung to all the parts... I had an electrical short on a memory module caused by dust about a year ago.
[13:45] <KenMan> Vaccuuming the chips solved the problem :o
[13:45] <greycat> 72/78 blocks this time.
[13:45] <greycat> NOW do you understand why I despise pcaching?
[13:45] <KenMan> yuck, where did they go ?
[13:45] <greycat> pcaching ate them.
[13:46] <hirvox> pcaching is still alive and well in datastores
[13:46] <KenMan> is this why activelinks on the main page can disappear 10 minutes after I have fetched them ??
[13:47] <KenMan> ( it is not clock rollover )
[13:47] <greycat> I'm assuming it's the same symptom.
[13:48] <greycat> toad always blows it off as "pcaching, and when we tested without it, the overall behavior was worse" kind of arguments
[13:48] * hirvox is no fan of pcaching, either
[13:49] <salahx> well toad was right
[13:49] <salahx> he tunred off pachace for a while and nothing improved
[13:50] <salahx> EVen if turning off pcaching would work; at best it simply hides the awful routing problem
[13:51] <greycat> salahx: then why have I inserted this *same* file over a dozen times and have fewer blocks of it inserted now than I did last time?
[13:51] <salahx> czu routing is broken
[13:51] <greycat> sure, granted, but why don't the blocks that have been inserted *stay inserted* in my own node?
[13:52] <greycat> if I wanted them to go away, they won't....
[13:52] <salahx> well that might be a bug, OTOH, if someone trying to probe datastores and you hace to whole; then...
[13:52] <salahx> pcache only effects download AFAIK; inserts are handeld by a differetn rule
[13:53] <hirvox> salahx:pcaching is applied when we're adding a key into the datastore
[13:53] <hirvox> in StoreData AFAIR
[13:54] <salahx> well last time i checked anything you locally request is ALWAYED checked regardless of pcaching
[13:55] * raged (raged@65.Red-81-35-69.pooles.rima-tde.net) has joined #freenet
[13:56] <Pascal666> pcaching only kicks in if your ds is almost full too
[13:56] <salahx> right
[13:56] <salahx> 90%
[13:59] * TheSeeker (~Fridlekh@cpe-65-28-158-211.socal.rr.com) has joined #Freenet
[13:59] <greycat> my data store has been full for at least a year now
[14:07] <Pascal666> greycat: time to buy a bigger hdd
[14:08] * tyesinclair (~tyesincla@170.252.64.1) has joined #freenet
[14:08] * tyesinclair (~tyesincla@170.252.64.1) has left #freenet
[14:15] * TheSeeker (~Fridlekh@cpe-65-28-158-211.socal.rr.com) Quit (Read error: 60 (Operation timed out))
[14:17] <hirvox> Pascal666:you're joking, right?
[14:19] <Pascal666> hirvox: what?
[14:19] <Ash-Fox> When freenet works, that's when it's broken
[14:22] <Pascal666> bbl
[14:22] * Pascal666 (~Pascal666@m067936d0.tmodns.net) Quit ()
[14:24] * TLF is away: Por ah?, de juerga.
[14:43] * TLF is back (gone 00:18:39)
[14:46] <lolo-laptop> s6 d5dn't d6 that
[14:46] <lolo-laptop> and numlock eats me.
[14:52] <wind789> wow it's taking hours to insert a few megs od stuff at htl 10, does that mean it's working?
[14:52] <salahx> maybe :)
[14:53] <wind789> well I sure hope so, do you think it would have made more sence to insert it twice at htl 5?
[14:53] <salahx> no
[14:54] <salahx> you cna only insert data once
[14:54] <salahx> (well, in theory anyway...)
[14:54] <wind789> local node bypass?
[14:54] <wind789> hehe
[15:00] <wind789> If I'm inserting somthing does my insert get bandwidth prioty over other transfers uploading from my node?
[15:00] <salahx> no
[15:00] <wind789> sheesh
[15:01] <salahx> tohguh kerrp in mind mosdy people can recived you iinsert FAR faster than you can send it
[15:02] <KenMan> about what ??
[15:02] <KenMan> oh...
[15:02] <KenMan> i still have computer maladies :( except now, when the machine locks up... this soundcard holds on the last frequency, like a EEG-flatline noise...
[15:02] <KenMan> instead of bwha-bwha-bwha-bwha-bwha-bwha-bwha-bwha ...
[15:02] <KenMan> damn. I really thought I might have cured this trouble. I think it is an issue with linux NAT. Anyway, a single ping on it's internal NIC brings the machine back to life.
[15:02] <KenMan> Or, if I have the patience, waiting for up to 200 seconds works too...
[15:02] <KenMan> often, when using NAT from an internal computer, this *causes* the lockups to occur...
[15:02] <KenMan> doesn't matter if I use a windows or a linux box through the linux NATter, eventually it goes goofy when I start browsing.
[15:02] <KenMan> oh, and NAT gets held up too. And , well, everything , I suppose. Freenet gets locked , just like all the other apps.
[15:02] <KenMan> okay, maybe i try switching one of the NICs - currently (1) NetGear FA-311 and (1) 3Com 595 vortex . Maybe exchanging the NetGear for another 3Com will solve it.
[15:02] * KenMan is slowly being driven towards running a mainstream distribution like fedora...
[15:02] <KenMan> i'll fight it kicking and screaming, until I am out of breath. Then i will just go run fedora.
[15:02] <KenMan> clearly this is an interrupt problem. When the machine freezes, no keystroke or mouse activity can bring it back. But ping the box, and fwoof! it is instantly alive again.
[15:02] <KenMan> does that sound logical to anyone ??
[15:02] <KenMan> I was running vmstat, and the count of interrupts went to half when this bug kept happening...
[15:02] <KenMan> sounds like a kernel bug to me :o
[15:02] * KenMan has amassed 150,000 keys to play with in his local freenet testnet
[15:02] <KenMan> on UNIX, can 'touch' be used to alter the 'create time' ??
[15:02] <KenMan> ah... the explanation is in 'info touch' :)
[15:02] <KenMan> fred rolls over and plays dead if it detects files in the DS that were created in the future...
[15:03] * KenMan (~chaziller@pcp403292pcs.mntcrm01.md.comcast.net) Quit ("ChatZilla 0.9.61 [Mozilla rv:1.7b/20040324]")
[15:03] <|UK-Monster|> heh
[15:03] * |UK-Monster| is now known as leex
[15:04] * srob99_ (~srob99@203-59-108-109.dyn.iinet.net.au) has joined #freenet
[15:04] * srob99 (~srob99@203-59-84-243.dyn.iinet.net.au) Quit (Read error: 110 (Connection timed out))
[15:09] * KenMan (~chaziller@pcp403292pcs.mntcrm01.md.comcast.net) has joined #freenet
[15:10] * jay (jcl@ool-18bf6dac.dyn.optonline.net) Quit (".")
[15:18] * KenMan suspects he may have fixed his clock problems anyway... but time will tell if this is true !!
[15:19] <KenMan> NTP is now reporting much lower offsets (under 1ms) for hosts that produce 20-50 ms transport delays !
[15:19] <KenMan> i changed how this system reads time, now it says it will use TSC for high res clock functions :)
[15:20] <KenMan> previously I had enabled a PIC that included a timer for CPU generated interrupts...
[15:21] <KenMan> so i turned off local APIC, AGP, DRM, and 2 or 3 other things...
[15:21] <greycat> Ah... the file *finally* inserted this time!
[15:22] <KenMan> patience is a virtue, if you plan to use FreeNet anyway
[15:22] <salahx> waht kernel are you running ?
[15:22] <KenMan> I be running 2.6.7 , now.
[15:22] <salahx> you need the patioce of a corpse on Freenet
[15:23] <greycat> <- Linux 2.4.26 on an Athlon XP 2000+, 512 MB RAM.
[15:23] <KenMan> oh my god. I have a tiny insect crawling INSIDE my LCD screen, behind the pixels !!!
[15:24] <KenMan> I can't squish him !!
[15:24] <greycat> consider it a pet.
[15:24] * KenMan doesn't use drugs , in case some are wondering...
[15:25] <greycat> is it really big, like Admiral Hopper's famous bug?
[15:25] * KenMan began wondering, for a minute there. It is a very disturbing thing. No, he is tiny. Like 4 pixels worth at max. And with a very fuzzy outline (out of focus)
[15:26] * iip_i2p (~changate@host.teitel.net) Quit ("Leaving")
[15:26] <hirvox> hrm, toad forgot to up the build number again
[15:27] <KenMan> this is a trip ! I just hope he doesn't die in there ! that would bring new meaning to a 'dead pixel' ...
[15:28] <KenMan> this little bugger is very active. Like the bee in the 'swarm' xscreensaver, only he doesn't travel that far.
[15:30] <Iakin> [17:34] <thelema> It seems a reasonable idea to have freenet-ext-x86
[15:30] <Iakin> [17:34] <thelema> ,freenet-ext-mac
[15:30] <Iakin> [17:34] <thelema> ,freenet-ext-wierdCPUs
[15:30] <Iakin> [17:35] <thelema> three .jars will cover everyone, and we can still keep the x86 jar (the one that nearly everyone will use) smaller in size.
[15:30] <Iakin> Well, sure.. but how would they be distributed? what would choose the correct version?
[15:30] <hirvox> the user?
[15:30] <Iakin> And what about all the code in there that is plain java?
[15:31] <hirvox> it would be a good idea to put all binaries in a separate jar
[15:31] <hirvox> btw, has anyone managed to compile native FEC lately?
[15:32] * iip_i2p (~changate@host.teitel.net) has joined #freenet
[15:34] <KenMan> what is native FEC ? I built freenet-ext from cvs a few days ago, no major problems...
[15:34] <salahx> is;nt there any way in Java got the architiure of the machine its running on witout have to resort to JNI?
[15:34] <KenMan> this is different than Onion Networks code ?
[15:34] * Pascal (Pascal@c-24-13-17-191.client.comcast.net) has joined #Freenet
[15:34] <Pascal> re
[15:34] <hirvox> KenMan:I mean the onion networks .sos
[15:35] <hirvox> after I tracked down all the dependencies, I got Onion's Java code to compile, but not the native implementations
[15:35] * TLF (francisco@122.Red-81-40-116.pooles.rima-tde.net) Quit ("http://www.vitaliano.esp.cc || Adi?s a todos y a todas y que les vaya bien.")
[15:36] <hirvox> salahx:well, there's os.arch, but it's _very_ inaccurate (as in: x86, powerpc, sparc etc)
[15:36] <KenMan> oh, neat. I didn't realize they had a native lib available.
[15:36] <hirvox> it's been there for ages
[15:37] * kers (~kers@205.ppp141.rsd.worldonline.se) has joined #freenet
[15:37] <hirvox> but I'd like to do it from scratch and make a yet another RPM of fec libraries
[15:38] <hirvox> I could get rid of some double dependencies (for example, the concurrent classes) while I'm at it
[15:39] <KenMan> you know, I'm gonna feel really bad if toad has broken his computer while trying to install an audio card... (it was my suggestion)
[15:40] * interrupt (~chatzilla@64.122.23.213) Quit ("Chatzilla 0.9.64b [Mozilla rv:1.7/20040707]")
[15:40] <hirvox> also, the native libraries do some silly stunts that don't make any sense when installing Freenet properly (read: according to LSB standards)
[15:41] <hirvox> for example, the FEC code extracts the native libraries from the Jar and puts them in ~/.onionnetworks/lib/linux/x86
[15:42] <salahx> acutally the jar is a mess; there stuff everywher in there...
[15:43] <hirvox> agreed, they have onion-fec, which contains the meat, there's onion-common, which contains support classes, concurrent and log4j, and there's the native lib jar
[15:43] <hirvox> our build.xml just stuffs everything from those jars into freenet-ext.jar
[15:43] <KenMan> meat ? onions ? now it's time for lunch...
[15:43] <hirvox> ..and compiles the wrapper classes, of course
[15:45] <hirvox> hm, my build_fred.sh script has ballooned to 12k
[15:46] <wind789> stable CHK@UGqWl--XXSk7g1lFTUWXD9hLsvgLAwI,A5I3Vy4JaAbcLvlJ2pu3qQ/Fenster-Free-Software-Song.ogg
[15:47] <hirvox> but it does export everything from CVS, build both unstable and stable and update both apt and yum repositories with new RPMs. It doesn't slice & dice yet, though :(
[15:49] <iip_i2p> <Sandworm> I can't find the source for the new freenet-ext in snapshots. Is it in the jar? Or elsewhere?
[15:49] <hirvox> sandworm:it's in CVS
[15:50] <Pascal> hirvox: what package name are you using for yum?
[15:50] <hirvox> fred
[15:50] <hirvox> http://nightwatch.mine.nu/freenet/
[15:51] <iip_i2p> <Sandworm> Thanks. Would be nice is somebody could put it up in snapshots. I'm well known for not accessing CVS for anonymity reasons. And I think it belongs into snapshots as well as freenet itself.
[15:52] <Pascal> oh, you arn't putting in the REAL repository
[15:52] <hirvox> Pascal:I have been thinking into submitting it into jpackage, because it's jpackage-compatible
[15:53] <hirvox> but I wanted to make sure that everything works first (yeah right) :)
[15:55] * kers (~kers@205.ppp141.rsd.worldonline.se) Quit ("Leaving")
[15:55] <hirvox> having some test sub^H^H^H^Hers would be nice, though
[15:57] <hirvox> sandworm:are you using stable or unstable?
[15:58] <iip_i2p> <Sandworm> hirvox: stable
[15:59] * greycat (~wooledg@192.35.79.70) Quit (Read error: 104 (Connection reset by peer))
[16:00] * TheSeeker (~Fridlekh@cpe-65-28-158-211.socal.rr.com) has joined #Freenet
[16:00] <hirvox> sandworm:sorry, my computer can only run one node (unstable) at a time :(
[16:01] <Pascal> When Fred's log ate all my disk Frost's boards.xml and knownboards.xml got lost. I copied boards.xml.default to boards.xml to get the default back, but how to I get frost to rescan the boards mailing list to learn about the other? The messages are all still there but no boards shown up in the known list
[16:01] <iip_i2p> <Sandworm> hirvox: ok, thinks making available source tarballs would be a task for toad :). thanks for the offer i read from your thoughts.
[16:02] * greycat (~wooledg@192.35.79.70) has joined #freenet
[16:04] <iip_i2p> <Newsbyte> toad, you there?
[16:04] * kers (~kers@169.ppp135.rsd.worldonline.se) has joined #freenet
[16:16] <iip_i2p> <Newsbyte> toad, you there?
[16:16] <hirvox> bbl, sleep
[16:17] <iip_i2p> <Newsbyte> you coder; no sleep!
[16:18] * Tailchaser (~tailchase@pool-68-162-16-41.nwrk.east.verizon.net) has joined #freenet
[16:22] * TheSeeker (~Fridlekh@cpe-65-28-158-211.socal.rr.com) Quit (Read error: 110 (Connection timed out))
[16:22] * interrupt (~chatzilla@64.122.23.213) has joined #freenet
[16:33] * insider75 (~Administr@pauguste-6-82-66-84-23.fbx.proxad.net) has joined #freenet
[16:34] <insider75> Hi all
[16:35] <insider75> Anyone to chat about the last unstable freenet ?
[16:36] <insider75> I'm currently testing it and can only get in touch with Freenet Help Index, other links cannot be reached.
[16:37] <salahx> hi
[16:37] <salahx> hang omn
[16:37] <salahx> I'm on unstable too
[16:37] <insider75> well what the results ?
[16:37] <salahx> I didn't blow up liek stable did :)
[16:38] <insider75> Me too, my first install of freenet since 0.4
[16:38] <salahx> plenty of RNF's though
[16:38] <insider75> Yes.
[16:40] <insider75> I can get the Freedom engine logo but cannot access the website.
[16:44] <insider75> He He, just get Yoyo working !!!
[16:46] <wind789> stable CHK@EtRYbTd71THRf-ElPO-BW4uFCRQMAwI,d3KCuwzSgC02n8~SwqNJ2Q/gnu.ogg
[16:48] <insider75> Slowly I can get some links, now Content of Evil is working (He just testing ;-)
[16:54] <Pascal> can I limit the size of freenet.log ?
[16:55] <Pascal> surely linux has a way..
[16:55] * TLF (francisco@186.Red-81-40-116.pooles.rima-tde.net) has joined #freenet
[16:56] <salahx> i think there a setting the freenet.conf file for that
[16:56] <salahx> Dunno if it works though
[16:59] <guido^pe> I am still trying to get Freenet running on my FreeBSD machine and it's like I am running against walls. Is there anyone here who recently has successfully run Freenet stable on FreeBSD?
[17:00] * TLF (francisco@186.Red-81-40-116.pooles.rima-tde.net) Quit ("http://www.vitaliano.esp.cc || Adi?s a todos y a todas y que les vaya bien.")
[17:03] * greycat (~wooledg@192.35.79.70) Quit ("Lick Bush in '04")
[17:03] <insider75> A question, does my seednodes.ref goes updated as long as I visit new sites and get data from peers ?
[17:04] <Pascal> no
[17:04] <Pascal> routing table is stored in different files
[17:06] <insider75> How can I manage to get new peers and get a smoother surfing ?
[17:06] <Pascal> visit new sites and get data from peers ?
[17:07] <Pascal> also help if they can get data from you
[17:07] <insider75> Get to know new nodes (or else I'm missing something BIG ?)
[17:09] <guido^pe> insider75: Your node will automatically acquire and store new noderefs, but the file seednodes.ref won't be touched again after the initial seeding (very first startup)
[17:10] <insider75> Seems like I'm now using 25 nodes while I had only 22 when starting freenet for the firt time. (just read your message: ok)
[17:11] <guido^pe> If you let it run some more, it should be at over 100 noderefs in a couple of hours
[17:11] * proteusguy (proteusguy@dialup-4.154.73.40.Dial1.Atlanta1.Level3.net) has joined #freenet
[17:12] <insider75> guido^pe: alright, let's leave it working a bit :)
[17:13] <KenMan> if toad doesn't come back soon, I think there is something wrong with his 'puter !! :(
[17:13] <KenMan> it will take me several days to decide if I've fixed NTP / system clock issues, but my only remaining likely culprit is kernel preemption...
[17:14] <Pascal> prob went to sleep
[17:14] <KenMan> I still get tiny little hiccups, but not long lockups (so far)...
[17:14] <KenMan> I think the hiccups are the preemption feature I left enabled in the kernel.
[17:15] <insider75> Something a bit annoying, there is no way to get a fresh install of FreeNet on win32, can't get refnode and NodeConfig.exe (error 404). I was obliged to resort to a backup of mine of freenet 0.4. How get new freenet win32 users with a situation like this ?
[17:16] <KenMan> good question. Is the win32 build not up to date ? What (4 digit) version does it claim to be ?
[17:16] <KenMan> it should be at least 5028, and you could update the jar and seednodes.ref after that.
[17:17] <KenMan> 5028 was the last good build we ever had ! I think we should immediately revert to that, and continue development from there :p
[17:17] <insider75> KenMan: I update to unstable starting with a 0.4 but I mean there is no complete freenet version to boostrat on win 32
[17:18] <guido^pe> now, is there really noone here who can advise me on Freenet on FreeBSD?
[17:18] <insider75> If I were a noob I could'nt be running Freenet at this moment !
[17:18] * proteusguy (proteusguy@dialup-4.154.73.40.Dial1.Atlanta1.Level3.net) Quit ("Leaving")
[17:19] <guido^pe> this is depressing...
[17:19] <insider75> guido^pe: sorry don't know BSD. What's the prog ?
[17:19] <insider75> guido: what's the problem ?
[17:20] <KenMan> well, thank goodness you aren't a noob, insider75. Install the windows package available from www.freenetproject.org , then come back and we help you upgrade.
[17:21] <KenMan> i think toad has temporarily quit the project in frustration, while he tries to figure out how to get a replacement computer ...
[17:22] <salahx> naah he's probably hard at work fixing the bug
[17:23] * toad_ (toad@82-32-16-91.cable.ubr03.azte.blueyonder.co.uk) has joined #freenet
[17:23] <insider75> KenMan: the current webinstaller (with or without java) when _run_, try to download two files (RefNode and NodeConfig.exe). But I keep having error 404 when trying and the installation fails miserably ;-)
[17:23] <KenMan> i picture a smoking heap of computer parts, with toad weilding a large hammer over top of it all...
[17:23] <KenMan> oh, he lives !!
[17:23] <KenMan> insider75: that is not good :(
[17:24] <insider75> I think these files are missing in the FN ftp server.
[17:26] <insider75> kenMan: A solution would be to put a zip archive of an installed version of Freenet and use it for bootstrapping freetnet. Then update the freenet.jar and seednodes.ref files by downloading them from unstable for example.
[17:26] <KenMan> hold on, I am looking for a CVS link where you can download NodeConfig.exe from ... yes you are right.
[17:27] <toad_> :)
[17:27] <KenMan> oh crap. The exe's don't live in CVS, only the sources... :(
[17:27] <KenMan> toad_: you live !!
[17:29] * raged (raged@65.Red-81-35-69.pooles.rima-tde.net) Quit (Read error: 54 (Connection reset by peer))
[17:30] <guido^pe> insider75: The problem is that I can't find a JVM that properly runs Freenet. I have tried Sun Java 1.4.2, with this Freenet will eat CPU without end and appearently do nothing besides
[17:30] <toad_> :)
[17:31] <guido^pe> I have tried 1.5 beta, that won't properly install on Freebsd, probably a problem with the linux compatibility layer
[17:31] * Ash-Fox (Hal-9000@aae32.neoplus.adsl.tpnet.pl) Quit (Success)
[17:32] <guido^pe> hm, I could try copying over the Java install from my linux machine...
[17:32] <KenMan> something is wrong with toad. All he can do is smile. Toad, are you okay ??
[17:32] <KenMan> he keeps smiling at bad news !!
[17:32] <toad_> KenMan: something I noticed... the linux QoS code supports network simulation
[17:32] <toad_> guido^pe: for how long?
[17:33] <KenMan> sure... may be useful for testnets !!
[17:33] <toad_> it might take an hour or two for it to settle after the startup
[17:33] <guido^pe> toad_: with do nothing, I mean fproxy and fcp ports never become available, the store directory never appears
[17:33] <KenMan> insider75: can you download http://mywebpages.comcast.net/jkcorson/NodeConfig.exe ??
[17:34] <toad_> guido^pe: ah, ouch
[17:34] <toad_> what are you trying to run it on?
[17:34] <toad_> is it an obscure platform?
[17:34] <guido^pe> FreeBSD 5.2, Sun Java 1.4.2
[17:34] <toad_> hmmm
[17:35] <guido^pe> I wouldn't call FreeBSD obscure
[17:35] <toad_> I would :)
[17:35] <insider75> KenMan: thx a lot but I have alaready NodeConfig.exe (from my freenet 0.4 backup) but you could put a link on the FN website if can.
[17:35] <KenMan> no, I'm not hosting the files, just trying to help out. Does your installer still complain with a NodeConfig.exe put in place ??
[17:36] <guido^pe> toad_: At the end of freenet.log, this line repeats a lot: "Jul 28, 2004 11:53:55 PM (freenet.support.BlockingQueue, PRNG/Yarrow entropy processing thread, DEBUG): Waiting... class java.util.LinkedList:[]"
[17:36] <toad_> guido^pe: hmmm
[17:36] <insider75> KenMan: yes the installer complains because it tries to download files that no more exists (the archive doesn't contain them).
[17:37] <toad_> set logLevelDetail=freenet.support.BlockingQueue:minor
[17:37] <toad_> you don't need the above
[17:37] <toad_> actually...
[17:37] <toad_> wait
[17:37] <KenMan> insider75: oh... gotcha.
[17:37] <toad_> guido^pe: is it producing that continuously?
[17:37] <toad_> or has it stopped?
[17:38] <guido^pe> hm, funny, it used to produce it continuesly, but now it seems to have stopped...
[17:39] <insider75> KenMan: I really should post a mail to the Freenet people explaining in details the problem, they dont seem to be aware of (all are linux freaks ;-) ?
[17:39] <toad_> insider75: what sort of problem?
[17:39] <KenMan> many of are linux freaks, but we have our windows freaks too, they just aren't here at this moment in time
[17:40] <toad_> Maximum memory the JVM will allocate 406,464 KiB
[17:40] <toad_> Memory currently allocated by the JVM 314,220 KiB
[17:40] <toad_> Memory in use 227,126,624 Bytes
[17:40] <toad_> Estimated memory used by logger 98,308,326 Bytes
[17:40] <toad_> Unused allocated memory 94,632,488 Bytes
[17:40] <insider75> Hope they read the logs so I don't have to repeat it.
[17:40] <toad_> woah
[17:40] <toad_> 227MB?!
[17:40] <toad_> why?
[17:40] <guido^pe> now they are coming again, along with java.lang.OutOfMemoryError
[17:40] <KenMan> mazzanet may be sleeping ...
[17:41] <mazzanet> mazzanet may be actually be awake
[17:41] <mazzanet> The time is now Thu Jul 29 07:59:11 2004
[17:41] <toad_> mazzanet: but you're not sure?
[17:41] <toad_> ahhh
[17:41] <KenMan> mazzanet - can you diagnose insider75's troubles ??
[17:41] <insider75> time
[17:41] <guido^pe> toad_: Ah, now it's going off... several minutes after startup the line above comes in several times per second
[17:42] <toad_> guido^pe: okay, change your config as i said above
[17:42] <guido^pe> (and nothing else)
[17:42] <toad_> what's the last NORMAL or ERROR in your log? tail +0f freenet.log | grep "\(NORMAL\)\|\(ERROR\)" ?
[17:42] <mazzanet> toad_: no i'm not really sure... was up till 1am last night...
[17:43] <mazzanet> KenMan: insider75's troubles being....?
[17:43] * toad_ wonders if there's a memory leak,
[17:43] <guido^pe> Jul 28, 2004 11:53:52 PM (freenet.node.Main, main, NORMAL): Starting Freenet (Fred) 0.5 node, build #5087 on JVM Sun Microsystems Inc.:Java HotSpot(TM) Client VM:1.4.2-p6-root_21_jul_2004_00_48
[17:43] * KenMan passes Mazzanet a bowl of *frozen* cereal ...
[17:43] <toad_> (with the exception of the nativebigint shouldn't cache so much issue, which i have fixed..)
[17:44] <KenMan> mazzanet - it looks like a file or two required by the win installer is not present in freenet's snapshots directory... :(
[17:44] <KenMan> NodeConfig.exe for one...
[17:44] <toad_> KenMan: woah
[17:44] <toad_> that's not good
[17:44] <mazzanet> ouch
[17:44] <salahx> The wininstlala problby coudl use a rewrite
[17:44] <KenMan> salahx - SHHH !!
[17:45] <toad_> hehe
[17:45] <guido^pe> toad_: these freenet.support.BlockingQueue lines started coming in in a flood after I saw the first java.lang.OutofMemory. Could that have something to do woth it?
[17:45] <mazzanet> KenMan: it probably would freeze, it's 3C outside
[17:45] <mazzanet> :O
[17:45] <KenMan> any snow ??
[17:45] <mazzanet> no
[17:45] <mazzanet> theres fog
[17:46] <toad_> guido^pe: very likely the OOM causes major problems
[17:46] <mazzanet> [ ] NodeConfig.exe 20-Mar-2004 02:48 70k
[17:46] <mazzanet> it's certainly in snapshots
[17:46] <KenMan> well, just don't forget to brag about life when we are all suffering through our cold winters 8-)
[17:46] * kers (~kers@169.ppp135.rsd.worldonline.se) Quit ("Leaving")
[17:47] <guido^pe> toad_: What should I change to give Java more mem?
[17:47] <KenMan> oh yeah, it sure it. So what is going on, insider ??
[17:47] <insider75> Mazzanet: I've downloaded the sources, will watch (but not now) if the installer links correctly to the ftp or website.
[17:48] <toad_> guido^pe: -Xmx
[17:48] <toad_> or -mx
[17:48] <toad_> on the command line
[17:48] <toad_> in start-freenet.sh
[17:49] <toad_> hmmm
[17:49] <toad_> Iakin didn't commit his new windows jbigi dlls :(
[17:49] <insider75> Mazzanet, KenMan: The simpliest would be for sure to fix the broken links if possible.
[17:49] <pwk_> insider75: how old is the windows installer your using?
[17:49] * KenMan will wait to ask toad about his successes with rearranged hardware(s) and software(s) ...
[17:50] <toad_> KenMan: it works well, so far
[17:50] <toad_> it appears to have been the sound card
[17:50] <insider75> KenMan: july 2002 (ouch)
[17:50] <KenMan> it did not cure me, yet... may have improved things though.
[17:50] <toad_> and I got esound working (except for playing 8kHz files)
[17:50] <mazzanet> umm
[17:50] <KenMan> insider75 - we have one from 2003 available...
[17:50] <mazzanet> download the latest installer from snapshots
[17:51] <mazzanet> it works fine here
[17:51] * toad_ thinks everyone should load the Fishman Papers freesite
[17:51] <guido^pe> How long would Freenet normally run before the store dir appears and fproxy port become available (800 Mhz Via C3 machine)
[17:51] <toad_> make sure it all propagates
[17:51] <toad_> guido^pe: hrrm
[17:51] <insider75> Mazzanet: does the installer completes witohut errors, including the downloading of RefNode and NodeConfig.exe ?
[17:52] <mazzanet> yes
[17:52] <toad_> sadly we don't have native optimizations for the C3 at the moment, this will slow down execution (but not loading)
[17:52] * mazzanet thinks he should install freenet again
[17:52] <toad_> but if you get OutOfMemory errors, all bets are off
[17:52] <toad_> also if you have logLevel=debug, not only does it drastically slow things down, it also increases the likelihood of OOMs :)
[17:52] <toad_> it's not a good idea
[17:52] <guido^pe> in any case, should it be any more than 5 minutes?
[17:52] <toad_> I don't use logLevel=debug
[17:53] <insider75> Mazzanet: the link on webinstaller on the Freenet website links to an installer that fails to complete.
[17:53] <mazzanet> :S
[17:53] <mazzanet> where did you download it from?
[17:53] <iip_i2p> <Sandworm> toad, would be nice to have the source of the new freenet-ext in snapshots
[17:53] <mazzanet> http://www.freenetproject.org/snapshots/freenet-webinstall.exe
[17:54] <insider75> Mazzanet: http://freenet.sourceforge.net/index.php?page=download
[17:54] <toad_> Sandworm: yes, it would...
[17:55] * interrupt (~chatzilla@64.122.23.213) Quit ("Chatzilla 0.9.64b [Mozilla rv:1.7/20040707]")
[17:55] <mazzanet> insider75: are you downloading the java webinstaller?
[17:55] <iip_i2p> <Sandworm> toad: ok. i see :)
[17:56] <mazzanet> is the new freenet-ext on cvs?
[17:56] <toad_> Sandworm: working on it
[17:56] <toad_> should be up soon
[17:56] <insider75> Mazzanet : no I've java installed bit I tried and it also failed. I'm testing your link and then the installer fails to download seednodes.ref. I bypass it and it then fails on NodeConfig.exe
[17:57] <iip_i2p> <Sandworm> toad: thanks!
[17:57] <mazzanet> whats the fail message
[17:57] * jay (jcl@ool-18bf6dac.dyn.optonline.net) has joined #freenet
[17:58] <insider75> Mazzanet: "Download of NodeConfig.exe failed: 'HTTP/1.1 404 Not Found'. Retry ?"
[17:58] <toad_> hmmm
[17:58] <toad_> quite a few RNFs...
[17:58] <mazzanet> ooooookay.....
[17:59] <mazzanet> i have no idea then
[17:59] <mazzanet> its a problem on your end
[17:59] <mazzanet> bbl
[17:59] <toad_> i haven't found any evidence to support the "fast RNFs" theory
[17:59] <insider75> Mazzanet: time for me to sleep. Nice to me them all. Hope to see you soon.
[18:00] * insider75 (~Administr@pauguste-6-82-66-84-23.fbx.proxad.net) has left #freenet
[18:00] <salahx> "fast RNF's" theory ?
[18:00] <toad_> salahx: yeah, mentioned on support
[18:01] <toad_> idea that 5087 RNFs a lot quicker/more easily than 5084 does
[18:01] <toad_> memory usage seems reasonable, not great but reasonable, after changes... hrrm
[18:02] * TheSeeker (~Fridlekh@cpe-65-28-158-211.socal.rr.com) has joined #Freenet
[18:05] <salahx> latest unstable RNF quickly too
[18:06] <toad_> bbiab
[18:23] * TheSeeker (~Fridlekh@cpe-65-28-158-211.socal.rr.com) Quit (Read error: 110 (Connection timed out))
[18:24] * Ash-Fox (Hal-9000@aag33.neoplus.adsl.tpnet.pl) has joined #FreeNET
[18:30] * salahx (salahx@sc1-24.217.174.147.charter-stl.com) Quit ()
[18:45] <jay> downloading the source to JDK 1.4
[18:45] <jay> whee
[18:46] <jay> http://wwws.sun.com/software/communitysource/j2se/java2/download.html
[18:53] <KenMan> yuck, only 40 peers on stable :(
[18:54] <KenMan> jay - it's probably worse than FreeNet, in the sense that you likely need 50 engineers in order to make heads or tails of it all...
[18:58] <verl> anyone got some seeds i can try which will allow me to start my stable node? or is there any others settings to fibble with?
[18:59] <verl> logging set to error, that is the lowest one right?
[19:01] <verl> or just wait for 5088 :)
[19:01] <jay> lowest meaning most information ?
[19:01] <jay> verl: err.. error is the lowest actually in verbage
[19:02] <verl> jay: i want the one which consumes the least amount of memory
[19:07] <jay> verl: dunno if it relates to memory usage
[19:12] * lolo-laptop (~lostlogic@68.251.84.226) Quit ("Client exiting")
[19:12] * guido^pe (~unknown@dsl-213-023-243-072.arcor-ip.net) Quit (Remote closed the connection)
[19:21] * Ash-Fox (Hal-9000@aag33.neoplus.adsl.tpnet.pl) Quit (Client Quit)
[19:23] <KenMan> stable rTCR - is averaging at about 99% (of total routes) over several hours :(
[19:27] * greycat (rfc1413@wooledge.org) has joined #freenet
[19:29] <greycat> Connections open (Inbound/Outbound/Limit) 68 (29/39/200)
[19:29] <greycat> Transfers active (Transmitting/Receiving) 303 (241/62)
[19:29] <greycat> Data waiting to be transferred 37 KiB
[19:29] <greycat> Total amount of data transferred 1,572 MiB
[19:29] <greycat> Connections are dropping, and transfers are growing. This is not a healthy trend. Oh, and did I mention I'm still getting RNF left and right?
[19:33] <KenMan> yes, there seems to be some 'issues' that need to be ironed out !!
[19:34] * Ash-Fox (Hal-9000@aag33.neoplus.adsl.tpnet.pl) has joined #FreeNET
[19:34] <Pascal> I just ran into a candidate for worst job. Door to door Kirby salesman. Knock on peoples doors trying to sell them a $2,000 vacum cleaner/stain remover, and offer to clean a room for them. One of my rooms has a large urine stain in it. I told them to clean it up :)
[19:34] <wind789> stable CHK@EtRYbTd71THRf-ElPO-BW4uFCRQMAwI,d3KCuwzSgC02n8~SwqNJ2Q/gnu.ogg
[19:34] <KenMan> Pascal hehe, good job. Did they get it out ??
[19:34] <KenMan> stable CHK@3I8M9jIukQp82GVnsTay8NKceT0QAwI,1a6zf97dtCqljRR7wqPcJQ
[19:35] <Pascal> KenMan: nope, but they just spent two hours trying :)
[19:35] <wind789> kenman: roger that
[19:35] <KenMan> Pascal - I hope they understood why you didn't purchase their product then...
[19:36] <Pascal> yup, if they (the pros) couldn't do it in 2 hours, this is supposed to be easy for the average person?
[19:36] <KenMan> I hate finding large urine stains on my floors, because I am the only one who lives here :o
[19:36] <Pascal> hahaha
[19:37] <Pascal> I'm always happy when the cat makes a mess, cause her bladder is smaller than the dogs :)
[19:39] <KenMan> I believe that MRIs are to blame for everything being so backed off now on stable (and probably unstable too)... they are now very low compared to what they used to be a month ago.
[19:40] <KenMan> anybody (that has been up 5+ hours) care to share their qph level ??
[19:41] <greycat> Smoothed local mean traffic (queries per hour): 5209.802
[19:41] <KenMan> is that high low or normal for you (compared to say a month ago) ??
[19:41] <greycat> I don't normally look at it, but it's higher than a couple years ago ;)
[19:42] <KenMan> Smoothed local mean traffic (queries per hour): 3991.663
[19:42] <Pascal> where is qph at now?
[19:42] <KenMan> I have about 44 peers right now
[19:42] <greycat> Pascal: network load
[19:42] <Pascal> Smoothed local mean traffic (queries per hour): 3726.397
[19:43] <KenMan> number of peers ?? (see http://localhost:8888/servlet/nodeinfo/networking/ocm?setMode=Connection )
[19:43] <Pascal> Number of distinct nodes connected 75
[19:44] * KenMan is jealous...
[19:44] * greycat doesn't see a "number of distinct nodes connected" field
[19:45] <Pascal> hit more details
[19:45] <Pascal> number I like best: Current probability of a request succeeding 5.1%
[19:45] <greycat> Number of distinct nodes connected66
[19:45] <greycat> Current probability of a request succeeding 1%
[19:46] <Pascal> :)
[19:48] <greycat> Freenet is Really Really Broken right now.
[19:48] <Pascal> right now?
[19:48] <greycat> I can't insert, and I can't retrieve. And since that's the only 2 things Freenet can do, it's not got much left.
[19:48] <Pascal> its been broken for years
[19:48] <jay> can anyone point me to an archive that would have redhat 6.x sources to download?
[19:48] <greycat> I mean "even more broken than usual".
[19:49] <Pascal> sometime I think we'd be better off reverting to the 0.1 codebase
[19:49] <greycat> jay: #redhat?
[19:49] <jay> ay
[19:49] <wind789> kenman: no luck on that stable chk you posted
[19:53] <wind789> scratch that, it just now came in at .4 megs
[19:53] <wind789> .04*
[19:53] * jay (jcl@ool-18bf6dac.dyn.optonline.net) has left #freenet
[20:03] * TheSeeker (~Fridlekh@cpe-65-28-158-211.socal.rr.com) has joined #Freenet
[20:05] <wind789> anybody here listen to .mod files?
[20:05] <wind789> or similar file types
[20:11] <KenMan> wind789 - that's pretty amazing that it made it through to you. I only inserted at HTL=8 !!
[20:11] <KenMan> let's try this again...
[20:11] <wind789> cool
[20:11] * KenMan still hasn't gotten wind's ogg though
[20:11] <wind789> hehe, I'll reinsert
[20:11] <greycat> ***how***?
[20:12] * greycat would LOVE to be able to reinsert FEC splitfiles!
[20:12] <KenMan> i keep getting RNF on insert attempts, prolly cause backoff is so high.
[20:12] <greycat> (Then again, I can't insert them the first time right now.)
[20:12] * Ash-Fox (Hal-9000@aag33.neoplus.adsl.tpnet.pl) Quit (Client Quit)
[20:12] <wind789> doesn't fuqid do it?
[20:12] <wind789> with local node bypass?
[20:12] <greycat> please try it and tell us
[20:13] <wind789> I do it all the time
[20:13] <greycat> (If you can insert files at all right now, you're in better shape than I am.)
[20:13] <lostlogic> I get network is busy on almost every request I try to make...
[20:13] <lostlogic> hate.
[20:13] <KenMan> same here... it's gonna be awhile before qpm drifts back down out of the stratosphere...
[20:13] <greycat> Oh, look! I actually got the index.html page of my own site! It only took about 20 inserts and god-only-knows how many requests!
[20:13] <wind789> fuqid has and check box that says "delete key from local node before inserting"
[20:14] <wind789> a check box*
[20:14] <greycat> wind789: "key"? Singular?
[20:14] <wind789> yeah but I'm guessing it does it too all the chunks
[20:14] <greycat> Fred's freenet.client.cli.Main has a --skipDS option that's supposed to do that same thing, but it DOES NOT WORK for the individual data/check block keys.
[20:15] <greycat> Or at least it didn't work yesterday when I was still able to insert.
[20:15] <wind789> darn
[20:15] <KenMan> 150 queries/minute is about 5-10 times what my node can handle :(
[20:15] <wind789> hehe
[20:15] * Ash-Fox (Hal-9000@aag33.neoplus.adsl.tpnet.pl) has joined #FreeNET
[20:15] <KenMan> therefore, backoff=no. of peers until queries falls back down, way down !
[20:15] <wind789> kenman: you inserting somthing else for me to download?
[20:16] <KenMan> I will, if I get the opportunity. Right now my node is inviting all the other nodes to bomb me with queries I cannot handle.
[20:16] <wind789> lol k
[20:16] <greycat> Mine was rejecting connections and queries this morning, but not now.
[20:17] <KenMan> I'm just telling everyone "RNF" 5000 times an hour.
[20:17] <wind789> I don't know what my node does, I just let it run, I monitor the bandwidth usage to make sure its productive
[20:18] <KenMan> well, the thing with Freenet, it can push a LOT of bytes around, and never really accomplish anything. That is un-productive in my book.
[20:18] <wind789> where do I see queries/minute?
[20:18] <greycat> network load (gives queries/hour)
[20:18] <wind789> well that sucks
[20:18] <KenMan> http://localhost:8888/servlet/nodestatus/diagnostics/graphs?image-type=image/bmp&period=minute&type=1&var=localQueryTraffic
[20:18] <lostlogic> hmm... I'm getting 10kqph or166qpmish with a 60k/s upstream bwlimit, my node is ... SATURATED TO HELL
[20:19] <wind789> max: 153.0
[20:19] <wind789> is that what I'm looking for?
[20:19] * toad_ (toad@82-32-16-91.cable.ubr03.azte.blueyonder.co.uk) Quit (Remote closed the connection)
[20:19] <lostlogic> max:257
[20:19] <greycat> max:132
[20:19] <KenMan> I don't know if more frequent MRI communications is to blame, but the MRI values are much lower than they used to be -> thus more queries
[20:20] <KenMan> perhaps now is the time to explore using "doReserveOutputBandwidthForSuccess=true" - it cuts down the number of queries
[20:21] <KenMan> i have no other suggestions for how to rein this beast back in :(
[20:21] <wind789> what exactly does that setting do?
[20:22] <KenMan> well, basically, it is supposed to reserve enough bandwidth for the case where all requests would succeed
[20:22] * TheSeeker (~Fridlekh@cpe-65-28-158-211.socal.rr.com) Quit (Read error: 110 (Connection timed out))
[20:22] <KenMan> or some percentage (which will be higher that whatever percentage you currently experience)
[20:23] <KenMan> the intention was to lower routedToChoiceRank so that routing could have the chance to be more effective...
[20:23] <wind789> I see
[20:23] <KenMan> it should also lower the 'number of backed off routes' value.
[20:24] <KenMan> (which you can see on the routing table page)
[20:24] * KenMan decides to be the first to give it a stab, perhaps others will follow his lead
[20:24] <lostlogic> KenMan: how to stab it?
[20:25] <KenMan> what are rabbits most vulnerable to ?? (a carrot) so stab it with a carrot.
[20:25] <lostlogic> mumble
[20:25] <KenMan> stop your node, add the line "doReserveOutputBandwidthForSuccess=true" to freenet.conf / freenet.ini , and restart
[20:25] <wind789> I'll give it a shot after the ogg finishes reinserting
[20:26] <lostlogic> contacted nodes: 117 backed off nodes: 75
[20:26] <lostlogic> nodes with successful transfers: 71
[20:26] <KenMan> hey, you are reaching the top of the heap. lostlogic is king of the hill.
[20:26] <lostlogic> Connection Attempts 4297
[20:26] <lostlogic> Successful Connections
[20:27] <lostlogic> 246
[20:27] <wind789> what's your uptime?
[20:27] <lostlogic> 20 hours
[20:27] <wind789> lol aah
[20:28] <lostlogic> crashed 20 hours ago, silly node... FEC a lot still blows it up :-(
[20:28] <wind789> only 11 hours for me
[20:28] <greycat> my node's uptime is 1 day 8 minutes.
[20:28] <lostlogic> damnit, I need to go shopping, but the store is so far away... what to do, what to do...
[20:28] <greycat> Ask me whatever you want to know about it...
[20:28] <KenMan> it tooks me 2+ days to get a whopping 60 peers on 5086... 5087 doesn't seem to be going much better
[20:28] <wind789> how much stuff do you guys have in your DS?
[20:28] <lostlogic> KenMan: what is your output bwlimit? what kind of CPU? you cpu limited?
[20:28] <lostlogic> 30G
[20:29] <greycat> wind789: mine's 27G and has been full for AGES.
[20:29] <greycat> This is a node that's been around for a year or two, except for my 6-month hiatus.
[20:29] <lostlogic> ok well 28GiB, 30GB
[20:29] * lostlogic like greycat.
[20:29] <KenMan> lostlogic - I have retained 10G of DS, after smoothly shaping it. CPU is never a problem here, for whatever reasons (1.4GHz atlon 1600+)
[20:29] <KenMan> lostlogic - hey, shouldn't you be in AL proposing to someone right about now ??
[20:30] <lostlogic> mine's an athlon 1800+ and it hits the cpu limit from time to time
[20:30] <wind789> Maximum size 100 GiB
[20:30] <wind789> Used space 45,674,532 KiB
[20:30] <wind789> Total keys 142403
[20:30] <KenMan> or was it TN ??
[20:30] <lostlogic> KenMan: *kick* I wish. I should be at the store, or hitting on the hot turkish girl at the restaurant down the street, while not thinking about AL
[20:30] <lostlogic> KenMan: you have a very good memory, repeatedly impresses me
[20:31] <KenMan> well, i secretly purveyed all your digital photos, saw your car too...
[20:31] <lostlogic> I love my car
[20:31] <KenMan> well, ok, not ALL of them, but a lot. Sometimes I have lots of time and nothign to do...
[20:31] <KenMan> It looked pretty nice.
[20:31] <lostlogic> KenMan: well I'm glad some people look at them... I like taking them :)
[20:32] <wind789> Least recent file access time Thu Feb 12 19:01:10 EST 2004
[20:32] <KenMan> so you aren't going to attempt to marry Heather ? Does saying "no!" put things in better perspective then ??
[20:32] <lostlogic> KenMan: oh, who knows... we'll see if she comes around, until then, I have to come up with something else to do with myself than pine over her.
[20:33] <KenMan> good enough.
[20:33] <KenMan> but don't wait too long if you are actually considering such a thing.
[20:33] <lostlogic> KenMan: was it you who told me to take down heatherminor.com? I did.
[20:33] <lostlogic> KenMan: I shan't, fear not.
[20:33] <KenMan> I think it was me.
[20:34] <lostlogic> wind789: nice DS size! Rock on!
[20:34] <KenMan> so, in the meantime, what is your diagnosis of freenet ??
[20:34] <lostlogic> wind789: still filling up, I see
[20:34] <wind789> indeed
[20:34] <wind789> I don't like data falling off
[20:34] <lostlogic> KenMan: it is coming along. requests that are not restarted have a chance of success which was unheard of a few short months ago.
[20:34] <KenMan> yeah, damn . 100G s ...
[20:35] <greycat> 100G is a very big data store. Thank you for being so generous.
[20:35] <wind789> hehe bought this sucker just for freenet
[20:35] <lostlogic> KenMan: CPU usage could still use work (as always) and I wish there was a way to prioritize local requests over requests we are routing.
[20:35] * Anvil_Vapre (~Cichli@63-227-158-112.omah.qwest.net) has joined #freenet
[20:36] <oierw> lolo: would that make things any better?
[20:36] <wind789> you guys remeber the datastore bug that would randomly strike?
[20:36] <oierw> i don't think our own requests are getting stuck somehow
[20:36] <KenMan> actually, I've given that some thought. In my bold opinion, local queries should not get more than 1/n+1 of query capacity (for n peers)
[20:36] <oierw> wind789: which one
[20:37] <wind789> I don't know, but it was pretty funny, some dude even made a fictional fresite about it
[20:37] <wind789> it would wide the DS clean I think
[20:37] <wind789> I think that was back in the monlithic days
[20:37] <oierw> okay. one of the monolithic ds bugs then :)
[20:37] <lostlogic> oierw: I'm not sure... but I do think that local requests might need to get different treatment than other requests. KenMan's idea sounds pretty good, because it would be another way to keep the network loaded only so much as it can handle...
[20:38] <oierw> i remember a certain ds bug that wiped out someone's mp3 collection :p
[20:38] <wind789> lol
[20:38] <oierw> and I swear to god that I had nothing to do with it
[20:38] * oierw sneaks away
[20:38] <KenMan> hey, this doReserveBW thing is already having an effect, very early on !! a good effect.
[20:39] <wind789> cool
[20:39] <greycat> oierw: hey, that wasn't a DS bug. That was an installer "feature".
[20:39] * lostlogic goes to the store, remind me to try it when I get back.
[20:39] <KenMan> it has caused my outgoingMRI to reach the limit of 600s , yet queries are still way high. Someone out there must be ignoring/abusing mRI.
[20:39] <wind789> if you wanna download a mp3 that has been inserted about 20 times on stable try CHK@bL35x8zPm6Arup2hbeMoD1sRjygTAwI,uaQeOnu1lA0SQV4hROhnpg/Richard_Stallman-Free_Software.mp3
[20:40] <oierw> greycat: sure, that one was
[20:40] <greycat> Windows installer asked the user where to put the data store, the user said something like "F:\MP3s", Freenet nukedi t.
[20:40] <oierw> but there was the same bug in fred
[20:40] <oierw> or one very similar that also wiped some files (not all of them though, i think)
[20:40] * KenMan heard the tale told once before, but wasn't around when DS was monolithic
[20:41] <wind789> anybody around in the .3 days?
[20:42] <oierw> i was
[20:42] <oierw> .2 days too
[20:42] <wind789> that's when I started using freenet but I don't remeber much about it
[20:42] <wind789> wow
[20:42] <KenMan> someone out there is definitely cheating the MRI scheme, without a doubt... let's peruse the logs...
[20:42] <wind789> lol
[20:42] <greycat> I started in the late 0.3 days
[20:42] <oierw> back int eh days when you could actaully go through your store and find the porn
[20:42] <wind789> sweet
[20:42] <KenMan> me started in around 5012... (0.5)
[20:42] <wind789> hehe
[20:43] <oierw> i signed up for freenet-dev on... april 19th 1999.
[20:43] <wind789> dang
[20:43] <oierw> only reason I remember that is because it's 2 days before my birthday (and I looked before)
[20:43] <wind789> hehe
[20:43] <oierw> i've since lost my email logs up to 3/23/2002
[20:44] <oierw> stupid mailbox corruption
[20:44] <wind789> imagine if you had a huge DS that had the keys from back then
[20:45] <KenMan> oh hell, it looks like 5084 nodes dpm respect the MRI from 5087 :( :(
[20:45] <KenMan> ts/dpm/don't
[20:45] <wind789> crazy
[20:45] <KenMan> there, i diagnosed what needs fixing in 5087. Now I must go rest for a little while.
[20:45] <wind789> hehe
[20:45] <KenMan> i'm serious
[20:46] <wind789> keep trying the gnu.ogg it's almost don reinserting
[20:46] <KenMan> i don't do actual coding fixes around here, I try to earn them with suggestions backed up with charts and grandiose descriptions...
[20:46] <KenMan> ;)
[20:46] * jay (jcl@ool-18bf6dac.dyn.optonline.net) has joined #freenet
[20:46] <wind789> hehe cool
[20:47] <wind789> I gotta go shop around for some gardening supplies myself, I will bbl
[20:47] <KenMan> grep your logs for the word 'violated' , and then report what version of fred is reported for each of the violators...
[20:48] <wind789> lol grep is hard to find in windows so I'll sit this one out
[20:48] <KenMan> see you soon wind...
[20:48] <wind789> cya
[20:48] <oierw> i've got you beat. I haven't done anythign that actually helps freenet (in any computer respect other than trying to help a user here or there) for 2 years
[20:48] * KenMan says "you can trust me"
[20:48] <oierw> err... a year atleast
[20:48] <greycat> freenet@griffon:~$ grep violated freenet.log | wc -l
[20:48] <greycat> 12105
[20:48] <greycat> freenet@griffon:~$ grep violated freenet.log | grep -v 'STABLE-1.50,5084' | wc -l
[20:48] <greycat> 0
[20:49] <KenMan> just pick out the actual version numbers and give us soem counts, like "5084" "5085" etc...
[20:49] <greycat> 5084: 12105. Everything else: 0.
[20:50] <greycat> (Sorry if I was too obscure.)
[20:50] <KenMan> excellent, so either the abusers are 5084 based, or my estimate is correct :) I ignored the '-v' in my haste. Sorry.
[20:50] <KenMan> estimate being - 5087 and 5084 just don't play nice together.
[20:52] <KenMan> time to up lastGoodBuild... it is nice when you can spot a definite problem without banging your forehead on all this 'sharp' code...
[20:53] <KenMan> it seems likely, as toad has twice just changed the way MRI's are communicated.
[20:56] * Hory (~Miranda@81.196.25.110) Quit ("Miranda IM! Smaller, Faster, Easier. http://miranda-im.org")
[20:57] * KenMan downloads the stable src, sets lastGoodBuild=5087, rebuilds and restarts. Oh, and he ALWAYS crosses his fingers.
[20:57] <greycat> doesn't that make it hard to type?
[20:58] <KenMan> yes, but the spirits protect me !
[21:00] * KenMan feels the need to hear Stevie Wonder's "Superstition" , but alas, he does not have it
[21:00] * greycat has it.
[21:05] <KenMan> i bet toad has crawled onto his lilypad for some shut eyes by now...
[21:05] <greycat> quite likely
[21:06] <oierw> just change your own lastGoodBuild
[21:08] <KenMan> done and done. and fingers still crossed.
[21:08] <KenMan> 19 peers already. Not bad.
[21:10] <KenMan> so there's no need to leverage that doReserveBW thingy against anyone just yet ...
[21:11] <lostlogic> man, I'm a newb... didn't show up till just before .5
[21:12] <KenMan> you were here before me, and already productive !! so there :p
[21:12] <oierw> you were definelt yone of the more productive members of the people who have hung out in #freenet
[21:12] <KenMan> you were hoping toad would just shrivel up at the time ;)
[21:13] <KenMan> "we've never got more work done than in the week toad was away on vacation" was the quote in my mind. How's THAT for memory, lolo ?
[21:14] <KenMan> don't worry, it can't be held against you (statute of limitations, plus good behavior)
[21:15] <lostlogic> *chuckle*
[21:15] <lostlogic> toad'll kick my ass...
[21:15] <KenMan> :)
[21:15] <lostlogic> but that shit happens a lot in projects ... not sure exactly why...
[21:15] <KenMan> he will only get upset if you don't contribute any more code.
[21:15] <lostlogic> when my boss at my job leaves, we get a ton more coding done...
[21:16] <lostlogic> heh, I'll uhh... do my best?
[21:16] <KenMan> bbl - go fix yer node with lastGoodBuild=5087.
[21:16] <lostlogic> no... actually, I'll just hang out here and be semi-useful from time-to-time
[21:16] <KenMan> like me ?
[21:16] <greycat> http://localhost:8888/servlet/nodestatus/diagnostics/graphs?image-type=image/bmp&period=minute&type=1&var=localQueryTraffic -> max=33.0
[21:16] <greycat> that's bad, right?
[21:16] <KenMan> no, not at all.
[21:17] <KenMan> 33 * 60 = 1800 qph, right ? that's not so bad.
[21:17] <KenMan> but if it wiggles all over the radar, like a liar attached to a polyscope, then that IS bad.
[21:17] <greycat> it is doing that.
[21:17] <KenMan> i know.
[21:18] * greycat looks around for the hidden cameras ;-)
[21:18] <KenMan> it is the broken MRI thingy :)
[21:19] <KenMan> wow, it sure was that. Now that I'm fixed up, my qpm is as smooth as a youngster's elbow.
[21:19] <greycat> hrm. I should start CVS-ing and ant-ing again.
[21:19] * Plato986 (~Silfone@ip68-0-21-119.hr.hr.cox.net) has joined #freenet
[21:19] * KenMan is cognizant of the potential presence of kiddy molesters, and thus avoided the usual term
[21:19] <lostlogic> KenMan: LoL
[21:20] <KenMan> unfortunately, i'm half serious...
[21:20] <lostlogic> KenMan: I know... you ever been on IIP? There nearly everyone is into underaged nude girls or so the conversation would seem to indicate.
[21:20] <Plato986> Does anyone know how to remove vocals from an MP3?
[21:20] <lostlogic> s/girls/people/
[21:21] <greycat> griffon:/usr/src/freenet$ ls -lart
[21:21] <greycat> drwxr-sr-x 11 greg greg 4096 2003-08-24 08:59 freenet/
[21:21] <greycat> about 11 months old :)
[21:21] <KenMan> actually, i need a lot more time and data to make this determination, but so far it looks good. No, I never been in IIP, i tried to once, probably just as well.
[21:21] <Plato986> Anyone?
[21:22] <greycat> Plato986: ask an audio channel?
[21:22] <lostlogic> Plato986: use a high/low pass filter to strip out the range of the human voice? or, use a bandpass filter to extract just the voice and the xor the result with the original...
[21:23] <greycat> of course you'd decode the mp3 first, then reencode when done. Which introduces another round of lossy encoding.
[21:23] <greycat> anyway, #hydrogenaudio would be a better starting point
[21:23] <Plato986> Thanks.
[21:23] <mikeDOTd> how do i send 3mB of data to a node who's address is "no addresses"?
[21:23] <KenMan> send out 200 mB of data, some of it will hit it.
[21:23] <Plato986> What server is that on grey?
[21:23] <lostlogic> mikeDOTd: it connected to you, didn't tell you it's address, but still you have the connection?
[21:24] <greycat> Plato986: is that a trick question?
[21:24] <mikeDOTd> lostlogic: wouldn't my node know the src address?
[21:24] <KenMan> yeah, but it's not telling :)
[21:24] <lostlogic> mikeDOTd: sure, your computer knows, but your node might not have asked your computer... *shrug* don't know where or when it calc.s the data for that connections page
[21:25] <greycat> griffon:/usr/src/freenet/freenet$ cvs -z3 update -dP <-- seems hung. connection timed out.
[21:25] <Plato986> Yeah grey, just kidding. Hehe. It's not that I'm new to mirc, and don't know what I'm doing or anything like that. ;)
[21:25] * greycat wonders whether cvs.freenet.sourceforge.net is still correct
[21:25] <lostlogic> greycat: that's what I was going to say.
[21:25] <KenMan> nope
[21:25] <lostlogic> greycat: change to cvs.sf.net
[21:26] * greycat wonders why that resolved and timed out, rather than just NXDOMAIN'ing.
[21:27] <greycat> is jikes 1.14 still considered "good"?
[21:27] * lostlogic decides to put latest unstable code on stable network with LKGB 5087
[21:27] <mikeDOTd> it seems broken behavor to me that i have active connections established for 15 minutes, that the connection page doesn't know how to display
[21:29] <greycat> U NativeBigInteger/gmp-4.1.3.tar.gz
[21:29] <greycat> huh?
[21:29] <lostlogic> greycat: huh what?
[21:29] <lostlogic> greycat: it's the source tarball for gmp
[21:29] <greycat> Why did they stuff a whole GPL'ed or LGPL'ed source tarball in the CVS?
[21:29] * Plato986 (~Silfone@ip68-0-21-119.hr.hr.cox.net) has left #freenet
[21:30] * jay (jcl@ool-18bf6dac.dyn.optonline.net) Quit ("Client exiting")
[21:30] * jay (jcl@ool-18bf6dac.dyn.optonline.net) has joined #freenet
[21:30] <jay> ugh
[21:30] <jay> compiling sun's jdk is a pain in the ass
[21:30] <lostlogic> greycat: you are asking the wrong people on that... Iakin or toad are your people
[21:30] <jay> maybe this time gcc wil compile
[21:31] <KenMan> Iakin is my guess
[21:31] * Lux (~Lux@82-37-17-24.cable.ubr03.wolv.blueyonder.co.uk) has joined #freenet
[21:31] <lostlogic> hey, Lux
[21:31] <Lux> My nickname is owned by someone else grrr
[21:31] <greycat> "ant jar" failed for me with jikes 1.14.
[21:31] <Lux> anyone use gcj at all round here?
[21:31] <greycat> [javac] *** Warning: "-source" is an invalid option; use: jikes [-bootclasspath path][....]
[21:32] <lostlogic> Lux: it doesn't build freenet, sadly
[21:32] <Lux> kk
[21:33] <KenMan> broken MRIs are destroying the stable network. Oh well, nothing out of the ordinary... wonder what the recovery time will be this time ??
[21:33] <greycat> KenMan: which compiler are you using to build freenet these days?
[21:34] <KenMan> build 1.4.2_03-b02, mixed mode
[21:34] <KenMan> Sun's JDK
[21:34] <greycat> what's that map to in build.xml?
[21:35] <KenMan> crap, it doesn't. Ant looks for $JAVA_HOME, I think.
[21:35] <greycat> <property name="build.compiler" value="jikes"/>
[21:36] <KenMan> shoot, I don't know enough ... i just run ant and it seems to work. Although I had some things to work out back when... then i promptly forgot whatever i learned about ant.
[21:37] <Lux> cant y'all just go back to makefiles :p
[21:37] <greycat> Ugh! My javac is pointing to /usr/lib/kaffe/bin/javac
[21:37] <lostlogic> Lux, there is a makefile in there for gcj, so if freenet ever builds on classpath again ...
[21:37] <KenMan> property name="build.compiler" value="modern"/ is the default. That works for me.
[21:38] <KenMan> ... we will fix up the makefile to make it work again
[21:38] <greycat> And I only have j2re1.4.2 installed, not a jdk.
[21:38] <KenMan> pisser !
[21:38] <greycat> let's see what happens if i install sid's jikes (1.15-2)
[21:39] <greycat> oops, no, it's 1.21-2
[21:40] <greycat> Um... well, it was *noisy*, whatever it was.
[21:40] <Lux> I like blackdown though it is really only Sun in the end
[21:40] <lostlogic> Lux: yeah, but they are too far behind right now...
[21:41] <Lux> they got java working on my Psion 5mx tho :p
[21:42] <greycat> jikes 1.21 gives me a metric shitload of "Semantic Warning" and "Semantic Error" lines
[21:42] <lostlogic> YAYA with LGB set like this, I get DNFs instead of RNFs finally
[21:43] <lostlogic> greycat: is a metric shitload more or less than a metric fuckton?
[21:43] <mikeDOTd> LGB?
[21:43] <greycat> lostlogic: it came very close to filling up my rxvt's scrollback buffer completely.
[21:43] <greycat> mikeDOTd: last good build
[21:43] <Lux> When in doubt... use Less :D
[21:43] <greycat> lostlogic: rxvt*saveLines: 2000
[21:44] <greycat> so, approximately 1950 lines of errors/warnings
[21:44] <lostlogic> greycat: yikes stripes!
[21:44] <Lux> cat /dev/zero | nc www.microsoft.com 80 :D
[21:44] <greycat> Lux: /dev/urandom might be more fun
[21:44] <Lux> its too slow
[21:45] <lostlogic> Lux: not on 2.6 with freenet running :):)
[21:45] <greycat> I'm sure /dev/urandom can saturate my uplink
[21:45] <Lux> so i can just emerge
[21:45] <Lux> or should i get the latest from somewhere else? ll?
[21:45] <lostlogic> Lux: emerge the package and then use the ebuild's update functionality...
[21:45] <lostlogic> Lux: I made it :-P
[21:45] <Lux> kk
[21:45] <Lux> ahh its 3am
[21:45] <Lux> slocate's kicked in
[21:46] <lostlogic> 9:00pm :-P
[21:46] <Lux> well im off to bed soon anyways
[21:48] <greycat> The first block of Semantic Warnings are in /usr/src/freenet/freenet/src/freenet/support/FileLoggerHook.java about the missing "break" statements in a big switch, but the code looks OK to me. The missing breaks appear to be intentional (despite a lack of comment indicating that We Know What We're About).
[21:48] <Lux> should i try stable or unstable?
[21:48] * Hadaka (naked@naked.iki.fi) Quit (Read error: 110 (Connection timed out))
[21:49] <greycat> [javac] *** Semantic Warning: Final field "maxSleepTime" is initialized with a constant expression and could be made static to save space. <-- looks ignore-able :)
[21:49] <lostlogic> Lux: hmm... up to you, stable will rape your connection right now, because of a bug, but unstable is a smaller network so not as much content to browse
[21:49] <Lux> so ur saying i should just wait?
[21:49] <Lux> :p
[21:49] <greycat> [javac] *** Semantic Warning: Local "hooks" shadows a field of the same name in "freenet.support.LoggerHookChain". <-- Lots of these. May be worth looking at.
[21:50] <lostlogic> Lux: in the next couple days, stable will sane up... the bug is known, kenman found out why, and toad'll deal with it one way or another soon
[21:50] <Lux> o its all crap! scrap it and start over :p
[21:50] <Lux> (just kidding)
[21:50] * lostlogic mutters something about inviting lux to this channel in the first place
[21:51] <Lux> OT: anyone know any good free ipv4->ipv6 tunnelers?
[21:51] <greycat> Another set of "missing break" warnings in /usr/src/freenet/freenet/src/freenet/config/Option.java for the switch(i) starting at line 84. This one may be a real bug. I can't tell at a glance.
[21:52] <greycat> switch (i) {
[21:52] <greycat> case 2:
[21:52] <greycat> r ^= b[i-2];
[21:52] <greycat> case 1:
[21:52] <greycat> r ^= b[i-1] << 8;
[21:52] <greycat> case 0:
[21:52] <greycat> r ^= b[i];
[21:52] <greycat> }
[21:54] <greycat> [javac] *** Semantic Caution: Overflow in int expression. (/usr/src/freenet/freenet/src/freenet/config/Params.java line 316)
[21:54] <Lux> does freenet require UDP port open too?
[21:54] <greycat> Lux: no, it's all TCP
[21:54] <Lux> kk t
[21:54] <Lux> x
[21:55] <greycat> [javac] *** Semantic Caution: Overflow in int expression. (ibid. line 317)
[21:55] <greycat> Err. I meant "[javac] *** Semantic Caution: The shift count of 40 is >= the 32-bit width of the type."
[21:55] * Naked (naked@naked.iki.fi) has joined #freenet
[21:55] * Naked is now known as Hadaka
[21:56] <greycat> Both of those seem a bit weird, because they're supposed to be initializing a long[] array.
[21:57] <greycat> Either jikes is complaining without just cause, or java requires some sort of explicit "This is a long" syntax in array initializers. I don't know Java.
[22:00] <greycat> "missing break" warning on /usr/src/freenet/freenet/src/freenet/node/states/request/TransferReply.java line 288-295, I can't tell whether it's intended or not.
[22:00] <greycat> ditto 297-300
[22:01] <jay> ./quit
[22:01] * jay (jcl@ool-18bf6dac.dyn.optonline.net) Quit ("Client exiting")
[22:01] <greycat> ditto /usr/src/freenet/freenet/src/freenet/node/states/request/SendingReply.java 89-91
[22:02] <greycat> ditto /usr/src/freenet/freenet/src/freenet/node/states/request/TransferInsert.java 364-373
[22:03] <greycat> ditto 374-387 same file
[22:03] <oierw> i'd really like to see a c or python port of freenet
[22:03] <oierw> or rather, c++
[22:03] * TheSeeker (~Fridlekh@cpe-65-28-158-211.socal.rr.com) has joined #Freenet
[22:04] <greycat> It's worth stating that there are other "missing break" warnings that I looked at and which *do* have "// fall through" type comments on them, so I didn't report them.
[22:04] <Lux> get it working with GCJ :P and at least u could have a binary version
[22:05] <greycat> "missing break" in /usr/src/freenet/freenet/src/freenet/crypt/ciphers/Twofish_Algorithm.java lines 416-420 and 421-425. That's *scary*.
[22:05] * Ash-Fox (Hal-9000@aag33.neoplus.adsl.tpnet.pl) Quit (Read error: 104 (Connection reset by peer))
[22:06] <oierw> Lux: gcj is probaly more work than redoing it in c++ :)
[22:06] * greycat stops for tonight.
[22:07] <lostlogic> oierw: well we just need to hope that kafe and gcc get classpath up to java 1.4 compliance, then gcj is a go for no work on our part
[22:07] <Lux> hehe
[22:07] <oierw> assuming it works like it should
[22:07] <Lux> just re-write it in perl :p
[22:07] <oierw> i want some maintainability
[22:07] * greycat beats Lux to death with a rabbit
[22:07] <Lux> hehe
[22:07] <Lux> ok PHP! im sure it could be done (some how)
[22:08] <greycat> nobody ever said it was impossible. Just... not very sensible.
[22:08] <Lux> lol
[22:08] <oierw> i'm getting sick of vm's
[22:08] <Lux> hehe
[22:08] <oierw> although I know that python has one...
[22:08] <Lux> ever done ps -A h
[22:08] <Lux> with freenet running
[22:08] <Lux> lol
[22:08] <Lux> why so many soddin' threads!
[22:08] * Ash-Fox (Hal-9000@aag33.neoplus.adsl.tpnet.pl) has joined #FreeNET
[22:08] <greycat> you should've seen it *before* NIO.
[22:09] <Lux> hehe
[22:10] <lostlogic> should have seen it soon after NIO ... thread flux in the range of 40k/hour
[22:10] <Lux> lol
[22:10] <thelema> <Iakin> Well, sure.. but how would they be distributed? what would choose the correct version?
[22:11] <thelema> The x86 version would come with the node. If it detected a non-x86 processor, it'd give instructions (mainly the URL) to download the right one.
[22:13] <greycat> OK, before I go to bed: the actual *errors* that caused jikes to stop seem to be of the form [javac] *** Semantic Error: The static type "freenet.client.http.filter.SaferFilter$ParsedTag" must use a qualified name to access the non-static member type "freenet.client.http.filter.SaferFilter$HTMLParseContext" of the enclosing type "freenet.client.http.filter.SaferFilter".
[22:14] <greycat> There are 23 of those in /usr/src/freenet/freenet/src/freenet/client/http/filter/SaferFilter.java, and that particular one is at line 580.
[22:15] * greycat (rfc1413@wooledge.org) Quit ("Lick Bush in '04")
[22:17] * Ash-Fox (Hal-9000@aag33.neoplus.adsl.tpnet.pl) Quit (Client Quit)
[22:18] * shendaras (~shendaras@ip68-97-113-208.ok.ok.cox.net) has joined #freenet
[22:18] * thelema is now known as thelema|away
[22:19] * Lux (~Lux@82-37-17-24.cable.ubr03.wolv.blueyonder.co.uk) Quit ("Leaving")
[22:21] * Ash-Fox (Hal-9000@aag33.neoplus.adsl.tpnet.pl) has joined #FreeNET
[22:24] * salahx (salahx@sc1-24.217.174.147.charter-stl.com) has joined #freenet
[22:25] <lostlogic> has anyone spent enough time collecting node refs to have a good estimate of the stable network size?
[22:28] <wind789> not really, but I can say that I've contacted 105 nodes in 13 hours if that helps
[22:29] <wind789> unique contacts: 718 inbound
[22:29] <lostlogic> my node know 389 nodes in 19 hours
[22:29] <wind789> unique contacts: 412 outbound
[22:29] <wind789> lol I've got the same number
[22:29] <wind789> Number of known routing nodes 389
[22:30] <wind789> probly has somthing to do with that hude seednodes.ref file
[22:30] <lostlogic> if the stable network is only 389 nodes, I'll kill someone.
[22:30] <lostlogic> :-P
[22:30] <lostlogic> where did you get unique contacts from?
[22:31] <wind789> http://127.0.0.1:8888/servlet/nodestatus/inboundContacts.txt
[22:31] <wind789> http://127.0.0.1:8888/servlet/nodestatus/outboundContacts.txt
[22:31] <lostlogic> oh, I don't have that on...
[22:31] <wind789> the links are on http://127.0.0.1:8888/servlet/nodestatus/
[22:32] * TheSeeker (~Fridlekh@cpe-65-28-158-211.socal.rr.com) Quit (Read error: 110 (Connection timed out))
[22:33] <salahx> i have it on
[22:33] <wind789> hehe
[22:33] <salahx> unique contacts: 387 (This is on unstable)
[22:34] <wind789> in or out?
[22:35] <salahx> out
[22:35] <wind789> big network
[22:36] <wind789> I wonder if stable and unstable are about the same size
[22:36] <salahx> i doubt it
[22:36] <salahx> Unstable is a lot smaller
[22:36] <wind789> how many inbound contacts do you have?
[22:37] <salahx> unique contacts: 196
[22:37] <wind789> aahh
[22:38] <wind789> did they do away with the whole transient node option?
[22:38] <salahx> yep
[22:38] <salahx> With bidi, ther are no more "traniets"
[22:38] <wind789> does freenet work on dial up?
[22:39] <mikeDOTd> freenet will "work" on dial up
[22:39] <wind789> I can't see how a permanent node on dial up can benifit the user or the network
[22:39] <mikeDOTd> the user can access freenet, that's a benefit
[22:39] <wind789> won't it swamp the users connection?
[22:40] <mikeDOTd> i'm sure it will :)
[22:40] <wind789> hehe
[22:40] <mikeDOTd> but you could rate limit it
[22:40] <wind789> the lucky downloadeer will be there for hours if it's a big key..
[22:41] <wind789> being sent form the limited dial up user
[22:42] <wind789> not hours but maybe 20+ minutes
[22:42] <wind789> on a 1 meg file
[22:42] <wind789> assuming he's teh only person downloing from the the dial up sender
[22:42] <wind789> sheesh
[22:45] * KenMan (~chaziller@pcp403292pcs.mntcrm01.md.comcast.net) Quit (Remote closed the connection)
[22:45] * KenMan (~chaziller@pcp403292pcs.mntcrm01.md.comcast.net) has joined #freenet
[22:45] <wind789> yo kenman, you there?
[22:45] <KenMan> oops. Didn't mean to waste my backlog...
[22:46] <wind789> anyluck on the gnu.ogg file?
[22:46] <KenMan> please share the CHK again, it looks like I'm now functional on stable.
[22:46] <wind789> CHK@EtRYbTd71THRf-ElPO-BW4uFCRQMAwI,d3KCuwzSgC02n8~SwqNJ2Q/gnu.ogg
[22:47] <wind789> stable...
[22:47] <KenMan> but of course ;)
[22:47] <wind789> hehe
[22:48] <KenMan> backoff is very reasonable now... sane query levels take time to settle, but they are "below the stratosphere" after 110 mins...
[22:48] <wind789> cool
[22:49] <lostlogic> splitfile came right up on that
[22:49] * salahx not geting it but I'm on unstable
[22:49] <KenMan> cool. I got a DNF, not an RNF on that ogg... retrying...
[22:49] <wind789> lol wtf
[22:49] <wind789> nevermind thougb you said getting it
[22:50] <lostlogic> shit, node just died on me
[22:50] <wind789> for unstable download CHK@bL35x8zPm6Arup2hbeMoD1sRjygTAwI,uaQeOnu1lA0SQV4hROhnpg/Richard_Stallman-Free_Software.mp3
[22:50] <wind789> and if you're feeling really luck on unstable download CHK@FwDwUWZKHDd3IeAzR95gCSifKVwMAwI,URWySMGDJK2BUbFK4s1JkQ/Richard_Stallman-Free_Software(remix).mp3
[22:52] <wind789> I'm reinserting the header file a couple times, I'll finish it off with one large htl insert of the header file
[22:52] <KenMan> lolo ?
[22:52] <lostlogic> ?
[22:52] <KenMan> death by taxes, or what ?
[22:52] <lostlogic> I can crash a node by doing splitfile downloads pretty reliably these days, but haven't found anything useful about why or how to report
[22:53] <wind789> that's unfortunate
[22:53] <KenMan> :( - I've never tried to do a splitfile insert, i don't think...
[22:53] <lostlogic> not inserts, just dls
[22:53] <salahx> Inserted Knoppix into usntable
[22:53] <salahx> CHK@4PQbOTcoW2KlKpQWZmxKHnlSn7oRAwI,TFGuhFRsHjzZqiB2UPftrw/KNOPPIX_V3.4-2004-05-17-EN.iso
[22:53] <KenMan> salahx - how did it go ? a little over a day's time ?
[22:53] <salahx> 3 days
[22:54] <KenMan> ouch... patience is for the dead
[22:55] * shendaras (~shendaras@ip68-97-113-208.ok.ok.cox.net) Quit ("Boom")
[22:55] <wind789> what htl are you trying to download at?
[22:55] <KenMan> me ? whatever the fproxy default is, 15 probably.
[22:55] <wind789> aahh
[22:55] <wind789> sounds resonale
[22:56] <wind789> ever used fuqid?
[22:56] <KenMan> no. Isn't that done in delphi ?
[22:56] <wind789> something like that
[22:57] <salahx> i was using fuqid
[22:57] <salahx> BTW, Fuqid 1.4 is now out
[22:57] <KenMan> we need more java tools, with swing or AWT or text or whatever...
[22:57] <KenMan> :p
[22:57] <wind789> any significant changes in Fuqid 1.4?
[22:58] <salahx> i'm told swing is bloatware and teh "free" havas are nowehre near AWT support
[22:58] <salahx> yeah
[22:58] <KenMan> so make it a console tool, just get more tools out there flexing the FCP API
[22:59] <salahx> just not the Onion NEtowrks API
[22:59] <KenMan> heh, you don't like using 7 jars to get FEC ??
[22:59] <wind789> ever tried talking to your node directly by manualy typing in FCP commands?
[23:00] <KenMan> not me, I'll leave that for the toads in the world...
[23:00] <salahx> that woudl be kinda hard ince all FCP requiest begin with the preamble \x00
[23:00] <salahx> err \x00\x00\x00\x02
[23:00] <KenMan> he manages to do it, somehow
[23:01] <wind789> in windows you can cold down the alt key and press the numberpad 0 key 4 times
[23:01] <wind789> and it gived you the null character
[23:01] <KenMan> thank you, saved me the typing. In vi, you can type ctrl-v and do something similar
[23:01] <wind789> cool
[23:01] <KenMan> maybe it is not just vi, but probably :(
[23:03] <KenMan> you know, maybe it is worth experimenting with that doReserveBW thing anyway... the other peers are still being assaulted with queries, and they have no shield from it. So they FORWARD...
[23:03] <KenMan> but I don't want to polarized to the other extreme... I'll just be "patient" :)
[23:03] <oierw> explain the doReserveBW stuff
[23:03] <KenMan> *to be
[23:03] <KenMan> toad can do a much better job, but i will try
[23:04] <oierw> and how goes the setting lastGoodBuild to > 5084?
[23:04] <salahx> ITs in Version.java
[23:04] <oierw> i'm more asking if his node is behaving better
[23:04] <lostlogic> oierw: goes pretty well
[23:04] <lostlogic> node runs smoother for sure.
[23:04] <KenMan> the theory is that we get , say, 3500 queries per hour (when it is stable anyway). This produces a certain level of backoff, and thus a certain level of routedToChoiceRank on average.
[23:05] <lostlogic> hmm... how about writing a node crawler. runs like a normal node, but only allows say 500 messages with any one node, then blacklists it permanently... over a few days or weeks (depending on network size) that should crawl the network, ya?
[23:05] <KenMan> If we scale back the queries (by leveraging MRI), we can reduce the rTCR, and thus have lower backoff. If NGR actually works (no one really knows) at ALL, this should have a positive effect on routing.
[23:06] <KenMan> Toad liked the idea, and built something very similar to this. But he built something with a 'target level of success' parameter. It controls the query rate dampening, basically.
[23:07] <KenMan> I said "shoot for 100%" and he doesn't believe we will ever come close to 100. So he used something like 'the highest success ever reported by another node' or something.
[23:07] <oierw> lolo: why not just have a huge number of nodes you can connect. you'd have a higher chance of getting a unknown node
[23:08] <KenMan> setting lastGoodBuild to 5087 has made a BIG difference on the charts I'm looking at...
[23:09] <wind789> hehe
[23:09] <wind789> how do you apply that setting?
[23:09] <KenMan> lostlogic has a workable idea there, i think... it should find a lot of nodes, if new connections are allowed to happen :o
[23:09] <wind789> is that in the config file?
[23:09] <lostlogic> oierw: because my poor box doesn't like running billions of connections? I think the slow crawl would be more effective
[23:09] <thelema|away> wind789: inside Version.java
[23:09] * thelema|away is now known as thelma
[23:09] <thelma> KenMan: what's the idea?
[23:09] <oierw> lolo: it would definetly give you an ending point for when you have traveled the entire network
[23:09] <KenMan> idea ?
[23:09] <oierw> when you run out of nodes to talk to :)
[23:10] <wind789> darn, hehe I don't compile my freenet, I just download it lol
[23:10] <thelma> crawling nodes?
[23:10] * lostlogic has 46% of gnu.ogg, 1 short of enough blocks in transit to finish the file.
[23:10] * thelma is now known as thelema
[23:10] <lostlogic> thelema: yes, what about it?
[23:10] <KenMan> aw, i thought thelma was a cute name
[23:10] <thelema> lostlogic: why wouldn't it run itsenf into a corner?
[23:10] <thelema> KenMan: :P typo
[23:11] <lostlogic> thelema: definitely
[23:11] <KenMan> that's a good bet, but it would be neat to see how far it got
[23:11] <lostlogic> ermh, ya mean like a local minimum basically?
[23:11] <thelema> lostlogic: more or less, but topologically based on the layout of node<->node links in the network.
[23:12] <thelema> once you start blacklisting parts of the network, you lose the ability to find out about nodes with similar specialization as them.
[23:12] <thelema> because you'll be bad at routing those kinds of queries, and so won't recieve them.
[23:12] <KenMan> he only wants to walk the whole net. That was his goal, i think
[23:12] <KenMan> or 'touch' the whole thing
[23:12] <lostlogic> thelema is correct though...
[23:13] <KenMan> he usually is :)
[23:13] <lostlogic> hmm... any other ideas on how to walk the net then?
[23:13] <oierw> thelema: that's assuming the network specializes and works :) as it is right now, it may hit every node
[23:13] * lostlogic chuckles
[23:13] <KenMan> wait, lostlogic , are you running with doReserveBW whatever thing ??
[23:14] <thelema> oierw: of course.
[23:14] <thelema> it's a curse, the always being right.
[23:15] <KenMan> so long as you don't think yourself INTO a box, OUT of boxes is okay...
[23:15] <lostlogic> KenMan: no
[23:15] <KenMan> cool
[23:15] <thelema> if you did a different blacklisting strategy; instead of 500 queries, maybe just assume spec is... doh, dynamic.
[23:16] <thelema> also, trusted links make walking the net tough.
[23:16] <KenMan> heh, what 'trusted links' ?? the ones that will be supported in Freenet 2.3 ?
[23:17] <salahx> ugh my rotuerToChoiceRank is in the 50's :(
[23:17] <KenMan> you have 50's peers ? you bastard. I only gots me the 38
[23:17] <thelema> KenMan: nah, probably right around 1.0 (before or after is ian's call)
[23:18] <thelema> the idea is to have your node only talk FNP to some nodes.
[23:19] <KenMan> my (brilliant) idea is to be able to form static links with trusted people's computers...
[23:19] <lostlogic> blah, I need sleep and the chat just got interesting.
[23:19] <lostlogic> g'night, peeps.
[23:19] <thelema> yeah, that idea has been around since before freenet had encryption.
[23:19] <KenMan> night. Dream of 'Maria' for a change... ;)
[23:20] <wind789> cya
[23:20] <KenMan> touche, thelema :)
[23:20] * thelema recalls back that far.
[23:21] <oierw> it was fun to go through the store
[23:21] <oierw> reading random documents and looking at random porn
[23:21] <thelema> oierw: yes, that was amusing.
[23:22] <thelema> ah, the good old days.
[23:22] <thelema> file *
[23:22] <KenMan> so it's a bold, cold, scary new world, huh ?
[23:23] <oierw> i still remember the fs before the monolithic fs. man ddi that thing have holes
[23:23] <thelema> no, it's more italic than bold
[23:23] <KenMan> we all know those italian mafioso's need something like Freenet, heh ?
[23:23] <thelema> It was a good attempt, but freenet having a filesystem built in is like my bike having a head sticking out the tire.
[23:24] <wind789> bumpy ride
[23:24] <oierw> it's all hobx's fault :)
[23:25] <KenMan> i was gonna ask where his trashcan had gotten off to. He doesn't seem to be popping his head up very often...
[23:25] <oierw> but it gave a bed for nativefs which is both fast and not monolithic
[23:26] <thelema> oierw: yes it is. Even the stupid state engine for handling requests.
[23:26] <salahx> geeit soudns liek a few tweaks FReenet coudl bei ts own operating system :)
[23:27] <KenMan> What is wrong with state engine ? the coding style or the actual operations ??
[23:27] <oierw> i think that up to before toad all the major developers spent atleast a portion of their time coding under the influence of drugs or alcohol
[23:27] <thelema> the way it's implemented, with java objects being reflected around.
[23:27] <salahx> well something wrong with it due to those "State Does Not Recieve" errors
[23:27] <thelema> with the way it's split up between a dozen files when it could be concisely done in one.
[23:27] <KenMan> hey man, it worked in berkeley 8)
[23:28] <oierw> the state engine is probably atleast a part of what was broken with announcements
[23:28] <KenMan> i can't seem to map announcements onto NGR. CPA yeah, but not NGR...
[23:28] <thelema> the way that the state engine is completely undocumented (and nearly undocumentable)
[23:29] <KenMan> oh, that's just par for the course
[23:29] <thelema> announcements aren't as needed with bidi
[23:30] <KenMan> other than the single 'censorship attack' i don't see why nodes can't just break down and tell the other side what it is guess at...
[23:31] <KenMan> to counter the censorship attack, nodes holding items that are strongly outside their spec could occasionally reinsert some of them ...
[23:31] <thelema> it's not nodes' job to decide their spec.
[23:31] <salahx> but nodes don;t knwo ther specializtion
[23:32] * Anvil_Vapre (~Cichli@63-227-158-112.omah.qwest.net) Quit ("Leaving")
[23:32] <KenMan> they could learn to measure it, anyway...
[23:32] <KenMan> from requests and DS contents...
[23:32] <thelema> with a very high level of uncertainty
[23:32] <thelema> mostly from requests.
[23:32] <KenMan> almost as high a level as the level of effectiveness ?
[23:32] <KenMan> sorry...
[23:32] <thelema> huh?
[23:33] <KenMan> i understand the clever concept that things will emerge, but I DON'T see it being highly scalable
[23:34] <KenMan> CPA looks doable. NGR , to date, doesn't.
[23:34] <KenMan> but there was no good evidence whether CPA would have scaled "big-time" ...
[23:35] <KenMan> one more point on announcements (from what you were teaching me), would more than one node adopt the same 'random' spec point chosen by node 1 in your A->1->2...->9 example ??
[23:36] <KenMan> or just node #9 ?
[23:36] <thelema> adpot the spec point?
[23:36] <thelema> there's one specialization point chosen in the process.
[23:36] <thelema> it's spec1 xor spec2 xor ... xor spec9
[23:36] <KenMan> the only likely way nodes can begin to coincide on their estimated point(s) of spec is by coinciding in the central node's DS...
[23:37] <thelema> they all add A to their routing table with initial specialization being that value
[23:37] <KenMan> yes, but the XOR'ed value is only to be used by a single requestor then
[23:37] <KenMan> oh, thank you. that's what i was driving at
[23:37] <thelema> good.
[23:38] <KenMan> the mud thins out a little :)
[23:38] <thelema> hmm, I wonder what would happen if nodes only introduced via announcements.
[23:38] <thelema> i.e. no more connecting to a node you don't know.
[23:38] <KenMan> total chaos ?
[23:38] <KenMan> or total order perhaps ?
[23:39] <thelema> it wouldn't be taht bad. nodes would still be able to restart, because both ends would still have references for the other.
[23:39] <KenMan> naw, it would be a 'web of trust' then , no ?
[23:40] <KenMan> how would "Ian's Kids" get connected ?? (his favorite charity that favors unwebbed newbies)
[23:41] <thelema> by announcing to a known node.
[23:41] <thelema> (that node wouldn't have to know them, but they'd annouce to a node, which would give their ref to a bunch of other nodes, and they'd all connect the new one.
[23:41] <KenMan> ah... it sounds almost workable then. But an announcement is only a single starting point, things would still spin off into chaos for the most part, over time....
[23:42] <thelema> StoreData would be interesting...
[23:42] <thelema> I don't think there would be that much chaos.
[23:42] <KenMan> or maybe something meaningful could "emerge" ! or not... we will probably never know :(
[23:42] <thelema> it'd just be like if everyone used trusted links.
[23:42] <KenMan> well, sort of.
[23:42] <thelema> except it was possible to announce your way into getting a connection.
[23:43] <KenMan> bbl, favorite cartoon is coming on :o
[23:43] <thelema> but the routing would be very developed, because connections would be nearly static
[23:43] <thelema> KenMan: what cartoon?
[23:43] <KenMan> "Aqua Teen Hunger Force"
[23:43] <KenMan> no. 1 in the hood, 'G'
[23:46] <KenMan> http://mywebpages.comcast.net/jkcorson/frylock.png
[23:47] <thelema> right.
[23:47] <KenMan> http://adultswim.toonzone.net/aqua.html - i think the target audience is 15 to 20-somethings that use a lot of drugs. But I find it rather comical...
[23:48] <thelema> KenMan: 5th Avacado. Bananaphone. Badgerphone.
[23:51] <KenMan> lemonizer ?
[23:51] <thelema> KenMan: *phone can be seen through lemonizer
[23:52] <thelema> 5th avocado is a sound-o-vision production.
[23:53] <KenMan> that bananaphone looks suspiciously like a demon remote-control that was shown on ATHF
[23:53] <thelema> nah, it's a gundam stuck in a banana.
[23:54] <thelema> check out 5th avocado. "And then all the fishy ghosts drowned"
[23:56] <KenMan> yes, it's pretty messed up, had to pause it !!
[23:57] <thelema> :)
[23:59] <oierw> i'm more of a fan of sealab 2021
These logs were automatically created by Jay Oliveri with his gimp hapi on irc.freenode.net.
