Timestamps are in GMT/BST.
[0:10] * attica (i=attic@) has joined #freenet
[0:23] * blimeyboy (n=chatzill@) Quit (Read error: 110 (Connection timed out))
[0:27] * attic (i=attic@) Quit (Read error: 110 (Connection timed out))
[0:27] * attic (i=attic@) has joined #freenet
[0:30] * attica (i=attic@) Quit (Read error: 110 (Connection timed out))
[0:40] <toad_> ok i will fix it tomorrow
[0:40] * toad_ (n=toad@) Quit (Remote closed the connection)
[0:44] * Andross (n=drew@) has joined #freenet
[0:58] * mozillaman_ (n=self@) has joined #freenet
[1:07] * MikeW (n=em@) has joined #freenet
[1:14] * mozillaman (n=self@) Quit (Read error: 110 (Connection timed out))
[1:25] * maelztrom (i=nobody@) has joined #freenet
[1:37] * Andross (n=drew@) Quit (Read error: 104 (Connection reset by peer))
[1:59] * mozillaman_ (n=self@) Quit (Read error: 104 (Connection reset by peer))
[2:00] * kinzite_ is now known as kinzite
[2:05] * lattt (n=gerhard@) Quit ("baba")
[2:13] * rebo123 (n=rebo123@) has joined #freenet
[2:41] * kinzite_ (n=kinzite@) has joined #freenet
[2:46] * mk (n=mk@) has joined #freenet
[2:52] * lokadin (n=loki@) has joined #freenet
[2:55] * kinzite (n=kinzite@) Quit (Read error: 110 (Connection timed out))
[3:24] * Pichu0102 (n=pichu010@) has joined #freenet
[3:24] * Pichu0102 (n=pichu010@) Quit ("Thoughtcrime does not entail death: Thoughtcrime IS death.")
[3:24] * Pichu0102 (n=pichu010@) has joined #freenet
[3:49] * needle (n=nevle@) has joined #freenet
[3:51] * MikeW (n=em@) Quit ()
[4:08] * Psilos (i=dapper@) has joined #freenet
[4:10] * peerless (n=Peerless@) has joined #freenet
[4:22] * needle (n=nevle@) Quit ("Leaving")
[4:35] * needle (n=nevle@) has joined #freenet
[4:52] * Urs_ShPo (n=gaim@) has joined #freenet
[5:15] * kinzite_ (n=kinzite@) Quit (Remote closed the connection)
[5:19] * needle (n=nevle@) Quit ("Leaving")
[5:50] * heph (n=heph@) has joined #freenet
[6:02] * forceflow307_ (n=forceflo@) has joined #freenet
[6:03] * PriYo (i=PriYo@) has joined #freenet
[6:03] <PriYo> ,14«ºôº» Priñcê OF PriYo,14«ºôº» ||,15|| ScRipT ,0-,1 PriYo JusT EñTêRêd Iñ Thê RooM *!* Wish YoU JoU åñD MåñY Blêssiñg s OF LiFê *!* Hêllo Hêåvêñly pêoplê ,0-,14«ºôº» Priñcê OF PriYo,14«ºôº» ||,15|| ScRipT
[6:06] * dickie (n=pirch@) has left #freenet
[6:08] * PriYo (i=PriYo@) Quit (Read error: 131 (Connection reset by peer))
[6:17] * darthmuss (n=kussefis@) has joined #freenet
[6:19] * forceflow307 (n=forceflo@) Quit (Read error: 110 (Connection timed out))
[6:23] <darthmuss> would it be possible to run ones own private darknet
[6:23] <Psilos> yes
[6:24] <darthmuss> only adding nodes you trust completely... like a gang of friends
[6:24] <Psilos> check out 'Waste'
[6:24] <darthmuss> cool
[6:24] <darthmuss> i know waste, was wondering if we could do it with freenet
[6:24] <darthmuss> waste is the p2p that the winamp people made, right?
[7:25] * Psilos (i=dapper@) Quit ()
[7:27] <sopues> my probeall is currently estimating 2000 nodes
[7:28] <sopues> also, why do the network need to be so big? I would _never_ recommend freenet to anyone at this point.. the _only_ reason being speed
[7:28] <sopues> not lack-of-opennet for sure
[7:32] <Ralith> sopues, the speed problem is a result of broken loadbalancing code
[7:32] <sopues> doesnt matter why that is
[7:32] <Ralith> sopues, it's a high priority fix and being worked on actively.
[7:32] <sopues> its just the only reason I see why the network does not grow
[7:33] <sopues> 3kb/s MAX speed is not "hot" today, trust me
[7:33] <Ralith> well, it'll be taken care of within a couple releases.
[7:33] <Ralith> and it's good enough for most of us :P
[7:33] <sopues> yeah... I also think this opennet idea is bad bad bad bad
[7:33] <sopues> atleast atm
[7:33] <sopues> oh well
[7:33] <Ralith> it's optional.
[7:33] <Ralith> you don't have to use opennet
[7:33] <Ralith> and opennet and darknet can talk to eachother
[7:33] <sopues> no, I mean spending effort on it/trying to get it to work
[7:33] <theseeker> whee, running the BackTrack livecd ... and found a remote desktop utility that interfaces properly with Windows remote desktop @_@ standards work!
[7:34] <Ralith> sopues, it will make it a lot easier for people to start using freenet
[7:34] <sopues> I guess thats true
[7:34] <Ralith> and once they get the hang of that they can set up a darknet over frost or something
[7:34] <sopues> I hope it wont fuck things up :)
[7:34] <Ralith> don't see how it could
[7:34] <sopues> I'm worried darknet will die once opennet gets out
[7:34] <Ralith> it'll make the network as a whole baloon, and the darknet grow, probably.
[7:34] <Ralith> sopues, there can always be a darknet, even if it's just one node
[7:35] <sopues> not very useful though ;)
[7:35] <Ralith> no, it works fine
[7:35] <Ralith> secure, too
[7:35] <Ralith> there will probably be an initial shrinking of the darknet
[7:35] <Ralith> but it'll def. stay around, and start to grow again too
[7:53] <darthmuss> why are you talking about darknet and opennet?
[7:53] <Ralith> why not?
[7:53] <darthmuss> i thought that 0.7 was NOT an opennet and never will be?
[7:54] <darthmuss> like 0.5 is an opennet because it just bootstraps to some unknown node in a ref file
[7:54] <darthmuss> and 0.7 is a darknet because we manually add "trusted" nodes
[7:55] <darthmuss> or are my definitions fubar
[7:55] <IMCensored1> darthmuss: http://wiki.freenetproject.org/OpennetDesign
[8:31] * paveq (i=paveq@) Quit (Read error: 110 (Connection timed out))
[8:46] <sopues> darthmuss, oh yes, opennet is in the works, apparently
[9:10] * sanity (n=ian@) Quit ()
[9:19] * Urs_ShPo (n=gaim@) Quit ("Leaving.")
[9:22] * needle (n=nevle@) has joined #freenet
[9:31] * paveq (i=paveq@) has joined #freenet
[9:39] * railk (n=railk@) has joined #freenet
[9:40] * peerless (n=Peerless@) Quit (Read error: 104 (Connection reset by peer))
[9:40] * peerless (n=Peerless@) has joined #freenet
[9:52] * whiterabbit (n=Whiterab@) has joined #freenet
[10:02] * railk (n=railk@) Quit (Remote closed the connection)
[10:03] * needle (n=nevle@) Quit ("Leaving")
[10:12] * Samu^Busy (n=mogul@) has joined #freenet
[10:12] <Samu^Busy> Hello
[10:12] * Samu^Busy is now known as Samulozy
[10:12] <Samulozy> Hello :)
[10:23] * Samulozy is now known as Samu^Away
[10:24] * Samu^Away (n=mogul@) has left #freenet
[10:29] * whiterabbit (n=Whiterab@) Quit ("KVIrc 3.2.5 Anomalies http://www.kvirc.net/")
[10:53] * Fenneszz (i=1234@) Quit (Remote closed the connection)
[11:14] * Treyvan (n=fred@) has joined #freenet
[11:18] * somerandomname (n=some@) has joined #freenet
[11:24] * vfervers (n=vfervers@) Quit (Remote closed the connection)
[11:28] <Treyvan> Hello, can anyone answer a question about the node "Location" value please?
[11:32] * somerandomname (n=some@) has left #freenet
[11:37] * somerandomname (n=some@) has joined #freenet
[11:37] * somerandomname (n=some@) has left #freenet
[11:39] <cyberdo> Treyvan: perhaps
[11:39] * rebo (n=rebo123@) has joined #freenet
[11:40] * rebo123 (n=rebo123@) Quit (Read error: 60 (Operation timed out))
[11:43] <Treyvan> Thanks cyberdo, the question is:- If I want to remove some noderefs ( I am getting rejected requests due to being too busy ) then can I identify the best nodes to remove by the distance value
[11:43] <Treyvan> and if so what value should I be looking to remove?
[11:43] <cyberdo> well.. you can't... not really
[11:44] <cyberdo> the "location" will change over time
[11:45] <cyberdo> but if you still want to.. I think you should just go for max(myval^2 - hisval^2)
[11:45] <Treyvan> I have been trying to find out how the location parameter is calculated
[11:45] <cyberdo> ;o)
[11:45] <cyberdo> short way: it's randomized during installation, and then it will be swapped between nodes to optimize the network layout
[11:46] <Treyvan> It also struck me that as I got a lot of references from freenode-ref in one session and there a group of us then presumably that group will have a very similar subset of nodes.
[11:46] <cyberdo> the location is kind of an IP... it's used both for routing and storing (since they go hand-in-hand in comm)
[11:46] <cyberdo> probably
[11:47] <cyberdo> that's the downside of the swapping-parties that goes on in freenet-refs
[11:47] <cyberdo> the net will form clusters with tiny interconnections
[11:47] <Treyvan> I suppose the best thing is to go onto #ref at different times and just get the one swap
[11:48] <cyberdo> otoh.. in time, some of the connections break up, and the arts will then join other clusters as well
[11:48] <cyberdo> probably
[11:48] <Treyvan> hadn't thought of that. Thanks
[11:50] <Treyvan> As a rather sneaky second question - is there any way that you know of to alter the SUB_MAX_PING_TIME?
[11:50] <cyberdo> no ;o)
[11:51] <Treyvan> I am also running a node under Microsoft VM and the overhead of the VM is causing packets to fail the check.
[11:51] <Treyvan> Ah well, I will have to live with it then. Thanks
[11:52] <cyberdo> np
[11:52] <cyberdo> Bombe: around?
[11:58] <Bombe> For a minute or two, yes.
[11:58] <Bombe> Lunch is coming up.
[12:01] * Treyvan (n=fred@) has left #freenet
[12:06] * whiterabbit (n=Whiterab@) has joined #freenet
[12:07] * railk (n=railk@) has joined #freenet
[12:12] * railk (n=railk@) Quit (Client Quit)
[12:35] * toad_ (n=toad@) has joined #freenet
[12:35] * ChanServ sets mode +o toad_
[12:52] * Fenneszz (i=1234@) has joined #freenet
[12:57] * forceflow307_ is now known as forceflow307
[13:23] * peerless (n=Peerless@) Quit (Read error: 104 (Connection reset by peer))
[13:31] <CIA-14> toad * r11477 /trunk/freenet/src/freenet/config/StringArrOption.java: Chop off last delimiter.
[13:59] * greycat (i=rfc1413@) has joined #freenet
[14:04] <CIA-14> toad * r11478 /trunk/freenet/src/freenet/config/StringArrOption.java:
[14:04] <CIA-14> Fix option encoding bug: we weren't decoding the initial value from the file, but we were encoding it when it was written to the file.
[14:04] <CIA-14> Now bookmarks= won't double in size on every restart.
[14:04] <CIA-14> toad * r11479 /trunk/freenet/src/freenet/clients/http/SymlinkerToadlet.java: What on earth is going on here?
[14:06] <toad_> hmmm CIA is overloaded or something
[14:06] <Zothar> yeah, it's been kinda funky in one way or another for several days now
[14:08] <Zothar> BTW, it looks like your location swap spying is causing little, if any, locations to "age", at least according to some uncommitted code; I'm gonna look at it some more when I get off work in 4-5 hours and then commit
[14:08] * Zothar is now known as Zothar_Work
[14:09] * pearl298 (n=pearl298@) has joined #freenet
[14:09] <CIA-14> toad * r11480 /trunk/freenet/src/freenet/node/fcp/ (5 files): Persistent inserts should work again. Hopefully.
[14:09] <toad_> Zothar_Work: please do
[14:34] <Beta_M> plugin interfaces in the SVN version are *very* different from the jar file being distributed
[14:34] <Beta_M> can the new jar version be released please
[14:36] <toad_> no, there are some problems with it still
[14:36] <toad_> also if i release it then all the plugins will break
[14:36] <toad_> i'm not sure how we're going to deal with this yet
[14:36] <toad_> ian is vehemently opposed to me finishing the job
[14:37] <toad_> so i suppose we'll just have to release it broken and ask everyone to upgrade their plugins manually
[14:38] <Beta_M> i didn't mean "right now"
[14:39] <Beta_M> i might hack on Librarian real quick if i have time
[14:39] <Beta_M> still flooded with the work for Freekiwiki
[14:39] <Beta_M> i wanted to release something by 21st
[14:40] <toad_> well we need to sort out the plugin API
[14:40] <toad_> and actually implement it
[14:40] <toad_> Bombe may be able to do it
[14:40] <toad_> if he gets any time
[14:41] <Beta_M> well, i'm coding for what exists in the SVN right now
[14:41] <toad_> freenet/plugin/api and freenet/support/api are the new API
[14:41] <toad_> there will be a jar containing just those two dirs to develop against
[14:42] <Beta_M> nice
[14:42] <toad_> right now i'm debugging some of the relatively drastic refactorings that went in
[14:43] <Beta_M> toad_: do you happen to know if there are any universities that have teams working on anonymity, censorship resistance, and distributed networks in UK
[14:44] <Beta_M> i'm thinking about doing a phd in the area, and am looking around if there are already research teams in this country
[14:45] <toad_> i dunno
[14:45] <Beta_M> i just thought i'd ask... thinking you might have ran across that stuff some time
[14:45] <Beta_M> thanks
[14:45] <Beta_M> (as in thanks for the time)
[14:47] * darthmuss (n=kussefis@) Quit ()
[14:52] * railk (n=railk@) has joined #freenet
[14:52] * railk (n=railk@) Quit (Client Quit)
[14:52] * railk (n=railk@) has joined #freenet
[14:54] * user1 (n=user1@) has joined #freenet
[14:57] * user1 (n=user1@) has left #freenet
[15:00] * pearl298 (n=pearl298@) Quit (Read error: 110 (Connection timed out))
[15:06] <CIA-14> toad * r11482 /trunk/freenet/src/freenet/node/fcp/ (ClientGet.java ClientPut.java): Really fix serialization of requests
[15:06] <CIA-14> toad * r11484 /trunk/freenet/src/freenet/support/io/SerializableToFieldSetBucketUtil.java: Logging
[15:13] * mozillaman (n=self@) has joined #freenet
[15:14] <mozillaman> This is a dark week for news :)
[15:15] <Zothar_Work> mozillaman: eh, depends on whether 1008 is released this week or not; there's some new stats toys for one (I'll probably be committing more on that later today)
[15:16] <mozillaman> Zothar_Work: Do tell :)
[15:17] <mozillaman> "stat toys" sounds like a story
[15:17] <mozillaman> Have screenshots? :)
[15:17] <Zothar_Work> mozillaman: well, download the testing build and see for yourself; do the same around 18:00 GMT for my addition (gotta debug it first, after I get off work)
[15:18] <Zothar_Work> (maybe that'll be 19:00 GMT, not sure when I'll be off work yet, etc.)
[15:20] * toad_ is fixing bugs in pre-1008
[15:20] <toad_> new bugs caused by me after 1007 :|
[15:21] <toad_> once that's done, we need to decide whether to break plugins and ask everyone to get the fixed ones, or to implement the new plugins and auto-deploy new ones
[15:41] <sopues> LOCATION 336: 1.0 - estimated nodes: 337.0
[15:45] <toad_> okay, that's the known bug
[15:45] <toad_> known bugs
[15:45] <toad_> that i knew about
[15:45] <toad_> there are probably more on the bugtracker
[15:45] <toad_> but that'll do for now
[15:45] <toad_> bbl
[15:46] <CIA-14> toad * r11485 /trunk/freenet/src/freenet/support/io/ (5 files): Make new serialization code backwards compatible.
[15:50] <CIA-14> toad * r11486 /trunk/freenet/src/freenet/support/io/SerializableToFieldSetBucketUtil.java: add FIXME
[15:55] <_ph00> would a second node on the same machine work, using the same store?
[15:55] <Zothar_Work> _ph00: not if they're running simultaneously
[15:55] <_ph00> so basically it wouldnt
[15:56] <_ph00> (work)
[15:56] <Zothar_Work> if all is separated, then it will; I've got three nodes on one box ATM
[15:56] <Zothar_Work> I think some might run their node in both Windows and Linux using the same store, etc. but not simultaneously
[15:56] <_ph00> yes, I had up to 4 nodes on one machine once, as an experiment.
[16:02] * paveq (i=paveq@) Quit (Read error: 110 (Connection timed out))
[16:21] * mozillaman (n=self@) Quit (Read error: 104 (Connection reset by peer))
[16:28] <nextgens> hi
[16:28] * nextgens reads backlog
[16:29] * PraiseChaos (n=kcecil@) has joined #freenet
[16:32] <nextgens> Beta_M> there is on in Brest (France)
[16:33] <nextgens> toad_> any news regarding you mail on opennet ?
[16:40] <nextgens> toad_> I lost my jobs queued on global queue, *again* :)
[16:42] * nextgens tries with the updated version
[16:43] <nextgens> ok, now it works
[16:43] <nextgens> I'm currious ; would it work from #1007 to #1008 ?
[16:46] * Urs_ShPo (n=gaim@) has joined #freenet
[16:49] <Beta_M> nextgens: you meant the research team
[16:49] <nextgens> Beta_M> yes
[16:50] <Beta_M> ok, thanks, i don't speak much french past the counting to 10, but that's a start non-the-less
[16:51] <nextgens> I know they are looking for someone to work on a distributed backup system atm
[16:52] <nextgens> they posted a hiring proposal on @p2p-hackers a few weeks ago
[16:53] <Beta_M> "@p2p-hackers" ???
[16:54] <nextgens> p2p-hackers@lists.zooko.com
[16:54] <nextgens> Beta_M> http://lists.zooko.com/mailman/listinfo/p2p-hackers
[16:54] <Beta_M> thanks
[17:03] * sanity (n=ian@) has joined #freenet
[17:03] * ChanServ sets mode +o sanity
[17:38] * lattt (n=gerhard@) has joined #freenet
[17:45] <CIA-14> nextgens * r11487 /trunk/freenet/src/freenet/keys/ (FreenetURI.java USK.java): Prevent a possible NPE on .getKeyType()
[17:57] <CIA-14> Jogy * r11488 /trunk/freenet/src/freenet/clients/http/StatisticsToadlet.java:
[17:57] <CIA-14> * Fix broken peercircle - *NumberFormat is localized!* This broke display on all non-englisch (anything else as '.' for comma) systems (like e.g. France)
[17:57] <CIA-14> * Remove this weird "me-offset" from peercircle
[17:59] <Zothar_Work> Jogy: if you were in this channel, I'd tell you that it was there for a reason
[18:00] <Zothar_Work> now I gotta see what you changed so I can fix it if needed before committing my changes currently in testing
[18:00] <Zothar_Work> oh well
[18:02] <Zothar_Work> I guess I need a non-localized version of Number/DecimalFormat
[18:15] * rob___ (n=rob@) has joined #freenet
[18:17] * nanaem (n=nanaem@) has joined #freenet
[18:17] * rob___ (n=rob@) has left #freenet
[18:20] * nanaem (n=nanaem@) Quit ()
[18:28] * Fenneszz is now known as Fennes
[18:30] <toad_> distributed backup sounds cool
[18:30] <sandos> yeah.. where did you hear it? =)
[18:31] <toad_> different requirements to freenet ... guaranteed recovery if at all possible...
[18:31] <sandos> o
[18:31] <sandos> ok
[18:33] <toad_> nextgens: yes we had a discussion about it
[18:33] <toad_> nextgens: that's part of the reason for my interest in PROBEALL traces
[18:34] <toad_> nextgens: it looks like we have either very few nodes (hundreds) and a lot more nodes at locations near 0.0/1.0, or we have many nodes (thousands) but very poor routing
[18:34] <toad_> nextgens: either way could well be caused by #freenet-refs
[18:34] <toad_> nextgens: it's designed to work from #1007 to #1008
[18:34] <toad_> nextgens: if it doesn't, tell me
[18:35] <toad_> Zothar_Work: so email him? :)
[18:35] <toad_> Zothar_Work: any chance of you looking into why spying on other ppl's swaps isn't giving any more locs, or shall i?
[18:35] <Zothar_Work> toad_: I'm reverting the change and fixing the problem; I'll explain the "wierd me-offset" in my commit message
[18:36] <toad_> Zothar_Work: and in comments, i hope?
[18:36] <toad_> Zothar_Work: what exactly does your uncomitted code say?
[18:36] <toad_> Zothar_Work: that we're not getting any more locations from spying on other ppl's refs?
[18:37] <Zothar_Work> Zothar_Work: Oh, the spying appears to be giving more locs; I'm just not sure if things are generally not aging because of the spying affecting things more often or if it's a bug in my display code; I'll probably commit and then fix, though I think I have an idea on what the problem might have been: "all or nothing"
[18:37] <sandos> toad_, uhm why exactly would #freenet-refs cluster around 0 specifically? My gut tells me that must be a bug, rather
[18:37] <Zothar_Work> I'll comment the "weird me-offset" thing
[18:37] <toad_> sandos: #freenet-refs produces bad topology
[18:38] <toad_> sandos: bad topology => bad routing
[18:38] <sandos> ok
[18:38] <toad_> sandos: so there are more nodes, but we don't see them
[18:38] <toad_> another, probably related theory:
[18:38] <toad_> natural selection
[18:38] <toad_> newbie nodes have few connections, are "peripheral"
[18:38] <sandos> well, a PROBEALL (afaict) will never visit all nodes, right?
[18:38] <toad_> so "unpopular" locations get swapped to them
[18:39] <toad_> so the well-connected node gets moved towards the center
[18:39] <toad_> then the newbie node drops off the network
[18:39] <sandos> even with a perfect topology there cant be any gurantees that I always have the next-highest/lowest node as my nieghbour
[18:39] <toad_> net result: part of the keyspace (0.0/1.0 in this case but could be anywhere) gets more crowded
[18:39] <Zothar_Work> toad_: I think you'll be interested in my commit; it should be coming in the next half hour or less
[18:39] <toad_> part of it gets less crowded
[18:39] <sandos> but its _always_ been around 0
[18:39] <toad_> sandos: probe requests are routed
[18:39] <sandos> it doesnt make sense?
[18:40] <toad_> sandos: yeah, it being around 0 makes no sense
[18:40] <toad_> but can you see a bug?
[18:40] <sandos> it should move slightly randomly if it was "just" a matter of clustering...
[18:40] <toad_> i can't see any systematic issue
[18:40] <toad_> it's possible that it's just a coincidence that all the devs' nodes are around 0.0
[18:40] <sandos> no, I cant.. unfortunately I dont have any time to spend on things other than work atm :(
[18:41] <toad_> well i don't like to implement opennet with this hanging over us but i have no idea how to progress forward
[18:41] <toad_> and it's just possible that it's an artifact of not having opennet causing a bad topology
[18:42] <toad_> nextgens: any ideas?
[18:43] <toad_> maybe i should ask mrogers, i've asked oskar...
[18:44] <toad_> Zothar_Work: hmmm
[18:44] <toad_> Zothar_Work: we are definitely spying on our peers and successfully parsing their traffic
[18:45] <toad_> oh
[18:45] <toad_> i see
[18:45] <toad_> bug
[18:45] <toad_> fix bug
[18:45] <Zothar_Work> Zothar_Work: and I we're learning from it; I first suspect a bug in my display code before I think it "breaks" aging................ I'll let you fix the bug you found... :)
[18:45] <CIA-14> toad * r11489 /trunk/freenet/src/freenet/node/LocationManager.java: Register the linked location as well as the link and the original location. DOH!
[18:46] <toad_> Zothar_Work: was that the same bug you fixed?
[18:46] <Zothar_Work> mine wasn't in LocationManager: I'll look at what you did to see if it might be related
[18:46] <toad_> okay, it's much faster picking up locations now
[18:46] <toad_> 134 nodes in 43 seconds uptime
[18:47] <Zothar_Work> yeah, that could do it... :)
[18:47] <toad_> :)
[18:48] <Zothar_Work> My seen bug is completely in my new stuff then
[18:49] <toad_> ah well
[18:49] <toad_> we'll see how this goes; if it turns out that the network is much bigger than the PROBEALL: traces show, then we'll have learned something
[18:49] <toad_> OTOH if it just gets to the same size more slowly, we'll also have learned something
[18:50] * SixNine (n=none@) has joined #freenet
[18:52] <sandos> is the new code displaying the results on the /stats page or only in logs?
[18:52] <toad_> on stats
[18:52] <sandos> ok
[18:52] <sandos> 180 nodes in 1m35s
[18:52] <toad_> 326 nodes in 6m38s
[18:52] <toad_> may be settling
[18:54] <sandos> yeah
[18:55] <toad_> there's a report of downloads not working...
[18:58] <toad_> hrrrrrm
[18:58] <toad_> there's something in this ...
[18:59] <toad_> CHK@XcbXaKgMly8zG63dUKxgorFjUZHLJPrWfow0mmLfpDM,tNnHgfeIx~TOFMR4X8nOqmvOUPTKN01LbBEMX15klv0,AAEC--8/01_genesis.zip
[18:59] <toad_> so far:
[18:59] <toad_> Total=7720
[18:59] <toad_> now, 7720 * 32kB = 247040 kB
[18:59] <toad_> no way can 01_genesis.zip possibly be that big
[19:00] <toad_> and it just keeps on growing
[19:00] <toad_> Total=12522
[19:00] <toad_> = 400MB
[19:00] <toad_> i think not
[19:05] <toad_> Dec 19, 2006 19:01:19:882 (freenet.client.async.SplitFileFetcher, Decoder for fr
[19:05] <toad_> eenet.client.async.SplitFileFetcherSegment@9fd062, MINOR): Algorithm: 1, blocks
[19:05] <toad_> per segment: 128, check blocks per segment: 64, segments: 66
[19:05] <toad_> hmmm
[19:05] <toad_> this seems unlikely ...
[19:06] <toad_> looks genuine though
[19:06] <toad_> i wonder what he uploaded - by mistake or on purpose to get me to fetch it - ?
[19:07] <toad_> DataFound
[19:07] <toad_> DataLength=272650872
[19:07] <toad_> Identifier=FProxy:01_genesis.zip
[19:08] <toad_> if that's really a zip file full of mp3's then freenet is working several orders of magnitude better than i thought
[19:08] <toad_> because he has dozens of such files on his page!
[19:09] <toad_> woah, it is
[19:10] * toad_ will try fetching the rest of the files!
[19:10] * Urs_ShPo (n=gaim@) Quit ("Leaving.")
[19:10] <sandos> its that fast downloading, or are you just impressed that he inserted them? =)
[19:10] <toad_> i'd already tried to fetch it
[19:10] <toad_> but if this is representative, then there's many many gigabytes of files on that page
[19:11] <toad_> well there are only 12 such zips actually
[19:11] <toad_> but 12 * 200 odd megs each = well over 2GB
[19:18] <CIA-14> toad * r11490 /trunk/freenet/src/freenet/node/fcp/FCPServer.java: Don't add extension twice.
[19:18] <toad_> the other files don't seem to work :|
[19:20] * paveq (i=paveq@) has joined #freenet
[19:20] * railk (n=railk@) Quit (Read error: 131 (Connection reset by peer))
[19:25] * railk (n=railk@) has joined #freenet
[19:29] * lattt (n=gerhard@) Quit ("baba")
[19:32] * Sugadude (i=Sugadude@) has joined #freenet
[19:38] * Beta_M (n=Somebody@) Quit ("http://freekiwiki.sf.net/")
[19:39] <CIA-14> zothar * r11491 /trunk/freenet/src/freenet/ (3 files in 2 dirs): (log message trimmed)
[19:39] <CIA-14> Re-introduce offsetMe to the peer circle to make the local node stand out better
[19:39] <CIA-14> and easier to find. The distance from the edge of the peer circle is a function
[19:39] <CIA-14> of the probability that the peer will reject our request (pReject). Add a
[19:39] <CIA-14> second circle for the node location distribution of swaps we've seen. The
[19:39] <CIA-14> distance from the edge of the node location distribution circle indicates how
[19:39] <CIA-14> long ago we last saw that location, up to 24 hours. Add a "center point" to the
[19:39] * Beta_M (n=Somebody@) has joined #freenet
[19:47] <CIA-14> toad * r11492 /trunk/website/pages/donate.php: Fix PHP warning on donate page.
[19:50] * attic (i=attic@) Quit ("bbl")
[19:54] * Randan_ (n=Randan@) has joined #freenet
[20:00] * Randan (n=Randan@) Quit (Read error: 60 (Operation timed out))
[20:14] * MikeW (i=Mike@) has joined #freenet
[20:17] <_ph00> hum
[20:17] <_ph00> looks like 28m20s ago all my peers got reconnected at the same time
[20:18] * Zothar_Work is now known as Zothar
[20:19] <Zothar> _ph00: node restarted maybe. Network hiccup perhaps.
[20:19] <_ph00> nopw
[20:19] <_ph00> not node restarted, it says uptime 21hrs
[20:20] <_ph00> only peer connections have 28min
[20:20] <_ph00> nothing in wrapper log since last restart 11PM+ yesterday
[20:21] <_ph00> 23:59:48
[20:21] <Zothar> musta been a network hiccup then
[20:22] <_ph00> network hiccups should get me disconnected here too, I've been online all the tiome
[20:22] <_ph00> time
[20:22] <Zothar> you might check your logs/freenet*.gz for that time period (filenames are local, but times in logs are GMT IIRC)
[20:22] <_ph00> no disconnessions ~half an hour ago
[20:22] <_ph00> hm
[20:22] <_ph00> checking
[20:22] <Zothar> the hiccup may have been packet loss that Freenet is more sensitive to than your other connections
[20:23] <Zothar> or a reallyl bad pick time for a few seconds/minutes, which is the same as packet loss as far as Freenet is concerned
[20:24] * mozillaman (n=self@) has joined #freenet
[20:31] <_ph00> Zothar: now this is weird
[20:31] <_ph00> took me some time to figure out what was the exact minute
[20:31] <_ph00> it was 20:47 local
[20:31] <_ph00> now, in the log file from :45 to :0
[20:32] <_ph00> 50
[20:32] <_ph00> there are no :47 entries
[20:32] <_ph00> it goes straight from 20:46 to 20:48
[20:32] <Zothar> I'm headed out for an hour-ish, so I'll have to read the back log (adjust the in-log file times for GMT, remember)
[20:32] * Zothar AFKs
[20:33] <_ph00> gmt would be 19:47
[20:33] <_ph00> my local times are gmt +1
[20:33] <toad_> _ph00: interesting
[20:33] <toad_> _ph00: a glitch or some sort
[20:33] <toad_> node glitch, cpu glitch, some high priority process running?
[20:34] <_ph00> hmm... I have freenet andf i2p running together, no changes
[20:34] <_ph00> I can't remember anything special hallening at 19:47 gmt
[20:34] <_ph00> happening
[20:35] <_ph00> right now I have the cpu running around 90% I wonder why...
[20:35] * _ph00 checks 'top'
[20:35] <CIA-14> Jogy * r11493 /trunk/freenet/src/freenet/clients/http/ToadletContextImpl.java:
[20:35] <CIA-14> Force the localization "SimpleDateFormat" to Locale.US. Otherwise it would lead to localized (and therefore corrupt) date: and last-modified: HTTP-Header fields. wget moaned about it and it should be fixed anyway.
[20:35] <CIA-14> Checked the code for other "SimpleDateformat"s, there are some but they don't use written names for months, etc. They are left untouched. Watch out :)
[20:36] <_ph00> I shutted down i2p, and saw that it *is* freenet (the other java process) that is taking a lot of cpu right now
[20:36] * MikeW (i=Mike@) Quit ()
[20:38] <_ph00> look like it was frost
[20:38] <_ph00> ...nope
[20:43] <_ph00> restarted the node (and once I was there I grebbed the latest r too)
[20:43] <_ph00> cpu use is back to normal
[20:46] <_ph00> soemthing's wrong with file inserton too
[20:46] <_ph00> I had one large file upload completed, now it's back to 93% completed
[20:46] <CIA-14> toad * r11494 /trunk/freenet/src/freenet/support/URLEncoder.java: Don't encode spaces; they are not dangerous, browsers etc know how to handle them.
[20:47] <_ph00> and two more files Ihad added to the insert queue are now gone
[20:47] <_ph00> (looks like the queue went back in time)
[20:49] <Beta_M> this is really annoying, uploads are *not* resuming for me
[20:49] <nextgens> hi
[20:50] <_ph00> hi
[20:50] <nextgens> toad_> well, aren't you worried by spaces "looking like" the same ?
[20:50] <_ph00> this is already the second time my quesues (up and down load) go back in time
[20:51] <_ph00> and also, at least the second time I notice that all my peers has reconnected symoultaneously
[20:51] <_ph00> (or whatever you spell that)
[20:51] <_ph00> have*
[20:52] <_ph00> quesues = queues (I never get use to use that word)
[20:52] <_ph00> d
[20:52] * sanity (n=ian@) Quit ()
[20:54] <nextgens> toad_> I suggest we enforce location checking on LocationManager : ie raise an exception if it's not in between 0 and 1
[20:55] <nextgens> [18:33] < toad_> | nextgens: yes we had a discussion about it
[20:55] <nextgens> toad_> what's the final answer then ? is opennet you top priority now ?
[20:56] <_ph00> oh, sorry if I bother, but I have one more detail to add: the lage file that got completed first and back to 9% uploaded than, it's showing the key anyways, not like the other incomplete uploads that show 'unknown'. I' tried to re-download the file using that key, it went straight to 80% downloaded, and it doesn't seem to have any problem... so far
[20:57] <_ph00> large* ; 39%*
[20:58] <toad_> nextgens: there's lots ot be done before i can even start on opennet; specifically i have to get 1008 out the door
[20:58] * whiterabbit (n=Whiterab@) Quit ("KVIrc 3.2.5 Anomalies http://www.kvirc.net/")
[20:59] <toad_> nextgens: i might include work on the New Plugins system - especially if Bombe does it for me
[20:59] <toad_> or I might not, in which case old plugins will break
[20:59] <toad_> and we'll have to ask people to upgrade manually
[20:59] <toad_> nextgens: as far as implementing opennet immediately goes, i'm still not entirely sure
[20:59] <_ph00> "N2NTM your peers!"
[21:00] <toad_> nextgens: i'm trying to find out what the cause of the Wierd Clustering Behaviour is; it seems quite plausible that it represents poor routing caused by a non-small-world and hierarchical topology caused by #freenet-refs
[21:01] <toad_> nextgens: which makes me think that implementing opennet soon may not be a bad idea
[21:01] * sanity (n=ian@) has joined #freenet
[21:01] * ChanServ sets mode +o sanity
[21:01] <nextgens> :/
[21:01] <toad_> rehi sanity
[21:01] <sanity> toad_: howdy
[21:01] <nextgens> and what about checking parameters of the funcion first ?
[21:01] <toad_> nextgens: what function?
[21:01] <nextgens> "just in case"
[21:02] <toad_> I asked mrogers what the problems were with opennet and he sent me several issues, none of them really specific to load management however
[21:02] <nextgens> the method on LocationManager
[21:02] * nextgens will do
[21:02] <toad_> my response is that a) lack of opennet seems to be causing bad topology and bad routing, though we can't yet be sure about it
[21:03] <toad_> hmmm
[21:03] * toad_ forgets b)
[21:03] * vfervers_ (n=vfervers@) has joined #freenet
[21:03] <toad_> but a) is fairly important imho IF it is true
[21:03] <toad_> i'd like to be more sure of that; that's why i've been working on probeall: traces etc
[21:05] <toad_> however i'm not comfortable with having a fixed, arbitrary and short timetable between the first non-working deployment of new load balancing, and the deployment of opennet which may mess things up even more
[21:05] <toad_> so i'm not really convinced either way
[21:05] <toad_> what is clear is that we need to get the major bugs out of 1008, and it deployed, before doing any of this
[21:05] <toad_> so i've done some work on that
[21:05] <toad_> if we deploy 1008 now everyone will have to replace their plugins manually
[21:06] <toad_> because of refactorings
[21:06] <toad_> and later on we will deploy the new plugins system, hopefully bombe will sort it out over xmas
[21:06] <toad_> hopefully that will include auto-update of plugins over freenet, so we will be able to auto-upgrade people from the existing system
[21:08] <toad_> sanity: any ideas on how to investigate the clustering issue?
[21:08] <toad_> sanity: other than "deploy opennet and see if it goes away" ? ;)
[21:11] <sanity> hmm
[21:11] <toad_> sanity: anything obvious and quick?
[21:11] <toad_> sanity: to prove that it's caused by defective network topology and therefore by lack of opennet?
[21:11] <sanity> set up a local network of, 5-10 nodes and see what happens
[21:11] <toad_> i doubt there'd be any swapping on a small network
[21:11] <toad_> i suppose it's a good sanity check though ... chain network?
[21:11] <sanity> its just very suspicious that it would cluster around 0.0 - there should be nothing special about that number
[21:11] <toad_> right
[21:12] <toad_> if there's something systematic relating to 0.0, then it might well show up on a small network
[21:12] <toad_> okay, i'll see what i can do ...
[21:12] <toad_> bbl, 1 hour
[21:13] <CIA-14> nextgens * r11495 /trunk/freenet/src/freenet/node/PeerManager.java: Ensure the location is in between 0.0 and 1.0 on PeerManager.distance()
[21:15] * vfervers_ (n=vfervers@) Quit (Remote closed the connection)
[21:33] <Beta_M> there is something *seriously* wrong with freenet
[21:33] <Beta_M> it's removing inserts/downloads now
[21:33] <Beta_M> all by itself
[21:33] <Beta_M> q;-/
[21:34] <Beta_M> it's just not remembering any changes
[21:35] <Beta_M> it's resetting priorities between restarts
[21:35] <Beta_M> and is doing a lot of other mean stuff
[21:37] * whiterabbit (n=Whiterab@) has joined #freenet
[21:37] * lokadin_ (n=loki@) has joined #freenet
[21:37] * whiterabbit (n=Whiterab@) Quit (Remote closed the connection)
[21:38] * lokadin (n=loki@) Quit (Read error: 113 (No route to host))
[21:42] <Ralith> ATravelingGeek, you there?
[21:46] * mozillaman (n=self@) Quit (Read error: 60 (Operation timed out))
[21:55] * mozillaman (n=self@) has joined #freenet
[21:55] * greycat (i=rfc1413@) Quit ("This time the bullet cold rocked ya / A yellow ribbon instead of a swastika")
[21:56] <JustMe> Beta_M: Build #1007 r11495 just reset a 107 meg insert that was at 24% to zero.
[21:57] <Beta_M> q;-/
[21:57] <Beta_M> mine got reset to 11% for the 3rd time
[21:57] <Beta_M> from 90%+
[21:57] <JustMe> Yikes!!
[21:58] <Beta_M> at least the file will be *very* well spread
[21:58] <JustMe> :)
[21:59] <Beta_M> 2 downloads disappeared, 1 finished download is starting over, 3 downloads where i've changed the priority are back to 'low', and a new insert is no longer there...
[21:59] <Beta_M> may i assume that something doesn't write the progress
[21:59] <JustMe> I noticed that sometimes it was removing all but the first insert on the list.
[22:00] <Beta_M> if it crashes due to starting more than one... yes
[22:01] <JustMe> Hmmm these were manual upgrades/restarts.
[22:02] <Beta_M> intersting
[22:03] <Beta_M> i didn't have *that* particular problem unless the crash was almost right after beginning of the follow up inserts
[22:03] <Beta_M> like if i tried to add 10 files to the queue, it added all, crashed, and then after restarting only one was there
[22:04] <JustMe> I used the stop.cmd. I wonder if using the web interface would make a difference. I'll test it.
[22:05] * Fennes is now known as Fennes[zzz]
[22:32] * railk (n=railk@) Quit ("Leaving")
[22:40] <_ph00> I had that problem, files in the queue diappearing, or going back to a previous point, one of them even after finishing a upload, went back to 93 or 94% done; and I'm talking with another guy who experienced similar problems. the only thing I can add to that, is that files that have been in the queu for a longer time don't seem to be affected by this problem; while newly added files tend to disappear or 'roll back', mostly at node restart
[22:46] * ffangel (n=freeform@) has joined #freenet
[22:46] * ffangel (n=freeform@) has left #freenet
[22:58] * sajdhsjkd (i=kuponka@) has joined #freenet
[23:01] * lattt (n=gerhard@) has joined #freenet
[23:06] <sajdhsjkd> i can't get freenet to install, would anyone be able and willing to help?
[23:07] <toad_> Beta_M: it sounds like it's just not writing the downloads.dat.gz, is this correct?
[23:07] <toad_> Beta_M: is this the most recent build? I thought I'd fixed serialization ..
[23:07] <Beta_M> i would assume that it's the problem
[23:07] <Beta_M> i'm using SVN build
[23:08] <Beta_M> built it a couple of hours ago
[23:09] <toad_> JustMe: / Beta_M: are you sure it's because of the recent build? it might be an older build you were running that didn't save the data? i'm reasonably sure i fixed it ...
[23:10] <Beta_M> toad_: it wasn't happening until yesterday
[23:10] <Beta_M> and now it's still going on
[23:17] <JustMe> toad_ I can't say that I am sure about the latest build. I did restart it once and the insert went from 24% to zero when I upgraded to r11495. Now I have a PROBEALL: running and don't want to restart.
[23:18] * mozillaman_ (n=self@) has joined #freenet
[23:19] * mozillaman (n=self@) Quit (Read error: 104 (Connection reset by peer))
[23:19] * mozillaman_ is now known as mozillaman
[23:20] <Beta_M> i'm going to test it right now by restarting after changing priorities on some files
[23:22] * mozillaman (n=self@) Quit (Client Quit)
[23:23] <Beta_M> hmmm.. interesting
[23:23] <Beta_M> it's actually changed correctly
[23:24] <Beta_M> toad_: it might be something with my particular setup, because i had the node "update" to the older version this morning
[23:24] <Beta_M> (i've upgraded to the SVN build, and then autoupdate decided that it didn't like that)
[23:24] <Beta_M> so you might be right, it could be something up with the older version
[23:25] <Beta_M> if i'll have more problems i'll let you know
[23:29] * horatius (n=gold@) has joined #freenet
[23:30] * horatius (n=gold@) has left #freenet
[23:31] <sajdhsjkd> man, that installer is stuck at "downloading thaw" what do i do
[23:50] * sajdhsjkd (i=kuponka@) Quit ()
Irc logs of #freenet : 2008 2007 2006 2005
These logs were automatically created by FreenetLogBot on chat.freenode.net using the Java IRC LogBot.