Timestamps are in GMT/BST.
[0:01] <toad_> port number A is the official port number
[0:01] <toad_> and get this
[0:01] <toad_> packets that are sent to port B are ignored !
[0:02] <toad_> so we have a clear violation of the rule that if you remap the port number, you should remap it for both incoming and outgoing traffic
[0:02] * toad_ wonders if it could be a symmetric NAT doing this?
[0:02] <toad_> or maybe it's the result of having both the port forwarded and it being NATed? but if it's forwarded anyway why change the port number?
[0:03] <toad_> anyway, it's totally bizarre
[0:04] <toad_> the obvious fix is to send packets to both ports
[0:04] <toad_> at the moment, the first phase of connection setup is sent to both ports
[0:04] <toad_> and the rest we just reply to the address that replied to us
[0:05] <toad_> we could send all packets during auth to both ports, but that wouldn't help a great deal, because then you have to consider the situation once it's actually connected
[0:05] <toad_> well actually it might work
[0:06] <toad_> because once we're connected we send to ... hmm, where do we send to?
[0:06] * Jase (i=jase@) Quit (zelazny.freenode.net irc.freenode.net)
[0:07] * Jase (i=jase@) has joined #freenet
[0:07] * xxxxx is now known as Apophis2
[0:07] <toad_> we send them to peernode.detectedPeer
[0:08] <toad_> which would be the port from which we are receiving packets :<
[0:08] <toad_> this works very well for most nodes
[0:08] <toad_> a further complication is we need to handle the case where the port has actually changed
[0:08] <toad_> normally taking the reply address is the only way to make it work
[0:09] <toad_> so what can we do?
[0:09] <toad_> for a single handshake, only send it to one address
[0:09] <toad_> pick that address when originating (if we are the originator)
[0:10] <toad_> if it completes, we know that the address is viable
[0:10] <toad_> this means we have to maintain a) the current detected address and b) the current validated address
[0:11] <toad_> ... and if we start getting packets from a third address (as we would when the IP changes), what do we do?
[0:12] <toad_> detected address = address we most recently received a packet from
[0:12] <toad_> validated address = address we know that it will receive packets from, because we've had round-trips when sending packets to that address
[0:13] <toad_> okay, so if we start getting packets from a third address ...
[0:14] <toad_> if the detected address == the validated address, pick up the new detected address
[0:15] <toad_> and once we have a round-trip (a packet we send is acknowledged), copy that to the validated address too
[0:16] <toad_> if the new address is a new address but the same port as the detected address, try sending to the new address with the validated port
[0:16] <toad_> if the new address is the same address but with a different port, panic
[0:16] <toad_> if the new address is the same address but with the same port, panic
[0:16] <toad_> errr
[0:17] <toad_> if the new address is the same address with a different port, panic
[0:17] <toad_> :)
[0:17] <Zothar> does sbc have other connections?
[0:17] <toad_> Zothar: i believe so
[0:17] <toad_> Zothar: his node detects his IP perfectly
[0:17] <Zothar> so how/why do they work, but the one with you doesn't?
[0:17] <toad_> it also a) includes it in his ref with his real specified port
[0:18] <toad_> and b) detects the changed port number he is sending out packets on
[0:18] <toad_> Zothar: good question
[0:18] <toad_> but here's the evidence:
[0:18] <toad_> phase 0 packets always get through to zothar's node
[0:18] <toad_> phase 2 packets never get through to it
[0:18] <toad_> why? because we send phase 0 to both addresses
[0:19] <toad_> in fact, not even phase 1's get through
[0:19] <toad_> because they're only sent to the reply address also
[0:20] <toad_> but amphibian does see sbc's responses
[0:20] <toad_> i dunno, would a NAT remap the port on outgoing but not on incoming?
[0:21] <toad_> is it a plausible theory?
[0:22] <toad_> source NAT but not destination NAT
[0:22] <toad_> hrrrm
[0:24] * toad_ notes that if such a crappy NAT were combined with another port restricted NAT, no communication would be possible at all
[0:25] * oierw (n=oierw@) has joined #freenet
[0:27] <Zothar> OK, so A is Toad, B is sbc: A0 -> By, B1 -> Ay, A2 -> Bn; and: B0 -> Ay, A1 -> B?
[0:27] * Iceman_B^Ltop (n=ice@) Quit ("Swan is hotter than Jasmine ^_^ -=SysReset 2.53=-")
[0:28] <toad_> Zothar: not sure I follow
[0:28] <toad_> Zothar: it receives on one port (the port set in the config), and sends on another
[0:29] <Zothar> -> is packet direction; y is received, n is not received and ? is I'm asking/don't know
[0:29] <Zothar> 0,1, 2 are phase
[0:29] <toad_> phase 0: A sends to B1 and B2
[0:29] <toad_> phase 1: B picks up B1 and sends back to A. packet source address is set to B2
[0:30] <toad_> phase 2: A receives packet, apparently from B2, and sends a reply to B2
[0:30] <toad_> phase 3: B does not receive the packet because B is only listening on B1
[0:30] <Zothar> what about when B sends phase 0?
[0:31] <toad_> phase 0: B sends to A. packet source address is set to B2
[0:31] <toad_> phase 1: A receives packet, apparently from B2, and sends a reply to B2
[0:31] <toad_> phase 2: B does not receive packet, because it's only lisening on B1
[0:32] <Zothar> so B has two different ports and there's no logical way to know that 1 is input, 2 is output; maybe some way to share this in the ref?
[0:33] <toad_> yes
[0:33] <toad_> there are hacks we could do
[0:33] <toad_> but if we're going to hack it and require user intervention then we should just detect it and tell the user
[0:33] <toad_> what would be better would be to detect it and handle it transparently
[0:34] <toad_> but that throws up the problem above - what if the IP changes?
[0:34] <toad_> still, the principle of sending to one address and using that if it works, is attractive
[0:35] <toad_> Zothar: we DO know the two port numbers. one was specified in the node's config, and is therefore appended to the STUN-detected IP. the other was detected by other nodes sending it the port number after it connected to them
[0:35] <toad_> although how it managed to connect to anyone i'm not quite clear!
[0:36] <Zothar> we've got x1:y1 and x1:y2 becomes x2:y3(maybe) and x2:y2; if we know B2, (the node's FNP port), then we know that y2 is probably constant at least
[0:36] <nextgens> toad_> load balancers are causing such behaviours
[0:36] * phrosty (n=phrosty@) Quit ("baseball is wrong: man with four balls cannot walk.")
[0:36] <nextgens> toad_> btw, that's what is preventing me from connecting to sanity
[0:37] <nextgens> it's not PAT but NAT
[0:37] <toad_> Zothar: for most NATed nodes, we want to reply to the reply address, because that was remapped along with the source address
[0:37] <toad_> nextgens: so it's common? ugh
[0:37] <nextgens> not common
[0:37] <nextgens> not that common
[0:37] <nextgens> ask him whether he is in a corporate environement
[0:37] <Zothar> is either B1 or B2 sbc's FNP port or are neither somehow?
[0:38] <toad_> i think he's at Google
[0:38] <nextgens> ;)
[0:38] <toad_> nextgens: it's only the port that changes, not the address
[0:38] <toad_> nextgens: both ports are on the same address
[0:38] <nextgens> toad_> that doesn't change anything
[0:38] <nextgens> if you're rich enough you do NAT
[0:38] <toad_> nextgens: please explain how this occurs then?
[0:38] <nextgens> otherwise you do PAT
[0:38] <toad_> why would you want to do something demonstrably inferior?
[0:38] <nextgens> NAT is less expensive but requires more than one ipv4 address
[0:39] * edt (n=Ed@) has joined #freenet
[0:39] * ChanServ sets mode +o edt
[0:39] <toad_> remapping the source address but not then accepting packets from the remapped address is demonstrably inferior; it will sabotage almost any UDP-based service
[0:39] <Zothar> toad_: you have a class C of real IPs and your network was already using a class C of non-Internet-routed IPs is one...
[0:39] <nextgens> toad_> that's on purpose
[0:39] <toad_> nextgens: so why not just block UDP entirely?
[0:40] <nextgens> toad_> it allows the network appliance to do stateless things
[0:40] <nextgens> because dns relies on it ? :)
[0:40] <toad_> how can DNS even survive in such an environment?
[0:40] <nextgens> dns and multicast are the why they keep it
[0:40] <toad_> does it have a fixed sender port ?
[0:41] <nextgens> on multicast it doesn't matter as it's on the lan (before the appliance)
[0:41] <toad_> it has a fixed destination port, but if the requestor's request is remapped then the requestor will miss the reply
[0:41] <Zothar> globally, no, but it can be set for all requests from a particular resolver
[0:41] <nextgens> and for dns it's the same
[0:41] <toad_> nextgens: in which case, why not block UDP completely?
[0:41] <nextgens> I dunno
[0:41] <Zothar> toad_: is either B1 or B2 sbc's FNP port or are neither somehow?
[0:41] <toad_> nextgens: he does receive the packets which are sent to both ports
[0:41] <nextgens> well, I don't see any strong reason why not
[0:42] <toad_> Zothar: B1 is the official specified FNP port
[0:42] <nextgens> toad_> but not all of them, right ?
[0:42] <toad_> on which it listens
[0:42] <toad_> B2 is the remapped port
[0:42] <toad_> outgoing packets have their source address remapped to B2
[0:42] <toad_> nextgens: not all of whaty?
[0:42] <nextgens> not all packets
[0:42] <toad_> nextgens: I completely fail to see the point of this; if they want to block UDP, they can just block UDP
[0:43] <nextgens> the point is not blocking it
[0:43] <toad_> nextgens: ALL outgoing packets from sbc's node to me have the source address set to B2
[0:43] <Zothar> ok, that's what I thought; so we send to the port we received from and if that attempt times out, we try a ref declared FNP port (the first one in the physical.udp, if we decide that standard)
[0:43] <nextgens> the point is not spending too much ressources filtering it
[0:43] <toad_> nextgens: yes but ... there's nothing to filter
[0:43] <nextgens> ^^-^^
[0:44] <toad_> only apps which are heavily over-engineered, or which completely ignore the source address (meaning they won't work on PATs) will be able to function in such an environment
[0:44] <nextgens> it's probably fragmenting/defragmenting things
[0:44] <nextgens> I guess it's on purpose anyway
[0:44] <toad_> if it was on purpose they'd just block all UDP
[0:44] <toad_> there has to be some reason for leaving UDP open
[0:44] <Zothar> toad_: what do you think?
[0:45] <toad_> it's much easier to block all UDP than to go through this mess
[0:45] <toad_> and it's not stateless anyway
[0:45] <toad_> the source port is consistently remapped to a specific port number different from the original one
[0:45] <nextgens> imo it is
[0:46] <toad_> nextgens: hmm?
[0:46] <nextgens> they just randomize the source port within a range given to a "client"
[0:46] <toad_> nextgens: one-way function you think? would collisions be a problem?
[0:46] <toad_> nextgens: the source port is always the same
[0:47] <nextgens> what I don't understand is why it's not working for you : you're port-forwarded, aren't you ?
[0:48] * nextgens will re-read backlog more carefully
[0:48] <toad_> toad/dark is directly connected to the internet
[0:48] <toad_> i can see the incoming packets
[0:48] <toad_> the problem is that they have their source port rewritten to a port on which the node does not listen
[0:48] <nextgens> because that ought to matter only if your node isn't reachable
[0:48] <toad_> no, it matters because i reply to the incoming address
[0:48] <Zothar> nextgens: problem is receive from B2, but we can't send back to B2, we have to send to B1
[0:48] <nextgens> well, that shouldn't be a problem
[0:48] <toad_> i could change it to always send to the official address
[0:49] <toad_> but if i did that then PATs wouldn't work
[0:49] <nextgens> the problem is we aren't "matching" it, right ?
[0:49] <toad_> no
[0:49] <toad_> the problem is that it tells us to send to port Y, but it listens on port X
[0:49] <toad_> so it only receives the initial packets
[0:49] <toad_> phase 0 packets are sent to all known addresses
[0:49] <toad_> which include both X and Y
[0:49] <nextgens> ah ok
[0:49] <toad_> but everything else is sent to the detected address
[0:49] <toad_> i.e. Y
[0:50] <toad_> and we have to do it that way for it to work on PATs
[0:50] <nextgens> some nats are rewritting the port, some aren't
[0:50] <toad_> well the problem is that it's asymmetric
[0:50] <toad_> it's rewriting the port in one direction only
[0:51] <toad_> apparently he did manage to connect to at least one other node though
[0:51] <toad_> otherwise it wouldn't know about Y
[0:51] <toad_> and we wouldn't even be able to send phase 0
[0:52] <toad_> hmmm wrong way around
[0:52] <TheSeeker|Gone> connected:17 backed off:1 ... o.O 1003 is looking good on the backoff front.
[0:52] <toad_> it's X
[0:52] <Zothar> toad_: maybe phase 1 should include the port we receive phase 0 on by including the destination port in the packet as well?
[0:52] <toad_> the first node he connects to works, because it doesn't know about Y
[0:52] <toad_> so it sends to X
[0:52] <toad_> except that the reply would still have the wrong port ... so it still wouldn't work ... hmmm
[0:53] <toad_> Zothar: how would that help?
[0:53] <toad_> brb
[0:55] <Zothar> toad_: toad sends phase 0 to sbc on X and Y, with X inside the FNPMessage oX and Y in FNPMessage on Y; sbc received the packet to X and includes that fact in it's phase 1 packet back to toad; toad now knows that sbc can see packets to X?
[0:58] <toad_> here's an idea
[0:58] <toad_> what it could do:
[0:58] <toad_> it doesn't remap the receiving ports
[0:59] <toad_> so conceivably you might be able to run UDP-based servers inside the firewall
[0:59] <toad_> but you can't access UDP-based servers outside it
[0:59] <toad_> hence, you can run corporate streaming servers
[0:59] <toad_> but your employees can't access external porn streams and udp-based filesharing
[0:59] <toad_> might that be it?
[1:00] <toad_> Zothar: hmmm
[1:01] * oCc (n=nodeOne@) Quit (Read error: 110 (Connection timed out))
[1:01] <toad_> Zothar: we can pick up which address we received it on in the packet, you say?
[1:01] <nextgens> toad_> yes, that's what is deployed usually
[1:01] <nextgens> toad_> but in our case, the thing is they don't block it
[1:01] <toad_> nextgens: hmm?
[1:01] <nextgens> I mean external udp connections are still possible
[1:02] <toad_> nextgens: the one-sided-remapping thing provides the above without you having to whitelist your corporate streaming servers
[1:02] * Latka2 (n=victor@) Quit ("using sirc version 2.211+KSIRC/1.3.12")
[1:02] <nextgens> keep in mind that it's likely there is several layers of firewalls in between
[1:03] <nextgens> and only borders one would filter that
[1:03] <toad_> no
[1:03] <toad_> it's not filtering
[1:03] <nextgens> the only point of doing port-remapping is imho :
[1:03] <toad_> simply by remapping the source port on outgoing traffic, while allowing incoming traffic as-is, you eliminate any possibility of employees using udp-based p2p
[1:03] <nextgens> 1) spare some space in the tracking table
[1:03] <toad_> while allowing the running of UDP servers internally on well-known ports
[1:04] <nextgens> 2) fight against udp holepunching
[1:04] <toad_> because the incoming port is fixed
[1:04] <nextgens> and the second point is what is disturbing you
[1:04] <toad_> 1) is wrong, as I have explained it has to be stateless as it has a stable changed source port on outgoing
[1:04] <nextgens> but assume it's on purpose
[1:04] <toad_> AFAICS the situation is that they have a separate NAT on outgoing packets
[1:04] <toad_> I mean, it's a different set of routers
[1:05] <toad_> a different outgoing route to the incoming one
[1:05] <TheSeeker|Gone> toad_: should I need to restart my node to allow external access to fproxy using the allowed hosts filter in the options?
[1:05] <nextgens> toad_> the static port can be the sum of all the members of the ip address ;)
[1:05] <toad_> TheSeeker|Gone: no idea
[1:05] <nextgens> I mean determined by the ip address
[1:05] <toad_> nextgens: can it?
[1:05] <nextgens> there is no need to keep track of it
[1:05] <toad_> nextgens: collisions are not a problem?
[1:05] <toad_> and it can't be recalculated on incoming?
[1:06] <toad_> no, it can't
[1:06] <toad_> because we don't know the destination address ...
[1:06] <nextgens> not if you use a power of two or something as a coefficient
[1:06] <toad_> except that we do know the destination internal address because THE PACKET IS DELIVERED
[1:07] <Zothar> toad_: I'm saying we can pick up what the destination port was for the packet because we put that in the FNPMessage part of the packet before sending it; essentially, we send different packets to each address for phase 0 so that we can get that info from phase 1
[1:07] * nextgens is experiencing difficulties to explain things over irc
[1:07] <toad_> okay so lets assume that there is 1:1 mapping of internal to external addresses
[1:07] <toad_> like nextgens said, NAT not PAT
[1:07] <toad_> so you don't need to track address mappings
[1:07] <nextgens> maybe because it's late here too
[1:07] <toad_> you can calculate the new port from the IP
[1:07] <toad_> well, from the IP, and the old port
[1:08] <nextgens> it's possible even on a PAT
[1:08] <nextgens> more complex and costy but still possible
[1:08] * nextgens will think about it tonight
[1:08] <nextgens> tomorrow maybe :)
[1:09] <toad_> nextgens: i don't see why it has to be an irreversible mapping to be stateless
[1:09] <nextgens> still, I don't think that's an urgent matter anyway :)
[1:09] <toad_> so that leaves us with #2
[1:09] <toad_> well, it prevents ian's node from working, it prevents sbc's node from working
[1:09] <nextgens> it doesn't prevent them from working
[1:09] <toad_> and on principle we ought to be able to get through corporate firewalls if our goal is to get through the Golden Shield
[1:09] <nextgens> it prevents them from connecting to some peers
[1:09] <toad_> it prevents them from connecting
[1:09] <toad_> to most peers
[1:10] <toad_> it prevents them from connecting to me
[1:10] <toad_> and I'm on the open internet
[1:10] <toad_> i'm not behind any kind of NAT
[1:10] <toad_> so how they manage to connect to anyone else is beyond me
[1:10] <nextgens> maybe you're just blacklisted ? :)
[1:10] <toad_> no
[1:10] <toad_> if that was the case I wouldn't see any UDP packets
[1:10] <toad_> but I do
[1:10] <nextgens> rate-limited ?
[1:10] <toad_> he sees my packets and I see his packets
[1:11] <toad_> but the source port is wrong on his packets
[1:11] <toad_> so we don't connect
[1:11] <nextgens> understood :)
[1:11] <nextgens> maybe we could have a fallback mechanism
[1:11] <toad_> which probably has more to do with anti-p2p than statelessness
[1:11] <toad_> nextgens: such as?
[1:11] <toad_> what I was thinking about was this:
[1:12] <toad_> we only send handshakes to one IP at a time
[1:12] <toad_> we remember which IP that was
[1:12] <Zothar> toad_: have you followed what I'm saying? (I realize you're intensely discussing with nextgens :)
[1:12] <nextgens> after three failled handshakes on post phase2, we ignore source port
[1:12] <nextgens> so that we keep on sending packets where they work
[1:12] <toad_> if the setup succeeds, we remember that; we have the validated address and the detected address
[1:13] <nextgens> btw, has he stun enabled ?
[1:13] <toad_> problem is what happens when we get a valid packet from an address other than the current detected address?
[1:13] <nextgens> what does it say ?
[1:13] <toad_> nextgens: it detects his IP address
[1:13] <toad_> nextgens: I dunno what it says about his firewall
[1:13] <toad_> well maybe he doesn't have it enabled
[1:13] <nextgens> that's what I'm enquiering
[1:13] <toad_> if they have real internet addresses internally
[1:13] <toad_> dunno, would be worth asking, he hasn't told me
[1:13] <nextgens> don't you have his node's logs ?
[1:14] <toad_> some of them
[1:14] <toad_> STUN wouldn't normally run if we know our IP
[1:14] <nextgens> ok, anyway I'm off to bed
[1:14] <nextgens> cya tomorrow
[1:14] <toad_> ok, cya
[1:14] <nextgens> but I suggest you spend time on binary blobs or anything else insteed :)
[1:15] * nextgens might write some code to test his nat tomorrow
[1:15] <toad_> well, connectivity issues are a concern
[1:15] <toad_> nextgens: you have an Evil Corporate NAT?
[1:15] <nextgens> yep :)
[1:15] <nextgens> more than one
[1:15] <toad_> hehe
[1:16] <toad_> well there are ways we can get around it
[1:16] <toad_> we've got two above
[1:16] <toad_> one is to remember the address we _SENT_ the handshake to
[1:16] <toad_> the other is to try ignoring the source port/address after some number of failures after having received packets
[1:17] <nextgens> toad_> get a recent version of inkscape ... so that we can discuss it in front of a whiteboard tomorrow
[1:17] <nextgens> :)
[1:17] <toad_> inkscape does whiteboarding?
[1:17] <nextgens> yes, over jabber
[1:17] <toad_> cool
[1:17] <toad_> ok
[1:17] <toad_> i'll file a bug
[1:18] <toad_> "Work over evil corporate NATs designed to keep people in not hackers out"
[1:18] <toad_> :)
[1:18] <toad_> good night
[1:18] <nextgens> you too
[1:21] * mozillaman (n=chatzill@) Quit ("Chatzilla 0.9.77 [Firefox 2.0/2006101023]")
[1:25] * electrosy (n=steo@) has left #freenet
[1:26] * DebolazX (n=DebolazX@) Quit (Read error: 60 (Operation timed out))
[2:07] * IMCensored1 (n=KMIntern@) Quit (Read error: 145 (Connection timed out))
[2:09] * Danilo (i=Danilo@) Quit ()
[2:16] <CIA-14> toad * r11128 /trunk/freenet/src/freenet/node/ (Node.java PacketSender.java): getPeer() not getDetectedPeer()
[2:16] <CIA-14> toad * r11129 /trunk/freenet/src/freenet/node/PeerNodeStatus.java: getPeer() not getDetectedPeer(), also minor optimisation
[2:17] <CIA-14> toad * r11130 /trunk/freenet/src/freenet/node/TextModeClientInterface.java: getPeer() not getDetectedPeer()
[2:17] <Zothar> toad_: got an N2NM/N2NTM update in a moment
[2:17] <toad_> okay, wait 1min
[2:18] * Zothar waits for toad's go ahead before committing
[2:20] <CIA-14> toad * r11131 /trunk/freenet/src/freenet/ (4 files in 3 dirs):
[2:20] <CIA-14> Add new peer option ignoreSourcePort. If this is enabled, we will search for
[2:20] <CIA-14> another address with the same address as the detected peer but a different port,
[2:20] <CIA-14> and use that if we find one. This enables us to connect through corporate NATs.
[2:20] <CIA-14> Not currently auto-detected; must be set manually.
[2:20] <CIA-14> Document the "On selected peers" options in the dropdown a bit.
[2:20] <toad_> now you can commit
[2:20] <CIA-14> Delete superfluous PeerNode.getDetectedPeer() in favour of getPeer()
[2:20] <Zothar> OK. Will test build with changes and then commit
[2:27] <toad_> Zothar: what are your changes?
[2:27] <Zothar> stuff to trace N2NMs, which we can use to figure out what's going on if the flood bug is still there (somehow)
[2:28] <toad_> ok cool
[2:28] <toad_> good night
[2:28] * toad_ (n=toad@) Quit (Remote closed the connection)
[2:32] * oierw (n=oierw@) Quit (Read error: 110 (Connection timed out))
[2:36] <CIA-14> zothar * r11132 /trunk/freenet/src/freenet/ (3 files in 2 dirs): (log message trimmed)
[2:36] <CIA-14> Fix multiple extraPeerDataType bug on N2NTMs that were queued by the sender and
[2:36] <CIA-14> then given a second extraPeerDataType when storing on the receiver. This
[2:36] <CIA-14> appears it might have caused the deletion of the N2NTM to fail possibly. Also
[2:36] <CIA-14> send a N2NTM as a N2NM when we can when sending queued N2NTMs, not just when not
[2:36] <CIA-14> queued. Add some tracing information to help debug the N2NTM flood bug if
[2:36] <CIA-14> this+toad_'s changes haven't already fixed it; some can be used for other
[2:36] * Zothar (n=Zothar@) Quit ("Chatzilla 0.9.74 [Firefox 1.5.0.8/2006102516]")
[2:47] * sleon (n=sleon@) Quit (Read error: 110 (Connection timed out))
[3:00] * edt (n=Ed@) Quit ("Leaving")
[3:12] * xxfooxx (n=foo@) has joined #freenet
[3:32] * sleon (n=sleon@) has joined #freenet
[3:32] * ChanServ sets mode +o sleon
[3:47] * timmy2chk (n=violent@) Quit (Read error: 110 (Connection timed out))
[4:07] * phrosty (n=phrosty@) has joined #freenet
[4:07] * xxfooxx (n=foo@) Quit ("Leaving")
[4:42] * PraiseChaos (n=kcecil@) Quit (Remote closed the connection)
[5:26] * Mac109 (n=mac109@) has joined #freenet
[5:27] * Mac109 (n=mac109@) has left #freenet
[5:45] * Aranjedeath (n=aranjede@) has joined #freenet
[5:49] * Ash-Fox (i=UNKNOWN@) Quit (Remote closed the connection)
[5:52] * Ash-Fox (i=UNKNOWN@) has joined #FreeNET
[6:48] * mazzanet (n=mazzanet@) Quit (Read error: 110 (Connection timed out))
[7:07] <_ph00> nextgens; around?
[7:08] <CIA-14> jflesch * r11133 /trunk/apps/Thaw/src/thaw/plugins/index/IndexManagementHelper.java: Fix index adding
[7:08] <Aranjedeath> he aint on the list
[7:08] <_ph00> he's on "my" list
[7:08] <Aranjedeath> not logged into this channel
[7:08] <_ph00> (as "away" tho)
[7:08] <Aranjedeath> oh
[7:08] <Aranjedeath> wait
[7:08] <Aranjedeath> hes admin of sorts
[7:08] <Aranjedeath> or op
[7:08] <Aranjedeath> heh
[7:09] <Aranjedeath> didnt see him
[7:09] <_ph00> hes a channel op and one of the major freenet devs
[7:11] <CIA-14> jflesch * r11134 /trunk/apps/Thaw/src/thaw/i18n/ (thaw.properties thaw_fr.properties): Fix text for JLabel just before the search field
[7:18] <Aranjedeath> woah
[7:18] <Aranjedeath> freenet has log bots
[7:18] <Aranjedeath> nice
[7:20] * Aranjedeath (n=aranjede@) Quit ()
[7:22] * outkid (n=marijuan@) Quit (Read error: 54 (Connection reset by peer))
[7:27] * outkid (n=marijuan@) has joined #freenet
[7:29] * LucisFerens (i=user@) Quit (Remote closed the connection)
[7:41] * IMCensored1 (n=KMIntern@) has joined #freenet
[7:49] * LucisFerens (i=user@) has joined #freenet
[7:49] * Aranjedeath (n=aranjede@) has joined #freenet
[7:56] * hottuna (n=robbansw@) Quit (Read error: 104 (Connection reset by peer))
[7:58] <_ph00> [OT] how many of you guys actually thinks this is true?
[7:58] <_ph00> http://www.abortiontv.com/Glitch/AbortedBabiesSoldAsHealthFood.htm
[7:58] <_ph00> (ev. answersn on #freenet-chat plz)
[7:59] * Aranjedeath (n=aranjede@) Quit ()
[8:09] * mazzanet (n=mazzanet@) has joined #freenet
[8:49] * whiterabbit (n=Whiterab@) has joined #freenet
[9:08] * sannes (i=ace@) Quit (zelazny.freenode.net irc.freenode.net)
[9:08] * christl_s (i=christin@) Quit (zelazny.freenode.net irc.freenode.net)
[9:08] * Jflesch (n=jflesch@) Quit (zelazny.freenode.net irc.freenode.net)
[9:08] * paveq_ (n=paveq@) Quit (zelazny.freenode.net irc.freenode.net)
[9:08] * lokadin_ (n=loki@) Quit (zelazny.freenode.net irc.freenode.net)
[9:08] * Baughn (n=svein@) Quit (zelazny.freenode.net irc.freenode.net)
[9:08] * mazza[W] (i=mazzanet@) Quit (zelazny.freenode.net irc.freenode.net)
[9:08] * nextgens (n=nextgens@) Quit (zelazny.freenode.net irc.freenode.net)
[9:08] * fridim_ (n=fridim@) has joined #freenet
[9:08] * nextgens (n=nextgens@) has joined #freenet
[9:08] * Jflesch (n=jflesch@) has joined #freenet
[9:08] * sannes (i=ace@) has joined #freenet
[9:08] * Baughn (n=svein@) has joined #freenet
[9:08] * paveq_ (n=paveq@) has joined #freenet
[9:08] * lokadin_ (n=loki@) has joined #freenet
[9:08] * christl_s (i=christin@) has joined #freenet
[9:08] * mazza[W] (i=mazzanet@) has joined #freenet
[9:27] <_ph00> After trying to reach him many times both here and using a n2ntm, I eventually decided to disable anonymouse's node until I hear from him
[9:27] <_ph00> according to my darknet page, he's still running #994
[9:28] <_ph00> I sent him at least 2 n2ntm's and tried to make his irc client blink by using his name here, but I didn't get any answer, yet he seems to be on line here, and his node seems to be up all the time
[9:28] <_ph00> BUT
[9:29] <_ph00> as he never answered to my messages, I assume he's not paying any attention to the node, and it's been that way for some time
[9:30] <_ph00> ...which made me kinda "susicious": I don't understand why someone would let a node run that long without even looking the the web interface once in 10 days or more
[9:30] <_ph00> suspicios*
[9:31] <_ph00> (and no, I don't usually wear tin foil hats...)
[9:31] <_ph00> (but I'm considerng making one)
[9:31] <_ph00> :P
[9:37] * Savoy (n=chatzill@) has joined #freenet
[9:45] <_ph00> I'm so damn bad at typing/spelling...
[9:55] * railk (n=railk@) has joined #freenet
[9:59] * mazzanet (n=mazzanet@) Quit ("Off to toy with routing")
[10:06] * anonick (i=anonymou@) has joined #freenet
[10:15] * anonick (i=anonymou@) Quit (Client Quit)
[10:16] * Savoy (n=chatzill@) Quit (Read error: 104 (Connection reset by peer))
[10:21] * Savoy (n=chatzill@) has joined #freenet
[10:27] * railk (n=railk@) Quit (Read error: 110 (Connection timed out))
[10:28] <IMCensored1> i did that once... leave my node running in the bground for a few days ... but thats cause i went out of town for a few days... and then i needed to dl some music quick (for a couple parties) and hopped on morpheus for a couple days
[10:30] <nextgens> _ph00> hi
[10:31] <_ph00> hey... (I forgot what I wanted to say, I'll be back as soon as I remember)
[10:31] <_ph00> ah, right
[10:32] * nextgens (n=nextgens@) has left #freenet
[10:33] <IMCensored1> but ya Ph00:... there was a time when ure node was connected to me for a couple weeks... then all the sudden i never saw it connected... but i watched your version rise from 993 to 999 while the whole time the disconnection time ascended above 2 weeks
[10:35] <_ph00> weird
[10:35] <_ph00> maybe id *did* get *some* packet every now and then
[10:36] <_ph00> they disconnected right after connecting, or something
[10:36] <_ph00> ...I think I saw something like that on one of my peers...
[10:36] <CIA-14> mrogers * r11135 /trunk/apps/load-balancing-sims/phase7/sim/Sim.java: Ten percent slow nodes
[11:13] * IMCensored1 (n=KMIntern@) Quit ("Peace and Protection 4.22")
[11:14] * IMCensored1 (n=KMIntern@) has joined #freenet
[11:19] * whiterabbit (n=Whiterab@) Quit (Read error: 110 (Connection timed out))
[11:31] * railk (n=railk@) has joined #freenet
[11:39] * MineHaunter (n=van85@) has joined #freenet
[11:41] * toad_ (n=toad@) has joined #freenet
[11:41] * ChanServ sets mode +o toad_
[11:50] * _ph00 (n=z@) Quit ("Leaving")
[11:50] * LucisFerens (i=user@) Quit (Client Quit)
[12:16] * mime (n=6d696d65@) has joined #freenet
[12:17] * mime is now known as alberech
[12:19] * mazzanet (n=mazzanet@) has joined #freenet
[12:21] <alberech> I am -ing anonymous :)
[12:21] <alberech> at last never to be whois`d
[12:23] <alberech> just thought I would say that
[12:23] <alberech> got to go thousands of other channels to visit
[12:23] * alberech (n=6d696d65@) Quit ("www.cotse.net")
[12:25] * DebolazX (n=DebolazX@) has joined #freenet
[12:27] * nextgens (n=nextgens@) has joined #freenet
[12:27] * ChanServ sets mode +o nextgens
[12:28] * Zothar (n=Zothar@) has joined #freenet
[12:37] * whiterabbit (n=Whiterab@) has joined #freenet
[13:05] <CIA-14> zothar * r11136 /trunk/apps/pyFreenet/ (fcp/node.py refbot.py): refbot: Added a version command. Re-ordered a few methods. Expanded the help text with more information about running a refbot.
[13:09] * edt (n=Ed@) has joined #freenet
[13:09] * ChanServ sets mode +o edt
[13:10] <edt> toad just rebooted here. Noticed that one node's name changed after the restart of freenet. Before it looked like random garbage, after all the names are normal again.
[13:11] <edt> this was in the darknet peers frame
[13:13] <CIA-14> zothar * r11137 /trunk/apps/pyFreenet/ (fcp/node.py refbot.py): refbot: Add message about registering the owner's IRC nick to the setup process. Set the svn:keywords on fcp/node.py for Revision substitution.
[13:16] <edt> on another topic. The various fixes from about 999 up are helping here. My bandwidth limit is still set high. Difference is that now the node _does_ use more bw on once in a while, the peeks are much higher too (20+ k/s vs 8-9 before the recient work)
[13:16] <edt> btw the run showing the name corruption was v1004
[13:19] * ferrox (i=ferrox@) Quit (Read error: 110 (Connection timed out))
[13:22] * greycat (i=rfc1413@) has joined #freenet
[13:27] * railk (n=railk@) Quit ("Cya, wouldn't want ta be ya!")
[13:41] * sbc (n=sbc@) has joined #freenet
[13:48] * Savoy (n=chatzill@) has left #freenet
[13:49] * Fennes (i=1234@) Quit (Read error: 131 (Connection reset by peer))
[13:50] <toad_> edt: so the node uses more than the limit? or what?
[13:52] * bitch22 (n=chatzill@) has joined #freenet
[13:52] <CIA-14> zothar * r11138 /trunk/freenet/src/freenet/node/fcp/ModifyPeer.java: Fix setting IgnoreSourcePort via FCP.
[13:53] * bitch22 (n=chatzill@) Quit ("Chatzilla 0.9.77 [Firefox 2.0/2006101023]")
[13:54] * mozillaman (n=chatzill@) has joined #freenet
[14:02] * pinkopalla (n=chatzill@) has joined #freenet
[14:07] * sleon (n=sleon@) Quit (Read error: 110 (Connection timed out))
[14:10] <toad_> woah, mrogers' simulations are looking rather interesting
[14:10] <toad_> if we understand what's wrong with the current mechanism we can tweak it in the short run
[14:11] <toad_> and in the long run design something better
[14:11] <Zothar> so when can we implement... :)
[14:12] <toad_> we could implement the bottom level changes now
[14:12] <toad_> the mid level congestion control stuff needs to be written up
[14:12] <toad_> after that, the simulations may suggest some optimisations that can be made to the current algorithm
[14:12] <toad_> and after that, it's a matter of sorting out token passing / some other new load limiting algorithm
[14:13] <sbc> toad_: I can't find the ignoreSourcePort you talked about in you mail.
[14:14] <toad_> sbc: hmmm?
[14:14] <toad_> sbc: you've upgraded to the latest testing build?
[14:14] <sbc> The drop down menu. Cant find it.
[14:14] <toad_> it doesn't matter anyway, it's your peers who should set the option
[14:15] <toad_> not you
[14:15] <sbc> ohh, it's not in 1004 yet?
[14:15] <sbc> toad_: I know it's my peers, but I can't really expect them to set a option I myself can't find.
[14:15] <Zothar> we have the needed details for the bottom level changes now then? are we planning to do these changes all at once or as we get details?
[14:15] * sbc tries testing.
[14:18] <toad_> Zothar: probably do the low-level stuff with the mid-level stuff
[14:18] <toad_> we need mrogers to write up the mid-level stuff
[14:18] <toad_> or me to
[14:18] <toad_> but the simulations are going well, so i'm happy to give him some more time
[14:19] <toad_> bbiab
[14:20] * Urs_ShPo (n=gaim@) has joined #freenet
[14:20] * Urs_ShPo (n=gaim@) Quit (Remote closed the connection)
[14:20] * Urs_ShPo (n=gaim@) has joined #freenet
[14:20] * mozillaman_ (n=chatzill@) has joined #freenet
[14:21] * mozillaman (n=chatzill@) Quit (Nick collision from services.)
[14:21] * mozillaman_ is now known as mozillaman
[14:51] * _Danilo_ (i=Danilo@) has joined #freenet
[14:51] * _Danilo_ (i=Danilo@) Quit (Client Quit)
[14:58] * mozillaman (n=chatzill@) Quit ("Chatzilla 0.9.77 [Firefox 2.0/2006101023]")
[15:01] * xxfooxx (n=foo@) has joined #freenet
[15:01] * Aranjedeath (n=aranjede@) has joined #freenet
[15:01] <xxfooxx> what's the panic button do?
[15:03] <MineHaunter> it makes your computer explode
[15:03] <xxfooxx> hm
[15:03] <MineHaunter> :P
[15:03] <MineHaunter> it deletes everything from the global queue
[15:03] <xxfooxx> why would i want to do that?
[15:04] * Aranjedeath (n=aranjede@) Quit (Client Quit)
[15:04] <MineHaunter> dunno...
[15:05] <MineHaunter> but you know, when one is in panic does stupid things :P
[15:05] <nextgens> hey toad_
[15:05] <makomk> Hmmm... I see the FreeNetWatch lot have discovered that Firefox 2.0 happily accepts RSS feeds sent as text/plain (which seems like an odd thing for it to do).
[15:06] <nextgens> that's dangerous imo
[15:06] <nextgens> as they can bundle external links, can't they ?
[15:09] <makomk> And possibly images, too.
[15:15] <makomk> I think it's most problematic for Firefox 2.0 users with no default RSS reader set (those with it set to Live Bookmarks can only be tricked into subscribing and then clicking an entry with the URL set to an external site).
[15:15] <CIA-14> zothar * r11139 /trunk/apps/pyFreenet/ (refbot.py updater.py): refbot/updater: Fix the updater when run with 'python updater.py' instead of './updater.py'. Fix refbot so that it won't accept a node name that will result in a bot IRC nick that's too long.
[15:17] * whiterabbit (n=Whiterab@) Quit (Read error: 60 (Operation timed out))
[15:19] <makomk> Whereas if you don't have a default RSS reader, it'll happily display any text file on Freenet that's a valid RSS feed as one - including referenced external images :-(
[15:20] <xxfooxx> sorry i guess im just not understanding, why would i want to clear the global queue with the panic button? Can anyone provide an example of when it would be used?
[15:21] <Zothar> 273 downloads got started and you don't want to have to delete each one?
[15:22] <xxfooxx> oh, so it's a quick way to purge/cancel lots of downloads?
[15:22] * kasvio (n=kasvio@) has joined #freenet
[15:24] <Zothar> I believe so, yes
[15:24] <Zothar> (I haven't read that code, or at least not in awhile)
[15:26] <xxfooxx> can anyone confirm the bandwidth settings are actually in kiloBYTES per second? as opposed to kbit/s
[15:26] <nextgens> makomk> btw, "not honoring content-type" is why we have banned ie users from using fproxy
[15:26] <nextgens> makomk> that's a security bug
[15:26] <nextgens> makomk> have you reported it to mozilla's team ?
[15:26] <nextgens> makomk> shall I ?
[15:27] <nextgens> in the meantime we might ban rss feeds using the content filter
[15:27] <nextgens> but that's a hack
[15:27] <makomk> I haven't reported it. I suppose you could try reporting it, but outside of Freenet it's not really very useful.
[15:27] * outkid (n=marijuan@) Quit ("Lost terminal")
[15:28] <makomk> The RSS preview seems to filter the HTML quite tightly (no JavaScript, and it looks like CSS might be stripped too...)
[15:28] <nextgens> well, things like privoxy might be affected too
[15:28] * _ph00 (n=z@) has joined #freenet
[15:29] <nextgens> and still, there is no reason for the browser not to honor the content-type
[15:29] <nextgens> too bad I can't start a node from here and check
[15:31] <Zothar> xxfooxx: bytes per second: 15K is 15KiBps
[15:33] <nextgens> brb
[15:34] <xxfooxx> thanks Zothar
[15:40] <CIA-14> zothar * r11140 /trunk/apps/pyFreenet/refbot.py: refbot: Explain that the bot's IRC nick will have _bot added to the end of the given usernick value.
[15:41] * mozillaman (n=chatzill@) has joined #freenet
[15:41] * mozillaman wants to know what's new in #1004 :)
[15:43] <Zothar> mozillaman: a CSS filter fix
[15:43] <Zothar> it was an exploitable vulnerability and the sum total of the differences with #1003 I believe
[15:45] <mozillaman> Thanks :)
[15:46] * Fennes (i=1234@) has joined #freenet
[15:47] * Zothar is now known as refbottest1_bot
[15:47] * refbottest1_bot is now known as Zothar
[15:48] <toad_> makomk: does it auto-detect images? html?
[15:48] <makomk> toad_: don't think so, no.
[15:51] * sleon (n=sleon@) has joined #freenet
[15:51] * ChanServ sets mode +o sleon
[15:56] <makomk> Hmmm... looks like it auto-detects RSS feeds served as image/jpeg though, but not as text/html.
[16:18] * fridim_ (n=fridim@) Quit ("Quitte")
[16:20] * Zothar is now known as refbottest1_bot
[16:21] * refbottest1_bot is now known as Zothar
[16:29] * NullAcht15 (n=NullAcht@) has joined #freenet
[16:31] * bitch22 (n=chatzill@) has joined #freenet
[16:36] * LucisFerens (i=z@) has joined #freenet
[16:40] * christl_s (i=christin@) Quit (zelazny.freenode.net irc.freenode.net)
[16:40] * sannes (i=ace@) Quit (zelazny.freenode.net irc.freenode.net)
[16:40] * NullAcht15 (n=NullAcht@) Quit (zelazny.freenode.net irc.freenode.net)
[16:40] * Fennes (i=1234@) Quit (zelazny.freenode.net irc.freenode.net)
[16:40] * Baughn (n=svein@) Quit (zelazny.freenode.net irc.freenode.net)
[16:40] * paveq_ (n=paveq@) Quit (zelazny.freenode.net irc.freenode.net)
[16:40] * lokadin_ (n=loki@) Quit (zelazny.freenode.net irc.freenode.net)
[16:40] * Jflesch (n=jflesch@) Quit (zelazny.freenode.net irc.freenode.net)
[16:40] * mazza[W] (i=mazzanet@) Quit (zelazny.freenode.net irc.freenode.net)
[16:40] * nextgens (n=nextgens@) Quit (zelazny.freenode.net irc.freenode.net)
[16:40] * nextgens (n=nextgens@) has joined #freenet
[16:40] * NullAcht15 (n=NullAcht@) has joined #freenet
[16:40] * Fennes (i=1234@) has joined #freenet
[16:40] * Jflesch (n=jflesch@) has joined #freenet
[16:40] * sannes (i=ace@) has joined #freenet
[16:40] * Baughn (n=svein@) has joined #freenet
[16:40] * paveq_ (n=paveq@) has joined #freenet
[16:40] * lokadin_ (n=loki@) has joined #freenet
[16:40] * christl_s (i=christin@) has joined #freenet
[16:40] * mazza[W] (i=mazzanet@) has joined #freenet
[16:44] <mozillaman> Nice bounce :)
[16:48] * LucisFerens (i=z@) Quit (Client Quit)
[16:50] <CIA-14> zothar * r11141 /trunk/apps/pyFreenet/minibot.py: refbot: Fix the bot not continuing after successfully registering with nickserv.
[16:50] * _ph00_ (n=z@) has joined #freenet
[16:54] * _ph00 (n=z@) Quit (Nick collision from services.)
[16:54] * _ph00_ is now known as _ph00
[17:00] <CIA-14> zothar * r11142 /trunk/apps/pyFreenet/ (minibot.py refbot.py): refbot: Remove some unused variables
[17:01] * Iceman_B^Ltop (n=ice@) has joined #freenet
[17:07] * pinkopalla_ (n=chatzill@) has joined #freenet
[17:14] * pinkopalla (n=chatzill@) Quit (Read error: 145 (Connection timed out))
[17:16] * christl_s (i=christin@) Quit (zelazny.freenode.net irc.freenode.net)
[17:16] * sannes (i=ace@) Quit (zelazny.freenode.net irc.freenode.net)
[17:16] * NullAcht15 (n=NullAcht@) Quit (zelazny.freenode.net irc.freenode.net)
[17:16] * Fennes (i=1234@) Quit (zelazny.freenode.net irc.freenode.net)
[17:16] * Baughn (n=svein@) Quit (zelazny.freenode.net irc.freenode.net)
[17:16] * paveq_ (n=paveq@) Quit (zelazny.freenode.net irc.freenode.net)
[17:16] * lokadin_ (n=loki@) Quit (zelazny.freenode.net irc.freenode.net)
[17:16] * Jflesch (n=jflesch@) Quit (zelazny.freenode.net irc.freenode.net)
[17:16] * mazza[W] (i=mazzanet@) Quit (zelazny.freenode.net irc.freenode.net)
[17:16] * nextgens (n=nextgens@) Quit (zelazny.freenode.net irc.freenode.net)
[17:17] * nextgens (n=nextgens@) has joined #freenet
[17:17] * NullAcht15 (n=NullAcht@) has joined #freenet
[17:17] * Fennes (i=1234@) has joined #freenet
[17:17] * Jflesch (n=jflesch@) has joined #freenet
[17:17] * sannes (i=ace@) has joined #freenet
[17:17] * Baughn (n=svein@) has joined #freenet
[17:17] * paveq_ (n=paveq@) has joined #freenet
[17:17] * lokadin_ (n=loki@) has joined #freenet
[17:17] * christl_s (i=christin@) has joined #freenet
[17:17] * mazza[W] (i=mazzanet@) has joined #freenet
[17:17] * Zothar is now known as refbottest1_bot
[17:18] * refbottest1_bot is now known as Zothar
[17:36] * pinkopalla_ (n=chatzill@) Quit ("Chatzilla 0.9.77 [Firefox 1.5.0.8/2006102516]")
[17:44] * MikeW (i=Mike@) has joined #freenet
[17:45] * Zothar is now known as refbottest1_bot
[17:45] * Urs_ShPo (n=gaim@) Quit ("Leaving.")
[17:45] * refbottest1_bot is now known as Zothar
[17:46] * TheShado (n=no@) Quit (Read error: 113 (No route to host))
[17:53] <CIA-14> zothar * r11143 /trunk/apps/pyFreenet/ (minibot.py refbot.py): refbot: Tweak nickserv identify process.
[17:56] * TheShado (i=no@) has joined #freenet
[18:24] * mickymouse (n=chatzill@) has joined #freenet
[18:31] * mickymouse (n=chatzill@) has left #freenet
[18:40] * Yukishiro (n=zhaan@) has joined #freenet
[18:43] * MikeW (i=Mike@) Quit ()
[18:44] * bitch22 (n=chatzill@) has left #freenet
[19:23] * Ash-Fox (i=UNKNOWN@) Quit ("Do not meddle in the affairs of kitsune, for you are crunchy and good with ketchup.")
[19:25] * Ash-Fox (i=UNKNOWN@) has joined #FreeNET
[19:45] <mozillaman> I have a question
[19:46] <mozillaman> If I make a freesite, then decide I don't want to use it, and never tell anyone about it
[19:46] <Zothar> I have an answer; let's see if the two match
[19:46] <mozillaman> The key or anything
[19:46] <CIA-14> zothar * r11144 /trunk/freenet/src/freenet/ (3 files in 2 dirs): Added input/output rate limits to /stats/ and /darknet/. Added useNonBreakingSpace option to SizeUtil.formatSize(). Use the new option in several places to get the 'word wrap' to be a little more friendly.
[19:46] <mozillaman> Would it be possible for someone else to find it two months later?
[19:46] <mozillaman> Considering the inserting node was destroyed.
[19:47] <Zothar> not without _a lot_ of computing power to be able to decrypt all blocks until whatever unknown item was found
[19:47] <mozillaman> Well, it's happened.
[19:47] <mozillaman> A freesite I inserted, never mentioned, and deleted the data from the same day
[19:47] <mozillaman> is now on indeica three months later.
[19:47] <Zothar> the inserting node being destroyed doesn't really have anything to do with the longevity of the item and that's one of the things Freenet is trying to achieve
[19:47] <toad_> mozillaman: if you don't tell anyone about it, nobody can find it
[19:48] <mozillaman> toad_: I thought so.
[19:48] <toad_> if it showed up on indicia then logically somebody did find out about it
[19:48] <Zothar> right, I said it a different way :)
[19:48] * mozillaman isn't exceptionally happy it's there.
[19:48] <toad_> well, freenet didn't do it
[19:48] <Zothar> apparently something knew about it somewhere
[19:48] <mozillaman> I mean
[19:48] <mozillaman> somehow, it was 'discovered'
[19:48] <toad_> did you tell anyone about it?
[19:48] <toad_> even a top developer who should have known not to post it publicly?
[19:49] <toad_> you sure you didn't post it to the index?
[19:49] <toad_> or mention it on frost?
[19:49] <Zothar> unfortunately, you can't ask the index operator to remove it or that draws attention to it :(
[19:50] <toad_> yeah, he'd probably post your request, then everyone would visit it
[19:54] * Iceman_B (n=Ice@) has joined #freenet
[19:54] * Aladin (n=lars@) has joined #freenet
[19:57] <Zothar> any estimates on how far away 1005 is?
[20:01] <toad_> xxfooxx: bandwidth settings are in BYTES per second
[20:01] <toad_> so "20k" is 20,000 bytes / second
[20:01] <CIA-14> jflesch * r11145 /trunk/apps/Thaw/src/thaw/ (core/MainWindow.java plugins/QueueWatcher.java): Use a JSplitPane to split downloads / uploads in the transfer queue tab
[20:01] <toad_> 20K is 20,480 bytes/sec :)
[20:02] <toad_> nextgens: / makomk we need to talk about ff 2.0 :<
[20:03] <toad_> makomk: does it auto-detect HTML served as text/plain?
[20:04] * zxcv_ (i=geoff@) has joined #freenet
[20:06] * Zothar is now known as refbottest1_bot
[20:06] * refbottest1_bot is now known as Zothar
[20:09] <toad_> nextgens: sadly there is a reason for the browser not to honor the content-type. it's called web servers. :<
[20:11] * toad_ supposes we will just have to whitelist browsers which can handle plain text, and to every other browser serve a <pre>'ed HTMLEncode'ed version of the original plain text
[20:11] <toad_> that's actually very easy to implement, and it would make freenet compatible with Internet Exploder
[20:11] * whiterabbit (n=Whiterab@) has joined #freenet
[20:11] <toad_> well with the exception of application/octet-stream, but you get a warning with that, and you get force-download headers
[20:12] * zxcv_ (i=geoff@) Quit ("I love you all!")
[20:13] <makomk> Nope; it seemos to correctly treat HTML served as text/plain as plain text...
[20:13] <toad_> so it only sniffs for RSS
[20:13] <toad_> in theory we could just check for anything that looks like RSS
[20:13] <toad_> we'd have to do that on _all_ documents
[20:13] <toad_> actually we'll have to do that anyway ... hmmm
[20:13] <toad_> because it will pick up RSS regardless of what type it's served as, correct?
[20:14] <toad_> an rss stream saved as blah.jpeg and served as image/jpeg will still be picked up as an RSS feed, yes?
[20:15] <makomk> Yep. text/plain and image/jpeg are treated as RSS, but text/html is still treated as HTML. IIRC, it sniffs the first 512 bytes (it's documented somewhere).
[20:16] <toad_> where is it documented?
[20:17] <makomk> https://bugzilla.mozilla.org/show_bug.cgi?id=338621 (the 512 bytes bit, anywhere - the content types probably aren't)
[20:17] <toad_> makomk: this is the shipped firefox 2.0?
[20:17] <makomk> *anyway
[20:18] <toad_> makomk: you're sure? /join #firefox
[20:21] <makomk> Nope; Gentoo version (but no interesting/relevant-looking patches). Hang on...
[20:26] <makomk> Yep; it works in the official Windows release of 2.0 too...
[20:29] * mozillaman (n=chatzill@) Quit (Read error: 110 (Connection timed out))
[20:31] * mozillaman (n=chatzill@) has joined #freenet
[20:32] * xxfooxx (n=foo@) Quit ("Leaving")
[20:33] <mozillaman> *sneeze*
[20:34] <CIA-14> zothar * r11146 /trunk/apps/pyFreenet/refbot.py: refbot: Now check that the node actually added the ref. Lengthen the default 'trade with me' and 'how to get me' message intervals. Tweak a error message.
[20:42] <CIA-14> jflesch * r11147 /trunk/apps/Thaw/src/thaw/plugins/QueueWatcher.java: Fix splitpane behavior
[20:53] <CIA-14> zothar * r11148 /trunk/apps/pyFreenet/refbot.py: refbot: Mention to the user the need to refresh their ref URL once they get a physical.udp line in their ref. Uncomment an exception handler that had gotten commented during debugging.
[20:54] * whiterabbit (n=Whiterab@) Quit ("KVIrc 3.2.5 Anomalies http://www.kvirc.net/")
[20:57] <toad_> okay, it looks like we're going to need some more work on the content filter(s)
[20:57] <toad_> i propose to:
[20:58] <toad_> 1. whitelist web browsers that don't autodetect anything from text/plain. all other browsers - IE, Firefox 2, in particular - get HTML-encoded versions of any text/plain we see. in other words, we <pre> it and escape all dangerous characters
[20:59] <toad_> 2. parse image types. not necessarily fully, but certainly the first 512 bytes, and certainly enough to determine that it's not really XML/RSS/etc (ff2's rss detector algorithm goes like this: look for <rss or <atom etc, check that all tags from the beginning to that tag are <! or <?)
[20:59] <toad_> 3. go through the CSS2 spec, carefully, and update it to be a proper whitelist filter.
[21:00] <toad_> 4. update the HTML spec to support XHTML properly
[21:00] <toad_> 5. write an RSS filter, if it's easy to do so
[21:01] <toad_> 6. write an SVG filter, if it's easy to do so
[21:01] <toad_> not necessarily all at once, or all before 0.7
[21:01] <toad_> but #1 and #2 are essential before 0.7
[21:01] <toad_> #3 is essential before 1.0 and important before 0.7
[21:13] * mozillaman (n=chatzill@) Quit ("Chatzilla 0.9.77 [Firefox 2.0/2006101023]")
[21:13] * MikeW (i=Mike@) has joined #freenet
[21:13] <CIA-14> jflesch * r11149 /trunk/apps/Thaw/src/thaw/plugins/ (QueueWatcher.java index/Index.java index/IndexTree.java): Bug https://bugs.freenetproject.org/view.php?id=914 should be fixed
[21:17] * Khhdl (n=chatzill@) has joined #freenet
[21:28] * toad_ changes topic to 'Do not use Firefox 2 with Freenet (until further notice) | http://freenetproject.org/download.html (990 - mandatory), please read that page before asking for help here. If you want to exchange references, join #freenet-refs | http://wiki.freenetproject.org/FrequentlyAskedQuestions | logs: http://emu.freenetproject.org/irc/ | Tor blocked due to spam (get an op to voice you) | #freenet-fr #freenet-se #freenet-es #freenet-it'
[21:30] <CIA-14> zothar * r11150 /trunk/apps/pyFreenet/refbot.py: refbot: The bot now checks for the existence of the same fields the Freenet node does before trying to add a ref and will now complain with the name of the field it first checked but didn't find.
[21:30] * auto98765 (n=auto9876@) has joined #freenet
[21:31] * toad_ writes email to devl ...
[21:32] <NullAcht15> Jflesch: Are you around? I'm having trouble using Thaw. I just checked it out from svn and compiled it. There's an exception at startup and then it hangs
[21:35] <CIA-14> jflesch * r11151 /trunk/apps/Thaw/src/thaw/plugins/ (IndexBrowser.java index/IndexCategory.java): By default, index "Thaw" is now added to the tree
[21:35] <Jflesch> NullAcht15: what exception exactly ?
[21:35] <NullAcht15> Jflesch: java.lang.NoSuchMethodError
[21:36] <Jflesch> NullAcht15: do an "ant clean" before compiling
[21:36] <NullAcht15> one moment, I'll paste it in a pastebin
[21:36] <NullAcht15> Jflesch: Okay, I'll try that
[21:37] <NullAcht15> ah yes, it works now. Thx!
[21:37] <Jflesch> NullAcht15: but you must be aware that thaw put the database where it is launched => if you run it from the build directory and then do an "ant clean", it will erase your thaw database
[21:37] <Jflesch> NullAcht15: you're welcome :)
[21:38] * auto98765 (n=auto9876@) Quit (Remote closed the connection)
[21:38] * PraiseChaos (n=kcecil@) has joined #freenet
[21:40] * Yukishiro (n=zhaan@) Quit ("Chatzilla 0.9.75 [Firefox 1.5.0.8/2006102516]")
[21:40] <NullAcht15> Jflesch: are you sure? I did just that and the index I added before i still there
[21:41] <Jflesch> NullAcht15: are you running thaw from the jar in the "lib" directory, or with "ant run" (and so from the build directory) ?
[21:41] <Jflesch> the 'lib' directory is not erased, but the 'build' directory is
[21:42] <Jflesch> (I should remove the target "run" in the build.xml ...)
[21:42] <NullAcht15> I'm running in apps/Thaw using java -jar lib/Thaw.jar
[21:42] <Jflesch> k, so then there is no problem :)
[21:44] * sbc (n=sbc@) Quit ("Ex-Chat")
[21:49] * greycat (i=rfc1413@) Quit ("This time the bullet cold rocked ya / A yellow ribbon instead of a swastika")
[21:52] <CIA-14> zothar * r11152 /trunk/apps/pyFreenet/minibot.py: refbot: Make the bot more accepting of different ways of addressing it.
[21:54] * phrosty (n=phrosty@) Quit ("baseball is wrong: man with four balls cannot walk.")
[21:59] <CIA-14> zothar * r11153 /trunk/apps/pyFreenet/refbot.py: refbot: Clean up some debugging.
[22:01] <CIA-14> jflesch * r11154 /trunk/apps/Thaw/src/thaw/plugins/Hsqldb.java: Ensure once again that database is written to disk when exit
[22:06] * railk (n=railk@) has joined #freenet
[22:06] * Iceman_B^Ltop (n=ice@) Quit (Read error: 113 (No route to host))
[22:07] * Zorix_ (n=Brandon@) has joined #freenet
[22:07] * phrosty (n=phrosty@) has joined #freenet
[22:08] * Zorix (n=Brandon@) Quit (Read error: 104 (Connection reset by peer))
[22:25] <CIA-14> zothar * r11155 /trunk/apps/pyFreenet/ (minibot.py refbot.py): refbot: Get the node's identity once instead of for every ref adding attempt. Tweak the bot's nickserv identify process some more.
[22:26] * NullAcht15 (n=NullAcht@) Quit (Remote closed the connection)
[22:32] <toad_> makomk: could you check something for me? will firefox detect an RSS stream which is encoded as UTF-16 ?
[22:32] <makomk> toad_: got an example of one?
[22:32] <toad_> no...
[22:33] <toad_> there are tools to convert though
[22:33] <toad_> some come with libc iirc
[22:33] * Iceman_B^Ltop (n=ice@) has joined #freenet
[22:35] <toad_> makomk: recode
[22:36] <makomk> ...has the worst-written and most cryptic manpage I've seen in a while.
[22:36] <toad_> iconv then
[22:36] <toad_> iconv -f iso-8859-1 -t utf-16 something.rss
[22:36] <toad_> iconv -f iso-8859-1 -t utf-16 something.rss -o something-utf16.rss
[22:39] <toad_> bbiab
[22:39] * toad_ has started work on the EvilHack
[22:39] <toad_> which will force the browser to save to disk anything that firefox might pick up as RSS
[22:41] <makomk> Nope, looks like it doesn't detect it as RSS if it's UTF-16...
[22:42] <CIA-14> jflesch * r11156 /trunk/apps/Thaw/src/thaw/plugins/ (4 files in 2 dirs): Replace the two JLabel in the 'state' tab by two buttons
[22:46] * kasvio (n=kasvio@) Quit ("Leaving")
[22:52] * Beta_M (n=Somebody@) Quit (Read error: 113 (No route to host))
[23:03] * railk (n=railk@) Quit ("Cya, wouldn't want ta be ya!")
[23:06] * poyt (n=anon@) has joined #freenet
[23:07] <poyt> What's the problem with ff2 & freenet?
[23:08] <Zothar> ff2 sees things that are RSS but don't have the RSS MIME type as RSS, bypassing FProxy's anonymity filtering
[23:09] <Zothar> (same with IE as well, I believe)
[23:09] <poyt> Ah, OK.
[23:20] * Beta_M (n=Somebody@) has joined #freenet
[23:45] * poyt (n=anon@) Quit ()
[23:49] * phrosty (n=phrosty@) Quit ("baseball is wrong: man with four balls cannot walk.")
[23:51] * phrosty (n=phrosty@) has joined #freenet
[23:57] * Viperb0y (n=Christia@) has joined #freenet
[23:58] <Viperb0y> Hi :) Is there a way to remove the freenet service at win xp by hand? The unstinstaller didn't work
Irc logs of #freenet : 2008 2007 2006 2005
These logs were automatically created by FreenetLogBot on chat.freenode.net using the Java IRC LogBot.