Timestamps are in GMT/BST.
[0:02] * JustMe_ (JustMe_@cs68206255-46.houston.rr.com) Quit ()
[2:10] * Sugadude (~Sugadude@202-63-160-59.broadband.isp.exatt.net) has joined #freenet
[2:17] <mikeDOTd> root: until kaffee meets the sun-java 1.4 spec (specificly full NIO support), freenet won't run under it
[3:05] * tyesinclair (~tyesincla@170.252.64.1) has joined #freenet
[3:06] * tyesinclair (~tyesincla@170.252.64.1) has left #freenet
[3:40] * pupok_ (~r00t@81-178-106-165.dsl.pipex.com) Quit ("Leaving")
[3:55] * sanity (~ian@81-178-106-165.dsl.pipex.com) Quit ()
[4:07] * sanity (~ian@81-178-106-165.dsl.pipex.com) has joined #freenet
[4:07] * sanity (~ian@81-178-106-165.dsl.pipex.com) Quit (Client Quit)
[4:08] * sanity (~ian@81-178-106-165.dsl.pipex.com) has joined #freenet
[4:23] * d-ArkAngel (~Rob@213-131-104-86.onyx.net) has joined #freenet
[4:24] * plixed (~plixed@i3ED6230C.versanet.de) has joined #freenet
[4:40] * plixed_ (~plixed@i3ED6212D.versanet.de) Quit (Read error: 110 (Connection timed out))
[5:34] * Ash-Fox (HAL-9000@abb65.neoplus.adsl.tpnet.pl) Quit ()
[5:39] * Hory (~Miranda@82.76.81.56) has joined #FreeNet
[5:54] * Hory (~Miranda@82.76.81.56) Quit (Read error: 104 (Connection reset by peer))
[6:12] * kevlAway (~kevloral@CZ1-RAS-8-u-0179.du.onolab.com) Quit (Read error: 110 (Connection timed out))
[6:27] * Hory (~Miranda@82.76.81.56) has joined #FreeNet
[6:38] <sanity> toad: here?
[6:42] <hobx> si it just me, or does "a virgin Mary sandwich" sound like porn to everybody else?
[6:42] <sanity> :-)
[6:42] <sanity> hobx: take a look at this: http://dijjer.sf.net/
[6:44] <hobx> dijjer do it, didn't jer?
[6:45] <sanity> i dij
[6:46] <hobx> can you dijjer that file for me?
[6:46] <hobx> Are you are dijjerer?
[6:46] <sanity> do you dij?
[6:46] <sanity> do you dij it?
[6:46] <hobx> Man this download is slow? Dude, just dijjer it!
[6:46] <sanity> dij it work?
[6:47] <hobx> I don't have time now
[6:47] <hobx> but I'll have a look at it tonight.
[6:47] <hobx> Bet your dijjer!
[6:48] <hobx> What happened with that chick? I dijjer!
[6:49] <sanity> you should work in marketing
[6:50] <hobx> I didn't come up with the name. I just appreciating your naming skillz!
[6:50] <sanity> i dijn't do it, steven dij
[6:51] <hobx> well give Steven my regards. I think he dij great!
[6:52] <hobx> off to work.
[6:52] <hobx> These early mornings are going to kill me!
[6:52] <sanity> i can imagine
[6:52] <sanity> i'm still awake from last
[6:52] <sanity> night
[6:59] * Hory (~Miranda@82.76.81.56) Quit ("CyberLore.net - Recommendations on the best games, freeware and websites.")
[7:11] * guido^pe (~unknown@dsl-213-023-244-006.arcor-ip.net) has joined #freenet
[7:11] * tyesinclair (~tyesincla@170.252.64.1) has joined #freenet
[7:12] * tyesinclair (~tyesincla@170.252.64.1) has left #freenet
[7:38] * sdogi (~java@84-50-20-115-dsl.kjj.estpak.ee) has joined #freenet
[7:49] * sdogi_ (~java@84-50-18-168-dsl.kjj.estpak.ee) Quit (Read error: 110 (Connection timed out))
[8:03] * deckard73 (~deckard73@D4788.d.pppool.de) has joined #freenet
[8:03] * deckard73 (~deckard73@D4788.d.pppool.de) has left #freenet
[8:23] <root> <man-di> root: I asked the freenet devs to give me simple testcases to reproduce the problems they have
[8:23] <root> can you give them that?
[8:34] * man-di (~man-di@dynadsl-080-228-94-035.ewetel.net) has joined #freenet
[8:34] <man-di> hi all
[8:35] <man-di> some freenet hacker here currently ?
[8:35] <sanity> man-di: whats up?
[8:35] <man-di> I'm involved in kaffe and other free Java efforts
[8:36] <man-di> freenet doesnt work with kaffe yet
[8:36] <sanity> i know, IIRC its due to problems with kaffe's java.nio implementation
[8:36] <man-di> can one of you send simple testcases that reproduce the problems so I can fix java.nio ?
[8:37] <man-di> I looked into freenet sources but they seem to be really chaotic and not well designed
[8:37] <man-di> sorry, but this was my impression
[8:38] <sanity> yeah, that is probably a fair criticism, due to the experimental nature of the project they were evolved rather than designed :-)
[8:38] <sanity> you need to talk to toad
[8:38] <sanity> he normally hangs out here, but doesn't seem to be awake right now
[8:38] <sanity> you may want to email him at toad@amphibian.dyndns.org
[8:39] <sanity> i am personally not familiar with the exact problem they encountered
[8:39] <man-di> yeah, I know him, he was in #kaffe several times already
[8:39] <man-di> I will contact him later here then
[8:39] <sanity> although i think toad was working with another kaffe developer on the problem, i think they met at FOSDEM
[8:39] <man-di> sanity: thanks for your help
[8:39] <sanity> man-di: no problem - good luck!
[8:39] <man-di> sanity: that is surely dalibor
[8:40] <man-di> I will go to FOSDEM too
[8:40] <sanity> i don't think that was his IRC nic
[8:40] <man-di> robilad ?
[8:40] <sanity> ah yes, that was it
[8:40] <man-di> hehe, thats the same person
[8:40] <sanity> ah, ok
[8:40] <sanity> gotta love irc :-)
[8:40] <man-di> one nick for from home, the other from work
[8:41] <man-di> one is the other reversed
[8:41] <sanity> so it is :-)
[8:41] <sanity> if you can't reach toad, he is probably up-to-date on the status of their debugging effort
[8:42] <sanity> if its urgent i can give you his cell phone number
[8:42] <man-di> well, dalibor just acts as interface to me as I maintian the whole NIO code
[8:42] <sanity> ah, i see
[8:42] <man-di> no, its not SOOOOO urgend
[8:42] <man-di> urgent
[8:42] <sanity> i will text him anyway and let him know that you were looking for him - we would really like to see kaffe work with Freenet :-)
[8:43] * man-di too
[8:43] <man-di> we got this request very often
[8:43] <man-di> is toad from UK ?
[8:44] <sanity> yes
[8:44] <sanity> as am i
[8:44] <sanity> well, not *from* the UK, but I live here now
[8:45] <man-di> that is what I meant, sorry
[8:45] <sanity> i just sent him a text letting him know that you are looking for him
[8:45] <man-di> I'm from Germany an dlive there too, so toad should available here at reasonable times for me
[8:45] <man-di> sanity: THANKS
[8:45] <sanity> yeah, he should be here now, he normally works tuesday-saturday
[8:46] <man-di> he works on freenet fulltime ?
[8:46] <sanity> yes, he is our fulltime developer
[8:46] <sanity> we pay for him out of donations through our website
[8:46] <man-di> aaah
[8:46] <sanity> so you can blame him for most of the messy code :-)
[8:47] <man-di> I know he write the code, he told me once
[8:47] <sanity> although in fareness, freenet is doing some pretty complicated things - i think even in an ideal world the code would be quite convoluted
[8:47] <sanity> s/fareness/fairness
[8:48] <sanity> man-di: slightly OT, but i am currently working on some code that is designed to work with Kaffe - i haven't tested it but it compiles and runs with GCJ, if you are interested i have a quick intro here :- http://dijjer.sf.net/
[8:49] <sanity> man-di: it is somewhat inspired by Freenet's design
[8:49] <sanity> man-di: but *much* simpler
[8:49] <sdogi> sanity: what is kaffe?
[8:49] <sanity> it doesn't use java.nio :-)
[8:49] <sanity> sdogi: its an open source implementation of the Java Runtime Environment
[8:49] <sanity> sdogi: see www.kaffe.org
[8:49] * sdogi is hoping that freenet can be announced ready for massive usage some day :)
[8:50] <sanity> sdogi: we are aiming for 2015 :-)
[8:51] <sanity> sdogi: at least we are making progress, slowly but surely...
[8:51] <man-di> sanity: at least you get enough donations for toad it seems
[8:51] <sdogi> well, maby on that year i am even participating :)
[8:51] <toad_> hi
[8:51] <sanity> man-di: just about
[8:52] <sanity> ah, speak of the devil
[8:52] <man-di> hi toad_
[8:52] <man-di> toad_: we really need to make kaffe/gij/jamvm and all the other free java vms support freenet
[8:53] <man-di> toad_: can you write some simple testcases that show the bugs you trigger in java.nio so I can more easily fix them and include them in out testsuite (mauve)
[8:55] <toad_> <sanity> so you can blame him for most of the messy code :-)
[8:55] <toad_> yup :)
[8:56] <toad_> man-di: hmmm
[8:56] <toad_> yes, i believe it's possible, but i don't know what the bugs are off hand
[8:56] <toad_> well, I might be able to..
[8:56] <toad_> if I had dalibor that might help significantly
[8:57] <man-di> dalibor is currently in boston
[8:57] <man-di> on some RedHat Java meeting
[8:57] <toad_> sanity: dijer looks nice
[8:57] <toad_> nod
[8:58] <toad_> okay
[8:58] <toad_> let me try starting it up..
[8:58] <sanity> toad: yeah, well, lots of work still to do
[8:58] <toad_> sanity: life has been rather high overhead lately..
[8:58] <sanity> also, its not easy getting people to test it without risking a slashdotting which would be premature
[8:58] <sanity> toad: for you?
[8:58] <toad_> yeah
[8:58] <sanity> why?
[8:59] * FallingBuzzard (~srademach@207.152.112.129) has joined #freenet
[9:00] <toad_> well, one reason is that I have an aikido grading on saturday, so I've been putting in quite a lot on that.. there are issues with dad and church, there are other things, there's the DTI meeting...
[9:00] <toad_> okay
[9:00] * toad_ starts freenet using kaffe...
[9:00] <sanity> well, i hope none of those things is too negative
[9:00] <toad_> not negative but often distracting
[9:01] <toad_> man-di: I'll get back to you
[9:01] <man-di> toad_: what version of kaffe are you suing ?
[9:01] <man-di> toad_: no problem, take your time
[9:02] <toad_> debian sid version... calls itself 1.1.x-cvs...
[9:02] <toad_> I can rebuild from CVS if you like?
[9:02] * toad_ does so..
[9:02] <toad_> is CVS reasonably functional at present?
[9:02] * spaetz (~spaetz@80-218-147-20.dclient.hispeed.ch) has joined #freenet
[9:03] <man-di> toad_: the debs are from 4 days ago afaik, so very recent
[9:04] <toad_> hmmm strange conflicts merging configure...
[9:04] * toad_ wonders how old his cvs tree for kaffe is..
[9:04] * toad_ will do a diff in a mo
[9:05] <toad_> man-di: I think I updated within that window, but not sure
[9:05] <toad_> ChangeLog head : 2004-10-15 Dalibor Topic <robilad@kaffe.org>
[9:05] <toad_> Compile date : Sat Oct 30 22:34:00 UTC 2004
[9:05] <toad_> Compile host : cyberhq
[9:05] <toad_> Install prefix : /usr/lib/kaffe/pthreads
[9:06] * toad_ updating..
[9:06] <toad_> Barbara Squyres hasn't got back to me..
[9:08] <toad_> if she doesn't get back to me by friday i'll just have to get a fax account somewhere
[9:11] <sanity> couldn't you call her?
[9:12] <toad_> i suppose...
[9:12] * Gunnar (~gunnar@34.80-203-209.nextgentel.com) Quit (Remote closed the connection)
[9:13] <toad_> sanity: I don't think she provided a phone number actually
[9:13] <toad_> just the fax line
[9:13] <toad_> and if I want to send a fax, I probably should be set up to receive them
[9:14] <toad_> know any good fax aggregation companies?
[9:14] <sanity> i guess
[9:14] <sanity> not in the uk
[9:14] <sanity> but then, i have never looked
[9:15] * toad_ wonders why -z isn't on by default for cvs...
[9:21] <toad_> okay, Kaffe grumbles that it can't load the native bigint acceleration...
[9:21] <toad_> no big deal, since kaffe doesn't need it..
[9:22] <toad_> timer granularity is 2
[9:22] <toad_> aha
[9:22] <man-di> are you on x86 ?
[9:22] <toad_> SEC 17/11/2004 14:21:59.021 (freenet.node.Node, main, ERROR): Encountered an une
[9:22] <toad_> xpected error when locking the lockfile '/var/lib/freenet-unstable/lock.lck'. Sh
[9:22] <toad_> utting down.
[9:22] <toad_> java.io.IOException: java.nio.FileChannelImpl.lock(): not implemented
[9:22] <toad_> at gnu.java.nio.channels.FileChannelImpl.lock (FileChannelImpl.java)
[9:22] <toad_> at gnu.java.nio.channels.FileChannelImpl.tryLock (FileChannelImpl.java:344)
[9:22] <toad_> at java.nio.channels.FileChannel.tryLock (FileChannel.java:279)
[9:22] <toad_> at freenet.node.Node.init (Node.java:3583)
[9:22] <toad_> at freenet.node.Main.main (Main.java:476)
[9:22] <toad_> man-di: yes
[9:22] <man-di> oh
[9:22] <man-di> isnt that merged into kaffe yet ?
[9:22] <toad_> man-di: we have some native code for BigInteger that we load..
[9:22] <man-di> damn
[9:22] <toad_> man-di: it may be in CVS
[9:22] <toad_> this is the debian build
[9:22] * pittaman (~mistery_b@d51A410CA.kabel.telenet.be) has joined #freenet
[9:22] * toad_ building cvs...
[9:23] <man-di> toad_: the debian package uses CVS from 4 days ago
[9:23] <toad_> we use a lock file since it's pretty common for users to have problems caused by accidentally starting the node twice
[9:23] <man-di> and I dont think dalibor merged it in that time
[9:23] <toad_> okay..
[9:23] <man-di> toad_: okay, I will tackle this
[9:23] <toad_> thanks
[9:23] <man-di> when I can reach dalibor he can merge i
[9:23] <man-di> it
[9:24] <man-di> toad_: then we can see for more errors, are you okay with this ?
[9:24] <toad_> man-di: sure; otherwise I can just take the lock call out :)
[9:24] <toad_> or turn it into a nonfatal ERROR
[9:24] <man-di> toad_: you can do this to, if you like
[9:24] <toad_> actually that's harder than it sounds (the latter) - it throws an IOE NOT an UnsupportedOpException
[9:24] <man-di> might be good
[9:25] <man-di> toad_: that got changed to IOE due to Eclipse 3 ...
[9:25] <toad_> hmm?
[9:25] <toad_> explain?
[9:26] <man-di> Eclipse uses file locking too and when we threw any non-IOE eclipse didnt started up
[9:26] <toad_> ah
[9:26] * guido^pe (~unknown@dsl-213-023-244-006.arcor-ip.net) Quit (Remote closed the connection)
[9:26] <toad_> but there's a patch to implement locking that hasn't been merged yet?
[9:27] <man-di> its implemented in libgcj and partly in classpath
[9:27] <man-di> normally kaffe just copies this sources
[9:28] <man-di> sometimes they rewrite it to use KNI, istead of JNI or CNI
[9:28] <toad_> what's the (approximate) difference?
[9:29] <toad_> native methods are actually called via JNI etc? that's interesting..
[9:29] <toad_> i mean built in methods
[9:30] <man-di> JNI is C, CNI is C++ and KNI is C too, but simpler then JNI
[9:30] <toad_> okay our JNI still doesn't work, but it's not important to freenet
[9:30] <toad_> it's detected its network address
[9:30] <toad_> good sign
[9:30] <toad_> does Kaffe work stably with pthreads? (sun doesn't, not for me...)
[9:31] <toad_> loading the datastore...
[9:32] <toad_> i've compiled kaffe cvs, should i install it?
[9:33] <toad_> loading estimators etc...
[9:33] <man-di> toad_: it should be relatively stable on x86
[9:33] <man-di> yeah, it should be installed
[9:33] <toad_> okay, after this run
[9:34] <toad_> starting selectors... aha
[9:34] <toad_> SEC 17/11/2004 14:34:38.254 (freenet.transport.ReadSelectorLoop, Network reading
[9:34] <toad_> thread, ERROR): Caught throwable in AbstractSelectorLoop!: java.lang.ArrayIndex
[9:34] <toad_> OutOfBoundsException
[9:34] <toad_> java.lang.ArrayIndexOutOfBoundsException
[9:34] <toad_> at java.lang.String.toLowerCase (String.java:555)
[9:34] <toad_> at java.lang.String.toLowerCase (String.java:549)
[9:34] <toad_> at freenet.transport.AbstractSelectorLoop.mySelect (AbstractSelectorLoop.java
[9:34] <toad_> :586)
[9:34] <toad_> at freenet.transport.AbstractSelectorLoop.loop (AbstractSelectorLoop.java:766
[9:34] <toad_> )
[9:34] <toad_> at freenet.transport.ReadSelectorLoop.run (ReadSelectorLoop.java:544)
[9:34] <toad_> at java.lang.Thread.run (Thread.java:666)
[9:34] <toad_> at java.lang.VMThread.run (VMThread.java:124)
[9:34] <toad_> hmmmm
[9:34] * toad_ will try to find what's happening there since it's in our code
[9:35] <toad_> L586 says "String msg = e.getMessage().toLowerCase();"
[9:35] <toad_> where e is an IOException
[9:35] <toad_> thrown by the Selector
[9:35] * greycat (~wooledg@192.35.79.70) has joined #freenet
[9:35] <man-di> hmmm
[9:36] <man-di> toad_: can you write a simple testcase that reproduces this ?
[9:36] <toad_> man-di: maybe, once i find out what's happening
[9:36] <man-di> toad_: that would help a lot
[9:36] <toad_> ;)
[9:37] <toad_> for all i can see here it might be a string handling bug
[9:38] <man-di> maybe a native string that gets handled wrong
[9:38] <man-di> I just wonder why this exception occurs
[9:39] <toad_> Caught java.io.IOException: Interrupted system call
[9:39] <toad_> java.io.IOException: Interrupted system call
[9:39] <toad_> at gnu.java.nio.VMSelector.select (VMSelector.java)
[9:39] <toad_> at gnu.java.nio.SelectorImpl.select (SelectorImpl.java:227)
[9:39] <toad_> at freenet.transport.AbstractSelectorLoop.mySelect (AbstractSelectorLoop.java
[9:39] <toad_> :577)
[9:39] <toad_> that's the exception then...
[9:39] <man-di> hmmm
[9:40] * kevloral (~kevloral@CZ1-RAS-8-u-0179.du.onolab.com) has joined #freenet
[9:40] <kevloral> g'afternoon all
[9:40] <toad_> and there's another one even wierder...
[9:40] <toad_> wait
[9:40] <toad_> that's not it
[9:40] <toad_> ok
[9:40] * toad_ restartsd
[9:42] <toad_> SEC 17/11/2004 14:43:15.389 (freenet.node.Main, main, ERROR): Failed to load service: distribution
[9:42] <toad_> ouch
[9:43] <toad_> ReadSelectorLoop is resetting constantly...
[9:43] <toad_> but the ArrayIndexOutOfBounds error isn't recurring
[9:43] <toad_> and it refuses connections to 8888...
[9:43] <toad_> ah
[9:44] <toad_> it failed to load BOTH services
[9:44] * toad_ finds out why...
[9:44] <toad_> SEC 17/11/2004 14:43:15.389 (freenet.node.Main, main, ERROR): Failed to load ser
[9:44] <toad_> vice: distribution
[9:44] <toad_> java.lang.NullPointerException
[9:44] <toad_> at freenet.transport.tcpListeningAddress.getValString (tcpListeningAddress.ja
[9:44] <toad_> va:45)
[9:44] <toad_> hmmm
[9:46] <toad_> this is wierd i have no idea why this NPE doesn't happen under sun :(
[9:47] <man-di> hmm
[9:50] <toad_> presumably it gets fed a null at some point...
[9:50] <man-di> back in some minutes
[9:53] * m[a]zzanet (~mazzanet@mazzanet.user) has joined #freenet
[10:04] * jay (jcl@ool-18bf6dac.dyn.optonline.net) has joined #freenet
[10:06] <toad_> hmmm
[10:06] <toad_> looks like a mess up on our end..
[10:06] <toad_> somebody decided it'd be a good idea to bind addresses even when we don't need to
[10:06] * mazzanet (~mazzanet@mazzanet.user) Quit (Client Quit)
[10:11] <toad_> okay
[10:11] <toad_> we do need to bind, but it is perfectly valid for bindAddr to be null
[10:11] <toad_> so it's a bug this end
[10:11] * toad_ fixes
[10:13] <toad_> now, back to the real bugs...
[10:15] <toad_> woah
[10:15] <toad_> it actually responds!
[10:15] <toad_> it doesn't close incoming HTTP connections after sending data, and it throws a few exceptions, but...!
[10:16] <toad_> okay, sometimes it responds ;)
[10:16] <toad_> SEC 17/11/2004 15:15:03.877 (freenet.transport.ReadSelectorLoop, Network reading
[10:16] <toad_> thread, ERROR): Caught throwable in AbstractSelectorLoop!: java.io.IOException:
[10:16] <toad_> gnu.java.nio.SelectorIm
[10:16] <toad_> java.io.IOException: gnu.java.nio.SelectorIm
[10:16] <toad_> at gnu.java.nio.VMSelector.select (VMSelector.java)
[10:16] <toad_> at gnu.java.nio.SelectorImpl.select (SelectorImpl.java:227)
[10:16] <toad_> at freenet.transport.AbstractSelectorLoop.mySelect (AbstractSelectorLoop.java
[10:16] <toad_> :577)
[10:16] <toad_> at freenet.transport.AbstractSelectorLoop.loop (AbstractSelectorLoop.java:767
[10:17] <toad_> )
[10:17] <toad_> at freenet.transport.ReadSelectorLoop.run (ReadSelectorLoop.java:544)
[10:17] <toad_> at java.lang.Thread.run (Thread.java:666)
[10:17] <toad_> at java.lang.VMThread.run (VMThread.java:124)
[10:17] <toad_> and the ArrayIndexOutOfBoundsException
[10:17] <toad_> hrrm
[10:18] <toad_> the above might be a string args glitch or something?
[10:19] <toad_> hmmm actually the second error is a different java.lang.ArrayIndexOutOfBoundsException :(
[10:19] <toad_> SEC 17/11/2004 15:15:21.763 (freenet.transport.ListenSelectorLoop, interface th
[10:19] <toad_> read, ERROR): Caught throwable in AbstractSelectorLoop!: java.lang.ArrayIndexOut
[10:19] <toad_> OfBoundsException
[10:19] <toad_> java.lang.ArrayIndexOutOfBoundsException
[10:19] <toad_> at java.lang.System.arraycopy (System.java)
[10:19] <toad_> at java.lang.String.getChars (String.java:335)
[10:19] <toad_> at java.lang.StringBuffer.append (StringBuffer.java:314)
[10:19] <toad_> at java.lang.Throwable.toString (Throwable.java:260)
[10:19] <toad_> at java.lang.Throwable.stackTraceString (Throwable.java:422)
[10:19] <toad_> at java.lang.Throwable.printStackTrace (Throwable.java:396)
[10:19] <toad_> at freenet.support.FileLoggerHook.log (FileLoggerHook.java:633)
[10:19] <toad_> at freenet.support.LoggerHookChain.log (LoggerHookChain.java:43)
[10:19] <toad_> at freenet.support.LoggerHook.log (LoggerHook.java:76)
[10:19] <toad_> at freenet.transport.AbstractSelectorLoop.mySelect (AbstractSelectorLoop.java
[10:19] <toad_> :586)
[10:19] <toad_> at freenet.transport.AbstractSelectorLoop.loop (AbstractSelectorLoop.java:767
[10:19] <toad_> )
[10:19] <toad_> at freenet.transport.ListenSelectorLoop.run (ListenSelectorLoop.java:149)
[10:19] <toad_> at java.lang.Thread.run (Thread.java:666)
[10:20] <toad_> freenet.support.FileLoggerHook.log (FileLoggerHook.java:633) just says "e.printStackTrace(bpw);"
[10:20] <toad_> where bpw is a printwriter on a bytearrayoutputstream
[10:20] <toad_> and lots and lots of timeouts
[10:21] <toad_> on both read and write...
[10:21] <toad_> man-di: here
[10:21] <toad_> ?
[10:24] <toad_> the web interface partially responds.. it writes the headers and not the content..
[10:24] <toad_> writing the headers is slow though
[10:24] <man-di> toad_: back now
[10:24] <toad_> man-di: hi
[10:24] <toad_> man-di: you see backlog?
[10:24] <toad_> two interesting exceptions
[10:24] <man-di> just looking
[10:25] <toad_> after that, major timeouts
[10:25] <toad_> probably indicates serious problems in the selector loops, but not sure what exactly... could investigate further, but can we do anything with the 2 obvious exceptions?
[10:26] <man-di> looks strange
[10:26] <man-di> I wonder why this happens
[10:26] <toad_> yeah
[10:27] <man-di> I have never seen this with gcj
[10:27] <toad_> hmmm
[10:27] <toad_> you think it might be a kaffe specific problem?
[10:28] <man-di> I need to take a look into kaffe native code to be sure
[10:29] <toad_> it would be quite time consuming to get an up to date gcj, compile it, compile freenet, and test, correct?
[10:29] * man-di checking out kaffe CVS
[10:29] <man-di> toad_: yeah
[10:30] <man-di> I need around 2 hours on my Athlon 1.2 GHz/256 MB ram
[10:30] * toad_ enables heavy logging on NIO, maybe can chase down the timeouts..
[10:30] <toad_> athlon 2800+ XP, 1.5GB...
[10:31] * d-ArkAngel (~Rob@213-131-104-86.onyx.net) Quit (Read error: 104 (Connection reset by peer))
[10:31] * d-ArkAngel (~Rob@213-131-104-86.onyx.net) has joined #freenet
[10:32] <toad_> hi d-ArkAngel
[10:32] <d-ArkAngel> hullo
[10:33] <d-ArkAngel> appologies for vanishing last week, I've been kinda absorbed trying to get things ready for my house move.
[10:34] <man-di> toad_: then I guess it will need around an hour for you, or so
[10:34] <d-ArkAngel> how is the estimator passing code working now? are the graphs looking a little more sane now?
[10:35] <toad_> d-ArkAngel: not sure
[10:37] <toad_> SEC 17/11/2004 15:34:58.560 (freenet.transport.ReadSelectorLoop, Network reading thread, ERROR): Caught throwable in AbstractSelectorLoop!: java.lang.ArrayStoreException: source not an array `freenet/node/states/request/SendFinished'
[10:37] <toad_> hmmm
[10:37] <toad_> different one
[10:38] <toad_> SEC 17/11/2004 15:34:58.560 (freenet.transport.ReadSelectorLoop, Network reading
[10:38] <toad_> thread, ERROR): Caught throwable in AbstractSelectorLoop!: java.lang.ArrayStore
[10:38] <toad_> Exception: source not an array `freenet/node/states/request/SendFinished'
[10:38] <toad_> java.lang.ArrayStoreException: source not an array `freenet/node/states/request/
[10:38] <toad_> SendFinished'
[10:38] <toad_> at java.lang.System.arraycopy (System.java)
[10:38] <toad_> at java.lang.String.getChars (String.java:335)
[10:38] <toad_> at java.lang.StringBuffer.append (StringBuffer.java:314)
[10:38] <toad_> at java.lang.Throwable.toString (Throwable.java:260)
[10:38] <toad_> at java.lang.Throwable.stackTraceString (Throwable.java:422)
[10:38] <toad_> at java.lang.Throwable.printStackTrace (Throwable.java:396)
[10:38] <toad_> at freenet.support.FileLoggerHook.log (FileLoggerHook.java:633)
[10:38] <toad_> that's peculiar
[10:38] <toad_> man-di: what's the next step?
[10:40] <man-di> toad_: I just build kaffe myself
[10:40] <man-di> let me finish this then I have a good thing to test against
[10:41] <toad_> ok
[10:42] * Ash-Fox (HAL-9000@aaz191.neoplus.adsl.tpnet.pl) has joined #FreeNET
[10:42] <toad_> meanwhile i'll try to get a better idea of what's happening with the timeouts
[10:42] <toad_> though with heavy logging, timeouts are inevitable...
[10:44] <toad_> SEC 17/11/2004 15:44:28.322 (freenet.transport.ReadSelectorLoop, Network reading
[10:44] <toad_> thread, ERROR): Caught throwable in AbstractSelectorLoop!: java.io.IOException:
[10:44] <toad_> While deleting, didn't
[10:44] <toad_> java.io.IOException: While deleting, didn't
[10:45] <toad_> at gnu.java.nio.VMSelector.select (VMSelector.java)
[10:45] <toad_> at gnu.java.nio.SelectorImpl.select (SelectorImpl.java:227)
[10:45] <toad_> at freenet.transport.AbstractSelectorLoop.mySelect (AbstractSelectorLoop.java
[10:45] <toad_> :577)
[10:45] <toad_> yet another wierd IOException from the selector...
[10:45] * root (~r00t@ool-18be3a49.dyn.optonline.net) Quit (Read error: 110 (Connection timed out))
[10:45] <toad_> and the "source not an array" error, (with a different object)
[10:47] * root (~r00t@ool-18be3a49.dyn.optonline.net) has joined #freenet
[10:52] <toad_> hmmm
[10:52] <man-di> toad_: looks all very strange
[10:52] <toad_> it wrote the first chunk successfully
[10:53] <toad_> and then it queued the second chunk
[10:53] <toad_> the socket apparently never became writable
[10:53] <toad_> so it eventually timed out
[10:53] <man-di> toad_: if you could write a little testcase that reproduces one of them that would be great
[10:53] <toad_> how can i write a test case when i don't know what the problem is? :|
[10:54] <toad_> the logging thing i probably could do
[10:54] <man-di> toad_: yeah, its not easy
[10:54] <toad_> do you have any standard NIO tests?
[10:55] <toad_> things that involve multiple writes to the same socket?
[10:55] <man-di> toad_: not really
[10:55] <man-di> I tried the testcases from the JavaNIO book
[10:55] * Hory (~Miranda@82.76.81.56) has joined #FreeNet
[10:55] <toad_> dalibor mentioned some test cases which seemed to work
[10:55] <toad_> i think they were really simple
[10:55] <toad_> they worked?
[10:55] <man-di> toad_: yeah, thats the ones from JavaNIO
[10:55] <man-di> yes
[10:56] <man-di> at least on gcj
[10:56] <toad_> got a url?
[10:56] <man-di> I never tried kaffe myslef
[10:56] <man-di> http://javanio.info/
[10:57] * toad_ nods, found it
[10:57] <toad_> have you tried running freenet on GCJ recently?
[10:57] <toad_> what happens?
[10:57] <man-di> I need to write some testcases for our own testsuite
[10:57] <man-di> no, I have tried and current GCC HEAD CVS doenst work for me currently
[10:58] <man-di> havent tried
[10:58] <toad_> http://www.javanio.info/filearea/bookexamples/unpacked/com/ronsoft/books/nio/channels/ChannelAccept.java
[10:58] <toad_> hmmm
[10:58] <toad_> how do i run it? kaffe ChannelAccept?
[10:59] <man-di> toad_: yes, that should do it
[10:59] <toad_> it doesn't work
[10:59] <toad_> i compiled it to channelaccept.class
[10:59] <man-di> I will try it when my kaffe build is complete
[11:00] <toad_> maybe i have to move it to com/ronsoft/books/nio/channels/ ?
[11:01] <man-di> or edit it and remove the package line
[11:02] <d-ArkAngel> there's some NIO examples in the sun guides for 1.4.2, not sure if they'de help any, I've not got a linux box to hand to test them to see if the have issues with gjc tho...
[11:02] <d-ArkAngel> http://java.sun.com/j2se/1.4.2/docs/guide/nio/example/index.html
[11:02] <toad_> okay I connect to it
[11:02] <toad_> it sends one line
[11:02] <toad_> and then disconnects
[11:03] <toad_> of course it sends it in blocking mode :(
[11:03] <toad_> or does it? hmm
[11:04] <man-di> d-ArkAngel: thanks for reminding me on them, last time I wanted to try we missed java/util/regex but noe we have it ...
[11:05] <d-ArkAngel> anyone know why regular expressions excite me so?
[11:05] <d-ArkAngel> it's most worrying
[11:06] <toad_> SelectSockets seems to work...
[11:08] <toad_> not getting anywhere with the examples
[11:10] <toad_> what i do know is i connect to it and do a request, it then sends me the headers (~ 50 bytes), and then sends the payload, but the socket never becomes writable after the payload is registered
[11:11] * toad_ tries something...
[11:11] <man-di> this gives me an idea ....
[11:12] <toad_> if it's a selector problem, lets just ignore the selector and try to write on every channel
[11:13] <toad_> if that works we know it's a selector problem
[11:13] <toad_> if that doesn't we know it probably isn't
[11:14] * Hory (~Miranda@82.76.81.56) Quit (Read error: 54 (Connection reset by peer))
[11:15] <man-di> toad_: good strategy
[11:21] <toad_> that's more or less what that echo server does..
[11:28] * root (~r00t@ool-18be3a49.dyn.optonline.net) Quit (Read error: 110 (Connection timed out))
[11:31] <toad_> hmmm
[11:31] <toad_> it doesn't seem to have worked
[11:31] <toad_> most peculiar..
[11:31] <toad_> brb
[11:31] * toad_ (toad@82-32-18-233.cable.ubr03.azte.blueyonder.co.uk) has left #freenet
[11:33] * root (~r00t@ool-18be3a49.dyn.optonline.net) has joined #freenet
[11:34] * tyesinclair (~tyesincla@170.252.64.1) has joined #freenet
[11:34] * toad (toad@82-32-18-233.cable.ubr03.azte.blueyonder.co.uk) has joined #freenet
[11:34] * toad back
[11:36] * Hory (~Miranda@82.76.81.56) has joined #FreeNet
[11:37] <ShaunMacPherson> wb
[11:39] * ShaunMacPherson (~ShaunMac@Sudbury-HSE-ppp3979244.sympatico.ca) Quit ("Reboot Time!")
[11:40] * toad adds more logging...
[11:43] <toad> logging: the cause of, and solution to, all life's problems...
[11:43] <toad> :)
[11:43] <toad> to paraphrase homer simpson
[11:44] <toad> Caught throwable in AbstractSelectorLoop!: java.io.IOException: NzRr-Pj88cVT0bN~1urLVgA
[11:44] <toad> heh, that's a new one...
[11:54] <toad> looks like the second write job isn't on the selector for some reason
[12:01] <toad> okay it definitely isn't on sel.keys()
[12:01] <toad> that's probably why it never gets written...
[12:01] * tyesinclair (~tyesincla@170.252.64.1) has left #freenet
[12:01] <toad> man-di:
[12:05] * d-ArkAngel (~Rob@213-131-104-86.onyx.net) Quit (Read error: 104 (Connection reset by peer))
[12:06] * greycat checks slashdot.... aww, shit. It's begun. http://www.msnbc.msn.com/id/6504024/
[12:06] <greycat> I guess I won't be buying any movies, either.
[12:08] <toad> greycat: doesn't seem especially draconian to me
[12:08] <toad> I might boycot movies due to them trying to pass stupid laws
[12:08] <toad> but suing filesharers?
[12:08] <toad> of course they're going to do that
[12:09] <toad> they're not suing filesharing networks yet are they?
[12:09] <greycat> I haven't seen what network's users got hit yet.
[12:09] <greycat> But that's not the point. If they're going to sue me, I'm not going to give them any money any more. I've boycotted RIAA music for over 1 year now. I'll boycott MPAA movies as well.
[12:10] <toad> why would you sue them if you are buying stuff from them? ;)
[12:10] <toad> "I refuse to deal with them if they are going to sue me when I copy their stuff"
[12:11] <greycat> because downloading/sharing movies and buying movies are NOT mutually exclusive activities.
[12:11] <greycat> sometimes you download something, and you like it so much that you go buy it on DVD
[12:12] <greycat> this is *especially* relevant for movies in the current generation of internet technology. There are some really crappy encodings out there.
[12:13] <toad> hmmm
[12:13] <toad> it appears that it is possible for register() to return null
[12:13] <toad> despite that not being documented
[12:17] * Ash-Fox (HAL-9000@aaz191.neoplus.adsl.tpnet.pl) Quit (Nick collision from services.)
[12:17] * Ash-Fox (HAL-9000@aaz191.neoplus.adsl.tpnet.pl) has joined #FreeNET
[12:22] * kyll (~jacob@cpe.atm2-0-1061095.0x50a3d342.arcnxx16.customer.tele.dk) Quit ("leaving")
[12:22] <toad> no, not that.. it did NOT return null..
[12:22] <toad> it was successfully registered
[12:23] <toad> but somehow it didn't go into keys
[12:23] <toad> maybe a HashSet bug?
[12:25] <man-di> toad: hmmm
[12:28] * toad hacks kaffe to find out...
[12:29] <toad> more likely it got removed at some point..
[12:29] <toad> due to some bizarre interaction with my code..
[12:29] <toad> Collections.unmodifiableSet just wraps it, correct?
[12:29] <toad> it does NOT copy?
[12:30] <toad> it is not likely to accidentally drop an element?
[12:31] <man-di> it should wrap it afaik
[12:32] * toad restarting again...
[12:38] <toad> brb
[12:43] <toad> restarting
[12:45] <toad> okay
[12:46] <toad> it apparently gets cancelled
[12:46] <toad> so where does it get cancelled?
[12:46] <toad> that is the next question..
[12:46] <toad> fortunately all cancellations go through one method on AbstractSelector
[12:46] <toad> so I'm just putting a trace in there...
[12:47] <toad> logically it must be being cancelled right after the first selection
[12:47] <toad> because it only appears in the selection the very first time
[12:48] * Hory (~Miranda@82.76.81.56) Quit ("CyberLore.net - Recommendations on the best games, freeware and websites.")
[12:50] <toad> restarting...
[12:54] * kevloral is now known as kevlAway
[12:54] * leexgx looks at toad
[12:55] <leexgx> haveing fun then toad
[12:56] * d-ArkAngel (~Rob@cpc2-midd2-5-0-cust16.midd.cable.ntl.com) has joined #freenet
[12:57] <man-di> toad: thats strange
[13:00] <toad> according to this, it was cancelled by my code, after a send attempt
[13:00] <toad> but that should have shown up in the logging
[13:06] * abeot (aeot@a80-186-42-1.elisa-laajakaista.fi) has joined #freenet
[13:06] * abeot (aeot@a80-186-42-1.elisa-laajakaista.fi) has left #freenet
[13:07] <toad> if I have a timestamp that might help...
[13:08] * kevlAway is now known as kevloral
[13:09] <kevloral> g'afternoon all
[13:10] <toad> hi
[13:12] <toad> restarting...
[13:18] <toad> hmm
[13:18] <toad> man-di: i think i may have some idea what's happening here
[13:19] <toad> at 18:14:18.199, the first write job completed
[13:19] <toad> and therefore we cancelled the key
[13:19] <toad> we immediately queued the next write job
[13:19] <toad> and somehow that managed to get cancelled by the first one's cancel queue item..
[13:19] <toad> this is very peculiar as they have their own SKs...
[13:20] <toad> but the cancellation happens in that exact millisecond..
[13:21] <toad> hmmm
[13:21] <toad> it definitely got the same SelectionKey
[13:22] <toad> hmmm
[13:22] <toad> no.. maybe.. hmm
[13:23] <toad> man-di: are SocketChannelSelectionKey's pooled somehow?
[13:28] <toad> okay.. what is happening: we register, we execute the write successfully, we add the key to the list of keys to be cancelled
[13:28] <toad> but it's still on the key set somewhere
[13:28] <toad> then when we register again, SelectionKey reuses the previous key
[13:28] <toad> which is then cancelled!
[13:33] <toad> so it's a one line fix.. hopefully!
[13:33] <toad> (in kaffe)
[13:34] <toad> AbstractSelectableChannel.locate() needs to check that the SelectionKey isValid() before reusing it
[13:35] <toad> this would affect any app which did one write right after the previous one, using NIO
[13:35] <toad> on classpath
[13:38] <man-di> toad: sorry, I was away for dinner
[13:39] <toad> yes!
[13:40] <toad> it's slow.. because of all the logging
[13:41] <toad> but it seems to work
[13:41] <man-di> aaah, I think I get ir
[13:41] <man-di> it
[13:41] <toad> first thing to try: turn off the check-all-conns-hack
[13:42] <man-di> hmmm
[13:42] <man-di> locate is only a little helper to shorten thhe code
[13:43] <toad> locate scans the list of SelectionKeys for one to re-use
[13:43] <toad> it needs to check whether the key is cancelled
[13:43] <toad> because they don't get removed until the deregistering function in Selector is called
[13:44] <man-di> no, its not for reuse
[13:44] <man-di> I added the check in register
[13:44] * toad tries an actual freesite..
[13:44] <toad> hmmm
[13:44] <toad> are we talking about the same function?
[13:45] <toad> private SelectionKey locate(Selector selector)
[13:45] <toad> ListIterator it = keys.listIterator();
[13:45] <man-di> yep
[13:45] <toad> while (it.hasNext())
[13:45] <toad> SelectionKey key = (SelectionKey) it.next();
[13:45] <man-di> toad: note that this method is called from "keyFor" too
[13:45] <toad> if (key.selector() == selector) return key;
[13:45] <man-di> toad: I applied this here locally now:
[13:45] <man-di> diff -u -r1.16 AbstractSelectableChannel.java
[13:45] <man-di> --- java/nio/channels/spi/AbstractSelectableChannel.java 21 Apr 2004 10:46:24 -0000 1.16
[13:45] <man-di> +++ java/nio/channels/spi/AbstractSelectableChannel.java 17 Nov 2004 18:49:25 -0000
[13:45] <man-di> @@ -228,7 +228,7 @@
[13:45] <man-di> if (key != null)
[13:45] <man-di> {
[13:45] <man-di> - if (att != null)
[13:45] <man-di> + if (att != null && key.isValid())
[13:45] <man-di> key.attach(att);
[13:45] <man-di> }
[13:46] <man-di> else
[13:46] <man-di> this is in register() method
[13:47] <toad> same thing, isn't it? except that mine allows for reuse of keys when one is cancelled but another isn't
[13:47] <toad> you don't delete the old key until much later
[13:47] <man-di> in your case when calling keyFor() will not work as expected
[13:48] <toad> ok
[13:48] <man-di> keys may only be deleted under special circumstances
[13:48] * toad will try your patch next
[13:48] <man-di> according to SUN javadocs
[13:48] * toad nods
[13:48] <man-di> and to my JavaNIO book
[13:49] <toad> ummm
[13:49] <toad> your patch means it won't get registered at all if key is not valid
[13:50] <toad> the old key will be deleted when we get around to deleting cancelled keys
[13:50] <toad> and there won't be a new key because you haven't asked for one
[13:50] <toad> because of the if() nesting...
[13:51] <toad> surely you want if(key != null && key.isValid()) ?
[13:51] <man-di> damn, you are right
[13:52] <toad> also the if(att != null) is correct? if there's something valid already registered, we must keep the old attachment?
[13:52] <toad> man-di: so do it on the if(key != null) ??
[13:52] <man-di> I should go back to my old behavior and change source code after writing testcases for the correct behaviour
[13:53] <toad> hmm?
[13:53] <man-di> I will write a mauve testcase for the key handling now
[13:53] <toad> what are you going to do with the source?
[13:53] <sanity> you guys think you have cracked it?
[13:54] <man-di> will be the easiest way to get sure this behaves the same then sun
[13:54] <toad> nod
[13:54] <man-di> toad: I will add your proposed change for now and test what side effects it has
[13:54] <toad> the case in question is when you register another write after the first one has completed, the key has been cancelled, but the cancelled keys haven't been zapped yet
[13:54] <man-di> toad: can you stress test it bit more ?
[13:55] <toad> man-di: which?
[13:56] <man-di> the isValid() check in locate
[13:56] <toad> man-di: I'm just about to restart it with a more normal level of logging, and the change you just agreed to
[13:56] <toad> in locate? i thought that was bad?
[13:56] <man-di> toad: you said mine is bad too
[13:56] <man-di> and you said it made something work for you
[13:56] <toad> I said if(key != null && key.isValid()) might work better than either proposed option
[13:56] <toad> am I wrong about that?
[13:57] <toad> i.e. shift it up one so the else works
[13:57] <man-di> toad: oh, then I misread
[13:57] <toad> key = locate(selector);
[13:57] <toad> if (key != null && key.isValid())
[13:57] <toad> {
[13:57] <toad> if (att != null)
[13:57] <toad> key.attach(att);
[13:57] <toad> }
[13:57] <toad> else
[13:57] <toad> {
[13:57] <toad> key = selector.register(this, ops, att);
[13:57] <toad> that make sense?
[13:58] * man-di shakes head against the wall
[13:58] <toad> sanity: looks like we have fixed it, there's some controversy over exactly how to do the fix; there are some major bugs still in e.g. logging
[13:58] <toad> man-di: no? why not?
[13:58] <man-di> toad: I added the isValid to the wrong line
[13:58] <man-di> toad: you are obviously correct
[13:58] <toad> okay, i'll test it
[13:58] <toad> with freenet
[13:58] <sanity> toad: there is no controversy that cannot be resolved by mudfight
[13:58] <toad> would you like me to run any other tests?
[13:59] <man-di> toad: not yet
[13:59] <toad> well there are some other major bugs but none of them are showstoppers
[13:59] <man-di> I will come back when I have some testcases and file locking implemented
[13:59] <toad> okay ttyl
[13:59] <man-di> thanks for your help
[13:59] <toad> sanity: oh, and you have to turn file locking off
[14:00] <toad> which can't as yet be done by a config option :|
[14:00] <sanity> due to another bug?
[14:00] <toad> man-di: sure, it didn't take TOO long, and it'd be really great if we could package freenet properly instead of having to rely on the sun stuff - for practical as much as ideological reasons
[14:01] <toad> sanity: file locking isn't implemented by kaffe
[14:01] <man-di> toad: yeah, many users will appreciate it, I think
[14:01] <toad> in order to let eclipse load, it throws an IOE instead of an UnsupportedOpException
[14:01] <sanity> i see
[14:01] <toad> thus we terminate :(
[14:02] <man-di> toad: when filelokcing is fixed it should be working with freenet and eclipse3
[14:02] <toad> that would be nice
[14:03] <toad> there's some wierd error messages that suggest some serious problems in stack dumping
[14:03] <toad> also
[14:03] <toad> those need looking at at some point
[14:05] * sanity (~ian@81-178-106-165.dsl.pipex.com) Quit ()
[14:06] * sanity (~ian@81-178-106-165.dsl.pipex.com) has joined #freenet
[14:06] <man-di> toad: yep, but these need dalibor/robilad
[14:06] * toad_ (toad@82-32-18-233.cable.ubr03.azte.blueyonder.co.uk) has joined #freenet
[14:06] * toad (toad@82-32-18-233.cable.ubr03.azte.blueyonder.co.uk) Quit ("[BX] Who ate my nuggets?!")
[14:06] <toad_> hi
[14:06] * toad_ still here :)
[14:07] <toad_> yikes
[14:07] <toad_> now it produces TONS of error messages
[14:07] <toad_> I wonder what I did?
[14:07] <toad_> SHIT
[14:07] <toad_> NPEs without stack traces... I thought only Sun did that :<
[14:08] <toad_> lots of cancellation related messages too
[14:08] <man-di> toad_: still using kaffe ?
[14:08] <toad_> yup
[14:10] <toad_> those NPEs occur pretty much continuously
[14:10] <toad_> hmmm
[14:10] <toad_> it's our fault that we don't show the stack trace
[14:11] <toad_> java.lang.NullPointerException
[14:11] <toad_> at gnu.java.nio.VMSelector.select (VMSelector.java)
[14:11] <toad_> at gnu.java.nio.SelectorImpl.select (SelectorImpl.java:227)
[14:11] <toad_> at freenet.transport.AbstractSelectorLoop.mySelect (AbstractSelectorLoop.java
[14:11] <toad_> :577)
[14:11] <toad_> at freenet.transport.AbstractSelectorLoop.loop (AbstractSelectorLoop.java:767
[14:11] <toad_> )
[14:12] <toad_> at freenet.transport.WriteSelectorLoop.run (WriteSelectorLoop.java:797)
[14:12] <toad_> at java.lang.Thread.run (Thread.java:666)
[14:12] <toad_> at java.lang.VMThread.run (VMThread.java:124)
[14:12] <toad_> ouch
[14:12] <toad_> alternates between being caused by select and selectNow
[14:13] <toad_> are you sure that putting it in locate() is wrong?
[14:17] * toad_ rebuilds with the change in locate instead
[14:18] * nanane (~nanane@d117252.adsl.hansenet.de) has joined #freenet
[14:18] <man-di> toad_: not 100%
[14:20] * toad_ starting node..
[14:22] <toad_> that seems to work..
[14:22] <toad_> also the string errors seem to be grabbing random data from anywhere in the VM
[14:22] <toad_> SEC 17/11/2004 19:23:26.529 (freenet.transport.ReadSelectorLoop, Network reading thread, NORMAL): null key for gnu.java.nio.SocketChannelImpl@c3fbe8c
[14:23] <toad_> hmmm new error
[14:24] <toad_> performance looks reasonably promising
[14:25] <toad_> e.g. some pagrs load
[14:25] <man-di> you mean it somehow works with kaffe ?
[14:25] <toad_> the ones that don't return DNF, which is what they should do
[14:25] <toad_> man-di: freenet seems to work with kaffe with the change to locate rather than to register
[14:25] <toad_> bbl, going to get food
[14:26] <man-di> toad_: thanks for testing
[14:29] * nanane (~nanane@d117252.adsl.hansenet.de) Quit ("Verlassend")
[14:34] * _Ribs_ (~freenode@riblet.plus.com) has joined #freenet
[14:34] * Ribs (~freenode@riblet.plus.com) Quit (Read error: 54 (Connection reset by peer))
[14:34] * _Ribs_ is now known as Ribs
[15:02] * kevloral (~kevloral@CZ1-RAS-8-u-0179.du.onolab.com) Quit ("Peace and Protection 4.22")
[15:06] * m[a]zzanet (~mazzanet@mazzanet.user) Quit (Client Quit)
[15:07] * mazzanet (~mazzanet@mazzanet.user) has joined #freenet
[15:09] * sdogi_ (~java@84-50-16-198-dsl.kjj.estpak.ee) has joined #freenet
[15:09] * guido^pe (~unknown@dsl-213-023-244-187.arcor-ip.net) has joined #freenet
[15:15] * sdogi (~java@84-50-20-115-dsl.kjj.estpak.ee) Quit (Read error: 110 (Connection timed out))
[15:20] * Zorix (Brandon@fl-67-76-222-71.sta.sprint-hsd.net) Quit (Read error: 110 (Connection timed out))
[15:52] <toad_> man-di: it seems to not work very efficiently; 100% cpu usage constantly
[15:52] <toad_> this results in it being very slow, timeouts, and so on
[15:53] <man-di> toad_: I guess I know why, our select code has a bug
[15:53] <toad_> hmm?
[15:53] <toad_> what sort of bug?
[15:54] <man-di> I think it runs in an endless loop instead of letting the OS wait
[15:54] <toad_> oops :)
[15:56] <toad_> http://amphibian.dyndns.org/simulations2-epass-0.025-100x25-new.png
[15:58] <toad_> http://amphibian.dyndns.org/simulations2-epass-0.025-400x50-new.png
[15:58] <toad_> hrrm
[15:58] <toad_> the latter is rather less promising :(
[16:02] <toad_> load average: 4.28, 5.03, 5.59
[16:03] <man-di> ehe
[16:03] <toad_> cpu usage at 100% except for 1 second in the last ~ 11 minutes
[16:04] <man-di> hehe
[16:04] <man-di> I will have to look into some bad C code
[16:04] <toad_> ok
[16:04] <toad_> i'm going to restart and go back to sun
[16:04] <toad_> contact me when you have some interesting changes
[16:04] <man-di> okay
[16:04] <toad_> root@amphibian:~# ssh toad@sbc.lir.dk
[16:04] <toad_> ssh: connect to host sbc.lir.dk port 22: Connection timed out
[16:04] <toad_> argh
[16:05] <man-di> toad_: I think I will need some days
[16:05] <toad_> yet another simulation host that ceases to be available :<
[16:05] <spaetz> toad, I'm back online
[16:05] <spaetz> feel free to use it again :)
[16:06] <spaetz> I was in Germany last week, at my granddads funeral, so I shut down everything
[16:06] <toad_> thanks
[16:08] <toad_> but judging by that graph I don't need processing power, I need to figure out what the fsck is wrong with the new simulations code
[16:08] <toad_> and i have no idea where to begin, as usual
[16:10] * spaetz (~spaetz@80-218-147-20.dclient.hispeed.ch) Quit (Read error: 113 (No route to host))
[16:11] * spaetz (~spaetz@80-218-147-20.dclient.hispeed.ch) has joined #freenet
[16:11] <toad_> wb spaetz
[16:11] <spaetz> hi
[16:11] <spaetz> *sigh* crappy laptop battery was emtpy :(
[16:12] <toad_> http://amphibian.dyndns.org/simulations2-epass-0.025-400x50-new.png
[16:12] <toad_> is that (either blue) going to recover?
[16:12] <toad_> and if it does will it recover anywhere near the range we want?
[16:13] <toad_> and will it do it in a reasonable number of reqs (the horiz scale is inaccurate; needs to be multiplied by 4) ?
[16:17] <toad_> anyone got any insights into how you debug simulations?
[16:23] * toad_ (toad@82-32-18-233.cable.ubr03.azte.blueyonder.co.uk) Quit ("using sirc version 2.211+KSIRC/1.3.11")
[16:27] * toad_ (toad@82-32-18-233.cable.ubr03.azte.blueyonder.co.uk) has joined #freenet
[16:27] <toad_> gr
[16:29] <toad_> grr
[16:30] <toad_> grrr
[16:31] [toad_ PING]
[16:34] <toad_> grrrr
[16:35] * pittaman (~mistery_b@d51A410CA.kabel.telenet.be) Quit (Client Quit)
[16:36] * pittaman (~mistery_b@d51A410CA.kabel.telenet.be) has joined #freenet
[16:37] <toad_> IS THERE ANYBODY OUT THERE?
[16:38] <man-di> toad_: we are here
[16:39] <toad_> man-di: any ideas how to debug simulators?
[16:40] <man-di> I ahve no idea what your simulators are
[16:40] <toad_> :|
[16:44] <toad_> i suppose i could simulate the backpassbug on 400x50
[16:44] * Zorix (Brandon@fl-67-76-222-71.sta.sprint-hsd.net) has joined #freenet
[16:44] <toad_> that might help
[16:44] <d-ArkAngel> does that show the kind of line we want? (in the smaller 100x sims you tested that on the other day?
[16:45] <toad_> huh?
[16:45] <toad_> http://amphibian.dyndns.org/simulations2-epass-0.025-100x25-new.png
[16:45] <toad_> very dark blue
[16:45] <toad_> not especially interesting
[16:47] <d-ArkAngel> so on the smaller network the updated code is in fact performing better?
[16:48] <toad_> hmm?
[16:48] <d-ArkAngel> sorry, I'm having trouble seeing what all the lines are since the words scroll of the side of the image :-)
[16:48] <toad_> it is performing significantly worse with settings that should emulate the old code
[16:48] <toad_> and a bit better with the new don't-drop-anything-till-we-have-data-on-everyone code
[16:49] * greycat (~wooledg@192.35.79.70) Quit ("This time the bullet cold rocked ya / A yellow ribbon instead of a swastika")
[16:49] <toad_> oddly, the occasionally-drop-a-node-when-we-don't-need-to bug (purple) seems to be quite beneficial
[16:53] * toad_ runs simulation of backpassbug on 400x50 on hal
[16:55] * toad_ (toad@82-32-18-233.cable.ubr03.azte.blueyonder.co.uk) Quit ("using sirc version 2.211+KSIRC/1.3.11")
[17:18] * pittaman (~mistery_b@d51A410CA.kabel.telenet.be) Quit (Client Quit)
[17:22] * toad_ (toad@82-32-18-233.cable.ubr03.azte.blueyonder.co.uk) has joined #freenet
[17:27] * cv` (~user@ts2-a168.Nsk.dial.rol.ru) has joined #freenet
[17:27] * cv` (~user@ts2-a168.Nsk.dial.rol.ru) Quit (Client Quit)
[17:27] <spaetz> toad_: how dare you simulating bugs on my machine :-)?
[17:27] <spaetz> i want the improved code man
[17:28] <spaetz> BTW: ignore me, I'm tired
[17:28] <spaetz> I did run simulations in my manufacturing course, I have never debugged any...
[17:41] <toad_> whole bunch of lines from that graph have disappeared...
[17:41] <toad_> grrrrrrr
[17:42] <spaetz> how can they disappear?
[17:42] <spaetz> lost the data file?
[17:42] <toad_> don't know
[17:42] <toad_> possibly they were deleted on the other end; not sure what rsync does with single files
[17:42] <toad_> i doubt it deletes them, but perhaps with -a...
[17:48] * FallingBuzzard (~srademach@207.152.112.129) Quit ("Download Gaim: http://gaim.sourceforge.net/")
[17:49] * becllam (~zrzerz@pauguste-3-81-57-114-186.fbx.proxad.net) has joined #freenet
[17:49] <becllam> hi
[17:49] <becllam> just download the 5099 but behind a firewall
[17:50] <becllam> is there any stuff to make to run freenet correctly?
[17:50] <becllam> behind this?
[17:50] <toad_> becllam: how much control do you have over the firewall?
[17:50] <becllam> all this is mine
[17:51] <toad_> then just forward the port
[17:51] <becllam> In fact I managed to go to freenet engine
[17:51] <becllam> no pb
[17:51] <becllam> but now would be sure to be considered as a permanant node
[17:51] <toad_> no such thing as transient any more
[17:51] <toad_> just forward the port and let the node detect its IP
[17:52] <becllam> how do i know that the ip is detected?
[17:53] <toad_> it should just work as far as ip detection goes, if you have any connections
[17:54] <becllam> (Inbound/Outbound/Limit)
[17:54] <becllam> 5 (0/5/200)
[17:54] <becllam> inbound is 0 :(
[17:54] <becllam> that mean that nobody get any info from my node?
[17:55] <toad_> you didn't forward the port?
[17:55] <toad_> what build number?
[17:56] <becllam> last one 5099
[17:57] <becllam> (stable)
[17:57] <becllam> what port should i forward?
[17:57] <toad_> well, did you forward the port?
[17:57] <toad_> the listenPort
[17:59] <becllam> ok will do that thanks !
[17:59] <toad_> becllam: listenPort is in the config file and in the config dialogs if you have them
[18:07] * guido^pe (~unknown@dsl-213-023-244-187.arcor-ip.net) Quit (Remote closed the connection)
[18:07] * jay (jcl@ool-18bf6dac.dyn.optonline.net) Quit (".")
[18:08] <toad_> GGGGGGGGGGGGGGGGGGGGRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRR
[18:09] <spaetz> lol, need some punching-bag?
[18:09] <leexgx> need an new keybord ?
[18:09] <leexgx> mouse
[18:10] <spaetz> I just read about that study which examines violence against computers
[18:11] <toad_> it's not the computer's fault that the simulations don't work, and there's little chance of them ever working before we run out of money
[18:11] <spaetz> amazingly many have said to have hit their keyboard or their monitor
[18:11] <toad_> it's first and foremost my fault
[18:11] <toad_> followed by ian's
[18:11] <spaetz> why that?
[18:12] <toad_> i suppose if i just rewrite it from scratch every time something goes wrong, sooner or later i'll make something that works
[18:13] <spaetz> rewriting the simulations? Or Freenet :)?
[18:13] <toad_> the simulations
[18:13] <toad_> :)
[18:14] * Hory (~Miranda@82.76.81.56) has joined #FreeNet
[18:16] * d-ArkAngel glances side long at the threaded sim code...
[18:17] * gi (~zrzerz@pauguste-3-81-57-114-186.fbx.proxad.net) has joined #freenet
[18:17] * man-di (~man-di@dynadsl-080-228-94-035.ewetel.net) Quit ("Client exiting")
[18:18] <leexgx> i had an cool web site that you can smash an pc by clicking in dif parts of the pc
[18:18] <leexgx> an pic of an computer click on dif parts and soon it starts to fall to bits
[18:18] <spaetz> heh
[18:18] <d-ArkAngel> kinda reminds me of my last keyboard.
[18:20] * tyesinclair (~tyesincla@195.157.199.85) has joined #freenet
[18:20] * tyesinclair (~tyesincla@195.157.199.85) has left #freenet
[18:21] * toad_ proposes to sulk until midnight and then go to bed
[18:21] <d-ArkAngel> :-)
[18:21] <KenMan> so long as it is productive sulking...
[18:22] * toad_ is not convinced there is any such thing
[18:22] <KenMan> try getting real angry at the sim code. Then rewrite it later, feeding off that angry energy
[18:23] <toad_> then debug it
[18:23] <toad_> then debug it some more
[18:23] <toad_> then find that the obvious bugs appear to have gone away but the psuccess is still absurdly low
[18:23] * becllam (~zrzerz@pauguste-3-81-57-114-186.fbx.proxad.net) Quit (Read error: 110 (Connection timed out))
[18:23] <KenMan> just keep it as simple as possible
[18:23] <toad_> i tried to
[18:24] <toad_> but the only way to make a working simulation is lucj
[18:24] <toad_> luck
[18:24] <toad_> and for some strange reason, luck isn't on my side at the moment
[18:24] <d-ArkAngel> :-)
[18:24] * spaetz wishes toad all the luck he needs
[18:25] <toad_> i don't believe in luck :)
[18:25] <KenMan> but you accept it should it come your way, right
[18:25] <spaetz> darn, that's a bad thing to rely on something one doesn't believe in :)
[18:25] <toad_> heh
[18:25] <d-ArkAngel> that's why it's not on your side :-)
[18:26] <KenMan> i have a more consumer-based approach to luck - you need to BUY it
[18:26] <toad_> LOL
[18:26] <toad_> human sacrifice anyone? ;)
[18:26] <spaetz> toad_: any more non-coding support that I can give you?
[18:27] <spaetz> I could sacrifice my neighbour's cat, I wonder if that suffices
[18:27] <spaetz> or if freenet needs something stronger
[18:28] * spaetz is really tired
[18:28] <toad_> heh
[18:28] <spaetz> I better go to bed right now. See you guys around
[18:28] <d-ArkAngel> I could probably manage the guy next door... and possibly the whole family on the other side...
[18:29] * spaetz is now known as spaetzzZ
[18:29] * toad_ believes in luck actually, after a fashion... "providence". unfortunately Boss seems to want me to figure out for myself what to do about the simulations crashing and burning.
[18:29] <gi> sorry to disturb you but do you know how to put a bigger datastore instead of the 256 mb one?
[18:30] <toad_> still, if he wanted me to abandon the project, he wouldn't sabotage the sims so they contradicted each other, he'd make them show that it can't scale :)
[18:30] <toad_> gi: storeSize
[18:30] <toad_> in the config file
[18:30] <toad_> takes multipliers
[18:30] <toad_> for example
[18:30] <toad_> storeSize=20.5G
[18:30] <toad_> remove the % at the beginning of the line if there is one
[18:30] <gi> great :) thank you
[18:30] <toad_> bbl
[18:31] <gi> can i change it without restarting freenet?
[18:31] <toad_> not sure
[18:31] <toad_> increasing it should not result in significant data loss from the store
[18:33] <gi> ok
[18:34] * Redb3ard (~oylerj@c-24-125-12-101.va.client2.attbi.com) has joined #freenet
[18:50] <toad_> hi Redb3ard
[18:50] <KenMan> here are some pictures to ponder... http://mywebpages.comcast.net/jkcorson/FreeNet/500x25x20x100-400.png
[18:51] <KenMan> http://mywebpages.comcast.net/jkcorson/FreeNet/nl.png
[18:51] <KenMan> http://mywebpages.comcast.net/jkcorson/FreeNet/keydist.png
[18:51] <toad_> KenMan: what's the top supposed to show?
[18:51] <KenMan> the notches along the bottom of the nodeload (nl.png) demark 8M CHK requests
[18:52] <KenMan> top of what ?
[18:52] <toad_> the first one
[18:52] <KenMan> just the typical stuff i'm measuring
[18:52] <KenMan> oh, you mean purple ? that is the number of queries (htl*unique CHK requests)
[18:53] <KenMan> keydist shows that, for whatever reason, more keys are having fewer copies as time progresses
[18:54] <KenMan> it is very common for a key to hang around somewhere with only one copy
[19:08] * toad_ investigating a radical hypothesis..
[19:10] <toad_> hmmm
[19:10] <toad_> probably not that
[19:11] <toad_> the theory was that the good lines have native bigints
[19:11] <toad_> and the bad lines don't
[19:11] <spaetzzZ> *yawn* that would indeed be radical
[19:11] * spaetzzZ goes back to sleep
[19:12] <toad_> but it looks like that's not it :(
[19:12] <toad_> middle blue, for example, has native big ints, on the 100x25
[19:13] <KenMan> good thing i'm not really using my internet connection right now - it's pretty much shot to hell
[19:13] <KenMan> oh, rather, it DIED! Cable modem losing signal, fading in and out...
[19:16] <KenMan> how could that possibly impact success ?
[19:17] <KenMan> faulty native vs java implementations ?
[19:18] * KenMan suspects toad is past his bedtime ;)
[19:35] * Hory (~Miranda@82.76.81.56) Quit ("CyberLore.net - Recommendations on the best games, freeware and websites.")
[19:40] * spaetzzZ (~spaetz@80-218-147-20.dclient.hispeed.ch) Quit ("Lost terminal")
[19:51] * d-ArkAngel (~Rob@cpc2-midd2-5-0-cust16.midd.cable.ntl.com) Quit (Read error: 104 (Connection reset by peer))
[20:19] <toad_> Key Not Found in Manifest
[20:19] <toad_> Unexpected Key
[20:19] <toad_> Couldn't retrieve key: SSK@Sc6qV%7eD6iFhaYord6HtbjJ8MaEYPAgM/YoYo//PrivSec.jpg
[20:19] <toad_> The key you are trying to fetch does not exist in the freesite it is supposed to be contained in, which was found. This is a permanent error. You can click here to go to the parent site or return to gateway page.
[20:26] <leexgx> bug then?
[20:27] <toad_> bug in the new yoyo
[20:29] <root> yo toad
[20:29] <root> can you send a test case to the kaffee peeps
[20:29] <root> they say they'll fix their thing then
[20:30] <root> as they'll know what to do
[20:30] <toad_> hmm?
[20:30] <toad_> is that from man-di?
[20:33] <root> yes
[20:34] <toad_> there are actually very few NIO code examples for writing :(
[20:34] <toad_> e.g. those book examples all do writing in silly ways
[20:35] <root> ive found that book examples always suck
[20:35] <root> i have to go look at other peeps code to understand anything "new"
[20:35] <root> oh well
[20:36] <toad_> the time server example only does one write, so can't show the problem i have
[20:36] <root> btw is that the only NIO prob in kaffee
[20:36] <root> <toad_> actual behaviour: it re-uses the cancelled SelectionKey, which is then cancelled, and your app never gets any sort of notification
[20:37] <root> i don't think that would take too much to fix
[20:37] <toad_> I don't know if that's the ONLY NIO problem in Kaffe; actually I'm pretty sure it's not, man-di thought the selectors were busy-looping
[20:37] <toad_> performance was pretty bad
[20:37] <root> oo
[20:37] <toad_> but with that fixed, it does more or less work
[20:37] <toad_> slowly, with lots of errors, but it responds
[20:37] <root> man-di wants more colaberation between you guys and kaffee devs to fix the thing
[20:37] <KenMan> really ? it keeps running ? that is a milestone, to be sure
[20:37] <root> could you oblige him for a bit?
[20:38] <toad_> sure
[20:38] <toad_> tomorrow :)
[20:38] * toad_ thinks it's probably time for him to go to bed now
[20:38] <root> id think a week would sort out the big issues
[20:38] <toad_> hmmm
[20:38] <root> and then they could optimise once that's done
[20:38] <root> :)
[20:38] <root> thanks
[20:38] <toad_> well if it's more than a couple days, ...
[20:38] <toad_> we'll see
[20:38] <KenMan> there is a certain payoff, to be sure!
[20:38] <toad_> I certainly can help you tomorrow
[20:39] <toad_> beyond that we'll have to see
[20:39] <root> bbl vid game
[20:39] <root> msg man-di
[20:39] <root> he's the code guy :P
[20:39] <toad_> I think getting it to work on kaffe is valuable, maybe it's worth a sustained effort
[20:39] <root> im just a messinger
[20:39] <toad_> thanks
[20:39] <root> I agree
[20:39] <root> as the ibm and sun bins don't work on some arches
[20:39] <toad_> you have man-di's address?
[20:40] <root> (amd M procs have probs)
[20:40] <toad_> root: very few arches; the real payoff is packaging IMHO
[20:40] <toad_> not to mention ideological issues that can put off the very people most likely to put up with freenet's currently dire performance ;)
[20:40] <root> unfortunatly I dont have his addy,
[20:40] <toad_> root: really? that's strange
[20:40] <root> he's in #kaffe often tho
[20:40] <toad_> no problems on my XP..
[20:40] <root> yea my AMD M machine will not run the bins
[20:41] <root> (its a deb box)
[20:41] <KenMan> M as in mobile ? or M in MP ?
[20:41] <root> and the prob isnt with the libs or anything
[20:41] <root> M as in MP
[20:41] <toad_> hmmm
[20:41] <toad_> multi opterons run it fine though
[20:41] <root> also im idelogcal too :) so i like kaffe :P
[20:41] <toad_> most peculiar
[20:41] <root> ah opts are too expensive for me :P
[20:41] <toad_> :)
[20:42] <root> im waiting till XP-64s get into 8ghz
[20:42] <toad_> i've heard bad things about athlon MPs
[20:42] <toad_> heh
[20:42] <root> then ill put up a 4 or 8 way box
[20:42] <toad_> i'd settle for 2x4000+'s running at say 3GHz on a single chip, then 2 chips on a mobo :)
[20:42] <root> free shells ofcoures
[20:42] <KenMan> it looks like AMD wants to sell more Athlon64 chips than opteron anyway
[20:42] <toad_> what's 4000+? more like 2.5GHz iirc..
[20:42] * root likes giving out free shels
[20:42] <root> lol
[20:42] <root> <toad_> what's 4000+? more like 2.5GHz iirc..
[20:43] <root> they better ramp up the things
[20:43] <toad_> more like 2.5 than 3 i mean
[20:43] <toad_> root: clock isn't everything. amd is kicking ass despite lower clocks.
[20:43] <root> they are slow to push up the clockspeeds
[20:43] <toad_> when the software adapts, they'll kick even more ass
[20:43] <KenMan> with a 9 metal layer chip, how far can they go ??
[20:43] <root> yes I know... but i need braging rights if ima shell out 10k
[20:43] <toad_> e.g. there is no retail 64-bit windows yet
[20:43] <toad_> heh
[20:44] <KenMan> i thought there was an athlon64 version of XP
[20:44] <root> toad i think you allready have a shell on my system
[20:44] <toad_> i do? which box? can i use it for simulations? :)
[20:44] <root> remeber a year ago, cat2, trying to get the jvms to work
[20:44] <root> yes you may :)
[20:44] <root> cat2.ath.cx
[20:44] <toad_> hmmm
[20:44] <toad_> nope
[20:44] <toad_> no password recorded
[20:45] <toad_> not in passwords file, not in freenet shells file
[20:45] * toad_ identifies...
[20:45] <root> might have been one of the other devs
[20:45] <root> but this can be remidied
[20:46] <root> username and temp pass?
[20:47] <KenMan> ah, windows XP 64bit is still in pre-release / beta
[20:47] <toad_> KenMan: and we only got a 64-bit jvm with 1.5
[20:47] <toad_> AAAAAGES after the opterons came out
[20:47] <toad_> something like 9 months after even the a64s
[20:48] <root> i need to go shoot some fake people in a fake world
[20:48] <root> bbl fps
[20:52] <KenMan> Windows XP 64-Bit Edition does not provide support for a number of older subsystems, including: POSIX
[20:53] <toad_> eh?
[20:53] <root> like they couldnt have inclueded the 32bit bins and have the compat layer on the CPUs do the job
[20:54] <KenMan> they run 32bit apps in emulation mode, which results in lower performance than on a 32bit platform
[20:54] <root> maby the passing of 64bit ints to a 32bit program would cause problems?
[20:54] <root> (it has caused problem in european rockets)
[20:54] <toad_> not ints
[20:54] <toad_> pointers
[20:54] <toad_> that's the big issue
[20:54] <toad_> but there are ways around it
[20:55] <toad_> they cost performance, of course
[20:55] <toad_> not necessarily a huge amount of it though
[20:55] <root> (the rocket nozzles went haywire, tech destructed it in mid air, payload lost)
[20:55] <root> (10 mil down the drain)
[20:56] <root> I'm guessint M$ just dosn't really care about being *nix compliant
[20:56] <toad_> root: american rockets
[20:56] <toad_> unit conversions :(
[20:57] <toad_> and we're talking high profile interplanetary space probes here too :<
[20:57] <root> toad: you sure? It was launched in south america
[20:57] <toad_> not so recent
[20:57] <root> oo we talking about diff ones i think
[20:57] <toad_> the 1999 mars craft, remember? all 3 were lost, at least 2 were because of software issues
[20:57] <root> Im talking about the launch of (something) 5 rocket (not saturn 5!)
[20:57] <toad_> oh, the ariane 5 maiden voyage?
[20:58] <toad_> that was horrible, yeah.. and i think it was more than $10M..
[20:58] <toad_> the satellite it carried was really important
[20:58] <root> yes
[20:58] <toad_> fortunately they built another one
[20:58] <toad_> now i go bed
[20:58] <toad_> bbl zzz
[20:58] <root> it was measuring gravity fields right?
[20:59] <toad_> i thought it was a solar thingy
[20:59] <toad_> dunno
[20:59] <root> seeya
[20:59] <toad_> was it a COBE type thing?
[20:59] <toad_> dunno.. seeya
[20:59] <toad_> why guess when you can google?
[20:59] <root> heh
[21:08] <leexgx> !google COBE
[21:08] <leexgx> i need to fix my self
[21:08] <leexgx> or turn it on
[21:09] <root> !google ariane 5
[21:09] <root> ...
[21:09] * root bites leexgx's head off!
[21:21] <leexgx> LOL
[21:22] <leexgx> this mod does not have an script for it old one did
[21:39] <toad_> http://www.around.com/ariane.html
[21:39] <toad_> re an earlier conversation with root
[21:39] <toad_> |Fortunately, he points out, really important software has a reliability of 99.9999999 percent. At least, until it doesn't.
[21:40] <toad_> nine nines? in software? /me is deeply skeptical!
[21:40] <toad_> bbl
[21:41] <toad_> really important software doesn't include the ariane's guidance system :)
[21:45] <KenMan> i wrote a program once that worked with at least 99.999% correctness. It added 2 and 2, and printed the result. Usually, when it failed, a thorough investigation would find that the software was not at fault.
[21:46] <KenMan> as software gets more complex, the number of conditions that can trigger a fault tend to increase. But it you don't subject it to widely variable input, the chances of triggering a fault decrease.
[22:37] * Zorix (Brandon@fl-67-76-222-71.sta.sprint-hsd.net) Quit (Read error: 110 (Connection timed out))
[23:38] * plixed (~plixed@i3ED6230C.versanet.de) Quit (Connection timed out)
[23:54] * Zorix (Brandon@fl-67-76-222-71.sta.sprint-hsd.net) has joined #freenet
These logs were automatically created by Jay Oliveri with his gimp hapi on irc.freenode.net.
