IRC log for #bzflag on 20100411

00:05.58*** join/#bzflag chaoscon (~jeremy@smartserv/ceo/chaoscon)
00:14.59*** join/#bzflag chaoscon (~jeremy@smartserv/ceo/chaoscon)
00:15.34*** join/#bzflag McSpider (~McSpider@xplr-ts-v10-208-114-184-180.barrettxplore.com)
00:36.21*** join/#bzflag xaver__ (~xaver@pD9ED5E26.dip.t-dialin.net)
00:39.07*** join/#bzflag McSpider (~McSpider@xplr-ts-v10-208-114-184-180.barrettxplore.com)
00:42.49*** join/#bzflag spldart (~spldart2@bzflag/contributor/spldart)
00:42.49*** mode/#bzflag [+v spldart] by ChanServ
01:05.32*** join/#bzflag chaoscon (~jeremy@smartserv/ceo/chaoscon)
01:21.13*** join/#bzflag Pimpinella (~frank@gondolin.pimpi.org)
01:23.08*** join/#bzflag allejo (~allejo@pool-173-51-205-154.lsanca.fios.verizon.net)
01:48.27*** join/#bzflag JeffM1 (~Jeff@adsl-75-50-184-13.dsl.lsan03.sbcglobal.net)
01:51.58*** join/#bzflag allejo (~allejo@pool-173-51-205-154.lsanca.fios.verizon.net)
02:16.20*** join/#bzflag McSpider (~McSpider@xplr-ts-v10-208-114-184-180.barrettxplore.com)
02:21.37*** join/#bzflag Albino_ (~62d91f2d@gateway/web/freenode/x-locppsgccqavcfpx)
02:42.55*** join/#bzflag TD-Linux (~wheeeeeee@about/essy/indecisive/TD-Linux)
02:50.16*** join/#bzflag chaoscon (~jeremy@smartserv/ceo/chaoscon)
02:56.49*** join/#bzflag chaoscon (~jeremy@smartserv/ceo/chaoscon)
03:02.33*** join/#bzflag chaoscon (~jeremy@smartserv/ceo/chaoscon)
03:14.56*** join/#bzflag chaoscon (~jeremy@smartserv/ceo/chaoscon)
03:22.48*** join/#bzflag chaoscon (~jeremy@smartserv/ceo/chaoscon)
03:40.12*** join/#bzflag mrapple (~mrapple@bzflag/player/MrAppleComputer)
03:52.18*** join/#bzflag Think_Differentl (~js@bzflag/player/ThinkDifferent)
04:31.23*** join/#bzflag chaoscon (~jeremy@smartserv/ceo/chaoscon)
04:49.22*** join/#bzflag FyRe_STaRTeR (fyre@adsl-64-217-16-108.dsl.okcyok.swbell.net)
04:56.32*** part/#bzflag FyRe_STaRTeR (fyre@adsl-64-217-16-108.dsl.okcyok.swbell.net)
04:58.46*** join/#bzflag Legoguy (~Legoguy@S0106001ee54ac005.ed.shawcable.net)
05:12.42*** join/#bzflag think_tank (~45aba331@gateway/web/freenode/x-fvmfyilcjjjbgbfy)
05:13.53think_tankis it worth asking what the problem is or has that horse already been beaten to death
05:14.29blast007random hosts from China keep hitting the brlcad.org site
06:26.05*** join/#bzflag Marzipan- (~Marzipan@p5B227898.dip.t-dialin.net)
06:27.31*** join/#bzflag Marzipan (~Marzipan@bzflag/player/Marzipan)
07:13.49*** join/#bzflag Upsetter (~Ups@89.246.172.128)
07:36.09*** join/#bzflag Swigg_ (~Default@pool-74-97-222-150.grnvsc.dsl-w.verizon.net)
07:42.51*** join/#bzflag chaoscon (~jeremy@smartserv/ceo/chaoscon)
08:00.45*** join/#bzflag chaoscon (~jeremy@smartserv/ceo/chaoscon)
08:09.11*** join/#bzflag chaoscon (~jeremy@smartserv/ceo/chaoscon)
08:17.58*** join/#bzflag chaoscon (~jeremy@smartserv/ceo/chaoscon)
08:31.27*** join/#bzflag chaoscon (~jeremy@smartserv/ceo/chaoscon)
08:44.13*** join/#bzflag chaoscon (~jeremy@smartserv/ceo/chaoscon)
08:58.02*** join/#bzflag chaoscon (~jeremy@smartserv/ceo/chaoscon)
09:05.32*** join/#bzflag chaoscon (~jeremy@smartserv/ceo/chaoscon)
09:11.12*** join/#bzflag chaoscon (~jeremy@smartserv/ceo/chaoscon)
09:36.48*** join/#bzflag chaoscon (~jeremy@smartserv/ceo/chaoscon)
09:46.17*** join/#bzflag chaoscon (~jeremy@smartserv/ceo/chaoscon)
09:52.54*** join/#bzflag chaoscon (~jeremy@smartserv/ceo/chaoscon)
10:08.46*** join/#bzflag Dontkillme (~chatzilla@frnk-590de0da.pool.mediaWays.net)
10:26.10*** join/#bzflag xaver (~xaver@pD9ED5E26.dip.t-dialin.net)
12:32.59bradfyi: mysql down
12:38.11*** join/#bzflag randomparticle (~randompar@about/essy/snick/randomparticle)
12:38.42randomparticleglobal is a bit unwell
12:39.14*** join/#bzflag menotume (~menotume@bzflag/contributor/menotume)
12:39.14*** mode/#bzflag [+v menotume] by ChanServ
12:40.27menotumeglobal down, mysql error :(
12:41.01menotumebrlcad awake, or should i hack at it? :P
12:41.09randomparticleis that a coding bug?
12:42.03bradmenotume: care for a pm?
12:42.39randomparticleshouldn't you wait until gordon has finished?
12:42.46menotumewhat is gordon doing ?
12:42.52randomparticlehe's being pm
12:42.55menotumei'm not doing anything yet, just looking
12:43.05randomparticlebut possibly won't be much longer
12:43.17brad>.<
12:43.21bradthat was terrible snick
12:43.31*** join/#bzflag kierra (~jolie@unaffiliated/kierra)
12:43.33menotumebeing pm ?
12:43.36randomparticlesorry. i can't resist the feeble efforts
12:44.16kierrayou should talk at my place ;)
12:45.16menotumeglobal fixed :P
12:45.38kierra:P
12:52.52*** part/#bzflag kierra (~jolie@unaffiliated/kierra)
12:55.36*** join/#bzflag randomparticle (~randompar@cpc2-wilm1-0-0-cust142.bagu.cable.ntl.com)
12:55.36*** join/#bzflag randomparticle (~randompar@about/essy/snick/randomparticle)
13:31.13*** part/#bzflag randomparticle (~randompar@about/essy/snick/randomparticle)
14:20.43blast007brlcad: could those PDF files be moved to a different web server...?
14:21.15blast007fighting a losing battle here trying to block all of the hosts that are hitting them.
14:23.44*** join/#bzflag bier (~bier@pD954F858.dip.t-dialin.net)
14:24.24*** join/#bzflag bier|tp (~bier@pD954F858.dip.t-dialin.net)
14:42.04*** join/#bzflag mrapple (~mrapple@c-69-180-173-220.hsd1.mn.comcast.net)
14:42.04*** join/#bzflag mrapple (~mrapple@bzflag/player/MrAppleComputer)
14:59.02*** join/#bzflag Dontkillme (~chatzilla@frnk-590de0da.pool.mediaWays.net)
15:11.38*** join/#bzflag bryjen (~bryjen@cpe-75-81-201-131.we.res.rr.com)
15:11.38*** mode/#bzflag [+v bryjen] by ChanServ
15:35.54*** join/#bzflag temporalD (~ATD@bzflag/serverop/TemporalDistraction)
16:00.46*** join/#bzflag chaoscon (~jeremy@smartserv/ceo/chaoscon)
16:16.43*** join/#bzflag chaoscon (~jeremy@smartserv/ceo/chaoscon)
16:23.25*** join/#bzflag chaoscon (~jeremy@smartserv/ceo/chaoscon)
16:24.39BulletCatcherblast007, brlcad: an Apache module for limiting connections per IP address is at http://dominia.org/djao/limitipconn2.html
16:32.18*** join/#bzflag chaoscon (~jeremy@smartserv/ceo/chaoscon)
16:39.56*** join/#bzflag chaoscon (~jeremy@smartserv/ceo/chaoscon)
16:52.17*** join/#bzflag chaoscon (~jeremy@smartserv/ceo/chaoscon)
16:57.17*** join/#bzflag JeffM2501 (~JeffM@bzflag/projectadmin/JeffM)
16:57.17*** mode/#bzflag [+o JeffM2501] by ChanServ
16:59.16*** join/#bzflag chaoscon (~jeremy@smartserv/ceo/chaoscon)
17:07.16*** join/#bzflag chaoscon (~jeremy@smartserv/ceo/chaoscon)
17:11.22*** join/#bzflag jcp (~jw@bzflag/contributor/javawizard2539)
17:19.29CIA-32BZFlag: 03macsforme * r20994 10/branches/v2_0branch/bzflag/src/bzfs/CmdLineOptions.cxx: Incorrect player count errors (for -mp) should be outputted to stderr, not stdout.
17:21.58CIA-32BZFlag: 03macsforme * r20995 10/trunk/bzflag/src/bzfs/CmdLineOptions.cxx: Incorrect player count errors (for -mp) should be outputted to stderr, not stdout.
17:24.18JeffM2501Constitution, that work on windows?
17:25.04Constitutionthat's how the same file handles other errors... I have no idea about windows
17:25.10JeffM2501ok
17:25.21JeffM2501it may work since it's a CLI app
17:25.32JeffM2501in the main client stderr dosn't output anything
17:25.57Constitutionyeah it's used widely like that throughout the file
17:26.23JeffM2501k
17:26.27JeffM2501then all is well
17:27.13JeffM2501well as well as can be for bzflag :)
17:27.54Constitutionit came to light because bzfs was failing to start, and was not reporting the error since stdout is redirected to a log file
17:29.01Constitutionand I just cat stderr to the startup web page (web admin interface) if bzfs fails to start... so anyway, less confusion now
17:30.30Constitutionnow that bzflag is good and well, as well as it can be, how is P2501? :)
17:31.05JeffM2501heh "good and well"
17:31.15JeffM2501I need to do the movement code
17:31.19JeffM2501but I am stuck on it a little
17:31.23JeffM2501so I may skip it
17:32.15JeffM2501and just move on to shot and weapon transport
17:34.21JeffM2501I'm sure I won't be able to finish it if I have to do it all on my own
17:36.15Constitutionsounds like it's time to form "Jeffery Myers Gaming Innovations Ltd." and start recruiting
17:36.32JeffM2501heh
17:36.44JeffM2501recruting needs money :)
17:37.16Constitutionsad how that is
17:37.25JeffM2501people gotta eat
17:37.27*** join/#bzflag me1 (~ausom@bzflag/player/Me1)
17:37.29JeffM2501and sadly food is not free
17:37.41JeffM2501change that and the world would be a very different place
17:38.04Constitutionis not sure he would want to know what the world would be like then
17:38.21JeffM2501hopefully it'd be a lot better
17:38.31cygalwindows writes to stderr.txt, no?
17:38.40JeffM2501in the client yeah
17:39.02JeffM2501had to add that since apps that start with a winmain don't bring up a console window
17:40.38JeffM2501you are more then welcome to grap the code for p2501 Constitution it's all in svn repos
17:41.00Constitutionok
17:41.05ConstitutionI have looked at it
17:41.13JeffM2501and I think I found out why it didn't work well on the mac and linux
17:41.19JeffM2501I forgot a file that they need :)
17:45.55*** join/#bzflag chaoscon (~jeremy@smartserv/ceo/chaoscon)
17:52.39*** join/#bzflag chaoscon (~jeremy@smartserv/ceo/chaoscon)
17:59.30*** join/#bzflag chaoscon (~jeremy@smartserv/ceo/chaoscon)
18:07.19BulletCatcherI have been looking at the trunk lag/jitter numbers problem.
18:07.20BulletCatcherSo far, the only interesting thing I have found is that bzfs frequently processes two MsgPlayerUpdateSmall packets in a single main loop pass, which causes the second one to increase the reported jitter because it has the same server timestamp.
18:08.15JeffM2501don't they have an ID?
18:08.26JeffM2501so they don't get taken out of order?
18:08.39BulletCatcherIt isn't an out-of-order problem.
18:08.48JeffM2501just 2 at once?
18:08.54cygalwhy does it process two packets?
18:09.09JeffM2501it probably got 2 during a frame
18:09.13BulletCatcherJitter is computed by comparing client timestamp changes to server timestamp changes.
18:09.22JeffM2501we should have it save all the updates it gets untill after the message loop
18:09.34JeffM2501then process just the newest ones at
18:09.35BulletCatcherWhen bzfs processes two at once, they have the same server timestamp.
18:09.39JeffM2501yeah
18:10.26BulletCatcherI see.  I'll look at changing that.  Fewer updates will make it smoother.
18:10.27JeffM2501or if the current packets timestamp is the same as the last update then ignore it for jitter
18:10.52BulletCatcherThat will be easy to do.  I'll try that first.
18:11.26JeffM2501did anyone review the lag code to see if it was changed for trunk ?
18:11.51BulletCatcherI see no relevant changes in the LagInfo class.
18:12.42BulletCatcher"diff" reports very few differences of any kind.
18:13.38JeffM2501odd that the lag would be so much larger in trunk then
18:13.56JeffM2501unless it was my changes to how the server handles messages
18:14.48JeffM2501perhaps the map lookup takes a very long time
18:15.12*** join/#bzflag I_Died_Once (~I_Died_On@c-98-244-158-51.hsd1.ga.comcast.net)
18:23.53*** join/#bzflag Swigg (~Default@bzflag/player/Swigg)
18:29.13BulletCatcherIgnoring jitter updates when the server timestamp is unchanged cuts the average jitter from 10ms to 5ms via loopback.
18:30.00BulletCatcherJitter numbers range from 4.2 to 6.5.
18:31.52BulletCatcherIf 5 is the new zero then we could live with that, I suppose, but there must be something else in the way.
18:31.54BulletCatcherThe new network handling code is my prime suspect.
18:33.41BulletCatcherI still haven't really looked at lag, the computation of which is completely independent of jitter.
18:40.03trepanan easy thing to try is to revert (completely), how timestamps are used
18:40.34trepan(ex: TimeKeeper::getCurrent() vs TimeKeeper::getTick(), and remove the getDRTime() stuff)
18:42.12trepan(note that v2_0branch's Player.cxx contains no getCurrent() calls)
18:45.58BulletCatcherI haven't gotten as far as looking at the client, yet.  Indeed, there are many opportunities there for packet delays.
18:47.29BulletCatcherIt could even be as simple as sending updates to the server at the end of the frame instead of earlier.
18:53.08*** part/#bzflag Upsetter (~Ups@89.246.172.128)
18:54.36BulletCatcherIn the bzfs main loop we call MSGMGR.update() near the end.  IIRC that flushes outgoing data that will have been been queued earlier in the loop.
18:55.34trepanwhy is outbound data being queued?
18:55.47trepan(with the exception of filled system buffers ...)
18:56.09BulletCatcherThat's the new network code.  (If I recall correctly.)
18:56.47blast007also, I had noticed that some messages are either being sent, received, or processed out of order.  bzfs was processing the MsgPlayerData before the MsgEnter, which failed since it didn't have a player to assign the player data to
18:58.14JeffM2501trepan, so that large groups of messages don't kill the transport
18:58.25JeffM2501like banlist and the like, lets them trickle out
18:58.52trepanonly need to queue them based on system buffers
18:59.09trepan(can use select() to detect when you can send more)
18:59.24JeffM2501IIRC windows dosn't have a big system buffer
18:59.27JeffM2501but whatever you say
19:03.25trepanbuffer sizes seem to be set in NetHandler.cxx
19:03.31trepanUDP is 128K
19:03.34trepanTCP is 64K
19:03.53JeffM2501then yank that code and fix it up
19:04.08trepanyou did it, you fix it  ;)
19:04.15BulletCatcherIt would help if we could send UDP messages immediately rather than queuing.
19:04.49JeffM2501too much has changed around it
19:04.59trepanmight as well do the same for TCP, tbh
19:06.30JeffM2501probalby can make a small change and have the buffer system send right away
19:07.14JeffM2501cus a TON of stuff got moved over to the buffered network message class
19:07.45BulletCatcherMore calls to MSGMGR.update() would be a proof-of-concept.
19:08.14JeffM2501there may be a simpler thing
19:08.25trepan(like stick one in BufferedNetworkMessageManager::queueMessage())
19:08.42JeffM2501yeah
19:09.25trepanit's odd that that call looks up the 'msg' in the current 'pendingOutgoingMessages'  (and places it in the back if it's already in there)
19:09.35trepanshould that ever happen?
19:09.49JeffM2501probably me just being paranoid
19:10.41trepanwonder if that has anything to do with the packet re-order blast mentionned, debug message  :)
19:11.05JeffM2501this is why I should not be allowed to write network code
19:11.27*** join/#bzflag Yassen (Yassen@unaffiliated/yassen)
19:12.30CIA-32BZFlag: 03JeffM2501 * r20996 10/trunk/bzflag/src/game/BufferedNetworkMessage.cxx: send messages when they are put into the buffer to see if that helps lag and crap
19:12.44JeffM2501there, that should send right away
19:14.24BulletCatcherUnfortunately, it didn't help much.
19:15.07JeffM2501if message manager is the problem then it will be hard to pull it's quite pervasive in the code
19:17.44CIA-32BZFlag: 03trepan * r20997 10/trunk/bzflag/src/game/BufferedNetworkMessage.cxx: * added a debug message
19:19.13trepannote that NetHandler already seems to have buffered output
19:19.38JeffM2501well then it wasn't working
19:19.50trepanbzfs.cxx/processConnectedPeer() -> sendBufferedNetDataForPeer() -> peer.netHandler->bufferedSend()
19:19.53JeffM2501the message buffer was added to fix lockups when the server sent out big things
19:20.35trepanhm, sendBufferedNetDataForPeer also has it's own buffer, 'sendChunks'
19:20.43trepan(well, the 'peer' does)
19:21.56*** join/#bzflag kierra (~jolie@unaffiliated/kierra)
19:22.10trepantriple buffering?  isn't that what they do for smoother gfx?  :)
19:23.09*** join/#bzflag Think_Differentl (~js@bzflag/player/ThinkDifferent)
19:25.34BulletCatchertrepan: That debug message fires every time, as far as I can tell.
19:28.12trepanah, yes, my bad
19:28.28trepanpendingOutgoingMessages vs. outgoingQueue ...
19:31.43CIA-32BZFlag: 03trepan * r20998 10/trunk/bzflag/src/game/BufferedNetworkMessage.cxx:
19:31.43CIA-32BZFlag: * removed the debug message from r20997,
19:31.43CIA-32BZFlag:  (pendingOutgoingMessages vs. outgoingQueue, mucho queueing)
20:00.56*** join/#bzflag xaver (~xaver@pD9ED5E26.dip.t-dialin.net)
20:01.31*** join/#bzflag xaver__ (~xaver@pD9ED5E26.dip.t-dialin.net)
20:09.36*** join/#bzflag xaver (~xaver@pD9ED5E26.dip.t-dialin.net)
20:16.48*** part/#bzflag kierra (~jolie@unaffiliated/kierra)
20:22.55CIA-32BZFlag: 03bullet_catcher * r20999 10/trunk/bzflag/src/game/LagInfo.cxx:
20:22.55CIA-32BZFlag: Log details of lag and jitter updates.
20:22.55CIA-32BZFlag: Ignore jitter updates when the server timestamp is unchanged.
20:24.36BulletCatcherThe jitter updates occur every update, so bzfs -ddddd is quite noisy.
20:42.46temporalD5 d's ?
20:45.14BulletCatcherIt enables level 5 debug messages.
20:49.45temporalDok - that's new to me - thx
20:50.52BulletCatcherI run my server with -dd to get level 2 messages for the log file.
20:51.13BulletCatcherjoins, chat, etc.
20:51.13*** join/#bzflag Swigg (~Default@bzflag/player/Swigg)
20:54.06*** join/#bzflag R0b0t1 (~Enigma@unaffiliated/r0b0t1)
21:02.50trepanfwiw, -d5 = -ddddd  (v2.99, server & client)
21:15.34*** join/#bzflag Packman (~chatzilla@125-236-135-85.jetstream.xtra.co.nz)
21:16.54Packmanis there a config or server variable option that specifies how often super flags are automatically reset?
21:17.31blast007_maxFlagGrabs controls the maximum number of times a flag can be picked up before it respawns somewhere else
21:18.01Packmanbut there's no /reset flags every minute?
21:18.05blast007no
21:18.22blast007a plugin could do it though
21:18.28blast007but I wouldn't recommend resetting them *all*
21:18.33blast007at least at the same time
21:19.05Packmanis there a way to keep flags in a zone meaning they can't be dropped outside?
21:19.29Packmanyou can make flags spawn in a zone but you can't prevent them from being 'taken' outside, if you have several maxflaggrabs
21:19.49blast007there's a flagStay plugin that makes you drop specific flag types if you leave a certain area
21:20.14blast007that one is even bundled with the source, afaik
21:20.50Packmanok, thank you
21:21.08blast007might be a wiki page on it
21:21.27blast007there doesn't seem to be any documentation on it in the provided README.txt
21:22.59*** join/#bzflag Yassen (Yassen@unaffiliated/yassen)
21:23.56Packmani found it at http://my.bzflag.org/w/FlagStay
21:24.55*** join/#bzflag McSpider (~McSpider@xplr-ts-v10-208-114-184-180.barrettxplore.com)
21:28.17*** join/#bzflag chaoscon (~jeremy@smartserv/ceo/chaoscon)
21:31.01temporalDPackman:  there is a really really buggy "bzcron" plugin that can run commands on a schedule
21:31.19temporalDmofo runs it and it fails to load about 70% of the time
21:31.35Packmanheh
21:31.43blast007really hackish too
21:32.21blast007it basically makes the server connect to itself as an observer, then gives that observer admin rights so it can run stuff ;)
21:32.49Packmanany idea if bboxes can be combined within the same plugin-supported object?
21:33.06CIA-32BZFlag: 03bullet_catcher * r21000 10/trunk/bzflag/src/game/MsgStrings.cxx: Clarify the output when messages are not decoded in detail.
21:33.21blast007Packman: you'd have to check the plugin source
21:58.59CIA-32BZFlag: 03bullet_catcher * r21001 10/trunk/bzflag/src/ (bzfs/bzfs.cxx game/BufferedNetworkMessage.cxx):
21:58.59CIA-32BZFlag: With -DDEBUG, log sent and received messages except for MsgPlayerUpdateSmall.
21:58.59CIA-32BZFlag: Add FIXME on r20996 change.
21:58.59CIA-32BZFlag: Log when player updates are out of order and ignored.
22:00.43BulletCatcherThat should be the server side of what is needed to help analyze out-of-order message processing, blast007.
22:05.28blast007k
22:24.47*** join/#bzflag JeffM1 (~Jeff@adsl-75-50-184-13.dsl.lsan03.sbcglobal.net)
23:04.49*** join/#bzflag chaoscon (~jeremy@smartserv/ceo/chaoscon)
23:05.17*** join/#bzflag JeffM (~Jeff@bzflag/projectadmin/JeffM)
23:05.17*** mode/#bzflag [+o JeffM] by ChanServ
23:15.25*** join/#bzflag joevano (~joevano@bzflag/developer/JoeVano)
23:15.25*** mode/#bzflag [+v joevano] by ChanServ
23:16.59*** join/#bzflag allejo (~allejo@pool-173-51-205-154.lsanca.fios.verizon.net)
23:20.05*** join/#bzflag chaoscon (~jeremy@smartserv/ceo/chaoscon)

Generated by irclog2html.pl Modified by Tim Riker to work with infobot.