#bitfighter IRC Log

IRC Log for 2012-03-28

Timestamps are in GMT/BST.

00:00:02 raptor Quit ()
01:11:42 koda Quit (Quit: koda)
01:33:23 FlynnnNT has joined
02:02:52 FlynnnNT Quit (Quit: This computer has gone to sleep)
02:08:22 FlynnnNT has joined
02:12:38 FlynnnNT Quit (Client Quit)
03:19:45 FlynnnNT has joined
03:23:57 FlynnnNT Quit (Client Quit)
03:25:43 raptor has joined
03:25:44 ChanServ sets mode +o raptor
03:26:10 raptor good evening!
05:00:37 BFLogBot - Commit f8519f05be49 | Author: buckyballreaction | Log: Comment out some debug messages
05:00:39 BFLogBot - Commit e9af84091517 | Author: buckyballreaction | Log: More game indicators in UIInstructions
05:04:19 sam686 Quit (Read error: Connection reset by peer)
05:07:38 sam686 has joined
05:07:38 ChanServ sets mode +v sam686
05:13:17 FlynnnNT has joined
05:41:36 raptor Quit ()
05:45:53 BFLogBot - Commit 7504f290a57c | Author: buckyballreaction | Log: Game release notes.  I think I got them all...
06:03:04 sam686 Quit (Ping timeout: 245 seconds)
06:03:33 sam686 has joined
06:03:33 ChanServ sets mode +v sam686
06:40:23 koda has joined
07:01:51 koda Quit (Quit: koda)
07:03:32 FlynnnNT Quit (Quit: This computer has gone to sleep)
07:12:28 FlynnnNT has joined
07:12:55 FlynnnNT Quit (Client Quit)
08:03:09 watusimoto has joined
08:03:09 ChanServ sets mode +o watusimoto
08:16:24 sam686 Quit (Ping timeout: 245 seconds)
09:18:09 kodaws has joined
11:33:51 LordDVG has joined
15:49:08 raptor has joined
15:49:09 ChanServ sets mode +o raptor
15:49:13 raptor good day!
16:06:13 raptor watusimoto: i did some testing last night..  didn't get it to crash; but i wasn't thorough - I was working on updating the release notes mostly
16:10:34 kodaws Quit (Ping timeout: 245 seconds)
16:11:04 raptor also I cannot for the life of me reproduce the crash that Little_Apple says happens on Mac
16:11:22 raptor I don't think I will be able until I get more specific data
16:30:48 watusimoto hi
16:31:01 watusimoto thanks for doing the notes
16:31:09 watusimoto I was pretty slack on that this time round
16:32:19 watusimoto I'm heading home now; my mother in law is here, so I don't know how much time I'll have tonight, but my only agenda is testing and bug fixing.
16:32:35 watusimoto LA did not provide much useful information
16:32:55 watusimoto so there's only so much you can do
16:39:07 watusimoto Quit (Ping timeout: 264 seconds)
16:41:10 FlynnnNT has joined
16:53:40 FlynnnNT Quit (Quit: This computer has gone to sleep)
18:27:39 FlynnnNT has joined
19:59:54 FlynnnNT Quit (Quit: This computer has gone to sleep)
20:08:14 LordDVG Quit (Read error: Connection reset by peer)
20:13:23 FlynnnNT has joined
20:26:21 sam686 has joined
20:26:25 ChanServ sets mode +v sam686
20:59:02 BlackBird_ has joined
20:59:09 BlackBird_ Guten Tag
20:59:17 raptor hello
20:59:28 BlackBird_ i have an idea
21:00:06 BlackBird_ You know how there are usually so few players that play bit fighter?
21:00:15 raptor y
21:00:42 FlynnnNT Quit (Quit: This computer has gone to sleep)
21:01:52 BlackBird_ what if when only one player is on, there should be like a message sent to the player saying nobody is on and suggesting to invite a friend or something
21:02:26 raptor like one of the popup messages?
21:02:29 raptor red ones
21:02:38 BlackBird_ yea i think
21:03:11 BlackBird_ and like make some in game emailing thing to email a friend to go and download bitfighter
21:03:13 BlackBird_ i dono
21:03:35 BlackBird_ I'm just kinda thinking out loud here
21:03:55 raptor "you are the lone player right now...  invite a friend!"
21:04:19 BlackBird_ HA
21:05:20 BlackBird_ but then i would have to open mail, then creat a new message, then like type stuff, and pick who i want to send it to, but then if he dosnt know how to download it i would have to provide a link, etc etc
21:06:10 raptor yeah, a bit complicated...
21:06:28 raptor maybe you want a form on bitfighter.org that just has a field for 'share bitfighter with a friend!'
21:06:50 raptor and we could program the mailing list to send it to the specified address
21:08:43 BlackBird_ its just that i see a lot of new players, but when i do see them, its on bitfightger.org, and nobody else is on
21:10:17 raptor you mean on the forums as opposed to in-game?
21:20:42 Watusimoto has joined
21:21:35 BlackBird_ wait wha?
21:21:46 BlackBird_ yea
21:21:51 BlackBird_ i think
21:22:36 Little_Apple has joined
21:22:44 Little_Apple RAPTOR
21:22:48 Little_Apple im on the ppc
21:22:52 Little_Apple LEMMIE TEST.
21:22:57 raptor hi Little_Apple
21:23:01 Little_Apple helloo
21:23:05 raptor ok, let me get you a build
21:23:10 Little_Apple k
21:23:13 Little_Apple thanx
21:23:26 Watusimoto hi
21:23:30 raptor hi Watusimoto
21:23:46 BlackBird_ hi
21:24:04 Little_Apple BlackBird_
21:24:07 Little_Apple :P
21:25:27 raptor Watusimoto:  Little_Apple suggested that a kinetic module for Cloak could be 'Spy', i.e. become another team's color
21:25:39 Little_Apple :3
21:25:53 Little_Apple until a weapon is activated
21:26:06 BlackBird_ like
21:26:11 BlackBird_ an extension?
21:26:20 raptor and BlackBird_ suggests some sort message saying you are the only player online, go and find friends
21:26:23 Little_Apple like a special aility
21:26:27 Little_Apple ability*
21:26:28 raptor uh, to play
21:26:31 BlackBird_ would it be more of an "add-on" for the cloak module?
21:26:32 Watusimoto interesting, but everyone would know
21:26:48 raptor but we should focus on releasing
21:26:51 raptor i think  :)
21:26:59 BlackBird_ oookie
21:26:59 Watusimoto of course
21:27:15 Little_Apple ITS FOR THE FUTUER
21:27:18 Little_Apple GOH GAWSH
21:28:12 Little_Apple durr durr durr
21:28:16 BlackBird_ bluh
21:28:35 BlackBird_ so, when is v017 going to be released?
21:28:38 Little_Apple now
21:29:50 raptor Watusimoto: will older clients work with newer master server protocol number?
21:30:02 Little_Apple :o
21:30:44 raptor we should probably update that protocol number too...
21:30:52 raptor i have done anything though - been busy at work
21:31:42 BlackBird_ RRGG
21:31:46 BlackBird_ WHENNNNNNUH
21:32:11 Watusimoto if, when we added new m2s and m2c and c2m and s2m methods we used an updated version number, then yes
21:32:37 Watusimoto that is, look at the definitions of one of those funtcions -- each contains a version number
21:33:14 BlackBird_ u talkin to me?
21:33:40 Watusimoto if client and server agree on the methods in a given version number, they're ok.  older clients will ignore functions with a newer number.
21:34:07 Watusimoto the master version number in version.h is not that important
21:34:19 sam686 016 and 017 still works with the same new master... (015 may have used old master)
21:34:23 Watusimoto I think that's mainly for us to keep track of... something?
21:34:48 sam686 the master knows what version the client/server is, through mCMProtocolVersion in MasterServerConnection::readConnectRequest
21:36:04 raptor in masterInterface.cpp, i see a '1' is tied to the high score stuff
21:36:07 Little_Apple i never got that build ._.
21:36:07 raptor instead of 0
21:36:14 raptor but the master protocol is '4'
21:36:20 Little_Apple HEEEY
21:36:25 sam686 as for masterInterface.cpp, there is RPCVersion (a version 0 only client, ignores any version 1 or later commands)
21:36:33 raptor Little_Apple: http://sam686.maxhushahn.com/upload/Bitfighter-017rc-MacOSX-32bit-Universal.dmg
21:36:41 sam686 master protocol was "3" for 015 and earliear
21:36:42 Little_Apple thanchsh
21:37:27 sam686 oh, i mean mCSProtocolVersion the master also knows
21:38:04 sam686 which is different between 017 and 016 (which is also sent to master for master to know)
21:39:10 Little_Apple HOLY CRUDS!
21:39:14 Little_Apple MENU MUSIC!!
21:39:48 Watusimoto hey kids, could you keep the chatter down?
21:39:52 Little_Apple NO
21:39:52 BlackBird_ wut
21:39:57 Little_Apple ksure
21:40:35 Watusimoto I think things got a little confused when we migrated the master and we reduced all the rpc versions to 0
21:40:59 Watusimoto what sam said was correct
21:42:27 Watusimoto the protocol version defined in version.h is used for defining what gets sent during handshaking, I think
21:42:51 Watusimoto or something
21:42:58 raptor heh
21:43:31 Watusimoto no, maybe it's not
21:44:07 Watusimoto it's only used once, here
21:44:09 Watusimoto       if(mCMProtocolVersion < 4 || mCMProtocolVersion > MASTER_PROTOCOL_VERSION) // check for unsupported version
21:44:20 Watusimoto if that evaluates to true, conn is rejected
21:44:30 raptor ah good
21:44:38 raptor then increasing it won't affect older clients
21:44:57 Watusimoto and mCMProtocolVersion is defined as
21:44:57 Watusimoto      bstream->read(&mCMProtocolVersion);    // Version of protocol we'll use with the client
21:44:58 sam686 it is mostly been coded to reject incompatible version (as it won't work anyway due to deleting / changing all individual RPCversion version to 0
21:45:10 Watusimoto a value sent to the master during the handshaking
21:45:31 BlackBird_ Quit (Quit: Page closed)
21:45:31 raptor i remember that
21:45:38 Watusimoto so... in the context of your question, we're fine
21:45:49 raptor the only way to make it incompatible was to remove something in the middle of the bitstream
21:45:57 raptor like we did with joystick data
21:47:06 raptor is the windows packaging script ok with the new menu music/etc.?
21:47:06 sam686 zap/masterConnection.cpp line 407 (MasterServerConnection::writeConnectRequest) sends most of client's versions
21:47:15 sam686 then the master reads it in readConnectRequest
21:47:22 raptor mac should be fine now - Little_Apple just confirmed
21:47:31 Little_Apple :P
21:47:57 raptor and i'll always do linux last...  but weeks before Ubuntu, i'm sure
21:56:30 BlackBird_ has joined
21:57:17 BFLogBot - Commit dfb42011bd5e | Author: sam8641 | Log: Fix dedicated compile
21:57:46 raptor linky: http://bitfighter.org/wiki/index.php?title=Release_checklist
21:58:35 BlackBird_ fer mee?
21:58:40 raptor for us
21:58:43 raptor for releasing
21:59:10 BlackBird_ oh
21:59:22 Little_Apple lol
21:59:59 BlackBird_ waaaa
22:04:29 Little_Apple doodeedoo
22:06:59 kodaws has joined
22:11:18 kodaws Quit (Ping timeout: 245 seconds)
22:13:10 BlackBird_ ZAP!
22:13:27 BlackBird_ Quit (Quit: Page closed)
22:15:08 sam686 I found a rather strange problem, go in a server, addbots, leave server, and rejoin the same server, all the bots is still there, with several points, as if the level didn't switch and reset robots score...
22:16:09 Watusimoto are those two 017 test servers running the latest code?
22:16:23 raptor no let me recompile
22:16:39 Watusimoto more specifically, the code I comitted last night
22:16:47 raptor not form last night
22:16:53 raptor compiling...
22:16:59 sam686 my "sam test" is using the latest changes, at the moment in 017
22:17:08 Watusimoto holy crap sam!
22:17:20 sam686 leave and rejoin, and check scoreboard
22:17:20 Watusimoto your server looks like a jelly fish party
22:17:44 Watusimoto all 0's
22:17:48 sam686 it doesn't want to reset scores (and change levels like it does on older 016)
22:17:59 Watusimoto sorry, that's ping, not score
22:18:21 raptor server isn't suspending with only bots left?
22:18:53 Watusimoto correct
22:18:54 sam686 when all players leave (bots don't count) level switches and reset scores on 016, the problem is it doesn't happen on 017, but just pauses when all player leave
22:19:08 Watusimoto but it does pause?
22:19:30 sam686 yes, i even see my server CPU usage go to 0 % when no player in it
22:19:55 Watusimoto ok, even without bots it doesn't reset
22:20:11 Watusimoto so it's not a bot problem
22:21:19 sam686 yes, even with no bots, time limit simply pauses when all players leave, and doesn't switch level like it does on 016.
22:21:39 raptor ok, other server is back up
22:27:06 Watusimoto in server's idle loop is this check:
22:27:08 Watusimoto    if(getPlayerCount() == 0 && !mGameSuspended && mCurrentTime != 0)
22:27:08 Watusimoto       suspendGame();
22:27:42 Watusimoto suspendGame() then runs cycleLevel
22:27:47 Watusimoto clearly this isn't being hit
22:28:05 raptor are bots accidentally counted as players somehow?
22:28:11 Watusimoto so I'm guessing one of the other mechanisms we (I) added for 017 is suspending before this is hit
22:28:19 Watusimoto no -- the bug happens even with no bots
22:28:24 raptor ah
22:28:40 Watusimoto and that other mechanism is NOT running suspendGame()
22:28:44 Watusimoto that's my theory
22:29:07 Watusimoto in reality, this should not be in idle, so perhaps the other mechanism is in a more logical place
22:30:08 Watusimoto and...
22:30:09 Watusimoto void ServerGame::suspendIfNoActivePlayers()
22:30:14 Watusimoto sets mSuspendGame
22:30:22 Watusimoto without cycling the level
22:30:36 Watusimoto so that's the problem!
22:32:15 raptor tada!
22:33:15 Watusimoto will checkin a possible fix
22:33:46 Little_Apple Quit (Ping timeout: 245 seconds)
22:35:18 Little_Apple has joined
22:35:20 sam686 when using profiler on bitfighter server, 50 robots, i find that "realloc" uses 11.30% of CPU, all of which comes from LUA looking at all call stacks
22:35:25 Watusimoto new code is in
22:35:42 Little_Apple ooh?
22:35:52 Watusimoto if you could recompile and restart your servers...
22:36:12 Watusimoto we need to fix lua resource consumption
22:36:20 Little_Apple :o
22:36:33 raptor recompiled and restarted
22:36:52 sam686 compiling mine (takes about 60-120 seconds)
22:37:21 BFLogBot - Commit 3d04dcb633c3 | Author: watusim...@bitfighter.org | Log: Untested possible fix for game not advancing when suspended
22:37:23 BFLogBot - Commit 2043c91fbd51 | Author: watusim...@bitfighter.org | Log: uh... something?
22:37:24 BFLogBot - Commit b6389bea6e3e | Author: watusim...@bitfighter.org | Log: Merge
22:39:03 raptor looks like it works
22:40:27 Little_Apple yay
22:43:20 Little_Apple soooo... whats going on?
22:43:22 Watusimoto here is where it (might) fail
22:43:29 Watusimoto 2 players on server
22:43:34 Watusimoto both players go idle
22:43:49 Watusimoto even  1 player on server goes idel
22:43:52 Watusimoto closes client
22:43:58 Watusimoto server may not advance
22:46:04 sam686 oh, the admin pass for my "sam test" is 2  (i saw raptor attemt to guess the passwords on sam test)
22:46:15 raptor ok, thanks
22:46:36 sam686 or so i think it is 2, let me check
22:46:48 raptor yes, it is
22:47:37 sam686 ok, so i guess it just doesn't log or display on console that a password was successful (like it does when password is wrong)
22:48:00 Watusimoto probably not -- we could add a log level for that
22:48:30 Watusimoto crash
22:48:34 raptor ccrash!
22:48:46 Watusimoto crap
22:49:17 sam686 i didn't crash, because i guesss i didn't idle long enough
22:49:28 raptor getting stack trace...
22:50:00 Watusimoto I see the problem
22:50:01 Watusimoto sort of
22:50:09 Watusimoto can't find a client with a specific name
22:50:13 raptor crash: http://pastie.org/3688460
22:50:21 Watusimoto but the stringtableentry just gives me the id, not the name it's looking for
22:50:32 sam686 i still didn't crash, for some reason... (that was a release build, though)
22:50:33 raptor brb
22:50:53 Watusimoto guessing it was me or raptor, since sam was not idle
22:50:56 raptor Assert: Could not find clientInfo! in gameConnection.cpp line 1176
22:51:02 Watusimoto but only tripped assert
22:51:09 Watusimoto would not have crashed game
22:51:11 sam686 you hit a breakpoint, i guess..., raptor, can you continue past the breakpoint?
22:51:26 Watusimoto I totally see the problem
22:51:37 Watusimoto look at TNL_IMPLEMENT_RPC(GameConnection, s2cSetIsIdle
22:51:42 Watusimoto he hit the assert in there
22:51:52 Watusimoto the question is why, and is it ok
22:51:55 Little_Apple POOTIS
22:51:59 Little_Apple Quit (Quit: Page closed)
22:52:04 Watusimoto if we remove the assert, the game won't crash
22:52:15 Watusimoto but want to make sure it;s ok
22:52:34 Watusimoto ok
22:52:45 Watusimoto it crashed for me on trying to find clientInfo for raptor
22:52:58 Watusimoto ah
22:53:08 Watusimoto I do not have clientinfo for him yet because game just cycled
22:53:17 Watusimoto so this is a problem
22:53:18 sam686 maybe, thats because, the client removes all client info, cause the level is changing, and haven't received new client infos
22:53:23 Watusimoto exactly
22:53:46 sam686 and was trying to setisidle in between clear client infos, and waiting to get new client infos
22:53:49 Watusimoto so what happens is if I ignore this, then when i do get his clientinfo (CI) I won't know he's idle
22:54:57 Watusimoto so we can 1) delay calling this; 2) cache name locally; 3) ignore and send idle status with new client connection or 4) panic
22:55:01 Watusimoto 2 is probably out
22:55:08 Watusimoto maybe the solution is 3
22:56:22 Watusimoto that's what I'll do
22:56:43 raptor i've gotta go soon and won't be back for a few hours :(
22:56:46 Watusimoto only costs 1 bit in s2caddclient
22:56:48 Watusimoto bye
22:57:00 Watusimoto hasta manana
22:57:02 raptor 1 bit
22:57:05 sam686 maybe it might help to move GameConnection, s2cSetIsIdle to GameType, s2cSetIsIdle, because that is where the  RemoteClientInfo is (GameType, s2cAddClient)
22:57:44 sam686 or at least that is where RemoteClientInfo is created
22:58:08 sam686 or, both s2cAddClient and s2cSetIsIdle can both be in GameConnection, or both in GameType
22:59:11 sam686 that way, RPCGuaranteedOrdered will work within the same class
23:02:05 Watusimoto sam686: do refPtrs need to be deleted?
23:02:28 raptor Quit ()
23:02:32 sam686 RefPtr auto deletes when no more reference to it..
23:02:38 Watusimoto ok
23:03:23 sam686 while RefPtr does auto delete (kind of like boost::SharedPtr), SafePtr does not auto delete...
23:03:36 Watusimoto right -- I always get confused
23:03:42 Watusimoto safeptr goes to null when deleted
23:04:17 sam686 refptr cannot be deleted (will get errors), you just need to let it go away, and it will auto delete
23:04:54 Watusimoto sam -- you can move the rpcs after I do my next checkin
23:05:02 sam686 ok
23:14:45 Watusimoto compiling now
23:17:50 Watusimoto code is in
23:20:02 sam686 let me try my changes first (moving s2cSetIsIdle to GameType) without your changes and see if that alone fixes the problem...
23:22:29 BFLogBot - Commit 019888004884 | Author: watusim...@bitfighter.org | Log: Fixes for assert with spawn delayed players after level cycles
23:23:33 Watusimoto ok
23:24:21 Watusimoto but
23:24:28 Watusimoto actually we might need them regardless
23:24:39 Watusimoto if player is spawn delayed, and game is suspended
23:24:42 Watusimoto and a new player joing
23:24:47 Watusimoto game will unsuspend
23:24:56 Watusimoto but new player needs to know that existing player is delayed
23:25:06 Watusimoto s2csetidle won't be called
23:26:16 sam686 oh, i guess thats another problem, i see, new clients join won't know if existing player is idling...
23:29:58 Watusimoto now I found another problem -- after waking up, clients think they themselves are still asleep on the scoreboard!
23:30:54 Watusimoto should be an easy fix
23:34:58 sam686 sort of found a new problem, the buzy "<< >>" stays there even though it is actually not on any menus or anything...
23:35:19 FlynnnNT has joined
23:36:44 Watusimoto fixed
23:36:53 Watusimoto my problem, not yours
23:36:57 Watusimoto really
23:41:06 sam686 no wonder why client don't recieve it when moved s2cSetIsIdle to gameType, it is not ghosted yet on client, there is GameType::onGhostAvailable()
23:43:27 Watusimoto well... tracing everything through... the only place the << >> gets turned off is when cancelChat() is called on client
23:43:46 Watusimoto so you should be able to turn off << >> by entering and exiting chat
23:44:04 Watusimoto if so, the problem is that other places don't disable busy signal
23:44:14 Watusimoto we turn it on in a bunch of places
23:45:11 Watusimoto aha!
23:45:18 Watusimoto a major flaw in the system!
23:45:38 Watusimoto busy status only gets updated when the ship moves
23:45:47 Watusimoto I never liked sending it via ship object
23:45:50 Watusimoto will fix that now
23:55:59 sam686 i am about to push a change that moves s2cAddClient to GameType and away from gameConnection, the advantags is it makes FullClientInfo::setSpawnDelayed go down to a simple single line, instead of looping through all client infos
23:58:17 sam686 i mean, moves s2cSetIsIdle to gametype and away from gameConnection

Index Search ←Prev date Next date→

These logs were automatically created by BFLogBot on irc.freenode.net.

00:00:02		raptor Quit ()
01:11:42		koda Quit (Quit: koda)
01:33:23		FlynnnNT has joined
02:02:52		FlynnnNT Quit (Quit: This computer has gone to sleep)
02:08:22		FlynnnNT has joined
02:12:38		FlynnnNT Quit (Client Quit)
03:19:45		FlynnnNT has joined
03:23:57		FlynnnNT Quit (Client Quit)
03:25:43		raptor has joined
03:25:44		ChanServ sets mode +o raptor
03:26:10	raptor	good evening!
05:00:37		BFLogBot - Commit f8519f05be49 \| Author: buckyballreaction \| Log: Comment out some debug messages
05:00:39		BFLogBot - Commit e9af84091517 \| Author: buckyballreaction \| Log: More game indicators in UIInstructions
05:04:19		sam686 Quit (Read error: Connection reset by peer)
05:07:38		sam686 has joined
05:07:38		ChanServ sets mode +v sam686
05:13:17		FlynnnNT has joined
05:41:36		raptor Quit ()
05:45:53		BFLogBot - Commit 7504f290a57c \| Author: buckyballreaction \| Log: Game release notes. I think I got them all...
06:03:04		sam686 Quit (Ping timeout: 245 seconds)
06:03:33		sam686 has joined
06:03:33		ChanServ sets mode +v sam686
06:40:23		koda has joined
07:01:51		koda Quit (Quit: koda)
07:03:32		FlynnnNT Quit (Quit: This computer has gone to sleep)
07:12:28		FlynnnNT has joined
07:12:55		FlynnnNT Quit (Client Quit)
08:03:09		watusimoto has joined
08:03:09		ChanServ sets mode +o watusimoto
08:16:24		sam686 Quit (Ping timeout: 245 seconds)
09:18:09		kodaws has joined
11:33:51		LordDVG has joined
15:49:08		raptor has joined
15:49:09		ChanServ sets mode +o raptor
15:49:13	raptor	good day!
16:06:13	raptor	watusimoto: i did some testing last night.. didn't get it to crash; but i wasn't thorough - I was working on updating the release notes mostly
16:10:34		kodaws Quit (Ping timeout: 245 seconds)
16:11:04	raptor	also I cannot for the life of me reproduce the crash that Little_Apple says happens on Mac
16:11:22	raptor	I don't think I will be able until I get more specific data
16:30:48	watusimoto	hi
16:31:01	watusimoto	thanks for doing the notes
16:31:09	watusimoto	I was pretty slack on that this time round
16:32:19	watusimoto	I'm heading home now; my mother in law is here, so I don't know how much time I'll have tonight, but my only agenda is testing and bug fixing.
16:32:35	watusimoto	LA did not provide much useful information
16:32:55	watusimoto	so there's only so much you can do
16:39:07		watusimoto Quit (Ping timeout: 264 seconds)
16:41:10		FlynnnNT has joined
16:53:40		FlynnnNT Quit (Quit: This computer has gone to sleep)
18:27:39		FlynnnNT has joined
19:59:54		FlynnnNT Quit (Quit: This computer has gone to sleep)
20:08:14		LordDVG Quit (Read error: Connection reset by peer)
20:13:23		FlynnnNT has joined
20:26:21		sam686 has joined
20:26:25		ChanServ sets mode +v sam686
20:59:02		BlackBird_ has joined
20:59:09	BlackBird_	Guten Tag
20:59:17	raptor	hello
20:59:28	BlackBird_	i have an idea
21:00:06	BlackBird_	You know how there are usually so few players that play bit fighter?
21:00:15	raptor	y
21:00:42		FlynnnNT Quit (Quit: This computer has gone to sleep)
21:01:52	BlackBird_	what if when only one player is on, there should be like a message sent to the player saying nobody is on and suggesting to invite a friend or something
21:02:26	raptor	like one of the popup messages?
21:02:29	raptor	red ones
21:02:38	BlackBird_	yea i think
21:03:11	BlackBird_	and like make some in game emailing thing to email a friend to go and download bitfighter
21:03:13	BlackBird_	i dono
21:03:35	BlackBird_	I'm just kinda thinking out loud here
21:03:55	raptor	"you are the lone player right now... invite a friend!"
21:04:19	BlackBird_	HA
21:05:20	BlackBird_	but then i would have to open mail, then creat a new message, then like type stuff, and pick who i want to send it to, but then if he dosnt know how to download it i would have to provide a link, etc etc
21:06:10	raptor	yeah, a bit complicated...
21:06:28	raptor	maybe you want a form on bitfighter.org that just has a field for 'share bitfighter with a friend!'
21:06:50	raptor	and we could program the mailing list to send it to the specified address
21:08:43	BlackBird_	its just that i see a lot of new players, but when i do see them, its on bitfightger.org, and nobody else is on
21:10:17	raptor	you mean on the forums as opposed to in-game?
21:20:42		Watusimoto has joined
21:21:35	BlackBird_	wait wha?
21:21:46	BlackBird_	yea
21:21:51	BlackBird_	i think
21:22:36		Little_Apple has joined
21:22:44	Little_Apple	RAPTOR
21:22:48	Little_Apple	im on the ppc
21:22:52	Little_Apple	LEMMIE TEST.
21:22:57	raptor	hi Little_Apple
21:23:01	Little_Apple	helloo
21:23:05	raptor	ok, let me get you a build
21:23:10	Little_Apple	k
21:23:13	Little_Apple	thanx
21:23:26	Watusimoto	hi
21:23:30	raptor	hi Watusimoto
21:23:46	BlackBird_	hi
21:24:04	Little_Apple	BlackBird_
21:24:07	Little_Apple	:P
21:25:27	raptor	Watusimoto: Little_Apple suggested that a kinetic module for Cloak could be 'Spy', i.e. become another team's color
21:25:39	Little_Apple	:3
21:25:53	Little_Apple	until a weapon is activated
21:26:06	BlackBird_	like
21:26:11	BlackBird_	an extension?
21:26:20	raptor	and BlackBird_ suggests some sort message saying you are the only player online, go and find friends
21:26:23	Little_Apple	like a special aility
21:26:27	Little_Apple	ability*
21:26:28	raptor	uh, to play
21:26:31	BlackBird_	would it be more of an "add-on" for the cloak module?
21:26:32	Watusimoto	interesting, but everyone would know
21:26:48	raptor	but we should focus on releasing
21:26:51	raptor	i think :)
21:26:59	BlackBird_	oookie
21:26:59	Watusimoto	of course
21:27:15	Little_Apple	ITS FOR THE FUTUER
21:27:18	Little_Apple	GOH GAWSH
21:28:12	Little_Apple	durr durr durr
21:28:16	BlackBird_	bluh
21:28:35	BlackBird_	so, when is v017 going to be released?
21:28:38	Little_Apple	now
21:29:50	raptor	Watusimoto: will older clients work with newer master server protocol number?
21:30:02	Little_Apple	:o
21:30:44	raptor	we should probably update that protocol number too...
21:30:52	raptor	i have done anything though - been busy at work
21:31:42	BlackBird_	RRGG
21:31:46	BlackBird_	WHENNNNNNUH
21:32:11	Watusimoto	if, when we added new m2s and m2c and c2m and s2m methods we used an updated version number, then yes
21:32:37	Watusimoto	that is, look at the definitions of one of those funtcions -- each contains a version number
21:33:14	BlackBird_	u talkin to me?
21:33:40	Watusimoto	if client and server agree on the methods in a given version number, they're ok. older clients will ignore functions with a newer number.
21:34:07	Watusimoto	the master version number in version.h is not that important
21:34:19	sam686	016 and 017 still works with the same new master... (015 may have used old master)
21:34:23	Watusimoto	I think that's mainly for us to keep track of... something?
21:34:48	sam686	the master knows what version the client/server is, through mCMProtocolVersion in MasterServerConnection::readConnectRequest
21:36:04	raptor	in masterInterface.cpp, i see a '1' is tied to the high score stuff
21:36:07	Little_Apple	i never got that build ._.
21:36:07	raptor	instead of 0
21:36:14	raptor	but the master protocol is '4'
21:36:20	Little_Apple	HEEEY
21:36:25	sam686	as for masterInterface.cpp, there is RPCVersion (a version 0 only client, ignores any version 1 or later commands)
21:36:33	raptor	Little_Apple: http://sam686.maxhushahn.com/upload/Bitfighter-017rc-MacOSX-32bit-Universal.dmg
21:36:41	sam686	master protocol was "3" for 015 and earliear
21:36:42	Little_Apple	thanchsh
21:37:27	sam686	oh, i mean mCSProtocolVersion the master also knows
21:38:04	sam686	which is different between 017 and 016 (which is also sent to master for master to know)
21:39:10	Little_Apple	HOLY CRUDS!
21:39:14	Little_Apple	MENU MUSIC!!
21:39:48	Watusimoto	hey kids, could you keep the chatter down?
21:39:52	Little_Apple	NO
21:39:52	BlackBird_	wut
21:39:57	Little_Apple	ksure
21:40:35	Watusimoto	I think things got a little confused when we migrated the master and we reduced all the rpc versions to 0
21:40:59	Watusimoto	what sam said was correct
21:42:27	Watusimoto	the protocol version defined in version.h is used for defining what gets sent during handshaking, I think
21:42:51	Watusimoto	or something
21:42:58	raptor	heh
21:43:31	Watusimoto	no, maybe it's not
21:44:07	Watusimoto	it's only used once, here
21:44:09	Watusimoto	if(mCMProtocolVersion < 4 \|\| mCMProtocolVersion > MASTER_PROTOCOL_VERSION) // check for unsupported version
21:44:20	Watusimoto	if that evaluates to true, conn is rejected
21:44:30	raptor	ah good
21:44:38	raptor	then increasing it won't affect older clients
21:44:57	Watusimoto	and mCMProtocolVersion is defined as
21:44:57	Watusimoto	bstream->read(&mCMProtocolVersion); // Version of protocol we'll use with the client
21:44:58	sam686	it is mostly been coded to reject incompatible version (as it won't work anyway due to deleting / changing all individual RPCversion version to 0
21:45:10	Watusimoto	a value sent to the master during the handshaking
21:45:31		BlackBird_ Quit (Quit: Page closed)
21:45:31	raptor	i remember that
21:45:38	Watusimoto	so... in the context of your question, we're fine
21:45:49	raptor	the only way to make it incompatible was to remove something in the middle of the bitstream
21:45:57	raptor	like we did with joystick data
21:47:06	raptor	is the windows packaging script ok with the new menu music/etc.?
21:47:06	sam686	zap/masterConnection.cpp line 407 (MasterServerConnection::writeConnectRequest) sends most of client's versions
21:47:15	sam686	then the master reads it in readConnectRequest
21:47:22	raptor	mac should be fine now - Little_Apple just confirmed
21:47:31	Little_Apple	:P
21:47:57	raptor	and i'll always do linux last... but weeks before Ubuntu, i'm sure
21:56:30		BlackBird_ has joined
21:57:17		BFLogBot - Commit dfb42011bd5e \| Author: sam8641 \| Log: Fix dedicated compile
21:57:46	raptor	linky: http://bitfighter.org/wiki/index.php?title=Release_checklist
21:58:35	BlackBird_	fer mee?
21:58:40	raptor	for us
21:58:43	raptor	for releasing
21:59:10	BlackBird_	oh
21:59:22	Little_Apple	lol
21:59:59	BlackBird_	waaaa
22:04:29	Little_Apple	doodeedoo
22:06:59		kodaws has joined
22:11:18		kodaws Quit (Ping timeout: 245 seconds)
22:13:10	BlackBird_	ZAP!
22:13:27		BlackBird_ Quit (Quit: Page closed)
22:15:08	sam686	I found a rather strange problem, go in a server, addbots, leave server, and rejoin the same server, all the bots is still there, with several points, as if the level didn't switch and reset robots score...
22:16:09	Watusimoto	are those two 017 test servers running the latest code?
22:16:23	raptor	no let me recompile
22:16:39	Watusimoto	more specifically, the code I comitted last night
22:16:47	raptor	not form last night
22:16:53	raptor	compiling...
22:16:59	sam686	my "sam test" is using the latest changes, at the moment in 017
22:17:08	Watusimoto	holy crap sam!
22:17:20	sam686	leave and rejoin, and check scoreboard
22:17:20	Watusimoto	your server looks like a jelly fish party
22:17:44	Watusimoto	all 0's
22:17:48	sam686	it doesn't want to reset scores (and change levels like it does on older 016)
22:17:59	Watusimoto	sorry, that's ping, not score
22:18:21	raptor	server isn't suspending with only bots left?
22:18:53	Watusimoto	correct
22:18:54	sam686	when all players leave (bots don't count) level switches and reset scores on 016, the problem is it doesn't happen on 017, but just pauses when all player leave
22:19:08	Watusimoto	but it does pause?
22:19:30	sam686	yes, i even see my server CPU usage go to 0 % when no player in it
22:19:55	Watusimoto	ok, even without bots it doesn't reset
22:20:11	Watusimoto	so it's not a bot problem
22:21:19	sam686	yes, even with no bots, time limit simply pauses when all players leave, and doesn't switch level like it does on 016.
22:21:39	raptor	ok, other server is back up
22:27:06	Watusimoto	in server's idle loop is this check:
22:27:08	Watusimoto	if(getPlayerCount() == 0 && !mGameSuspended && mCurrentTime != 0)
22:27:08	Watusimoto	suspendGame();
22:27:42	Watusimoto	suspendGame() then runs cycleLevel
22:27:47	Watusimoto	clearly this isn't being hit
22:28:05	raptor	are bots accidentally counted as players somehow?
22:28:11	Watusimoto	so I'm guessing one of the other mechanisms we (I) added for 017 is suspending before this is hit
22:28:19	Watusimoto	no -- the bug happens even with no bots
22:28:24	raptor	ah
22:28:40	Watusimoto	and that other mechanism is NOT running suspendGame()
22:28:44	Watusimoto	that's my theory
22:29:07	Watusimoto	in reality, this should not be in idle, so perhaps the other mechanism is in a more logical place
22:30:08	Watusimoto	and...
22:30:09	Watusimoto	void ServerGame::suspendIfNoActivePlayers()
22:30:14	Watusimoto	sets mSuspendGame
22:30:22	Watusimoto	without cycling the level
22:30:36	Watusimoto	so that's the problem!
22:32:15	raptor	tada!
22:33:15	Watusimoto	will checkin a possible fix
22:33:46		Little_Apple Quit (Ping timeout: 245 seconds)
22:35:18		Little_Apple has joined
22:35:20	sam686	when using profiler on bitfighter server, 50 robots, i find that "realloc" uses 11.30% of CPU, all of which comes from LUA looking at all call stacks
22:35:25	Watusimoto	new code is in
22:35:42	Little_Apple	ooh?
22:35:52	Watusimoto	if you could recompile and restart your servers...
22:36:12	Watusimoto	we need to fix lua resource consumption
22:36:20	Little_Apple	:o
22:36:33	raptor	recompiled and restarted
22:36:52	sam686	compiling mine (takes about 60-120 seconds)
22:37:21		BFLogBot - Commit 3d04dcb633c3 \| Author: watusim...@bitfighter.org \| Log: Untested possible fix for game not advancing when suspended
22:37:23		BFLogBot - Commit 2043c91fbd51 \| Author: watusim...@bitfighter.org \| Log: uh... something?
22:37:24		BFLogBot - Commit b6389bea6e3e \| Author: watusim...@bitfighter.org \| Log: Merge
22:39:03	raptor	looks like it works
22:40:27	Little_Apple	yay
22:43:20	Little_Apple	soooo... whats going on?
22:43:22	Watusimoto	here is where it (might) fail
22:43:29	Watusimoto	2 players on server
22:43:34	Watusimoto	both players go idle
22:43:49	Watusimoto	even 1 player on server goes idel
22:43:52	Watusimoto	closes client
22:43:58	Watusimoto	server may not advance
22:46:04	sam686	oh, the admin pass for my "sam test" is 2 (i saw raptor attemt to guess the passwords on sam test)
22:46:15	raptor	ok, thanks
22:46:36	sam686	or so i think it is 2, let me check
22:46:48	raptor	yes, it is
22:47:37	sam686	ok, so i guess it just doesn't log or display on console that a password was successful (like it does when password is wrong)
22:48:00	Watusimoto	probably not -- we could add a log level for that
22:48:30	Watusimoto	crash
22:48:34	raptor	ccrash!
22:48:46	Watusimoto	crap
22:49:17	sam686	i didn't crash, because i guesss i didn't idle long enough
22:49:28	raptor	getting stack trace...
22:50:00	Watusimoto	I see the problem
22:50:01	Watusimoto	sort of
22:50:09	Watusimoto	can't find a client with a specific name
22:50:13	raptor	crash: http://pastie.org/3688460
22:50:21	Watusimoto	but the stringtableentry just gives me the id, not the name it's looking for
22:50:32	sam686	i still didn't crash, for some reason... (that was a release build, though)
22:50:33	raptor	brb
22:50:53	Watusimoto	guessing it was me or raptor, since sam was not idle
22:50:56	raptor	Assert: Could not find clientInfo! in gameConnection.cpp line 1176
22:51:02	Watusimoto	but only tripped assert
22:51:09	Watusimoto	would not have crashed game
22:51:11	sam686	you hit a breakpoint, i guess..., raptor, can you continue past the breakpoint?
22:51:26	Watusimoto	I totally see the problem
22:51:37	Watusimoto	look at TNL_IMPLEMENT_RPC(GameConnection, s2cSetIsIdle
22:51:42	Watusimoto	he hit the assert in there
22:51:52	Watusimoto	the question is why, and is it ok
22:51:55	Little_Apple	POOTIS
22:51:59		Little_Apple Quit (Quit: Page closed)
22:52:04	Watusimoto	if we remove the assert, the game won't crash
22:52:15	Watusimoto	but want to make sure it;s ok
22:52:34	Watusimoto	ok
22:52:45	Watusimoto	it crashed for me on trying to find clientInfo for raptor
22:52:58	Watusimoto	ah
22:53:08	Watusimoto	I do not have clientinfo for him yet because game just cycled
22:53:17	Watusimoto	so this is a problem
22:53:18	sam686	maybe, thats because, the client removes all client info, cause the level is changing, and haven't received new client infos
22:53:23	Watusimoto	exactly
22:53:46	sam686	and was trying to setisidle in between clear client infos, and waiting to get new client infos
22:53:49	Watusimoto	so what happens is if I ignore this, then when i do get his clientinfo (CI) I won't know he's idle
22:54:57	Watusimoto	so we can 1) delay calling this; 2) cache name locally; 3) ignore and send idle status with new client connection or 4) panic
22:55:01	Watusimoto	2 is probably out
22:55:08	Watusimoto	maybe the solution is 3
22:56:22	Watusimoto	that's what I'll do
22:56:43	raptor	i've gotta go soon and won't be back for a few hours :(
22:56:46	Watusimoto	only costs 1 bit in s2caddclient
22:56:48	Watusimoto	bye
22:57:00	Watusimoto	hasta manana
22:57:02	raptor	1 bit
22:57:05	sam686	maybe it might help to move GameConnection, s2cSetIsIdle to GameType, s2cSetIsIdle, because that is where the RemoteClientInfo is (GameType, s2cAddClient)
22:57:44	sam686	or at least that is where RemoteClientInfo is created
22:58:08	sam686	or, both s2cAddClient and s2cSetIsIdle can both be in GameConnection, or both in GameType
22:59:11	sam686	that way, RPCGuaranteedOrdered will work within the same class
23:02:05	Watusimoto	sam686: do refPtrs need to be deleted?
23:02:28		raptor Quit ()
23:02:32	sam686	RefPtr auto deletes when no more reference to it..
23:02:38	Watusimoto	ok
23:03:23	sam686	while RefPtr does auto delete (kind of like boost::SharedPtr), SafePtr does not auto delete...
23:03:36	Watusimoto	right -- I always get confused
23:03:42	Watusimoto	safeptr goes to null when deleted
23:04:17	sam686	refptr cannot be deleted (will get errors), you just need to let it go away, and it will auto delete
23:04:54	Watusimoto	sam -- you can move the rpcs after I do my next checkin
23:05:02	sam686	ok
23:14:45	Watusimoto	compiling now
23:17:50	Watusimoto	code is in
23:20:02	sam686	let me try my changes first (moving s2cSetIsIdle to GameType) without your changes and see if that alone fixes the problem...
23:22:29		BFLogBot - Commit 019888004884 \| Author: watusim...@bitfighter.org \| Log: Fixes for assert with spawn delayed players after level cycles
23:23:33	Watusimoto	ok
23:24:21	Watusimoto	but
23:24:28	Watusimoto	actually we might need them regardless
23:24:39	Watusimoto	if player is spawn delayed, and game is suspended
23:24:42	Watusimoto	and a new player joing
23:24:47	Watusimoto	game will unsuspend
23:24:56	Watusimoto	but new player needs to know that existing player is delayed
23:25:06	Watusimoto	s2csetidle won't be called
23:26:16	sam686	oh, i guess thats another problem, i see, new clients join won't know if existing player is idling...
23:29:58	Watusimoto	now I found another problem -- after waking up, clients think they themselves are still asleep on the scoreboard!
23:30:54	Watusimoto	should be an easy fix
23:34:58	sam686	sort of found a new problem, the buzy "<< >>" stays there even though it is actually not on any menus or anything...
23:35:19		FlynnnNT has joined
23:36:44	Watusimoto	fixed
23:36:53	Watusimoto	my problem, not yours
23:36:57	Watusimoto	really
23:41:06	sam686	no wonder why client don't recieve it when moved s2cSetIsIdle to gameType, it is not ghosted yet on client, there is GameType::onGhostAvailable()
23:43:27	Watusimoto	well... tracing everything through... the only place the << >> gets turned off is when cancelChat() is called on client
23:43:46	Watusimoto	so you should be able to turn off << >> by entering and exiting chat
23:44:04	Watusimoto	if so, the problem is that other places don't disable busy signal
23:44:14	Watusimoto	we turn it on in a bunch of places
23:45:11	Watusimoto	aha!
23:45:18	Watusimoto	a major flaw in the system!
23:45:38	Watusimoto	busy status only gets updated when the ship moves
23:45:47	Watusimoto	I never liked sending it via ship object
23:45:50	Watusimoto	will fix that now
23:55:59	sam686	i am about to push a change that moves s2cAddClient to GameType and away from gameConnection, the advantags is it makes FullClientInfo::setSpawnDelayed go down to a simple single line, instead of looping through all client infos
23:58:17	sam686	i mean, moves s2cSetIsIdle to gametype and away from gameConnection