#bitfighter IRC Log

Index Search ←Prev date Next date→

IRC Log for 2012-08-06

Timestamps are in GMT/BST.

00:00:50raptorgood read: http://gafferongames.com/networking-for-game-programmers/what-every-programmer-needs-to-know-about-game-networking/
00:01:32raptorwe could actually improve our network latency problems with some ideas in there..
00:01:49raptorunless TNL already does it all....
00:02:01raptorand i'm just unfamiliar with it..
00:26:13Watusimotowill look tomorrow
00:26:29WatusimotoI read a lot of that stuff when I was just starting out
00:28:26Watusimototnl does no latency handling; that's all done by zap code
00:29:13Watusimotook, pushed some code, going to bed
00:29:15Watusimotolater!
00:31:09BFLogBot - Commit 09ceba299154 | Author: watusim...@bitfighter.org | Log: Whitespace
00:31:11BFLogBot - Commit 1b06ff2d60a3 | Author: watusim...@bitfighter.org | Log: Comments
00:31:12BFLogBot - Commit 977962ec92aa | Author: watusim...@bitfighter.org | Log: Comment
00:31:14BFLogBot - Commit 845b91007d39 | Author: watusim...@bitfighter.org | Log: Left things a little messy, but HS now only steers towards targets in a cone it's "facing", changed projectile dynamics. Will clean up a little tomorrow, but want to give folks a chance to try this.
00:31:15BFLogBot - Commit 4fdc0bcd1b33 | Author: watusim...@bitfighter.org | Log: Merge
00:34:46Watusimoto Quit (Ping timeout: 246 seconds)
00:56:19Little_Apple has joined
00:56:23Little_Applehellooo
00:56:28raptorhi
00:56:35Little_Applewazzap
00:56:37raptorsam686: have you tried watusimoto's new heat seeker changes?
00:56:50sam686hi
00:57:17raptorheat seeker is constant speed now, and much much faster...
00:57:23Little_AppleHEAT SEEEEEKKKERRRRRRR
00:58:25Little_AppleCAN I TEST TO FIND ANY BUGS FOR MAC????
00:59:06raptorLittle_Apple: not much has changed since last build I gave you (I think)
00:59:16Little_AppleBUT… HEAT SEEKER!!
00:59:18raptordid you have heat seeker in your last build?
00:59:40Little_Appleno...
00:59:43raptoroh..
00:59:45Little_Apple._.
00:59:49raptorok, i'll give you a build to test
00:59:57Little_Apple:D
00:59:57sam686you need to select loadout to see heat seeker weapon..
01:00:11Little_Appleyep
01:00:15sam686and its only for 018, not 017
01:00:37Little_Appleyep
01:00:37raptorthat's right - you last tested teleporter for me
01:00:43Little_Appleyea :P
01:02:03raptorwell... heat seeker is being completely changed now
01:02:12raptori'm not sure i like the new changes sam686
01:02:14Little_Appleheh
01:02:31sam686still compiling, i haven't seen the latest changes
01:02:48raptorthere is a lot less 'seeking' to it...
01:03:11Little_Appleheh
01:08:13Little_Appleerr...
01:10:11Little_Appleerrrrrrrrrrrrrrrrrrrrrrr...
01:10:25raptorbuilding now...
01:10:46Little_Applewoo
01:20:56Little_Applehelloo?
01:21:01raptorhi
01:21:05raptoroops, forgot...
01:22:14raptorok Little_Apple: http://sam686.maxhushahn.com/upload/Bitfighter-018-beta3.zip
01:22:28Little_Appleyay!
01:22:33raptorsorry, i got distracted asking a question on a forum...
01:22:40Little_Applelol :P
01:24:10raptoryou still around sam686?
01:24:51Little_Applei think the heat seaker should be a little slower...
01:25:15Little_Appleright now it just zooms by someone and bends a little torwards them
01:25:28raptorit's was waaay different before
01:25:35sam686I think heat seeker goes too fast, its more like a super-fast bullet instead..
01:26:21BFLogBot - Commit 097cb5eb5e60 | Author: sam8641 | Log: Fix a runtime error, and fix blank game list on lobby list for quick rejoin.
01:38:32Little_Apple Quit (Quit: Page closed)
02:30:26Little_Apple has joined
02:34:13Little_Apple Quit (Client Quit)
03:57:05raptor Quit (Ping timeout: 264 seconds)
04:42:52raptor has joined
04:42:52ChanServ sets mode +o raptor
05:18:04raptoroh boy oh boy
05:18:09raptormars rover going to land!
06:09:10raptorit landed! (about 30 min. ago...)
06:09:14raptoryay!
06:12:48raptor Quit ()
06:54:42sam686 Quit (Ping timeout: 245 seconds)
07:08:07watusimoto has joined
07:08:08ChanServ sets mode +o watusimoto
07:26:59watusimoto1 has joined
07:28:02watusimoto Quit (Ping timeout: 272 seconds)
09:55:29watusimoto1 Quit (Ping timeout: 244 seconds)
10:07:15watusimoto has joined
10:07:16ChanServ sets mode +o watusimoto
12:05:14Watusimoto_ has joined
12:24:28-jtrucks- [GLOBAL NOTICE] The human eye is a wonderful device. With a little effort, it can fail to see even the most glaring injustice.
14:19:36Watusimoto_ Quit (Ping timeout: 240 seconds)
14:53:07watusimoto Quit (*.net *.split)
14:55:43watusimoto has joined
15:19:06watusimoto Quit (Ping timeout: 240 seconds)
15:22:18raptor has joined
15:22:18ChanServ sets mode +o raptor
15:25:28watusimoto has joined
15:25:28ChanServ sets mode +o watusimoto
15:25:46raptorgood morning!
15:25:50watusimotohi
15:25:55raptorsoo...
15:26:00watusimotook, so I read that article you linked to last night
15:26:05watusimotoand can answer any questions
15:26:42raptori understand all of it except whether or not we do the client-side prediction with a circular buffer like they mentioned
15:27:14watusimotoI don't think we do
15:27:14raptori know we do prediction in many areas
15:27:31watusimotowe run a parallel simulation of events on client and server
15:27:36raptorbut not sure if we use the trick about rebasing the object state on the returned server state
15:27:39watusimotoand update the cleint when we et new info
15:28:07watusimotoand when there is disagreement we don't rewind and resimulate (though there are hints here and there that suggest that might be supported)
15:28:24watusimotoinstead we slowly nudge the two simulations closer together
15:28:32raptoris that what 'ActualState' and 'RenderState' stuff is for?
15:28:40watusimotoactualstate is what the server says
15:28:43raptorbecause i was never quite clear on it...
15:28:50watusimotorenderstate is what the client is showing
15:28:55watusimotothere is no renderstate on server
15:28:59raptoryes
15:29:06raptori mean, i understand
15:29:30raptorbut does our client-side actually do recalculations between the two?
15:29:32watusimotoon client, if those states are different, that means there is a problem that is rectified by adding apparent movement to bring them back together
15:29:35watusimotoyes
15:29:47watusimotowhen a ctf flag is returned to base, the server simply moves it
15:30:13watusimotoit looks like it is flying across the screen because the client is trying to bring the renderpos back to the actualpos
15:30:27raptorha
15:30:28watusimotoit does it by moving the flag rather than just having it jump
15:30:54watusimotowith the falg it doesn't make sense, but with a ship, it would look lame if ships kept jumping here and there
15:30:59raptorinteresting... i thought that was for the coolness factor
15:31:00watusimotoinstead they look like they are flying
15:31:08watusimotono, I think it is actually a bug
15:31:15raptorha! ok
15:31:21raptoreverything seems clearer now...
15:31:24raptor:)
15:31:26watusimotothe server *should* send a bit suggesting the flag has warped
15:31:40watusimotoinstead it looks to the client like it just moved
15:32:08watusimotobut it does look kind of cool, but can be confusing if you pick up the flag when it looks like it is flying half a map away
15:32:40raptorok, that means for heat-seeker - i don't think i coded client-side prediction, because i return early in the idle() loop if it isn't the server...
15:32:44watusimotoso we use a much simpler algo for rectifying divergent states than rewinding and replyaing
15:33:01watusimotobullets are handled differently
15:33:11watusimotothey are all server side, with no client side prediction
15:33:14watusimotothat is
15:33:28watusimotowhen you fire, we don't forecast a new bullet and create it
15:33:35watusimotowe wait for a message from the server to create the bullet
15:33:53watusimotowhich is why when you are lagging, firing seems delayed
15:33:58watusimotobut mvememt does not
15:34:11watusimotobut movement does seem wierd -- when you stop, you move backwards a little
15:34:35watusimotosometime try setting a high simulated lag and you can see how the algos manage divergent states
15:35:09watusimotothat whole mess with the soccer ball was an attempt to reduce lag by doing more client side prediction for picking the ball up, if you remember that
15:35:20watusimotosam undid most of the code I wrote for that
15:35:27raptori remember the mess, yes..
15:35:28watusimotobecause it didn't quite work
15:35:36watusimotoand now pickup is goine anyway
15:35:43raptorbut i never understood what was going on...
15:36:19watusimotoI was trying to let the client predict you were picking up the ball rather than waiting for the server to tell the client it had picked up the ball
15:36:40watusimotobasically, lag sucks
15:36:44raptorheh
15:36:46raptoryep
15:36:52raptorok, thanks for clarifying
15:36:54watusimotodoes that help?
15:36:56raptoryes
15:37:09raptorso, now... on to heat seeker
15:37:15watusimotoI read lots of artilces on that site while starting on bg
15:37:17watusimotobf
15:37:22watusimotoyes, heat seeker
15:37:31raptorhonestly, i think your idea is way different than what i had in mind...
15:37:39raptorthere is a lot less 'seeking' involved :)
15:37:41watusimotoin what sense
15:37:46watusimotoah
15:37:53raptoralso, it goes waay fast
15:37:54watusimotowell, what's there is not what I think it shoul dbe
15:37:59raptorok
15:38:11watusimotothe major thing I did that I do like is that it only seeks objects that are in front of it
15:38:18watusimotorather than, say, behind
15:38:35watusimotowhich seems more realistic somehow
15:38:43raptorafter testing with sam686 and Little_Apple, it was mentioned that it is little different than a mildly more-on-target bullet
15:38:56raptorif you get lucky
15:38:58watusimotoyes, the seeking effect is weak
15:39:04watusimotoand not terribly useful
15:39:14watusimotomaybe slowing it will help
15:39:23watusimotomaybe letting it turn more sharply will help
15:39:39watusimotomaybe increasing the "cone" of detection will help
15:39:53raptorwhat did you think of my idea to have it increase in speed the straighter it was going?
15:40:12watusimotoI didn't like it -- it made me feel there was a problem with my connection
15:40:16watusimotowhich there was :-)
15:40:21raptorha
15:40:51watusimotobecause all other projectiles are constant speed, this just felt wrong
15:40:52watusimotobut your code is still there, only commented out
15:41:11raptorok
15:41:21watusimotobut if you are suggesting my mods made it worse, I might agree
15:41:25raptorideas for graphics?
15:41:47watusimotomy kids want a elongated purple triangle with white flames
15:42:04watusimotobut that's the least of the issues at the moment
15:42:21raptorheh
15:42:34watusimotoit might be a cool effect to show some sort of seeking or locking action
15:42:43watusimotolike a light line btwn missle and target
15:42:56watusimotoor sonar like effect
15:43:09watusimotosomething to instill fear
15:43:44raptori'd say the cone of detection should be pi or close to pi in front
15:43:49watusimotoso give me another go-round with it tonight, and let's see if I can bring it back into line with what you are thinking
15:43:52raptortau /2
15:43:56watusimotopi is 90deg
15:44:01watusimotono 180 deg
15:44:04watusimotoyes, maybe
15:44:11watusimotothat's what I started with, and made a math error
15:44:15watusimotowhich is why it is what it is
15:44:38watusimotoI checked in at 2:15 -- I just wanted to get the code in that wasn't broken
15:44:42raptorhmmm
15:45:09watusimoto180 deg might be good
15:45:18raptorthat normalizeAngle function should return between -tau/2 and tau/2
15:45:39watusimotoI though it only returned positive, which was the source of my error
15:45:50raptormaybe i wrote it wrong...
15:45:54watusimotoyou didn't
15:46:03watusimotoit was my mistake
15:46:10raptoroh i get it now
15:47:49watusimotoso I think I can fix it pretty easily and make it work more like we both seem to think it should
15:48:06watusimotobut get rid of some of what I saw as abberant behavior with the first try
15:48:09raptorok, i'll let you tinker more before i get any more ideas...
15:48:14watusimoto:-)
15:48:47raptoralso... curiosity landed!
15:50:23watusimotoI saw
15:50:24watusimotohttp://www.youtube.com/watch?v=Le2cWsISOE4
15:50:27watusimoto0:36
15:50:32watusimotoshows a heat seeker
15:50:35watusimotoit goes slow
15:51:03watusimotosorry 0:43
15:51:21raptorinteresting - looks like constant speed
15:51:28raptorok, i capitulate
15:51:37watusimotojust looking for ideas
15:52:42watusimotohere's one that changes speed
15:52:47watusimoto:0:50ish
15:53:11watusimoto0:47, watch in the right hand side
15:54:17watusimotoI call that one "when heat seekers get silly"
15:55:06raptori'm not sure i'm finding what you're looking at
15:55:45watusimotohttp://www.youtube.com/watch?v=rcU689Pvt4w
15:55:46watusimotosorry
15:55:57watusimotostupid linux and their crappy clipboard
15:56:35raptorhaha
15:56:45raptorthat's exactly what one of my first iterations did in game
15:56:54watusimotoI really don't like the implementation, but there are some good ideas
15:57:13watusimotothey should go faster, not turn 180 degrees, but I like the idea of some sort of tracer effect
15:57:56watusimotowas also thinking we could add a proximity fuse; don't need a direct hit to explode and damage
15:58:50raptori did the proximity fuse in my first iteration
15:59:04raptordid work so well when it got going really fast...
15:59:07raptor*didn't
16:00:17raptorthat was one of the questions i wanted to ask you - whether you thought it should a point damage weapon, or area damage
16:02:02watusimotonot sure yet
16:02:12watusimotoif it works well enough, point would be better
16:02:14watusimotoI think
16:02:36raptoralso, another question was: highenergy/high-yield, or lower energy, lower yield?
16:02:51raptoryield = damage
16:03:59raptorhey, my question got 5 votes already!: http://stackoverflow.com/questions/11821723/keeping-code-in-a-client-server-game-organized
16:04:12raptori should use a different username... that is my work user
16:05:25watusimotojust upvoted you
16:05:32raptordid i make sense?
16:05:40watusimotohaven't read it yet :-)
16:05:48raptorha! shame shame
16:06:10raptori don't hink i understand the highest rated answer...
16:06:22watusimotowell written
16:07:41watusimotoI don't like either answer
16:08:09raptori purposefully didn't put in any ideas i had so i could get answers clear of bias
16:09:01raptorthat bytemaster guy wrote the mace-rpc library - and i still don't understand its purpose
16:10:05watusimotoforgot parens here:
16:10:06watusimotovoid Ship::renderExplosion
16:10:17watusimotoin .cpp and .h
16:10:17raptoroops
16:11:23raptorfixed, thanks
16:11:32raptorand that's why you don't write questions late at night...
16:52:12watusimotook, back later
16:58:16watusimoto Quit (Ping timeout: 246 seconds)
17:20:35Watusimoto has joined
17:28:43Watusimoto Quit (Ping timeout: 246 seconds)
18:12:01sam686 has joined
18:12:02ChanServ sets mode +v sam686
18:58:27sam686 Quit (Ping timeout: 245 seconds)
18:59:22raptor Quit ()
19:00:05sam686 has joined
19:00:05ChanServ sets mode +v sam686
19:03:46krummi1999 has joined
19:04:04krummi1999hello everybody
19:04:48krummi1999 Quit (Client Quit)
20:05:48Watusimoto has joined
20:39:44raptor has joined
20:39:44ChanServ sets mode +o raptor
20:39:49raptorhi
21:01:47sam686 Quit (Ping timeout: 245 seconds)
21:02:46sam686 has joined
21:02:46ChanServ sets mode +v sam686
21:06:50raptorbe back in a bit..
21:07:46sam686 has left
21:10:23sam686 has joined
21:10:23ChanServ sets mode +v sam686
21:25:13sam686 has left
21:29:38sam686 has joined
21:29:38ChanServ sets mode +v sam686
21:52:29raptor Quit ()
22:27:13sam686 Quit (Ping timeout: 245 seconds)
22:42:54sam686 has joined
22:42:54ChanServ sets mode +v sam686

Index Search ←Prev date Next date→

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