#bitfighter IRC Log

Index Search ←Prev date Next date→

IRC Log for 2011-10-19

Timestamps are in GMT/BST.

00:01:00raptoryay everything compiles!
00:03:00BFLogBot - Commit d4934b6afdba | Author: buckyballreaction | Log: Fix issue 74. Vertices are now treated as a normal move when dragged to the dock
00:08:00raptor47 issues left!
00:08:00raptorbeen a while since we
00:08:00raptorve been so low..
01:16:00koda Quit (Quit: k thx bai)
01:18:00raptor Quit (Remote host closed the connection)
01:31:00karamazovapylots of talky in here
02:29:00Flynnn Quit (Quit: Leaving)
03:40:00raptor has joined
03:40:00ChanServ sets mode +o raptor
03:42:00Kumul Quit (Quit: gone)
03:46:00raptorlively tonight
03:54:00raptori still can't believe how enlightening little_apple's video is...
04:24:00raptorsam686: karamazovapy: do we want this?: http://code.google.com/p/bitfighter/issues/detail?id=124
04:43:00raptor Quit (Remote host closed the connection)
05:14:00sam686 Quit (Ping timeout: 258 seconds)
05:14:00sam686 has joined
05:14:00ChanServ sets mode +v sam686
07:51:00kodax has joined
12:45:00kodax Quit (Remote host closed the connection)
13:35:00watusimoto has joined
13:43:00watusimoto Quit (Read error: Connection reset by peer)
15:08:00karamazovapythe Playa zapped ChumpChange! message separation is kind of interesting...basically, it's giving kills their own message feed directly above other game messages
15:08:00karamazovapyI think they also have a slightly longer decay time versus other game messages
15:09:00karamazovapymight make it easier to read what's going on. I don't have much of an opinion, though
15:09:00raptor has joined
15:09:00ChanServ sets mode +o raptor
15:11:00raptordo you think the killed messages should be separated somehow? Or would a combination like Zoomber's issue suggests be enough
15:11:00karamazovapyyou mean raptor zapped 1, 2, 3, 4! ?
15:12:00raptoryes
15:12:00raptor(if this is what you're talking about...)
15:14:00karamazovapywell it's adding to each line instead of having individual lines that makes things more readable
15:14:00raptorwas that part of the original zap! demo?
15:15:00karamazovapyit was in at least one of the versions of Zap!...zoomber and I noticed it when we were playing around
15:21:00raptori guess i just wasn't sure that it was an 'approved' feature request - i don't remember talking about it
15:22:00karamazovapyit's not an "approved" feature request
15:22:00karamazovapyI think some people are fuzzy on that rule
15:24:00raptorok, so, due to the ambiguity around the issue - I'm going to just pretend I didn't see it for now...
15:32:00karamazovapyI don't think it would be hard or particularly controversial to give it a shot, but it also isn't high priority
15:41:00raptor"small but friendly community"
15:41:00raptorhehe
15:45:00watusimoto has joined
15:45:00watusimotohello
15:45:00raptorgood day
15:48:00raptortesting issue 96
15:49:00raptorlooks like it's fixed...
15:52:00watusimotogreat! what was it?
15:52:00raptoreditor undo glitch
15:52:00watusimotofixed the random green text
15:52:00raptorhooray!
15:52:00raptorwhat was it?
15:52:00watusimotostring -> const char * confusion
15:52:00watusimotojust as I predicted
15:53:00raptoryeah
15:53:00raptorso
15:53:00raptori was meaning to ask about that
15:53:00raptori see lots of methods with 'const char *'
15:53:00raptorin their signature
15:53:00watusimotoyes
15:53:00raptorshould I switch them to string?
15:53:00raptorto be tidy/consistent?
15:54:00watusimotoHere's how I've been approaching it:
15:54:00watusimotosome methods inherently take a const char * at their bottom
15:54:00watusimotolike, say logprintf
15:54:00raptoryup
15:54:00raptorand that's where i force the .c_str()
15:55:00watusimotoif we are working with a method like that, and just passing a value around without working on it, I just create it as a const char * and leave it at that
15:55:00watusimotobut if it's something we manipulate, then I use string because it's safer
15:55:00watusimotoand, usually, easier
15:55:00raptorok
15:56:00watusimotobut I figure why create a string object with its (admittedly small) overhead when it doesn't make something easier
15:56:00raptormake sense
15:56:00watusimotobut when it does make something easier, then no reason not to use string
15:56:00watusimotowhen I first started on bf, there were no strings
15:56:00raptor\i was just thinking more along the lines of passing references & also
15:56:00watusimotoso every string was either ported from a const char * or is a new bit of code
15:57:00watusimotoi always pass a const string &ref when passing strings
15:57:00watusimotounless there is a good reason not to, and then I try to comment it
15:57:00raptoryes
15:57:00raptorok
15:58:00watusimotothere's only a couple of places we do that, such as when we deliberatley want to make a copy of the object, or when we're rturning a value in the string param, but those are very rare
15:58:00watusimotoso that's how I see it anyway
15:58:00watusimotocontext and all that :-)
15:59:00raptorok
16:00:00raptorso i've been just constantly fixing issues and minor bugs...
16:00:00raptor(and ask about issues)
16:00:00watusimotofun, eh?
16:00:00raptori started the object arrows one, but I got lost
16:00:00raptorobjective arrows
16:00:00watusimotoI'm working on bugs tonight, too
16:00:00watusimotoleave it and do something else, if you like
16:01:00raptorwell that's the thing - i've done that with 4 or 5 issues already :)
16:01:00raptorone i ant to complete is http://code.google.com/p/bitfighter/issues/detail?id=119
16:01:00raptorwant
16:02:00raptorI thought I already had... but I didnt' quite understand your response
16:05:00BFLogBot - Commit 0d7763af3227 | Author: watusim...@bitfighter.org | Log: Remove comments
16:05:00BFLogBot - Commit f61b761373f4 | Author: watusim...@bitfighter.org | Log: BUGIX: No more random green chars on the menus; also used simpler constructor where appropriate
16:05:00BFLogBot - Commit a7595038d009 | Author: watusim...@bitfighter.org | Log: Merge
16:06:00raptorah yes, null terminated character arrays
16:06:00raptoryay
16:06:00watusimotook
16:06:00watusimotomaybe you did fix this
16:07:00watusimotowhat's an example of a command you consider fixed by this case?
16:07:00raptorlooking
16:07:00raptori added one recently...
16:08:00raptoraach, where'd they go?
16:09:00raptori had them all inserted into that struct at one point... i thought
16:10:00watusimotowell, maybe you answered the question
16:10:00raptortake a look at this revision: http://code.google.com/p/bitfighter/source/browse/zap/UIGame.cpp?spec=svnb79ec30e0563e10b65029ce1fccde0b00d379490&name=b79ec30e0563&r=b79ec30e0563e10b65029ce1fccde0b00d379490
16:10:00watusimotoit is my opinion that they should not work without being added to the struct
16:11:00raptori agree with you
16:11:00watusimotothe struct should be the essence of the command
16:11:00raptorsearch for 'banip' on that page
16:11:00watusimotoand it was my impression that you added them to the struct, but that you could have removed them from the struct again without breaking the command
16:12:00raptori had added: ban, banip, gmute, resetscore, a couple others
16:12:00raptormust have been a abad merge somewhere...
16:12:00watusimotoprobably my fautl :-)
16:12:00watusimotobut let's use that as an example
16:12:00raptorok
16:13:00watusimotoif you added that to the struct without changing anything else, then in my mind it still breaks the design
16:13:00raptorso for instance, i added 'resetscore' (which survived teh merge)
16:13:00raptorit calles: GameUserInterface::serverCommandHandler
16:15:00watusimotoso maybe instead of putting the string in serverCommandHandler, maybe we shoudl create a separate function that gets called directly without having to search for the string "banip"
16:15:00watusimotoif that makes sense
16:16:00watusimotoit seems that serverCommandHandler somewhat bypasses the design for these commands
16:17:00watusimotothough now I'm wondering if I've confused myself
16:17:00watusimotobecause serverCommandHandler isn't hardly any code at all
16:18:00watusimotoThe command I don't like is processServerCommand
16:18:00watusimotobecause it's a huge mess of code doing all sorts of things
16:18:00watusimotobut the command struct doesn't really interact with that at all
16:19:00watusimotook, I got to make dinner for my grumpy kids
16:20:00watusimotoif what I wrote doesn't make sense, we can discuss it later
16:41:00raptorok, i'm back
16:41:00raptorpeoples have left my officle
16:42:00raptoryes, that makes perfect sense - i don't like 'processServerCommand' either
16:42:00raptorofficle = office + cubicle
17:05:00raptorsigh, opti
17:05:00BFLogBot - Commit c73c759b88ac | Author: buckyballreaction | Log: Restore commands to CommandInfo struct that were lost somehow
17:07:00raptorhttp://code.google.com/p/bitfighter/issues/detail?id=130 <- karamazovapy, watusimoto, sam686, what do you think about this? (Have I asked before?)
17:17:00watusimoto Quit (Ping timeout: 258 seconds)
17:28:00LordDVG has joined
17:44:00watusimoto has joined
17:45:00watusimotoI wonder if we canuse that same struct on the client and server sides??
17:45:00raptorhi
17:45:00watusimotoactually, that probably wouldn't work
17:46:00raptoryeah
17:46:00raptorso, from the client
17:46:00raptorGameUserInterface::serverCommandHandler -> c2sSendCommand -> processServerCommand
17:47:00raptorand processServerCommand handles a dozen or so in a big if statement
17:47:00sam686there need to be some kind of transfer between server and client, as in server send struct, client receive struct, similar to gameStatistics.
17:47:00raptorwatusimoto: you are thinking about creating a dozen or so new 'c2s' methods?
17:48:00watusimotothe way to fix it would be to create separate commands for each possibility in proccessServerCommand... but that would be a lot of commands.
17:48:00watusimotoback soon
17:48:00raptorI can do that
17:48:00watusimotonot sure we want to...
17:48:00watusimotoback in a sec
17:48:00raptorunless there is a performance problem somewhere...
17:48:00watusimotoin a minute
17:48:00raptorok
18:36:00watusimotoback
18:36:00raptorwelcome
18:36:00watusimotoso... do multiple methods make sense?
18:36:00raptormaybe
18:37:00raptorit seems that having only one c2s command is efficient, even though if server-side it is ugly
18:38:00watusimotoperhaps, but not necessarily
18:38:00watusimotonot sure how the methods work, but it may be as simple as sending an int to specify which c2sxxx is run
18:38:00watusimotoin which case running one would be as simple as running another
18:38:00watusimotoif that's the case, then multiple methods would be both clearer and more efficient
18:39:00watusimotoand would be easier to maintain
18:39:00raptorah - i don't really understand teh ?2? architecture, but the clearness is better
18:39:00sam686another problem is if a new, server side command is added during mid-life of a version as it happened to version 015
18:40:00sam686c2s is client to server
18:40:00raptoryes, i get that bit :)
18:40:00raptorjust not how it is ultimately wrapped
18:40:00raptori'm thinking about doing lots of commands: follow our current model
18:43:00raptori'll get that done today, probably
18:43:00raptorthen i can close another issue!
18:43:00[1]watusimoto has joined
18:43:00[1]watusimotowhy don't we leave the framework for processServerCommand in place, but move the existing commands out into new single-purpose commands
18:43:00watusimoto Quit (Ping timeout: 260 seconds)
18:43:00[1]watusimoto is now known as watusimoto
18:43:00watusimotodid that last msg get through?
18:44:00raptorabout leaving framework in place?
18:44:00raptorgot that
18:45:00watusimotoyes
18:45:00watusimotook; had a connection hiccup
18:45:00watusimotoanyone see a downside to that approach (aside from the work involved)?
18:47:00raptornot really
18:47:00sam686i guess there is mostly no downside...
18:50:00raptorand besides, we can break compatibility with anything we want!
18:50:00watusimotook, let's do it
18:53:00raptorwe can use the processCommandHandler for NOT breaking compatibility if needed...
18:55:00watusimotoexactly -- for mid-cycle commands
18:55:00watusimotosam, this was on your buglist:
18:55:00watusimoto2. SDL + Windows: While in menus, Pressing down arrow or up arrow will put mouse location back to the middle.
18:55:00watusimotodo you know if that's been fixed? I can't seem to reproduce it
18:56:00watusimotoit definitely is not a problem for me, either in window or in fullscreen
18:56:00raptorhow is fake fullscreen?
18:57:00watusimotomuch better
18:57:00raptordoes it cover your task bar?
18:57:00watusimotonot as slick as GLUT, but definitely does the trick
18:57:00watusimotoyes
18:57:00watusimotoit's good
18:57:00raptordefine 'slick as GLUT'
18:57:00sam686some of "fake fullscreen" problems makes some real fullscreen problems go away, including "While in menus, Pressing down arrow or up arrow will put mouse location back to the middle."
18:57:00watusimotoah, good!
18:58:00watusimotoGLUT was super fast. You can see SDL recomposing the screen. Not a problem, just feels a little slower
18:58:00watusimotobut I'm satisified
18:58:00raptorthat may be my fault, too
18:59:00raptori do a test run of the video hardware first to see if we can use hardware opengl
18:59:00watusimotoMaybe it resizes->redraws->recomposes?
18:59:00raptorthen i set the videomode
18:59:00raptoroh actually that is a different problem
18:59:00watusimotowhy do you do that test?
18:59:00raptori don't know what your porblem is
18:59:00watusimotoit's not really a problem
18:59:00raptorI start by initializing hardware SDL, if it fails, I use software
19:00:00watusimotointeresting
19:00:00watusimotowhat if you always used software?
19:00:00watusimotowell, who cares. it's great
19:01:00sam686i get a result with it completely ignoring that HWOPENGL and swOpengl flags on both windows and linux (ubuntu)
19:01:00sam686and it just does hardware opengl no matter what the game does...
19:01:00sam686or software if hardware is not available..
19:01:00raptoryea, i may not need my test, actually...
19:02:00raptorif so, it will get rid of the momentary blink upon start-up
19:03:00sam686it just says "Using hardware opengl" even when I have hardware opengl disabled...
19:04:00raptoryeah, SDL may do it automatically already
19:04:00raptormore research needed...
19:05:00raptorfor some reason on Mac, GLUT always used software, i think
19:05:00raptorso SDL with hardware is really giving mac users a performance boost
19:06:00sam686i don't know, if you only used virtual machine for mac, it might not do hardware opengl, where real mac might have been doing hardware opengl, i guess...
19:06:00sam686otherwise, all mac users will complain the game runs too slow..
19:07:00raptori wasn't testing on VM - i was relating zoomber's experiences
19:08:00sam686what did zoomber say?
19:09:00raptoron his Mac machines he says the framerate for 016 builds can get up to a 1000 - with 015a he could only get to 60 or 90, I think
19:09:00sam686i thought zoomber can get up to 1000 fps in version 015..
19:10:00sam686only when window size is small, i think..
19:10:00raptorinteresting - maybe we'll have to ask him and get facts sorted out
19:10:00sam686must be more like the vsync problem
19:11:00raptoryeah, could be
19:11:00raptoranyways, it's bugfixin' time!
19:26:00LordDVG Quit (Ping timeout: 240 seconds)
19:32:00LordDVG has joined
19:40:00LordDVG Quit (Ping timeout: 260 seconds)
19:44:00raptorwatusimoto: did i ever get your opinion on: http://code.google.com/p/bitfighter/issues/detail?id=130
19:56:00watusimotono
19:56:00watusimotoI question the utility myself
19:56:00watusimotorestart attempts to add players in order of skill, making theoretically even teams
19:56:00watusimotoexcept our measure of skill is bogus
19:56:00raptor:)
19:57:00watusimotoscramble would just be random
20:00:00watusimotoI'd like to hear from sam and k on the utility of such a function
20:00:00watusimotothey have more experience than I on these matters
20:01:00raptorme too
20:01:00raptorkaramazovapy? sam686?
20:01:00karamazovapyraptor? raptor?
20:01:00raptorread ^^
20:02:00raptorabout /scramble
20:02:00karamazovapyI think this was prompted by the times we've had a long series of games where teams didn't really change
20:03:00karamazovapyyou could make it random - or - you could have it swap every other player
20:03:00raptoryou mean: /pseudoscramble
20:03:00raptor:)
20:03:00karamazovapylike riffle shuffling a deck of cards
20:17:00sam686what will /shuffle_teams do, make team change completely random, or based on rating?
20:17:00watusimotoif we do it, it will be random
20:18:00watusimotowe already have ratings based, and it's not so hot
20:18:00karamazovapyI vote riffle shuffle
20:19:00sam686rating system don't seem to be reliable with more then 4 players, all it does is keep the 2 highest reting from being together in a team..
20:19:00karamazovapyswap 3&4, swap 7&8
20:19:00raptoris the 'proper' solution to fix the rating system? (e.g. will we ever actually need /scramble if rating system was really good?)
20:21:00karamazovapyI think kills/death is a reasonably good metric, although we might have to adjust so suicides don't count
20:25:00watusimotodeaths/kills is easy to calculate
20:25:00karamazovapyas a general rule, I think better players get killed less
20:25:00watusimotoI'd be happy to replace our existing metric with that
20:26:00watusimotowe currently have points scored by player : total points scored by everyone
20:26:00karamazovapyif you make that move, like I said, I'd remove suicides from the equation
20:29:00sam686|2 has joined
20:29:00sam686 Quit (Disconnected by services)
20:29:00sam686|2 is now known as sam686
20:29:00ChanServ sets mode +v sam686
20:33:00watusimotoI concur
20:54:00watusimotoso let's hold off on a shuffle command, but try the kill/death ratio as a replacement for the current rating scheme
20:56:00raptori say yes
20:56:00raptorand don't count suiceides
20:56:00raptorand learn to spell..
21:13:00watusimotocorrect
21:14:00karamazovapyanybody want to give opinion(s) on my logo?
21:15:00raptorwhere can i find this?
21:15:00karamazovapyokay - I'm working to open a music venue. the local area code is 517
21:15:00karamazovapyhttp://206.217.217.165
21:15:00karamazovapyblack and white is sort of the base logo I'm going with
21:16:00karamazovapythe design studio suggested a bunch of color schemes (I think looking forward to future web and print work)
21:16:00karamazovapyI wasn't wild about them and made a couple mock ups that are more what I had in mind
21:16:00raptori'm trying to see old nintendo mario characters in that logo...
21:17:00raptori think both mid-century ones are my least favorites
21:19:00raptorsophisticatic looks too... sophisticated, or business-like; but, I don't know what you are going for
21:20:00karamazovapywe'll be hosting all genres of music, trying to appeal to the year-round crowd early in the week (and over the summer) and college kids the rest of the time
21:20:00karamazovapyconsidering Michigan State University students our target demographic
21:21:00raptori kind of like energetic
21:21:00karamazovapyI'm assuming you were able to pick out the "hidden" 517 in the EQ on my version 2
21:22:00raptori see it now! :)
21:22:00karamazovapyit's supposed to be one of the things where only people who "get it" really see it. sort of an inside-code-type-thing
21:23:00sam686|2 has joined
21:23:00sam686 Quit (Disconnected by services)
21:23:00sam686|2 is now known as sam686
21:23:00ChanServ sets mode +v sam686
21:24:00karamazovapyI just showed these to a college girl and she thought they all looked too "starbucks/furniture store"
21:24:00karamazovapywell...the first four
21:24:00raptorthat's the thing with logos - few actually look good after a few years
21:25:00raptorbrb
21:25:00karamazovapyI think the logo itself is solid...but I like the black and white better than any of the studio's color schemes
21:26:00BFLogBot - Commit fc6931cd84b3 | Author: watusim...@bitfighter.org | Log: BUGFIX: Fix joining wall segments in editor
21:26:00BFLogBot - Commit 6bdba202769c | Author: watusim...@bitfighter.org | Log: Merge
21:36:00koda has joined
21:47:00raptoryeah, the original black and white seems the best of the studio options...
21:51:00BFLogBot - Commit 252c0911b500 | Author: watusim...@bitfighter.org | Log: BIGFIX: Can now add lines again in editor; fixed joining line segments
21:52:00raptor Quit (Remote host closed the connection)
21:53:00koda Quit (Quit: koda)
21:53:00watusimotoI like the blue ones; the last two
21:54:00watusimotobut they do look corporate
21:54:00watusimotoor rather like a super upscale logo
21:54:00watusimotolike a $12 cocktail bar
21:54:00watusimotofull of ex-frat boys
21:54:00koda has joined
21:57:00watusimotoWhat about a graffitti style font -- create your logo here
21:57:00watusimotohttp://www.graffiticreator.net/
22:14:00watusimotowhat kind of music will the venue have?
22:21:00BFLogBot - Commit 38753ef42370 | Author: sam8641 | Log: Fix ZAP_DEDICATED, added some "Not a TeamEditor" checks in debug mode..
22:22:00sam686there is an editor problem with team renaming after testing from editor..
23:28:00raptor has joined
23:28:00ChanServ sets mode +o raptor
23:31:00raptoryay for bugfixing
23:39:00Flynnn has joined
23:46:00koda Quit (Quit: k thx bai)
23:55:00Zoomber has joined
23:55:00ChanServ sets mode +v Zoomber

Index Search ←Prev date Next date→

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