Timestamps are in GMT/BST.
| 00:08:19 | raptor | be back later.. |
| 00:08:23 | | raptor Quit () |
| 01:47:07 | | koda Quit (Quit: I used to be chatting like you. Then I took an arrow in the knee) |
| 02:46:40 | | raptor has joined |
| 02:46:41 | | ChanServ sets mode +o raptor |
| 03:47:27 | | zoomber_mbp has joined |
| 04:02:25 | raptor | sam686: are we pretty my guaranteed to have 32bit color for anyone playing bitfighter? |
| 04:07:03 | zoomber_mbp | hihi |
| 04:07:29 | raptor | hi |
| 04:15:17 | sam686 | i think so, but I can switch my windows xp graphics to 16 bit color (5 bit red, 6 bit green, 5 bit blue) |
| 04:17:04 | sam686 | i often wonder why 32 bit if there is 8 bit red, 8 bit green, 8 bit blue, and there is 8 bit unused, it seems kind of a waste of memory... |
| 04:17:12 | raptor | 8 bit alpha |
| 04:17:28 | raptor | but for color only, it is really only 24 bit |
| 04:17:45 | raptor | 32bit color always = 24bit color + 8bit alpha |
| 04:18:10 | sam686 | what does the extra 8 bit alpha do in windows xp? |
| 04:19:05 | sam686 | almost, nothing, as i said, that 8 bit alpha is mostly unused |
| 04:19:18 | raptor | probably nothing.. |
| 04:19:42 | sam686 | but some parts of alpha is used in 3D rendering (in texture for transparency) |
| 04:20:31 | sam686 | even if the display graphics settings is 16 bit, there can still be 32 bit texture for 3D, rendering to 16 bit display graphics in that case.. |
| 04:25:48 | sam686 | http://sam686.maxhushahn.com/upload/16_bit_graphics_editor.png if you look closely, dithering of my Radeon 9200 for 16 bit graphics tries to hide the effects of lower number of colors |
| 04:27:00 | sam686 | my radeon 9200 have an option to use error difusion (as the screenshot above), or ordered dithering |
| 04:27:10 | raptor | very interesting. |
| 04:28:45 | sam686 | also, 16 bit graphics appears to run about 1.5 times faster then 32 bit graphics for my Radeon 9200.. |
| 04:30:23 | sam686 | but, it hardly makes a difference on bitfighter (except on ultra high resolutions), due to hardware graphics being so fast.. |
| 04:31:49 | raptor | the reason i've been asking about this, is because i need to solve the mac fullscreen problems |
| 04:31:51 | sam686 | oh, and my radeon 9200 error difusion is random for every next frame, farther hiding the 16 bit limited colors.. |
| 04:31:59 | raptor | neat |
| 04:33:04 | sam686 | i don't know about how mac does it... as i don't have mac.. |
| 04:34:25 | sam686 | obviously, only some hardware graphics have such a good 16 bit error diffusion dithering, most hardware, including my intel mobile graphics 945 express use a simple dithering.., |
| 04:40:11 | raptor | can you test something for me? |
| 04:40:51 | sam686 | test what? |
| 04:41:02 | raptor | main.cpp:860 |
| 04:41:12 | raptor | change all '5' to 8 |
| 04:41:22 | raptor | and change 16 to 32 |
| 04:41:34 | raptor | compile, then try to start in 16bit color |
| 04:44:34 | sam686 | it starts fine, and it still uses 16-bit color (i see it with ctrl + Q, looking at screenshot), with 5 bit red, 6 bit green, 5 bit blue... |
| 04:44:44 | raptor | ok |
| 04:44:47 | raptor | thanks! |
| 04:45:12 | sam686 | almost like all that SDL_GL_RED_SIZE and similar are all being ignored for my windows computer.. |
| 04:52:41 | sam686 | heres the fun part, starting bitfighter with 16 bit display graphics, switching graphics to 32 bit, and it keeps rendering bitfighter as 16 bit... |
| 04:53:22 | raptor | ha! |
| 04:53:36 | raptor | i remember windows 95 doing that.. |
| 04:55:45 | sam686 | ok, after some more testing, SDL_GL_DEPTH_SIZE = 32 goes very very slow frame rate when desktop is at 32 bit graphics |
| 04:56:10 | raptor | really? |
| 04:56:13 | raptor | hmm... |
| 04:56:14 | sam686 | SDL_GL_DEPTH_SIZE = 16 is fine, i think the SDL_GL_DEPTH_SIZE is something to do with z-depth buffering |
| 04:56:47 | sam686 | and, SDL_GL_RED_SIZE / green / blue appears to be completely ignored for my Radeon 9200.. |
| 04:57:13 | raptor | yeah - it's doing the same on my mac vm |
| 05:14:39 | zoomber_mbp | ahh good morrow |
| 05:28:07 | sam686 | SDL_GL_DEPTH_SIZE = 24 works just as fast for my computer, but i guess it is safe to use SDL_GL_DEPTH_SIZE = 16 |
| 05:28:32 | raptor | ok, i wont touch the settings then |
| 05:28:42 | raptor | none of my changes have been working with mac anyways.. |
| 05:28:55 | sam686 | I put in logprintf("%s, %s", glGetString(GL_RENDERER), glGetString(GL_VENDOR)); right after SDL_SetVideoMode |
| 05:29:22 | sam686 | shows GDI Generic, Microsoft Corporation when SDL_GL_DEPTH_SIZE = 32 |
| 05:29:41 | sam686 | RADEON 9250/9200 Series DDR x86/SSE2, ATI Technologies Inc. ... when SDL_GL_DEPTH_SIZE = 24 or 16 |
| 05:29:44 | raptor | ah. so it must switch to software? |
| 05:29:49 | raptor | that is bad |
| 05:29:50 | sam686 | probably yes |
| 05:31:08 | sam686 | ok... SDL_GL_DEPTH_SIZE = 0 works too, i might be guessing that means, let the hardware pick the best depth |
| 05:31:43 | sam686 | SDL_GL_DEPTH_SIZE does make a difference on how accurate the Z depth is, one place is used in bitfighter is pressing TAB in editor with multiple spy bugs |
| 05:31:45 | raptor | i think if SDL detects hardware for opengl, it just lets hardware do most of the work.. |
| 05:36:49 | | sam686 Quit (Ping timeout: 245 seconds) |
| 05:40:21 | | sam686 has joined |
| 05:40:21 | | ChanServ sets mode +v sam686 |
| 05:40:59 | zoomber_mbp | hey raptor |
| 05:41:07 | raptor | hi |
| 05:41:12 | zoomber_mbp | i'm back in the irc buisness |
| 05:41:14 | zoomber_mbp | with a friend |
| 05:41:15 | zoomber_mbp | this time |
| 05:41:19 | raptor | ? |
| 05:41:21 | raptor | welcome back? |
| 05:41:40 | zoomber_mbp | mono, i mean, we're making an irc network |
| 05:42:12 | zoomber_mbp | i think he's reselling VPS servers at the moment too, so the personal irc network will help him as well |
| 05:43:46 | raptor | makeing one? or adding a server to an established one? |
| 05:44:05 | zoomber_mbp | sort of kind of making one, but we've already got it going |
| 05:44:11 | zoomber_mbp | not established socially, if thats what you mean |
| 05:50:43 | raptor | ok, fixed a stupid bug on Mac |
| 05:50:53 | raptor | window was going black when resizing |
| 05:51:28 | | BFLogBot - Commit 72c5f55362a9 | Author: buckyballreaction | Log: Add missing class to Xcode project |
| 05:51:29 | | BFLogBot - Commit b33bc4bb178b | Author: buckyballreaction | Log: Add patch to restore OpenGL context saving on Mac, and recompile the framework |
| 05:54:23 | zoomber_mbp | did you fix mac's full screen changer bug? |
| 05:55:32 | raptor | i hate that bug |
| 05:55:37 | raptor | no |
| 05:55:46 | raptor | but there is a workaround |
| 05:56:17 | raptor | which i coded to be default - use fake fullscreen |
| 05:56:25 | raptor | you just have to hide the dock |
| 05:59:00 | sam686 | i see a problem, entering "High score" menu seems to freeze (become unresponsive) for one second... |
| 05:59:27 | sam686 | i mean, freeze the master for one second.. |
| 06:00:29 | sam686 | so if you repeatedly enter "high score" menu, the master will appear not responsive for a few seconds (about one second for every time the high score menu is opened) |
| 06:00:29 | raptor | it takes me about 3 seconds to load the data from master |
| 06:00:47 | raptor | oh, freeze the master... |
| 06:00:53 | raptor | that's not good |
| 06:01:05 | raptor | maybe the master needs to use a thread? |
| 06:01:18 | raptor | while it's waiting to retrieve the database info? |
| 06:01:20 | sam686 | now i made the master freeze, i think... |
| 06:01:32 | sam686 | or now it unfreeze.. |
| 06:02:00 | sam686 | most likely its the retrieving the stats from mysql in master which may be the biggest slowdown |
| 06:02:04 | raptor | yes |
| 06:04:16 | sam686 | it might help to have a second thread, the main thread does all the stuff as usual, and the second thread does mysql work, although not as problematic as game server (there was a 3 second freeze in game server writing to sqlite or mysql, until i moved that to deperate thread to fix freezing) |
| 06:04:37 | sam686 | oh, look what time it is, i think it is time to get to sleep... |
| 06:06:16 | sam686 | since the stats is only being read, it can be catched (hold in memory) for about 20 minutes, or until a new game report comes in.. |
| 06:06:54 | raptor | yeah, i've been thinking about a cache on master |
| 06:07:00 | raptor | that would only update when a game ends |
| 06:08:16 | raptor | ok, i have to go to bed, too |
| 06:08:18 | raptor | good night! |
| 06:11:16 | zoomber_mbp | anyways |
| 06:11:22 | zoomber_mbp | oh, is everyone off now? |
| 06:11:26 | raptor | almost |
| 06:11:32 | | BFLogBot - Commit 7426776b5933 | Author: buckyballreaction | Log: Add 017 update directives, starting with useFakeFullScreen for Mac |
| 06:11:35 | zoomber_mbp | i guess i missed the convo |
| 06:11:42 | raptor | good night! |
| 06:11:48 | zoomber_mbp | aww |
| 06:11:48 | sam686 | probably yes, it is 12:11 AM for my clock |
| 06:11:49 | zoomber_mbp | good ngiht |
| 06:11:59 | | raptor Quit () |
| 06:12:03 | | zoomber_mbp Quit (Quit: zoomber_mbp) |
| 06:16:24 | | sam686 Quit (Ping timeout: 245 seconds) |
| 10:04:24 | | watusimoto Quit (Quit: Leaving.) |
| 12:14:28 | | Watusimoto has joined |
| 12:31:27 | | LordDVG has joined |
| 14:28:31 | | watusimoto1 has joined |
| 14:52:16 | | Watusimoto Quit (Ping timeout: 276 seconds) |
| 16:41:59 | | raptor has joined |
| 16:42:01 | | ChanServ sets mode +o raptor |
| 16:49:41 | raptor | buenos |
| 17:14:52 | watusimoto1 | hi |
| 17:14:56 | raptor | hello |
| 17:14:56 | watusimoto1 | heading home now |
| 17:15:08 | watusimoto1 | I'll implement a stats cache tonight |
| 17:15:21 | raptor | you're coherent! |
| 17:15:38 | raptor | ok |
| 17:16:08 | raptor | you think that is better than a db thread? |
| 17:16:38 | watusimoto1 | no |
| 17:16:44 | raptor | or maybe both would be good.. |
| 17:16:49 | watusimoto1 | both |
| 17:17:29 | raptor | i spent most of last night determining why SDL was losing the openGL context since the release of 1.2.15 |
| 17:17:33 | raptor | on Mac |
| 17:17:47 | watusimoto1 | did you figure it out? |
| 17:17:48 | raptor | and found out that an SDL dev explicitly disabled it for unknown reasons... |
| 17:17:59 | watusimoto1 | interesting |
| 17:18:03 | watusimoto1 | accidentally? |
| 17:18:04 | raptor | so i reverted his change and recompiled for our purposes |
| 17:18:06 | raptor | no |
| 17:18:37 | watusimoto1 | I'm beginning to feel sdl is an ameteur production |
| 17:18:42 | watusimoto1 | at least in some ways |
| 17:19:09 | raptor | yeah - me too... i think they are really trying to solve it in SDL 2 |
| 17:19:27 | watusimoto1 | back to the stats for a sec -- I think the thread is needed if the master stalls for 3 secs |
| 17:19:30 | raptor | i guess it's popularity arose because there was nothing else as *good*? |
| 17:19:38 | watusimoto1 | caching will help somewhat , but probably not much |
| 17:19:45 | raptor | ok |
| 17:19:56 | raptor | TNL::Thread should be good to use now on linux... |
| 17:19:57 | watusimoto1 | glut was pretty good in its way, but it was completely dead and rather limited |
| 17:20:20 | watusimoto1 | I'm hoping sam will do it as he is the only one with experience |
| 17:21:09 | watusimoto1 | one alternative to caching would be to have a webservice provide the stats |
| 17:21:24 | raptor | ahh webservices |
| 17:21:24 | watusimoto1 | and have the clients get them "out of band" from the master |
| 17:21:42 | watusimoto1 | well something to consider anyway |
| 17:21:52 | raptor | i could whip up a JPA/JAX-WS/J2EE service |
| 17:22:00 | raptor | and have it use up all the system memory |
| 17:22:12 | watusimoto1 | the service is easy; the question is how to get the game to use it |
| 17:22:27 | watusimoto1 | not sure how to use http from tnl |
| 17:22:34 | raptor | we have RPC? |
| 17:22:48 | watusimoto1 | yes, but that is not "out of band" |
| 17:23:02 | watusimoto1 | ok, well I'm out of here -- I have german course tonight, but I'll be on in a few hours |
| 17:23:06 | raptor | ok later |
| 17:23:14 | watusimoto1 | bis spaeter! |
| 17:23:18 | watusimoto1 | (until later) |
| 17:23:26 | raptor | hasta la pasta |
| 17:23:40 | raptor | until pasta (dinner) |
| 17:27:39 | | watusimoto1 Quit (Ping timeout: 260 seconds) |
| 17:33:03 | | sam686 has joined |
| 17:33:03 | | ChanServ sets mode +v sam686 |
| 18:05:46 | | Watusimoto has joined |
| 18:42:22 | | Watusimoto Quit (Ping timeout: 276 seconds) |
| 20:40:26 | | Watusimoto has joined |
| 20:42:25 | Watusimoto | hi |
| 20:42:33 | raptor | hello |
| 20:42:50 | | BFLogBot - Commit 2941021e6dd6 | Author: watusim...@bitfighter.org | Log: whitespace |
| 20:42:52 | | BFLogBot - Commit 011b90c2cc07 | Author: watusim...@bitfighter.org | Log: Check for robot status more directly |
| 20:42:53 | | BFLogBot - Commit e179b58dc436 | Author: watusim...@bitfighter.org | Log: Cache stats on master (untested) probably will still need threaded access to db |
| 20:42:55 | | BFLogBot - Commit 5c131dee859a | Author: watusim...@bitfighter.org | Log: Merge |
| 20:43:39 | Watusimoto | for some strange reason, tortoise hg thinks I am using a swiss keyboard |
| 20:43:43 | Watusimoto | I really don't get it |
| 20:43:48 | Watusimoto | all my keys are remapped |
| 20:44:07 | raptor | there are probably multiple localization layers in the OS |
| 20:45:37 | | LordDVG Quit (Remote host closed the connection) |
| 20:50:34 | Watusimoto | and one of them has activated itself for me in hg |
| 20:50:41 | Watusimoto | oh well |
| 20:50:48 | Watusimoto | I'm learning where they keys are |
| 21:17:39 | | sam686 Quit (Ping timeout: 245 seconds) |
| 21:18:14 | | sam686 has joined |
| 21:18:14 | | ChanServ sets mode +v sam686 |
| 23:07:40 | raptor | ok, i'm out for the rest of the afternoon |
| 23:07:49 | raptor | so good night! |
| 23:13:54 | | raptor Quit () |
