| *** StocksRoomm <StocksRoomm!~StocksRoo@37.19.211.45> has joined #bzflag | 00:45 | |
| *** StocksRoomm <StocksRoomm!~StocksRoo@37.19.211.45> has quit IRC (K-Lined) | 00:45 | |
| macsforme | global illumination with the fixed-function OpenGL pipeline is rather impractical (impossible?) | 01:59 |
|---|---|---|
| SpringTank | you might be able to make a "lightobject" using some code from the bullet lighting, but too many lights would probably begin to slow things down. | 02:15 |
| SpringTank | off topic question: does windows handle intergers and floats different from linux? I'm doing a multi-arch program and I get more bugs on windows than I do linux it seems. Weird little things | 02:16 |
| SpringTank | doing subtraction between pixels on an image and linux the result is clean, but on windows it leaves little bits here an there. | 02:18 |
| Agatha | Does bz even use shaders? I think it doesn't. (I've been told this before but I forgot.) | 02:27 |
| Agatha | @SpringTank any differences would most likely be due to using different compilers, but reasonably written programs should work the same: if they aren't perhaps you have uninitialized variables somewhere. Run it through valgrind. | 02:27 |
| SpringTank | I should provably do that. And no bzflag does not use shaders | 02:28 |
| SpringTank | that helped. fixed some problems and found an array I wasn't initializing | 02:58 |
| SpringTank | so it was starting with garbage and that's probaby where my strays were coming from because now they are gone | 02:59 |
| SpringTank | thanks! | 02:59 |
| *** _I_Died_Once <_I_Died_Once!~I_Died_On@c-73-184-170-223.hsd1.ga.comcast.net> has quit IRC (Ping timeout: 248 seconds) | 04:48 | |
| *** StocksRoomm <StocksRoomm!~StocksRoo@37.19.211.45> has joined #bzflag | 05:13 | |
| *** StocksRoomm <StocksRoomm!~StocksRoo@37.19.211.45> has quit IRC (K-Lined) | 05:13 | |
| *** StocksRoomm <StocksRoomm!~StocksRoo@185.76.9.58> has joined #bzflag | 05:40 | |
| *** StocksRoomm <StocksRoomm!~StocksRoo@185.76.9.58> has quit IRC (K-Lined) | 05:40 | |
| tupone | SpringTank: There is a branch on my repository that uses shaders. | 06:13 |
| tupone | well was for Agatha | 06:14 |
| SpringTank | wait, really? does it work? | 06:15 |
| *** Sgeo <Sgeo!~Sgeo@user/sgeo> has quit IRC (Read error: Connection reset by peer) | 06:32 | |
| *** FastLizard4 is now away: IN-GAME - Outer Wilds | 07:21 | |
| *** StocksRoomm <StocksRoomm!~StocksRoo@185.76.9.58> has joined #bzflag | 08:36 | |
| *** StocksRoomm <StocksRoomm!~StocksRoo@185.76.9.58> has quit IRC (K-Lined) | 08:36 | |
| *** FastLizard4 is back | 09:17 | |
| *** FastLizard4 is now away: AWAY from keyboard | 09:29 | |
| *** FastLizard4 is back | 09:53 | |
| *** FastLizard4 is now away: AWAY from keyboard | 10:53 | |
| *** FastLizard4 is now away: GONE - Screen Detached and Disconnected from IRC (I'm probably asleep, at work, or doing something in real life) | 11:11 | |
| *** _I_Died_Once <_I_Died_Once!~I_Died_On@c-73-184-170-223.hsd1.ga.comcast.net> has joined #bzflag | 12:02 | |
| *** Sgeo <Sgeo!~Sgeo@user/sgeo> has joined #bzflag | 13:31 | |
| *** StocksRoomm <StocksRoomm!~StocksRoo@185.76.9.58> has joined #bzflag | 13:43 | |
| *** StocksRoomm <StocksRoomm!~StocksRoo@185.76.9.58> has quit IRC (K-Lined) | 13:43 | |
| *** StocksRoomm <StocksRoomm!~StocksRoo@185.76.9.58> has joined #bzflag | 17:04 | |
| *** StocksRoomm <StocksRoomm!~StocksRoo@185.76.9.58> has quit IRC (K-Lined) | 17:04 | |
| Agatha | What's funny is that now that I'm reminded, I remember that that exists; looking at it was unfortunately but honestly one of those things that got popped from my stack in the constant page-thrashing that is pandemic life. | 17:11 |
| Agatha | Using shaders requires moving support from GL 1 -> GL 2 at least AFAIK. Since computers have been GL 4+ for well over a decade (and GL 2 support for nearly two decades), I will again state my opinion that this is a safe change. | 17:11 |
| blast007 | the eventual goal was to target 3.2 or 3.3 | 17:15 |
| blast007 | I don't know enough about OpenGL and graphics to know if it's easier for us to gradually transition to GL 2.0 and then to 3.2/3.3, or just go straight to 3.2/3.3 | 17:16 |
| Agatha | Transitioning GL 1 -> GL 2 is mostly a question of removing some ancient (and almost certainly nonfunctional anyway) GL 1 functionality, which has to be done regardless. | 17:45 |
| Agatha | I thought jumping to GL 3 was a non-starter portability-wise, but if that's on the table then I'll say it's a significant undertaking. Doing it 'properly' would basically amount to rewriting most of the rendering system. With the current resources of bz development, that therefore seems inadvisable. | 17:45 |
| blast007 | https://gist.github.com/blast007/917f72b59fac372fb2f09cd2f555b94f | 17:58 |
| blast007 | if by "portability-wise" you mean hardware support, that's really a non-issue | 17:58 |
| blast007 | most stuff since ~2007 has supported OpenGL 3.2 | 17:59 |
| blast007 | the nVidia and AMD stuff went straight from GL 2.1 to 3.3. Only Intel had some chips that did 3.1. And after those, Intel went to GL 4. | 18:06 |
| blast007 | so.. the limit is really what we have the time/knowledge to implement | 18:10 |
| blast007 | I've never really progressed beyond the spinning cube tutorials :) | 18:12 |
| Agatha | @blast007 I thought the reluctance about moving to newer GL versions was that there are allegedly current bz players who don't have / can't afford hardware-newer-than-two-decades-old. (I fully believe this is unlikely, and that there's no portability issue in practice, but that was the objection, I thought.) | 18:22 |
| Agatha | Note that GL 2.0 support usually implies GL 2.1 support too, GL 3.0 almost always implies 3.3 (with Intel being the notable exception, but their contexts are broken anyway, as @SpringTank found out), and to a lesser extent, GL 4 implies GL 4.6 (at the very least GL 4.0-only support is very rare). That's sortof the idea behind breaking up the major versions; within a major version you keep compatibility. | 18:22 |
| Agatha | From a quick skim of the code it basically `SDL_GL_CreateContext(...)`s a GL context and then `glewInit()`s the rest of the context. On my system, that seems to give me a OpenGL 4.6.0 (compatibility) context. Bzflag runs fine for me; thus one can say that the port from GL 1 to GL 4.6 is already complete :) | 18:22 |
| Agatha | More seriously, the path to GL 3 is probably to request a GL 2.1 *core* context, then fix stuff so everything works, then request a GL 3.3 *core* context, then fix stuff so everything works. | 18:22 |
| SpringTank | "I've never really progressed beyond the spinning cube tutorials :)" Further than I made it, but again I'm using Intel graphics so that's probably a big part of my problem. | 18:29 |
| SpringTank | its an old laptop that I need to upgrade anyways so no skin off my nose | 18:30 |
| blast007 | Agatha: when I force HTTPS to the server list, that'll already take care of some of those people :P | 18:34 |
| blast007 | (people on Windows XP won't be able to play anymore) | 18:34 |
| Agatha | It's worth noting that { GL 2 with shaders } is a totally viable way to improve the graphics. Moving to GL 3.* (or GL 4.*) would mainly enable cleaner graphics code and faster rendering if certain features were used, but otherwise it's not necessary for anything AFAICT | 18:36 |
| blast007 | well, unless they recompile their cURL to use OpenSSL instead of Window's stuff | 18:36 |
| *** FastLizard4 is back | 18:37 | |
| Agatha | Jokes aside, as before, I don't think those people literally even exist, and I strongly doubt GL 1-only support even still works. Bumping minimum version to 2.1 (or higher) just formalizes the current state of affairs without actually affecting anybody. | 18:42 |
| blast007 | I know one person still running XP | 18:43 |
| blast007 | cuz I messaged them about upgrading their client after checking server list logs for non-HTTPS accesses | 18:43 |
| Agatha | I mean, unironically as their main computer? Even so, AFAICT XP supports https and GL 2+, with the right hardware. | 18:45 |
| Agatha | Maybe statistics could be collected on this by the list server or something? | 18:45 |
| *** FastLizard4 is now away: AWAY from keyboard | 18:46 | |
| blast007 | the issue with HTTPS was that we are using WinSSL/schannel, which uses Window's built-in support for SSL/TLS, and the support in XP is a bit lacking | 18:47 |
| blast007 | I didn't want to use OpenSSL because it required Perl to configure the build process for it :P | 18:48 |
| Agatha | I appreciate that. Additional build complication suuuux (especially on windows) | 18:52 |
| * Agatha supposes someone could send them a $40 computer off ebay to make the discussion moot . . . | 19:03 | |
| blast007 | or they can just play Microsoft Hearts | 19:04 |
| Agatha | Scratches the same itch, since bzflag on an xp-era machine is slow enough to qualify as turn-based /s | 19:16 |
| tupone | I have ported the code to GL1.5 using only buffer objects and after that I added suport to shader, about 20/30 separate commits | 19:29 |
| tupone | so now is 2.1 | 19:29 |
| tupone | I used to play bzflag with that changes. I also added per pixel lighting but it is adding load to my poor hardware | 19:31 |
| *** Optic_Delusion_ <Optic_Delusion_!~Optic_Del@2600:4041:19d:2d00:a1b3:350d:3ee2:7064> has joined #bzflag | 20:42 | |
| *** Optic_Delusion <Optic_Delusion!~Optic_Del@2600:4041:19d:2d00:88af:72ae:8272:eea7> has quit IRC (Ping timeout: 260 seconds) | 20:45 | |
| *** StocksRoomm <StocksRoomm!~StocksRoo@185.76.9.58> has joined #bzflag | 21:19 | |
| *** StocksRoomm <StocksRoomm!~StocksRoo@185.76.9.58> has quit IRC (K-Lined) | 21:20 | |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!