00:02.07 | BulletCatcher | Having now done that, I see the triangle in FULLSCREEN_MODE 0 (windowed) but just blackness in 1 and 2. |
00:02.50 | blast007 | :/ |
00:03.03 | BulletCatcher | It's a pretty blue triangle. |
00:03.23 | blast007 | yeah.. dunno why it's solid blue.. |
00:06.10 | blast007 | comment out all the code between SDL_GL_MakeCurrent(window, context); and glBegin(GL_TRIANGLES); |
00:09.18 | BulletCatcher | Oooh, the colors! (windowed mode only) |
00:10.51 | blast007 | why isn't it working in fullscreen mode? :/ |
00:11.57 | BulletCatcher | I'm willing to say my computer is somehow at fault, but wouldn't be able to be more specific than that. |
00:13.29 | blast007 | is it the same if you set the delay to 5000? |
00:14.43 | BulletCatcher | Yup. It just runs for 2 seconds longer. |
00:20.09 | *** join/#bzflag infobot (ibot@rikers.org) |
00:20.09 | *** topic/#bzflag is https://BZFlag.org || http://ohloh.net/p/bzflag || https://bzflag.org/help || Channel Logs: http://infobot.rikers.org/%23bzflag/ || 2.4.14 released! https://forums.bzflag.org/viewtopic.php?f=62&t=19953 |
00:20.09 | *** mode/#bzflag [+v infobot] by ChanServ |
00:20.36 | blast007 | bleh |
00:23.41 | blast007 | on the plus side, it's nice that we now have a super minimal test case that doesn't work right, though I'm not sure if it's the same kind of doesn't work right... |
02:04.23 | *** join/#bzflag Kutakizukari (~Kutakizuk@207.189.2.20) |
03:15.22 | macsforme | blast007: what is FULLSCREEN_MODE == 1 supposed to represent? are you creating a standard window and then setting it fullscreen afterwards, rather than passing SDL_WINDOW_FULLSCREEN as a flag to SDL_CreateWindow()? |
03:18.33 | macsforme | you also have no glClear() or glViewport() calls... I'm uncertain whether adding those would make any difference |
03:20.57 | macsforme | for all those state calls, are you trying to set up something similar to what bz has, or just trying to set up some arbitrary drawing state? |
03:33.35 | macsforme | you also disabled GL_COLOR_MATERIAL, then proceeded to specify colors |
03:53.17 | macsforme | the singular color appears to be due to glShadeModel(GL_FLAT) |
03:58.24 | BulletCatcher | Confirmed: I see the colorful triangle (still only in windowed mode) when that glShadeModel() call is the only line commented out. |
04:04.39 | macsforme | can you add this around line 78 and see if anything changes: SDL_DisplayMode displayMode; SDL_GetWindowDisplayMode(window, &displayMode); glViewport(0, 0, displayMode.w, displayMode.h); |
04:05.57 | macsforme | actually, around line 82, after the GL context is initialized |
04:07.39 | BulletCatcher | It makes no difference for me. |
04:14.09 | macsforme | how about this after line 79: if(context == nullptr) cerr << "Unable to initialize OpenGL Context: " << SDL_GetError() << endl; |
05:14.25 | *** join/#bzflag FastLizard4 (~fastlizar@wikipedia/pdpc.active.FastLizard4) |
05:43.37 | *** join/#bzflag FastLizard4 (fastlizard@wikipedia/pdpc.active.FastLizard4) |
05:50.20 | allejo | macsforme: fwiw, I was able to replicate the F4 minimizing issue. 2.4.10 vs 2.4.14 |
05:51.27 | allejo | I have the same behavior as you on 2.4.14. for 2.4.10 it's the same behavior as you but after it goes behind the menu bar, it minimizes |
05:51.55 | allejo | 10.13.4 for me |
10:09.44 | blast007 | macsforme: I was trying to replicate the GL setup that BZFlag uses |
10:11.23 | blast007 | macsforme: as for FULLSCREEN_MODE == 1, I believe it wasn't working right when I was just passing SDL_WINDOW_FULLSCREEN to SDL_CreateWindow, but maybe that was before I was using SDL_GetClosestDisplayMode to get a proper resolution |
10:50.06 | blast007 | macsforme: yeah.. creating the window fullscreen still isn't working for me. It changes resolution, but I see the window title bar and border and it's partially offscreen on the right/bottom. |
10:50.33 | blast007 | I'm probably doing something wrong, but SDL's documentation leaves a bit to be desired.. |
10:52.38 | blast007 | I added a new FULLSCREEN_MODE (2 tries to create the window fullscreen from the start, 3 is what 2 was before) and replaced the GL setup code with that from NeHe Lession 03's Linux/SDL port |
10:52.42 | blast007 | https://gist.github.com/blast007/57f6d95522932035ed57db27b6db7070 |
10:54.02 | blast007 | I commented out the gluPerspective call because that just results in a black screen |
10:55.48 | blast007 | ah, that was because I left out a glTranslatef |
10:58.50 | blast007 | so for me, modes 0, 1 and 3 work |
14:46.37 | macsforme | allejo: I wonder if we use the same SDL 2 version that we used in 2.4.10, it could work... will try later (or if you get to it first, let me know how it goes) |
14:46.48 | macsforme | will still be worth a bug report to the SDL team either way |
14:54.21 | blast007 | (I also added a nullptr check on the context, but didn't try compiling that) |
14:55.39 | blast007 | macsforme: I used SDL_GL_GetDrawableSize instead of SDL_GetWindowDisplayMode to set the glViewport |
16:24.21 | *** join/#bzflag McSpider (~McSpider@li1055-178.members.linode.com) |
17:11.36 | *** join/#bzflag tupone (~alfredo@gentoo/developer/tupone) |
21:47.10 | *** join/#bzflag Zehra (~Zehra@unaffiliated/zehra) |
22:34.45 | Zehra | Quick question, why can't we we do '/localset panelheight 10'? |
22:35.25 | Zehra | I tried to do something like this to sort of try to make a per-map configuration, but it didn't work. :p |
22:36.50 | blast007 | probably because the menu calls some method to actually make that size change effective |
22:37.30 | blast007 | I *think* BZDB variables can have callbacks assigned, and if so that one may be able to be hooked up to it |
22:38.11 | blast007 | but I don't think /localset is a particularly useful feature |
22:38.49 | Zehra | :/ |
22:39.43 | blast007 | it's kinda the same as hand editing a config file. it shouldn't be necessary (nor should it provide any additional options/ranges than what the menus provide) |
22:41.11 | Zehra | Could multiple configurations or configuration switching be made an option in the future? |
22:41.46 | blast007 | maybe |
22:42.06 | blast007 | I'd like to split the configuration into separate files, for instance, so that would make that kind of thing easier |
22:43.20 | Zehra | Like in different files for graphics, radar, contrast/color, password/callsign? |
22:47.42 | blast007 | something like that |
22:48.24 | blast007 | I think we could do something different for store user credentials - instead of storing the username/password, store something more akin to a cookie |
22:48.33 | blast007 | storing* |
22:49.13 | blast007 | oo, and then cloud sync some of the settings :P |
22:50.10 | Zehra | I'm probably a bit slow on this, but how would that work? |
22:51.11 | blast007 | which part? |
22:52.50 | Zehra | The storing user credentials as something similar to a cookie. |
22:55.01 | JeffM | store a token not the actual credentials |
22:55.41 | Zehra | oh, okay. |
22:56.03 | JeffM | and that token can be invalidated very easlay |
22:56.16 | JeffM | you could also local crypt the stored credentials on windows pretty easy |
22:57.26 | Zehra | This might sound a bit stupid, but could players still be authenticated even if auth is down? |
22:58.14 | Zehra | Like even if it doesn't auth, but I have the right current token, I'd still have a + by my name? |
22:58.36 | JeffM | the server would have to have a way to know it's a valid token |
22:58.56 | JeffM | ther are ways to do it, but at the risk of more security holes |
22:59.04 | JeffM | better to make the auth system redundant |
22:59.37 | JeffM | but there are ideas to put cryptographic signatures into the one shot tokens used for auth to make the game server not have to check back with auth |
23:00.23 | blast007 | Zehra: auth should just be designed to not go down |
23:00.33 | JeffM | basicly the game servers give the auth server a key, when you say "join this server" you login to the auth server, get a one shot key that is signed with the key for the server you are going to, and it's good. |
23:00.54 | JeffM | you still need auth in that case, but it's not a back and forth. |
23:01.02 | blast007 | I don't plan to engineer offline authentication fallbacks into the system :) |
23:01.17 | blast007 | the issue with auth going down hasn't been a real problem for a number of years |
23:01.18 | JeffM | yeah full offline auth is just asking to be hacked |
23:02.04 | Zehra | eek |
23:02.08 | JeffM | the advantage of the signing system is that you can easaly have multiple auth servers and the game server doens't have to care what one it was. |
23:02.51 | blast007 | we'll just have the client launch a web browser to log in |
23:03.21 | JeffM | facebook will work for that |
23:03.30 | JeffM | it's good enough for oculus |
23:03.49 | blast007 | I don't use a facebook login for oculus |
23:06.05 | JeffM | myspace? |
23:06.14 | JeffM | or livejournal? |
23:07.35 | blast007 | Friendster |
23:30.25 | *** join/#bzflag The_Noah (~The_Noah@cpe-97-99-253-228.tx.res.rr.com) |
23:38.11 | macsforme | blast007: I am most curious if there will be a context initialization error in fullscreen for BulletCatcher... since the drawing code works in windowed mode, a context failure is my best guess from what we're seeing |