| 00:21.01 | *** join/#bzflag infobot (~infobot@rikers.org) |
| 00:21.01 | *** topic/#bzflag is https://BZFlag.org || http://ohloh.net/p/bzflag || https://bzflag.org/help || Channel Logs: http://infobot.rikers.org/%23bzflag/ || 2.4.10 released! https://forums.bzflag.org/viewtopic.php?f=8&t=19401 |
| 00:21.01 | *** mode/#bzflag [+v infobot] by ChanServ |
| 00:23.36 | *** join/#bzflag blastis (~BLASTIS@c-24-8-43-250.hsd1.co.comcast.net) |
| 01:50.03 | *** join/#bzflag bertman (~bert@unaffiliated/bertman) |
| 01:58.55 | *** join/#bzflag bertman_ (~bert@unaffiliated/bertman) |
| 04:13.50 | *** join/#bzflag Periapsis (~nadir@unaffiliated/znadir) |
| 05:04.08 | *** join/#bzflag Periapsis (~nadir@unaffiliated/znadir) |
| 06:55.10 | *** join/#bzflag Dracos85 (~dracos85@67-60-40-148.cpe.cableone.net) |
| 07:39.36 | *** join/#bzflag BZnotify (~BZnotify@192.30.252.45) |
| 07:39.36 | BZnotify | [13bzflag] 15macsforme pushed 1 new commit to 062.4: 02https://git.io/vQ6m4 |
| 07:39.36 | BZnotify | 13bzflag/062.4 1433bfdcc 15Joshua Bodine: Implemented a workaround for an issue with scaled retina screens under macOS where SDL reports a resolution that the display can't actually use. |
| 07:39.36 | *** part/#bzflag BZnotify (~BZnotify@192.30.252.45) |
| 10:11.07 | *** join/#bzflag Dracos85 (~dracos85@67-60-40-148.cpe.cableone.net) |
| 12:38.53 | blast007 | macsforme: "On Apple's OS X you must set the NSHighResolutionCapable Info.plist property to YES, otherwise you will not receive a High DPI OpenGL canvas." |
| 12:40.56 | blast007 | and there's also the SDL_WINDOW_ALLOW_HIGHDPI window flag, though they don't indicate if that's needed for macOS http://wiki.libsdl.org/SDL_WindowFlags |
| 13:43.47 | blast007 | https://discourse.libsdl.org/t/macbook-pro-with-retina-display-strange-behaviour/22570/9 |
| 19:12.38 | *** join/#bzflag Zehra (~Zehra@unaffiliated/zehra) |
| 19:24.51 | *** join/#bzflag rob1n (~rob1n@unaffiliated/rob1n) |
| 21:08.32 | *** join/#bzflag Delusional_ (~Delusiona@unaffiliated/delusional) |
| 22:05.45 | macsforme | blast007: okay, but I don't see what that has to do with the problem in question |
| 22:06.55 | macsforme | we don't use a high-dpi canvas |
| 22:11.12 | JeffM | maybe you should? |
| 22:14.13 | macsforme | it would be a very invasive change... our graphics code has a very top-down approach, asking for a specific resolution and then expecting it to be exactly that |
| 22:17.12 | JeffM | and the issue is that the resolution reported is not the one that can be given? |
| 22:24.25 | macsforme | yes, you ask for the "logical" resolution and the actual resolution you get back may be double that, or some other multiplier |
| 22:24.48 | JeffM | can you detect the resolution from the context that you get back? |
| 22:26.50 | JeffM | or get the resolution of the desktop at least? |
| 22:28.15 | macsforme | you can ask SDL for it (that's easy)... the issue is that bz asks "make me a window with this resolution," and right now there is no way of telling it "okay, but what I actually got back was this," at least with our current window initialization routine |
| 22:28.32 | JeffM | sure, the current stuff is very outdated. |
| 22:28.53 | JeffM | can you think of any valid reason on a modern computer (aka one that doesn't use a CRT) to allow bzflag to change the screen res? |
| 22:29.34 | JeffM | would it be viable to remove all the resolution changing code, and just have two options, window or full screen, and when you are full screen you get whatever your native resolution is |
| 22:30.02 | JeffM | or at least make that the default and have a bzdb option to force a specific res, but only change reses if that is set by the user. |
| 22:30.45 | JeffM | as long as the game can query how big the context is after it's created, the screen size globals can be set and the rest of the game will work. |
| 22:31.50 | macsforme | only if we care about performance on very low-end hardware (which I personally don't particularly), where they might want to scale down the resolution to squeeze out a little more performance |
| 22:32.05 | JeffM | there are other options to make it run faster |
| 22:32.32 | macsforme | but even with that, the more modern way would be to render to a low-resolution framebuffer, and blit that to the main one, rather than doing a modeset |
| 22:32.39 | JeffM | scaling the res on an LCD makes it look like crap |
| 22:32.52 | JeffM | sure, but that would be for later ;) |
| 22:33.17 | JeffM | really you are talking about people with old intel onobard trying to run games in HD/4k :) |
| 22:33.31 | JeffM | and they are not a big group |
| 22:34.17 | JeffM | it sounds like it may be resonable to always create the context at desktop res, and go from there. If many people complain, then later do the smaller sized back buffer thing. |
| 22:34.26 | macsforme | but yes, basically I think all the resolution switching code could go away |
| 22:34.32 | JeffM | or those people can run in a window |
| 22:34.40 | JeffM | it woudl make things simpler |
| 22:34.51 | JeffM | as long as SDL can tell you how big the context is |
| 22:34.54 | macsforme | or adjust the resolution of their display using system settings prior to launching the client |
| 22:35.03 | JeffM | sure |
| 22:35.20 | JeffM | I don't think that would be too bad of a client change |
| 22:35.31 | JeffM | sure it'd hit the entire startup but it'd streamline it for sure |
| 22:38.12 | macsforme | I looked once, and I think I found that our window setup code creates and destroys the main window about 5 times during the initial startup |
| 22:38.24 | JeffM | yeah it was not good code :) |
| 22:38.37 | JeffM | I did not write it but I remember it from making it work on windows. |
| 22:38.47 | JeffM | the "find a good default" stuff is not efficient |
| 22:39.31 | JeffM | the default settings can probalby just be hardcoded, and then people can be expected to change them as needed. |
| 22:39.43 | JeffM | all that stuff is from when hardware accelerators were optional |
| 22:46.23 | JeffM | at that point would it not make sense to always ask for the high DPI context? |
| 22:46.41 | macsforme | sure |
| 22:50.32 | BulletCatcher | I use a 4x3 full-screen resolution on my 16x9 display to get a taller field of view. |
| 22:50.34 | BulletCatcher | That works better for me than running in a window. |
| 22:51.10 | JeffM | rotate the display, or set the FOV ;) |