| 00:26.55 | macsforme | are we going to do an RC release? there are some big changes in this one |
| 00:30.09 | macsforme | the other thing was disabling the menu entry for starting servers when --disable-server is given... but I don't think I'll get to that soon and I don't think we should hold up the release for that |
| 00:36.34 | blast007 | is that mostly about the keyboard layout stuff? |
| 01:14.21 | macsforme | the use of an OpenGL extension to such an extent is a big change too |
| 01:46.01 | JeffM | What extension did you use? And how long has it existed? How many verified testers have you had of the code? |
| 02:02.48 | *** join/#bzflag bertman (~bert@unaffiliated/bertman) |
| 07:44.11 | blast007 | GL_ARB_framebuffer_object - approved Aug 4, 2008 - OpenGL 1.1 is required |
| 14:29.08 | JeffM | oh god...that's nothing new |
| 14:32.25 | JeffM | all you need to worry about is if you've had an acceptable number of people test it on the various platforms. |
| 16:24.59 | *** join/#bzflag nadir (uid134094@gateway/web/irccloud.com/x-guxrnzxrcnejpdgp) |
| 17:03.17 | macsforme | it should also continue working without the feature on systems where the extension doesn't exist... although I haven't been able to actually try it on a system without the extension, and it might be hard to find one |
| 17:05.52 | macsforme | but it follows the same extension testing pattern used by GL_EXT_texture_filter_anisotropic |
| 17:14.40 | JeffM | honestly anyone who dosn't have GL 1.1 should not be running the game |
| 17:17.02 | blast007 | my Pentium M laptop from 2006 with Intel 855GM iGPU (which only fully supports OpenGL 1.3) running XP lacks the extension ;) |
| 17:17.17 | JeffM | it should not run the game |
| 17:17.24 | blast007 | oh it doesn't, not anymore |
| 17:17.33 | macsforme | the game won't run without GL 1.1... https://github.com/BZFlag-Dev/bzflag/blob/2.4/include/bzfgl.h#L34-L36 |
| 17:17.33 | blast007 | using an OpenGL extension viewer |
| 17:17.38 | JeffM | what serious age of computer to you want to support? |
| 17:17.51 | JeffM | 10 years? |
| 17:17.54 | blast007 | 2009 |
| 17:18.05 | blast007 | so my thinkpad x200 can still play :) |
| 17:18.41 | JeffM | you want that as a fixed date? or are you willing to let it be a sliding scale of 8 years? |
| 17:19.13 | blast007 | I mean, this thing can do OpenGL 2.1 |
| 17:19.39 | JeffM | but at what age do you stop having to care if a developer breaks it? |
| 17:22.55 | blast007 | is there a point that we're going to need more than 8GB of RAM or require more than OpenGL 2.1? |
| 17:23.22 | JeffM | probalby not the former, but depending on how much you change, possibly the later |
| 17:23.38 | JeffM | do you ever want fancy shaders? |
| 17:24.12 | JeffM | but as for testing your extension thing, if you've shown that an acceptable number of users have tested it with out issue, it's probalby good for release. |
| 17:24.15 | blast007 | if we do redo the graphics, I'd hope we could split it out to a lib - then when Apple keeps breaking GL we can use Metal |
| 17:24.26 | macsforme | nah, tessellating quad faces for lighting is sooo much fun :P |
| 17:24.43 | JeffM | how long has the extension been in ? |
| 17:25.03 | JeffM | the code not GL :) |
| 17:25.04 | blast007 | it shows up on this system, but I don't know if the Windows driver would have it |
| 17:25.12 | blast007 | might be mesa adding it |
| 17:27.02 | blast007 | doesn't show any AA levels in the game though, so it might not do AA :) |
| 17:27.29 | macsforme | just using shaders at all (OpenGL 2.0) could bring some significant graphical improvements... for example, per-pixel lighting, normal maps, soft shadows... |
| 17:27.46 | blast007 | JeffM: oh, in our code - it's not been in a release yet afaik |
| 17:28.12 | macsforme | I don't think we're anywhere near requiring stuff that isn't in the latest OpenGL version that Apple supports |
| 17:28.32 | JeffM | I know not a release, but how long has it been in the repo |
| 17:28.50 | blast007 | macsforme: I mean how they keep breaking it |
| 17:31.52 | macsforme | JeffM: since July 21... https://github.com/BZFlag-Dev/bzflag/commit/1798d351595530d805ef7164bbc823e3fa6d6a61 |
| 17:32.10 | JeffM | and how many people to you KNOW have run a client built with that code? |
| 17:32.50 | macsforme | three :P |
| 17:33.02 | JeffM | so 10% of the population.. ship it |
| 17:33.48 | blast007 | if it breaks, we release a 2.4.14 |
| 17:33.52 | JeffM | do you feel the risk of it failing is signifigant? |
| 17:34.50 | blast007 | it's optional anyway |
| 17:35.03 | blast007 | if the system doesn't have the extension, it doesn't try to read the multisample levels |
| 17:35.13 | JeffM | is it enabled by default? |
| 17:35.23 | blast007 | the check for the extension? |
| 17:35.28 | JeffM | yes |
| 17:35.31 | blast007 | yes |
| 17:35.41 | blast007 | just like our check for aniso levels |
| 17:35.52 | JeffM | so everyone who gets the new release will execute the code from the feature |
| 17:36.11 | JeffM | is there a risk of the code in that PR, causing the game to be unplayable or crash? |
| 17:36.23 | blast007 | don't see how |
| 17:36.31 | JeffM | then the risk assement is very low |
| 17:36.34 | blast007 | like, if they don't have the extension? |
| 17:36.44 | blast007 | cuz if they don't have it, it won't do much |
| 17:36.46 | JeffM | like there is a bug in what he wrote and the app goes BOOM |
| 17:36.53 | JeffM | for any code path |
| 17:37.09 | JeffM | if all code paths will function at some acceptable level, the risk is low |
| 17:37.22 | JeffM | that means you can get away with shipping it with minimal testing |
| 17:37.48 | JeffM | optimaly you'd want to have some level of confidence that it's been run "in end user situations" |
| 17:38.51 | blast007 | I've run it on this linux laptop from 2009, a more recent third-gen i3 with linux, and on my Windows 7 desktop |
| 17:38.52 | macsforme | that was the whole point of suggesting an RC release |
| 17:39.05 | JeffM | indeed |
| 17:39.20 | JeffM | blast007: how much testing of the entire release do you know has happened? any? |
| 17:39.48 | JeffM | and what is your acceptable level of testing for QA? |
| 17:40.16 | JeffM | is it "it works for me" level? or do you need/want some level of confidence using other people? |
| 17:42.51 | blast007 | even if we do an RC there's no guarantee |
| 17:43.12 | blast007 | I don't think doing an RC really gets us much for these smaller changes |
| 17:43.16 | JeffM | there never is a guarantee, there is only levels of confidence. |
| 17:43.27 | JeffM | so you are of the "it works for me" camp |
| 17:43.38 | JeffM | if that's your policy, then that's your policy |
| 17:43.44 | JeffM | just have one and stick to it. |
| 17:44.04 | JeffM | high levels of confidence ARE hard in a community this small |
| 17:44.09 | JeffM | so it's an understandabale policy |
| 17:44.10 | blast007 | the options: release an RC, and if it works, release 2.4.12 -or- release a 2.4.12 and it it doesn't work release a 2.4.14 |
| 17:44.23 | JeffM | there is a third option |
| 17:44.23 | blast007 | the work involved is mostly the same |
| 17:45.10 | JeffM | Do a dev build, ask specific people to test it, then you get some level of overall confidence |
| 17:45.13 | JeffM | like a closed RC. |
| 17:45.29 | blast007 | that's a similar amount of work to a RC build |
| 17:45.37 | JeffM | I understand not wanting to do an RC, it means stilling around and waiting for people to test it, and that NEVER happens. |
| 17:45.43 | blast007 | if we just release 2.4.12 and everything works, that's less work :) |
| 17:46.02 | JeffM | remind me to never hire you as a development manager, but whatever floats your boat :) |
| 17:46.09 | JeffM | you are so very much a web developer |
| 17:46.11 | blast007 | heh |
| 17:46.41 | JeffM | macsforme: he's going for super agile, not more traditonal waterfal |
| 17:46.55 | JeffM | if it's busted, ship another, who cares if it is busted in the field for a day/week |
| 17:47.06 | JeffM | it's a common thing used with websites and services. |
| 17:47.23 | JeffM | where I think you and I are more used to some form of QA pass :) |
| 17:48.10 | blast007 | *shrugs* it's been tested on all three platforms with a range of hardware |
| 17:48.23 | JeffM | then it's good enough for you |
| 17:48.30 | JeffM | it is just a silly tank game |
| 17:48.48 | blast007 | and hasn't the mac build been broken on retina systems for a while? |
| 17:48.56 | JeffM | would you do some form of public test/RC with larger structural changes? |
| 17:48.58 | blast007 | 2.4.12 is supposed to fix that |
| 17:49.05 | blast007 | oh sure |
| 17:49.17 | blast007 | there's only been 50 commits since the last release, and most of those are to the bzfs API |
| 17:49.19 | JeffM | so you see this as like a security fix |
| 17:49.26 | JeffM | then the risk is probalby low |
| 17:49.30 | JeffM | as long as you've thought about it |
| 17:49.41 | JeffM | then if it bites you in the ass, you take responsibility for it :) |
| 17:49.57 | JeffM | I'm just here to make sure you've thought about it |
| 17:50.25 | blast007 | an RC release will get us a smaller sample size of testing |
| 17:51.15 | JeffM | sure, a full release forces people to test your code |
| 17:51.24 | macsforme | are there any linux distros that are close to a cutoff for their next release, that might get this release but might not get the next bugfix release if we have to do one? |
| 17:51.31 | blast007 | was the idea to test the extension thing, or the the key events? |
| 17:51.34 | JeffM | in a lot of industries, that would be considered BAD, because paying customers should not be forced to test :) |
| 17:51.59 | blast007 | macsforme: let me go ask all 500 of them :) |
| 17:52.05 | blast007 | odds are, ONE of them is |
| 17:52.12 | JeffM | macsforme: with the way distro's work that's just going to happen |
| 17:52.26 | JeffM | yeah, they are NOT conducive to any kind of release schedule. |
| 17:53.11 | blast007 | pretty sure I've seen at least one that used a pre-release version from VCS |
| 17:54.32 | blast007 | debian had three different 2.0.13 builds |
| 17:59.23 | macsforme | well, as has been said, there won't be a significant difference either way |
| 18:00.12 | JeffM | it's good to bring it up though, then at least everyone knows what the possible ramifications of just releasing are. |
| 18:03.23 | macsforme | are we going to release today, then? |
| 22:16.29 | *** join/#bzflag clinicallyinsane (uid258957@gateway/web/irccloud.com/x-izgbxcaqbrjpjqwj) |