Time Nick Message 03:59 celeron55 packer sounds vague enough to make someone look up what its actual purpose is. serialization tends to have more implications that you don't want 08:22 MTDiscord what would it take for #7865 to be merged? I feel it is a very important feature that improves the Android control scheme significantly 08:22 ShadowBot https://github.com/minetest/minetest/issues/7865 -- Add crosshair support for Android by srifqi 09:52 sfan5 reviews probably? 09:52 sfan5 more importantly someone needs to figure out why android is so crashy on the play store, we can't release new versions without this fixes 09:52 sfan5 fixed* 10:02 ROllerozxa fair enough 10:26 sfan5 just found out the default chat limit also applies to commands 10:27 erle by doing what? 10:34 sfan5 by sending more chat commands than the limit allows? 10:35 erle yeah but for what purpose 10:43 rubenwardy I wasn't able to get the tracebacks from the play store. We need better tooling 10:43 sfan5 did you upload debug symbols? 11:34 MTDiscord Hey, my phone crashes on 5.5 can I be useful? 11:47 sfan5 perhaps 11:47 sfan5 going by what ruben said if you can build a debug build, run it and extract the backtrace from logcat that'd be useful 12:05 rubenwardy The build didn't generate the .zip that the docs said it should, and I remember having problems making my own 12:06 rubenwardy My workaround was to use a tool to deobscure the traces locally, but that didn't work on the 5.5.0 crashes 12:06 Zughy[m] considering you can barely follow the PC version of MT, have you considered dropping Android support for now? 12:07 Zughy[m] Because, I mean, it's bad anyway 12:07 sfan5 is bad therefore we should make it worse? 12:08 sfan5 rubenwardy: from what I've seen you can loosely throw the unstripped .so's into a ZIP and google will take it 12:11 Zughy[m] no, it's is bad and we don't seem able to follow it properly anyway, so it might be better to just drop it till we actually have time/manpower for that, whilst focusing on other things we can actually handle in the meantime 13:40 MTDiscord I'm all for leaving the old version up and trying to put more resources into android. 13:49 MTDiscord please, do not drop android support. That would be awful. Many blockbomber players can only use android, and the game is surprisingly playable on android 13:57 Zughy[m] " I'm all for leaving..." <- and where are you going to take those resources exactly? As far as I know there are no core devs focusing specifically on Android, and they're already pretty busy with the general engine. I'm sorry but Android version is bad, plenty of people uses Multicraft instead for a reason. Last time I checked they had a full team of people working solely on an Android version, how can you compare 13:57 Zughy[m] with no people at all? 13:59 MTDiscord most people I know that play Multicraft were those who use iOS since we don't have any official release on there, those on Android actually play on official Minetest app 14:00 Zughy[m] IMHO choosing to keep supporting both things when there is evidently not enough time and energy to support even just one thing (devs are in charge of Irrlicht too now, contrary to some months ago) means badly supporting both of them 14:01 Zughy[m] is there some active contributor active on making PRs for Android? 14:03 sfan5 the premise seems to be that android is taking away time from us doing more important things, but in terms of actual time spent this is not the case 14:09 erle zughy where does your idea of “focus” come from and how did you not go from “it's a manpower problem” to “add more devs or reviewers” (as i adcovated) ? do you think a smaller team is superior or do you think no one would help? 14:09 erle i mean you have been saying it for a while but i don't really see that some tasks take away from others, exactly as sfan5 says. 14:10 erle many devs have their special area 14:11 erle rubenwardy are there non-play-store builds that could help you with debugging? i have an old phone with a shattered screen (i.e. it is unusable), it does run android, but i doubt i can get a play store on it. 14:12 sfan5 the release section contains APKs 14:12 erle well the important question is if running those *helps* 14:12 erle i.e. how to get the debug info you need 14:13 erle for example, the thing has no internet connection 14:13 erle so whatever non-play-store telemetry is there, you can't get to it 14:21 ROllerozxa adb logcat I guess? I don't think play services crash telemetry would be able to magically create anything more than what already gets printed to there 14:24 sfan5 correct, for example this is what google play gives us https://0x0.st/ob6_.txt 14:28 sfan5 although a small detail is still missing #11131 is now ready for review 14:28 ShadowBot https://github.com/minetest/minetest/issues/11131 -- [MANUAL MERGE] Async environment for mods to do concurrent tasks by sfan5 15:20 MTDiscord is there a good guide to building minetest for android on windows? I have found some horribly out-of-date guides 15:25 MTDiscord Isn't it done on Linux? 15:25 ROllerozxa Dunno about Windows but on Linux you can build it by installing OpenJDK 11+ and Gradle, downloading the Android SDK to somewhere and setting the ANDROID_SDK_ROOT env-variable to that location, and doing `./gradlew build`. After that Gradle will install the necessary dependencies (e.g. the NDK) and start building Minetest 15:25 ROllerozxa this was how I set up to build Minetest for Android at least 15:27 ROllerozxa Gradle, OpenJDK and the Android SDK all exist on Windows but it's probably slightly more complicated there 15:45 Zughy[m] erle: my idea of focus comes from the 1.1k+ issues and 120+PRs, which show that devs can't handle the current situation already. About the "add more devs or reviewers", I've already expressed my suggestions on the forum (I saw your comment this morning) 16:01 erle Zughy[m] so are you saying that adding more people would make it worse? 16:02 erle a.k.a. “the mythical man month” 16:04 erle wuzzy i think your point about the style of reviews mattering is spot-on. it can be *very* frustrating for people to first get advice from a “human linter” but no advice on if the approach is even correct. 16:05 erle i have alienated contributors in other projects with this style of review myself, so i know what i'm talking about ;> 16:22 Zughy[m] erle: no, I'm saying that having a method will improve the situation no matter whether there are new maintainers or not 16:23 Zughy[m] I think nobody despises new maintainers as long as they're competent and they don't ghost the project after a month 17:20 MTDiscord Thank you for adding async environment sfan5, you made my day 17:21 erle what are you planning to use it for? 17:33 v-rob[m] Just a note: if there's a meeting tomorrow at 6 PM UTC, I'll likely be able to come partway through at 6:30 PM. 17:36 MTDiscord Zughy: Minetest continues to grow in popularity, and with the kinds of people that would file issues and make PRs. I would say that in order to handle the workload, we do need more people, but by no means are the current priorities just wrong. Hopefully when I start selling this as a retail box I can help give us some resources to pay some people to help out. We'll see 17:37 v-rob[m] Retail? 17:38 erle lol what 17:38 erle minetest retail box when and where 19:03 erle zughy if you think that your ideas are sound, why haven't you forked minetest? 19:04 erle this is not a troll question. if you would ask me, the answer is: i do not have the energy right now. 19:32 ROllerozxa erle: you develop a minetest hacked client though right? 19:33 erle no, cora does. 19:36 erle ROllerozxa, waspsaliva (cora's client) is basically a) new GUI for enabling/disabling rendering options and CSMs (i think mostly made by fleckenstein) b) new lua APIs c) CSMs that use these APIs. otherwise she tries to follow vanilla minetest, but i am not sure if she has managed to get a ws variant of 5.5 out yet. 19:37 ROllerozxa ah 19:39 erle i do think that some of those things would be very useful for vanilla minetest, like toggling rendering opitons at runtime. but every time i ask people what APIs would be useful for vanilla minetest, the immediate answer is something along the lines of “it's mostly useful for cheaters”. 19:40 erle ROllerozxa if you find something that you think the minetest engine could need, tell! maybe i can convince cora to make a PR :P 20:22 MisterE[m] toggling rendering options (especially client shadow settings! and other graphics related settings) would be amazing 20:25 erle well, it is mostly used for a) avoiding laggy rendering b) fullbright & brightnight shenanigans c) xray hax d) tracers 20:26 erle not sure how it is implemented though 20:26 erle i may remember something wrong, so better check it out yourself! 20:27 erle what i find especially funny is that CSMs can in fact identify entities or nodes by texture and act upon them like that 20:27 erle wanna have a tracer to everything that uses a specific texture? i guess it's possible 21:24 MTDiscord erle: that is the reason why anyone has not forked the engine who is dissatisfied: they do not have enough manpower to make the situation any better in a fork. 21:33 erle MisterE i believe that is actually mostly because it isn't *that* bad. it is just slow. if minetest would, for example, increase hardware requirements to the point it no longer runs on my machine or those of my friends, i would probably not be the only one wanting to maintain a client to make it work. 21:34 erle doing that just to get “proper” handling of integer overflows though? no chance anyone puts in time 21:34 erle better just wait until it is fixed in minetest proper, or scare debian to fix it for you ;) 21:36 erle rubenwardy since the CVEs made debian (and other distributions probably) patch holes in old minetests for free, do you have any opinion on me writing a bunch of notes about crashes fixed in 5.5? 21:37 erle i believe there is great value in letting distributions mantain old versions (as in: someone else does the work, yay), they only need a justification and patches. 21:50 sfan5 will distributions also take our bug reports? 21:51 sfan5 because I can assure you I *will* be upset when I waste an hour tracking down a reported bug that only exists in Debian's patched version but none of our git revisions 21:56 sfan5 though most often than not I do not look at bugs that do not reproduce in the latest version anyway