Time Nick Message 15:22 rubenwardy @grorp: is #13850 ready to be merged? 15:22 ShadowBot https://github.com/minetest/minetest/issues/13850 -- Notify users that they need to reinstall MTG if they still want it by grorp 15:35 MTDiscord yes! :) 15:36 rubenwardy merging #13850 and #13739 in 10 then (the particle one seems to be waiting on feedback?) 15:36 ShadowBot https://github.com/minetest/minetest/issues/13850 -- Notify users that they need to reinstall MTG if they still want it by grorp 15:36 ShadowBot https://github.com/minetest/minetest/issues/13739 -- Add script to update/generate mod translations by Wuzzy2 15:48 [MTMatrix] rubenwardy: I was only curious about velartrill's take but if it's only for that I think you can proceed 15:51 [MTMatrix] Anyone against removing the ASCII logo? https://github.com/minetest/minetest/issues/13287#issuecomment-1752052646 15:51 rubenwardy Zughy: well, I'll wait for what Desour wants to do 15:55 MTDiscord Zughy: I'm in favor of removing the ASCII logo, actually. It's just spam in the terminal. 16:25 lhofhansl Hey all... I would like to bring X2048's "god-rays" to Minetest, behind a proper setting. I have extracted a clean patch from his branch, and had it on for a quite a while. 16:25 lhofhansl What's the best way to go about it? The easiest for me would be to create a branch from the patch in my fork, and then attribute it all to x2048. 16:27 MTDiscord bfs branch? 16:53 rubenwardy yeah that makes sense, use your fork but keep the author info 17:01 grorp I'd like to have #13858 (or a different solution) in 5.8.0 so that Android users aren't greeted with an oversized hypertext[] when they open Minetest after the update. What do you think? 17:01 ShadowBot https://github.com/minetest/minetest/issues/13858 -- Make hypertext[] respect font size settings by grorp 17:02 grorp (The "reinstall MTG" dialog contains a hypertext[].) 17:06 rubenwardy sounds good to me 17:07 rubenwardy using (* font_size / 16) makes sense to me 17:09 rubenwardy but a multiplier could be better, that would be like rem on the web 18:02 [MTMatrix] maybe make just like ffmpeg --hide-banner, but for minetest? 🤔 18:02 [MTMatrix] maybe make just like ffmpeg -hide-banner, but for minetest? 🤔 18:24 sfan5 I vote for keeping the ascii art 18:24 sfan5 removing it to solve a minor visual bug with an uncommong option is a bad choice 18:27 [MTMatrix] Neutral 18:33 Krock the ASCII logo is nice. should be fixed instead of removing 18:40 MTDiscord I've rebased #13609. 18:40 ShadowBot https://github.com/minetest/minetest/issues/13609 -- Upgrade client active object mgr tests to Catch2 by JosiahWI 18:46 rubenwardy I don't think it's worth fixing just for --terminal, just close the issue 18:47 rubenwardy either that or get someone on hacktoberfest to look at it 18:47 rubenwardy or maybe not, there's so many more important things to be working on 18:47 rubenwardy the ASCII logo has close to 0 positive user impact 18:48 pgimeno it helps me distinguish minetest boots visually in the logs quickly 18:48 rubenwardy the --- is supposed to be for that 18:49 pgimeno it's bigger and more prominent 18:49 pgimeno easier to catch if you're paging quickly 19:03 sfan5 do we want to use catch2 for unittests in the future? 19:03 nrz lhofhansl wow god rays, that sounds amazing, if we have water reflection after it'll be nice 19:04 MTDiscord the later is already sorta done 19:04 nrz sfan5, what do we gain using catch2 instead of current internal framework ? 19:04 nrz i'd say if it's a popular framework and license compatible with us, go ahead 19:05 sfan5 I don't know, I am asking due to 13609 19:05 sfan5 the PR only makes sense if we are open to using catch2 for everything in the future 19:06 MTDiscord I created an issue #13610 where I presented my case for it. 19:06 ShadowBot https://github.com/minetest/minetest/issues/13610 -- Migrate unit tests to Catch2 framework. 19:08 nrz ah catch2 is already used by MT, i missed that part, then... the gain is just uniformization then 19:08 celeron55 for what it's worth i like catch2, i've used it in other projects 19:08 celeron55 but i wasn't the one who added it to MT 19:09 nrz i hope at a moment C++ will have modules & unittesting in the STL but it sound to be lost, C++ commitee is a bit bloated by their interia 19:09 nrz inertia 19:09 nrz sfan5, i'd say why not, but we should make a plan to use the same way everywhere after , if we decide to move forward on this 19:10 MTDiscord GTest is also popular, but I have a feeling maybe the licensing is not as friendly since it's Google. And it would be a completely new dependency, whereas we already use Catch2 for benchmarking. 19:12 MTDiscord The C++ commitee is considering adding unittesting support to the STL? Interesting. 19:16 MTDiscord Bing can't find anything about that. Is there a proposol for it somewhere? 19:16 MTDiscord proposal* 20:43 [MTMatrix] #13875 20:43 ShadowBot https://github.com/minetest/minetest/issues/13875 -- Remove Minetest ASCII art at launch by Zughy 20:45 ROllerozxa heh brilliant, controversial label 20:46 [MTMatrix] Looks like I'm a living magnet for drama 20:46 erle it can be both trivial and controversial though 20:48 [MTMatrix] and of course I worked on the PR missing the discussion after localhost ffmpeg suggestion 20:49 MTDiscord Consider: Make the ASCII art a configuration option 20:50 [MTMatrix] enum: none, one line, expanded, one line colorised and expanded colorised 20:50 MTDiscord Theres a colored version? 20:51 [MTMatrix] oh no, what have I done 20:51 ROllerozxa oh my god please do not add a configuration option for something as miniscule as console ASCII art, either keep it or remove 20:51 ROllerozxa it 20:51 MTDiscord who cares 20:51 [MTMatrix] (I think they were ironic) 20:52 MTDiscord I just think a simple on or off would be faster than bikeshedding this 20:52 [MTMatrix] ..oh, you were serious. 20:52 ROllerozxa the fastest would be to keep it and consider the fact it messes up in particular terminals to be wontfix 20:52 erle better features have been axed for less 20:53 [MTMatrix] but you should also consider that, if we change name, someone will have to redo it 20:53 MTDiscord It was generated 20:53 ROllerozxa oh no, someone has to go onto patorjk and regenerate it... 20:54 [MTMatrix] that I didn't know, there's no need to use sarcasm 20:54 ROllerozxa well I'm unsure what ASCII font it uses, though, so that would need to be figured out 20:54 erle zughy do you know about figlet(1)? you can use it to make such logos. http://www.figlet.org/ 20:54 MTDiscord Apparently people exist that like the ascii art, so a simple config option to disable it to fix the --terminal render issue would be the best of both worlds 20:54 MTDiscord and trivial 20:55 MTDiscord https://patorjk.com/software/taag/#p=display&f=Graffiti&t=minetest 20:55 MTDiscord Mind, the default generated version has some stupidities that I fixed, but that is trivial 20:55 [MTMatrix] Well, you know what? I've spent 30 minutes on that PR (20 minutes just to understand how to make the text fancy/useful). Considering the debate is already creating, I'm just gonna close the PR so people can focus on more important things 20:56 MTDiscord haha 20:57 [MTMatrix] done, please back to discuss on godrays, lagging particles, MTG ruining the user first impression and the like. I'm not sarcastic 20:57 MTDiscord closing the PR over a debate? thats a first :^) 20:57 erle what is a godray? 20:57 erle PR? 20:57 MTDiscord sunshine through dust 20:57 erle does the thing have a PR and screenshot? 20:57 MTDiscord no idea 20:58 erle well lhofhansl is not here rn 20:58 [MTMatrix] actually, since we're so talkative. Thoughts on #13874? 20:58 ShadowBot https://github.com/minetest/minetest/issues/13874 -- Too many open issues 20:59 MTDiscord >some projects with over 10k issues be like 21:01 MTDiscord It is worth noting that over a third of the open bugs are audiovisual issues. And we just lost our graphics dev (again) 21:02 ROllerozxa I hope legitimate issues won't be closed just to deflate the issue count 21:02 erle zughy there are a number of approaches to reduce open issues. some projects (like GNOME) go through all issues of a component after it has been refactored and close all issues related to it. some projects have stale bots. both lead to more work, not less. for example, microsoft's stale bot has closed at least 3 isses asking for ed25519 support (the standard key algorithm for openssh) that users opened since 2019. 21:02 ROllerozxa the issue is not that there are a lot of open issues, the issue is that realistically there's not enough persons working on the engine to fix them all 21:02 MTDiscord ^ 21:03 erle what ROllerozxa says 21:03 sfan5 I PR'd a better fix 21:03 MTDiscord sweet 21:09 sfan5 I unfortunately don't have a good suggestion how to deal with so many open issues 21:09 sfan5 closing valid problems is obviously not a workable solution 21:10 sfan5 Minetest could be very good if it had lots of man hours thrown at it 21:10 sfan5 but alas that's not where we are 21:11 erle in my opinion mozilla conclusively proved that merely having open issues is not a problem, unless people are afraid of big numbers. mozilla has bugs that get fixed after 8 or 13 or 18 years. 21:12 erle zughy is there an underlying issue, does having many open issues the triaging more difficult maybe? 21:21 [MTMatrix] Nah, you just label new things and ignote the rest, mostly 21:21 [MTMatrix] *ignore 21:22 MTDiscord As for attracting more contributors, if more people could dig up content for the blog, that would be helpful 21:22 sfan5 that may be correct but what's your point? 21:23 [MTMatrix] Then from time to time you check "possible close" issues and the like, hoping you can actually close something 21:24 [MTMatrix] greenxenith: the blog is kinda dead at the moment, even there we need more contributors 21:24 MTDiscord People other than MisterE can write posts (myself, rubenwardy), its just a matter of finding content 21:26 erle greenxenith i did submit content for the blog and i have no idea, do you need more? i thought it was a monthly thing. no longer? 21:26 MTDiscord MisterE is the main one writing posts, so if he doesnt get to it, it extends to encompass however many months get missed 21:26 MTDiscord I think we have enough stuff from august, mostly just need stuff for september 21:31 MTDiscord Ive sent MisterE a message about current progress, if he cant get to it ill try writing the post 21:32 [MTMatrix] pretty sure rubenwardy should not be involved in the blog, as he's already busy in many other engine aspects, and I don't want to see the guy burning out 21:32 MTDiscord I just said he can, not that he does 21:42 ROllerozxa throwing a thought out there, since a comment on the issue tracker reminded me that minetest's updating UX is still awful on windows because it's distributed as a RUN_IN_PLACE=1 build in a .zip archive. I know there is the WIX installer, but it requires MSVC and the official builds use MinGW, so could we compromise by shipping official RUN_IN_PLACE=0 builds in a .zip file that save user data in AppData instead? separating user 21:42 ROllerozxa and share data would make it easier to update by just throwing away the old install while the user data stays intact in AppData, even if it's not distributed as an installer 21:43 MTDiscord sfan5, I appreciate your recognition of the reality that we just don't have enough man hours to do everything we want to. Thank you for everything you do even with your limited time. 21:44 erle ROllerozxa you would need an update strategy as well for that. does an issue exist? 21:45 MTDiscord Zughy, if you're running out of things to triage, you could always remove the Rebase Needed label from my Catch2 PR. 😉 21:46 sfan5 ROllerozxa: i agree in principle but that has the chance to confuse people more 21:46 sfan5 if we're changing this anyway we should also get rid of the need to extract the ZIP, find 'bin' and click on some random executable 21:47 ROllerozxa yes that's fair 21:48 sfan5 I was thinking perhaps instead of an installer we could NIH some sort of self-extracting executable that 1) uses RUN_IN_PLACE=0 and 2) also hides the actual files somewhere else in %appdata% where people don't see them 21:48 sfan5 alternatively we could see if WIX can me made to work on Linux/MinGW 21:49 sfan5 [NIH solution] that way you'd get 1) upgradeability 2) one-click launch 3) bare files are hidden from users 21:51 sfan5 (the distributed files would go to e.g. %appdata%/minetest-bin/5.8.0/ while path_user is %appdata%/minetest as usual) 21:52 erle i wonder, does anyone here use windows and this package manager? https://community.chocolatey.org/packages/minetest 21:53 ROllerozxa I believe CPack supports generating NSIS installers, which would work when building with MinGW 22:02 ROllerozxa erle: chocolatey is neat but rather niche. Microsoft even has their own "package manager" (glorified installer executable downloader) for windows called winget 22:04 ROllerozxa it can't exactly perform magic with the windows environment but it's a slightly more ergonomic way of installing things without going out and searching for downloads yourself 22:04 erle ROllerozxa oh yeah i remember lol https://www.theverge.com/2020/5/28/21272964/microsoft-winget-windows-package-manager-appget-copied 22:05 ROllerozxa haha yeah 22:05 erle well the question would be if they have figured out the “keep your worlds and savegames” issue 22:07 ROllerozxa well choco's minetest install script just downloads the official portable minetest builds, unzips it and makes a start menu shortcut for it 22:08 ROllerozxa (I think you can tell it's generally designed for proprietary software only shipped in binary form hehe) 22:10 ROllerozxa anyways, bit derailed from the original topic. CPack's NSIS generator is worth looking into as an alternative to WIX 22:11 erle ROllerozxa that sounds like something you could write into an issue 22:16 ROllerozxa yes. maybe someone does it while I'm asleep or I do it tomorrow. either way I'm willing to look into it myself to see if it would work