Time Nick Message 01:40 MTDiscord Won't be a bug until SDL is a feature. And true that would be in 5.11 with new workflow, NOT a concern unless our core devs get release fatigue because the process could be as short as 2 months or less 12:29 rubenwardy I don't think we're going with that release workflow. But if scancodes is to have a chance, a draft pr needs to happen within the first few weeks of 5.10-dev 12:29 rubenwardy I think the workflow we're going with is to just release every 2-3 months without caring what makes it in 13:22 celeron55 i think that's been the unwritten goal since forever, but people can't help themselves getting attached to their favourite features 13:41 MTDiscord and trying to merge a PR and an attached developer takes time 13:43 sfan5 bonus problem: we can't merge anything during freeze 14:09 rubenwardy Could do feature freeze on a different branch. Ideally it shouldn't take more than a week or two 14:09 rubenwardy Entered with way too many regressions this time 17:52 rubenwardy "temporary fix" lol #12608 17:52 ShadowBot https://github.com/minetest/minetest/issues/12608 -- Use legacy image implementation when not using 9-slice image by x2048 17:54 sfan5 I just found out that by github automatically changing the runner the macOS build we produce is now arm64 and not x86_64 like before 17:54 sfan5 I assume we want to still support both? 17:55 sfan5 support as in provide official builds 17:55 [MTMatrix] Isn't MacOS broken anyway? 17:55 sfan5 is it? 17:56 [MTMatrix] I thought it was, I haven't got Macs to check 17:57 [MTMatrix] Maybe I'm wrong 🤷 18:03 sfan5 according to SFENCE in #minetest the current build result doesn't run anyway 18:21 sfan5 -> #14937 18:21 ShadowBot https://github.com/minetest/minetest/issues/14937 -- Revert macOS workflow to run on x86 by sfan5 18:30 SFENCE sfan5: Can I somehow download a built app from #14937 to test a result? 18:30 ShadowBot https://github.com/minetest/minetest/issues/14937 -- Revert macOS workflow to run on x86 by sfan5 18:30 sfan5 https://github.com/minetest/minetest/actions/runs/10307417898?pr=14937 at the bottom 18:33 SFENCE Thanks. I have to be loggen into github to make it works. 18:36 Krock wasn't there a proxy website somewhere that would ease that? 18:36 SFENCE Ok, now results if x86_64 binary, file command says. Can you add a codesign command to github action? 18:36 Krock I remember seeing a link recently 18:39 Krock I'd assume that "codesign --force" uses a computer-specific certificate 18:43 sfan5 Krock: https://nightly.link 18:44 SFENCE Krock: good to know. 18:44 sfan5 I can try adding a codesign invocation 18:45 sfan5 what's the entire command? 18:45 Krock sfan5: definitely bookmarked now. Thanks. 18:50 SFENCE looks like it is litle compicated to do regular codesign. Apple developer ID is required. 19:06 SFENCE I belive I start to understand it better. Fix looks to works well on Mac with M CPU. 19:12 SFENCE Maybe for next release, it is a good idea to find a way to correctly sign minetest.app for Mac users. 19:13 sfan5 sure 19:26 [MTMatrix] Daily reminder, quoting sfan5: "sooo what will be the solution for 5.9.0?" 19:26 [MTMatrix] Problem won't fix itself 19:40 Krock @Zughy so how are we going to fix it? 19:41 Krock actually I could open a forum poll for this. But I need to know what would stop working with USE_SDL=0, or why the forums don't load. 19:42 Krock both wikis are dead too. 19:44 sfan5 nothing really "breaks". we lose some niche features but those didn't exist in 5.8 either 19:45 Krock niche features = touch on desktop builds? 19:46 Krock OpenGL 3 support requires USE_SDL, but there's yet no use for it 19:48 Krock Android requires SDL2, although keyboard input is niche there. 19:59 Krock The Irr win32 and osx devices do not support e.TouchInput at all, but there is some code for the Linux device 20:13 [MTMatrix] @Krock: I'm not a core dev nor I know anything about SDL2, IME and the like :^] I'm just the annoying reminder that we should release 5.9 as soon as possible because it's been 8 months now 20:13 [MTMatrix] all I can see is that this discussion is taking days and it doesn't seem to be a real solution 20:14 Krock well thanks. I am pretty sure the core devs know it at this point. After all, the milestone "overdue" indicator is also present. 20:14 Krock so I'll open a forum post 20:15 [MTMatrix] see? Did my part :^] 20:15 Krock I did not get any input on my idea though 20:16 [MTMatrix] forum works for me btw 20:45 Krock https://forum.minetest.net/viewtopic.php?p=438186#p438186 based on these: https://irc.minetest.net/minetest-dev/2024-07-27#i_6188330 (*) 20:46 Krock (*) where option 2 would need checking whether there's no conflicts when both are enabled. where option 3 turned out to be too less reliable due to more keys, such as +/- and vastly different keyboard layouts. 20:53 MTDiscord another bug which is apparently fixed by not using sdl (but which has not been thoroughly analyzed yet): #14932. 20:53 ShadowBot https://github.com/minetest/minetest/issues/14932 -- Cursor bug, game unplayable 20:55 MTDiscord I believe this bug should be reproducible by using WSL OpenSUSE Tumbleweed with wslg. 20:56 MTDiscord I think I might have also had it in the spring on my old Linux computer running Debian, but I don't have access until end of August or beginning of September to confirm that and get system specs. 22:18 sfan5 Krock: yes 22:18 sfan5 we wouldn't be removing it from android anyway 22:18 sfan5 just the other platforms 22:51 MTDiscord smalljoker/krock: should add native wayland support as a benefit to https://forum.minetest.net/viewtopic.php?p=438186#p438186 22:58 MTDiscord What are the end user implications of native Wayland support though? 23:02 MTDiscord no xwayland usage needed, better controls iirc. idk as all the people who have been asking about native wayland support and using 5.9-dev