💾 Archived View for gemini.ctrl-c.club › ~phoebos › logs › kisslinux-2022-09-21.txt captured on 2023-01-29 at 03:45:30.
-=-=-=-=-=-=-
[2022-09-21T02:03:42Z] <testuser[m]1> rust [2022-09-21T02:04:51Z] <testuser[m]1> illiliti: ok [2022-09-21T02:04:54Z] <ioraff> ha [2022-09-21T02:05:00Z] <testuser[m]1> Hi [2022-09-21T02:08:23Z] <ioraff> hi [2022-09-21T02:18:40Z] <noocsharp> testuser[m]1: i would consider starting kiss-ng from nkiss because a lot of the functionality is already in place [2022-09-21T02:19:14Z] <noocsharp> but i am biased of course [2022-09-21T02:21:34Z] <virutalmachineus> rust kiss!!! [2022-09-21T02:21:49Z] <virutalmachineus> is for security reasons? [2022-09-21T02:26:27Z] <testuser[m]1> noocsharp i prefer to do from scratch cuz I can learn more [2022-09-21T03:29:04Z] <noocsharp> fair [2022-09-21T03:29:10Z] <noocsharp> virutalmachineus: he's joking about rust [2022-09-21T03:40:16Z] <rohan> hey guys [2022-09-21T03:41:28Z] <rohan> mesa build is broken for some reason [2022-09-21T03:41:30Z] <virutalmachineus> <noocsharp> "virutalmachineuser: he's..." <- what? we all know testuser loves rust. he's just joking. [2022-09-21T03:41:32Z] <rohan> heres the log [2022-09-21T03:41:45Z] <rohan> https://termbin.com/qgan [2022-09-21T03:43:59Z] <virutalmachineus> mesa is broken again [2022-09-21T03:44:04Z] <virutalmachineus> that thing is hot mess [2022-09-21T03:46:06Z] <ioraff> sorry. on it [2022-09-21T03:49:08Z] <illiliti> i suppose you just need to null vulkan-drivers option [2022-09-21T03:49:27Z] <illiliti> -Dvulkan-drivers='[]' [2022-09-21T03:50:40Z] <illiliti> https://gitlab.freedesktop.org/mesa/mesa/-/blob/main/meson.build#L685 [2022-09-21T03:51:47Z] <illiliti> rohan: you have amd gpu right? [2022-09-21T04:00:46Z] <rohan> yeah [2022-09-21T04:04:42Z] <ioraff> we could null it by default [2022-09-21T04:06:23Z] <wael[m]> aren't vulkan drivers important [2022-09-21T04:06:37Z] <ioraff> for games [2022-09-21T04:06:46Z] <wael[m]> well, you would still need vulkan headers and loader which don't exist in the kiss repos so [2022-09-21T04:07:00Z] <ioraff> indeed [2022-09-21T04:08:51Z] <wael[m]> it is errroring because of missing glslang, which is opengl thing [2022-09-21T04:10:10Z] <ioraff> yes, but it only tries to find it if vulkan is enabled [2022-09-21T04:11:05Z] <illiliti> so disable vulkan [2022-09-21T04:11:16Z] <illiliti> we don't use or need it anyway [2022-09-21T04:11:25Z] <illiliti> -Dvulkan-drivers='[]' [2022-09-21T04:12:04Z] <wael[m]> true [2022-09-21T04:12:44Z] <illiliti> muon equivalent: -Dvulkan-drivers= [2022-09-21T04:13:02Z] <illiliti> if anyone need it [2022-09-21T04:16:47Z] <ioraff> pushed [2022-09-21T04:20:55Z] <illiliti> thanks [2022-09-21T04:26:54Z] <wael[m]> is there a way to quickly make a patch for a package without having to git init, git add, git commit, git diff in the source dir? [2022-09-21T04:32:34Z] <wael[m]> also in muon, there is no i18n so i cannot build pulseaudio for that [2022-09-21T04:32:42Z] <wael[m]> eg. i18n = import('i18n') [2022-09-21T04:32:53Z] <ioraff> well, you only run git init once. git diff is optional, and you can combine add and commit into commit -a if all the changes in the tree belong to one commit [2022-09-21T04:33:06Z] <wael[m]> i didnt know that was possible [2022-09-21T05:15:12Z] <wael[m]> `/usr/lib/libGLdispatch.so -> libGLdispatch.so.0.0.0c_1248` [2022-09-21T05:15:15Z] <wael[m]> why did muon do this [2022-09-21T05:16:14Z] <wael[m]> `/usr/lib/libEGL.so.1 -> libEGL.so.1.1.0?C?D?Q??D?D?P` [2022-09-21T05:16:15Z] <wael[m]> wtf [2022-09-21T05:17:12Z] <ioraff> https://lists.sr.ht/~lattis/muon/patches/35430 [2022-09-21T05:17:45Z] <illiliti> lattis merged this patch [2022-09-21T05:17:50Z] <illiliti> update muon [2022-09-21T05:18:04Z] <wael[m]> in what commit/version? [2022-09-21T05:18:19Z] <wael[m]> 5 hours ago ok [2022-09-21T05:19:02Z] <wael[m]> yep seems good [2022-09-21T05:25:55Z] <wael[m]> illiliti: sway is failing to build because of glib? https://termbin.com/6ivz [2022-09-21T05:27:59Z] <illiliti> rebuild glib [2022-09-21T05:28:08Z] <wael[m]> with muon? [2022-09-21T05:28:45Z] <illiliti> yes [2022-09-21T05:31:13Z] <illiliti> rebuild everything you build with muon [2022-09-21T05:31:35Z] <illiliti> i guess you have corrupted symlinks [2022-09-21T05:31:52Z] <illiliti> somewhere [2022-09-21T05:32:21Z] <wael[m]> well, i am converting grepo's current meson packages to muon, and sway relies on glib, but the glib muon hasnt been merged in the primary repos so i have to wait for that sorta [2022-09-21T05:33:01Z] <illiliti> sway doesn [2022-09-21T05:33:06Z] <illiliti> 't need glib [2022-09-21T05:33:14Z] <wael[m]> so why is it failing to build whar [2022-09-21T05:33:27Z] <illiliti> because pango need it [2022-09-21T05:33:41Z] <wael[m]> yeah i think i will wait for the muon pr to get merged [2022-09-21T05:34:25Z] <wael[m]> in the meanwhile ill try to get all the muon packages in the pr and test the ones that my system needs [2022-09-21T05:35:19Z] <wael[m]> btw, what about mesa? is it still failing to build with muon? [2022-09-21T05:35:54Z] <illiliti> it builds fine if your gpu don't need llvm [2022-09-21T05:36:03Z] <wael[m]> ..nvidia [2022-09-21T05:36:10Z] <ioraff> amd does too [2022-09-21T05:41:59Z] <illiliti> i'll work on mesa once i fix all regressions and get freetype-harfbuzz to work [2022-09-21T05:42:54Z] <wael[m]> freetype muon works? harfbuzz doesnt? [2022-09-21T05:44:38Z] <illiliti> they are merged into one package [2022-09-21T05:44:56Z] <wael[m]> I know, I was asking if harfbuzz muon doesn't work [2022-09-21T05:45:11Z] <wael[m]> for me freetype muon builds [2022-09-21T05:45:44Z] <illiliti> hm [2022-09-21T05:45:47Z] <ioraff> btw -Dvulkan-drivers= doesn't work [2022-09-21T05:46:06Z] <illiliti> for me second build of freetype is broken [2022-09-21T05:46:26Z] <illiliti> maybe my system is dirty [2022-09-21T05:46:39Z] <illiliti> ioraff: will look into it [2022-09-21T05:46:46Z] <wael[m]> the freetype muon I've tested is a 32-bit package, which harfbuzz never builds in so.. [2022-09-21T06:50:56Z] <wael[m]> about muon missing i18n required for pulseaudio to build, i have attempted to disable it entirely (po) but now im getting missing glib.h error, which is handeled by po [2022-09-21T06:51:02Z] <wael[m]> https://gitlab.freedesktop.org/pulseaudio/pulseaudio/-/blob/master/po/meson.build [2022-09-21T06:51:17Z] <wael[m]> weird [2022-09-21T06:52:58Z] <wael[m]> disabling glib via muon option breaks in https://gitlab.freedesktop.org/pulseaudio/pulseaudio/-/blob/master/po/meson.build [2022-09-21T06:54:48Z] <illiliti> nuke it [2022-09-21T06:54:54Z] <wael[m]> got it sir [2022-09-21T06:54:57Z] <wael[m]> as in remove it [2022-09-21T06:55:42Z] <illiliti> send error log [2022-09-21T06:57:59Z] <wael[m]> nuking glib fixed it [2022-09-21T06:59:01Z] <wael[m]> https://github.com/wael444/gkiss-repo/commit/cad1e9f2e5f04dfe5e292edfc5f0c0af5c83389d [2022-09-21T06:59:54Z] <wael[m]> also, when a feature is used as a boolean (eg. glib=false) muon seems to crash [2022-09-21T07:00:14Z] <wael[m]> now last package is at-spi2 which requires meson GNOME bindings and whatnot [2022-09-21T07:00:21Z] <wael[m]> probably isnt important, ignoring [2022-09-21T07:00:22Z] <illiliti> exit with 1 without output? [2022-09-21T07:00:27Z] <wael[m]> yep [2022-09-21T07:00:38Z] <wael[m]> 'Terminated' is all kiss gives me [2022-09-21T07:00:44Z] <illiliti> i've reported it already [2022-09-21T07:00:53Z] <wael[m]> yay cool [2022-09-21T07:00:57Z] <illiliti> btw [2022-09-21T07:01:20Z] <illiliti> consider sending muon.patch to upstream [2022-09-21T07:01:39Z] <illiliti> it's their bug [2022-09-21T07:02:24Z] <wael[m]> wdym [2022-09-21T07:02:56Z] <wael[m]> booleans are required to not be in quotes, as specified in manuals [2022-09-21T07:03:10Z] <wael[m]> pulseaudio has this already, just those 3 for some reason [2022-09-21T07:03:25Z] <illiliti> booleans must not be in quotes [2022-09-21T07:03:28Z] <wael[m]> OHHHHHHHHHH i see what you mean yeah ur right [2022-09-21T07:03:33Z] <wael[m]> on it [2022-09-21T07:03:46Z] <illiliti> https://gitlab.gnome.org/GNOME/gtk/-/merge_requests/4615 [2022-09-21T07:03:59Z] <wael[m]> 5 months ago wtf [2022-09-21T07:04:26Z] <illiliti> yep, gtk is slow to ship fixes [2022-09-21T07:04:43Z] <illiliti> i bet pulseaudio is slower [2022-09-21T07:04:48Z] <wael[m]> lets find out [2022-09-21T07:06:12Z] <wael[m]> i will be copying the gtk pr's title and description since idk what to write [2022-09-21T07:06:53Z] <illiliti> fine [2022-09-21T07:07:02Z] <wael[m]> and for mangohud as well lol [2022-09-21T07:07:11Z] <illiliti> mangohud? [2022-09-21T07:07:14Z] <illiliti> link? [2022-09-21T07:07:20Z] <wael[m]> oh yeah i forgot i dropped it [2022-09-21T07:07:28Z] <wael[m]> its pretty useless nvm [2022-09-21T07:07:30Z] <illiliti> ok [2022-09-21T07:07:38Z] <wael[m]> https://github.com/flightlessmango/MangoHud here is the link anyway [2022-09-21T07:07:53Z] <wael[m]> its just a overlay for games, i barely even use it anyway, as my gpu already has one built in [2022-09-21T07:10:12Z] <wael[m]> ~~opens a pr for mangohud anyway~~ [2022-09-21T07:26:39Z] <wael[m]> now libglvnd with 32-bit support is failing to build, ...with assembler errors https://termbin.com/tf8b [2022-09-21T07:27:14Z] <wael[m]> other lib32 packages work with muon so hmm [2022-09-21T07:29:10Z] <illiliti> no idea what's wrong [2022-09-21T07:29:40Z] <wael[m]> it is required for lib32 mesa, which is still meson, so i guess i can ignore it and keep meson until mesa gets fixed [2022-09-21T07:29:47Z] <wael[m]> normal libglvnd with muon works however [2022-09-21T07:33:29Z] <illiliti> wait, can meson build 32 libglvnd? [2022-09-21T07:33:36Z] <wael[m]> yep [2022-09-21T07:33:55Z] <illiliti> so it is muon bug, ok ... [2022-09-21T07:34:23Z] <wael[m]> how is it muon bug if 32 libdrm works [2022-09-21T07:34:24Z] <wael[m]> hmm [2022-09-21T07:34:40Z] <illiliti> idk [2022-09-21T07:34:56Z] <illiliti> need to investigate [2022-09-21T07:44:08Z] <testuser[m]1> What does muon even have to do with 32 [2022-09-21T07:45:40Z] <illiliti> maybe it doesn't work well with multilib [2022-09-21T07:45:47Z] <wael[m]> probably [2022-09-21T07:46:16Z] <testuser[m]1> But all it cares about is value of $CC right [2022-09-21T07:52:11Z] <illiliti> nope [2022-09-21T07:53:14Z] <illiliti> there might be a problem with lib32/lib64 split [2022-09-21T07:53:49Z] <illiliti> muon resolves libs using libpkgconf instead of passing -l<lib> to CC [2022-09-21T07:55:07Z] <illiliti> considering that muon doesn't support cross-compilation, i think that multilib is not supported too [2022-09-21T10:06:53Z] <testuser[m]1> https://github.com/kiss-community/repo/issues/94#issuecomment-1253472611 [2022-09-21T11:13:01Z] <phoebos> can someone on musl do `nm /lib/libc.a | grep backtrace` [2022-09-21T11:16:10Z] <phoebos> oh dw [2022-09-21T11:16:32Z] <phoebos> execinfo.h isn't owned by musl [2022-09-21T13:15:49Z] <testuser[m]1> ioraff: why is nproc changed to grep core id in rust [2022-09-21T16:06:13Z] <ioraff> testuser[m]1: nproc isn't posix [2022-09-21T16:16:48Z] <testuser[m]1> ioraff: how posix is core id in cpuinfo [2022-09-21T16:16:57Z] <testuser[m]1> android doesn't seem to have it [2022-09-21T16:17:06Z] <ioraff> good point [2022-09-21T16:17:07Z] <testuser[m]1> linux 4.4 [2022-09-21T16:17:23Z] <testuser[m]1> Not that you'd run kiss on android [2022-09-21T16:17:25Z] <testuser[m]1> But i was just checking [2022-09-21T16:28:20Z] <testuser[m]1> Also with sed -i and install removal [2022-09-21T16:28:33Z] <testuser[m]1> Tons of makefiles and build scripts would still be making use of them [2022-09-21T16:29:40Z] <testuser[m]1> I'm sure about install, not sure how common sed -i is [2022-09-21T16:34:20Z] <ioraff> the file of course isn't posix, but we still avoid requiring a non-posix utility. I suppose I could add '|| printf 1'. [2022-09-21T16:36:28Z] <ioraff> and yes, that's a potential issue. at least we're trying not to contribute to it [2022-09-21T16:41:50Z] <testuser[m]1> ioraff: grep || nproc || 1? [2022-09-21T16:44:03Z] <ioraff> why both grep and nproc? [2022-09-21T16:47:45Z] <testuser[m]1> If the file isnt present [2022-09-21T16:47:45Z] <testuser[m]1> nproc doesn't rely on proc [2022-09-21T17:17:38Z] <ioraff> it's still unportable under the hood [2022-09-21T17:19:02Z] <testuser[m]1> But no harm in checking if it exists and possibly getting correct core count rather than 1