gemini - kennedy.gemi.dev

💾 Archived View for gemini.ctrl-c.club › ~phoebos › logs › kisslinux-2022-09-06.txt captured on 2024-05-12 at 16:01:57.
-=-=-=-=-=-=-
[2022-09-06T01:22:58Z] <testuser[m]12> Hi
[2022-09-06T01:27:06Z] <ioraff> hi
[2022-09-06T01:27:23Z] <virutalmachineus> hi
[2022-09-06T07:02:31Z] <wael[m]> hi
[2022-09-06T09:27:49Z] <testuser[m]12> ioraff: are u ok with adding pipewire as a mandatory build (not runtime, dlopen'd) dependency for chromium? it can probably save a few hrs if someone decided that they need  screen sharing
[2022-09-06T09:28:30Z] <testuser[m]12> ofc they'd need dbus and stuff too
[2022-09-06T11:35:36Z] <wael[m]> cant pipewire be built without dbus
[2022-09-06T11:38:24Z] <testuser[m]12> i meant screen casting
[2022-09-06T11:39:39Z] <wael[m]> so screen casting needs dbus 100%
[2022-09-06T11:41:49Z] <testuser[m]12> yeah cuz browsers communicate to abstraction layer over ayyland compositor's cast implementation which also communicates via dbus, via dbus
[2022-09-06T11:42:30Z] <testuser[m]12> u can write another webrtc backend that skips all this abstraction
[2022-09-06T11:42:48Z] <wael[m]> its sad to see that dbus is needed
[2022-09-06T11:42:48Z] <wael[m]> i hope there is a backend yea
[2022-09-06T11:43:01Z] <testuser[m]12> it's not gonna appear by itself lol
[2022-09-06T11:43:32Z] <testuser[m]12> alternatively u can cast to v4l2 camera via wf-recorder and use that as a webcam in firefox
[2022-09-06T11:43:50Z] <testuser[m]12> and do whatever config to pass speaker to mic device
[2022-09-06T11:43:59Z] <testuser[m]12> but then obv u cant use either of those with real defices
[2022-09-06T11:44:52Z] <wael[m]> software is magic it'll appear and save ur life
[2022-09-06T13:09:59Z] <macslash1[m]> hey uh is there any way to test if my pc could realistically compile a kernel, I want my own custom kernel for KISS but I really don't know if my system could compile it
[2022-09-06T13:10:37Z] <raph_ael> macslash1[m]: i compiled my kernels back in 2001, the thing is time
[2022-09-06T13:11:48Z] <macslash1[m]> Yeah true
[2022-09-06T13:12:15Z] <macslash1[m]> also just remembered I'm gonna have to use GKISS for this too
[2022-09-06T13:21:10Z] <ioraff> testuser[m]12: hm. not sure
[2022-09-06T13:21:16Z] <wael[m]> @raph_ael poor guy's 4g ram laptop couldn't compile it in a chroot
[2022-09-06T13:21:26Z] <wael[m]> probably because of systemd oomd moment
[2022-09-06T13:21:38Z] <raph_ael> oh
[2022-09-06T13:22:11Z] <raph_ael> some dev should get rpi to dev and taste most people real life computing feeling
[2022-09-06T13:24:37Z] <ioraff> testuser[m]12: I'm holding off on llvm 15 as rust can't build against it yet
[2022-09-06T13:25:01Z] <macslash1[m]> wael: well I'm compiling on a 16gig ram pc now
[2022-09-06T13:25:26Z] <macslash1[m]> quadcore AMD smth CPU I forget
[2022-09-06T13:29:22Z] <testuser[m]12> ioraff: ok
[2022-09-06T13:30:28Z] <testuser[m]12> a new rust release should be coming in few days anyway so no point in bumping old ver
[2022-09-06T13:30:56Z] <macslash1[m]> rust linux my beloved
[2022-09-06T13:33:06Z] <ioraff> 1.64 is due on 9/22. maybe there'll be a 1.63.1
[2022-09-06T15:52:04Z] * macslash1[m] posted a file: glibc-2022-09-06-15:24-338301 (8305KiB) < https://libera.ems.host/_matrix/media/r0/download/matrix.org/WweIjXqnbkjUpLKKDuBiSXGl >
[2022-09-06T15:52:11Z] <macslash1[m]> Hello, so I'm trying to update glibc 2.36-1 => 2.36-3 and whenever I do it always gets stuck at a line with the error (.text+0xe5e9): undefined reference to `strcpy'
[2022-09-06T15:52:11Z] <macslash1[m]> this is the whole log file
[2022-09-06T15:54:34Z] <testuser[m]12> macslash1: cflags?
[2022-09-06T15:55:06Z] <testuser[m]12> env
[2022-09-06T15:55:55Z] <macslash1[m]> okay weirdly whenever I try `echo $CFLAGS` it doesn't output anything I'll try fix it and retry
[2022-09-06T15:56:21Z] <testuser[m]12> maybe u didn't set any then
[2022-09-06T15:56:21Z] <testuser[m]12> It should work fine
[2022-09-06T15:56:24Z] <testuser[m]12> Hmm
[2022-09-06T15:57:49Z] <macslash1[m]> I added it maybe it'll work now
[2022-09-06T15:57:49Z] <wael[m]> what CFLAGS did you set?
[2022-09-06T15:58:46Z] <macslash1[m]> should be okay on 16 gigs
[2022-09-06T15:58:46Z] <macslash1[m]> -O3 -pipe -march=native
[2022-09-06T15:58:59Z] <wael[m]> yep
[2022-09-06T15:59:14Z] <testuser[m]12> macslash1: can u tell the exact steps u used before rebuilding glibc
[2022-09-06T15:59:17Z] <macslash1[m]> that's probably why it was so damn slow it was all on one core before
[2022-09-06T15:59:36Z] <testuser[m]12> you need to set MAKEFLAGS=-j$(nproc) for slowness
[2022-09-06T16:00:05Z] <macslash1[m]> yeah ik I set it to 8 threads
[2022-09-06T16:00:09Z] <macslash1[m]> 8 is output by nproc so
[2022-09-06T16:02:19Z] <macslash1[m]> testuser[m]12: I just exported CFLAGS, CXXFLAGS, and MAKEFLAGS
[2022-09-06T16:02:19Z] <macslash1[m]> CFLAGS with `export CFLAGS="-O3 -pipe -march=native",
[2022-09-06T16:02:19Z] <macslash1[m]> `export CXXFLAGS=$CFLAGS` and
[2022-09-06T16:02:19Z] <macslash1[m]> `export MAKEFLAGS="-j8"
[2022-09-06T16:02:22Z] <macslash1[m]> arg weird codeblocks
[2022-09-06T16:02:50Z] <macslash1[m]> > <@testuser:matrix.pratham.cc> macslash1: can u tell the exact steps u used before rebuilding glibc
[2022-09-06T16:02:50Z] <macslash1[m]>  * I just exported CFLAGS, CXXFLAGS, and MAKEFLAGS
[2022-09-06T16:02:50Z] <macslash1[m]> CFLAGS with `export CFLAGS="-O3 -pipe -march=native"`, `export CXXFLAGS=$CFLAGS`and`export MAKEFLAGS="-j8"`
[2022-09-06T16:02:56Z] <macslash1[m]> it's working now anyway
[2022-09-06T16:03:58Z] <macslash1[m]> ty btw
[2022-09-06T16:08:33Z] <macslash1[m]>  * ty
[2022-09-06T16:08:45Z] <midfavila> fwiw you should avoid O3
[2022-09-06T16:08:50Z] <midfavila> O2 is enough
[2022-09-06T16:09:07Z] <macslash1[m]> sorry forgot this was an IRC with the reply lmao
[2022-09-06T16:09:07Z] <midfavila> O3 can cause subtle errors because it enables unsafe optimizations
[2022-09-06T16:10:07Z] <macslash1[m]> I see, I'll probably change it then
[2022-09-06T16:10:15Z] <midfavila> probably a good idea
[2022-09-06T16:10:30Z] <midfavila> O3 can also cause binaries to actually be larger and slower
[2022-09-06T16:10:41Z] <midfavila> it enables a lot of weird tricks
[2022-09-06T16:10:58Z] <midfavila> if you want to fiddle with CFLAGS, it's best to read the GCC manual and look at what Red Hat and Debian set
[2022-09-06T16:11:12Z] <macslash1[m]> really?
[2022-09-06T16:11:14Z] <midfavila> yes
[2022-09-06T16:11:25Z] <macslash1[m]> did not know that ty
[2022-09-06T16:11:37Z] <midfavila> O2 with a handful of hardening options is really all you need tbqh
[2022-09-06T16:13:26Z] <midfavila> alternatively if you were a based cool kid like me you would just use a non-optimizing compiler
[2022-09-06T16:13:28Z] <midfavila> :smug:
[2022-09-06T16:15:52Z] <macslash1[m]> lmao
[2022-09-06T16:23:47Z] * macslash1[m] uploaded an image: (16KiB) < https://libera.ems.host/_matrix/media/r0/download/matrix.org/eaYLPKfQoiUwKIVnUwlbxOqf/image.png >
[2022-09-06T16:23:50Z] <macslash1[m]> did not realise this until just now
[2022-09-06T16:25:55Z] <wael[m]> @midfavila : been using O3 ever since I got into kiss
[2022-09-06T16:28:27Z] <testuser[m]12> midfavila: O3 only breaks UB
[2022-09-06T16:28:32Z] <testuser[m]12> Of which there is a lot
[2022-09-06T16:32:07Z] <macslash1[m]> should I get the mainline or stable kernel
[2022-09-06T16:32:09Z] <macslash1[m]> probably stable but
[2022-09-06T16:32:30Z] <macslash1[m]> 6.0-rc4 sounds fun xd
[2022-09-06T16:35:42Z] <midfavila> i use a LTS kernel
[2022-09-06T16:35:59Z] <midfavila> 5.10 with hardening and performance patches
[2022-09-06T16:36:12Z] <midfavila> wael[m] anecdote means nothing in this case
[2022-09-06T16:36:29Z] <midfavila> O3 is known to cause problems and GCC even warns against using it iirc
[2022-09-06T16:36:59Z] <midfavila> i will repeat myself: O2 with hardening flags is usually enough
[2022-09-06T16:37:02Z] <macslash1[m]> midfavila: fair but bleeding edge kernel sounds fun :3
[2022-09-06T16:37:15Z] <midfavila> upstream will enable O3 and other optimizations as needed
[2022-09-06T16:37:27Z] <midfavila> macslash1[m] eeeeeeeeeeeh, you do you
[2022-09-06T16:40:17Z] <macslash1[m]> inb4 my system lasts 2 hours 
[2022-09-06T16:57:16Z] <testuser[m]12> Hmm it seems to be trivial to parse elf and get rpath and NEEDED from dynamic section within 100 loc
[2022-09-06T16:57:16Z] <testuser[m]12> But does library search logic in various libc's differ to such a level that just searching in /usr/lib wouldn't suffice?
[2022-09-06T16:57:18Z] <testuser[m]12> illiliti: 
[2022-09-06T17:04:42Z] <testuser[m]12> I mean the default search path
[2022-09-06T17:04:53Z] <testuser[m]12> not including rpath
[2022-09-06T17:08:35Z] <testuser[m]12> macslash1: https://sourceware.org/bugzilla/show_bug.cgi?id=29454
[2022-09-06T17:08:38Z] <testuser[m]12> Found the bug
[2022-09-06T17:09:34Z] <macslash1[m]> I see, so my system was probably just defaulting to -O1 so when I changed it it fixed the issue
[2022-09-06T17:10:48Z] <testuser[m]12> Yeah
[2022-09-06T17:42:31Z] <wael[m]> midfavila: just curious, what flags do you use?
[2022-09-06T17:42:44Z] <midfavila> none
[2022-09-06T17:42:50Z] <midfavila> i wasn't joking when i said i use a non-optimizing compiler
[2022-09-06T17:43:22Z] <midfavila> although on my current main workstation...
[2022-09-06T17:44:22Z] <midfavila> http://0x0.st/ofqc.txt
[2022-09-06T17:44:29Z] <midfavila> some of these are redundant because they're already enabled by O2
[2022-09-06T17:47:45Z] * macslash1[m] uploaded an image: (19KiB) < https://libera.ems.host/_matrix/media/r0/download/matrix.org/JoFFJztALQEHgSeqvDaeqkBb/image.png >
[2022-09-06T17:47:56Z] <macslash1[m]> I finally got a screenie with max on all cores lmao
[2022-09-06T17:48:12Z] <midfavila> >he only has eight cores
[2022-09-06T17:48:18Z] <macslash1[m]> can u tell I'm compiling a kernel
[2022-09-06T17:48:20Z] <midfavila> ara ara
[2022-09-06T17:48:25Z] <wael[m]> no
[2022-09-06T17:48:31Z] <macslash1[m]> nerd smh
[2022-09-06T17:49:15Z] <macslash1[m]> ik
[2022-09-06T17:49:28Z] * macslash1[m] uploaded an image: (5KiB) < https://libera.ems.host/_matrix/media/r0/download/matrix.org/zyWChMbIwmcXKheWolsGZpsn/image.png >
[2022-09-06T17:49:30Z] <midfavila> do you use a swap file or partition?
[2022-09-06T17:49:32Z] <macslash1[m]> I need an upgrade
[2022-09-06T17:49:38Z] <midfavila> and is it on a hard disk or SSD?
[2022-09-06T17:49:44Z] <macslash1[m]> oh shit it finished compiling
[2022-09-06T17:51:41Z] <macslash1[m]> partition and HDD, I also need to get an SSD at some point
[2022-09-06T17:51:55Z] <midfavila> did you make sure your swap partition was the first one?
[2022-09-06T17:51:59Z] <macslash1[m]> lol
[2022-09-06T17:52:24Z] <midfavila> it's important = w=
[2022-09-06T17:54:41Z] <macslash1[m]> Uh im not sure, cant fix it rn im about to eat
[2022-09-06T17:54:52Z] <midfavila> np
[2022-09-06T17:55:24Z] <midfavila> anyway yeah for future reference: keep in mind that the lower a track is, the higher its angular velocity and the lower the latency to seek is
[2022-09-06T17:55:43Z] <midfavila> so by placing your swap partition at the start of your partition table, you minimize the time necessary to complete a full rotation of the track(s) it's on
[2022-09-06T17:55:50Z] <midfavila> thereby increasing performance when swapping
[2022-09-06T17:56:27Z] <midfavila> (you might have already known that though idk)
[2022-09-06T18:02:22Z] <testuser[m]12> -fdelete-null-pointer-checks sounds sussy
[2022-09-06T18:02:36Z] <midfavila> iirc that's one of the redundant flags enabled by O2
[2022-09-06T18:05:23Z] <midfavila> yup
[2022-09-06T18:05:25Z] <midfavila> it totally is
[2022-09-06T18:05:43Z] <midfavila> https://gcc.gnu.org/onlinedocs/gcc/Optimize-Options.html
[2022-09-06T18:13:34Z] <ioraff> kernel patch to remove the bash dependency got merged. woooo
[2022-09-06T18:14:10Z] <midfavila> based and posixpilled
[2022-09-06T18:17:20Z] <macslash1[m]> ioraff: WHAT I literally just had to install bash to compile my kernel
[2022-09-06T18:17:24Z] <macslash1[m]> 😭
[2022-09-06T18:17:39Z] <macslash1[m]> ughhhh
[2022-09-06T18:18:28Z] <ioraff> it might be in a later version of 6.0. for now, it's only in the kbuild branch
[2022-09-06T18:18:45Z] <macslash1[m]> ah I see
[2022-09-06T18:18:54Z] <ioraff> https://0x0.st/o2_n.patch
[2022-09-06T18:20:20Z] <macslash1[m]> 6.0 is shaping up to a quite nice kernel update
[2022-09-06T18:20:33Z] <midfavila> doesn't it have rust?
[2022-09-06T18:20:39Z] <macslash1[m]> yes.
[2022-09-06T18:20:51Z] <midfavila> i'm sticking with 5.10 = w=
[2022-09-06T18:20:55Z] <midfavila> hBSD can't come soon enough
[2022-09-06T18:20:56Z] <macslash1[m]> :)
[2022-09-06T18:23:31Z] <ioraff> I don't think that patchset was merged yet
[2022-09-06T18:23:53Z] <testuser[m]12> midfavila: h what
[2022-09-06T18:24:04Z] <midfavila> HyperbolaBSD
[2022-09-06T18:24:20Z] <macslash1[m]> I wonder how many will actually move to BSD because of rust in the linux kernel
[2022-09-06T18:24:25Z] <midfavila> probably not many
[2022-09-06T18:24:34Z] <ioraff> probably 0
[2022-09-06T18:24:44Z] <testuser[m]12> 0 unless it's hard dep for common modules
[2022-09-06T18:24:51Z] <macslash1[m]> Personally I don't see an issue in it and a version of the gnu coreutils has been created in rust and it's actually faster than the current implementation
[2022-09-06T18:24:59Z] <testuser[m]12> Few decades left for that
[2022-09-06T18:25:03Z] <midfavila> that's irrelevant, macslash1[m]
[2022-09-06T18:25:23Z] <macslash1[m]> idk but I've seen some people who really hate rust and I can't understand it
[2022-09-06T18:25:38Z] <midfavila> rust is a poorly-defined and non-portable language with licensing flaws and really shitty build UX
[2022-09-06T18:25:43Z] <midfavila> do they even have a spec yet?
[2022-09-06T18:25:49Z] <midfavila> or is it still "the compiler is the spec"?
[2022-09-06T18:26:13Z] <midfavila> is the opinion on PPC still "get fucked nerds"?
[2022-09-06T18:26:28Z] <midfavila> do you still need to either rely on binaries or bootstrap literally years' worth of compilers?
[2022-09-06T18:26:29Z] <midfavila> etc
[2022-09-06T18:26:37Z] <midfavila> these are all massive issues
[2022-09-06T18:26:45Z] <testuser[m]12> Firefox constantly needs small patches for rust breakage every 2 major releases
[2022-09-06T18:26:57Z] <midfavila> and all for what - so people can think a little less when writing programs?
[2022-09-06T18:27:06Z] <macslash1[m]> midfavila: huh.
[2022-09-06T18:27:35Z] <midfavila> like don't get me wrong, C is a pain in the ass
[2022-09-06T18:27:42Z] <midfavila> the syntax is obtuse compared to Scheme
[2022-09-06T18:27:52Z] <midfavila> and all the bullshit people do with pointers and casting is sinful
[2022-09-06T18:27:53Z] <macslash1[m]> Scheme?
[2022-09-06T18:28:02Z] <midfavila> minimalist Lisp implementation
[2022-09-06T18:28:14Z] <midfavila> very portable, very efficient, very easy to understand, and very easy to use
[2022-09-06T18:28:27Z] <midfavila> but, it also doesn't really have a standard library
[2022-09-06T18:28:35Z] <midfavila> (they're working on that though)
[2022-09-06T18:29:04Z] <macslash1[m]> sounds cool
[2022-09-06T18:29:10Z] <midfavila> it's very cool
[2022-09-06T18:29:28Z] <macslash1[m]> ohhh the (((((((()))))))))
[2022-09-06T18:29:29Z] <midfavila> although some of the stuff you can do with it is a little mind-bending at first
[2022-09-06T18:29:34Z] <macslash1[m]> I forgot about that
[2022-09-06T18:29:46Z] <midfavila> and yes, it uses parentheses for its primary punctuation
[2022-09-06T18:29:58Z] <macslash1[m]> lol, a little too much if you ask me
[2022-09-06T18:30:48Z] <midfavila> compared to <>{}[]();+-=/*==|&~||&&! and so on and so forth?
[2022-09-06T18:31:24Z] <macslash1[m]> oh yeah shellscript is weird like that I don't know too much about it
[2022-09-06T18:31:36Z] <midfavila> C has a bunch of grammatical rules to memorize that can act in unusal ways and the language itself for better or worse forces you to think about implementation more than the program itself
[2022-09-06T18:31:44Z] <midfavila> shell, I don't even care about
[2022-09-06T18:31:45Z] <midfavila> i use rc
[2022-09-06T18:31:54Z] <midfavila> but, point is,
[2022-09-06T18:32:13Z] <midfavila> replacing C with "C but handholding" doesn't really improve the situation much
[2022-09-06T18:32:18Z] <midfavila> and by adding a bunch of caveats to rust,
[2022-09-06T18:32:23Z] <midfavila> it's overall a much worse situation
[2022-09-06T18:32:24Z] <midfavila> like,
[2022-09-06T18:32:27Z] <midfavila> i could write in rust,
[2022-09-06T18:32:52Z] <midfavila> or i could write a program in scheme and transpile it to C and know that it'll be as fast or faster than rust, yet have all the niceties of scheme
[2022-09-06T18:33:31Z] <midfavila> or, if I needed something that would absolutely *not* fuck up *ever*, I could use haskell, or if I was concerned about space, FORTH... there are a bunch of languages that do what rust is trying to do but better
[2022-09-06T18:33:42Z] <midfavila> and they also don't have shitty communities
[2022-09-06T18:34:03Z] <midfavila> and they're also, you know, portable to more than x86, x86-64 and ARM
[2022-09-06T19:05:31Z] <ioraff> testuser[m]12: instead of shipping a pc file with llvm, should we instead try to add llvm-config support to muon?
[2022-09-06T19:11:02Z] <ioraff> tt