gemini - kennedy.gemi.dev

💾 Archived View for gemini.ctrl-c.club › ~phoebos › logs › kisslinux-2022-11-06.txt captured on 2024-03-21 at 15:51:55.
-=-=-=-=-=-=-
[2022-11-06T03:22:36Z] <wael_> Hi
[2022-11-06T04:22:22Z] <testuser[m]> Hi
[2022-11-06T05:31:46Z] <illiliti> testuser[m]: i'm against meta packages
[2022-11-06T05:32:19Z] <illiliti> if you want to avoid hardcoding a list i have a better idea
[2022-11-06T05:33:40Z] <illiliti> we could make /var/db/kiss/installed repo-aware
[2022-11-06T05:33:58Z] <illiliti> i.e install/copy packages to /var/db/kiss/installed/<repo>/<pkg> instead of /var/db/kiss/installed/<pkg>
[2022-11-06T05:34:12Z] <illiliti> this way we could fetch core packages at runtime
[2022-11-06T05:34:17Z] <illiliti> and exclude them
[2022-11-06T05:35:15Z] <testuser[m]> Hmm seems too intrusive, and now it requires hardcoding the repo "core" instead of the name of a meta pkg. someone could have gnugrep set as their *grep but it's in extra/ repo so
[2022-11-06T05:44:20Z] <illiliti> metapackages install dependencies that i don't want. without provides system they are even more distructive
[2022-11-06T05:45:03Z] <illiliti> i have had a bad time with them on other distros. i would prefer to not have them at all
[2022-11-06T05:45:40Z] <illiliti> let's keep things simple and just document when needed
[2022-11-06T05:46:25Z] <testuser[m]> I mean it would work like by default kiss would have a meta core package installed which includes baselayout, busybox, openssl, musl, gcc and whatever else. If someone switches their coreutils for example, they'll override the meta pkg with deps as toybox, gnugrep etc instead of busybox
[2022-11-06T05:46:31Z] <testuser[m]> But it's clunky without provides
[2022-11-06T05:48:07Z] <testuser[m]> For sandbox alone it could be solved by testing owners of a hardcoded list of posix specified utils but meh
[2022-11-06T05:51:01Z] <wael_> what us a use case for a meta package, just to have many packages bundled together at installation time?
[2022-11-06T05:51:11Z] <wael_> provides system can only get so useful
[2022-11-06T05:52:57Z] <testuser[m]> What
[2022-11-06T05:53:05Z] <testuser[m]> It's for grouping pkgs
[2022-11-06T05:53:39Z] <testuser[m]> Here it's group of pkgs that all pkg assumed are installed at build time
[2022-11-06T05:53:42Z] <testuser[m]> assume*
[2022-11-06T05:54:01Z] <testuser[m]> gcc, openssl, make etc. assumed
[2022-11-06T05:54:12Z] <testuser[m]> Now for a sandbox you only want to allow access to what the pkg explicitly requires
[2022-11-06T05:54:26Z] <testuser[m]> It can't explicitly request gcc or openssl cuz someone could be using clang or libressl
[2022-11-06T05:54:42Z] <testuser[m]> Ignore openssl part, that's mentioned in depfile
[2022-11-06T05:54:46Z] <testuser[m]> say busybox vs toybox
[2022-11-06T05:57:47Z] <illiliti> wait perhaps just having provides system solve this problem?
[2022-11-06T05:58:32Z] <illiliti> one would have to hardcode provides names tho: e.g ssl, cc, make etc
[2022-11-06T05:58:48Z] <testuser[m]> Yeah
[2022-11-06T05:58:51Z] <illiliti> but it still better
[2022-11-06T05:59:22Z] <testuser[m]> But why not have one pkg to provide all those instead of individual
[2022-11-06T05:59:41Z] <testuser[m]> They can be made a bit more general ig, like toolchain, coreutils
[2022-11-06T05:59:50Z] <testuser[m]> But every pkg needs them so i don't see a point
[2022-11-06T05:59:53Z] <testuser[m]> in splitting
[2022-11-06T06:00:31Z] <testuser[m]> Like there's no pkg that needs gcc but doesn't need cp/mv
[2022-11-06T06:03:16Z] <illiliti> because individual is more explicit than opaque meta package?
[2022-11-06T06:05:33Z] <illiliti> look, the fact that we hardcode core packages is a temporary
[2022-11-06T06:05:33Z] <testuser[m]> There should be a limit to explicitness else it'll become a mess
[2022-11-06T06:06:01Z] <illiliti> when we get provides system core packages are no longer explicit
[2022-11-06T06:06:06Z] <illiliti> implicit*
[2022-11-06T06:06:31Z] <testuser[m]> Like if you wanted to go super explicit you could have a provider for every single command
[2022-11-06T06:07:13Z] <testuser[m]> say for the cc package, does that include ld, strip etc
[2022-11-06T06:07:35Z] <testuser[m]> If so then it should be called toolchain
[2022-11-06T06:07:42Z] <testuser[m]> Which introduces some opaquenes
[2022-11-06T06:10:20Z] <illiliti> can't tell now
[2022-11-06T06:10:25Z] <illiliti> we need to have a clear picture which packages should be converted to provides
[2022-11-06T06:12:30Z] <illiliti> then i can tell whether metapackage is viable or not
[2022-11-06T06:13:51Z] <illiliti> maybe there's some way to abuse provides system to implement groups
[2022-11-06T06:14:16Z] <illiliti> if such exists i would prefer it
[2022-11-06T06:16:27Z] <illiliti> btw
[2022-11-06T06:16:47Z] <illiliti> there's already a request to make gmake explicit
[2022-11-06T06:17:52Z] <illiliti> how with your toolchain approach it would look?
[2022-11-06T06:18:43Z] <illiliti> fuck my net
[2022-11-06T06:18:49Z] <illiliti> slow as shit
[2022-11-06T06:21:39Z] <wael_> I wouldn't fuck a router
[2022-11-06T06:21:55Z] <testuser[m]> User would override (or provide) the toolchain meta pkg with deps set to bmake instead of `gmake`, and pkgs that explicitly need gnu features would depend on `gmake` instead of not having any mention of make at all
[2022-11-06T06:22:10Z] <wael_> like base and base-devel?
[2022-11-06T06:26:24Z] <illiliti> if package need posix make, would it still depend on toolchain? assuming that toolchain depends on gmake
[2022-11-06T06:26:46Z] <wael_> maybe a check at build time?
[2022-11-06T06:27:19Z] <wael_> idk anyone who uses posix make
[2022-11-06T06:27:28Z] <wael_> I doubt it would work with most packages a system needs
[2022-11-06T06:27:59Z] <illiliti> i use and strive to write Makefiles compatible with it
[2022-11-06T06:31:36Z] <illiliti> hm wait toolchain could depend on provided make instead of gmake
[2022-11-06T06:31:57Z] <illiliti> and package could explicitly depend on gmake
[2022-11-06T06:32:44Z] <illiliti> sounds like a big headache to implement it properly in dependency resolver
[2022-11-06T06:37:05Z] <illiliti> yeah these nested levels of provides would drive me mad
[2022-11-06T06:39:56Z] <illiliti> does our make depend on make? lol
[2022-11-06T06:40:13Z] <illiliti> https://codeberg.org/kiss-community/repo/src/branch/master/core/make/build#L9
[2022-11-06T06:40:16Z] <illiliti> lol it does
[2022-11-06T06:43:32Z] <testuser[m]> bruh
[2022-11-06T06:54:58Z] <illiliti> aha there's build.sh available
[2022-11-06T06:55:32Z] <testuser[m]> even gcc depends on itself so 
[2022-11-06T06:56:26Z] <illiliti> yeah i mean
[2022-11-06T06:56:40Z] <illiliti> it uses host gcc if available
[2022-11-06T06:56:57Z] <illiliti> if not, it uses bundled one
[2022-11-06T06:57:02Z] <illiliti> so it's fine
[2022-11-06T06:59:36Z] * illiliti still think that weak dependencies is a nice addition for kiss
[2022-11-06T07:10:43Z] <illiliti> testuser[m]: have you applied cherry pick script? does it work for you?
[2022-11-06T07:11:50Z] <illiliti> no rush, just curious
[2022-11-06T07:11:58Z] <testuser[m]> I'll check it now when i get on pc
[2022-11-06T07:12:21Z] <testuser[m]> There's no bundled gcc tho 
[2022-11-06T07:12:24Z] <testuser[m]> Wdym by that
[2022-11-06T07:12:28Z] <testuser[m]> U always need a compiler 
[2022-11-06T07:14:15Z] <illiliti> what does --disable-bootstrap do then?
[2022-11-06T07:14:41Z] <testuser[m]> Prevents multi stage build 
[2022-11-06T07:15:31Z] <testuser[m]> It's to test for miscompilations
[2022-11-06T07:21:10Z] <illiliti> ah right
[2022-11-06T07:27:58Z] <wael_> Beep boop
[2022-11-06T08:08:07Z] <virutalmachineus> hi
[2022-11-06T08:09:19Z] <illiliti> hi
[2022-11-06T09:15:19Z] <testuser[m]> illiliti: works fine
[2022-11-06T09:15:32Z] <testuser[m]> thanks
[2022-11-06T09:23:25Z] <illiliti> nice
[2022-11-06T10:43:04Z] <macslash1[m]> 🐱
[2022-11-06T11:49:59Z] <testuser[m]> The machine is 1000x faster - so a 60 second boot time in 1985 should be 60 ms in 2014
[2022-11-06T12:49:09Z] <testuser[m]> USB-C actually sucks. Government should not decide about things they have no clue.
[2022-11-06T13:11:03Z] <mmatongo> Isn't apple just planning on supporting the port and not the protocol?
[2022-11-06T14:37:03Z] <wael_> can mesa provide GL libraries without libglvnd on AMD/Intel systems?
[2022-11-06T14:37:35Z] <wael_> or am i mistaken
[2022-11-06T14:39:17Z] <wael_> mesa doesnt seem to build libGL tho
[2022-11-06T14:39:46Z] <wael_> yeah i think it might be a X thing not a wayland thing
[2022-11-06T14:43:43Z] <illiliti> wayland uses egl
[2022-11-06T14:43:59Z] <wael_> yeah
[2022-11-06T15:02:25Z] <wael_> how do i make kiss force overwrite etc files
[2022-11-06T15:02:51Z] <wael_> and delete if possible since some files remained
[2022-11-06T15:05:26Z] <illiliti> not possible atm
[2022-11-06T15:10:54Z] <wael_> t
[2022-11-06T15:10:55Z] <wael_> far
[2022-11-06T15:49:30Z] <midfavila> degstop lives :DDD
[2022-11-06T15:49:50Z] <wael_> whats a degstop
[2022-11-06T15:50:05Z] <midfavila> degstop is degstop
[2022-11-06T15:52:11Z] <wael_> but what is a degstop in a degstop is a degstop
[2022-11-06T15:52:17Z] <midfavila> - w-
[2022-11-06T15:52:30Z] <midfavila> i'll tell you when you're older
[2022-11-06T15:54:05Z] <wael_> im old
[2022-11-06T15:54:12Z] <wael_> im so so old
[2022-11-06T15:54:21Z] <midfavila> not old enough kiddo
[2022-11-06T15:54:55Z] <wael_> frick
[2022-11-06T15:56:51Z] <wael_> degstop isnt in any dictionary 
[2022-11-06T15:57:06Z] <wael_> is just literally just locked by reality itself until im older
[2022-11-06T15:59:03Z] <testuser[m]> Desktop
[2022-11-06T16:00:04Z] <wael_> DESKTOP
[2022-11-06T16:00:32Z] <wael_> ok thank you old person
[2022-11-06T16:44:02Z] <saturn[m]> hello everyone
[2022-11-06T16:44:07Z] <saturn[m]> i am having a VERY odd problem
[2022-11-06T16:44:21Z] <saturn[m]> i’m using 500mb of ram without any user processes running
[2022-11-06T16:44:32Z] <saturn[m]> i’ve tried different kernel configurations
[2022-11-06T16:44:39Z] <saturn[m]> three, to be exact, and nothing changes
[2022-11-06T17:00:25Z] <testuser[m]> 500mb without sway
[2022-11-06T17:01:23Z] <saturn[m]> yep
[2022-11-06T17:01:54Z] <wael_> without anything at all
[2022-11-06T17:02:41Z] <wael_> no running process takes this much memory it could be a kernel process
[2022-11-06T17:02:51Z] <testuser[m]>   Check kworkers
[2022-11-06T17:02:56Z] <wael_> ps sort?
[2022-11-06T17:04:10Z] <wael_> yeah sorting by %mem doesnt show the process
[2022-11-06T17:04:10Z] <testuser[m]> top
[2022-11-06T17:04:17Z] <wael_> top
[2022-11-06T17:04:56Z] <wael_> yeah neither does top show a process with much VSZ other than user processes like services which dont take that puch
[2022-11-06T17:06:35Z] <wael_> something in dmesg i did notice (which is the only starting point available) is 'used greatest stack depth: XXXXX bytes left'
[2022-11-06T17:06:49Z] <wael_> the dmesg is a bit interesting 
[2022-11-06T17:07:40Z] <wael_> https://termbin.com/fvg8
[2022-11-06T17:11:53Z] <illiliti> amdgpu
[2022-11-06T17:11:58Z] <illiliti> btrfs
[2022-11-06T17:12:03Z] <illiliti> microcode
[2022-11-06T17:12:07Z] <illiliti> firmware
[2022-11-06T17:12:15Z] <illiliti> it could be anything
[2022-11-06T17:12:34Z] <wael_> i removed firmware for this specific purpose
[2022-11-06T17:12:46Z] <wael_> microcode doesnt exist
[2022-11-06T17:12:57Z] <wael_> me and others use btrfs and memory usage would never get this high
[2022-11-06T17:13:09Z] <illiliti> did you build kernel with builtin modules?
[2022-11-06T17:13:22Z] <wael_> amdgpu is a module 
[2022-11-06T17:13:31Z] <wael_> so that firmwrae stuff doesnt get built into the kernel ig
[2022-11-06T17:14:01Z] <wael_> as for btrfs the 'nocal' option is on
[2022-11-06T17:14:07Z] <wael_> noacl
[2022-11-06T17:14:46Z] <testuser[m]> My server is at 400mb with synapse and postgres running l
[2022-11-06T17:15:04Z] <illiliti> i don't know man
[2022-11-06T17:15:21Z] <illiliti> it is either btrfs or amdgpu or other modules
[2022-11-06T17:15:27Z] <wael_> yeah me neither this is confusing
[2022-11-06T17:15:27Z] <wael_> testuser: atlesat you have a clear idea of what takes memory
[2022-11-06T17:15:43Z] <wael_> dont know how to test that
[2022-11-06T17:16:57Z] <illiliti> try to switch to more basic fs like ext4 or f2fs
[2022-11-06T17:17:19Z] <illiliti> or disable amdgpu if it's possible
[2022-11-06T17:17:26Z] <illiliti> and check
[2022-11-06T17:17:59Z] <illiliti> can i get your lsmod output just in case?
[2022-11-06T17:18:48Z] <testuser[m]> ~~lsmod | xargs rmmod~~
[2022-11-06T17:22:23Z] <illiliti> btw is this problem reproducible on previous kernels?
[2022-11-06T17:22:54Z] <illiliti> might be worth to check if you didn't already
[2022-11-06T17:25:10Z] <wael_> what do you mean on previous
[2022-11-06T17:35:15Z] <illiliti> before 6.0
[2022-11-06T17:48:17Z] <wael_> illiliti: just an offtopic question: what does illiliti mean and where did it come from
[2022-11-06T17:54:06Z] <illiliti> just a dumb nick, nothing more
[2022-11-06T17:57:08Z] <wael_> what noo
[2022-11-06T17:57:14Z] <wael_> it sounds cool that cant be the backstory
[2022-11-06T17:57:24Z] <wael_> you are hiding the cool backstory aren't you
[2022-11-06T17:58:31Z] <illiliti> maybe
[2022-11-06T18:02:13Z] <wael_> give cool backstory
[2022-11-06T18:06:26Z] <wael_> illiliti: you were right its firmware
[2022-11-06T18:06:31Z] <wael_> its now a healthy(?) 136m
[2022-11-06T18:08:13Z] <illiliti> i was playing some online game and came up with this nick
[2022-11-06T18:08:31Z] <illiliti> i don't remember how my brain even managed to come up with it, but here it is
[2022-11-06T18:09:24Z] <wael_> GAME
[2022-11-06T18:09:25Z] <wael_> gamemA GMAE
[2022-11-06T18:09:25Z] <wael_> but what name did you have at the time you were playing the 'some online gmae'
[2022-11-06T18:10:24Z] <illiliti> i had many names
[2022-11-06T18:11:05Z] <illiliti> those were less "original" than this one
[2022-11-06T18:12:27Z] <testuser[m]> What game
[2022-11-06T18:12:39Z] <wael_> yes what game
[2022-11-06T18:13:36Z] <illiliti> browser game
[2022-11-06T18:14:50Z] <illiliti> but that game died already
[2022-11-06T18:15:29Z] <illiliti> it doesn't have same spirit like it was before
[2022-11-06T18:19:23Z] <wael_> what about you testuser 
[2022-11-06T18:27:16Z] <testuser[m]> I just stole it from some guy on frogging family server
[2022-11-06T18:27:36Z] <testuser[m]> That tkg kernel and proton server thingy
[2022-11-06T18:27:47Z] <testuser[m]> before it was abcdefgh...z
[2022-11-06T18:33:53Z] <wael_> what about paratham
[2022-11-06T18:33:59Z] <wael_> that's a kick cool name
[2022-11-06T18:36:01Z] <testuser[m]> Lol
[2022-11-06T18:36:14Z] <testuser[m]> Why do u always butcher it 
[2022-11-06T19:05:47Z] <wael_> I'm not
[2022-11-06T19:05:53Z] <wael_> Its just a cool interesting
[2022-11-06T19:05:57Z] <wael_> Also um
[2022-11-06T19:06:15Z] <wael_> elinks https://termbin.com/ihxj no workie
[2022-11-06T19:11:22Z] <illiliti> ld: no input files
[2022-11-06T19:12:12Z] <testuser[m]> make changes ig 
[2022-11-06T19:12:19Z] <testuser[m]> 4.4
[2022-11-06T19:14:38Z] <wael_> maybe ld broke because rust failed with ld rated error too
[2022-11-06T19:14:50Z] <wael_> termbin.com/9frlu
[2022-11-06T19:15:31Z] <wael_> should I try mold
[2022-11-06T19:16:40Z] <wael_> nvm that won't help
[2022-11-06T19:22:10Z] <testuser[m]> bruh ld has nothing to do with make
[2022-11-06T19:24:46Z] <wael_> ok still my rust and elink broken
[2022-11-06T19:24:52Z] <wael_> hmmmm
[2022-11-06T19:29:22Z] <testuser[m]> https://git.savannah.gnu.org/cgit/make.git/tree/NEWS
[2022-11-06T19:30:49Z] <wael_> o
[2022-11-06T19:30:56Z] <wael_> so rust need new updtae toom?
[2022-11-06T19:31:22Z] <wael_> Also ehrr where ostree gpgme GPGME_PTHREAD becaus ostree explode compile with no GPGME_PTHREAD
[2022-11-06T19:33:08Z] <saturn[m]> ^ https://termbin.com/olat
[2022-11-06T21:24:05Z] <midfavila-mobile> https://xed.sourceforge.net/
[2022-11-06T21:24:21Z] <midfavila-mobile> ^the only proper choice of editor
[2022-11-06T21:24:41Z] <Torr> > The most minimalist editor, written in the most bloated language!
[2022-11-06T21:24:43Z] <Torr> Lol
[2022-11-06T21:25:19Z] <midfavila-mobile> the *only* problem I have with it is being a Tcl/Tk program
[2022-11-06T21:25:38Z] <midfavila-mobile> otherwise, phenomenally comfy
[2022-11-06T21:28:37Z] <Torr> I remember when I 1st programmed in Tcl and discovered its function identifiers could have spaces, and I was like: "WTF"
[2022-11-06T21:31:40Z] <midfavila-mobile> i've never programmed in tcl myself, although i've heard it's a fairly capable language
[2022-11-06T21:33:00Z] <Torr> Indeed, it is.
[2022-11-06T21:56:58Z] <saturn[m]> anyone think they can help me out with compiling `ostree`? i keep getting an error about GPGME_PTHREAD
[2022-11-06T21:57:32Z] <saturn[m]> https://termbin.com/m6qa8
[2022-11-06T23:31:32Z] <midfavila-mobile> https://ssnf.xyz/blog/net-unixs-unofficial-default-network-manager.html
[2022-11-06T23:31:42Z] <midfavila-mobile> ^you guys might find this neat
[2022-11-06T23:31:54Z] <midfavila-mobile> seems very KISS, conceptually
[2022-11-06T23:36:32Z] <Torr> Nice
[2022-11-06T23:36:50Z] <Torr> I remember that guy's Git server implementation.
[2022-11-06T23:37:25Z] <Torr> Even simpler than Cgit.
[2022-11-06T23:37:56Z] <Torr> Tiny and pretty.