💾 Archived View for gemini.ctrl-c.club › ~phoebos › logs › kisslinux-2022-08-30.txt captured on 2023-01-29 at 03:45:54.
-=-=-=-=-=-=-
[2022-08-30T00:29:54Z] <midfavila> anyone get docbook set up?
[2022-08-30T00:30:04Z] <midfavila> need it to build Xaw's documentation (because ofc)
[2022-08-30T00:41:03Z] <phoebos> midfavila: it's compiled to various stuff at https://x.org/releases/X11R7.7/doc/libXaw/
[2022-08-30T00:46:47Z] <midfavila> erm, maybe i'm missing something, but they only seem to have the XML source and the libXaw paper
[2022-08-30T00:51:33Z] <midfavila> could probably pull docs out of an old Xaw release actually...
[2022-08-30T01:01:20Z] <phoebos> what docs did you want
[2022-08-30T01:02:24Z] <midfavila> those, but in, you know, roff format
[2022-08-30T01:02:40Z] <midfavila> as far as I'm aware docbook exists to translate from XML-based documentation to other formats
[2022-08-30T01:03:44Z] <phoebos> similar to roff
[2022-08-30T01:03:52Z] <phoebos> good luck :p
[2022-08-30T01:04:11Z] <phoebos> i don't think that documentation is good enough to be worth reading like manpages
[2022-08-30T01:04:51Z] <midfavila> maybe
[2022-08-30T01:05:08Z] <midfavila> just sucks that I'm gonna have to page through the Athena paper every time I want a preview of something
[2022-08-30T01:05:15Z] <midfavila> could always cut stuff out from that, too...
[2022-08-30T01:06:40Z] <phoebos> it's a bit of a PITA
[2022-08-30T01:07:14Z] <phoebos> for example, a note to self: XawTextReplace()'s widget needs to be the parent of the textSource, not the source itself. Otherwise it just segfaults.
[2022-08-30T01:07:26Z] <phoebos> not documented in the paper or in the source code
[2022-08-30T01:08:45Z] <phoebos> that i could find
[2022-08-30T01:09:13Z] <midfavila> hrm
[2022-08-30T01:49:28Z] <midfavila> working on a twm package rn, it's pretty comfy
[2022-08-30T01:49:44Z] <midfavila> could see myself dailying it
[2022-08-30T02:00:31Z] <noocsharp> anyone know what in firefox requires perl to build?
[2022-08-30T02:02:27Z] <testuser[m]12> nss has 2 scripts and 5-6 more. In others
[2022-08-30T02:02:34Z] <testuser[m]12> Shouldn't be hard to get rid of
[2022-08-30T02:02:36Z] <testuser[m]12> Ho
[2022-08-30T02:02:38Z] <testuser[m]12> Hi
[2022-08-30T02:04:14Z] <noocsharp> hi
[2022-08-30T02:08:42Z] <lylac> hi o/
[2022-08-30T02:14:09Z] <lylac> anyone have CPU usage issues with Firefox? its reporting the networking task to be using significant CPU power, as in 10-13% of a 5950x significant
[2022-08-30T03:26:48Z] <noocsharp> testuser[m]12: do you know how to detect when perl is run as part of the build?
[2022-08-30T03:49:10Z] <midfavila> greetings, fellow children of the lamb
[2022-08-30T07:29:02Z] <testuser[m]12> noocsharp: wdym detect
[2022-08-30T07:29:02Z] <testuser[m]12> Like dynamically or what
[2022-08-30T08:16:49Z] <vouivre> Hi
[2022-08-30T08:17:54Z] <vouivre> testuser[m]12: are you online ?
[2022-08-30T08:21:00Z] <testuser[m]12> vouivre: no
[2022-08-30T08:22:40Z] <vouivre> ok, I had a look about libgphoto2, I'll make a comment in the issue "Outdated packages"
[2022-08-30T08:27:17Z] <testuser[m]12> non dbus notifications for ayyland?
[2022-08-30T08:40:38Z] <wael[m]> (herbe wayland alternative here)
[2022-08-30T08:45:38Z] <vouivre> wael[m]: what is the alternative ?
[2022-08-30T08:48:34Z] <wael[m]> *just imagine it
[2022-08-30T08:51:49Z] <vouivre> ?????
[2022-08-30T08:52:02Z] <vouivre> you mean there is no alternative ?
[2022-08-30T08:57:43Z] <wael[m]> there maybe is idk
[2022-08-30T08:58:51Z] <wael[m]> @testuser :, why do you need dbus-less notifications?
[2022-08-30T08:59:46Z] <testuser[m]12> i need
[2022-08-30T09:01:48Z] <vouivre> https://github.com/Vixeliz/Wayherb
[2022-08-30T09:02:10Z] <vouivre> wayland alternative
[2022-08-30T09:06:24Z] <wael[m]> Bloody knew it
[2022-08-30T09:09:49Z] <testuser[m]12> its unmaintained
[2022-08-30T09:19:35Z] <vouivre> on my system it works
[2022-08-30T09:20:15Z] <vouivre> that's right, that's not perfect
[2022-08-30T10:02:09Z] <illiliti> framework for device manager sounds like a nice idea but hard to get right
[2022-08-30T10:02:26Z] <illiliti> because framework implies that device manager need to fork/exec scripts. that's where things get out of control
[2022-08-30T10:03:00Z] <illiliti> fork/exec has cost and can easily fork-bomb your system
[2022-08-30T10:04:25Z] <illiliti> that's why people abandoned legacy /proc hotplugger
[2022-08-30T10:07:18Z] <illiliti> and switched to udev which doesn't have a problem of this magnitude
[2022-08-30T10:09:01Z] <illiliti> but udev has other problems
[2022-08-30T10:09:36Z] <illiliti> it has fucking goto in its rules!
[2022-08-30T10:13:35Z] <illiliti> i very often see bug reports where people report that udev consumes all cpu due to some dumb rule which uses goto
[2022-08-30T10:14:20Z] <illiliti> ok, back to the topic
[2022-08-30T10:15:20Z] <illiliti> if you have idea how to avoid fork/exec while still keeping modularity, let me know
[2022-08-30T10:19:21Z] <testuser[m]12> what about multiple mdev.conf's
[2022-08-30T10:22:37Z] <illiliti> mdevd author against this idea
[2022-08-30T10:22:46Z] <illiliti> if you meant /etc/mdev.d/*.conf
[2022-08-30T10:23:09Z] <testuser[m]12> yeah
[2022-08-30T10:23:10Z] <testuser[m]12> why
[2022-08-30T10:23:43Z] <illiliti> he says it's already possible via wrapper
[2022-08-30T10:24:16Z] <testuser[m]12> what wrapper
[2022-08-30T10:24:34Z] <testuser[m]12> example
[2022-08-30T10:24:48Z] <illiliti> one could combine all configs into one and pass to mdevd
[2022-08-30T10:25:44Z] <illiliti> to implement this wrapper is needed which is meh
[2022-08-30T10:25:52Z] <illiliti> i don't like wrappers
[2022-08-30T10:25:57Z] <testuser[m]12> isnt it less effort than framework
[2022-08-30T10:27:36Z] <illiliti> wdym?
[2022-08-30T10:28:01Z] <testuser[m]12> pkgs can just install one liners to /etc/mdev.conf.d or whatever
[2022-08-30T10:29:57Z] <illiliti> yeah
[2022-08-30T10:30:14Z] <illiliti> we can agree that mdev config is a standard
[2022-08-30T10:30:57Z] <illiliti> instead of implementing full-blown framework
[2022-08-30T10:32:57Z] * illiliti thinking about freebsd devd config
[2022-08-30T10:38:08Z] <illiliti> but, the point of framework is modularity and hardcoding some specific format defeats it..
[2022-08-30T10:39:08Z] <illiliti> look at kiss's build files
[2022-08-30T10:39:30Z] <illiliti> you can write them in python btw if you like
[2022-08-30T10:39:48Z] <illiliti> because kiss doesn't hardcode specific format, i.e shell
[2022-08-30T10:40:20Z] <testuser[m]12> hmm
[2022-08-30T10:40:38Z] <illiliti> how we can achieve that without fork/exec?
[2022-08-30T10:41:18Z] <illiliti> it's impossible i think
[2022-08-30T10:42:01Z] <testuser[m]12> but we would have to implement support for generic framework in every device managerrrr
[2022-08-30T10:42:03Z] <testuser[m]12> manager
[2022-08-30T10:43:11Z] <illiliti> perhaps
[2022-08-30T10:46:45Z] <illiliti> i think the best option is to agree that mdev config is a standard and implement wrappers for mdev/mdevd to support .d
[2022-08-30T10:46:59Z] <testuser[m]12> yewah
[2022-08-30T10:47:43Z] <illiliti> we have to accept tradeoffs if we are going to implement framework
[2022-08-30T10:48:15Z] <testuser[m]12> what impl other than eudev and mdev exists anyway
[2022-08-30T10:50:17Z] <illiliti> considering how easy is to implement device manager, there are quite a lot
[2022-08-30T10:51:06Z] <illiliti> oasis hotplugd
[2022-08-30T10:51:21Z] <illiliti> ndev by TAAPArthur
[2022-08-30T10:51:38Z] <illiliti> smdev
[2022-08-30T11:16:26Z] <illiliti> people at alpine thinking how to translate udev rules to mdev configs
[2022-08-30T11:16:57Z] <illiliti> it's pretty hard to do because udev has goto
[2022-08-30T11:24:13Z] <illiliti> fuck, who decided to put goto into it
[2022-08-30T11:25:02Z] <illiliti> the all thing looks like an intentional vendor lock
[2022-08-30T11:42:10Z] <testuser[m]12> https://github.com/M0Rf30/android-udev-rules
[2022-08-30T11:44:47Z] <illiliti> "distro-agnostic"
[2022-08-30T12:29:09Z] <testuser[m]12> } elsif (($predefined_C{__GNUC__} // -1) >= 3
[2022-08-30T12:29:09Z] <testuser[m]12> && !($predefined_C{__APPLE_CC__} && !$predefined_C{__clang__})) {
[2022-08-30T12:29:10Z] <testuser[m]12> wtf is this language
[2022-08-30T12:32:29Z] <wael[m]> gnu
[2022-08-30T12:43:53Z] <illiliti> ruby?
[2022-08-30T12:58:56Z] <testuser[m]12> perl
[2022-08-30T13:05:44Z] <illiliti> where is use strict; then? xd
[2022-08-30T13:16:43Z] <testuser[m]12> why isnt make setting $RANLIB
[2022-08-30T13:18:55Z] <illiliti> why would it
[2022-08-30T13:19:34Z] <illiliti> ranlib is deprecated
[2022-08-30T13:38:15Z] <testuser[m]12> illiliti: $(AR) s $file is equivalent?
[2022-08-30T13:40:42Z] <illiliti> yep
[2022-08-30T13:45:37Z] <testuser[m]12> illiliti: https://github.com/kiss-community/openssl
[2022-08-30T14:22:31Z] <testuser[m]12> https://github.com/kiss-community/openssl/releases/tag/generated
[2022-08-30T14:25:48Z] <wael[m]> testuser: whats that for?
[2022-08-30T14:31:15Z] <testuser[m]12> pregenerating files that require perl
[2022-08-30T14:31:47Z] <wael[m]> wait is perl just not liked here or something
[2022-08-30T14:37:41Z] <illiliti> reducing the amount of crap needed to have usable system is the main goal of kiss
[2022-08-30T14:37:52Z] <illiliti> perl is not an exception
[2022-08-30T14:37:59Z] <illiliti> as well as python
[2022-08-30T14:39:46Z] <illiliti> and everything else which deemed to be unnecessary
[2022-08-30T14:41:16Z] <testuser[m]12> https://github.com/kiss-community/repo/pull/84
[2022-08-30T14:41:57Z] <illiliti> nice
[2022-08-30T14:43:20Z] <wael[m]> but what if another program i use needs it
[2022-08-30T14:43:22Z] <wael[m]> would tha make it an exception
[2022-08-30T14:43:42Z] <illiliti> no
[2022-08-30T14:44:00Z] <illiliti> package it locally or to community repo
[2022-08-30T14:44:30Z] <illiliti> main repo has bare minimum to have usable system
[2022-08-30T14:44:54Z] <midfavila> hmmmmmmmmmmmmmmmmmmmmmmmmmm
[2022-08-30T14:45:02Z] <midfavila> i bet you could bootstrap gcc with tcc
[2022-08-30T14:45:11Z] <midfavila> maybe kiss should switch :^)
[2022-08-30T14:47:04Z] <illiliti> if cproc can bootstrap gcc, then no doubt tcc can as well
[2022-08-30T14:47:15Z] <illiliti> gcc 4 to be exact
[2022-08-30T14:49:14Z] <illiliti> compiling everything with tcc would be amazing, but we are not in ideal world sadly
[2022-08-30T14:49:41Z] <illiliti> crapware such as firefox and chromium makes it impossible
[2022-08-30T14:50:00Z] * midfavila laughs in Links2
[2022-08-30T14:50:36Z] <illiliti> meh
[2022-08-30T14:50:48Z] <illiliti> i did use links before
[2022-08-30T14:50:58Z] <illiliti> it was pita
[2022-08-30T14:52:50Z] <testuser[m]12> netsurf
[2022-08-30T14:53:45Z] <sad_plan> cant even get netsurf to build on my on laptop
[2022-08-30T14:54:07Z] <illiliti> when i retire, i'll spent my time working on a sane browser till last day
[2022-08-30T14:54:13Z] <midfavila> netsurf is *actually* garbage
[2022-08-30T14:54:32Z] <midfavila> i mean, i appreciate that they have a teeny development team and all
[2022-08-30T14:54:33Z] <sad_plan> how so?
[2022-08-30T14:54:40Z] <midfavila> it literally just doesn't work
[2022-08-30T14:54:47Z] <midfavila> like, it doesn't render ISO15445 HTML properly
[2022-08-30T14:55:10Z] <sad_plan> hm
[2022-08-30T14:56:35Z] <wael[m]> sad_plan: have you gotten toybox init/mdev/otherstuff working? im currently getting toybox to replace busybox in my system
[2022-08-30T14:56:58Z] <illiliti> toybox mdev is incomplete afaik
[2022-08-30T14:59:25Z] <sad_plan> never tried toybox mdev. but no on init. I was going to ask dilyn, be he's not here as often. other stuff, for the most part, yes
[2022-08-30T14:59:35Z] <illiliti> > Some man page symlinks rely on perl
[2022-08-30T14:59:40Z] <illiliti> testuser[m]12: what does it mean? do i still need perl installed?
[2022-08-30T15:01:05Z] <testuser[m]12> illiliti: no
[2022-08-30T15:01:23Z] <testuser[m]12> Symlinks require perl so we just don't make them
[2022-08-30T15:01:35Z] <testuser[m]12> They rely on parsing man page
[2022-08-30T15:01:36Z] <testuser[m]12> there's thousands of them though
[2022-08-30T15:03:26Z] <sad_plan> wael[m]: I have an issue about it on my repo, explaining some of the ..crufts on some of toyboxes utilities.
[2022-08-30T15:03:40Z] <sad_plan> wael[m]: https://github.com/hovercats/kiss-dumpsterfire/issues/8
[2022-08-30T15:04:30Z] <sad_plan> its by no means complete im sure, but it atleast explains some of the wierdness you might encounted when switching to toybox
[2022-08-30T15:15:12Z] <wael[m]> mfw toybox adds features i really want in busybox but breaks everything else
[2022-08-30T15:15:31Z] <wael[m]> ill look at busybox config and hope it has the features i need
[2022-08-30T15:15:38Z] <sad_plan> lol. what feature might that be? I find toybox to be way less complete, and lack features one would find elsewhere
[2022-08-30T15:18:04Z] <wael[m]> well for 1, ps -k/--sort sort
[2022-08-30T15:18:26Z] <wael[m]> i have psutils or whatever it was just for this specific flag
[2022-08-30T15:18:35Z] <sad_plan> hm
[2022-08-30T15:18:57Z] <midfavila> >he uses non-POSIX extensions
[2022-08-30T15:19:00Z] <midfavila> reeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeee
[2022-08-30T15:19:08Z] <sad_plan> oh noo
[2022-08-30T15:19:27Z] <wael[m]> if busybox has chattr, then chattr as well
[2022-08-30T15:19:32Z] <sad_plan> run midfavila
[2022-08-30T15:19:38Z] <wael[m]> xxd, iotop
[2022-08-30T15:19:49Z] <wael[m]> i believe toybox has a dhcp client so, dhcpd
[2022-08-30T15:19:56Z] <midfavila> xxd isn't even standard on unix
[2022-08-30T15:20:00Z] <midfavila> but you can get it separately
[2022-08-30T15:20:07Z] <wael[m]> s/toybox/busybox
[2022-08-30T15:20:09Z] <midfavila> it's got its own package under community i think
[2022-08-30T15:20:26Z] <sad_plan> I belive there is some dhcp in toybox, but I dont belive its done. the busybox one doesn work properly either iirc
[2022-08-30T15:20:30Z] <wael[m]> i will convert you to non-POSIX
[2022-08-30T15:20:31Z] <wael[m]> run for your life muhahahha
[2022-08-30T15:20:56Z] <midfavila> i'll kill you in minecraft
[2022-08-30T15:21:04Z] <midfavila> swear to god
[2022-08-30T15:21:20Z] <midfavila> if I see one more non-posix part of your userland i'm hiring a bitcoin assassin
[2022-08-30T15:21:23Z] <sad_plan> iirc sabotage linux has some scripts to use with busybox's builtin dhcp client, but ive never gotten it to work. not that Ive taken a whole lot of time on it though
[2022-08-30T15:21:26Z] <wael[m]> i use non-posix stuff for my own stuff
[2022-08-30T15:21:36Z] <sad_plan> > bitcoin-assassin
[2022-08-30T15:21:48Z] <wael[m]> pvp me
[2022-08-30T15:21:48Z] <wael[m]> whats the server
[2022-08-30T15:22:31Z] <midfavila> lmao why would I pvp you personally, I might lose that way
[2022-08-30T15:22:40Z] <midfavila> much better to get my legions of assassins to pvp you
[2022-08-30T15:22:50Z] <wael[m]> i just like convenience in my stuff!! i try to go POSIX on stuff i believe others would look at
[2022-08-30T15:23:02Z] <midfavila> posix isn't even that inconvenient tho
[2022-08-30T15:23:35Z] <wael[m]> yes
[2022-08-30T15:23:36Z] <wael[m]> do it
[2022-08-30T15:24:17Z] <wael[m]> just certain flags
[2022-08-30T15:25:13Z] <midfavila> do any of you know of a decent standalone cpp?
[2022-08-30T15:26:29Z] <illiliti> i'm interested too
[2022-08-30T15:27:38Z] <midfavila> alternatively, hrm
[2022-08-30T15:27:45Z] <noocsharp> http://mcpp.sourceforge.net/
[2022-08-30T15:27:48Z] <midfavila> it'd be neat if you could implement a cpp on top of m4
[2022-08-30T15:27:52Z] <midfavila> oop, thanks noocsharp
[2022-08-30T15:28:01Z] <midfavila> quick search didn't pull anything up other than SO questions
[2022-08-30T15:28:13Z] <noocsharp> never used it, but there was a discussion in #oasis about c preprocessors yesterday
[2022-08-30T15:29:02Z] <midfavila> hmm
[2022-08-30T15:29:13Z] <midfavila> well, it looks like it'd implement everything necessary for tcc
[2022-08-30T16:29:50Z] <testuser[m]12> Has anyone gotten iwd or anything that isn't wpa supplicant (with -Dwext) to work with r8188eu?
[2022-08-30T19:04:50Z] <testuser[m]12> does anyone know the rationale for bundling libepoxy in gtk and nss/nspr in firefox
[2022-08-30T19:06:17Z] <testuser[m]12> Making nss nspr seperate can save 5 mins in firefox build, and libepoxy is required by other stuff sometimes like chromium if building with pipewire support
[2022-08-30T19:15:06Z] <sad_plan> isnt the general rule of if something is needed elsewhere, dont bundle it?
[2022-08-30T23:12:16Z] <ioraff> testuser[m]12: probably just because dylan wanted less packages. I'm definitely in favor of separating nss, nspr, and libepoxy. libretls, too, for that matter.