💾 Archived View for gemini.ctrl-c.club › ~phoebos › logs › freenode-kisslinux-2019-12-01.txt captured on 2024-03-21 at 16:10:34.
⬅️ Previous capture (2021-12-17)
-=-=-=-=-=-=-
2019-12-01T09:56:45 #kisslinux <chrempson> dylanaraps: rust & firefox now building on ppc: https://ibb.co/WWk6rC6
2019-12-01T10:04:20 #kisslinux <dylanaraps> Nice!
2019-12-01T10:05:20 #kisslinux <dylanaraps> What's the build time like?
2019-12-01T10:05:41 #kisslinux <dylanaraps> Takes about 2 1/2 hours on my machine.
2019-12-01T10:45:14 #kisslinux <chrempson> 25 min
2019-12-01T14:16:54 #kisslinux <chrempson> Remaining failing packages on ppc: go, yajl, i3-gaps, luajit, neovim. Everything else now builds. I put a package status page here: https://github.com/jdavies-dev/kiss-ppc64le
2019-12-01T15:02:52 #kisslinux <dylanaraps> Great work. :)
2019-12-01T15:03:41 #kisslinux <dylanaraps> I wonder if I can fix the "OK. Patched (specify build type)" on my end.
2019-12-01T15:03:45 #kisslinux <dylanaraps> chrempson:
2019-12-01T15:04:18 #kisslinux <dylanaraps> What's the yajl error?
2019-12-01T15:06:03 #kisslinux <dylanaraps> qtpie: I've had zero freezes in Firefox with 'xcompmgr'.
2019-12-01T15:47:33 #kisslinux <chrempson> regarding the build type, it's normally on autoconf-based projects. Mostly just adding --build=powerpc64le-musl-linux. Unfortunately $(cc -dumpmachine) -> powerpc64le-linux-musl.
2019-12-01T16:11:41 #kisslinux <dylanaraps> chrempson: I wonder if the triplet issue is something you can fix in GCC?
2019-12-01T16:12:49 #kisslinux <dylanaraps> The structure is meant to be this: machine-vendor-operatingsystem
2019-12-01T16:13:16 #kisslinux <dylanaraps> This is what it looks like on my machine: x86_64-pc-linux-musl
2019-12-01T16:13:26 #kisslinux <dylanaraps> Hm.
2019-12-01T16:13:28 #kisslinux <dylanaraps> Actually.
2019-12-01T16:13:38 #kisslinux <dylanaraps> Why doesn't powerpc64le-linux-musl work?
2019-12-01T16:16:24 #kisslinux <qtpie> dylanaraps, nice. it seems not having a compositor was the issue. i haven't experienced any issues either after installing picom
2019-12-01T16:57:07 #kisslinux <KayW> hey anyone here do AoC?
2019-12-01T16:59:28 #kisslinux <dylanaraps> qtpie: This is still a Firefox bug though. It should still work without a compositor. This is just a workaround.
2019-12-01T17:08:15 #kisslinux <qtpie> dylanaraps, i'm inclined to think it's caused by something else because even FF69 freezes for me now and I'm pretty sure it never froze for me back in the day it was the latest release
2019-12-01T17:08:45 #kisslinux <qtpie> i wasn't using any compositor back then either
2019-12-01T17:21:32 #kisslinux <chrempson> got a build of qemu working if anyone's interested: https://github.com/jdavies-dev/community/tree/master/community/qemu
2019-12-01T17:21:36 #kisslinux <chrempson> should work on x86
2019-12-01T17:22:00 #kisslinux <chrempson> ncurses and serial io work, it segfaults when you try to boot a graphical guest though
2019-12-01T17:22:05 #kisslinux <chrempson> working on it
2019-12-01T17:53:42 #kisslinux <dylanaraps> chrempson: I've been using this: https://termbin.com/0vx4
2019-12-01T17:54:07 #kisslinux <dylanaraps> It's a minimal build exactly as you describe.
2019-12-01T17:56:04 #kisslinux <chrempson> actually seems ok now - bad cmdline args
2019-12-01T17:56:32 #kisslinux <chrempson> ok yeah, booting x64_64 linux iso in graphical mode
2019-12-01T17:56:40 #kisslinux <chrempson> dylanaraps: thanks will check out your config
2019-12-01T19:05:28 #kisslinux <dylanaraps> So much for iwd being portable:
2019-12-01T19:05:30 #kisslinux <dylanaraps> > checking systemd unit directory... configure: error: systemd unit directory is required
2019-12-01T19:05:50 #kisslinux <dylanaraps> (Forgetting the fact it also needs dbus)
2019-12-01T19:06:12 #kisslinux <E5ten> --disable-systemd-service
2019-12-01T19:06:40 #kisslinux <dylanaraps> Yeah
2019-12-01T19:06:47 #kisslinux <dylanaraps> I just found it.
2019-12-01T19:06:57 #kisslinux <dylanaraps> Going to patch dbus out of iwd.
2019-12-01T19:07:20 #kisslinux <E5ten> I will be overjoyed when you do
2019-12-01T19:07:33 #kisslinux <dylanaraps> > wiphy.c:(.text+0x1908): undefined reference to `rawmemchr'
2019-12-01T19:07:39 #kisslinux <dylanaraps> Normal compiler errors first.
2019-12-01T19:08:47 #kisslinux <E5ten> damn iwd you really out here using glibc-specific functions :(
2019-12-01T19:10:10 #kisslinux <dylanaraps> Alpine have a patch.
2019-12-01T19:10:19 #kisslinux <E5ten> nice
2019-12-01T19:10:39 #kisslinux <dylanaraps> Literally just: rawmemchr -> strchr
2019-12-01T19:11:10 #kisslinux <E5ten> hopefully that gets upstreamed lol...
2019-12-01T19:11:25 #kisslinux <E5ten> also nice, sed-able no patch needed
2019-12-01T19:11:45 #kisslinux <dylanaraps> Yup
2019-12-01T19:11:50 #kisslinux <dylanaraps> OK. Builds fine.
2019-12-01T19:12:20 #kisslinux <dylanaraps> Will setup a repository instead of one giant patch for this.
2019-12-01T19:12:49 #kisslinux <E5ten> the iwd guy said he'd "fix this soonish" like 2 weeks ago so hopefully that happens not too long from now...
2019-12-01T19:12:53 #kisslinux <E5ten> rawmemchr thing I mean
2019-12-01T19:13:11 #kisslinux <dylanaraps> lol
2019-12-01T19:13:30 #kisslinux <dylanaraps> There's no rush for him until next release anyway.
2019-12-01T19:15:26 #kisslinux <mforney> what will you replace the IPC between iwd and iwctl with? a while ago, i was looking at getting iwd to work without dbus, too
2019-12-01T19:15:37 #kisslinux <E5ten> I know I'm living in a fantasy world but I have a small amount of hope that once you patch out dbus with uh I guess some other IPC thing? that they'd react to that like "oh nice we can actually be depless" and then merge it upstream
2019-12-01T19:15:58 #kisslinux <dylanaraps> iwd depends on dbus for things like networkmanager.
2019-12-01T19:15:59 #kisslinux <dylanaraps> :(
2019-12-01T19:16:33 #kisslinux <mforney> iwd uses networkmanager?
2019-12-01T19:16:38 #kisslinux <dylanaraps> No no
2019-12-01T19:16:48 #kisslinux <dylanaraps> Clients can be written for it using dbus as the bridge.
2019-12-01T19:16:56 #kisslinux <E5ten> oh right yeah it has its dbus API for networkmanager to use so they'd never accept that upstream
2019-12-01T19:16:59 #kisslinux <E5ten> forgot about that
2019-12-01T19:17:02 #kisslinux <dylanaraps> Yup
2019-12-01T19:17:12 #kisslinux <dylanaraps> Will use Unix domain socket probably.
2019-12-01T19:17:27 #kisslinux <E5ten> but I assume you'll just be like ripping that out and the only stuff getting actually replaced will be the interaction between iwctl and iwd?
2019-12-01T19:17:35 #kisslinux <dylanaraps> I need to understand the code first.
2019-12-01T19:17:37 #kisslinux <mforney> with some kind of custom protocol on top of that?
2019-12-01T19:17:37 #kisslinux <dylanaraps> Yeah
2019-12-01T19:18:26 #kisslinux <mforney> cool
2019-12-01T19:19:04 #kisslinux <dylanaraps> It depends on how the dbus code is implemented though.
2019-12-01T19:19:11 #kisslinux <dylanaraps> It might be a pain in the ass.
2019-12-01T19:21:20 #kisslinux <dylanaraps> -> grep -r dbus src/ | wc -l
2019-12-01T19:21:22 #kisslinux <dylanaraps> 886
2019-12-01T19:22:15 #kisslinux <dylanaraps> Hang on.
2019-12-01T19:22:17 #kisslinux <dylanaraps> -> grep -r dbus src/ client/ wired/ monitor/ linux/ tools/ | wc -l
2019-12-01T19:22:19 #kisslinux <dylanaraps> 1510
2019-12-01T19:22:21 #kisslinux <dylanaraps> oof
2019-12-01T19:22:39 #kisslinux <mforney> when i was looking at this, it seemed pretty easy to just remove the code that registered the dbus interfaces
2019-12-01T19:22:48 #kisslinux <mforney> how much of that is the ell dbus client library?
2019-12-01T19:23:44 #kisslinux <dylanaraps> None of it.
2019-12-01T19:23:50 #kisslinux <dylanaraps> ell doesn't seem to be in this git tree(?)
2019-12-01T19:24:59 #kisslinux <mforney> yeah, it's only in the tarball i think
2019-12-01T19:25:04 #kisslinux <dylanaraps> Yup
2019-12-01T19:25:09 #kisslinux <dylanaraps> So this is all iwd code.
2019-12-01T19:27:28 #kisslinux <dylanaraps> It'd be nice if I'm to just swap in my own version of dbus.c so none of the codebase needs changing.
2019-12-01T19:28:08 #kisslinux <dylanaraps> iwd has code included for connman already fyi.
2019-12-01T19:29:26 #kisslinux <mforney> you mean emulate the ell dbus api on top of your own custom protocol? might be tricky
2019-12-01T19:31:02 #kisslinux <dylanaraps> Exactly.
2019-12-01T19:32:16 #kisslinux <mforney> the diff i had from a month ago is http://ix.io/23jN. it was enough to get it started without a dbus daemon, but i didn't get further than that
2019-12-01T19:33:07 #kisslinux <mforney> i figured if i could get it working just with a config file and no iwctl client, that'd be a start
2019-12-01T19:33:12 #kisslinux <dylanaraps> That's what I'm working on now.
2019-12-01T19:34:08 #kisslinux <dylanaraps> It also opens normal netlink sockets fyi: l_error("Failed to open generic netlink socket");
2019-12-01T19:34:35 #kisslinux <mforney> do you have an strace log? might be a kernel config issue
2019-12-01T19:34:55 #kisslinux <dylanaraps> This is from the source.
2019-12-01T19:35:06 #kisslinux <dylanaraps> Just posting to show it uses netlink sockets. :)
2019-12-01T19:35:37 #kisslinux <mforney> oh, of course it uses netlink, that's how you configure wireless interfaces on linux
2019-12-01T19:36:25 #kisslinux <mforney> wpa_supplicant uses libnl for this
2019-12-01T19:37:13 #kisslinux <dylanaraps> Yup
2019-12-01T19:38:44 #kisslinux <dylanaraps> OK. Daemon runs.
2019-12-01T19:44:42 #kisslinux <dylanaraps> brb
2019-12-01T20:22:43 #kisslinux <dylanaraps> Man iwd is funky.
2019-12-01T20:39:08 #kisslinux <dylanaraps> OK. iwd hangs here: genl = l_genl_new();
2019-12-01T20:39:29 #kisslinux <dylanaraps> Endless epoll_pwait();
2019-12-01T20:42:03 #kisslinux <dylanaraps> Hang on.
2019-12-01T20:42:07 #kisslinux <dylanaraps> Hangs lower down.
2019-12-01T20:44:22 #kisslinux <dylanaraps> Hangs here: exit_status = l_main_run_with_signal(signal_handler, NULL);
2019-12-01T20:46:26 #kisslinux <dylanaraps> ooo
2019-12-01T20:59:56 #kisslinux <dylanaraps> OK. So it enters the main loop and does... nothing. I'm hoping that it doesn't require dbus to actually function once the main loop starts.
2019-12-01T21:00:23 #kisslinux <dylanaraps> I've setup the main config file and a config file for my current network.
2019-12-01T21:03:52 #kisslinux <chrempson> dylanaraps: want a pr for qemu in community? or do you already have something in the works for qemu?
2019-12-01T21:04:41 #kisslinux <dylanaraps> Sure
2019-12-01T21:07:34 #kisslinux <dylanaraps> E5ten: From the iwd manpages. > iwd [...] The result is a self-contained environment that only depends on the Linux Kernel and the runtime C library.
2019-12-01T21:07:50 #kisslinux <dylanaraps> So much for that. :(
2019-12-01T21:08:50 #kisslinux <E5ten> yeah we already agreed about that being lies lol
2019-12-01T21:09:29 #kisslinux <E5ten> I think what they mean is that the binary doesn't require any extra libs, like it doesn't use libdbus, but that's dumb and misleading
2019-12-01T21:09:50 #kisslinux <dylanaraps> Yup
2019-12-01T21:10:05 #kisslinux <dylanaraps> Their ELL library also has an optional dependency on glib
2019-12-01T21:10:09 #kisslinux <dylanaraps> for reasons
2019-12-01T21:10:12 #kisslinux <E5ten> it does?
2019-12-01T21:10:20 #kisslinux <dylanaraps> Void compiles it with glib.
2019-12-01T21:10:29 #kisslinux <dylanaraps> https://github.com/void-linux/void-packages/blob/master/srcpkgs/ell/template
2019-12-01T21:10:38 #kisslinux <E5ten> I thought the glib thing was only for an example command that uses the lib?
2019-12-01T21:10:43 #kisslinux <dylanaraps> "--enable-glib"
2019-12-01T21:11:14 #kisslinux <E5ten> yeah it's for an example, dumb that it's got a configure flag like that making it seem like part of the lib...
2019-12-01T21:11:38 #kisslinux <dylanaraps> 10/10
2019-12-01T21:11:44 #kisslinux <E5ten> examples/glib-eventloop.c
2019-12-01T21:12:22 #kisslinux <E5ten> the other problem here is uh why would void enable a flag that literally makes an example and nothing else lol
2019-12-01T21:12:41 #kisslinux <dylanaraps> Yup
2019-12-01T21:12:59 #kisslinux <dylanaraps> The whole maintainer-ship of the distribution is weird.
2019-12-01T21:14:17 #kisslinux <dylanaraps> https://github.com/void-linux/void-packages/commit/a7eec1dc4316c3e62b2811bc15c1957d05a6e59e
2019-12-01T21:14:20 #kisslinux <dylanaraps> Bug "fixed"
2019-12-01T21:15:46 #kisslinux <E5ten> uh
2019-12-01T21:15:54 #kisslinux <E5ten> that's some high quality fixing there
2019-12-01T21:16:06 #kisslinux <dylanaraps> There's a lot of stuff like that.
2019-12-01T21:16:15 #kisslinux <dylanaraps> They also haven't released an ISO since 2018.
2019-12-01T21:16:21 #kisslinux <dylanaraps> And the signature for it has expired.
2019-12-01T21:16:45 #kisslinux <E5ten> 😬
2019-12-01T21:16:58 #kisslinux <dylanaraps> iwd just loops doing nothing after launch.
2019-12-01T21:17:21 #kisslinux <dylanaraps> The main loop literally just watches a socket.
2019-12-01T21:18:39 #kisslinux <E5ten> yikes
2019-12-01T21:19:22 #kisslinux <dylanaraps> > exit_status = l_main_run_with_signal(signal_handler, NULL);
2019-12-01T21:19:42 #kisslinux <dylanaraps> The function is from ell and literally just loops forever on an interval.
2019-12-01T21:20:16 #kisslinux <dylanaraps> The signal handler just handles INT/TERM.
2019-12-01T21:20:42 #kisslinux <E5ten> so is this gonna require an ell-no-dbus?
2019-12-01T21:21:02 #kisslinux <dylanaraps> The main loop code in ell doesn't use dbus either iirc.
2019-12-01T21:21:18 #kisslinux <atkka> stuff, like you mentioned is why I went from void to alpine
2019-12-01T21:21:31 #kisslinux <dylanaraps> The goal is to get iwd to work with config files (like wpa_supplicant w/o wpa_cli).
2019-12-01T21:21:54 #kisslinux <dylanaraps> (at first)
2019-12-01T21:22:16 #kisslinux <dylanaraps> But iwd just loops doing nothing :@
2019-12-01T21:23:04 #kisslinux <E5ten> I'm so confused why iwd itself needs dbus, like it's the daemon, how can it need an IPC mechanism to communicate with itself?
2019-12-01T21:23:14 #kisslinux <mforney> no error messages or anything?
2019-12-01T21:23:17 #kisslinux <dylanaraps> Nope
2019-12-01T21:23:22 #kisslinux <dylanaraps> I'm watching strace.
2019-12-01T21:24:47 #kisslinux <mforney> E5ten: presumably you want some way to control the daemon to tell it to connect to a certain network, check its status, etc. same things you use wpa_cli for
2019-12-01T21:25:09 #kisslinux <mforney> dylanaraps: what's your current diff?
2019-12-01T21:25:22 #kisslinux <dylanaraps> Pushing it now.
2019-12-01T21:25:33 #kisslinux <dylanaraps> mforney: This is the daemon needing dbus to talk to the... daemon.
2019-12-01T21:25:53 #kisslinux <E5ten> that's what I'm talking about yeah
2019-12-01T21:26:16 #kisslinux <mforney> are you sure it is doing that?
2019-12-01T21:26:17 #kisslinux <E5ten> I get that it would need it to talk to iwctl and vice versa, but iwctl isn't coming into play here, so I don't get how the lack of dbus is causing it to not work
2019-12-01T21:27:39 #kisslinux <dylanaraps> https://github.com/dylanaraps/iwd-no-dbus/commit/d958e4febb57a97be05066b49c52ea3f7e01921a
2019-12-01T21:28:05 #kisslinux <dylanaraps> Unpatched it fails here: l_error("Failed to initialize D-Bus");
2019-12-01T21:28:26 #kisslinux <dylanaraps> Patched it hits the main loop and stays there doing nothing.
2019-12-01T21:28:28 #kisslinux <dylanaraps> > exit_status = l_main_run_with_signal(signal_handler, NULL);
2019-12-01T21:29:38 #kisslinux <dylanaraps> There's no code in main.c which would make iwd actually do anything from what I can see.
2019-12-01T21:29:38 #kisslinux <mforney> dylanaraps: look at request_name_callback. you still need to call l_genl_request_family to request nl80211 interfaces
2019-12-01T21:29:46 #kisslinux <dylanaraps> ooo
2019-12-01T21:29:52 #kisslinux <dylanaraps> heh
2019-12-01T21:30:51 #kisslinux <dylanaraps> Hang on. That's called through a dbus related function.
2019-12-01T21:30:59 #kisslinux <dylanaraps> Let me try it regardless.
2019-12-01T21:31:02 #kisslinux <mforney> yeah, it's a callback once it initialized dbus
2019-12-01T21:31:33 #kisslinux <mforney> but since we're not doing that, you need to call it in place of the dbus_init call you did
2019-12-01T21:31:42 #kisslinux <mforney> you removed*
2019-12-01T21:35:55 #kisslinux <dylanaraps> ooo
2019-12-01T21:35:57 #kisslinux <dylanaraps> OK
2019-12-01T21:36:17 #kisslinux <dylanaraps> Progress.
2019-12-01T21:36:20 #kisslinux <E5ten> is that a "that's a good idea" ooo or a "it's working more now" ooo
2019-12-01T21:36:22 #kisslinux <E5ten> oh fuck yeah
2019-12-01T21:36:22 #kisslinux <dylanaraps> Thanks :D
2019-12-01T21:37:04 #kisslinux <mforney> it probably was doing nothing since it detected no wireless interfaces on your system :)
2019-12-01T21:37:15 #kisslinux <dylanaraps> Bingo
2019-12-01T21:38:30 #kisslinux <dylanaraps> > Module agent failed to start: -5
2019-12-01T21:39:26 #kisslinux <E5ten> I think agent has to do with iwctl and like authentication or something?
2019-12-01T21:39:45 #kisslinux <mforney> probably more dbus interface registrations you need to remove. take a look at the diff i posted earlier
2019-12-01T21:39:49 #kisslinux <dylanaraps> Yup
2019-12-01T21:43:02 #kisslinux <mforney> btw, i think `(uint8_t *) strchr(s, 0) - s` can be written more succinctly as `strlen(s)`. not really sure why they made it so complicated with rawmemchr and pointer arithmetic
2019-12-01T21:43:18 #kisslinux <dylanaraps> OOOOOOOOOOOOO
2019-12-01T21:43:24 #kisslinux <dylanaraps> We might be working.
2019-12-01T21:43:39 #kisslinux <dylanaraps> Hang on.
2019-12-01T21:43:39 #kisslinux <dylanaraps> Need to kill wpa_supplicant/dhcpcd.
2019-12-01T21:49:35 #kisslinux <dylanaraps> OK.
2019-12-01T21:49:45 #kisslinux <dylanaraps> > No usable resolvconf found on system
2019-12-01T21:49:50 #kisslinux <dylanaraps> > src/resolve.c:resolve_resolvconf_init() Trying to find resolvconf in $PATH
2019-12-01T21:50:01 #kisslinux <dylanaraps> > EAP_MSCHAPv2 init: MD4 support not found, skipping
2019-12-01T21:50:04 #kisslinux <E5ten> I mean, that's not a dbus thing that's a you don't have openresolv thing
2019-12-01T21:50:05 #kisslinux <dylanaraps> > rfkill id 1 can't be matched to a wiphy
2019-12-01T21:50:10 #kisslinux <dylanaraps> I know
2019-12-01T21:50:17 #kisslinux <dylanaraps> Just listing out the errors.
2019-12-01T21:50:26 #kisslinux <dylanaraps> or warnings
2019-12-01T21:53:17 #kisslinux <mforney> i think that's as far as i got
2019-12-01T21:53:29 #kisslinux <dylanaraps> brb again
2019-12-01T21:56:48 #kisslinux <E5ten> wait is this a "compile my kernel with MD4 support" brb
2019-12-01T21:57:46 #kisslinux <dylanaraps> no
2019-12-01T21:57:48 #kisslinux <dylanaraps> openresolv brb
2019-12-01T21:57:57 #kisslinux <E5ten> ah
2019-12-01T21:58:02 #kisslinux <dylanaraps> Another brb to try iwd again
2019-12-01T22:13:03 #kisslinux <dylanaraps> This is iwd now: https://termbin.com/9klk
2019-12-01T22:15:07 #kisslinux <dylanaraps> iwd is finding the wrong wlan device basically.
2019-12-01T22:16:01 #kisslinux <dylanaraps> It's using ID 0 instead of ID 1: https://termbin.com/hgbpq
2019-12-01T22:16:10 #kisslinux <dylanaraps> phy0 instead of ideapad_wlan
2019-12-01T22:18:10 #kisslinux <mforney> regarding `scan_periodic_start called without scan_wdev_add`, i just figured that one out. device_init just fails if it can't register the dbus interface. try http://ix.io/23kE
2019-12-01T22:18:42 #kisslinux <mforney> i'm now getting as far as seeing the scan results. not sure if i configured iwd correctly though
2019-12-01T22:19:11 #kisslinux <dylanaraps> Nice
2019-12-01T22:19:12 #kisslinux <dylanaraps> Will try.
2019-12-01T22:20:47 #kisslinux <dylanaraps> brb
2019-12-01T22:21:08 #kisslinux <mforney> `src/station.c:station_enter_state() Old State: connecting, new state: connected` :D
2019-12-01T22:21:17 #kisslinux <E5ten> oh shit, nice
2019-12-01T22:23:27 #kisslinux <mforney> looks like it's working
2019-12-01T22:23:37 #kisslinux <mforney> got an ip over dhcp
2019-12-01T22:42:07 #kisslinux <dylanaraps> My iwd segfaults with your changes which I think is related to it using the wrong interface in the first place.
2019-12-01T22:42:10 #kisslinux <dylanaraps> Also: open("/sys/class/rfkill/rfkill1/device/index", O_RDONLY) = -1 ENOENT (No such file or directory)
2019-12-01T22:42:18 #kisslinux <dylanaraps> This is why the rfkill error appears.
2019-12-01T22:42:24 #kisslinux <dylanaraps> The index ile doesn't exist.
2019-12-01T22:43:17 #kisslinux <mforney> the segfaults are all due to iwd trying to change dbus properties (i.e. making progress on connecting), but dbus==NULL
2019-12-01T22:43:53 #kisslinux <dylanaraps> Ah
2019-12-01T22:44:27 #kisslinux <mforney> look for l_dbus_property_changed in src/knownnetworks.c
2019-12-01T22:45:48 #kisslinux <dylanaraps> Thanks
2019-12-01T22:45:52 #kisslinux <dylanaraps> brb
2019-12-01T22:50:12 #kisslinux <mforney> dylanaraps: here is my diff against your iwd-no-dbus repo: http://ix.io/23kJ
2019-12-01T22:52:42 #kisslinux <dylanaraps> Is the client/ code used in the daemon?
2019-12-01T22:53:17 #kisslinux <mforney> i don't believe so. i built with --disable-client
2019-12-01T22:57:34 #kisslinux <E5ten> I think client/ is iwctl?
2019-12-01T22:57:51 #kisslinux <mforney> yes
2019-12-01T23:00:18 #kisslinux <dylanaraps> Thanks
2019-12-01T23:00:55 #kisslinux <dylanaraps> I pushed a commit.
2019-12-01T23:11:52 #kisslinux <dylanaraps> I segfault much later now.
2019-12-01T23:12:00 #kisslinux <dylanaraps> src/scan.c:scan_notify() Scan notification Trigger Scan(33)
2019-12-01T23:12:02 #kisslinux <dylanaraps> Segmentation fault
2019-12-01T23:12:45 #kisslinux <dylanaraps> I get scan results now.
2019-12-01T23:26:21 #kisslinux <dylanaraps> WOO
2019-12-01T23:26:25 #kisslinux <dylanaraps> iwd works
2019-12-01T23:26:36 #kisslinux <dylanaraps> Hello from iwd connected wifi
2019-12-01T23:26:45 #kisslinux <dylanaraps> E5ten:
2019-12-01T23:26:47 #kisslinux <dylanaraps> mforney:
2019-12-01T23:27:53 #kisslinux <E5ten> fuck yeah
2019-12-01T23:28:07 #kisslinux <E5ten> now comes the hard part I guess lol
2019-12-01T23:28:10 #kisslinux <dylanaraps> Heh
2019-12-01T23:28:26 #kisslinux <dylanaraps> I think we get rid of all dbus code and then build our own interface on top.
2019-12-01T23:28:50 #kisslinux <dylanaraps> I've removed so much of it already.
2019-12-01T23:29:00 #kisslinux <E5ten> makes sense
2019-12-01T23:29:47 #kisslinux <E5ten> so entirely redoing the interaction between iwd and iwctl, not making functions equivalent to the ell ones but using normal IPC instead of dbus
2019-12-01T23:30:05 #kisslinux <dylanaraps> Yup
2019-12-01T23:30:41 #kisslinux <E5ten> makes sense, I hope it doesn't lead to it being harder to maintain if they change stuff frequently
2019-12-01T23:31:15 #kisslinux <dylanaraps> Yup
2019-12-01T23:31:18 #kisslinux <E5ten> but I guess making functions meant for dbus use with non-dbus would be pretty damn hard
2019-12-01T23:31:22 #kisslinux <dylanaraps> Yup
2019-12-01T23:31:31 #kisslinux <dylanaraps> dbus is so butt ugly though
2019-12-01T23:31:46 #kisslinux <E5ten> dbutt
2019-12-01T23:35:25 #kisslinux <dylanaraps> https://github.com/dylanaraps/iwd-no-dbus
2019-12-01T23:35:31 #kisslinux <dylanaraps> README has instructions now.
2019-12-01T23:37:28 #kisslinux <E5ten> oh man I can't wait for iwctl to be working so I can finally ditch dbus, for iwd-no-dbus among other things I'm putting you in my non-existent "software royalty" hall of fame
2019-12-01T23:37:59 #kisslinux <dylanaraps> Heh
2019-12-01T23:38:04 #kisslinux <dylanaraps> This will take me some time.
2019-12-01T23:38:21 #kisslinux <dylanaraps> I also want to remove the resolvconf dependency.
2019-12-01T23:39:05 #kisslinux <dylanaraps> What's funny is that the key needs 'wpa_passphrase' to be generated manually and resolvconf is by the dhcpcd guy.
2019-12-01T23:39:36 #kisslinux <mforney> what is resolvconf? i didn't run into that
2019-12-01T23:39:53 #kisslinux <dylanaraps> openresolv
2019-12-01T23:40:06 #kisslinux <dylanaraps> iwd uses it to handle resolv.conf
2019-12-01T23:40:23 #kisslinux <dylanaraps> Unless you're using systemd which also handles it.
2019-12-01T23:40:52 #kisslinux <mforney> oh, i just ran sdhcp which did dhcp and set up resolv.conf
2019-12-01T23:41:11 #kisslinux <dylanaraps> iwd can replace sdhcp as well.
2019-12-01T23:41:26 #kisslinux <dylanaraps> openresolv is 100kb though.
2019-12-01T23:41:27 #kisslinux <mforney> yeah, just saw that it has its own dhcp client
2019-12-01T23:41:31 #kisslinux <dylanaraps> Not a big dea.
2019-12-01T23:41:33 #kisslinux <dylanaraps> deal*
2019-12-01T23:42:00 #kisslinux <atkka> you'll still want openresolve if use dns though a vpn anyway
2019-12-01T23:42:06 #kisslinux <E5ten> I think it uses resolvconf cuz it trusts it more to handle multiple things writing to resolv.conf than whatever they might write to try to do the same
2019-12-01T23:42:55 #kisslinux <dylanaraps> It stays then.
2019-12-01T23:43:15 #kisslinux <dylanaraps> You never know, upstream iwd might eat up openresolv's functionality too.
2019-12-01T23:44:46 #kisslinux <E5ten> I'm pretty sure I remember denkenz saying it wouldn't for the reason I mentioned
2019-12-01T23:48:46 #kisslinux <dylanaraps> All systemd code will go too as users with systemd may as well use upstream.
2019-12-01T23:49:02 #kisslinux <E5ten> makes sense
2019-12-01T23:49:35 #kisslinux <E5ten> but if it's not being used wouldn't removing it just make the diff larger and therefore keeping up with upstream slightly more annoying?
2019-12-01T23:50:02 #kisslinux <atkka> I'm pretty sure openresolv is the leanest way to automatically handle resolv.conf anyway
2019-12-01T23:52:04 #kisslinux <dylanaraps> E5ten: systemd stuff is mostly in the configure.ac/makefile.am.
2019-12-01T23:52:16 #kisslinux <dylanaraps> The only systemd code in the daemon is for resolv stuff iirc.
2019-12-01T23:52:19 #kisslinux <E5ten> ah nevermind then, I thought you meant like systemd-resolved stuff
2019-12-01T23:52:30 #kisslinux <E5ten> oh wait unnevermind
2019-12-01T23:58:52 #kisslinux <dylanaraps> I've most likely already purged code from the systemd portions of resolve.c
2019-12-01T23:58:53 #kisslinux <dylanaraps> lol
2019-12-01T23:59:32 #kisslinux <E5ten> lol nice