2023-02-01 01:39:43 <Waitingfortherain[m]> for macbook wifi afaik it uses a proprietary broadcom-wl driver, which I don't think supports glibc
2023-02-01 01:40:11 <Waitingfortherain[m]> s/glibc/musl/
2023-02-01 01:40:11 <alpine-meetbot> Waitingfortherain[m] meant to say: for macbook wifi afaik it uses a proprietary broadcom-wl driver, which I don't think supports musl
2023-02-01 01:41:56 <psykose> it's a kernel module
2023-02-01 02:14:27 <Saijin_Naib> <tomk> "Did anyone sucessfully connect..." <- A while back in the channel someone mentioned the config shipped with Alpine doesn't work, and that they used one from another distro and it worked fine
2023-02-01 02:14:36 <Saijin_Naib> I've never been able to get xRDP to work under Alpine, no
2023-02-01 07:52:09 <we11en> hi. any idea why i can boot apline iso in UEFI mode in qemu, but fail to boot in UEFI mode (baremetal) when put the iso on a usbstick?
2023-02-01 07:55:22 <we11en> unfortunately im unable to boot legacy bios on this machine :/ so only UEFI will boot.
2023-02-01 08:32:46 <we11en> exit
2023-02-01 11:29:58 <JanJ[m]> we11en: how did you create the usbstick?
2023-02-01 13:13:51 <ikke> JanJ[m]: they left 
2023-02-01 13:23:02 <emersion> hi! i'm hitting a broken dep when installing xwayland-dev on edge: https://builds.sr.ht/~emersion/job/932391
2023-02-01 13:23:25 <emersion> it seems like font-util-dev is missing from the repos
2023-02-01 13:23:30 <emersion> it's supposed to be provided by https://gitlab.alpinelinux.org/alpine/aports/-/blob/master/main/font-util/APKBUILD
2023-02-01 13:23:48 <emersion> but the latest build logs don't mention anything about -dev: https://build.alpinelinux.org/buildlogs/build-edge-x86_64/main/font-util/font-util-1.3.3-r2.log
2023-02-01 13:24:45 <psykose> apologies, was a typo in that commit last on it
2023-02-01 13:24:48 <psykose> 0a1bb39cef8cf44896fec26ddf9974e2e6c56e99
2023-02-01 13:24:53 <psykose> give it like 20 min and should be fixed
2023-02-01 13:25:23 <psykose> friendlier link https://git.alpinelinux.org/aports/commit/?id=0a1bb39cef8cf44896fec26ddf9974e2e6c56e99
2023-02-01 13:25:46 <emersion> ah yes i see it
2023-02-01 13:26:05 <emersion> added an extra subpackages which overwrites the existing one
2023-02-01 13:26:09 <psykose> yep
2023-02-01 13:26:14 <emersion> ah cool, thanks!
2023-02-01 15:07:11 <Fxzxmic> What is the use of the install command?
2023-02-01 15:08:20 <minimal> Fxzxmic: to install stuff ;-)
2023-02-01 15:10:32 <Fxzxmic> When to use it?
2023-02-01 15:10:40 <psykose> to install stuff
2023-02-01 15:10:48 <quinq> When you want to install stuff
2023-02-01 15:11:02 <Fxzxmic> I can't think of when I will use it.
2023-02-01 15:11:12 <psykose> then don't use it :)
2023-02-01 15:11:13 <quinq> Then never
2023-02-01 15:11:21 <psykose> it doesn't do anything other commands can't do
2023-02-01 15:11:30 <quinq> Also note that while it can be handy, that's a not portable tool
2023-02-01 15:12:32 <Fxzxmic> So does Alpine have an application that depends on this command?
2023-02-01 15:13:06 <minimal> Fxzxmic: why do you ask? why do you care?
2023-02-01 15:13:30 <psykose> if you phrase it to 'does anything use it', sure, grep aports for "install"
2023-02-01 15:15:06 <Fxzxmic> minimal: I just want to know under what circumstances this command will be used. It would be better to have an example.
2023-02-01 15:16:06 <minimal> Fxzmic: I've heard a rumour of a magical thing called Google Search that can provide answers to most questions ;-)
2023-02-01 15:16:10 <elly> well like, for example: install -D -m 755 binary -t $PREFIX/bin
2023-02-01 15:16:34 <psykose> install -Dm755 -o xyz x a/y
2023-02-01 15:16:37 <elly> is equivalent to: mkdir -p $PREFIX/bin ; cp binary $PREFIX/bin/binary ; chmod 755 $PREFIX/bin/binary
2023-02-01 15:16:43 <elly> which is basically what install does
2023-02-01 15:17:09 <psykose> there's even an -s for strip too
2023-02-01 15:17:11 <minimal> "under what circumstances this command will be used" - when someone wants to install a file/directory...
2023-02-01 15:17:35 <elly> the real benefit of install, imo, is that it makes your intent very clear to the reader
2023-02-01 15:17:41 <elly> vs writing out the mkdir / cp / chmod
2023-02-01 15:17:58 <minimal> you quote the Ronseal advert catchphrase "It does exactly what it says on the tin" lol
2023-02-01 15:18:30 <Fxzxmic> elly: psykose: thanks
2023-02-01 15:19:00 <Hello71> it also removes the target beforehand which cp doesn't do
2023-02-01 15:19:27 <psykose> what was the point of that again
2023-02-01 15:20:09 <elly> hm
2023-02-01 15:20:20 <elly> I guess it kinda ensures the new file has a different inode number and ctime
2023-02-01 15:23:56 <Fxzxmic> I just haven't seen this command. The information found in the search engine still makes me confused, so I ask here.
2023-02-01 15:29:35 <psykose> if you're writing scripts for stuff generally you avoid it as -d/-D mean different things on bsd
2023-02-01 15:29:51 <psykose> if you know that's not an issue (aports is linux)then you can use all the options just fine :p
2023-02-01 15:37:02 <Fxzxmic> :)
2023-02-01 15:40:22 <JanJ[m]> hmz
2023-02-01 15:52:59 <scorpion2185[m]1> is musl-dev on alpine for glibc gcc-libs?
2023-02-01 15:53:53 <scorpion2185[m]1> gcc-libs (fastgcc, gccrs-libs-git, gcc-libs-git)
2023-02-01 15:55:09 <psykose> there is no glibc
2023-02-01 15:55:16 <psykose> musl-dev are the libc headers
2023-02-01 15:55:38 <psykose> i have no idea what "gcc-libs (fastgcc, gccrs-libs-git, gcc-libs-git)" is
2023-02-01 15:56:19 <psykose> libgcc is probably gcc-libs, which is the libgcc_s runtime support library
2023-02-01 15:57:49 <psykose> which you generally never have to care about or even know exists
2023-02-01 15:57:57 <psykose> compiler semantics
2023-02-01 16:03:15 <scorpion2185[m]1> when you are building how can you put something in build () to print the workind dir?
2023-02-01 16:03:44 <psykose> pwd
2023-02-01 16:05:05 <scorpion2185[m]1> will that appear on when you are running abuiild -r?
2023-02-01 16:05:19 <scorpion2185[m]1> *on the terminal
2023-02-01 16:06:40 <psykose> find out in rp
2023-02-01 16:07:38 <psykose> (abuild doesn't filter anything and you see the normal stdout/err)
2023-02-01 16:11:36 <Fxzxmic> I found that the built-in man command of busybox cannot be used.
2023-02-01 16:11:54 <Fxzxmic> It said "sh: nroff: not found".
2023-02-01 16:15:23 <minimal> Fxzxmic: what "built-in man command of busybox"? Alpine has Busybox's man disabled
2023-02-01 16:19:04 <Fxzxmic> I compiled busybox myself.
2023-02-01 16:19:34 <minimal> Fxzxmic: and you installed nroff?
2023-02-01 16:19:54 <Fxzxmic> trying
2023-02-01 16:20:26 <minimal> Fxzxmic: isn't the error message obvious? It can't find "nroff" likely because nroff is not installed
2023-02-01 16:21:47 <psykose> which comes from the `groff` package
2023-02-01 16:22:41 <Fxzxmic> I thought busybox didn't need any extra dependence.
2023-02-01 16:23:04 <Fxzxmic> Now I'm wrong.
2023-02-01 16:23:45 <minimal> Fxzxmic: that's not an Alpine issue - you've decided to build/use Busybox's man functionality yourself
2023-02-01 16:24:26 <psykose> a busybox without extra dependencies makes me think of some busybox-giga-turbo-xl
2023-02-01 16:24:38 <psykose> with a C compiler and web browser and terminal all baked in or something
2023-02-01 16:25:11 <minimal> psykose: imagine a wedding between busybox and systemd lol
2023-02-01 16:25:12 <psykose> busybox doesn't really implement manpage stuff and i don't recommend even using that one, just install the standard mandoc (and groff for that stuff if it's used elsewhere)
2023-02-01 16:25:22 <Fxzxmic> minimal: Just chatting, I didn't report an issue to Alpine.
2023-02-01 16:25:30 <elly> new alpine design, all packages are linked into one gargantuan busybox binary
2023-02-01 16:25:40 <minimal> systemd-busybox ;-)
2023-02-01 16:25:44 <elly> apk simply adds and removes symlinks to it
2023-02-01 16:25:59 <psykose> elly: if you think about it it almost makes sense until you realise you are reimplementing userspace inside a binary
2023-02-01 16:26:21 <psykose> full desktop inside a binary, lets go
2023-02-01 16:27:18 <minimal> unikernel desktop?
2023-02-01 16:28:14 <invoked> quick bug report, maybe: in the contents of '/etc/init.d/net.lo' from package 'netifrc-openrc', the path for SHDIR and MODULESDIR should be /usr/libexec/.. not /libexec/..
2023-02-01 16:28:15 <pj> minimal: i would like to get back to working on systemd package
2023-02-01 16:28:41 <minimal> Fxzxmic: was something unclear in the "nroff: not found" error message?
2023-02-01 16:29:21 <elly> psykose: you've heard about static linking, now get ready for ball-of-string linking where your entire OS is all linked together
2023-02-01 16:29:50 <Fxzxmic> minimal: No, I know exactly what happened, but I'm surprised.
2023-02-01 16:30:38 <minimal> Fxzxmic: surprised that nroff wasn't installed?
2023-02-01 16:31:42 <Fxzxmic> minimal: surprised for "<Fxzxmic> I thought busybox didn't need any extra dependence."
2023-02-01 16:32:04 <psykose> invoked: fixed
2023-02-01 16:32:18 <invoked> psykose: cheers
2023-02-01 16:32:59 <minimal> Fxzxmic: surprised as you thought that Busybox's ~"man" didn't need any extra dependencies? or that you thought that Busybox in general didn't need any extra dependencies?
2023-02-01 16:33:04 <psykose> Fxzxmic: extra dependence on what? by definition that does not make sense
2023-02-01 16:34:27 <Fxzxmic> minimal: The latter.
2023-02-01 16:35:04 <psykose> busybox implements a few utilities not the entirety of linux
2023-02-01 16:35:21 <psykose> type `busybox` to find the list
2023-02-01 16:35:39 <psykose> a few things are disabled in the alpine build, but practically everything is enabled that is useful
2023-02-01 16:35:46 <psykose> (a few things are in busybox-extras instead)
2023-02-01 16:35:49 <minimal> Fxzxmic: well for a start the Alpine Busybox package is depend on the musl package lol
2023-02-01 16:35:59 <minimal> s/depend/dependant/
2023-02-01 16:35:59 <alpine-meetbot> minimal meant to say: Fxzxmic: well for a start the Alpine Busybox package is dependant on the musl package lol
2023-02-01 16:37:50 <Fxzxmic> psykose: I mean that the applet it has implemented does not need external dependencies, like nroff, not the whole Linux.
2023-02-01 16:38:43 <Fxzxmic> minimal: By dependency, I don't mean libraries, but programs.
2023-02-01 16:38:48 <minimal> Fxzxmic~: as I pointed out Alpine's busybox requires musl....that's a rather fundamental dependency
2023-02-01 16:39:25 <minimal> Fxzxmic: ah, but you're only now defining what you mean by dependency
2023-02-01 16:40:07 <minimal> if you search through the Busybox source there may well be other programs that it may depend on for some applets
2023-02-01 17:02:07 <us3rbd> Can i get help here?
2023-02-01 17:02:37 <ikke> us3rbd: yes, if you ask a question
2023-02-01 17:06:02 <us3rbd> Okay thanks. and also thanks to devs for this awesome alpine os. My problem is everytime i boot my pc i have to stop networking and wpa_supplicant from rc-service and restart networkmanager to connect to wifi in kde plasma. 
2023-02-01 17:08:06 <psykose> you're supposed to remove wpa_supplicant and networking from the default services to use nm
2023-02-01 17:08:27 <psykose> rc-update del -a networking
2023-02-01 17:08:31 <psykose> rc-update del -a wpa_supplicant
2023-02-01 17:09:10 <us3rbd> Thanks @psykose :)
2023-02-01 17:13:59 <us3rbd> Its working :-D
2023-02-01 17:20:44 <boombim> how to run polkit gnome agent on alpine? I got this error:
2023-02-01 17:20:45 <boombim> https://0x0.st/oF5y.txt
2023-02-01 17:29:23 <psykose> no idea but generally that stuff is elogind
2023-02-01 17:48:54 <scorpion2185[m]1> psykose it didn't print it. do I need -v?
2023-02-01 17:49:17 <psykose> either it never ran the line you wrote or you didn't read your terminal
2023-02-01 17:49:22 <psykose> abuild does not mangle the output
2023-02-01 17:49:34 <psykose> if you type
2023-02-01 17:49:36 <psykose> build() {
2023-02-01 17:49:37 <psykose> pwd
2023-02-01 17:49:37 <psykose> }
2023-02-01 17:49:40 <psykose> you will get a pwd
2023-02-01 17:49:51 <psykose> anything else is your own code doing something wrong
2023-02-01 17:52:22 <scorpion2185[m]1> it's the same a line after pwd
2023-02-01 17:52:32 <scorpion2185[m]1> ah wait
2023-02-01 17:52:39 <scorpion2185[m]1> sorry I missed it
2023-02-01 21:25:59 <boombim[m]> <psykose> "no idea but generally that stuff..." <- I've configured sway with seatd by alpine sway wiki 
2023-02-01 21:26:09 <boombim[m]> Looks like it was changed
2023-02-01 21:26:13 <boombim[m]> Strange
2023-02-01 21:26:38 <boombim[m]> Is seatd not recommended anymore?
2023-02-01 21:26:44 <boombim[m]> It was...
2023-02-01 21:26:46 <psykose> it's not related to using elogind or not
2023-02-01 21:27:48 <boombim[m]> May you clarify it please. I swear it was recommended option instead of elogind
2023-02-01 21:27:49 <boombim[m]> Before
2023-02-01 21:28:11 <psykose> there is no "instead of"
2023-02-01 21:28:56 <psykose> sway requires seatd
2023-02-01 21:29:03 <psykose> one can use elogind (as well) if they want
2023-02-01 21:30:31 <boombim[m]> Stop. As I know sway can work with elogind without seatd
2023-02-01 21:31:15 <psykose> i have no idea where you "know" that from but i can repeat that you are wrong if you want
2023-02-01 21:32:40 <psykose> or was it libseat actually
2023-02-01 21:33:08 <psykose> ah, indeed it was that direction
2023-02-01 21:33:32 <psykose> good point, indeed it's one of them
2023-02-01 21:34:00 <psykose> thinking about it again i don't think it would fix the polkit agent, but you can try
2023-02-01 21:34:47 <boombim[m]> So no differences between seatd and logind in terms of goals?
2023-02-01 21:34:59 <boombim[m]> elogind*
2023-02-01 21:35:45 <psykose> there are quite a few differences but you could find out what they do and what they are yourself, i don't have a good reference myself
2023-02-01 21:38:20 <boombim[m]> Got it.
2023-02-01 21:38:21 <boombim[m]> Thanks anyway
2023-02-01 21:39:11 <psykose> in this specific case it might work by magic or might not :-)
2023-02-01 21:39:27 <psykose> aside from that i never got the weird pk stuff to work myself
2023-02-01 22:04:48 <boombim[m]> I just would like to try this app and it says that it should be started with pkexec
2023-02-01 22:04:52 <boombim[m]> https://github.com/waycrate/swhkd
2023-02-01 22:08:53 <psykose> yes, and indeed that pkexec doesn't work
2023-02-01 22:19:41 <boombim[m]> Yes...
2023-02-01 22:34:14 <boombim[m]> Is situation with broken screen sharing in chromium changed somehow? It is broken several months...
2023-02-01 22:51:38 <Quillith> question, is it possible to install an application via tarball? I downloaded one from github, decompressed it, changed permissions, and tried to run, but it didn't work. Figured I'd ask if it was doable before I spent more time on it
2023-02-01 22:52:19 <Borley> Probably that tarball was a glibc program, Alpine is musl
2023-02-01 22:52:51 <Quillith> Good point, I'd forgot about that
2023-02-02 07:32:41 <hellno> After installing the last update (the kernel was updated), after rebooting I got this error: mount: mounting UUID=blah-*-blah on /systemroot failed: no such device Mounting root failed. initramfs emergency recovery shell launched
2023-02-02 07:32:59 <hellno> alpine 3.17, how to fix this?
2023-02-02 07:34:36 <ikke> What kind of storage do you have? 
2023-02-02 07:36:09 <hellno> ikke: hdd disk, sda1 for / on ext4, sda2 for other ext4 too
2023-02-02 07:39:27 <aron> hellno: can you mount it manually from the recovery shell? maybe ext4 module is missing?
2023-02-02 07:42:34 <hellno> aron: i can mount it if boot from usb flash
2023-02-02 07:42:47 <hellno> keyboard not work in recovery shell 
2023-02-02 07:45:00 <hellno> but not sure how to fix it :)
2023-02-02 07:56:45 <hellno> aron: ext4 modules are in place, nothing is disabled
2023-02-02 07:58:59 <ikke> Maybe something went wrong generating the new initramfs 
2023-02-02 08:15:41 <hellno> ikke: how to fix this? im not famous with all this staff
2023-02-02 08:15:55 <hellno> i can boot from usb and what next?
2023-02-02 08:16:07 <hellno> mount root and
2023-02-02 08:28:45 <hellno> For some reason, mkinitfs was trying to find modules from a previous kernel version. i did a chroot and just reinstalled the kernel and everything worked. How could this happen? Other than updating packages, I haven't touched anything on this system.
2023-02-02 10:01:06 <pj> does ash has pipefail?
2023-02-02 10:01:20 <pj> asking for a friend
2023-02-02 10:05:35 <aron> pj: I got no error for "set -o pipefail"
2023-02-02 10:06:36 <pj> yeah, me too, just a sanity check :)
2023-02-02 10:12:59 <ikke> Yes, it has 
2023-02-02 12:16:52 <ptrc> did anyone ever get "ip: can't send flush request: Address not available" when trying to use ifdown?
2023-02-02 12:17:05 <ptrc> i'm trying to figure out what's even happening, and google isn't really helpfuk
2023-02-02 12:17:08 <ptrc> helpful*
2023-02-02 12:17:09 <ikke> Don't recall
2023-02-02 14:19:28 <elly> is it time for strace?
2023-02-02 15:47:37 <scorpion2185[m]1> https://paste.centos.org/view/47cbf210 error when building
2023-02-02 15:48:05 <scorpion2185[m]1> https://github.com/calculon102/xgifwallpaper.git
2023-02-02 15:58:40 <scorpion2185[m]1> never mind
2023-02-02 15:58:47 <scorpion2185[m]1> I wonder why that error on alpine
2023-02-02 16:16:21 <macktronics> anyone have audio break after updating yesterday / today using firefox?
2023-02-02 16:20:28 <psykose> scorpion2185[m]1: because the code is wrong
2023-02-02 16:22:39 <tjohej> hey, is it possible to backup the whole linux install with a diskless config, not just /etc?
2023-02-02 16:23:31 <macktronics> okay, now it just started working after I installed pipewire-alsa; never mind :-)
2023-02-02 16:23:42 <ikke> tjohej: with lbu add you can specify more files / directories that should be included in the apkvol
2023-02-02 16:23:46 <ikke> apkovl
2023-02-02 16:24:01 <ikke> tjohej: but, you should not just do lbu add /
2023-02-02 16:27:08 <tjohej> ikke why not?
2023-02-02 16:28:13 <ikke> at boot, apk will automatically install packages from /etc/apk/world, including those files in the overlay would cause conflicts
2023-02-02 16:29:27 <tjohej> so you say Iwould add the rest and exlude /etc/?
2023-02-02 16:32:40 <tjohej> ikke:
2023-02-02 16:35:05 <pj> you need to not add anything that is installed by packages
2023-02-02 16:35:24 <ikke> and typically you don't want to add things in /tmp and /run either
2023-02-02 16:35:47 <pj> and etc can be included because it will create just .apk-new files
2023-02-02 16:36:10 <pj> although you probably don't want whole etc either
2023-02-02 16:36:20 <ikke> /etc is include by default
2023-02-02 16:36:21 <pj> just the files that you changed
2023-02-02 16:47:28 <tjohej> pj: do I have to be aware of amount of RAM and total file sizes I guess?
2023-02-02 16:56:12 <tjohej> ikke: what happens if not enough ram at boot?
2023-02-02 16:57:06 <tjohej> ikke: ah sorry, realized it will not even fit before during upload
2023-02-02 17:18:25 <FinkleWrinkle> Hello, I am trying to expand the filesystem of my RPi SD card (/dev/mmcblk0p1) remotely but am struggling to get the filesystem size to match the partition size.  I can expand the partition successfully using sfdisk but the filesystem size remains the same (comparing output of fdisk and df).  resize2fs doesn't seem to be the ticket as it's a vfat partition.  Alpine does match them when doing a fresh install 
2023-02-02 17:18:25 <FinkleWrinkle> but I can't seem to have them match with an existing system.  Any suggestions?
2023-02-02 17:19:59 <Habbie> why do you want to expand p1?
2023-02-02 17:20:10 <Habbie> (usually people want to expand p2)
2023-02-02 17:23:23 <ikke> FinkleWrinkle: there is fatresize in testing
2023-02-02 17:26:44 <FinkleWrinkle> No p2 as it's diskless. The p1 partition was setup to be as small as possible for whatever reason.  Trying to bump it up to match the ram size and allow new apks to be cached.
2023-02-02 17:27:06 <FinkleWrinkle> I will check fatresize out.
2023-02-02 17:32:23 <boombim> Is it possible to run sway with elogind only without seatd?
2023-02-02 17:33:40 <boombim> according to void linux doc user need to set one of seatd or elogind
2023-02-02 17:34:00 <boombim> but looks like alpine doesnt work with elogind only
2023-02-02 17:34:34 <macktronics> i use sway + seatd only.   I login from the tty by 'dbus-run-session sway'
2023-02-02 17:35:19 <boombim> Yeat it work for me as well but I cannot run polkit with seatd
2023-02-02 17:37:33 <macktronics> ack, understand.   i don't have polkit so didn't run into that - good luck
2023-02-02 18:01:12 <psykose> you enable elogind and you disable seatd
2023-02-02 19:01:15 <scorpion2185[m]1> I can't use blueman-applet
2023-02-02 19:02:16 <scorpion2185[m]1> https://paste.centos.org/view/0d17136e
2023-02-02 19:08:43 <ptrc> looks like blueman crashes
2023-02-02 19:18:12 <scorpion2185[m]1> is there something likt tpaste but for sharing files?
2023-02-02 21:28:31 <elagost> scorpion2185[m]1: uguu.se or catbox.moe? Both have APIs.
2023-02-02 22:09:34 <pj> transfer.sh or gofile.io
2023-02-03 07:25:01 <bulle_> hey, how can I control how my drives are named (sda etc). its not always consistent and running diskless gives a WARNING when the usb with ISO on is not named the same at boot and running apk
2023-02-03 07:28:06 <psykose> probably using uuids instead
2023-02-03 07:28:31 <bulle_> where do I edit this?
2023-02-03 14:40:44 <takkie> o/
2023-02-03 14:44:58 <takkie> hi all, looking for some help or push into right direction.
2023-02-03 14:45:21 <takkie> I'm running alpine in a container and have an host on the same network with 192.168.x.x address.
2023-02-03 14:45:33 <takkie> nslookup can resolve the hostname fine, but curl cannot. Even `curl -H "Host: somehost.name" http://192.168.x.x` trows an error not being able to resolve a hostname, eventhough I try to connect to an IP.
2023-02-03 16:17:09 <lopid> docker?
2023-02-03 16:17:27 <lopid> weird
2023-02-03 17:02:48 <bovis> Has anyone here had success with creating custom maps using packages available in Alpine? My efforts with packages not in the repository (python-mapnik, tileserver, tilemill, qgis) have ended in dependency hell with lots of installation errors. Mapnik is available and installs fine, but without python or ruby bindings working I haven't made any progress.
2023-02-03 17:03:01 <bovis> Is there a package I'm not finding? How do you do it?
2023-02-03 17:28:09 <invoked> no clue about mapnik, but looking at their repo, it wants python 2.4-2.7
2023-02-03 17:30:03 <bovis> invoked: Unfortunately, it seems that all the user-created addons to Mapnik are old. Ruby bindings are from 2015. Python was 2017 if I remember correctly. Neither installs well when you try it today.
2023-02-03 17:30:30 <bovis> It's like there was a huge surge of interest around 2015. Then nothing.
2023-02-03 17:30:56 <invoked> well, yeah, it's probably not going to work with python3 at any rate
2023-02-03 20:14:16 <Saijin_Naib> <bovis> "Has anyone here had success with..." <- I want to package OGC stuff, but I'm not doing well so far. Geoserver or QGIS-server, ideally 
2023-02-03 20:14:26 <Saijin_Naib> Tons to package for that 
2023-02-03 20:15:02 <Saijin_Naib> Leaflet not up to the task? 
2023-02-03 20:31:29 <bovis> Saijin_Naib: Yes QGIS looks great but has a ton of dependencies. I hadn't resorted to trying to build/package it since it's so large.
2023-02-03 20:32:04 <bovis> Leaflet looks interesting. I hadn't found that yet. Seems like it has all the features I'd want.
2023-02-03 20:33:05 <Saijin_Naib> It's phenomenal, and has tons of add-ons to provide extra features 
2023-02-03 20:33:35 <Saijin_Naib> If you don't need crazy stuff and are serving static, I don't think it can be beat 
2023-02-03 20:34:12 <Saijin_Naib> If you're doing 3D, maybe PoTree? 
2023-02-03 20:34:30 <Saijin_Naib> Not a map, but a viewer 
2023-02-03 20:35:08 <bovis> Saijin_Naib: I don't want to host on a website for public viewing. Can it be run on a local server and accessed through my browser?
2023-02-03 20:36:21 <Saijin_Naib> Which? 
2023-02-03 20:36:33 <Saijin_Naib> Both, as far as I've used them 
2023-02-03 20:36:43 <Saijin_Naib> That's what we use for OpenDroneMap 
2023-02-03 20:37:12 <Saijin_Naib> Leaflet and PoTree hosted locally and accessed by localhost 
2023-02-03 20:38:30 <Saijin_Naib> We render GeoTIFF to OGC Cloud Optimized GeoTIFF so it has support for http range requests and streaming, as well as internal tiles/overviews
2023-02-03 20:40:51 <Saijin_Naib> Another good option could be OGC GeoPackage, which can hold both vector and raster (and aspatial tabular) in whatever CRS with tiles, and it's basically amped up SQLite3 or SpatialLite, so you have single file DB without needing a full DB server 
2023-02-03 20:42:04 <Saijin_Naib> These are all on my list to learn how to build and package 🤣
2023-02-03 20:42:05 <Saijin_Naib> Long list 
2023-02-03 20:43:29 <bovis> Saijin_Naib: My only goal currently is to view tiles and create output (static png, svg) in a style I create, like turning the roads all black, orange water, yellow bridges or whatever style I prefer.
2023-02-03 20:44:21 <bovis> Fetching areas with OSM is easy enough on its website, but I can't separate the labels from the roads or change the color scheme.
2023-02-03 20:44:27 <Saijin_Naib> OGC GeoPackage also has a spec for storing styles 
2023-02-03 20:45:30 <Saijin_Naib> Hmm 🤔 ODM should also have a vector tile endpoint, unless I'm misremembering
2023-02-03 20:45:31 <Saijin_Naib> OSM, rather 
2023-02-03 20:47:04 <Saijin_Naib> That's what MapTiler provides. And you can override styles client side. I've done it with QGIS 
2023-02-03 20:51:17 <bovis> Saijin_Naib: Thanks for all the information. I'll try out Leaflet later today.
2023-02-03 20:52:34 <Saijin_Naib> Any time! 
2023-02-04 00:45:13 <Quillith> Hi, when I run Steam via flatpak I get a lot of errors about wrong ELF class (elfclass32). It doesn't seem to be keeping a game from running, but I don't want to assume an error is meaningless and ignore it. I have installed proton via flatpak and instructed steam to run the game through proton. Are there 32-bit libraries I should be installing? I don't see mention on the wiki page.  I tried 
2023-02-04 00:45:19 <Quillith> installing all packages with "mesa" in the name (AMD GPU), but that doesn't seem to have helped
2023-02-04 00:51:26 <psykose> flatpak does not use any host libraries
2023-02-04 00:51:48 <psykose> you also can't install any 32-bit libraries on a 64-bit host
2023-02-04 00:52:00 <Quillith> So this is less a 'my computer' problem and more a 'steam flatpak' problem
2023-02-04 00:52:39 <psykose> it is a "no problem"
2023-02-04 00:53:42 <Quillith> gotcha, thank you. Is it bad to have multiple mesa packages installed? I learned on my first day with alpine that too many audio managers cause errors; does that also apply to video drivers?
2023-02-04 00:55:03 <psykose> it does not, and all the gl drivers are just in mesa-dri-gallium
2023-02-04 00:55:16 <psykose> the vulkan drivers are the split ones i guess
2023-02-04 00:55:43 <Quillith> Okay, thank you
2023-02-04 01:07:55 <Quillith> I'm seeing on the Steam wiki page that the 'eventfd: Too many open files' error is fixed by raising user limits, which requires installing PAM and switching to a PAM-enabled login. I installed util-linux-login, but I'm having trouble finding instructions on how to switch my login from default to util-linux-login
2023-02-04 01:08:55 <psykose> you have it after installing it
2023-02-04 01:09:00 <psykose> nothing to switch
2023-02-04 01:12:33 <Quillith> Oh, perfect, thank you
2023-02-04 09:41:40 <deluga90> Hey folks I've got a problem. I'm rescuing a Dell Latitude D810 from the dead and are currently trying to get Wifi to work, but the driver keeps crashing on startup, any ideas? 
2023-02-04 09:41:41 <deluga90> /home/elisabeth # lspci -kk | tail -n 2
2023-02-04 09:41:43 <deluga90> 03:03.0 Network controller: Intel Corporation PRO/Wireless 2200BG [Calexico2] Network Connection (rev 05)
2023-02-04 09:41:45 <deluga90> 	Subsystem: Intel Corporation Dell Latitude D600
2023-02-04 09:41:47 <deluga90> /home/elisabeth # dmesg | grep ipw2200
2023-02-04 09:41:49 <deluga90> [   26.827679] ipw2200: Intel(R) PRO/Wireless 2200/2915 Network Driver, 1.2.2kmprq
2023-02-04 09:41:51 <deluga90> [   26.827689] ipw2200: Copyright(c) 2003-2006 Intel Corporation
2023-02-04 09:41:53 <deluga90> [   26.828043] ipw2200: Detected Intel PRO/Wireless 2200BG Network Connection
2023-02-04 09:41:55 <deluga90> [   26.828132] ipw2200 0000:03:03.0: Direct firmware load for ipw2200-bss.fw failed with error -2
2023-02-04 09:41:57 <deluga90> [   26.828139] ipw2200: ipw2200-bss.fw request_firmware failed: Reason -2
2023-02-04 09:41:59 <deluga90> [   26.828143] ipw2200: Unable to load firmware: -2
2023-02-04 09:42:01 <deluga90> [   26.828309] ipw2200: probe of 0000:03:03.0 failed with error -5
2023-02-04 09:49:53 <deluga90> I might want to add that the driver was working properly on the Ubuntu 16.04 the system had on it before 
2023-02-04 10:29:44 <bulle_> hey, how do I add /var to my lbu? is it 'lbu include var' ?
2023-02-04 10:32:42 <deluga90> lbu add/include /var
2023-02-04 10:32:44 <deluga90> yes
2023-02-04 10:33:12 <bulle_> add and include is the same?
2023-02-04 10:33:20 <deluga90> yes they are the same
2023-02-04 10:33:38 <bulle_> is there a config where this can be edited?
2023-02-04 10:34:36 <bulle_> why is bot home and home/user in the list?? if home is in the list then no need for a specific user?
2023-02-04 10:35:07 <deluga90> /etc/lbu/lbu.conf but im just guessing at this point.. details here -> https://wiki.alpinelinux.org/wiki/Alpine_local_backup
2023-02-04 10:36:30 <deluga90> from my experience adding the folder containing the others is enough 
2023-02-04 10:38:28 <bulle_> lbu.conf is very limited
2023-02-04 10:38:43 <bulle_> but maybe I am missing something
2023-02-04 10:40:53 <bulle_> how do I list folders that is listed for commit, not only the incremental difference?
2023-02-04 10:41:52 <bulle_> is it lbu list?
2023-02-04 10:44:30 <bulle_> I did lbu include /var, bit it doas not add everything, its like it did a selection of folder....why???
2023-02-04 11:17:25 <dwfreed> deluga90: as your dmesg indicates, you need the intel wifi firmware
2023-02-04 12:09:01 <deluga90> dwfreed -> ipw2200 is the kernel module/driver for my card 
2023-02-04 12:09:31 <deluga90> i guess because it's a really old intel card (?)
2023-02-04 12:10:29 <deluga90> apk list | grep linux-firmware-intel 
2023-02-04 12:10:29 <deluga90> linux-firmware-intel-20230117-r0 x86 {linux-firmware} (custom:multiple) [installed]
2023-02-04 12:15:48 <dekedro> deluga90: looks like ipw2200 requires separate non-free firmare https://wiki.debian.org/ipw2200, I think it is not packaged in Alpine
2023-02-04 12:39:46 <deluga90> thanks for putting me on the right track, I'll try to figure it out 
2023-02-04 12:50:25 <dekedro> I tried this: https://github.com/VorpalBlade/ipw2x00-firmware, it loads on my system with FIRMWAREDIR=/lib/firmware
2023-02-04 16:09:19 <bulle_> hey, I cant fint grub.conf. DO you kow where its to to be edit?
2023-02-04 16:10:20 <quinq> Maybe you're rather thinking about /boot/grub/grub.cfg
2023-02-04 16:27:15 <bulle_> quinq: yes! I guess its not there because diskless mode. is there a different way to edit in this case?
2023-02-04 16:29:03 <bulle_> where can I find grub.cfg in diskless mode??
2023-02-04 16:32:47 <bulle_> now I found it in my ISO on the boot USB, but is there a way to edit this or have a file that is read instead of this file at boot?
2023-02-04 16:33:10 <bulle_> because this file is read only...
2023-02-04 17:11:01 <quinq> That's the concept of diskless…
2023-02-04 19:26:55 <elly> this is somewhat odd: a USB stick that previously worked fine is behaving very strangely now; it shows as having a capacity of 28GB instead of 32GB, and trying to mount it yields "I/O error" with dmesg complaining about timeouts
2023-02-04 19:27:02 <elly> the same stick was working fine a couple of days ago
2023-02-04 19:28:13 <elly> also, after the failed mount the device disappears
2023-02-04 19:30:39 <elly> this device, which I know for a fact contains a valid filesystem, reads as all 0s when I try to dd from it as well
2023-02-04 19:30:44 <elly> extremely suspect
2023-02-04 19:32:04 <Ermine> Maybe it broke during these two years?
2023-02-04 19:32:08 <Ermine> days
2023-02-04 19:32:20 <elly> it's possible
2023-02-04 19:33:25 <Ermine> I'd check with other OSes/distros/kernel versions
2023-02-04 19:39:26 <elly> yeah, it doesn't seem to work in my Win11 device either
2023-02-04 19:39:29 <elly> well, that's irritating
2023-02-04 19:42:11 <quinq> Try on a lower USB controller version
2023-02-04 19:42:35 <Waitingfortherain[m]> sometimes this happens
2023-02-04 19:42:45 <Waitingfortherain[m]> Its quite unfortunate
2023-02-04 19:42:48 <elly> quinq: how does one do that?
2023-02-04 19:43:23 <Waitingfortherain[m]> I knew a few people who had this problem, best thing you can do is use another one
2023-02-04 19:43:23 <quinq> You plug it in another socket, connected to a controller with a lower USB version
2023-02-04 19:43:31 <elly> oh
2023-02-04 19:43:39 <quinq> Maybe you'll be able to make a backup of it before it gets worse
2023-02-04 19:43:42 <elly> Waitingfortherain[m]: yeah, it's just irritating :) it is one of my backup devices
2023-02-04 19:43:48 <quinq> ah ^^
2023-02-04 19:43:49 <elly> quinq: fortunately all it has on it is backup tarballs anyway
2023-02-04 19:44:18 <quinq> Just out of curiosity, do you make a heavy usage of it?
2023-02-04 19:44:52 <elly> no, I write new backups to it every week or so
2023-02-04 19:45:03 <elly> which are like, 4 tarballs
2023-02-04 19:45:16 <quinq> Yeah, that doesn't sound like wear
2023-02-04 19:46:40 <vkrishn> elly: first thing  advisable is dd if=/dev/sda of=sda.backup.img (on hdd)
2023-02-04 19:46:57 <elly> it is literally a stick already used for storing backups, I have other copies of these :P
2023-02-04 19:47:03 <elly> I am just puzzled by its failure mode
2023-02-04 19:47:41 <quinq> Yeah, usually crappy USB stick are crappy from the start, but I'm not USB storage expert
2023-02-04 19:49:08 <vkrishn> is it a micro-sd card in adapter ?
2023-02-04 19:49:25 <elly> it is a usb stick
2023-02-04 19:51:31 <Apachez> are you using the ram edition of alpine?
2023-02-04 19:51:37 <Apachez> highly recommended if you run from usb drive
2023-02-04 19:51:58 <Apachez> it will just read the content to ram and run from there and only write to it if you do lsbackup
2023-02-04 19:52:03 <Apachez> or whatever the sync command
2023-02-04 19:52:35 <Apachez> other than that there are "industrial" micro-sd and usb drives which claims extended endurance for writes but meh
2023-02-04 19:52:40 <Apachez> I prefer samsung or sandisk
2023-02-04 19:52:46 <Apachez> avoiding kingston like plague :)
2023-02-04 19:55:21 <elly> no, I am not using the ram edition of alpine :P
2023-02-04 19:57:45 <md1> i
2023-02-05 00:55:26 <EtherMan> So umm... In which package for 3.17 would I find the netlink_dev or ethertap kernel modules? Or how do I get a working tap device on alpine? I'm not really finding anything for it on google :/
2023-02-05 01:09:13 <psykose> probably `modprobe tap`
2023-02-05 01:09:34 <psykose> or tun
2023-02-05 01:09:35 <psykose> or both
2023-02-05 01:10:01 <psykose> i was initially confused because "netlink dev ethernet kernel module" is something from like the 2.6 days haha
2023-02-05 01:11:32 <EtherMan> Yea I'm just pointing to the two names I generally know them as. But neither exists. tun and tap does, but neither result in a /dev device for either.
2023-02-05 01:13:31 <EtherMan> lsmod doesn't list either either after a modprobe :/
2023-02-05 01:13:41 <psykose> https://docs.kernel.org/networking/tuntap.html
2023-02-05 01:31:34 <dwfreed> generally /dev/net/tun will exist regardless of whether the modules are loaded, and they'll be loaded on usage
2023-02-05 01:37:14 <EtherMan> Sorry, I'm a dumbass.... It probably helps if the two ssh connections are actually against... you know, the comp I'm trying to actually configure and not the one have one of them be a switch ^_^;
2023-02-05 01:45:41 <dwfreed> lol
2023-02-05 01:52:48 <ForeverNoob[m]> Oh no, the Alpine wiki switched to the new design too? :(
2023-02-05 04:53:31 <novorol[m]> is Enlightenment unavailable on Alpine due to lack of maintainer or some musl incompabilities?
2023-02-05 04:54:16 <novorol[m]> <ForeverNoob[m]> "Oh no, the Alpine wiki switched..." <- stupid mediawiki soydevs
2023-02-05 04:55:02 <dwfreed> let's not attack the mediawiki devs, shall we
2023-02-05 04:55:04 <novorol[m]> 1. increased the default font size
2023-02-05 04:55:04 <novorol[m]> 2. added sticky table headers
2023-02-05 04:55:04 <novorol[m]> If they really cared about accesibility, they would've:
2023-02-05 04:56:49 <novorol[m]> I'd personally love to port alpine wiki to gopher or gemini
2023-02-05 05:00:50 <novorol[m]> though I find it true for any website that things are okay as long as any page is <1MB sans media and most of the stuff works w/o JS
2023-02-05 05:02:41 <psykose> ForeverNoob[m]: don't think it's very new, it's been the default for a while
2023-02-05 05:03:20 <psykose> since 1.38 i guess
2023-02-05 05:06:47 <epad> Hi, I am trying to run renoise that has glibc as dependencie, but I installed gcompat but get error: Error loading shared library libmvec.so.1      any ideas?
2023-02-05 05:09:19 <psykose> it depends on libmvec.so.1 and that doesn't exist
2023-02-05 05:09:52 <epad> sorry I dont understand? 
2023-02-05 05:10:22 <psykose> you can't run that
2023-02-05 05:10:59 <raven523> I suppose gcompat doesn't provide a replacement for libmvec.so.1. It looks like it's part of glibc
2023-02-05 05:11:22 <epad> I tried using void linux with glic and it workd. but you say even with gcompat I am missing something?
2023-02-05 05:12:00 <epad> So I am doomed to fail in this case then....
2023-02-05 05:12:12 <psykose> gcompat doesn't magically let you run everything, yes
2023-02-05 05:12:22 <psykose> gcompat doesn't implement libmvec from glibc
2023-02-05 05:12:50 <epad> ok. do you see any way to make it work on Alpine?
2023-02-05 05:13:00 <novorol[m]> vm
2023-02-05 05:13:20 <psykose> you most likely cannot run it on alpine at all, no
2023-02-05 05:13:22 <ForeverNoob[m]> @epad: Running closed source glibc-dependent programs will always be a challenge on musl systems. Try to see if there's a Flatpak available somewhere. That, or try out other DAWs like Bitwig Studio (which does coincidentally have a Flatpak available :))
2023-02-05 05:13:26 <epad> hu....why would a wm work differently??
2023-02-05 05:13:40 <psykose> because a vm is a separate system and not alpine anymore
2023-02-05 05:14:22 <epad> I see
2023-02-05 05:14:58 <epad> Reaper works but not Renoise....damn. I will try wm then.
2023-02-05 05:15:45 <epad> Renoise are supposed to be open source btw
2023-02-05 05:16:03 <ForeverNoob[m]> *VM
2023-02-05 05:16:21 <ForeverNoob[m]> And Renoise is definitely not open source, nor is Reaper for that matter (or Bitwig Studio)
2023-02-05 05:18:45 <epad> what, reaper is not open source?
2023-02-05 05:23:34 <psykose> almost no daw is
2023-02-05 05:24:43 <psykose> ardour is i guess
2023-02-05 05:25:21 <psykose> there's also zrythm
2023-02-05 05:31:32 <epad> thanks
2023-02-05 05:37:32 <ForeverNoob[m]> This too, which is loosely inspired by Bitwig Studio - https://github.com/MeadowlarkDAW/Meadowlark - But unfortunately it's like pre-pre-pre alpha at this point.
2023-02-05 05:37:51 <ahills> anyone running alpine on digitalocean have trouble getting a dhcp lease?
2023-02-05 05:38:29 <ForeverNoob[m]> @ahills: What command are you running in order to get a lease?
2023-02-05 05:38:54 <ahills> ForeverNoob[m]: udhcpc, via the networking service
2023-02-05 05:39:30 <ForeverNoob[m]> I usually run `doas udhcpc -fq -i <iface>` so that I can see stuff happening in stdout.
2023-02-05 05:39:37 <ahills> I'm wondering if I built this image without the appropriate modules
2023-02-05 05:40:25 <ahills> it broadcasts discover a few times and then fails to get a lease, never hears back
2023-02-05 05:55:11 <ahills> aha, the network config comes through cloud-init, not dhcp, my bad
2023-02-05 06:11:44 <novorol[m]> <ahills> "anyone running alpine on..." <- digitalocean generally sucks
2023-02-05 06:22:12 <ahills> yeah, just trying to make a particular situation better with alpine
2023-02-05 07:19:05 <Saijin_Naib> Oh, that's news to me. Why does it suck? 
2023-02-05 07:40:57 <ahills> well, it's a bit inconvenient to run a nicer linux like alpine, for one...
2023-02-05 09:16:55 <Saijin_Naib> Ahhh, okay. Focused on Debian-derived? 
2023-02-05 09:21:54 <deluga90> Hi to the one that gave me the hint yesterday how to resolv my problem regarding the ipw2200 wifi driver, sorry I forgot your name -> what you'd suggested worked like a charm and you saved me hours of fiddeling, thank you very much!
2023-02-05 10:36:36 <mendelmaleh> hi, anyone run alpine on a macbook? i'm trying to install on a 2019 mba (intel), the installer image boots but the built in keyboard doesn't work. an external one works, but it's not a good long term solution xD
2023-02-05 14:42:02 <elly> I have alpine running on a macbook but it's significantly older than 2019
2023-02-05 14:42:08 <elly> and the keyboard did not give me any trouble
2023-02-05 14:55:13 <minimal> ahills: Hi, I'm the Alpine maintainer for cloud-init. Do you have the DigitalOcean DataSource enabled for cloud-init?
2023-02-05 14:59:13 <wyk72> hello everyone
2023-02-05 14:59:27 <wyk72> can you please help me with an APK issue ?
2023-02-05 14:59:35 <wyk72> for reason unknown to me
2023-02-05 15:00:04 <wyk72> my 3.17 (ugraded) Alpine still gets a BROKEN-WORLD because it asks for python2 (?!)
2023-02-05 15:00:13 <wyk72> root@imm-prototype:~# apk fix --no-cache fetch http://dl-cdn.alpinelinux.org/alpine/v3.17/main/x86_64/APKINDEX.tar.gz fetch http://dl-cdn.alpinelinux.org/alpine/v3.17/community/x86_64/APKINDEX.tar.gz fetch http://dl-cdn.alpinelinux.org/alpine/edge/main/x86_64/APKINDEX.tar.gz fetch http://dl-cdn.alpinelinux.org/alpine/edge/community/x86_64/APKINDEX.tar.gz fetch http://dl-cdn.alpinelinux.org/alpine/edge/testing/x86_64/APKINDEX.tar.gz ERROR: unable to s
2023-02-05 15:00:23 <wyk72> ERROR: unable to select packages:   python2 (no such package):     required by: world[python2] root@imm-prototype:~#
2023-02-05 15:00:32 <wyk72> I am tusck and I do not know what to do
2023-02-05 15:00:37 <wyk72> stuck
2023-02-05 15:00:58 <ikke> what does `apk del python2` return?
2023-02-05 15:01:20 <ikke> you are btw combining edge with 3.17, which also is a source of problems
2023-02-05 15:01:49 <wyk72> I use edge and the rest with a tag (@edge) usually I don't mix things
2023-02-05 15:02:09 <wyk72> root@imm-prototype:~# apk del python2 OK: 4972 MiB in 925 packages
2023-02-05 15:02:36 <ikke> What's in your /etc/apk/world?
2023-02-05 15:05:22 <wyk72> oh
2023-02-05 15:05:34 <wyk72> deleting python2 worked
2023-02-05 15:05:41 <wyk72> it was so simple yet kinda obscure
2023-02-05 15:05:49 <wyk72> sorry for the newb question
2023-02-05 15:05:53 <ikke> No worry
2023-02-05 15:06:07 <ikke> glad you could fix it
2023-02-05 15:07:07 <minimal> wyk72: when you previously upgraded did you specify the "-a" option, i.e. "apk -a upgrade"?
2023-02-05 15:07:28 <wyk72> yeah
2023-02-05 15:07:33 <wyk72> I did
2023-02-05 15:08:46 <minimal> wyk72: also even with a "tag" you are still mixing some stuff from Edge with 3.17, which still may cause problems
2023-02-05 15:08:59 <wyk72> I know
2023-02-05 15:09:26 <wyk72> I am hacking something with postmarketOS
2023-02-05 15:09:44 <wyk72> and I needed pmbootstrap and got stuck with this python2 request (?)
2023-02-05 15:10:06 <minimal> wyk72: even more problematic then if you are installing postmarketOS packages also
2023-02-05 15:10:11 <wyk72> this system maybe has beeen upgraded way too much, I think it was from 3.12
2023-02-05 15:10:26 <wyk72> I should containerize it into a new one I guess
2023-02-05 15:10:30 <ikke> We have servers around from 3.11
2023-02-05 15:10:31 <wyk72> and start clean edge
2023-02-05 15:10:33 <wyk72> repo
2023-02-05 15:10:58 <psykose> don't have to
2023-02-05 15:15:48 <wyk72> I'll just go edge then
2023-02-05 15:16:15 <wyk72> upgrading now
2023-02-05 17:03:32 <bovis> My APKBUILD failed at the package() step, but prepare() and build() were successful. I still have the build directory. How can I run abuild again without building the package a second time?
2023-02-05 17:04:05 <Saklad5> Is there a way to separate stable, edge, and testing while having packages in testing capable of using package in edge?
2023-02-05 17:05:17 <Saklad5> That is, I have stable-main and stable-community untagged, edge-main and edge-community tagged with "edge", and edge-testing tagged with "testing".
2023-02-05 17:05:48 <Saklad5> If example-package@testing has dependencies that can only be satisfied by something in edge-main or edge-community, will it just fail?
2023-02-05 17:08:29 <ptrc> Saklad5: unless you manually add something from tagged edge
2023-02-05 17:08:37 <Saklad5> True
2023-02-05 17:08:50 <Saklad5> It's a shame you can't have multiple tags on a repository
2023-02-05 17:09:03 <ptrc> but mixing edge and stable is a really Not Good Idea(tm)
2023-02-05 17:09:27 <Saklad5> I'm aware, and I do it very selectively
2023-02-05 17:09:50 <Saklad5> And I imagine testing can break at any moment
2023-02-05 17:10:09 <ptrc> at some point you might be better off just upgrading to edge
2023-02-05 17:10:27 <ptrc> bovis: you can run `abuild rootpkg` if i remember correctly
2023-02-05 17:10:55 <Saklad5> ptrc: Why? Aren't they basically the same thing, given six months or so?
2023-02-05 17:11:32 <Saklad5> I only use edge with packages I monitor directly
2023-02-05 17:12:01 <Saklad5> And only when they release an update that is only in edge
2023-02-05 17:12:33 <ptrc> Saklad5: if you need to pull in shared libraries from edge community onto a stable system, chances are it will be fragile enough to break with the next update of any package in the chain
2023-02-05 17:12:48 <Saklad5> What do you mean?
2023-02-05 17:13:10 <ptrc> like, a package in edge-community linking against a newer version of some really common library
2023-02-05 17:13:42 <Saklad5> That'd only happen if the edge package required the newer version, right?
2023-02-05 17:14:03 <ptrc> indeed
2023-02-05 17:14:13 <Saklad5> But in that scenario it wouldn't keep the older version?
2023-02-05 17:15:03 <ptrc> apk would also be really dissatisfied that it cannot satisfy your current world
2023-02-05 17:15:21 <Saklad5> I imagine
2023-02-05 17:15:58 <Saklad5> It wouldn't happen automatically, though: anything I monitor manually is also pinned to non-breaking releases
2023-02-05 17:16:03 <bovis> ptrc: Thank you. That worked.
2023-02-05 17:16:21 <Saklad5> So I'd tell it was no longer satisfiable
2023-02-05 17:16:40 <psykose> the thing pinned is not the thing changing
2023-02-05 17:17:02 <Saklad5> I'm operating under the assumption that developers consider new dependency requirements breaking changes
2023-02-05 17:17:14 <psykose> they're not requirements
2023-02-05 17:17:41 <psykose> easy example: there is a package abc1234
2023-02-05 17:17:51 <psykose> abc1234 was linked against libwhatever.so.1
2023-02-05 17:18:09 <psykose> in edge libwhatever gets upgraded and broke the ABI, so it's now libwhatever.so.2
2023-02-05 17:18:25 <psykose> everything gets rebuilt without caring, and now depends on libwhatever.so.2
2023-02-05 17:18:37 <psykose> libwhatever.so.2 only exists in edge
2023-02-05 17:19:03 <psykose> if you try add abc1234 from edge it won't find the libwhatever.so.2 from anywhere
2023-02-05 17:19:33 <Saklad5> abc1234 wouldn't explicitly depend on libwhatever.so.2?
2023-02-05 17:19:36 <psykose> it does
2023-02-05 17:19:41 <Saklad5> Then it'd find it in edge
2023-02-05 17:19:46 <psykose> no, because you're not on edge
2023-02-05 17:19:53 <psykose> you did apk add abc1234@mytaggededge
2023-02-05 17:20:11 <Saklad5> APK's documentation clearly says it will use dependencies from the tagged repository if it has to
2023-02-05 17:21:02 <ptrc> iirc it's gonna say "that dep is masked in repo @edge-community", but i haven't used that feature for a while
2023-02-05 17:21:12 <Saklad5> I'm probably misunderstanding something here, but I can't see what it is
2023-02-05 17:21:28 <Saklad5> It seems like APK's tagged repository feature is intended specifically to make this fairly safe
2023-02-05 17:22:22 <psykose> idk what documentation that is but it's not how it works
2023-02-05 17:22:45 <psykose> and what safety guarantees apk gives you is not relevant to what alpine supports or not
2023-02-05 17:23:06 <psykose> this is what you get specifically https://img.ayaya.dev/W5Qafw2hJsxl
2023-02-05 17:23:08 <ptrc> ah, no, i'm wrong; apk just happily adds stuff if it's in the same repo
2023-02-05 17:23:17 <Saklad5> https://wiki.alpinelinux.org/wiki/Alpine_Package_Keeper#Repository_pinning
2023-02-05 17:24:03 <Saklad5> Either you're wrong, the wiki is wrong, or I'm misunderstanding the wiki
2023-02-05 17:24:27 <psykose> this is not apk documentation, but sure
2023-02-05 17:24:28 <ptrc> so... worst case, packages `something` and `otherthing` depend on libxyz, you add something@edge-community, libxyz gets pulled from @edge-community too, you add otherthing from stable, libxyz gets an ABI-breaking update, everything breaks
2023-02-05 17:25:16 <psykose> specifically it will only auto-upgrade it if it's in the same repo
2023-02-05 17:25:27 <Saklad5> Auto-upgrade?
2023-02-05 17:25:38 <psykose> if community depends on libwhatever.so.2 that is actually in main, it only fails instead
2023-02-05 17:26:00 <psykose> so, in that hypothetical scenario, the pinned package just doesn't upgrade
2023-02-05 17:26:18 <Saklad5> Which is what I want
2023-02-05 17:26:27 <psykose> i don't care what you want or not :p
2023-02-05 17:26:34 <psykose> i'm just telling you it's not supported whatsoever
2023-02-05 17:26:38 <Saklad5> I got that bit
2023-02-05 17:26:54 <psykose> i used to be more sympathetic until the roughly 35th issue report of "help why it broken no work"
2023-02-05 17:27:33 <Saklad5> Understandable
2023-02-05 17:27:54 <psykose> if you know what you're doing, it mostly works quite well and safer than any other equivalent hackyness i've done with other package managers
2023-02-05 17:28:16 <Saklad5> I find it refreshingly straightforward, really
2023-02-05 17:28:17 <psykose> in practice you have to know the corner cases and how to fix it and how the solver works, more or less
2023-02-05 17:28:23 <psykose> i.e.
2023-02-05 17:29:32 <Saklad5> The main corner case I foresee being a problem is different packages requiring different versions of the same dependency, but I'm pretty sure that could happen within a single repo too
2023-02-05 17:29:44 <psykose> if you know what https://img.ayaya.dev/Z3k2ARuZjzZV means and how to fix it without opening issues, go nuts
2023-02-05 17:29:58 <psykose> the other corner case is the invisible one- apk does not warn when it holds back upgrades that should be upgraded
2023-02-05 17:30:26 <Saklad5> If I needed it upgraded, I would probably be pinning it and forcing a failure that way
2023-02-05 17:30:42 <Saklad5> And security updates go in stable
2023-02-05 17:31:08 <psykose> there wouldn't be a failure is the point, `apk upgrade` will just say nothing, if there is an upgrade held back because some dependency constraint on it isn't satisfied
2023-02-05 17:31:22 <psykose> and the world dependency constraints take precedence
2023-02-05 17:31:32 <Saklad5> All of which is what I'm expecting
2023-02-05 17:32:04 <psykose> so, for instance, up until a bit ago, if you had `openssl3` in world (back when it was a separate package from openssl)
2023-02-05 17:32:22 <psykose> after the openssl3 rebuilds, it would actually keep everything related to openssl behind forever, because openssl3 doesn't exist anymore
2023-02-05 17:32:34 <psykose> same goes for any missing package i guess
2023-02-05 17:32:50 <Saklad5> I have my world file small enough that I'm manually managing it and keeping track of each thing I need
2023-02-05 17:33:30 <Saklad5> There's like, a dozen things in it on most of my servers
2023-02-05 17:33:47 <psykose> the other thing is there is no "stable, edge, and testing", it's more just "stable and edge", they're release branches with "repos", and testing is only in edge, but aside from that the branches are all that matter there
2023-02-05 17:34:02 <Saklad5> I get that
2023-02-05 17:34:34 <Saklad5> Main depends on nothing, community depends on main, testing has no rules
2023-02-05 17:35:03 <psykose> no, depends on community and main, just 3rd in the order
2023-02-05 17:35:23 <Saklad5> Things in testing can vanish at any moment, right?
2023-02-05 17:35:57 <ptrc> things in *edge* can vanish at any moment
2023-02-05 17:36:11 <Saklad5> I mean, yes, but that's less routine
2023-02-05 17:36:18 <Saklad5> Things in testing are EXPECTED to vanish
2023-02-05 17:36:25 <ptrc> it's not really common
2023-02-05 17:36:32 <Saklad5> They graduate to edge, right?
2023-02-05 17:36:39 <Saklad5> That is, main or community?
2023-02-05 17:36:46 <ptrc> testing is in edge
2023-02-05 17:36:48 <ptrc> but yes
2023-02-05 17:36:58 <ptrc> they often go to community from testing
2023-02-05 17:37:11 <psykose> or stay in testing forever
2023-02-05 17:37:22 <Saklad5> I'm expecting things pinned to testing to break if the package graduates to community
2023-02-05 17:37:41 <psykose> that doesn't change anything in the package, but the pin itself would, sure
2023-02-05 17:38:05 <Saklad5> For now I'm just keeping an eye on it, the thing I am using isn't exactly essential anyway
2023-02-05 17:38:44 <Saklad5> By the way, the reason I'm separating main/community and testing here is that I have them pinned that way
2023-02-05 17:40:06 <Saklad5> Is there a specific impetus for moving out of testing?
2023-02-05 17:40:16 <ikke> Having it in the next stable release
2023-02-05 17:40:43 <Saklad5> Ah, so packages *never* exist solely in edge
2023-02-05 17:40:55 <ikke> testing is never built for stable brances
2023-02-05 17:41:04 <Saklad5> I know
2023-02-05 17:41:08 <ikke> so testing only lives in edge
2023-02-05 17:41:20 <Saklad5> I meant moving from testing to main/community
2023-02-05 17:41:24 <ikke> when a stable release is made, it will be branched of whatever is in master 
2023-02-05 17:42:33 <ikke> so things are moved from testing to community or main so that it will be built for the stable release
2023-02-05 17:42:42 <ikke> (before it's branched)
2023-02-05 17:43:17 <psykose> if you want to see it even simpler, the branches are just git branches and the repos are just folders in the repo
2023-02-05 17:43:28 <psykose> it works exactly like how you would expect in that sense
2023-02-05 17:43:34 <Saklad5> I thought that was literally was this was
2023-02-05 17:43:39 <Saklad5> what this*
2023-02-05 17:43:44 <psykose> it is
2023-02-05 17:45:13 <Saklad5> So decisions about moving things from testing to main/community are made roughly every six months, in advance of a new stable branch?
2023-02-05 17:49:30 <psykose> closer to just whenever
2023-02-05 17:49:41 <Saklad5> OK, that was my initial take
2023-02-05 17:50:42 <Saklad5> Is there a rhyme or reason to it?
2023-02-05 17:51:00 <Saklad5> Does it get moved at the maintainer's discretion?
2023-02-05 17:52:14 <ikke> yes
2023-02-05 17:52:22 <ikke> sometimes a user requests it
2023-02-05 17:52:34 <psykose> sometimes it's a dependency of something else
2023-02-05 17:54:09 <Saklad5> When should a user request it?
2023-02-05 17:54:22 <Saklad5> When it's been a while? When it seems to work for them?
2023-02-05 17:55:14 <Saklad5> What's the etiquette?
2023-02-05 18:00:52 <ikke> There is no time requirement
2023-02-05 18:01:06 <ikke> So typically they have tested it and report everything is working for them
2023-02-05 18:01:24 <Saklad5> Then the user should make an issue for it?
2023-02-05 18:12:27 <ikke> Typically yes
2023-02-05 18:23:21 <Borley> Doing my first Alpine install today. Anything you wish you knew during your first run of Alpine?
2023-02-05 18:24:25 <Saklad5> Make sure you install an NTP/NTS client
2023-02-05 18:24:45 <Saklad5> And if you are running it on a VPS or other form of VM, use the virtual image
2023-02-05 18:24:49 <ikke> setup-alpine asks you to install one
2023-02-05 18:25:50 <Saklad5> I'd also recommend paying particular attention to how OpenRC and APK work
2023-02-05 18:26:09 <Saklad5> In the case of OpenRC, specifically how /etc/conf.d files work
2023-02-05 18:29:40 <ikke> Saklad5: there you go
2023-02-05 18:29:48 <ikke> it has been moved
2023-02-05 18:33:39 <Saklad5> Thanks, that was remarkably fast
2023-02-05 18:36:19 <Borley> Thanks guys. I am partial to ntpsec, so that wouldn't have been a problem.
2023-02-05 18:36:57 <minimal> AFAIK setup-ntp doesn't configure NTS
2023-02-05 18:38:08 <Saklad5> Which is why I mentioned it
2023-02-05 18:38:13 <minimal> oops, misread "ntpsec" as NTS doh!
2023-02-05 18:38:45 <Saklad5> Busybox doesn't do NTS
2023-02-05 18:39:04 <minimal> Saklad5: yes but there's always chrony
2023-02-05 18:39:39 <minimal> and it seems that ntpsec also does NTS
2023-02-05 18:40:13 <Borley> I didn't know that NTS was used anywhere outside of ntpsec.
2023-02-05 18:40:31 <minimal> Borley: chrony has had it for a while
2023-02-05 18:42:06 <Saklad5> It's an IETF Proposed Standard, so it'd be weird if only one implementation existed
2023-02-05 18:42:11 <Saklad5> I'm using Chrony, myself
2023-02-05 18:42:59 <minimal> Borley: Chrony added it in 4.0, released Oct 2020
2023-02-05 18:44:23 <Saklad5> If you are running in a virtual environment, do make sure that the VM doesn't provide a reference clock already
2023-02-05 18:49:39 <Saklad5> In theory, it'd be nice to have a package that *only* uses reference clocks
2023-02-05 18:49:51 <Saklad5> Since that's presumably a common scenario
2023-02-05 18:50:13 <Saklad5> Seems silly to install Chrony just to watch a socket
2023-02-05 18:50:41 <minimal> Saklad5: you mean (sub)packages for various chrony configurations?
2023-02-05 18:50:50 <Saklad5> No
2023-02-05 18:51:04 <minimal> Saklad5: so what then?
2023-02-05 18:51:39 <Saklad5> Maybe getting that added into Busybox? I can't conceive of a scenario where using its insecure NTP client would be preferable
2023-02-05 18:52:33 <Saklad5> That's not the concern of Alpine Linux, at any rate
2023-02-05 18:54:33 <minimal> Saklad5: indeed, and how/which reference clock would you expect to see it use?
2023-02-05 18:56:10 <Saklad5> I suspect that's the problem: the number of options Chrony has implies a lot of fragmentation: https://chrony.tuxfamily.org/doc/4.3/chrony.conf.html#refclock
2023-02-05 18:58:05 <Saklad5> PHC is what KVM and such seem to use most, but that might be too niche to justify inclusion in Busybox.
2023-02-05 18:59:11 <mercenary> That is only a few options. Want to see fragmentation, look at the 127.x.y.z options in the reference ntpd
2023-02-05 18:59:18 <minimal> Saklad5: PHC can use /dev/ptp_kvm or /dev/ptp_hyperv, or /dev/ptp_vmware if they exist but that would either require manual config or else automatic-search-and-use code
2023-02-05 18:59:38 <Saklad5> Who said anything about not using manual config?
2023-02-05 18:59:46 <minimal> also PPS isn't (solely) useable as a reference clock
2023-02-05 18:59:58 <Saklad5> Yes
2023-02-05 19:00:15 <minimal> Saklad5: it wasn't clear what precisely you were suggesting
2023-02-05 19:00:53 <Saklad5> Doesn't the Busybox NTP client need to be told where to connect?
2023-02-05 19:02:29 <minimal> Saklad5: "it wasn't clear what precisely you were suggesting". when you mentioned getting reference clock support added to Busybox I guessed you meant for it to automagically detect a reference clock and use it
2023-02-05 19:02:50 <Saklad5> No, that'd be fundamentally impossible
2023-02-05 19:03:22 <minimal> wel lit could look for presence of /dev/ptp_* devices.but that wouldn't be foolproof
2023-02-05 19:03:30 <Saklad5> There could be conflicting ones
2023-02-05 19:03:55 <minimal> indeed, that why I mentioned it earlier
2023-02-05 19:04:43 <Saklad5> It just feels silly to install additional software to do something *less* complicated, that's all I'm saying
2023-02-05 19:05:33 <minimal> Saklad5: so you're going off to add PHC support to Busybox ntp now? ;-)
2023-02-05 19:06:01 <Saklad5> Might look into it eventually
2023-02-05 19:06:26 <minimal> once you get a Round Tuit? lol
2023-02-05 19:07:18 <minimal> https://www.etsy.com/uk/market/round_tuit
2023-02-05 19:07:24 <Saklad5> Alternatively, I might just propose dropping the NTP functionality on the grounds that NTS is preferable in pretty much every scenario
2023-02-05 19:08:29 <minimal> apart from RTC-less SBCs? <waves at RPI owners> :-)
2023-02-05 19:09:11 <Saklad5> I genuinely don't know what that means, and I'm curious
2023-02-05 19:09:54 <Saklad5> Single-board computers like Raspberry Pis without real-time clocks?
2023-02-05 19:10:57 <mercenary> Saklad5: NTP works fine inside local networks, no reason to drop it.
2023-02-05 19:11:39 <Saklad5> Putting aside the debate about trusting local networks, the question is whether using NTP is too niche to justify inclusion in Busybox
2023-02-05 19:12:43 <mercenary> I would certainly expect an NTP client in a basic OS yes.
2023-02-05 19:12:59 <minimal> Saklad5~: yes, not devices like RPIs do not have battery-backed RTCs and so at every power-on their system clock defaults to 1st Jan 1970
2023-02-05 19:13:06 <Saklad5> …Ouch
2023-02-05 19:13:13 <Saklad5> OK, so I can see how NTS would be a problem there
2023-02-05 19:13:32 <mercenary> (now if udhcpc would listen to the 'time servers' option and do something useful with it, that'd be ideal)
2023-02-05 19:13:40 <Saklad5> Though in a zero-trust threat model I don't think NTP is a *solution*, exactly
2023-02-05 19:14:27 <minimal> "whether using NTP is too niche to justify inclusion in Busybox"? NTP support is already in Busybox so no justification is required to justify its inclusion. Whether it should remain is another question...
2023-02-05 19:14:39 <Saklad5> That's the same question
2023-02-05 19:14:39 <mercenary> openbsd's implementation will use ntp, and https to set a +/- couple of seconds window where it trusts it
2023-02-05 19:15:10 <Saklad5> I find it somewhat irritating that some people seem to throw HTTPS at literally every problem
2023-02-05 19:15:11 <mercenary> That takes care of the RPi problem as well
2023-02-05 19:16:02 <Saklad5> NTP and DNS strike me as very similar protocols in numerous ways, and in both cases people try to contort HTTPS into providing security
2023-02-05 19:16:19 <minimal> mercenary: use of https doesn't necessarily work if the system clock defaults to 1970 and the TLS cert has a newer valid after date ;-)
2023-02-05 19:16:21 <mercenary> That was last year. This year, they throw udp:443/QUIC at everything
2023-02-05 19:16:51 <Saklad5> Rule of thumb: if you are proposing HTTPS as a security solution, you should be using TLS instead
2023-02-05 19:16:57 <Saklad5> Just TLS
2023-02-05 19:17:20 <mercenary> minimal: it actually works around that, by initially ignoring valid-before, then running again, or something along those lines
2023-02-05 19:17:31 <Saklad5> Yeah, but you could absolutely do that with NTS
2023-02-05 19:18:09 <Saklad5> https://www.ietf.org/rfc/rfc8915.html#name-initial-verification-of-ser
2023-02-05 19:19:36 <Saklad5> I do think there's some merit to throwing QUIC at every problem and seeing if it works. QUIC is incredibly capable
2023-02-05 19:21:16 <Saklad5> Virtually any UDP protocol can benefit from the anti-amplification mechanisms, virtually any TCP protocol can benefit from multiplexing without head-of-line blocking
2023-02-05 19:22:13 <Saklad5> It is deliberately designed to be generally useful. HTTPS, however, is very much not
2023-02-05 19:23:32 <mercenary> Quite so. From application point of view, QUIC is nice. From network point of view, it is a bit of a nightmare as one loses visibility into what type of traffic is hidden inside.
2023-02-05 19:23:50 <Saklad5> From user point of view, those are two benefits
2023-02-05 19:23:50 <minimal> mercenary: re udhcpc, can't you add "export ntpsrv" to udhcpc's default.script and then get it to configure ntp client appropriately?
2023-02-05 19:24:23 <Saklad5> And also that's true of anything with TLS, no? When ECH rolls it out it won't just be for QUIC
2023-02-05 19:24:31 <Saklad5> rolls out*
2023-02-05 19:24:49 <mercenary> minimal: probably. but that is manual config, and at the point where manual config is involved automation has taken over already.
2023-02-05 19:25:50 <mercenary> Saklad5: yup. which is why security people have this love/hate relationship with it, and have users install MITM certificates
2023-02-05 19:26:08 <minimal> mercenary: well its more that such a change would be a "improvement" (i.e. MR) to the Alpine busybox package (as it, not upstream Busybox) provides the default.script file
2023-02-05 19:26:27 <mercenary> But, for example, at network level I can give udp:123 lowest possible latency.
2023-02-05 19:27:10 <Saklad5> I don't believe in assigning any actual significance to port numbers anyway
2023-02-05 19:27:35 <Saklad5> Especially once SVCB records are finalized
2023-02-05 19:28:02 <mercenary> not for security purposes. But for optimization, surely
2023-02-05 19:28:44 <Saklad5> Maybe I'm naive, but I don't believe a convoluted port-based guessing game is worth the trouble
2023-02-05 19:29:19 <Saklad5> Besides, I believe in IPv6. Someday
2023-02-05 19:30:13 <mercenary> Hah. One day we will have IPv6 here. If I live that long.
2023-02-05 19:30:58 <Saklad5> For now, I make sure to at least bug people whenever I see an obstacle
2023-02-05 19:31:18 <Saklad5> And try never to present any myself
2023-02-05 19:31:57 <Saklad5> In terms of routing addresses, my priorities are IPv6, Tor single onion services, and IPv4, in that order.
2023-02-05 19:37:32 <minimal> Saklad5: perhaps the phc2sys program (from linuxptp) might be useful for you? It doesn't appear to be packaged for Alpine currently however
2023-02-05 19:41:55 <Saklad5> minimal: That's exactly the sort of thing I had in mind, thanks!
2023-02-05 19:44:44 <minimal> Saklad5: I might consider packaging it myself if you're not interested. I see RHEL docs referring to using it as a daemon
2023-02-05 19:45:07 <Saklad5> The question is whether it is more lightweight than Chrony. Seems likely
2023-02-05 19:47:48 <psykose> 
ACTION checks

2023-02-05 19:48:17 <minimal> Saklad5: I'd assume so - less functionality so should be more lightweight
2023-02-05 19:48:30 <Saklad5> The installation size seems slightly larger in other distros, somehow
2023-02-05 19:49:07 <Saklad5> Might be some form of package manager overhead, though
2023-02-05 19:53:57 <psykose> not the worst thing i've ever seen
2023-02-05 20:34:44 <Borley> Everything seems to install smoothly. lxqt doesn't want to start though
2023-02-05 20:34:55 <Borley> Something about qt platform plugin could not be initialized
2023-02-05 20:48:28 <psykose> which one
2023-02-05 20:48:47 <psykose> paste the log
2023-02-05 20:55:25 <minimal> Saklad5: the phc2sys binary is only 105K is size
2023-02-05 20:55:48 <psykose> what if you made it 40
2023-02-05 20:56:14 <Borley> I'm just kind of stumbling through things to get a feel. It looks like the other DEs have setup scripts
2023-02-05 20:56:47 <Borley> lxqt only recently came out of the edge repo so maybe that's why it doesn't have setup-desktop ?
2023-02-05 21:02:30 <psykose> nobody made one
2023-02-05 21:22:10 <Borley> Overall I like Alpine and will continue to watch its development
2023-02-06 00:33:06 <mendelmaleh> hi, anyone know how to make cryptsetup work? i set the cryptroot and cryptdm flags, and i added cryptsetup to the initramfs, but it doesn't seem to use it
2023-02-06 00:33:58 <minimal> mendelmaleh: is this for an existing installation?
2023-02-06 00:34:07 <mendelmaleh> i'm getting "mounting /dev/mapper/root on /sysroot failed: No such file or directory" from the initramfs, and it drops me into a shell
2023-02-06 00:34:33 <mendelmaleh> i can see there the crytpsetup binary, the crypt* flags, and nothing in /dev/mapper
2023-02-06 00:34:36 <mendelmaleh> minimal: no
2023-02-06 00:35:05 <minimal> mendelmaleh: so you are installing via setup-alpine?
2023-02-06 00:35:30 <mendelmaleh> i did the steps one at a time
2023-02-06 00:35:38 <minimal> what steps?
2023-02-06 00:35:46 <mendelmaleh> then setup-disk -m sys /mnt
2023-02-06 00:35:51 <mendelmaleh> minimal: setup-*
2023-02-06 00:36:12 <minimal> why? if you run setup-alpine you can select an encrypted install
2023-02-06 00:36:44 <mendelmaleh> triple boot
2023-02-06 00:36:54 <mendelmaleh> i'm trying to get alpine working on a mac lol
2023-02-06 00:37:24 <minimal> setup-disk -m sys /mnt doesn't specific encryption, that's a non-encrypted sys install
2023-02-06 00:38:20 <minimal> "-m cryptsys" would be the encrypted version of "-m sys"
2023-02-06 00:38:35 <minimal> setup-disk --help indicates this
2023-02-06 00:39:23 <minimal> what type of Mac? x86_64 or aarch64? with which bootloader?
2023-02-06 00:40:53 <mendelmaleh> minimal: for me it says -m  Use disk for MODE without asking, where MODE is either 'data' or 'sys'
2023-02-06 00:41:00 <mendelmaleh> i do see there is a -e flag now tho
2023-02-06 00:41:05 <minimal> setup-alpine/setup-disk is not really intended for multi-boot
2023-02-06 00:41:24 <mendelmaleh> i was mostly going off this https://wiki.alpinelinux.org/wiki/Full_disk_encryption_secure_boot#Installing_Alpine
2023-02-06 00:41:50 <mendelmaleh> it's x86_64 i assume, not sure what aarch64 is, it's not arm, it's 2019 intel
2023-02-06 00:42:03 <minimal> setup-alpine/setup-disk is not really intended for multi-boot, it assumes it is the only OS on the system
2023-02-06 00:42:04 <dwfreed> aarch64 is 64-bit arm
2023-02-06 00:42:05 <mendelmaleh> no bootloader, just efi entry
2023-02-06 00:42:16 <mendelmaleh> dwfreed: thanks
2023-02-06 00:42:35 <dwfreed> some distros call it that (and that's arm's preferred name for it), others call it arm64
2023-02-06 00:43:08 <dwfreed> arm64 looks kinda similar to amd64, so sometimes it's better to use aarch64
2023-02-06 00:43:54 <mendelmaleh> minimal: does it really matter if there are other os's? i don't really want the grub setup anyway
2023-02-06 00:44:04 <minimal> mendelmaleh: setup-disk for x86_64 UEFI will install and setup grub-efi which will add a EFI/Alpine/ directory to the ESP partition and (I believe) also add a EFI/boot/bootx64.efi (EFI Fallback file) which might overwrite any that other OSes have written (I assume MacOS setup up a fallback)
2023-02-06 00:44:24 <mendelmaleh> it just fails because i don't create /boot/efi
2023-02-06 00:44:24 <mendelmaleh> it just fails because i don't create /boot/efi
2023-02-06 00:44:53 <minimal> mendel,maleh@: setup-disk is only (currently) written to setup grub-efi for UEFI
2023-02-06 00:44:59 <mendelmaleh> yeah
2023-02-06 00:45:03 <mendelmaleh> so what should i do?
2023-02-06 00:45:45 <mendelmaleh> btw the dualbooting article points to setup-disk too
2023-02-06 00:45:47 <mendelmaleh> https://wiki.alpinelinux.org/wiki/Dualbooting#Install_Alpine
2023-02-06 00:46:08 <minimal> also setup-disk does not (currently) support/setup secure boot either
2023-02-06 00:46:23 <mendelmaleh> anyway it doesn't seem like the disk setup is the problem, the initramfs is not even mounting the disk yet
2023-02-06 00:47:05 <minimal> well how did you layout the partitions?
2023-02-06 00:49:22 <mendelmaleh> just a single root partition
2023-02-06 00:49:35 <minimal> plus a ESP partition?
2023-02-06 00:49:42 <mendelmaleh> that exists already
2023-02-06 00:50:04 <mendelmaleh> i set up root with cryptsetup, mounted it to /mnt, mounted boot to /mnt/boot
2023-02-06 00:50:04 <minimal> and that 2nd parition is LUKS? and you have then LVM on top of LUKs like that article?
2023-02-06 00:50:11 <mendelmaleh> no lvm
2023-02-06 00:50:43 <minimal> mounted the ESP partition to /mnt/boot ?
2023-02-06 00:51:05 <minimal> or to /mnt/boot/efi ?
2023-02-06 00:51:15 <mendelmaleh> the first
2023-02-06 00:51:43 <mendelmaleh> boot has the kernel and initramfs, and initramfs has cryptsetup, so that much worked
2023-02-06 00:52:24 <mendelmaleh> i didn't really follow that article
2023-02-06 00:53:14 <mendelmaleh> i looked at the initramfs docs and set cryptroot to the partition uuid, and cryptdm to root
2023-02-06 00:53:40 <mendelmaleh> i figured just setting them would be enough, but it doesn't seem like the initramfs does anything with them
2023-02-06 00:54:54 <minimal> and you rebuilt the initramfs after adding cryptsetup to the features list?
2023-02-06 00:55:01 <mendelmaleh> yes
2023-02-06 00:55:15 <mendelmaleh> and cryptsetup is available in the initramfs shell i got dropped in to
2023-02-06 00:55:46 <minimal> "cryptdm to root" - so you had the rootfs as /dev/mapper/root ?
2023-02-06 00:56:09 <mendelmaleh> wdym by rootfs
2023-02-06 00:56:20 <minimal> the root filesystem
2023-02-06 00:56:23 <mendelmaleh> efibootmgr -c -d /dev/nvme0n1 -p 1 -L "Alpine Linux" -l '\vmlinuz-lts' -u "initrd=\\initramfs-lts cryptroot=$(cat p3) cryptdm=root root=/dev/mapper/root roo
2023-02-06 00:56:24 <mendelmaleh> tfstype=ext4 modules=sd-mod,usb-storage,ext4 quiet"
2023-02-06 00:56:29 <mendelmaleh> this are the args i set
2023-02-06 00:57:54 <minimal> what is in the file "p3" ?
2023-02-06 00:58:10 <mendelmaleh> the uuid of the partition
2023-02-06 00:58:39 <mendelmaleh> oh do i need to add UUID=
2023-02-06 00:58:42 <mendelmaleh> like the kernel?
2023-02-06 00:58:45 <minimal> just the uuid or UUID=<uuid> ?
2023-02-06 00:58:54 <minimal> yes you need to prefix it with UUID=
2023-02-06 00:58:54 <mendelmaleh> just uuid
2023-02-06 00:58:56 <mendelmaleh> f
2023-02-06 00:59:33 <minimal> as it could be a UUID or a PARTUUId (or a LABEL also I think)
2023-02-06 00:59:54 <mendelmaleh> makes sense, i'll try now
2023-02-06 01:00:14 <mendelmaleh> shouldn't it give an error about not being able to find the device or something
2023-02-06 01:00:22 <mendelmaleh> it seems like it just skips ahead to mount root=
2023-02-06 01:01:24 <minimal> the mounting is handled by nlplug-findfs inside the initramfs
2023-02-06 01:07:09 <minimal> mendelmaleh: whatever you specify as "cryptroot=" is passed by the initramfs' init straight to nlplug-findfs
2023-02-06 01:09:52 <mendelmaleh> minimal: i added UUID=, still the same error
2023-02-06 01:10:05 <mendelmaleh> not sure what nlplug is 
2023-02-06 01:10:13 <mendelmaleh> efibootmgr -c -d /dev/nvme0n1 -L "Alpine Linux" -l '\vmlinuz-lts' -u 'initrd=\\initramfs-lts cryptroot=UUID=$(blkid /dev/nvme0n1p3 | cut -d'"' -f2) cryptdm=root root=/dev/mapper/root rootfstype=ext4 modules=sd-mod,usb-storage,ext4 quiet"
2023-02-06 01:10:23 <mendelmaleh> this is what i used
2023-02-06 01:10:53 <mendelmaleh> oops wait
2023-02-06 01:10:57 <minimal> nlplug-findfs is a binary in the initramfs that the initramfs' init used to mount the root filesystem (whether or not it is encrypted)
2023-02-06 01:15:12 <mendelmaleh> minimal: it worked!
2023-02-06 01:15:43 <mendelmaleh> now it gets to loading hardware drivers and it gets stuck there
2023-02-06 01:16:13 <mendelmaleh> i wonder what's the difference between the live usb and the install 
2023-02-06 01:16:38 <mendelmaleh> is there somewhere i can see the live usb kernel params?
2023-02-06 01:17:49 <minimal> it used the same kernel package, linux-lts
2023-02-06 01:18:45 <minimal> perhaps you have missing firmware? you said it is an old Mac - what type of wifi hardware does it have?
2023-02-06 01:19:08 <mendelmaleh> no clue
2023-02-06 01:19:16 <minimal> though missing wifi firmware normally woul dmean it wouldn't work, rather than fail to boot
2023-02-06 01:19:35 <mendelmaleh> but i remember setup-interfaces being able to show some wifi networks
2023-02-06 01:19:57 <minimal> we've had a few people in here recently with old Mac that we don't have Wifi firmware for (due ot license issues AFAIK)
2023-02-06 01:20:32 <mendelmaleh> btw how hard is it to make a custom live usb?
2023-02-06 01:20:48 <minimal> try removing "quiet" for the efibootmgr provided cmdline options? I might give some more info during boot
2023-02-06 01:21:14 <mendelmaleh> hm ok
2023-02-06 01:30:25 <mendelmaleh> minimal: didn't make any difference, it seems like it's only for the initramfs (that became a lot noisier, but no errors)
2023-02-06 01:48:07 <minimal> mendelmaleh: I'd expect without "quiet" you'd see a little more during boot (though maybe that only applies when syslinux/grub is used)
2023-02-06 01:49:46 <minimal> do you have "kms" in your /etc/mkinitfs/mkinitfs.conf features list?
2023-02-06 01:51:36 <mendelmaleh> i didn't add anything besides cryptsetup, so i guess not
2023-02-06 01:51:57 <mendelmaleh> later i'll mount it manually and look
2023-02-06 01:57:09 <mendelmaleh> minimal: i came back and it unstuck itself, amazing!
2023-02-06 01:57:52 <minimal> hmm..........try "dmesg" and try to figure out what happened around the delay?
2023-02-06 01:58:38 <minimal> the numbers of on the left are offsets in seconds from the kernel initialising - so you'll see some kernel modules loading - look for a jump in time
2023-02-06 02:00:16 <mendelmaleh> yeah, it seems to me that applesmc is what kept it hanging until it gave up 2.5 min later
2023-02-06 02:00:34 <minimal> was there an error?
2023-02-06 02:00:47 <mendelmaleh> probe of applesmc.768 failed with error -5
2023-02-06 02:00:48 <mendelmaleh> driver init failed (ret=-5)!
2023-02-06 02:00:55 <mendelmaleh> these two, twice
2023-02-06 02:03:16 <mendelmaleh> i'm gonna upgrade everything to edge for the lols
2023-02-06 02:03:38 <mendelmaleh> i don't see any dedicated apple packages, maybe it's somewhere there already
2023-02-06 02:03:41 <minimal> seem to be various linux users seeing that error (and fan-related issues)
2023-02-06 02:04:07 <mendelmaleh> Upgrading linux-firmware-qcom (20221214-r1 -> 20230117-r0)
2023-02-06 02:04:24 <mendelmaleh> and kernel was at 5.15
2023-02-06 02:04:32 <minimal> qualcomm? I'd doubt that would be related
2023-02-06 02:04:43 <mendelmaleh> hm i meant to paste -other
2023-02-06 02:04:54 <mendelmaleh> it has the same dates tho
2023-02-06 06:42:53 <novorol[m]> I'm trying to set up firewalld with ansible, but my playbook fails with... (full message at <https://matrix.org/_matrix/media/v3/download/matrix.org/HiibkRlwRdjraxAPmBkFBkNX>)
2023-02-06 06:45:10 <novorol[m]> nvm the problem is with firewalld apparently
2023-02-06 09:23:51 <JohannesJ[m]> heh
2023-02-06 11:47:18 <Leolo_3> so the standard ISO has 480mb of firmware bundles on it.  I'm sure I don't need 90% of that.  
2023-02-06 11:50:44 <mercenary> The problem is 'which 90%'.
2023-02-06 13:16:21 <bovox> hi
2023-02-06 13:54:46 <raspbeguy> hello, I was wondering if there were already some attemps to bring openstack components into Alpine repos. Currently I found nothing.
2023-02-06 13:56:18 <minimal> raspbeguy: you mean to run Openstack on Alpine? wouldn't that sort of go against Alpine's "lightweight" stance?
2023-02-06 13:57:04 <raspbeguy> well, lightweight doesn't mean stopping you installing anything you want
2023-02-06 13:58:21 <minimal> right, but Openstack is pretty large
2023-02-06 13:58:34 <raspbeguy> doesn't matter to me
2023-02-06 13:59:18 <raspbeguy> GNOME is in the repos, and you can't call that "lightweight"
2023-02-06 14:01:22 <raspbeguy> anyway, I was asking before trying to package it myself, in case there is a red flag somewhere
2023-02-06 14:02:09 <minimal> I haven't seen it mentioned before, I guess no-one until now has expressed an interest in doing it
2023-02-06 14:03:09 <raspbeguy> great, that'll be a hopefully enriching trip for me
2023-02-06 14:03:11 <minimal> also as openstack is "technology neutral" then most deployments are more or less unique
2023-02-06 14:03:58 <raspbeguy> I know. But also openstack is pretty well split into several components so I think this is a good candidate
2023-02-06 14:05:46 <minimal> so would you package up each layer so it works with only the single technology you chose for that layer, or for all the alternative technologies for the layer?
2023-02-06 14:09:31 <raspbeguy> I don't understand the question. the different "technologies" uses are defined by the configuration, so I don't see the point of several packages
2023-02-06 14:10:20 <raspbeguy> I was simply thinking about one package for nova, one for keystone, etc...
2023-02-06 14:11:27 <minimal> as an example for the Openstack Designate component (for DNS) - it supports at least Bind and PowerDNS. So were you to package it for Alpine would you only test/set it up for BIND (if you chose that) or would you test/set it up for all DNS servers supported by Designate?
2023-02-06 14:11:45 <minimal> the same applies for every other Openstack component
2023-02-06 14:12:40 <raspbeguy> I would let the set up to the user...
2023-02-06 14:13:28 <linuxuser> Hello, I'd like to change the keyboard layout on gnome gdm (the default is English-US), but I haven't found a way to do it yet. Could you help me please?
2023-02-06 14:14:14 <raspbeguy> linuxuser, well I hit that problem on another systemd-free distro, and it is not possible
2023-02-06 14:14:52 <raspbeguy> this is the cost of adapting gnome/gdm for systemd-free environment
2023-02-06 14:37:55 <linuxuser> raspbeguy thank you for responding, hopefully we will see this problem solved soon
2023-02-06 14:40:06 <raspbeguy> I honestly don't think so :/ GNOME is very deeply linked to systemd, patches to set it free from it are quite heavy so I think that it will be eventually too hard to maintain
2023-02-06 14:40:21 <raspbeguy> in the meantime you can une lightdm for instance
2023-02-06 14:42:59 <raspbeguy> Well, shiver me timbers, patches don't seem so heavy actually
2023-02-06 14:44:07 <raspbeguy> I might have spoken too fast.
2023-02-06 15:47:52 <ikke> raspbeguy: FYI, linuxuser left
2023-02-06 15:55:12 <raspbeguy> oh no, they will only remember my unfortunate lies
2023-02-06 15:59:24 <ikke> They left before you talked about patches 
2023-02-06 17:52:31 <Leolo_3> mercenary - lsmod + a simple matter of programming and I can find out
2023-02-06 17:59:41 <Leolo_3> wait. these are kernel modules :/
2023-02-06 18:07:42 <Leolo_3> I can't find what loads the firmware
2023-02-06 18:10:51 <minimal> Leolo_3: what problem exactly are you trying to resolve?
2023-02-06 18:15:21 <Leolo_3> I want to boot my router from a small CD-R.  Most of the CD-R is filled with firmwares.  95% of those firmwares aren't being used.  I would like to trim the list down to only necessary firmware
2023-02-06 18:16:25 <wart_> not sure if this is alpine or pmos, but i noticed that fc-match doesn't list my font in xorg/dwm but it does in xwayland/sway, so i wonder if we need to mod xinitrc or something?
2023-02-06 18:16:54 <minimal> Leolo_3: ok, but the ISO is generally only used to boot a machine before then running setup-alpine to install onto another media
2023-02-06 18:18:04 <minimal> Leolo_3: if you boot the ISO on your router it won't actually be a router until you set things up (and how will you keep the settings with a CD-R?)
2023-02-06 18:18:12 <Leolo_3> My router currently boots 2.4.5 from CD-R, which then loads config from USB
2023-02-06 18:18:32 <Leolo_3> is they functionality removed in 3.17?
2023-02-06 18:18:37 <minimal> Alpine 2.4.5?
2023-02-06 18:18:52 <Leolo_3> yes
2023-02-06 18:19:10 <Leolo_3> I save the config with lbu
2023-02-06 18:19:17 <ikke> That's still there
2023-02-06 18:19:23 <minimal> as you're using a ISO then why do you want to trim it down? it is RO
2023-02-06 18:20:11 <Leolo_3> because a CD-R has a max of 700mb, 480mb of which is firmware.  and I can't fit all the tools I want on it
2023-02-06 18:20:24 <Leolo_3> a secondary goal is to discover all I can about Alpine
2023-02-06 18:20:47 <ikke> Leolo_3: you'd have to create a custom iso then
2023-02-06 18:21:27 <minimal> Leolo_3: I assume the 3.17 Standard ISO fits on a CD, are you using the Extended ISO?
2023-02-06 18:22:07 <Leolo_3> yes, I have created a custom ISO
2023-02-06 18:22:18 <Leolo_3> it is based on the standard ISO
2023-02-06 18:22:40 <minimal> so what's the problem then? you decide which firmware to have on your custom ISO
2023-02-06 18:23:13 <ikke> by default linux-lts pulls in all firmware
2023-02-06 18:23:21 <ikke> (by means of linux-firmware-any)
2023-02-06 18:23:35 <Apachez> next level is to do a iso but write that to a usb flash drive for example samsung pro or similar
2023-02-06 18:23:40 <ikke> If you want none of the firmware packages, make sure it pulls in linux-firmware-none
2023-02-06 18:23:41 <Apachez> will boot much faster than cd :)
2023-02-06 18:23:51 <minimal> so then he needs to add linux-firmware-none ?
2023-02-06 18:23:56 <ikke> yes
2023-02-06 18:25:39 <Leolo_3> ikke - I want "some" of the firmware.  I don't know which ones.  I'd like to find out what is actually loading the firmwares
2023-02-06 18:26:14 <ikke> The kernel modules
2023-02-06 18:27:56 <minimal> Leolo_3: for any given module do "modinfo <module name> | grep -i firmware"
2023-02-06 18:28:13 <minimal> that will show you the firmware files (if any) that module needs
2023-02-06 18:29:26 <minimal> so check the modules that your "router" needs for its hardware
2023-02-06 18:36:34 <Leolo_3> for n in `lsmod | cut -d\  -f 1 ` ; do modinfo $n | grep -i firmware ; done # lists nothing :D
2023-02-06 18:38:28 <ikke> modinfo -F firmware $(awk '{print $1}' /proc/modules
2023-02-06 18:40:43 <Leolo_3> for some reason I never reach for awk
2023-02-06 19:35:05 <donoban> I noticed that my user is not on video group if I use 'docker exec -it .... ash', but it appears on /etc/groups
2023-02-06 19:35:16 <donoban> is this an expected behaviour? 
2023-02-06 19:36:22 <donoban> if I use -u root, and then 'su user' the groups look fine
2023-02-06 19:36:56 <ikke> I don't think /etc/groups is consulted
2023-02-06 19:37:00 <donoban> hmm, if I use '-u user' it works too, user is the default group
2023-02-06 19:37:12 <donoban> s/group/user/
2023-02-06 19:37:12 <alpine-meetbot> donoban meant to say: hmm, if I use '-u user' it works too, user is the default user
2023-02-06 19:39:39 <donoban> looks very similar to https://github.com/moby/moby/issues/3012
2023-02-06 19:40:12 <donoban> wow this very old
2023-02-06 19:42:44 <donoban> "Note that when specifying a group for the user, the user will have only the specified group membership. Any other configured group memberships will be ignored."
2023-02-06 19:43:11 <donoban> I'm not specifying any group but looks the same behaviour
2023-02-06 19:57:21 <donoban> ahh, I found the problem
2023-02-06 19:57:27 <ikke> do tell
2023-02-06 19:57:31 <donoban> is a custom script for launch it that does
2023-02-06 19:57:34 <donoban> --user=$(id -u):$(id -g) $@
2023-02-06 19:57:56 <donoban> i guess that if I remove the :$(id -g) part
2023-02-06 19:57:57 <ikke> as argument to docker run?
2023-02-06 19:58:01 <donoban> yes
2023-02-06 19:58:02 <ikke> aha
2023-02-06 19:59:52 <donoban> / $ groups
2023-02-06 19:59:55 <donoban> user video
2023-02-06 19:59:57 <donoban> :D
2023-02-06 20:00:35 <donoban> interesting behaviour
2023-02-06 20:05:14 <Leolo_3> so the answer is : i915 :D
2023-02-06 20:05:37 <Leolo_3> also : thank you ikke for your pointers
2023-02-06 20:05:45 <ikke> Leolo_3: you're welcome
2023-02-06 22:11:28 <LAC-Tech> I'm trying to update my long neglected alpine system. can't escape UNTRUSTED signature
2023-02-06 22:11:30 <LAC-Tech> https://pastebin.com/Q4v3UaN6
2023-02-06 22:11:56 <ikke> LAC-Tech: add alpine-keys from 3.14 first
2023-02-06 22:12:08 <ikke> that contains new keys
2023-02-06 22:12:15 <ikke> but still trusted by older installs
2023-02-06 22:12:46 <LAC-Tech> thanks! that's working. yeah haven't updated this since like 2020
2023-02-06 22:13:01 <ikke> https://wiki.alpinelinux.org/wiki/Release_Notes_for_Alpine_3.15.0#New_package_signing_keys
2023-02-06 22:13:07 <LAC-Tech> it's my little WSL island of sanity on windows
2023-02-06 22:14:36 <ikke> I use that too, but I do keep it up-to-date :P
2023-02-06 22:17:10 <LAC-Tech> yeah I love alpine, it's on my server too. but I love rolling release distros for ease of update
2023-02-06 22:17:29 <ikke> I use edge for that :)
2023-02-06 22:18:34 <LAC-Tech> hmm. how living on the edge is using edge? my desktop has void rolling release, not many problems, but that's also voids primary mode
2023-02-06 22:19:49 <elly> edge on laptop
2023-02-06 22:19:50 <elly> it's the way
2023-02-06 22:25:13 <LAC-Tech> https://pastebin.com/4ZFiMPcj how do I get rid of llvm-libs-14? I assume that's what the upgrade message is suggesting I might do
2023-02-06 22:25:31 <latex_> Hey, what do you guys do when a service on your Alpine server crashes? Manually restart it? I've read OpenRC cannot automatically restart services. I also noticed OpenRC falsely reports my service has crashed while it's still running :/
2023-02-06 22:26:28 <latex_> So when my service crashed for real, I was troubleshooting for 3 hours because I thought the process crashed when I started it, while it was actually OpenRC misreporting the status
2023-02-06 22:29:08 <LAC-Tech> (I just deleted both python packages. that seemed to work. think I just had python2 because node needs (needed?) python2 to build native dependencies. speaking of which, need to see if deno got round to a musl version yet)
2023-02-06 22:29:31 <pj> nope
2023-02-06 22:30:12 <pj> it is stuck on the same problem as years ago which is "we don't know how to patch chromium so it works on musl"
2023-02-06 22:31:48 <ikke> Node didn't need python2 for quite some time now
2023-02-06 22:32:28 <LAC-Tech> what a few months? :p
2023-02-06 22:32:42 <ikke> Years
2023-02-06 22:32:53 <LAC-Tech> pj: yeah, no progress
2023-02-06 22:33:11 <LAC-Tech> ikke: for building native deps? not building node
2023-02-06 22:33:18 <LAC-Tech> in 2020 it definitely needed python2
2023-02-06 22:33:58 <ikke> Native deps I don't know
2023-02-06 22:34:46 <LAC-Tech> oh wait 2020 was like 3 years ago. my god..
2023-02-06 22:42:39 <psykose> gyp runs with python3 for quite a while
2023-02-06 22:43:16 <psykose> if a project has devDependencies : gyp: some-version-from-2020 it probably won't work, but you can freely just raise it by hand
2023-02-06 22:43:49 <psykose> aside from that it invoked `python` verbatim so you need a symlink to python3 or some configuration file i forgot about, but since 3.17 on alpine `python` points to python3 anywya
2023-02-07 03:15:03 <upstart> Hello
2023-02-07 03:15:43 <upstart> I am trying to figure out what I need to adjust in the installation ISO to successfully boot it
2023-02-07 03:16:40 <upstart> i suspect it has to do with X server and nouveau, though I am not sure
2023-02-07 03:16:55 <upstart> I can't get far enough into the process to find out
2023-02-07 03:23:16 <psykose> the iso doesn't start an x server
2023-02-07 03:25:26 <upstart> How can I check to see what is holding up the boot process?
2023-02-07 10:44:15 <trees> Hi! can someone help me? I'm trying to run steam and this error shows up> doas flatpak run com.valvesoftware.St
2023-02-07 10:44:57 <trees> bwrap: Unable to open lock file /run/flatpak/per-app-dirs-ref: Permission denied
2023-02-07 10:56:27 <psykose> why are you running it as root
2023-02-07 11:00:41 <trees> Oh i thought ur suppose to
2023-02-07 11:01:01 <trees> > flatpak run com.valvesoftware.Steam
2023-02-07 11:01:04 <trees> (flatpak run:16596): dconf-CRITICAL **: 19:00:19.621: unable to create file '/tmp/-runtime-dir/dconf/user': Permission denied.  dconf will not work properly.
2023-02-07 11:01:11 <trees> (flatpak run:16596): dconf-CRITICAL **: 19:00:19.621: unable to create file '/tmp/-runtime-dir/dconf/user': Permission denied.  dconf will not work properly.
2023-02-07 11:01:46 <trees> (flatpak run:16596): dconf-CRITICAL **: 19:00:19.621: unable to create file '/tmp/-runtime-dir/dconf/user': Permission denied.  dconf will not work properly.
2023-02-07 11:02:06 <trees> Now how do I solve this?
2023-02-07 11:16:38 <donoban> something looks wrong with  '/tmp/-runtime-dir/dconf/user', i guess it should be  '/tmp/1000-runtime-dir/dconf/user' or something similar, maybe there is something wrong with ENV or it fails getting the user id
2023-02-07 11:19:49 <donoban> could you try 'echo $XDG_RUNTIME_DIR'
2023-02-07 11:20:47 <trees> > /tmp/-runtime-dir
2023-02-07 11:22:04 <donoban> and does it exist in your "host"?
2023-02-07 11:22:36 <donoban> take a look on it and check permissions
2023-02-07 11:22:51 <donoban> if there is no uid maybe it was mixed between root and your user
2023-02-07 11:23:28 <donoban> what desktop and session manager do you use?
2023-02-07 11:23:39 <trees> I don't know how to answer that but almost of the applications like wayland, sway, imv and pcmanfm has their cache thing there
2023-02-07 11:24:07 <donoban> check 'ls -lh /tmp/-runtime-dir'
2023-02-07 11:24:24 <donoban> is your user owner of everyhing?
2023-02-07 11:25:16 <trees> yes, cactus is the user are it says all of them belongs to cactus. Only one is from root which is dconf
2023-02-07 11:25:25 <trees> *and it says
2023-02-07 11:25:35 <donoban> probably you created it when tried to run flatpak as root
2023-02-07 11:25:50 <donoban> try removing it with sudo, and run flatpak again as user
2023-02-07 11:27:43 <trees> I did rm -rf /tmp/-runtime-dir/dconf
2023-02-07 11:28:25 <trees> used the command 'ls -lh /tmp/-runtime-dir' and it still says that root is the owner of dconf
2023-02-07 11:28:56 <donoban> meh, try chown
2023-02-07 11:29:15 <donoban> do you have some process running as root that could create it?
2023-02-07 11:32:43 <trees> doas chown cactus:cactus /tmp/-runtime-dir/dconf
2023-02-07 11:32:46 <trees> drwx------    2 cactus   cactus      4.0K Feb  7 16:42 dconf
2023-02-07 11:32:57 <donoban> uh
2023-02-07 11:32:59 <donoban> run again
2023-02-07 11:33:02 <donoban> with chown -R
2023-02-07 11:33:03 <trees> (flatpak run:28709): dconf-CRITICAL **: 19:31:57.827: unable to create file '/tmp/-runtime-dir/dconf/user': Permission denied.  dconf will not work properly.
2023-02-07 11:34:23 <trees> okay even with chown -R it stil didn't work
2023-02-07 11:35:19 <trees> wait the dconf error disappeared
2023-02-07 11:35:31 <donoban> great
2023-02-07 11:35:38 <trees> now this other error shows up bwrap: Can't mkdir parents for /tmp/-runtime-dir/app/com.discordapp.Discord: Permission denied
2023-02-07 11:35:54 <donoban> try
2023-02-07 11:36:04 <donoban> chown xx:xxx -R /tmp/-runtime-dir
2023-02-07 11:36:14 <donoban> chown -R xx:xx /tmp/-runtime-dir
2023-02-07 11:36:39 <trees> doas chown cactus:cactus -R /tmp/-runtime-dir
2023-02-07 11:37:09 <donoban> I wonder how are you loggin & starting desktop, it looks that you run something as root in the way
2023-02-07 11:37:11 <trees> flatpak run com.valvesoftware.Steam
2023-02-07 11:37:14 <trees> INFO:root:https://github.com/flathub/com.valvesoftware.Steam/wiki
2023-02-07 11:37:18 <trees> oops
2023-02-07 11:37:27 <trees> the steam gui SHOWED UP
2023-02-07 11:37:29 <trees> OMG
2023-02-07 11:37:40 <donoban> nice
2023-02-07 11:37:42 <trees> its currently downloading update, you're my hero!
2023-02-07 11:37:54 <donoban> hehe
2023-02-07 11:38:04 <trees> I've been having this problem for a year now since I switched to alpine, I thought i was going to finally quit alpine today
2023-02-07 11:38:15 <donoban> hope that it stills working after you reboot
2023-02-07 11:38:25 <trees> yeah lol
2023-02-07 11:38:29 <trees> thank you so much :))
2023-02-07 11:39:06 <donoban> np
2023-02-07 11:54:54 <psykose> a year huh
2023-02-07 12:06:01 <Fxzxmic> I think he should solve this problem: /tmp/-runtime-dir, otherwise he may still encounter other problems.
2023-02-07 14:11:55 <em> Hey, I dont have /home/*/.Xresources  ...any ideas why? I want to edit font settings
2023-02-07 14:12:22 <grawity> *very few* such files by exist by default
2023-02-07 14:12:37 <grawity> ~/.Xresources doesn't exist by default; you create it when you have something to put inside
2023-02-07 14:16:41 <psykose> it's your lucky day to birth a new file into existence
2023-02-07 14:21:07 <em> wow
2023-02-07 14:21:13 <em> haha
2023-02-07 14:24:46 <em> psykose: what does EOF mean?
2023-02-07 14:24:56 <em> end of file?
2023-02-07 14:25:03 <elly> yes
2023-02-07 14:25:16 <em> and should be exluded from example?
2023-02-07 14:26:17 <em> also, any recommandations of easy reaadable fonts without anti-alising? 
2023-02-07 14:26:35 <em> for peaople who cant spell apparently...
2023-02-07 14:27:04 <elly> fixed-width or proportional?
2023-02-07 14:27:08 <grawity> if an example has "EOF" at the end, I assume it also has "cat > something <<EOF" at the beginning
2023-02-07 14:27:23 <em> grawity: that is correct
2023-02-07 14:27:35 <elly> ah! in that case that's the end of a shell "heredoc"
2023-02-07 14:27:39 <elly> it doesn't go in the resulting file
2023-02-07 14:27:39 <em> fixed-width sounds nice
2023-02-07 14:27:45 <grawity> in which case "EOF" is some arbitrary label that just has to match on both ends
2023-02-07 14:27:46 <elly> em: I like Terminus, then
2023-02-07 14:27:50 <em> elly:got it
2023-02-07 14:27:56 <grawity> <<Foo ... Foo, or <<! ... !
2023-02-07 14:28:20 <em> so if I only install terminus, then this will automatically be used as default everywhere?
2023-02-07 14:29:39 <elly> no, you have to configure stuff to use it
2023-02-07 14:29:51 <em> ok
2023-02-07 14:29:53 <elly> in my .Xresources I have:
2023-02-07 14:29:54 <elly> UXTerm*faceName: terminus
2023-02-07 14:30:15 <em> what does it define?
2023-02-07 14:30:30 <elly> it tells XTerm to use terminus as its font
2023-02-07 14:30:43 <elly> you probably need something similar for your terminal emulator
2023-02-07 14:30:49 <em> ok
2023-02-07 14:30:53 <em> I will use st
2023-02-07 14:31:14 <em> st*faceName: terminus
2023-02-07 14:31:19 <em> like that you think?
2023-02-07 14:32:05 <elly> if I told you I would be straight up guessing
2023-02-07 14:32:11 <elly> you're better off reading the st documentation
2023-02-07 14:32:17 <elly> the whole 'faceName' thing is an Xterm thing
2023-02-07 14:32:29 <em> ait
2023-02-07 14:33:05 <psykose> please assume the mind reading position
2023-02-07 14:33:10 <psykose> what did i have for breakfast today?/
2023-02-07 14:33:34 <em> just want to share something crazy I found today: different power supplies makes hardware perform differently ....crazy
2023-02-07 14:34:56 <uniqNuuid> what hardware
2023-02-07 14:35:19 <em> x86 mashine but most difference on router
2023-02-07 14:54:45 <em> why cant I ssh into root?
2023-02-07 14:55:16 <elly> probably because the default value of PermitRootLogin in /etc/ssh/sshd_config is "prohibit-password" and you are using password auth
2023-02-07 14:55:29 <elly> if that is correct, you should cease and repent, and use public-key auth
2023-02-07 14:55:39 <em> would it be any disadvantage installing the system from user account?
2023-02-07 14:55:54 <elly> what do you mean by "installing the system"?
2023-02-07 14:55:58 <qaqland[m]> It is almost default setting now
2023-02-07 14:56:11 <em> sorry I mean X etc
2023-02-07 14:56:24 <elly> usually, what you do is log in as your user, then use su / sudo / doas to become root
2023-02-07 14:56:30 <elly> and install packages or whatever you need to do from there
2023-02-07 14:56:47 <em> and its exactly the same as beeing loged in as root?
2023-02-07 14:57:00 <elly> basically?
2023-02-07 14:57:59 <em> coming from windows have some messy permission ghosts stuck in me...
2023-02-07 14:58:19 <pj> you are effectively becoming root user for the time of execution of the command
2023-02-07 14:58:43 <elly> ya
2023-02-07 14:59:09 <qaqland[m]> You can turn off the password of user.
2023-02-07 15:00:07 <pj> also windows perm system is not messy, it's just ACL in comparison to default u:g:w on Linux
2023-02-07 15:00:10 <em> now I can past my long list of packages from ssh, very nice
2023-02-07 15:01:48 <em> pj: I dont know enough, but my experience is that linux is more straight on and clear...dont know why
2023-02-07 15:02:29 <pj> it's because on linux it's a spectrum of easy and broken to complex and still broken
2023-02-07 15:03:18 <em> is this a joke that I dont understand?
2023-02-07 15:03:22 <pj> on windows you have just one choice which is complex and broken, no matter if you want a very simple small server or full blown GUI desktop
2023-02-07 15:03:34 <pj> em: no?
2023-02-07 15:03:43 <elly> there's no joke really, but complaining about Linux is a major pastime of Linux users
2023-02-07 15:04:20 <elly> I wouldn't let it bother you too much
2023-02-07 15:04:31 <pj> as a sysadmin of both Windows and Linux, I'm just stating how it is in both worlds
2023-02-07 15:05:32 <pj> tl;dr: on linux you have a choice of the software blocks you use so you can build minimalistic system that's "simple", e.g. Alpine
2023-02-07 15:06:59 <em> my jurney: Ubuntu-Mint-Manjato-Arch-Void-Alpine......I would say that the "hard for me to understand" errors disapeard when I got to Void...
2023-02-07 15:07:41 <em> very happy to get behind the curtain of the wizard of oz
2023-02-07 15:14:13 <em> elly: what is hint in the .xresouces file?
2023-02-07 15:16:31 <em> and how do I set the settings globally?
2023-02-07 15:41:15 <em> I installed terminus font and it became default, then I installed noto but want to go back to terminus, how do I do that?
2023-02-07 15:42:44 <elly> no idea, sorry
2023-02-07 15:50:00 <latex_> em: insert this into ~/.config/fontconfig/fonts.conf : https://ttm.sh/S5Z.conf
2023-02-07 15:50:27 <latex_> Then, open your favorite qt5ct program, and set the font to Terminus
2023-02-07 15:52:58 <em> latex: this file sdoes not exist, shall I create it?
2023-02-07 15:55:42 <em> latex: It worked, great thank!! but whould this be ok as well? : ~/.fonts.conf
2023-02-07 15:57:28 <latex> idk. Try it and if it works I guess it does
2023-02-07 15:57:45 <latex> Personally, I use Terminus for everything. I force every program to use Terminus
2023-02-07 15:58:13 <em> will your syntax do that do that for me?
2023-02-07 15:58:56 <em> ~/.fonts.conf worked fine
2023-02-07 16:01:53 <em> latex: ?
2023-02-07 16:02:27 <latex> It sets Terminus as default font for every style
2023-02-07 16:02:37 <em> latex: cool
2023-02-07 16:02:40 <latex> e.g. sans-serif, seris, blah, it will all be Terminus
2023-02-07 16:03:05 <latex> However there are programs that don't listen to that. To force them to use Terminus, you can use qt5ct to force Qt to use Terminus, and there's a GTK thing as well
2023-02-07 16:03:17 <latex> and even Xft things can be forced to use Terminus
2023-02-07 16:03:33 <em> latex: to save time and money, I have aold 23" 1680x1050 screen and now the pixals are a bit big. what pixel dencity do you like?
2023-02-07 16:03:51 <latex> Xft.dpi: 96
2023-02-07 16:03:53 <latex> *.font:Terminus:pixelsize=14:antialias=true:autohint=true;
2023-02-07 16:04:03 <latex> you can put these two lines in ~/.Xresources
2023-02-07 16:05:18 <latex> Also, my DPI is around 90 because I use the standard 1920x1080 24 inch monitor
2023-02-07 16:05:39 <em> do you like it?
2023-02-07 16:06:08 <latex> I like it. I don't feel I need anything more than 1920x1080
2023-02-07 16:06:22 <latex> 4K is a meme
2023-02-07 16:06:33 <em> this is the line:
2023-02-07 16:06:34 <em> Terminus:pixelsize=14:antialias=true:autohint=true;
2023-02-07 16:06:35 <latex> I don't understand why people want to make pixels invisible
2023-02-07 16:06:59 <latex> Wait why is antialias true. Huh. Maybe disable that
2023-02-07 16:07:13 <em> I was just wondering why you had that
2023-02-07 16:07:27 <latex> I don't remember but it was like that
2023-02-07 16:07:30 <latex> I just disabled it
2023-02-07 16:08:01 <em> maybe you like it??
2023-02-07 16:08:33 <em> but wit, I might have conficting edits
2023-02-07 17:36:10 <Oyo> im new in Linux, and cant understand why Alpine Linux is using OpenRC with Busybox course runit utilities are already in Busybox due to https://bkhome.org/news/201802/busybox-1251-runit-applets.html. Why Alpine uses OpenRC?
2023-02-07 17:43:31 <elly> because alpine long predates runit being in busybox and migrating init systems would be a huge compatibility break for dubious gain
2023-02-07 17:43:42 <elly> (and because, if there is such a migration, it might not be to runit :P)
2023-02-07 17:46:32 <Oyo> extra entity is the best wayout?
2023-02-07 17:47:00 <elly> maybe?
2023-02-07 17:47:14 <elly> if you would like to argue that alpine should migrate to runit, there's probably a mailing list for that
2023-02-07 17:47:33 <ikke> runit is apparently no longer maintained
2023-02-07 17:47:41 <ikke> so not a good target to migrate to
2023-02-07 17:47:52 <Ermine> ikke: bb has runit too
2023-02-07 17:49:27 <Leolo_3> also, OpenRC is perfect
2023-02-07 17:50:06 <Newbyte> Leolo_3: I'm curious what you like about it
2023-02-07 17:51:55 <Leolo_3> I'm just really old school. OpenRC is sysvinit + dependencies.  and I really like an init system written in shell.
2023-02-07 17:53:25 <Ermine> It's written in both C and shell
2023-02-07 17:53:57 <Leolo_3> yes
2023-02-07 17:54:07 <Leolo_3> you wouldn't want PID=1 to be in shell :D
2023-02-07 17:57:17 <Oyo> https://github.com/OpenRC/openrc/blob/master/BUSYBOX.md
2023-02-07 18:00:14 <Ermine> pid 1 is a part of init system too
2023-02-07 18:02:19 <Leolo_3> yep
2023-02-07 18:02:43 <ikke> We do happen to use bb init
2023-02-07 18:02:47 <ikke> instead of openrc init
2023-02-07 18:21:10 <Oyo> ok, runit is already more than 4 years mantaining by bb's devs, why Alpine uses bb + openRC i still cant understand. 3.10.6 21.02.23 OpenRC appeared with BB as init https://distrowatch.com/table.php?distribution=alpine but in that time bb's devs were already maintaining runit in bb. That is why i
2023-02-07 18:21:10 <Oyo> have such question now.
2023-02-07 18:24:10 <minimal> Oyo: as already explained Alpine has existing for more than 4 years and used OpenRC/Busybox init, obviously none of the Alpine Devs saw a benefit to switch to using Busybox runit
2023-02-07 18:24:53 <minimal> plus there have been thoughts for Alpine to potentially switch over to s6 in the future
2023-02-07 18:35:31 <Oyo> ''..I have also compiled 's6' statically, and there are PETs, however, it is a solution that will add megabytes to the build. So, at this stage I am favouring Busybox Runit.." EasyOS main dev words
2023-02-07 18:38:52 <abby> runit is maintained by void now ikke 
2023-02-07 18:59:33 <em> hey, how do I edit GRUB in diskless mode?
2023-02-07 19:35:46 <psykose> there are plenty of distros with runit to choose from if you want runit
2023-02-07 19:36:07 <psykose> why people run in a circle and ask $current_bus_stop_distro to implement their favorite feature i don't really get
2023-02-07 19:39:58 <em> psykose: I dont understand what you mean!??
2023-02-07 19:40:15 <psykose> it's not a response to you
2023-02-07 19:40:33 <em> ah ok
2023-02-07 19:42:48 <em> void has runit
2023-02-07 20:02:29 <Oyo> psykose: im new in linux; i dont have anything favorite in linux, but if you use bb which already has full init applets why you dont use this?
2023-02-07 20:08:41 <Leolo_3> Oyo - are you asking "why not change?" or are you asking "why was this selected?"
2023-02-07 20:12:26 <Oyo>  Leolo_3: 2nd
2023-02-07 20:13:22 <elly> alpine predates runit being in busybox
2023-02-07 20:13:26 <elly> is why (imo)
2023-02-07 20:14:06 <minimal> Oyo: you have already been told the reason at least a couple of times before
2023-02-07 21:36:01 <pj> <psykose> "why people run in a circle and..." <- but muh systemd
2023-02-07 21:45:26 <Saijin_Naib> 6.1 just got LTS status 🎉
2023-02-07 21:50:24 <Saijin_Naib> Now time to ruin my good mood by reading the comments on Phoronix
2023-02-07 21:53:14 <elly> don't do that
2023-02-07 21:54:11 <Misthios> :)
2023-02-07 22:01:27 <Saijin_Naib> <elly> "don't do that" <- Too late, but thankfully the three comments there aren't awful. Time will fix that! 
2023-02-07 22:03:12 <Saijin_Naib> I always make the mistake of thinking maybe someone will share some more information or provide context on something I don't understand well and it will be great. And then I wade through pages of unhinged rants about X vs Wayland, Rust, or who knows what else 
2023-02-07 22:03:33 <Misthios> *looks at the xorg vuln comments*
2023-02-07 22:04:43 <psykose> pov: you are marching into battle https://img.ayaya.dev/L5aE6v8nlP7Z
2023-02-07 22:05:08 <Misthios> good luck
2023-02-07 22:42:59 <Saijin_Naib> Oh, that's got to be top tier dumpsterfire 
2023-02-07 22:44:56 <Saijin_Naib> 
ACTION uploaded an image: (201KiB) < https://matrix.org/_matrix/media/v3/download/matrix.org/XXkTGagfgXNkTfEsrYdUrRjZ/Screenshot_20230207-174346_1.png >

2023-02-07 22:44:59 <Saijin_Naib> First two comments are off on the right foot, at least 🤣
2023-02-07 22:49:37 <pj> i would prefer to not have to read phoronix outside of phoronix, thank youuuu
2023-02-07 22:55:07 <Saijin_Naib> Sorry 😭
2023-02-08 08:30:39 <JohannesJ[m]> what is phoronix?
2023-02-08 08:33:43 <abby> a linux news site
2023-02-08 08:59:19 <Ermine> with a toxic forum
2023-02-08 12:00:16 <consus[m]> wait, there is such a thing as non-toxic forum?
2023-02-08 12:01:13 <aron> sure it is: a closed forum with no registered users :P
2023-02-08 13:34:25 <qaqland[m]> Hi! Can offical build and add this package?
2023-02-08 13:34:27 <qaqland[m]> https://github.com/KDAB/codebrowser
2023-02-08 15:36:26 <eletrotupi> hello!
2023-02-08 15:36:52 <eletrotupi> I'm trying to get libwebcam to work on Alpine, and it's currently complaining on cannot execute cc1plus
2023-02-08 15:37:10 <eletrotupi> >gcc: fatal error: cannot execute 'cc1plus': execvp: No such file or directory
2023-02-08 15:37:17 <eletrotupi> does it ring any bells?
2023-02-08 15:37:59 <eletrotupi> duh, g++ wasn't installed
2023-02-08 16:37:41 <alexhaydock> Does anyone know what the Alpine packaging policy is for situations where an upstream package appears to have stopped producing "releases" and instead seems to develop as a mostly rolling-release model?... (full message at <https://matrix.org/_matrix/media/v3/download/matrix.org/PWstcgEwvsajpyfjTXCZzBDM>)
2023-02-08 16:39:25 <ikke> alexhaydock: firs thing you could do is asking whether if they are planning to make a release 
2023-02-08 16:41:27 <alexhaydock> Sadly that seems to be the one issue the current maintainers seem to have ignored, but some asked back in August last year and I did bump the thread fairly recently: https://github.com/radvd-project/radvd/issues/185
2023-02-08 16:41:27 <alexhaydock> It's what got me wondering about the general policy if they just simply... don't do releases anymore
2023-02-08 16:42:02 <ikke> ok
2023-02-08 16:42:29 <ptrc> alexhaydock: not sure if it's an official policy, but usually packages bump the version to one that doesn't exist yet, suffixed with _gitYYYYMMDD (date of the commit used)
2023-02-08 16:43:49 <ikke> It's currently in main, I suppose it should at least be moved to community
2023-02-08 16:43:52 <alexhaydock> ptrc:  Yeah I've seen that a few times (though not necessarily in Alpine). I wondered if there was some conditions that might have to be satisfied before that route got chosen. Or maybe it's just up to the maintainer? It's packaged in `main` though.
2023-02-08 16:44:10 <ikke> I don't see any dependencies on radvd
2023-02-08 16:44:34 <ikke> alexhaydock: there are no pre-written list of conditions
2023-02-08 16:45:08 <alexhaydock> I could ask the maintainer (Natanael Copa) directly and see what they think?
2023-02-08 16:54:58 <minimal> alexhaydock: as an aside, as an alternative "corerad" is also packaged for Alpine
2023-02-08 16:55:56 <alexhaydock> minimal: Thanks! I'll have to check that one out
2023-02-08 16:57:41 <alexhaydock> Looks very nice - the Prometheus metrics integration is a nice touch
2023-02-08 17:10:48 <alexhaydock> minimal: Well I got corerad up and running very quickly and it was remarkably easy to configure. Very nice, thanks. I notice it's in the Testing repo. Anyone know if I can contribute to the 'testing' that needs doing for it to get moved into community/main ?
2023-02-08 17:11:34 <ikke> alexhaydock: sometimes all it takes is poking the maintainer
2023-02-08 20:22:00 <l1x> hi
2023-02-08 20:22:15 <l1x> is it possible to install python 3.9 on alpine 3.17?
2023-02-08 20:22:22 <l1x> apk install python3=3.9.16
2023-02-08 20:22:27 <l1x> ERROR: unable to select packages:
2023-02-08 20:22:27 <l1x>   python3-3.10.9-r1:
2023-02-08 20:22:27 <l1x>     breaks: world[python3=3.9.16]
2023-02-08 20:23:19 <minimal> llx: no, Alpine 3.17 comes with Python 3.10.9 currently
2023-02-08 20:23:52 <minimal> l1x: ^^^
2023-02-08 20:25:44 <gmalsack> I dd the iso to usb. booted it on a minibox. unfortunately the minibox only has a serial console. can anyone tell me how to get the default alpine iso to output to a serial console?
2023-02-08 20:26:55 <l1x> minimal: thanks
2023-02-08 20:27:12 <l1x> so i need to downgrade alpine version if i want 3.9 right?
2023-02-08 20:29:00 <minimal> l1x: you'd have to go to Alpine 3.15 to get Python 3.9.16
2023-02-08 20:30:05 <minimal> (or Alpine 3.14)
2023-02-08 21:01:17 <l1x> minimal: thanks, i just did, it works
2023-02-08 21:01:40 <ikke> l1x: or build it yourself
2023-02-08 21:02:52 <l1x> ikke: no thanks, this is why we have packaging and package managers so I do not need to build it myself
2023-02-08 21:03:34 <abby> you could use a container
2023-02-08 21:03:34 <ikke> l1x: THen you should also accept what the package manager (repositories) provide
2023-02-08 21:03:57 <l1x> yep, i just was not sure how to look up which version of python is in which version of alpine
2023-02-08 21:04:37 <ikke> https://pkgs.alpinelinux.org/packages?name=python3&branch=v3.15&repo=&arch=&maintainer=
2023-02-08 21:04:38 <minimal> l1x: https://pkgs.alpinelinux.org/packages?name=python3&branch=edge&repo=&arch=&maintainer=
2023-02-08 21:05:11 <minimal> and change the 2nd field to whichever version of Alpine and click "Search"
2023-02-08 21:17:02 <l1x> thanks!
2023-02-09 05:38:33 <techt> Specifically, if there is a critical security issue with OpenJDK, which is included in 3.17.1 of Alpine Linux, how long will Alpine Linux take updates?
2023-02-09 05:38:33 <techt> I am looking for information on the support policy for applications included with Alpine linux when Alpine Linux goes EOL.
2023-02-09 05:38:33 <techt> Will Alpine Linux versions that are EOL not update their applications?
2023-02-09 05:42:38 <psykose> https://alpinelinux.org/releases/
2023-02-09 05:42:57 <psykose> specifically things in community/ are only updated on the latest release, so when 3.18 is released 3.17 won't be
2023-02-09 05:43:43 <psykose> it depends specifically per case sometimes though, openjdk is updated in 3.16 too by the maintainer (so 2 latest releases)
2023-02-09 05:44:00 <techt> Thank you for your infomation.
2023-02-09 13:56:34 <Ermine> Is gdb supposed to work in WSL 1 alpine?
2023-02-09 13:57:54 <psykose> probably not
2023-02-09 13:58:17 <psykose> wsl1 is basically wine but for linux on windows
2023-02-09 13:58:28 <psykose> i'm surprised anyone uses wsl1 still
2023-02-09 14:00:55 <Ermine> IIRC there are fs performance issues unless you create ext virtual drive in wsl2
2023-02-09 14:01:51 <Ermine> But I didn't look at this close;y
2023-02-09 14:01:55 <Ermine> closely
2023-02-09 14:59:52 <Saijin_Naib> The I/O performance issues are when you cross HV boundaries, so doing disk IO across the WSL2 instance to the Windows host or vice-versa, because this is done using Plan9 Network filesystem
2023-02-09 15:00:10 <Saijin_Naib> Disk I/O within the WSL2 instance is basically native performance
2023-02-09 15:07:40 <psykose> 9pfs mounting is a disaster
2023-02-09 15:08:08 <psykose> for all the elegance of various plan9 things, performance is not even within 20 fields
2023-02-09 16:47:48 <Saijin_Naib> Never used it, despite Bell Labs being big in my family 
2023-02-09 17:50:56 <Apachez> bell labs? you must have been early adopter of rfb (vnc) then? ;-)
2023-02-09 18:22:19 <Saijin_Naib> No, no Unix-likes at home
2023-02-09 19:42:39 <nicci> hey, is it possible to automate import of files/folders from disk to ram in diskless mode, to avoid incorperate the files/folder in lbu??
2023-02-09 19:44:16 <ptrc> you mean, fstab?
2023-02-09 19:44:38 <ptrc> or do you specifically need them loaded into ram
2023-02-09 19:46:21 <ikke> I guess overlay from disk 
2023-02-09 19:48:07 <nicci> I want to run the files from specifically from ram.
2023-02-09 19:48:36 <nicci> can you explain overlay from disk?
2023-02-09 19:53:27 <psykose> cat files &> /dev/null
2023-02-09 19:53:31 <psykose> there, now they're in page cache
2023-02-09 19:55:07 <JohannesJ[m]> heh
2023-02-09 22:07:04 <nicci> hey, I just installed dwm and trying to run mixxx but its not working. can someone have a look at the log from terminal?
2023-02-09 22:07:04 <nicci> https://logpaste.com/Z5TmBaD5
2023-02-09 22:09:41 <nicci> or I should say output from terminal...
2023-02-09 22:09:54 <nicci> dbus is running from start
2023-02-09 22:11:10 <psykose> you're missing `modprobe seq` i guess
2023-02-09 22:13:20 <psykose> aside from that everything is warnings, i guess the portaudio output would be the main cause
2023-02-09 22:13:27 <nicci> I serched for modprobe seq in "packages" on the alpine site and no hit
2023-02-09 22:13:37 <psykose> you literally type modprobe seq
2023-02-09 22:13:38 <dwfreed> that's a command, not a file
2023-02-09 22:13:58 <nicci> ah ok, to load this mod?
2023-02-09 22:14:34 <nicci> alsa is working fine with default card correct on other apps
2023-02-09 22:14:40 <nicci> like mpv...
2023-02-09 22:15:54 <nicci> modprobe: module seq not found in modules.dep
2023-02-09 22:17:31 <psykose> have to reboot after kernel upgrades
2023-02-09 22:18:34 <nicci> but this error came when typing modprobe seq
2023-02-09 22:20:33 <nicci> how do I install a Qt platform that might be missing?
2023-02-09 22:25:28 <nicci> this was spit back when I tried to start mixxx as root:
2023-02-09 22:25:29 <nicci> https://logpaste.com/9S1U1Zv4
2023-02-09 22:25:49 <nicci> could it be I am missing a Qt platform?
2023-02-09 22:26:23 <psykose> qt-qtbase-x11 i guess
2023-02-09 22:26:52 <psykose> which is required by mixxx already
2023-02-09 22:27:04 <psykose> ah
2023-02-09 22:27:05 <psykose> right
2023-02-09 22:27:09 <psykose> yeah it fails to connect to display
2023-02-09 22:27:32 <nicci> I have xf86-video-intel 
2023-02-09 22:28:11 <nicci> qt-qtbase-x11 is not avalable
2023-02-09 22:30:10 <nicci> qt5-qtbase-x11 works
2023-02-09 22:30:19 <psykose> whadya mean by "works"
2023-02-09 22:31:14 <nicci> ah sorry, I meant is avalable. but there is also a qt6-qtbase-x11
2023-02-09 22:31:31 <psykose> the issue is probably "Authorization required, but no authorization protocol specified"
2023-02-09 22:31:39 <psykose> i.e. weird xorg stuff you have to figure out yourself
2023-02-09 22:31:57 <nicci> should I intall gt5 or qt6?
2023-02-09 22:32:06 <nicci> ok
2023-02-10 12:44:18 <alexhaydock> Where am I going wrong if I can't get busybox cron to run anything? crond is actually running and enabled as a service. My run-parts scripts don't seem to run (I haven't touched the default crontab), even though they're 0755, executable, contain the right #!/bin/sh shebang at the beginning and otherwise work when I run the run-parts commands directly from the terminal
2023-02-10 12:45:43 <alexhaydock> I thought maybe it was a $PATH issue but even adding something super minimal like `/bin/echo > /tmp/test` as a cron entry with `* * * * *` as a schedule does nothing. I can see it supposedly executing in `/var/log/messages` if I configure crond to log verbosely but it doesn't actually do anything. I can't get anything to run
2023-02-10 12:46:48 <alexhaydock> I've noticed it prints child running /sbin/nologin after each execution when in very verbose mode, so I did wonder if it's trying to run as the cron user, which does have that null shell configured, but that's the default config for that user from the alpine-baselayout package, and these cron jobs belong to root so maybe that's not the issue
2023-02-10 14:19:50 <WhyNotHugo> When packages have man pages or openrc files, these get split into a seprate subpackage and the right files end up there. How does this magic happen?
2023-02-10 14:20:32 <ikke> abuild has default split functions defined that get used when you add openrc and doc subpackages 
2023-02-10 14:21:40 <WhyNotHugo> Ah, I see it now -> https://git.alpinelinux.org/abuild/tree/abuild.in
2023-02-10 14:21:42 <WhyNotHugo> Thanks.
2023-02-10 14:22:08 <WhyNotHugo> So if we wanted completions for a new shell, or for s6 or anything alike, it would need to be done there.
2023-02-10 14:34:00 <ikke> You can define your own split functions, but if you want something generic for all packages, then yes
2023-02-10 14:51:36 <minimal> alexhaydock: do you have a /var/log/cron.log file? that should have entries every time cron triggers
2023-02-10 14:52:37 <minimal> I just tested busybox cron and it worked as expected by creating a script and then using "crontab -e" to run that every 5 mins. I didn't use run-parts however
2023-02-10 14:52:39 <alexhaydock> minimal: I don't - seems like busybox cron feeds into /var/log/messages by default. But I can definitely see it triggering. It just doesn't have any useful output when it does trigger
2023-02-10 14:53:23 <alexhaydock> Seems like everything I try reports itself as if it's running fine based on all the logs I can see, I just can't find any evidence that it actually does run
2023-02-10 14:54:34 <minimal> do you have syslogd running?
2023-02-10 14:55:44 <alexhaydock> Yep, that's running. I can only assume it's syslogd that's writing /var/log/messages
2023-02-10 14:56:02 <minimal> ah, forgot I'm running rsyslog which is where the cron.log is coming from.
2023-02-10 14:56:09 <minimal> anyway, have you tried without run-parts?
2023-02-10 14:57:18 <alexhaydock> Yeah I've even tried super minimal cron entries like `* * * * * /bin/echo test > /tmp/test` and similar. It never creates anything or shows any evidence of doing anything even though crond does log to `/var/log/messages` to indicate it's supposedly running that cron entry every minute
2023-02-10 14:58:27 <minimal> I tested something similar (though I created a script and referenced it in the crontab entry rather than placing the commands there)
2023-02-10 14:59:18 <alexhaydock> I was trying to go as minimal as possible to cut down as many factors I thought might be causing the problem as possible, but I'm really stumped by this one :D
2023-02-10 15:01:07 <alexhaydock> I would have just swapped in a different cron daemon by now in the hope of skipping over the problem, but there's none in main so I thought I'd try fixing it
2023-02-10 15:01:18 <minimal> well it's working here. Which Alpine version are you using? I'm using Edge here
2023-02-10 15:02:38 <alexhaydock> Also edge, but specifically aarch64 in diskless mode on the off-chance that might somehow have an impact
2023-02-10 15:15:31 <pj> have you tried run-parts with debug mode
2023-02-10 15:15:33 <minimal> alexhaydock: an I also just tested via run-parts and again it works as expected
2023-02-10 15:16:02 <alexhaydock> Interesting
2023-02-10 15:16:09 <alexhaydock> pj: I'll have to try this and report back
2023-02-10 15:16:42 <alexhaydock> But I don't really think run-parts is the actual problem here, considering it won't run anything
2023-02-10 15:17:00 <minimal> pj: I don't see a debug option for Busybox run-parts
2023-02-10 15:17:00 <ikke> alexhaydock: haven't read chat history, but are there any dots it the script file names? 
2023-02-10 15:17:27 <ikke> strace :-p
2023-02-10 15:17:43 <minimal> ikke: he also tried putting shell commands straight into the crontab
2023-02-10 15:17:57 <alexhaydock> ikke: No, but I've definitely seen that as a suggested problem quite a few times when hunting around so I'll keep it in mind for the future haha
2023-02-10 15:18:20 <alexhaydock> Oh dear I hope I'm not at the stage of actually stracing things to see what they're doing but I suspect I might be
2023-02-10 15:18:32 <pj> minimal: --test is in bb
2023-02-10 15:19:07 <minimal> pj: "--help" shows that option as "Dry run"
2023-02-10 15:19:29 <pj> same thing really
2023-02-10 15:19:37 <alexhaydock> Oh --test definitely works. It shows me it will definitely run my script. And if I execute the run-parts command directly it definitely does. It's only when launched from cron that it doesn't run.
2023-02-10 15:19:37 <alexhaydock> But then neither does anything else 🤔
2023-02-10 15:20:26 <minimal> well personally I'd think of "dry run" as not run something but log which you might have run if "--test" was not specified, versus a debug option meaning run something and log detailed info about what/how you ran
2023-02-10 15:21:18 <minimal> alexhaydock: it can't be a run-parts issue as you said things also didn't work when you put shell commands directly in crontab
2023-02-10 15:22:01 <alexhaydock> Yeah that makes sense. I tried to see if it was an environment thing by trying stuff like `env > /tmp/env` as a cron entry, but of course I can't even see what env it runs under because it doesn't seem to even run that command 
2023-02-10 15:23:21 <alexhaydock> Like I say, the crond entries that end up in `/var/log/messages` when I increase the verbosity of crond's logging seem to claim it's running, but I'm not so sure it actually does get far enough to run these things
2023-02-10 15:23:21 <alexhaydock> I wonder if I am at the point of trying to use strace or something like auditd to see if it does actually get as far as running the execve for the script or whether it's not getting that far
2023-02-10 15:26:18 <alexhaydock> The child running /sbin/nologin might be the most interesting one to me since that is definitely the default cron user's shell, but since these are all in /etc/crontabs/root I wouldn't really expect it to be dropping privilege at any point. But maybe that's just me misunderstanding something about how crond works internally
2023-02-10 15:26:52 <alexhaydock> Worth noting that I have validated the crond service is definitely running as root
2023-02-10 15:32:05 <minimal> strace on crond shows it reading the "root" password entry, I don't think it uses the cron user at all
2023-02-10 15:32:52 <alexhaydock> Ohhh you know what
2023-02-10 15:33:04 <alexhaydock> I'm gonna SSH in but I bet I've set my root shell to /sbin/nologin
2023-02-10 15:33:12 <alexhaydock> Since I also locked the root user
2023-02-10 15:33:34 <alexhaydock> Yep... I have. That'll do it 🤦‍♂️
2023-02-10 15:41:15 <alexhaydock> Well I learned something at least. Looks like crond (at least this implementation) seems to try and launch everything in a shell so I can't even seem to trick it by making the cron entry something like /bin/sh run-parts ... .... Will just have to change root back to being able to spawn a proper shell I guess.
2023-02-10 15:43:07 <alexhaydock> Thanks all for the help anyway :) 
2023-02-10 15:47:42 <minimal> alexhaydock: I tried changing the shell of another user to nologin and its cronjobs still run fine, via strace I don't see any lookups of that user's /etc/passwd entry, only for root (as crond runs as that)
2023-02-10 15:47:51 <minimal> so you problem is a root-only one
2023-02-10 15:47:59 <minimal> s/you /your/.
2023-02-10 15:47:59 <alpine-meetbot> minimal meant to say: so yourproblem is a root-only one
2023-02-10 15:48:14 <alexhaydock> Quite interesting that it would only impact root
2023-02-10 16:07:54 <docmax_> anyone got alpine running on a steamdeck?
2023-02-10 16:11:20 <alexhaydock> docmax_:  I can't imagine it'd be too hard since the Steam Deck is really just an x86 PC and it supports UEFI and everything that a "normal" PC does.
2023-02-10 16:17:53 <invoked> steamos is just arch linux iirc
2023-02-10 16:26:10 <pj> it is
2023-02-10 16:29:55 <docmax> yeah, but the hardware, the graphics, input devices and so on. bit too much for little alpine?
2023-02-10 16:30:21 <docmax> also steamdeck got better support in kernel 6.1
2023-02-10 16:30:28 <elly> I would be surprised if it was too much for alpine?
2023-02-10 16:30:33 <pj> tiny alpine
2023-02-10 16:30:35 <elly> maybe gpu accel or something
2023-02-10 16:30:41 <docmax> and we are still on 5.xx
2023-02-10 16:31:08 <docmax> and what about wine dxvk vulkan and all that stuff?
2023-02-10 16:31:18 <pj> teenie weenie alpine
2023-02-10 16:31:18 <qaqland[m]> 5.19 is not old
2023-02-10 16:31:27 <elly> "we" aren't on 5.anything :)
2023-02-10 16:31:32 <elly> Linux esper 6.1.10-0-lts #1-Alpine SMP PREEMPT_DYNAMIC Mon, 06 Feb 2023 17:32:27 +0000 x86_64 Linux
2023-02-10 16:31:42 <docmax> oh nice
2023-02-10 16:31:46 <pj> 6.1 has been on edge for some time now
2023-02-10 16:32:14 <docmax> ok i would love to see alpine running steamdeck
2023-02-10 16:32:26 <elly> be the change you want to see in the world
2023-02-10 16:33:04 <ptrc> i'd be more than happy to port alpine to steamdeck if someone donated the hardware to me :)
2023-02-10 16:33:05 <qaqland[m]> Native Games maybe released on glibc🤔
2023-02-10 16:33:06 <docmax> i did a try with void linux some time ago
2023-02-10 16:33:19 <docmax> and went into trouble with packages wanting systemd functions
2023-02-10 16:33:42 <pj> smol alpine
2023-02-10 16:33:46 <ptrc> qaqland[m]: that's what steam linux runtime is for
2023-02-10 16:33:53 <pj> not able to hold this huge thicc steamdeck
2023-02-10 16:34:09 <ptrc> pj: 😳
2023-02-10 16:34:47 <pj> sorry, it's just too funny for me that someone would call "little alpine"
2023-02-10 16:35:45 <qaqland[m]> So sweet XD
2023-02-10 16:35:48 <ptrc> what do people usually say? "size doesn't matter, but how you use it"? :p
2023-02-10 16:37:13 <idkrn[m]> <invoked> "steamos is just arch linux iirc" <- Immutable
2023-02-10 16:42:55 <anonymouss> Hi! Why is nvidia-src in the alpine repository? Is it possible to run nvidia drivers on alpine?
2023-02-10 16:43:04 <docmax> compared to arch its little regarding resources
2023-02-10 16:43:32 <docmax> join my journey installing alpine on steamdeck
2023-02-10 16:43:48 <pj> anonymouss: yes but also no
2023-02-10 16:43:58 <anonymouss> ?
2023-02-10 16:44:01 <pj> don't bother, you won't be able to run the driver
2023-02-10 16:44:08 <anonymouss> (
2023-02-10 16:44:16 <anonymouss> Ok, thank you
2023-02-10 16:44:45 <docmax> nvidia not working in alpine?
2023-02-10 16:45:10 <anonymouss> Yes, the drivers are only compatible with glibc
2023-02-10 16:45:40 <anonymouss> Sadly
2023-02-10 16:45:55 <pj> nvidia gpus work "fine" with nouveau
2023-02-10 16:46:02 <docmax> maybe the open drivers of nvidia?
2023-02-10 16:46:49 <docmax> but AMD is ok
2023-02-10 16:47:26 <ptrc> fwiw, the nvidia-src drivers work only with specific nvidia cards and are more for tinkering than consumer use
2023-02-10 16:48:37 <elly> some nvidia gpus work fine with nouveau :(
2023-02-10 16:49:08 <ptrc> the proprietary userspace drivers don't work at all due to 'initial-exec TLS resolves to dynamic definition'
2023-02-10 16:49:21 <ptrc> but nouveau works fine, even for some basic gaming
2023-02-10 16:49:36 <docmax> why alpine decided to use musl libc?
2023-02-10 16:49:57 <pj> because it's small and secure
2023-02-10 16:50:39 <pj> it was also an upgrade from uclibc iirc
2023-02-10 16:52:02 <ptrc> it's a bit hard to find the original reason because of linkrot
2023-02-10 16:53:13 <ptrc> but going by the info from the leaf-devel ML, it was supposed to run from RAM, so one can assume glibc was too bloated for that usecase
2023-02-10 16:54:09 <qaqland[m]> Why vscode-remote-ssh  works when connect to alpine docker and alpine wsl while not in physical one. They are all musl libc🤔
2023-02-10 16:54:41 <qaqland[m]> * wsl while does not, * not work in physical
2023-02-10 16:55:01 <ptrc> ( fwiw, at peak there were 3 libcs supported by alpine - musl, uclibc and eglibc, the last one being "embedded glibc" - http://www.eglibc.org/home )
2023-02-10 16:55:03 <qaqland[m]> * Why vscode-remote-ssh  works when connect to alpine docker and alpine wsl while does not work with physical one. They are all musl libc🤔
2023-02-10 16:55:13 <anonymouss> "but nouveau works great, even for some basic games" <-- Agree, but for most modern games nouveau is not enough
2023-02-10 16:55:41 <anonymouss> I really like Alpine Linux but I just can't find a distro that uses glibc and apk package manager and doesn't use systemd :(
2023-02-10 16:56:54 <minimal> qaqland[m]: We had that question recently, vscode copies a glibc binary across to the destination system. Exactly which "alpine docker" are you using? I don't see how it would work either
2023-02-10 16:57:04 <ptrc> qaqland[m]: it would be helpful if you included something more than "not working"
2023-02-10 16:57:28 <Apachez> anonymouss: devuan?
2023-02-10 16:57:44 <Apachez> well it uses apt but close enough since you wont have the malware named systemd? :)
2023-02-10 16:58:08 <anonymouss> ;)
2023-02-10 17:01:34 <pj> minimal: qaqland: it's using custom alpine image
2023-02-10 17:02:01 <pj> mcr.microsoft.com/devcontainers/base:alpine-3.17
2023-02-10 17:03:17 <foul-secretion> hello
2023-02-10 17:04:40 <pj> so much for microsoft loving open source when they won't even help out with running nodejs/code-server on alpine
2023-02-10 17:07:46 <qaqland[m]> <ptrc> "qaqland: it would be helpful..." <- Error about missing Glibc. It is an old error 
2023-02-10 17:08:49 <qaqland[m]> ptrc  Error about missing Glibc. It is an old error 
2023-02-10 17:28:51 <minimal> qaqland[m]: pj's link about seems to "explain" things, the Microsoft devcontainer for Alpine has extra stuff in it to make vscode-remote-ssh work
2023-02-10 17:28:56 <minimal> s/about/above/
2023-02-10 17:28:56 <alpine-meetbot> minimal meant to say: qaqland[m]: pj's link above seems to "explain" things, the Microsoft devcontainer for Alpine has extra stuff in it to make vscode-remote-ssh work
2023-02-10 17:35:28 <qaqland[m]> Oh, I got it.
2023-02-10 17:36:06 <qaqland[m]> Hope MS would support musl oneday.
2023-02-10 17:36:29 <mdekstrand> qaqland[m]: i am sucessfully using vsode remote on a physical alpine box with the 'gcompat' package installed. vscode warns about glibc compatibility problems but it works.
2023-02-10 17:36:46 <docmax> will i get lots of problems if i switch to edge?
2023-02-10 17:37:20 <idkrn[m]> Depends on what you have installed
2023-02-10 17:37:43 <docmax> and why is wfi not comming up? i followed https://wiki.alpinelinux.org/wiki/Wi-Fi
2023-02-10 17:38:34 <docmax> rc-update add iwd boot && rc-update add dbus boot
2023-02-10 17:38:42 <docmax> allready added to boot
2023-02-10 17:41:01 <docmax> only works if i restart networking
2023-02-10 17:43:08 <minimal> docmax: running Alpine from ram? if so did you "lbu commit"?
2023-02-10 17:43:30 <docmax> no installed on disk
2023-02-10 17:43:46 <docmax> i see "eth0" "wlan0" no such device at boot
2023-02-10 17:44:01 <docmax> so i guess it needs to wait or something?
2023-02-10 17:48:15 <docmax> i mean i installed on external hard disk, is this also "diskless"?
2023-02-10 17:48:33 <docmax> booted from bios
2023-02-10 17:55:37 <minimal> docmax: you don't need "lbu2 if you did a Sys (disk mode) install
2023-02-10 17:56:01 <minimal> so iwd and dbus are not running after you boot?
2023-02-10 17:57:13 <docmax> they are, but no connection
2023-02-10 17:57:20 <docmax> only if i restart networking manually
2023-02-10 17:58:18 <docmax> yes i installed with sys
2023-02-10 18:14:16 <minimal> docmax: perhaps /etc/init.d/networking has a missing "after" entry for iwd?
2023-02-10 18:14:28 <docmax> elogind is running and i dont have  XDG_RUNTIME_DIR
2023-02-10 18:15:18 <docmax> after bootmisc hwdrivers modules
2023-02-10 18:15:24 <docmax> do i have to edit it?
2023-02-10 18:16:48 <minimal> ocmax: ah, no, it's ok , /etc/init.d/iwd defines a "after net"
2023-02-10 18:17:02 <minimal> s/after/before/
2023-02-10 18:17:02 <alpine-meetbot> minimal meant to say: ocmax: ah, no, it's ok , /etc/init.d/iwd defines a "before net"
2023-02-10 18:18:00 <minimal> if you edit /etc/rc.conf and set rc_logger to YES and reboot then /var/log/rc.log will record the each init.d service starting in sequence
2023-02-10 18:18:45 <minimal> were there no errors in any logfiles to show what might be going on?
2023-02-10 18:20:55 <docmax> https://wiki.alpinelinux.org/wiki/Elogind
2023-02-10 18:21:04 <docmax> installed this, no XDG_RUNTIME_DIR
2023-02-10 18:23:27 <minimal> docmax: I'm asking about you iwd/networking issue
2023-02-10 18:29:07 <docmax> minimal, http://ix.io/4nGm
2023-02-10 18:29:31 <docmax> have to go to the gym
2023-02-10 18:29:36 <docmax> will continue later
2023-02-10 18:31:03 <minimal> docmax: you did notice the error from udhcpc in that rc.log? ;-)
2023-02-10 18:32:07 <docmax> yes
2023-02-10 18:32:13 <docmax> what does it tell me?
2023-02-10 18:32:14 <minimal> so that's the problem
2023-02-10 18:32:32 <minimal> the eth0 device does not appear to exist at the time /etc/init.d/networking runs
2023-02-10 18:32:49 <docmax> eth0 never exists
2023-02-10 18:32:58 <docmax> but wlan0 exists
2023-02-10 18:33:10 <docmax> after network restart
2023-02-10 18:33:28 <docmax> the script has to wait for wlan0
2023-02-10 18:33:40 <minimal> ah, eth0 AND wlan0 - neither appear to exist at the time /etc/init.d/network initially runs during boot
2023-02-10 18:34:08 <docmax> no what can i do?
2023-02-10 18:34:22 <minimal> docmax: if eth0 does not exist at all (i.e. you have no ethernet card then where is eth0 referends? /etc/network/interfaces?
2023-02-10 18:34:39 <docmax> eth0 is a usb lan strick
2023-02-10 18:34:42 <docmax> stick
2023-02-10 18:34:52 <docmax> and wlan0 is build in
2023-02-10 18:35:02 <minimal> so it does exist? you earlier said "eth0 never exists"
2023-02-10 18:35:24 <docmax> it nevers exists because i never plugin the lan stick again
2023-02-10 18:35:33 <docmax> the lan stick was for install only
2023-02-10 18:36:48 <minimal> so is "eth0" mentioned in /etc/network/interfaces ?
2023-02-10 18:41:29 <minimal> alexhaydock: wondering why you were setting root's shell to nologin. Were you worried that locking root's account was not sufficient security?
2023-02-10 18:43:27 <alexhaydock> minimal: Pretty much, though not extremely concerned. It’s a router/gateway setup that I’m trying to make mostly immutable/stateless so I figured there’d (in theory) be no reason for root to ever need an actual shell. Today I learned though! :D
2023-02-10 18:46:17 <minimal> alexhaydock: I'm currently documenting differences in behaviour between busybox/shadow/util-linux login programs with regard to locked/disabled/expired accts etc
2023-02-10 18:47:12 <alexhaydock> minimal: Nice! Sounds like interesting research. Is it destined for the Alpine Wiki?
2023-02-10 18:48:27 <minimal> it may end up there eventually, like the UTMP stuff that I need to revise
2023-02-10 18:48:44 <Leolo_3> is /etc/modprobe.d still read?
2023-02-10 18:49:01 <Leolo_3> I have an aliases.conf in there with alias eth2 natsemi, but I replaced that NIC and it's now a skge
2023-02-10 18:49:16 <Leolo_3> I would expect the new nic to be eth3, but it's eth2
2023-02-10 18:49:50 <ikke> Leolo_3: is the modules service enabled?
2023-02-10 18:52:23 <Leolo_3> good question
2023-02-10 18:52:44 <Leolo_3> answer is "yes"
2023-02-10 20:12:48 <docmax> i'm back
2023-02-10 20:13:05 <docmax> and i want gamescope and steam to run
2023-02-10 20:13:18 <docmax> both installed
2023-02-10 20:14:27 <docmax> https://wiki.alpinelinux.org/wiki/Wayland
2023-02-10 20:14:48 <docmax> A login manager such as Elogind can configure this and other XDG environment variables automatically.
2023-02-10 20:15:09 <docmax> installed elogind but still no XDG_RUNTIME_DIR
2023-02-10 20:17:27 <ikke> I think that requires PAM
2023-02-10 20:17:34 <ikke> which means you should install shadow
2023-02-10 20:18:27 <docmax> why isn't this mentioned in the wiki? :-(
2023-02-10 20:19:11 <docmax> minimal ?
2023-02-10 20:21:38 <docmax> and what can i do about the wlan0 problem?
2023-02-10 20:21:50 <docmax> http://ix.io/4nGm
2023-02-10 20:22:28 <ikke> perhaps you are missing firmware for your NIC
2023-02-10 20:22:52 <docmax> no after doing netowrk restart after login wlan works
2023-02-10 20:23:16 <docmax> init needs to wait for wlan0 giong up
2023-02-10 20:26:05 <docmax> but are the init scripts really supposed to be modified?
2023-02-10 20:35:46 <docmax>  https://gitlab.alpinelinux.org/alpine/aports/-/issues/9079
2023-02-10 20:55:15 <docmax> please
2023-02-10 21:43:37 <docmax> i have no session with loginctl
2023-02-10 21:44:24 <ikke> I just use awesomewm without DM, so I have very little experience with those setups
2023-02-10 21:48:01 <docmax> if nobody is gonna help me i will throw alpine out the window
2023-02-10 22:00:07 <elly> go for it, then
2023-02-10 22:00:39 <elly> with that attitude odds are pretty good that within not very long you'll run into a problem that nobody knows how to fix and you'll end up uninstalling it anyway :P
2023-02-10 22:00:44 <elly> might as well save your own time
2023-02-10 22:08:14 <minimal> docmax: I use Alpine with udev and I have not had such problems. That issue you linked to is 4 years old, plus there is no /etc/conf.d/networking file presently in Alpine
2023-02-10 22:09:02 <docmax> i'm still stuck here https://wiki.alpinelinux.org/wiki/Wayland#XDG_RUNTIME_DIR
2023-02-10 22:09:18 <docmax> i installed elogind EXCACTLY as described
2023-02-10 22:09:33 <docmax> and i dont get those variables nor any seat!
2023-02-10 22:09:39 <docmax> i go crazy
2023-02-10 22:09:58 <docmax> loginctl says no sessions!!
2023-02-10 22:10:00 <minimal> Wiki articles are not official documents and so there is no guarantee that they are accurate
2023-02-10 22:10:25 <minimal> did you add the shadow package as ikke suggested?
2023-02-10 22:10:41 <docmax> yes no difference
2023-02-10 22:16:36 <docmax> i give up with stupid elogind and try seatd and pam_rundir
2023-02-10 22:16:51 <docmax> thanks for all your help!
2023-02-10 23:12:57 <docmax> heeeeeeeeellp
2023-02-10 23:19:33 <docmax> https://gitlab.alpinelinux.org/alpine/aports/-/issues/13619
2023-02-10 23:20:12 <ptrc> > We need help from someone who really understands PAM.
2023-02-10 23:20:23 <ptrc> i guess that's still true
2023-02-10 23:20:49 <docmax> XDG_RUNTIME_DIR simply does not work
2023-02-10 23:21:13 <docmax> session required pam_rundir.so is in /etc/pam.d/base-session
2023-02-10 23:21:23 <docmax> and it's not getting used?!?
2023-02-10 23:25:26 <docmax> the whole pam is messed up
2023-02-10 23:39:21 <docmax> got gamescope running
2023-02-10 23:40:52 <ptrc> docmax: you have util-linux-login installed, right?
2023-02-10 23:40:59 <docmax> yes
2023-02-10 23:44:57 <docmax> where is the execvp package?
2023-02-10 23:45:45 <docmax> ok installing gamescope with steam is a pain
2023-02-10 23:45:56 <docmax> xwm: execvp failed: No such file or directory
2023-02-11 00:43:18 <paper_> execvp is a function fro your libc, not a package
2023-02-11 02:10:28 <qaqland[m]> mdekstrand: Works now. It shows retry or close button before
2023-02-11 02:22:52 <mdekstrand> qaqland[m]: glad to hear it!
2023-02-11 09:28:48 <WhyNotHugo> Can I raise the ulimit or all my sessions without pam?
2023-02-11 09:29:16 <WhyNotHugo> E.g.: basically, what I do via /etc/security/limits.conf
2023-02-11 09:32:30 <psykose> afaik rc_ulimit on the getty instance or whatever
2023-02-11 09:32:43 <psykose> if you `ulimit` in a shell it sticks is the simple concept
2023-02-11 09:34:49 <psykose> so if you are root, call ulimit with whatever values, then doas into a user, it will keep it
2023-02-11 09:35:23 <psykose> you don't want to do that because it keeps around the doas the whole time for a shell of course, so an alternative could be `su-exec youruser /bin/sh`
2023-02-11 09:35:48 <psykose> and from there you can think of whatever you want, rc_ulimit sets ulimit for an openrc process, afaik if you do that on the getty that you log in with your shell will inherit it
2023-02-11 09:35:58 <psykose> i didn't test that though and just use regular limits.conf via pam
2023-02-11 09:36:37 <psykose> ah, right, getties are inittab and not through rc
2023-02-11 09:36:55 <psykose> so you probably have to change that as well :p
2023-02-11 10:03:08 <WhyNotHugo> That's the thing, not entirely sure how to raise it without inittab.
2023-02-11 10:03:13 <WhyNotHugo> Maybe I should just keep pam.
2023-02-11 10:12:08 <psykose> apk add agetty agetty-openrc
2023-02-11 10:12:20 <psykose> https://github.com/OpenRC/openrc/blob/master/agetty-guide.md
2023-02-11 10:12:24 <psykose> comment out the inittab
2023-02-11 10:14:17 <psykose> (or spawn this on only like 6 and comment out only 6 so you don't lock yourself out first)
2023-02-11 10:14:44 <psykose> then you can test rc_ulimit="-n 500000" in the conf.d/agetty.tty6
2023-02-11 10:15:01 <psykose> and ulimit -n from the session to see if it works
2023-02-11 12:37:57 <donoban> wow, after fixing the --user uid:guid problem with my script qutebrowser runs very smooth! I suppose that it was working without hardware acceleration and I didn't notice it
2023-02-11 12:38:51 <psykose> :)
2023-02-11 12:39:04 <psykose> that is good
2023-02-11 12:40:01 <donoban> yeah, the startup stills a pretty slow but the user experience is clearly better
2023-02-11 13:02:59 <novorol[m]> has anyone else experienced apk being stuck creating initramfs with the latest update for 3.17? how do I fix this without breaking grub?
2023-02-11 13:03:16 <ikke> haven't noticed / heard about it yet
2023-02-11 13:03:31 <psykose> do you mean the mkinitfs trigger being stuck? could post the console output
2023-02-11 13:59:36 <minimal> novorol[m]: why would a mkinitfs problem break grub?
2023-02-11 14:03:16 <novorol[m]> 
ACTION sent a code block: https://matrix.org/_matrix/media/v3/download/matrix.org/HKnmTfyNoKelArSpquCNVnau

2023-02-11 14:04:36 <c7s> novorol[m]: are you doing disk IO in general while this is happening?
2023-02-11 14:05:05 <psykose> initramfs-Its seems wrong
2023-02-11 14:06:11 <c7s> the script ends with "sync", so that's probably hanging
2023-02-11 14:06:27 <c7s> you can strace it to make sure
2023-02-11 14:06:28 <psykose> yea
2023-02-11 14:06:56 <psykose> strace -p on the pid of the script if you can find it will probably show it in fsync
2023-02-11 14:07:41 <minimal> wondering why the kmod trigger is warning about old versions of linux-edge and linux-lts
2023-02-11 14:43:28 <novorol[m]> <c7s> "novorol: are you doing disk IO..." <- nah, nothing happening in this matter
2023-02-11 14:43:44 <novorol[m]> * > <@_oftc_c7s:matrix.org> novorol: are you doing disk IO in general while this is happening?
2023-02-11 14:43:44 <novorol[m]> nah, nothing happening
2023-02-11 14:48:26 <psykose> i wonder if it should do `sync /boot` instead
2023-02-11 14:57:22 <psykose> sadly it is not recursive
2023-02-11 14:57:29 <psykose> annoying fsync design
2023-02-11 14:58:12 <psykose> why yes, you can sync your entire system, or fsync one single file. wonderful.
2023-02-11 14:58:44 <psykose> i guess sync -f is good enough
2023-02-11 21:17:02 <Barnerd> Struggling with overlay mount on a diskless install
2023-02-11 21:17:19 <Barnerd> Guess I must be missing some key info...
2023-02-11 21:17:58 <Barnerd> using `mount -t overlay -o lowerdir=/media/vda1,upperdir=/mnt/docker/cache/upper,workdir=/mnt/docker/cache/.work overlay /media/vda1`
2023-02-11 21:19:53 <Barnerd> /mnt/docker is an NFS mount, the directories exist there but getting error
2023-02-11 21:20:13 <Barnerd> mount: mounting overlay on /media/vda1 failed: No such file or directory
2023-02-11 21:22:38 <Barnerd> Based my config on https://wiki.alpinelinux.org/wiki/Raspberry_Pi#Loopback_image_with_overlayfs
2023-02-11 21:32:41 <Barnerd> It doesn't help that "
2023-02-11 21:32:48 <Barnerd>  overlay" is also the apkovl :D
2023-02-11 21:46:22 <Ermine> $pkgname installs locale data in /usr/lib/python3.11/site-packages/$pkgname/data/locale. I guess locales cannot be subpackaged in $pkgname-lang ?
2023-02-11 21:47:30 <psykose> if you write it manually sure
2023-02-11 21:47:38 <psykose> generally not detected and forgotten though
2023-02-12 00:08:45 <psykose> WhyNotHugo: funnily, harfbuzz 7.0 now seems to support colrv1 which i think was the requirement for more widespread rendering of that exact fancy font
2023-02-12 00:08:52 <psykose> so maybe one day it'll be in everything!
2023-02-12 00:45:51 <psykose> ah, freetype with cairo support has a ot-svg renderer too, technically
2023-02-12 10:47:43 <WhyNotHugo> psykose: Interesting. SVG-in-TrueType should have better scaling, so I'd be happy to be rid of my builds one day.
2023-02-12 10:51:06 <psykose> tis all over the place :)
2023-02-12 12:08:05 <psykose> WhyNotHugo: this comparison is quite cool https://colrv1-emoji-vs-cbdt.glitch.me/
2023-02-12 12:08:14 <psykose> (it's also a bit different in chromium vs ff)
2023-02-12 12:13:42 <Ermine> So I've migrated Alpine to WSL 2, and now man pages open very slowly. Bash completion for man is also slow af. Can man pages turn out to be outside of VM disk and require to be accessed with 9p?
2023-02-12 12:14:50 <ikke> I have no issues with man pages on WSL2
2023-02-12 12:15:29 <Ermine> tried strace on man, it tries to open/stat a lot of stuff from Windows (/mnt/c)
2023-02-12 12:16:11 <Ermine> MANPATH is empty
2023-02-12 12:21:35 <Ermine> Ah, that is because of PATH, which is also used by man
2023-02-12 12:21:49 <ikke> makes sense
2023-02-12 12:22:09 <psykose> man opens stuff in PATH? what for?
2023-02-12 12:22:23 <Ermine> *shrug*
2023-02-12 12:22:55 <Ermine> But this is what manpath(5) says
2023-02-12 12:23:52 <Ermine> Anyway, cleared path, and man performance is back to normal
2023-02-12 12:46:19 <psykose> poggers
2023-02-12 13:31:35 <Fxzxmic> I think Alpine is very suitable for WSL, isn't it?
2023-02-12 13:32:09 <ikke> I would say so
2023-02-12 13:33:22 <Fxzxmic> :)
2023-02-12 13:38:20 <Fxzxmic> I am using the Konversation on WSL1 on Windows 11 to chat here. :)
2023-02-12 13:39:45 <Fxzxmic> I am using the Konversation in Alpine on WSL1 on Windows 11 to chat here.
2023-02-12 13:39:48 <Fxzxmic> Correct it.
2023-02-12 13:46:25 <psykose> corrected
2023-02-12 14:03:55 <charmango> Hi. While upgrading this error occured "ERROR: php81-8.1.15-r0: trying to overwrite usr/bin/php owned by php7-7.4.33-r0." File /usr/bin/php does not exist. apk info shows that php81-8.1.15-r0 is installed.
2023-02-12 14:10:18 <lopid> i think you have to symlink it yourself
2023-02-12 14:10:33 <lopid> php7 is not installed now, right?
2023-02-12 14:16:50 <charmango> I don't think php7 is installed anymore. The binary is not there.
2023-02-12 14:18:35 <charmango> I need to symlink /usr/bin/php to /usr/bin/php81?
2023-02-12 14:18:55 <lopid> yes
2023-02-12 14:19:05 <lopid> if it is not already
2023-02-12 14:19:54 <lopid> i had to do the same, maybe because one package is replacing another, versus one package being upgraded
2023-02-12 14:21:32 <charmango> Great. Thank you. 'php --version' now works. 
2023-02-12 14:33:00 <psykose> you just had to `apk fix`
2023-02-12 14:39:41 <charmango> Thanks. Deleted the symlink that I added. Ran `apk fix php` which recreated the symlink.
2023-02-12 14:42:08 <charmango> Good to know both.
2023-02-12 15:54:17 <JohannesJ[m]> oh! psykose i just saw the release notes for 3.17.2 .. how on earth do you pull off so many fixes? Awesome!
2023-02-12 15:56:45 <psykose> the secret is that it's not actually very many
2023-02-12 15:58:33 <Ermine> Perception
2023-02-12 15:59:00 <JohannesJ[m]> well, perception or not, i think its awesome still
2023-02-12 16:04:17 <Saijin_Naib> Nah, psykose is a force to be reckoned with
2023-02-12 16:07:18 <JohannesJ[m]> :)
2023-02-12 16:31:15 <Saijin_Naib> I can't even figure out how to maintain like, 3 packages in Community yet, and they've got like 50 🤣
2023-02-12 16:31:40 <Saijin_Naib> That's more a reflection of my limits though 🤔
2023-02-12 16:31:55 <Saijin_Naib> Still! Big thanks to everyone and the hard work they do 
2023-02-12 16:35:09 <invoked> bumping a thousand packages is easy, but knowing whether they work as intended is a whole other thing
2023-02-12 17:44:54 <charmango> Just FYI as this is probably a zabbix thing but the upgrade from zabbix-5.4.12-r1 to zabbix-6.2.6-r1, after restarting zabbix-server trying to log into the web interface gives "The Zabbix database version does not match current requirements. Your database version: 6020000. Required version: 5040000. Please contact your system administrator." The database is postgres.
2023-02-12 18:38:09 <JohannesJ[m]> <Saijin_Naib> "Still! Big thanks to everyone..." <- Well, to you too ofcourse :-) i maintain 0 packages, so you’re still far ahead to most of us :)
2023-02-12 20:50:41 <pj> maintaining packages really depends just on how much free time you have
2023-02-13 02:25:13 <shoatd11[m]> Welcome to my Service,I sell fresh cvv (visa,master,amex,dis,bin,dob,fullz..) all country, Dumps track 1&2, Account Paypal, Bank Login, do WU transfer and Gift Card.
2023-02-13 02:25:13 <shoatd11[m]> - I sell cvv Fresh - Fast and Good price.looking for a long term partner who's willing to work , hit me up on telegram @GeeTrust . will be looking forward to hear from you 
2023-02-13 04:48:28 <elly> wow, I haven't seen one of those in a while
2023-02-13 06:30:49 <psykose> same
2023-02-13 17:11:22 <mayli> matrix ad spammer?
2023-02-13 17:11:58 <psykose> also that kind of specific message :p
2023-02-13 17:16:39 <elly> yeah, I meant the message really
2023-02-13 17:20:02 <em> hey! If I configure .xinitrc with 'exec mixxx --kiosk' then all good, but is there a way to set .xinitrc and run this with nice and taskset values?
2023-02-13 17:20:58 <psykose> exec nice .. mixxx ?
2023-02-13 17:21:05 <psykose> or did you mean something else
2023-02-13 17:21:43 <em> I tried ' exec nice -n -20 taskset -c 1 mixxx '
2023-02-13 17:22:13 <psykose> aha, taskset can't exec
2023-02-13 17:22:13 <em> but the nice value was not correct in htop in ssh
2023-02-13 17:22:26 <psykose> you probably don't care about nice there
2023-02-13 17:22:36 <psykose> you'd have more value in using `chrt` and making it realtime or something
2023-02-13 17:22:49 <em> ah sorry, this is what I tried: exec nice -n -20 taskset -c 1 mixxx --kiosk
2023-02-13 17:24:44 <em> I I have this command line in grub: intel_pstate=disable isolcpus=1 nohz_full=1 rcu_nocbs=1 transparent_hugepage=never mitigations=off       and then I execute: nice -n -20 taskset -c 1 mixxx
2023-02-13 17:25:36 <em> all to isolate a core as good as I can just for mix to give it lower latency
2023-02-13 17:26:21 <psykose> it works if i reverse it
2023-02-13 17:26:36 <psykose> i.e. taskset -c 1 nice -n -20
2023-02-13 17:26:55 <em> in .xinitrc?
2023-02-13 17:27:03 <psykose> just in terminal
2023-02-13 17:27:09 <psykose> if i run it the other way around the nice is not applied
2023-02-13 17:27:20 <psykose> xinitrc doesn't matter for that
2023-02-13 17:27:24 <em> yes in X it works fine
2023-02-13 17:27:25 <psykose> but i do mean the nice is probably meaningless there
2023-02-13 17:27:43 <em> is it??
2023-02-13 17:28:08 <em> I have no problem in this order, it show up as -20 in htop
2023-02-13 17:28:40 <psykose> sure, nice only means something within a process group iirc
2023-02-13 17:28:43 <psykose> and the whole x session is one
2023-02-13 17:28:52 <psykose> SCHED_FIFO is probably more useful
2023-02-13 17:28:58 <em> I am trying to find info now about chrt...can you give a hint on how to add this?
2023-02-13 17:28:58 <psykose> chrt -f 10 ..
2023-02-13 17:29:11 <psykose> this requires special permissions or root
2023-02-13 17:29:26 <psykose> you can also investigate if wayland works better, instead of .xinitrc just `cage mixxx` with cage installed iirc
2023-02-13 17:29:41 <em> so for example: sudo chrt -f 10 taskset -c 1 mixxx ?
2023-02-13 17:29:47 <psykose> i guess
2023-02-13 17:30:00 <em> what is the value 10?
2023-02-13 17:30:04 <psykose> 10 prio
2023-02-13 17:30:13 <em> and what is the -f flag?
2023-02-13 17:30:19 <psykose> sched_fifo
2023-02-13 17:30:25 <psykose> you can just read chrt --help or google it
2023-02-13 17:30:33 <em> sure okok
2023-02-13 17:30:44 <psykose> even 1 is higher than any other thread you can spawn
2023-02-13 17:31:06 <em> do I need real-time kernel for this?
2023-02-13 17:31:10 <psykose> no
2023-02-13 17:31:17 <em> why is that????
2023-02-13 17:31:23 <psykose> it sounds like you don't actually know anything about what you are doing
2023-02-13 17:31:31 <em> that is true
2023-02-13 17:31:31 <psykose> so i'm going to tell you to just try it and find out if it works or not
2023-02-13 17:32:06 <psykose> and if it doesn't (i.e. does not have your latency/whatever requirement), you can go from there
2023-02-13 17:32:14 <em> are we not all on our way to better understaing? 
2023-02-13 17:32:40 <psykose> for a start i doubt you want to lock to one core unless the program is broken in some fashion and not intended for this to begin with
2023-02-13 17:33:23 <em> I have a dual core, that why. giving one of the core to everything else...
2023-02-13 17:33:47 <psykose> that makes more sense
2023-02-13 17:33:57 <em> I talked to one of the guys at suckless who told me to look into how to isolate a core fully in this way. 
2023-02-13 17:34:19 <psykose> that's unfortunate, sorry that happened
2023-02-13 17:34:49 <em> and I know that its not easy to not have the kernel to schedule "anything" else on this dedicated core, but I will try
2023-02-13 17:35:10 <em> why is that unforunate?
2023-02-13 17:35:31 <em> you dont agree whit these guys?
2023-02-13 17:36:00 <psykose> isolcpus does prevent the kernel from using the core
2023-02-13 17:36:09 <psykose> so you're doing that well
2023-02-13 17:36:32 <em> not fully according to Erik Rigtorp    https://rigtorp.se/low-latency-guide/
2023-02-13 17:37:25 <elly> this all sounds kind of convoluted, what are you actually trying to do?
2023-02-13 17:40:06 <em> I am trying to give a specific program a fully uninterupted processing flow. from disk to soundcard.
2023-02-13 17:40:11 <elly> why?
2023-02-13 17:41:08 <elly> or rather: what's the specific problem you have that you're trying to fix? is the audio stuttering or buffering? are you not getting the throughput you expect? something else?
2023-02-13 17:41:42 <elly> (these are different problems that require different fixes)
2023-02-13 17:41:47 <em> I dont think this is the place for that discussion. but if you are willing to listen then its about reducing jitter
2023-02-13 17:42:08 <psykose> i think it is the place for that discussion because otherwise this is just a glorified search engine chat
2023-02-13 17:42:40 <psykose> you need a problem to fix before you do all this :p
2023-02-13 17:42:50 <psykose> not some hypothetical jitter that may or may not affect anything you are doing
2023-02-13 17:43:33 <elly> em: ok - are you experiencing jitter right now?
2023-02-13 17:43:42 <em> many say inte snakoil, but I can hear a difference when there is lower jitter, it just sound better. so there is no problem, its just a matter of doing it better
2023-02-13 17:43:57 <psykose> oh you're just high on drugs
2023-02-13 17:43:58 <psykose> apologies
2023-02-13 17:44:12 <elly> okay, yeah, never mind
2023-02-13 17:44:16 <elly> good luck with your problem
2023-02-13 17:45:46 <em> that hurt.
2023-02-13 17:47:00 <psykose> apologies then
2023-02-13 17:47:14 <psykose> but you are most likely imagining the difference unless you know how to test it otherwise
2023-02-13 17:48:21 <em> do you play aucustic instruments?
2023-02-13 17:49:27 <em> I guess not
2023-02-13 17:49:51 <elly> unless you have actual hard measurements, taken from inside the audio stack with a profiling tool, nobody here is likely to put a lot of stock in your subjective experience
2023-02-13 17:50:53 <elly> sorry, that's kind of just how it is :P since the problem is subjective the fix will be too, and you should do whatever will make you feel like you are getting a better result
2023-02-13 17:50:55 <em> All I asked for was to get some help on reducing jitter, I was not asking for Dr Phill to have a saying
2023-02-13 17:52:15 <elly> my #1 tip would be to not use software (mixxx) that is written partly in javascript then, since a single gc pause will annihilate anything you could conceivably do to the kernel
2023-02-13 17:52:20 <elly> but, that's really up to you
2023-02-13 17:54:10 <em> its just for prototyping, I am sure mixxx is far from clean code. and do I dare say it, that I can hear it
2023-02-13 17:54:48 <psykose> my favorite acoustic instrument is `cat /dev/urandom | aplay -`
2023-02-13 17:55:29 <psykose> doesn't sound very good however.. maybe i need some fancy quantum random thingie
2023-02-13 17:56:23 <psykose> em: i think you've done pretty much everything you can (in terms of args, if you also add the chrt), short of making a unikernel for it
2023-02-13 17:56:24 <psykose> shrug
2023-02-13 17:58:50 <em> I guess, I am looking into freeRTOS and other embeded stuff in the future. But I will tell sound people when they ask if a hardware synth sound the same as a software version, that they should not talk to you about it....
2023-02-13 18:05:14 <em> here is a good articale on how to get in touch with your feelings, I hope it can help:
2023-02-13 18:05:14 <em> https://www.lionsroar.com/how-to-drop-into-your-body-feelings/
2023-02-13 18:09:24 <ty3r0x> hello, I have installed alpine with linux-virt and the wireguard module isn't working. modprobe wireguard returns `modprobe: FATAL: Module wireguard not found in directory /lib/modules/5.15.90-0-virt`.
2023-02-13 18:09:47 <Barnerd> Anyone using NFS overlays of / ?
2023-02-13 18:09:58 <ikke> ty3r0x: what does uname -r return?
2023-02-13 18:10:14 <ty3r0x> ikke: 5.15.90-0-virt
2023-02-13 18:13:52 <psykose> are you sure? it's definitely built and normally there
2023-02-13 18:14:10 <psykose> the most common issue is you upgrade the kernel and the old modules are not kept so you need to reboot to the new one
2023-02-13 18:14:16 <ikke> indeed
2023-02-13 18:14:33 <psykose> you can `find /lib/modules -name "*wireguard*"` to see i guess
2023-02-13 18:15:03 <ikke> ty3r0x: what does /lib/modules return?
2023-02-13 18:15:09 <ikke> ty3r0x: what does ls /lib/modules return?
2023-02-13 18:16:18 <ty3r0x> no the issue was I forgot to restart
2023-02-13 18:16:27 <ikke> That's what we were trying to tell
2023-02-13 18:16:29 <ty3r0x> thanks for the help ikke and psykose
2023-02-13 18:17:23 <ty3r0x> hard to keep track on a cluster :^)
2023-02-13 18:18:07 <ikke> But I keep forgetting that the modprobe output by definition returns the same version as uname -r
2023-02-13 18:19:10 <psykose> same
2023-02-13 18:19:48 <ikke> It's even so bad that I recalled there was something with it but concluded that it was nothing :D
2023-02-13 18:24:54 <psykose> god same
2023-02-13 18:30:02 <HaikuUser> Is there a linux live boot cd that recognizes BFS-partitions?
2023-02-13 18:34:04 <psykose> probably not
2023-02-13 18:34:23 <psykose> i guess you'll have to grep for CONFIG_BFS_FS CONFIG_UNIXWARE_DISKLABEL on every distro kconfig
2023-02-13 18:34:35 <psykose> (.. who even uses that?)
2023-02-13 18:37:21 <HaikuUser> I used SuperGrub for it but it does not recognise BFS-partion.
2023-02-13 18:37:54 <HaikuUser> BFS-partition
2023-02-13 18:40:32 <psykose> you probably have to pay someone to use this haha
2023-02-13 18:41:01 <Ermine> psykose: fwiw CONFIG_MINIX_FS=m in virt kernel
2023-02-13 18:41:08 <psykose> is it that one?
2023-02-13 18:41:18 <Ermine> No, it's another
2023-02-13 18:41:36 <Ermine> Just on topic of old file systems
2023-02-13 18:44:29 <psykose> you could probably open an issue for that
2023-02-13 18:44:36 <psykose> and CONFIG_OMFS_FS
2023-02-13 18:44:37 <psykose> xD
2023-02-13 18:44:44 <psykose> for virt at least
2023-02-13 18:44:49 <psykose> (to remove)
2023-02-13 18:50:20 <elly> minix users in shambles
2023-02-13 18:54:46 <Ermine> elly: Linux does not support current version of Minix FS anyway
2023-02-13 18:55:23 <psykose> there's a current version ?
2023-02-13 18:56:46 <Ermine> Third version. Linux supports FSses shipped with Minix 1.x and 2.x AFAIK
2023-02-13 18:58:31 <Ermine> Also, Minix splits partitions further into slices, not sure whether Linux supports that
2023-02-13 18:58:59 <grawity> linux supports BSD_DISKLABEL, iirc
2023-02-13 18:59:03 <grawity> if that's what minix uses
2023-02-13 19:01:45 <Ermine> Probably
2023-02-13 19:21:40 <Ermine> psykose: done
2023-02-13 19:25:21 <psykose> ty
2023-02-13 20:52:21 <mbx> good evening. Say I want to install Alpine Linux on a bare metal MacBook Pro 2019 with T2 chip. I need to patch the kernel to support the T2 Intel Kernel chipset. I wonder if anybody in the community does this already?
2023-02-13 20:53:07 <psykose> patch it with what
2023-02-13 20:53:26 <mbx> t2linux.org
2023-02-13 20:53:52 <mbx> basically, the stock kernel is incompatible with MacBooks from 2016 to 2019
2023-02-13 20:54:20 <mbx> Some people maintain patched kernels for some distros, such as Fedora
2023-02-13 20:54:52 <elly> I'm surprised the stock kernel isn't compatible
2023-02-13 20:55:57 <psykose> alpine doesn't really ship patched kernels no
2023-02-13 20:56:04 <psykose> i guess m1 is an exception
2023-02-13 20:56:07 <mbx> Well it would boot I think, but hardware components won't work
2023-02-13 20:56:33 <psykose> you'd have to diy the whole process or merely download it from another distro
2023-02-13 20:57:39 <psykose> love outdated patched kernels
2023-02-13 20:57:40 <psykose> sheesh
2023-02-13 20:58:12 <elly> I don't really understand why this is a separate patchset and the stuff hasn't been upstreamed
2023-02-13 20:58:16 <elly> are there a bunch of binary blobs?
2023-02-13 20:59:05 <mbx> elly: no binary blobs, see https://github.com/t2linux/linux-t2-patches for the full patch list
2023-02-13 20:59:22 <elly> oh, but you do need binary blobs for a bunch of the devices it seems
2023-02-13 20:59:24 <mbx> the binary stuff comes on top, like the propriatary drivers
2023-02-13 20:59:54 <psykose> binary firmware loading is always fine for whatever in there
2023-02-13 20:59:58 <psykose> upstreaming is just hard
2023-02-13 21:00:26 <elly> wow these are some girthy patches
2023-02-13 21:01:14 <mbx> and I thought I was smart to get an Intel based one aha, if owning an Apple device was smart in the first place
2023-02-13 21:01:41 <psykose> but yeah if you want to use this "with alpine" you have to do it yourself or cp from another distro
2023-02-13 21:02:09 <mbx> They got nix configs which is nice, having the same for Alpine would be dope, but I guess it's not worth it
2023-02-13 21:02:31 <psykose> eh
2023-02-13 21:02:43 <mbx> is the Alpine Linux kernel untouched?
2023-02-13 21:03:05 <psykose> i would've said yes but we do carry 1 actual non-build-system patch now
2023-02-13 21:03:12 <psykose> so there's one single patch
2023-02-13 21:03:46 <psykose> it sets kexec to be disabled by default and enabled only with kexec_load_disabled=0 instead of the opposite
2023-02-13 21:04:33 <mbx> makes sense
2023-02-13 21:04:53 <mbx> could you please link the patch for me?
2023-02-13 21:05:17 <psykose> https://git.alpinelinux.org/aports/tree/main/linux-lts
2023-02-13 21:06:54 <mbx> alright, I was hoping I could avoid compiling the kernel myself, at least I know where the trip is going if I tackle this.
2023-02-13 21:07:38 <mbx> maybe I nag the t2linux community, maybe somebody is more motivated than me :D
2023-02-13 21:07:55 <psykose> i mean we don't exactly accept vendor/patched kernels
2023-02-13 21:08:11 <psykose> there's 2-3 that are some stale thing someone added ages ago that are now just bitrotted and that's about it
2023-02-13 21:08:24 <psykose> there is a downstream distro of alpine that is more focused on exactly this though
2023-02-13 21:08:32 <psykose> and you can just add the repo into repositories and use one from there
2023-02-13 21:08:35 <psykose> so if you wanted..
2023-02-13 21:08:39 <psykose> https://gitlab.com/postmarketOS/pmaports/
2023-02-13 21:08:52 <psykose> benefits of same package manager i guess
2023-02-13 21:09:10 <psykose> that is aside the obvious of "just unpack the arch linux pkg and cp the kernel"
2023-02-13 21:09:18 <psykose> because that would work, kernels aren't that special for the most part
2023-02-13 21:11:43 <mbx> I'm still very new to kernel patching. Say I cp my own kernel into the installer, how would I do it and how could I prevent apk from overwriting my current kernel? I really wish to stay on the edge and not just pick a downstream, unless it's maintained properly.
2023-02-13 21:13:18 <psykose> i don't mean install the downstream
2023-02-13 21:13:23 <psykose> i mean you literally install only a kernel
2023-02-13 21:13:37 <psykose> and you can't anyway, since there is no t2linux in pmos, i just mean it's where you'd put it
2023-02-13 21:13:53 <psykose> aside from that apk doesn't delete your kernels and what you boot is just what you select in your bootloader
2023-02-13 21:14:06 <psykose> if you create a /boot/fancy-kernel that would just stay there
2023-02-13 21:14:07 <mbx> ah, so installing Alpine normally, then compile my patched kernel and install it, pull, patch and install it again when apk overwrites my kernel with a new version
2023-02-13 21:14:17 <psykose> it wouldn't overwrite anything
2023-02-13 21:14:29 <psykose> they're just files
2023-02-13 21:14:46 <psykose> if you create a /boot/vmlinuz-lts and then install linux-lts it would overwrite it, sure
2023-02-13 21:14:59 <psykose> if you name it.. not the name of a file from a package though, so there's no conflict...
2023-02-13 21:15:06 <mbx> so an additional kernel is another file and another boot entry?
2023-02-13 21:15:35 <psykose> a kernel is literally just a file in the most basic way possible yes
2023-02-13 21:15:54 <psykose> (and then the whole module tree, but that is unique to the kernel version and doesn't conflict in the same way)
2023-02-13 21:16:06 <psykose> ((and potentially devicetree etc but none of this conflicts trust me))
2023-02-13 21:16:32 <psykose> and then you configure your bootloader however you want same as it always works
2023-02-13 21:16:40 <psykose> you either know how that works or you don't :
2023-02-13 21:16:41 <psykose> :p
2023-02-13 21:17:27 <mbx> haha yes, np, I'm not expecting a deep dive, but it's helpful to get an overview about things, so I can search properly
2023-02-13 21:18:35 <mbx> I will then dig into kernels and bootloaders if I decide to go that road, there will be a day this device will run Linux first, no matter what ;)
2023-02-13 21:18:51 <psykose> perhaps
2023-02-13 21:19:01 <psykose> afaik t2 macs work pretty ok if you get past these steps
2023-02-13 21:19:21 <psykose> personally to me it's easy, so i lost track of remembering how daunting it is for someone not me
2023-02-13 21:19:23 <mbx> I've seen Ubuntu and Fedora on it, absolutely
2023-02-13 21:20:00 <mbx> and that's predictable, everybody needs to start somewhere
2023-02-13 21:20:17 <psykose> first safe-r sanity test: get a device you run linux on that's just mainline, no patches, easy stuff, and build your own kernel
2023-02-13 21:20:30 <psykose> figure out how to install it (no helper scripts) and add it to the bootloader configuration in use
2023-02-13 21:20:47 <psykose> from there you'll learn the rest
2023-02-13 21:21:14 <psykose> and literally just after that, you'll know enough to just unpack an existing package from some other distro (no guarantees, there can be broken stuff, but in general..)
2023-02-13 21:21:18 <mbx> that is definitely the way to go. external keyboards work on mac no matter what, all that is necessary to start
2023-02-13 21:21:23 <psykose> and then do that, from another unpatched kernel :p
2023-02-13 21:21:31 <psykose> and *then* try the t2 stuff
2023-02-13 21:21:56 <psykose> maybe it's a day of learning, maybe it's a week
2023-02-13 21:21:58 <psykose> hope it's fun
2023-02-13 21:22:34 <mbx> if I do this, I have to share it, I can already feel the pain thinking about it aha
2023-02-13 21:24:06 <psykose> (also i totally forgot the whole "generate mkinitfs" extra, but that's usually just a single command once you know it (distro/tool specific))
2023-02-13 21:24:16 <psykose> if one is even in use anyway
2023-02-14 04:20:20 <Ignaty> hi! i really like this distro so far. I have one little issue: when I run "rc-service lightdm start" it says "Starting Display Manager... [ ok ]" and... Nothing happens. I also made it load on startup (rc-update add lightdm) and rebooted the machine, and the same thing happens, although it says that starting was ok, nothing happens. I tried, as wiki suggests, install alpine-desktop for an easy installation, but it says "no such 
2023-02-14 04:20:20 <Ignaty> package". Any suggestions? :)
2023-02-14 04:24:19 <psykose> i think there's a var/log/lightdm or something you could look at
2023-02-14 04:24:28 <psykose> also check if it didn't crash instantly (i.e. still running)
2023-02-14 04:25:43 <elly> you can also supply --debug to rc-service
2023-02-14 04:26:02 <psykose> does it stay attached for a bit?
2023-02-14 04:27:10 <elly> only one way to find out
2023-02-14 04:28:26 <psykose> hehe
2023-02-14 04:29:34 <psykose> answer: no
2023-02-14 04:47:37 <kwadroke[m]> maybe missing lightdm-gtk-greeter or similar? I've had that problem before
2023-02-14 07:47:20 <kreyren_> Looking for this guy https://gitlab.com/travmurav who was messing with https://wiki.postmarketos.org/wiki/Acer_Aspire_1_(A114-61) bcs my friend got the same hardware and needs help so ideally someone fetch him for me pleasee~ ^-^
2023-02-14 07:47:25 <kreyren_> oh wrong channel
2023-02-14 09:30:20 <telmich> Probably already seen by many, however still important: https://seclists.org/oss-sec/2023/q1/75
2023-02-14 11:57:42 <lavaball> i cna't find the supported devices list. can i install it on a bananapi?
2023-02-14 11:58:31 <lavaball> oh, never mind. found it. 
2023-02-14 14:41:28 <Ignaty> hi? 
2023-02-14 14:48:39 <lopid> ?
2023-02-14 15:37:13 <bleb> haven't been able to get X working on my new laptop
2023-02-14 15:38:55 <bleb> I ran setup-xorg-base and installed cwm, and put exec cwm in ~/.xinitrc
2023-02-14 15:39:27 <bleb> got some errors that seemed to be about firmware and installed all the intel firmware packages
2023-02-14 15:39:50 <lopid> intel integrated graphics?
2023-02-14 15:39:56 <bleb> yea
2023-02-14 15:40:39 <bleb> now the (first) error I'm getting in /var/log/Xorg.0.log is (EE) Unable to find a valid framebuffer device
2023-02-14 15:41:16 <bleb> then (WW) Falling back to old probe method for fbdev
2023-02-14 15:42:03 <bleb> then (EE) Screen 1 deleted because of no matching config section.
2023-02-14 15:43:05 <bleb> I have a couple earlier warnings before those: (WW) the directory "/usr/share/fonts/TTF" does not exist.  Entry deleted from font path.
2023-02-14 15:43:14 <bleb> and the same thing for /usr/share/fonts/Type1
2023-02-14 15:43:41 <bleb> and (WW) Open ACPI failed (var/run/acpid.socket) (No such file or directory)
2023-02-14 15:44:57 <bleb> based on the logs it seems to be an issue with fbdevhw submodule of the i915.ko driver, but no clue what specifically could be causing the issue
2023-02-14 15:47:14 <bleb> any ideas?  would it help to upload the entire log file?
2023-02-14 15:47:59 <lopid> you can ignore the fonts warnings
2023-02-14 16:53:50 <bleb> oh no the wiki :(
2023-02-14 16:54:04 <ikke> bleb: ?
2023-02-14 16:56:23 <bleb> they burgered it
2023-02-14 16:56:59 <ikke> ah, the default theme you mean
2023-02-14 17:03:55 <aron> (the new theme is not my favorite either)
2023-02-14 17:54:39 <bleb> due to the subject matter of the wiki one can expect that most users will still be on desktop/laptop rather than mobile
2023-02-14 17:55:01 <bleb> given that it would have been better to keep the theme the same
2023-02-14 17:55:27 <elly> lol, my main use of the wiki is from my phone because my laptop isn't working :P
2023-02-14 17:55:35 <ikke> :D
2023-02-14 17:55:35 <elly> that's generally when I look at it
2023-02-14 17:56:15 <bleb> and do you find the new theme to be better?
2023-02-14 17:56:25 <bleb> it seems to have shifted from desktop optimized to mobile optimized
2023-02-14 17:56:59 <bleb> but I'm always skeptical of "mobile optimized" sites even on mobile
2023-02-14 17:57:20 <elly> I haven't used it much recently tbh
2023-02-14 17:59:40 <Saijin_Naib> elly: 🤣
2023-02-14 17:59:40 <Saijin_Naib> Too real for me 
2023-02-14 18:00:59 <Saijin_Naib> Every so often through elinks because I trashed my DE or session 🤣
2023-02-14 18:01:00 <ikke> I use Vector (2022), works fine for me
2023-02-14 18:12:01 <bleb> back to my xorg issue... some resources mention using vesafb if the hardware-specific driver doesn't work
2023-02-14 18:13:46 <bleb> that could potentially get past the "no framebuffer" error but idk why I need a framebuffer device if I'm using X
2023-02-14 18:17:10 <bleb> and modprobe doesn't recognize the name vesafb so not sure how I would even use vesafb or if I'm using it already
2023-02-14 18:20:10 <lavaball> when i mount a usb drive in diskless, can i write on it?
2023-02-14 18:35:04 <bleb> xorg works now under linux-edge, looks like my graphics (intel integrated alder lake) are not supported on 5.15
2023-02-14 19:27:20 <vkrishn> compared to wikipedia wiki.a.o default font-size is bigger for same theme
2023-02-14 19:27:52 <vkrishn> oh, no its same
2023-02-14 19:31:58 <vkrishn> on firefox i reduce the font-size to 80%, and use "Dark background and light text" add-on, kinda better
2023-02-14 20:56:07 <lavaball> anyone installed alpine to a bananapi? 
2023-02-14 21:20:29 <lavaball> never mind. figured it out.
2023-02-14 22:10:51 <lavaball> can i not use ext4? setup-alpine tells me: No disks available. Try boot media /media/mmcblk0p1? but there is mmcblk0p2 with ext4 ready and waiting. 
2023-02-14 22:18:37 <lavaball> made a /media/mmcblk0p2 folder and mounted it there. seemed to have worked.
2023-02-14 22:20:24 <lavaball> though shouldn't there have been a diskless data sys question? because nobody asked me.
2023-02-14 22:33:25 <lavaball> so since there was the question for cache, i'm guessing data option is forced?
2023-02-14 22:38:56 <lavaball> do i also do the setup-bootable on an arm device?
2023-02-14 23:38:57 <minimal> lavaball: you are trying to install on the SDcard that you are booting Alpine from?
2023-02-15 19:09:20 <frag> does busybox date accept 'date -d YYYYMMDD' ? man busybox seem to suggest so(?)
2023-02-15 19:15:57 <quinq> $ busybox date -d 20230215
2023-02-15 19:15:57 <quinq> date: invalid date '20230215'
2023-02-15 19:17:54 <frag> man lies, kk
2023-02-15 19:19:11 <elly> does date(1) claim it supports -d like that?
2023-02-15 19:19:25 <elly> FTR, it does accept: date -d 2023-02-15
2023-02-15 19:19:55 <ikke> There is a difference between specifying the time to set and printing it
2023-02-15 19:20:01 <psykose> it displays a "TIME"
2023-02-15 19:20:09 <psykose> not a FORMAT from the default current TIME
2023-02-15 19:20:19 <psykose> TIME has formats; those are the ones you can pick to write it in
2023-02-15 19:20:24 <psykose> -d YYYYMMDD is not a time
2023-02-15 19:20:27 <psykose> it's just letters
2023-02-15 19:20:38 <psykose> -d 20230112 is a time, in a format of YYYYMMDD
2023-02-15 19:21:33 <psykose> -d @2 is accepted for example as a format too
2023-02-15 19:21:58 <psykose> it is a bit unintuitive- this is why nobody understands pointers :p
2023-02-15 19:22:03 <ikke> That's unixtime 2
2023-02-15 19:22:14 <psykose> unixtime was so good
2023-02-15 19:22:18 <psykose> they made unix time two !
2023-02-15 19:27:11 <frag> so is the man page wrong? :S
2023-02-15 19:28:01 <psykose> no
2023-02-15 19:28:03 <psykose> what is wrong about it
2023-02-15 19:29:29 <frag> i thought it said -d 20230215 would work, but i guess not :]
2023-02-15 19:30:00 <psykose> ah
2023-02-15 19:31:22 <psykose> date -d 199810111213
2023-02-15 19:31:37 <psykose> the pattern starts at `mm` as the required part
2023-02-15 19:31:39 <psykose> in the middle
2023-02-15 19:31:49 <psykose> so that works, as does 9810111213
2023-02-15 19:31:54 <psykose> and so on, removing from the left
2023-02-15 19:31:57 <psykose> do you see it?
2023-02-15 19:32:18 <psykose> i guess i also misread it :p
2023-02-15 19:32:30 <psykose> but "YYYYMMDD" is not listed as one, it's all inside [[]] from mm
2023-02-15 19:32:36 <frag> oooh..
2023-02-15 19:32:44 <ty3r0x> how do I enable verbose wireguard output to dmesg in alpine?
2023-02-15 19:33:19 <frag> no, you looking at this line psykose ? [[[[[YY]YY]MM]DD]hh]mm[.ss] 
2023-02-15 19:33:23 <psykose> ye
2023-02-15 19:33:27 <frag> isnt YY the required part?
2023-02-15 19:33:30 <psykose> no
2023-02-15 19:33:37 <psykose> the non[] parts are required
2023-02-15 19:33:41 <psykose> everything in [] is optional
2023-02-15 19:33:44 <frag> aha mm yeah
2023-02-15 19:33:54 <frag> ok, i see it :]
2023-02-15 19:33:56 <psykose> ty3r0x: probably the same way as everywhere else
2023-02-15 19:34:21 <ty3r0x> psykose: no dice, I get ash: can't create /sys/kernel/debug/dynamic_debug/control: nonexistent directory
2023-02-15 19:34:22 <psykose> i googled it and apparently wireguard only really has debugfs logging? so idk
2023-02-15 19:34:30 <psykose> do you even have debugfs mounted
2023-02-15 19:34:42 <quinq> frag, manpages says [[[[[YY]YY]MM]DD]hh]mm[.ss]
2023-02-15 19:34:53 <quinq> Erf ok, I'm a bit late, nevermind
2023-02-15 19:35:17 <ty3r0x> psykose: seems so
2023-02-15 19:35:29 <ty3r0x> resource busy
2023-02-15 19:36:12 <psykose> can you paste the console of what you typed somewhere
2023-02-15 19:37:58 <ty3r0x> echo 'module wireguard +p' > /sys/kernel/debug/dynamic_debug/control ?
2023-02-15 19:39:46 <psykose> you select the entire contents of your terminal, and you paste them into a pastebin
2023-02-15 19:41:20 <ty3r0x> you mean everything I typed so far?
2023-02-16 01:02:23 <Armote[m]> if you have a root user UID/GID 0 but with no CAPs, is there some way to trick an application into seeing a non-root UID? some apps don't like running as root, and in this instance due to some limitations, running as non root breaks permissions on files.
2023-02-16 01:02:39 <Armote[m]> this is inside a OCI compliant container launched as rootless.
2023-02-16 01:05:09 <Armote[m]> details in relevant: running google's gvisor runsc in rootless from inside a container apparently breaks ownership, no matter what you do everything is owned by root, so launching container as a non UID 0 means you can't touch any files.
2023-02-16 01:06:31 <elly> I can't really think of a way to do that... I was pondering whether there's some clever thing you can do with user namespaces, but no, they solve the reverse problem (make an app that isn't UID 0 believe it is)
2023-02-16 01:09:09 <Armote[m]> actually that may have worked, but gvisor doesn't implement linux perfectly. so that doesn't appear to work. bwrap: Can't read /proc/sys/kernel/overflowuid: No such file or directory
2023-02-16 01:11:01 <Armote[m]> i believe the primary issue here is the inability to blind apps to what UIG/GID they are running as. perhaps a clever LD_PRELOAD hack would trick them?
2023-02-16 01:14:02 <elly> it seems possible, but if they are statically linked then that won't work
2023-02-16 01:59:46 <Maxpm> I believe LD_PRELOAD is how fakeroot works. But fakeroot also solves the reverse problem, making a non-root program think it is root.
2023-02-16 02:02:37 <dwfreed> elly: with user namespaces, you can make it so a range of UIDs in the child namespace map to 1 UID in the parent, so you could make 0-1000 in the child all be 0 in the parent
2023-02-16 05:58:57 <idkrn[m]> <Maxpm> "I believe LD_PRELOAD is how..." <- I thought it just used bwrap
2023-02-16 06:01:26 <idkrn[m]> To use userns
2023-02-16 06:02:20 <psykose> fakeroot doesn't use bwrap
2023-02-16 06:02:27 <idkrn[m]> <Armote[m]> "i believe the primary issue here..." <- Why can't bwrap do this?
2023-02-16 06:02:37 <idkrn[m]> psykose: Just userns?
2023-02-16 06:03:24 <psykose> pretty sure it doesn't use that either
2023-02-16 06:03:34 <idkrn[m]> Oh
2023-02-16 06:03:38 <Armote[m]> idkrn[m]: seems like the kernel implementation gvisor is doing is not capable of supporting that functionality.
2023-02-16 06:03:38 <psykose> could be wrong
2023-02-16 06:04:14 <idkrn[m]> Armote[m]: What about when bwrap is run as root. Unprivileged userns might not exist in gvisor
2023-02-16 06:04:22 <idkrn[m]> * What about when bwrap is run as root? Unprivileged userns might not exist in gvisor
2023-02-16 06:04:23 <psykose> gvisor is more a pile of hacks than anything realistic
2023-02-16 06:06:02 <idkrn[m]> I guess that would already have been obvious if that was the issue
2023-02-16 06:06:02 <Armote[m]> idkrn: i did attempt it as root. i got this error and i suspect even if its resolves others will get int he way. gvisor is very limited in that regard. bwrap: Can't read /proc/sys/kernel/overflowuid: No such file or directory
2023-02-16 06:06:52 <Armote[m]> the problem seems to arise when gvisor runsc is launched rootless. with root rbind mounts preserve owner.
2023-02-16 06:07:04 <idkrn[m]> Did you --proc /proc? Just curious
2023-02-16 06:07:31 <Armote[m]> yes. ps displays processes and it is one of the mounts gvisor performs
2023-02-16 06:08:02 <idkrn[m]> You seem to know more than I do, so I don't think I can help lol
2023-02-16 06:08:37 <idkrn[m]> idkrn[m]: With bwrap
2023-02-16 06:08:49 <idkrn[m]> If that wasn't clear
2023-02-16 06:09:54 <Armote[m]> the error was generated from: bwrap --dev-bind / / bash  inside gvisor. i cant think of any reason why google would dedicate dev hours to implement linux kernel namespace capabilities.
2023-02-16 06:11:40 <Armote[m]> LD_PRELOAD seems like a viable solution. there are just a few libc functions in need of hijack.
2023-02-16 06:13:21 <idkrn[m]> Armote[m]: Does that make using --proc redundant?
2023-02-16 06:13:44 <Armote[m]> it just passes the /proc of the caller
2023-02-16 06:14:07 <Armote[m]> pid isnt unshared
2023-02-16 06:15:46 <idkrn[m]> Would choosing a specific UID help?
2023-02-16 06:16:54 <idkrn[m]> I've bwrapped bash before and it never needed /proc…
2023-02-16 09:22:26 <Biswa96> Q. Why is samurai used instead of ninja-build?
2023-02-16 09:27:14 <Biswa96> Never mind, found answer https://lists.alpinelinux.org/~alpine/devel/%3CCAAOiGNwnzTomzOQ6YjH3t7VHmbfMRZUNXap8GM0Gy_mGS7aKhg%40mail.gmail.com%3E
2023-02-16 13:02:45 <lavaball> i have the same problem as this guy: https://dev.alpinelinux.org/~clandmeter/other/forum.alpinelinux.org/forum/installation/usb-mount-problem.html    plugged in hdd, can't mount it.    
2023-02-16 13:08:30 <lavaball> -t ext4. got it. tanks.
2023-02-16 13:08:31 <lavaball> h
2023-02-16 14:06:24 <Wyk72> hello guys
2023-02-16 14:06:38 <Wyk72> I have an isse with Alpine that plagues me since 3.11 
2023-02-16 14:06:45 <ikke> That's been a long time
2023-02-16 14:07:18 <Wyk72> when I boot from USB stick in diskless mode the usb disk gets mounted as "/media/usb" or "media/sda1" RANDOMLY
2023-02-16 14:07:36 <Wyk72> this insanity breaks all the LBU stuff apks etc
2023-02-16 14:07:45 <Wyk72> just experienced 1 minute ago
2023-02-16 14:08:02 <Wyk72> this morning the machine booted with USB disk mounted on /media/sda1
2023-02-16 14:08:26 <Wyk72> rebooted 1 minute ago (same machine same all), It's mounted on /media/usb
2023-02-16 14:08:55 <Wyk72> I know there is a parameter for init script like UUID= something
2023-02-16 14:09:08 <Wyk72> but how come there is no fix for this since 3.10 ?
2023-02-16 14:09:32 <Wyk72> this makes usb booting impossibly unreliable
2023-02-16 14:11:33 <Wyk72> system boots but without any apk (since apkcache is lost)
2023-02-16 14:23:12 <ikke> I think it's mdev / eudev that mounts the disk
2023-02-16 14:28:54 <Wyk72> @ikke: It kills me that is  totally random.../media/sda1 or /media/usb with no apparent reason at all
2023-02-16 14:28:59 <Wyk72> can't pinpoint it
2023-02-16 14:29:19 <minimal> Wyk72: what's in your /etc/fstab?
2023-02-16 14:30:32 <Wyk72> nothing
2023-02-16 14:30:34 <Wyk72> the defaults
2023-02-16 14:32:22 <Wyk72> "/dev/usbdisk    /media/usb      vfat    noauto,ro 0 0"
2023-02-16 14:33:04 <minimal> so /dev/usbdisk will be mounted as /media/usb...
2023-02-16 14:33:09 <Wyk72> but sometimes I get /media/usb empty, and /media/sda1 with the OS
2023-02-16 14:33:30 <minimal> and in those situations I assume /dev/usbdisk doesn't exist
2023-02-16 14:34:11 <Wyk72> actually it's a symlink di sda
2023-02-16 14:34:13 <Wyk72> to
2023-02-16 14:34:23 <Wyk72> lrwxrwxrwx    1 root     root           3 Feb 16 15:23 /dev/usbdisk -> sda
2023-02-16 14:35:44 <minimal> right, this link is created by mdev
2023-02-16 14:37:54 <minimal> look at the bottom of /lib/mdev/persistent-storage
2023-02-16 14:40:15 <minimal> Wyk72: do you have more than 1 storage device attached?
2023-02-16 14:43:07 <minimal> Wyk72: also in your /etc/lbu.conf what is the value of LBU_MEDIA?
2023-02-16 14:44:49 <Wyk72> just 1 usb key
2023-02-16 14:44:58 <Wyk72> but lots of usb devices
2023-02-16 14:45:07 <Wyk72> 5 webcams
2023-02-16 14:45:14 <Wyk72> and the key
2023-02-16 14:45:30 <Wyk72> but happens on systems with just the usb key
2023-02-16 14:46:21 <Wyk72> i.e.: you boot from system, you get /media/usb, you boot to another, you get /media/sda1
2023-02-16 14:46:33 <Wyk72> all random
2023-02-16 14:47:11 <Wyk72> LBU_MEDIA=usb
2023-02-16 14:47:15 <Wyk72> always usb
2023-02-16 14:47:19 <Wyk72> never touched
2023-02-16 14:47:32 <minimal> ok, was just wondering if the persistent-storage "rule" had a bug, at a quick glance it looks like if you have more than a single USB storage device then as *EACH* such device is discovered then a link from /dev/usbdisk to the device is created in turn. ikke: do you agree?
2023-02-16 14:47:38 <Wyk72> lbu works even if I get /media/sda1, but apkcache does not
2023-02-16 14:48:17 <Wyk72> I mean
2023-02-16 14:48:44 <minimal> Wyk72: lbu and apkcache have references to mounting /media/$LBU_MEDIA (so /media/usb in otherwords)
2023-02-16 14:48:53 <Wyk72> it works at boot, when the system reads it and prepares the overlay diff
2023-02-16 14:49:06 <Wyk72> it fails at writing of course when it gets the wrong path
2023-02-16 14:49:15 <minimal> s/apkcache/setup-apkcache/
2023-02-16 14:49:15 <alpine-meetbot> minimal meant to say: Wyk72: lbu and setup-apkcache have references to mounting /media/$LBU_MEDIA (so /media/usb in otherwords)
2023-02-16 15:10:50 <markand> I have strange bugs since 3.17 regarding X.Org on a raspberry pi where even after closing Xorg seems to stay zombie and keeps reference to /dev/dri/* 
2023-02-16 16:32:32 <bovis> Having followed the instructions here (https://wiki.alpinelinux.org/wiki/Nginx_with_PHP#PHP7_Installation), I get to the point where I check that nginx and php are running, with "ps aux | grep 'nginx\|php-fpm'", but output doesn't show php (http://sprunge.us/EYo6WK).
2023-02-16 16:33:16 <bovis> Starting nginx and php-fpm works fine. Restarting php-fpm works http://sprunge.us/JiRH5b but doesn't change the output of ps aux ^^
2023-02-16 16:34:36 <ikke> bovis: it's probably php7.2-fpm or something like that
2023-02-16 16:35:01 <ikke> php8.2*
2023-02-16 16:41:02 <bovis> ikke: "ps aux | grep php" gives the same result. I've tried it with multiple versions, first 7 according to the doc and now 8.2. Visiting localhost/phpinfo.php (made as a test file in the nginx root directory) to check says "page not found."
2023-02-16 16:54:21 <ikke> bovis: what does rc-status return?
2023-02-16 17:12:27 <bleb> anyone have a favorite or good way of connecting to slack without the official client
2023-02-16 17:12:36 <ikke> I don't use slack
2023-02-16 17:13:20 <elly> I do but I use the official client
2023-02-16 17:21:54 <bovis> bleb: https://github.com/wee-slack/wee-slack I used this every day for years as part of a midsize company. It had some trouble with threads, like it wouldn't always notify me when a new one was created. Otherwise, it was a lifesaver.
2023-02-16 17:23:05 <pj> ripcord is nice but paid and not source available
2023-02-16 17:29:54 <bleb> pj: looks like ripcord is currently free because it's still alpha?
2023-02-16 17:29:58 <bleb> gratis I mean
2023-02-16 17:30:20 <bovis> ikke: rc-status shows that php-fpm crashed. Logs lead me to the first step here (https://wiki.alpinelinux.org/wiki/Nginx_with_PHP#Configuration_of_PHP7) in creating environment variables in /etc/profile.d/. The variables weren't set, so the sed commands didn't work as intended.
2023-02-16 17:30:47 <pj> bleb: "currently" since many years
2023-02-16 17:31:00 <bovis> ikke: A restart didn't appear to set the variables. Should it have? Running the .sh in that directory manually didn't appear to work either.
2023-02-16 17:32:32 <bleb> pj: cool ty
2023-02-16 17:32:34 <ikke> /etc/profile.d is only used for interactive users, not for services
2023-02-16 17:32:46 <bleb> bovis: ty I will check it out
2023-02-16 17:33:09 <ikke> bovis: the service should be able to run out of the box
2023-02-16 17:34:43 <bovis> ikke: The php-fpm crash was "ERROR: [pool www] please specify user and group other than root". In the php config file, the field is blank. Those sed commands are supposed to replace it with "www" as user.
2023-02-16 17:36:57 <ikke> user = nobody
2023-02-16 17:37:00 <ikke> group = nobody
2023-02-16 17:37:01 <ikke> that's the default for me
2023-02-16 17:38:10 <bovis> I possibly overwrote those values with the original sed commands. If sed found PHP_FPM_USER to be "" then it wrote over "nobody" with ""
2023-02-16 18:01:53 <bovis> ikke: Changed the values in the sed commands individually. I guess that's what I was supposed to do all along? Definitely didn't understand that at first.
2023-02-16 18:02:01 <bovis> It works now. Thanks for pushing me in the right direction.
2023-02-16 18:20:43 <ikke> bovis: like I said, it should generally work out of the box
2023-02-16 18:33:05 <mbx_> hey :) connecting through tmux on the installer media of alpine right now. I encrypted a partition using luksFormat and I luksOpen'ed it, so it's available at /dev/mapper/crypt, but I can't get setup-disk to work. I try setup-disk -m sys -s 1024M /dev/mapper/crypt
2023-02-16 18:33:23 <mbx_> it returns No disks found :/
2023-02-16 18:36:15 <mbx_> ok fdisk -l /dev/mapper/crypt reports no valid partition table, but I don't know if this is expected for some reason
2023-02-16 18:36:40 <frag> 'date -d YYYYMMDDhhmm' is not POSIX, but in busybox date? strange?
2023-02-16 18:36:43 <minimal> mbx_: setup-disk supports creating/setting up LUKS alread
2023-02-16 18:36:49 <mbx_> also I did mkfs.ext4 -m0 /dev/mapper/crypt after luksOpen
2023-02-16 18:37:10 <mbx_> minimal: but only full devices, not partitions
2023-02-16 18:37:20 <mbx_> this is dual boot, I can't wipe my other partitions
2023-02-16 18:37:22 <minimal> I guess it may not handle setting up "sys" on an already encrypted device, it has "cryptsys" (from memory) as an option to set up it all
2023-02-16 18:39:45 <mbx_> minimal: afaik, there is also crypt as type which will be used to abstract cryptsetup via setup-disk, it then will ask for sys, data, etc. without crypt for the encrypted partition itself, but this all only works when I want to format everything and start from scratch
2023-02-16 18:40:13 <minimal> mbx_: normally if you are partitioning/formatting yourself before called setup-disk then you'd mount the rootfs (and any other filesystems) before calling setup-disk
2023-02-16 18:40:19 <mbx_> accoding to the docs, I can choose to drop the type, since giving a mounting point will act like sys, so the outcome is the same
2023-02-16 18:40:48 <mbx_> minimal: ah ok, I thought the mounting is done by setup-disk
2023-02-16 18:40:58 <minimal> what mount point? /dev/mapper/crypt is a device
2023-02-16 18:41:16 <mbx_> because the example of setup-disk gives "/dev/sda" as an example argument
2023-02-16 18:41:31 <ikke> /dev/sda is also a block device
2023-02-16 18:41:33 <minimal> mbx_: yes setup-disk will normally mount stuff after it has partitions and formatted etc, but you don't want that as you're setting up dualboot
2023-02-16 18:41:35 <mbx_> yes, that was confusing ok
2023-02-16 18:41:49 <mbx_> ok I will mount it, makes sense :)
2023-02-16 18:41:51 <mbx_> thx
2023-02-16 18:41:59 <ikke> but afaik it will not use a disk that is already formatted
2023-02-16 18:42:13 <minimal> yeah is you give it a device the intention is it will slice-n-dice it (partition, format etc)
2023-02-16 18:43:14 <minimal> mbx_: look at the setup-disk --help output
2023-02-16 18:43:45 <minimal> "If MOUNTPOINT is specified, then do a traditional disk install with MOUNTPOINT as root."
2023-02-16 18:44:59 <mbx_> ok it installs, but I got grup-install error: failed to get canonical path of `' cant stat /EFI/alpine/grubx64.efi
2023-02-16 18:46:06 <mbx_> but I installed grub-efi and efibootmgr and I set BOOTLOADER=grub and USE_EFI=1
2023-02-16 18:46:25 <minimal> well did you manually mount the ESP partition on top of rootfs before running setup-disk?
2023-02-16 18:46:33 <mbx_> u right, I forgot that
2023-02-16 18:48:22 <minimal> also setup-disk tells grub-install NOT to write to EFI vars (via efibootmgr)
2023-02-16 18:48:54 <minimal> also setup-disk will also install EFI/boot/bootx64.efi (which your other OS might also have previously installed)
2023-02-16 18:50:55 <mbx_> yes there is an EFI/boot/bootx64.efi, I must not mess up now. will setup-disk destroy my other boot entries?
2023-02-16 18:51:35 <mbx_> I think I will look into efibootmgr
2023-02-16 18:52:17 <mbx_> if I install my boot entry manually via efibootmgr, I don't have to repeat setup-disk, right?
2023-02-16 18:52:35 <minimal> mbx_: no, typically each OS will install files into a named subdir of EFI (e.g. EFI/alpine, EFI/debian)
2023-02-16 18:53:05 <mbx_> good
2023-02-16 18:53:18 <minimal> however EFI/boot is "special"  as it is the directory where the fallback EFI file is stored (bootx64.efi for x86_64)
2023-02-16 18:53:39 <minimal> like Highlander there can only be one ESP/boot/bootx64.efi obviously
2023-02-16 18:53:59 <mbx_> ok, I get it, I need to add my bootentry first, so EFI/alpine/grubx64.efi exists
2023-02-16 18:54:05 <minimal> this file is used if no UEFI boot variables exist or none can be used and so it "fallback" to this file
2023-02-16 18:55:06 <minimal> my point is, if your existing OS is Windows (for example) and it has installed a EFI/boot/bootx64.efi file then setup-disk will replace that file with another one (which is Grub)
2023-02-16 18:55:42 <mbx_> my system is macOS
2023-02-16 18:55:47 <minimal> depending on what the other OS is you may not want the existing bootx64.efi touched
2023-02-16 18:55:59 <mbx_> and efibootmgr returns a bunch of entries of stuff that I installed in the past
2023-02-16 18:56:15 <mbx_> I still have BOOTCAMP entries for a Windows that is wiped
2023-02-16 18:56:21 <mbx_> and some Linux distros
2023-02-16 18:56:55 <minimal> https://gitlab.alpinelinux.org/alpine/alpine-conf/-/blob/master/setup-disk.in#L324
2023-02-16 18:57:02 <mbx_> macOS uses EFI as well
2023-02-16 18:57:29 <minimal> as grub-install is being called with "--no-nvram" then it will not call efibootmgr to add a boot entry for Alpine
2023-02-16 18:58:09 <minimal> so then setup-disk appears to be relying on the fallback file to boot Alpine instead
2023-02-16 18:59:25 <minimal> so you may want to (a) keep the existing ESP/boot/bootx64.efi file and possibly copy it back in place after setup-disk overwrites it, and/or (b) run efibootmgr manually after setup-disk to add a boot entry for Alpine (so bootx64.efi is then not actually used)
2023-02-16 19:00:05 <minimal> also I assume your machine is not using UEFI Secure Boot
2023-02-16 19:00:41 <mbx_> this is confusing to me, because setup-disk tried to stat /EFI/alpine/grubx64.efi, so it doesn't look like it tries to rely on the fallback file
2023-02-16 19:00:48 <mbx_> yes, my system isn't UEFI
2023-02-16 19:00:53 <mbx_> macOS uses EFI
2023-02-16 19:01:02 <mbx_> this is a MacBook
2023-02-16 19:01:25 <minimal> mbx_: confusing how? setup-disk will write TWO files, EFI/alpine/grubx64.efi and EFI/boot/bootx64.efi
2023-02-16 19:01:43 <minimal> in reality they are identical files
2023-02-16 19:02:56 <minimal> EFI normally relies in boot variables, one per boot option, plus a BootOrder variable to define the sequence of those boot entries
2023-02-16 19:03:39 <mbx_> ok, I have to read things up, right now, this makes no sense to me, I need to dig deeper into documentation of EFI to understand whats going on
2023-02-16 19:03:46 <minimal> if no suitable boot entry is found then the fallback is used. As I pointed out setup-disk calls grub-install with "-no-nvram" and so NO boot entry variable will be added for Alpine
2023-02-16 19:04:04 <mbx_> my goal is to get rid of the old EFI boot entries and to add alpine without distroying my macOS boot entry
2023-02-16 19:04:16 <minimal> if you do have other exiting boot entry variables, i.e. for MacOS, then those will be consulted
2023-02-16 19:05:10 <mbx_> consulted?
2023-02-16 19:05:30 <minimal> the BootOrder variable listed the sequence of boot entries to check in turn
2023-02-16 19:06:12 <minimal> as it is currently you (I assume) only have an entry for MacOS, after running setup-disk this will not change - you will not have a boot variable for Alpine
2023-02-16 19:06:29 <mbx_> ok, so it will add a new entry to the fallback and modify the boot order, but why the error? I also checked with efibootloader and there is no alpine entry
2023-02-16 19:06:51 <mbx_> i have a bunch of entries, not just macos, old entries
2023-02-16 19:06:58 <mbx_> just no alpine entry
2023-02-16 19:07:02 <minimal> no - there is never any boot entry for the fallback, it is a fallback when no suitable boto entry is found
2023-02-16 19:07:23 <minimal> if fallback needed to be a boot entry then it wouldn't be a fallback...
2023-02-16 19:08:03 <minimal> I assume on a Mac there is a key you can press when you switch it on to give to a list of boot options to pick from
2023-02-16 19:08:27 <mbx_> this is correct
2023-02-16 19:08:36 <mbx_> and it will change the BootOrder respectively
2023-02-16 19:08:43 <minimal> I don't know if this list includes the fallback
2023-02-16 19:10:28 <mbx_> Here is a summary of efibootmgr: Boot0000-0004 is all Windows (i need to remove that), Boot0005 is nobara (I need to remove that), Boot0080,0081 MacOS, Boot0082,FFFF PciRoot
2023-02-16 19:10:54 <mbx_> ls /boot/efi/EFI returns APPLE, Boot, Microsoft, fedora
2023-02-16 19:11:05 <mbx_> PciRoot looks like Apple Recovery
2023-02-16 19:13:58 <minimal> right, and BootOrder should be there too
2023-02-16 19:14:08 <mbx_> yes
2023-02-16 19:15:08 <minimal> which does exactly what you'd expect, it defines the order of booting
2023-02-16 19:15:25 <mbx_> I hope I didn't nuke myself now, I removed Windows and nobara with efibootmgr -b XXXX -B and rm -r fedora
2023-02-16 19:15:35 <minimal> so the 1st (valid) entry listed is the one your Mac will boot when you turn it on
2023-02-16 19:16:00 <minimal> did you update BootOrder to remove those from the list?
2023-02-16 19:16:18 <mbx_> yes, my question is just: How to setup-alpine properly so it won't error about the missing file? My conclusion is I have to add the alpine boot entry myself
2023-02-16 19:16:34 <mbx_> No I did not update BootOrder, but BootOrder points to macOS
2023-02-16 19:16:45 <mbx_> ah wait, you are right, I need to update it
2023-02-16 19:16:58 <mbx_> no I don't, the numbers are persistent
2023-02-16 19:18:34 <mbx_> ok I found it in the docs, will try to add the boot entry myself now
2023-02-16 19:18:51 <minimal> persistent? I meant to remove the Bootxxxx numbers for Windows and nobara if present in BootOrder
2023-02-16 19:19:11 <mbx_> they are are not present, so all looks right
2023-02-16 19:19:21 <mbx_> I was in MacOS previously
2023-02-16 19:20:04 <minimal> error about missing file? you mean about /EFI/alpine/grubx64.efi? wasn't that because you didn't have ESP partition mounted to efi inside the rootfs?
2023-02-16 19:20:28 <mbx_> no, I already did that, but I still get the error
2023-02-16 19:20:33 <mbx_> that is what I was trying to say
2023-02-16 19:21:01 <minimal> I don't know what EFI/boot/bootx64.efi is on a Mac so it may be important to backup that file before setup-disk overwrites it
2023-02-16 19:22:06 <mbx_> can I store something on the installation media persistently?
2023-02-16 19:24:00 <minimal> you could copy the file somewhere else in the ESP partition
2023-02-16 19:24:16 <minimal> can you show the error you're seeing?
2023-02-16 19:26:57 <mbx_> setup-disk -s 1024M /mnt/crypt output: Installing system on /dev/mapper/crypt, Installing for x86_64-efi platform., Install: can't stat '/EFI/alpine/grubx64.efi': No such file or directory, ... and then it just installs the files
2023-02-16 19:27:43 <mbx_> the efibootmgr example in the docs doesn't consider encryption, so I'm a bit stuck. I wish setup-disk would work, I'm so close ^^
2023-02-16 19:28:40 <minimal> so you mounted the ESP partition as /mnt/crypt/boot/efi before running setup-disk?
2023-02-16 19:29:33 <mbx_> ohhhh, no, I mounted it to /boot/efi, me dummy dum dum, ofc I need to mount it into the same fs, I totaly forgot that, sec
2023-02-16 19:30:34 <minimal> having setup-disk magically handle multiple OS boot would be very complicated and there would be many corner-cases for it to fail on (and so risk making a computer unbootable)
2023-02-16 19:31:43 <minimal> yes, as far as setup-disk is concerned /mnt/crypt == "/" for the installation you're doing so everything needs to be inside of/relative to there
2023-02-16 19:41:36 <mbx_> ok, now it says it canÄt create /mnt/crypt/boot/efi/EFI/boot: No such file or directory and it also can't create /mnt/efi/EFI/boot/bootx64.efi as well. The problem is the folder is called Boot, not boot lol
2023-02-16 19:42:49 <mbx_> I renamed the folder to boot and now it works
2023-02-16 19:44:11 <mbx_> ls /mnt/crypt/boot/efi now returns alpine as well, but efibootmgr does NOT list an entry for alpine. Could this be a problem?
2023-02-16 19:44:21 <minimal> hmm, the ESP is a vfat partition, perhaps some DOS-related upper/lowercase related issue...
2023-02-16 19:44:53 <minimal> mbx_: I told you twice already that there will be NO boot entry for Alpine as setup-disk calls grub-install with "--no-nvram" option
2023-02-16 19:45:08 <mbx_> ok, now I got it
2023-02-16 19:45:15 <minimal> and pointed you to the comment in the code that states that
2023-02-16 19:45:16 <mbx_> Boot folder is the fallback
2023-02-16 19:45:28 <mbx_> then I think it worked
2023-02-16 19:45:35 <minimal> that was why I suggest you might want to create a boot entry yourself using efibootmgr
2023-02-16 19:46:10 <minimal> and which point then you can decide whether to adjust BootOrder or not
2023-02-16 19:46:34 <minimal> side question,  LUKS v1 or v2?
2023-02-16 19:46:57 <mbx_> ok, so I need to add that entry, so I can control the BootOrder, got it
2023-02-16 19:47:19 <mbx_> I use the cryptsetup that alpine ships with
2023-02-16 19:47:34 <minimal> mbx_: yes but did you create a v1 or v2 LUKS device?
2023-02-16 19:47:40 <mbx_> cryptsetup 2.5
2023-02-16 19:47:55 <minimal> or did not not specify and relied on default?
2023-02-16 19:48:34 <mbx_> I didn't specify any version, the partition was a LUKS partition from a previons installation and all I did was luksFormat, so I don't know if it's v1 or v2, can I figure that out afterwards?
2023-02-16 19:48:57 <minimal> do you have a separate /boot partition?
2023-02-16 19:49:04 <mbx_> yes
2023-02-16 19:49:12 <mbx_> no
2023-02-16 19:49:16 <mbx_> yes
2023-02-16 19:49:20 <mbx_> sorry, yes
2023-02-16 19:49:30 <minimal> final answer? ;-)
2023-02-16 19:49:55 <mbx_> I have a boot/esp partition
2023-02-16 19:50:08 <minimal> I didn't ask that...
2023-02-16 19:50:22 <minimal> do you have a separate /boot partition?
2023-02-16 19:50:42 <mbx_> then I don't
2023-02-16 19:50:44 <minimal> in other words, is the /boot directory part of the rootfs or a separate partition
2023-02-16 19:51:04 <mbx_> it's part of the rootfs, sorry for the confusion
2023-02-16 19:51:11 <minimal> well then if you created a LUKS v2 you're not going to get very far trying to boot lol
2023-02-16 19:51:55 <lavaball> no nnn for arm? this saddens me greatly.
2023-02-16 19:52:08 <mbx_> ok I somehow confused this with the boot esp partition that has my EFI records. I think I get it now, I need a /boot partition, so luks can decrypt the other partition to boot
2023-02-16 19:52:28 <minimal> Grub 2.06 allegedly supports booting from a LUKS v2 partition, in reality that code doesn't work (upstream *may* have finally fixed it in Git master a few weeks ago, I haven't tested yet)
2023-02-16 19:52:42 <mbx_> but interestingly, I only have this partition in my previous installation and it was using luks as well
2023-02-16 19:52:46 <mbx_> and it booted
2023-02-16 19:52:55 <minimal> was it v1 or v2, that is the point...
2023-02-16 19:53:09 <minimal> Grub boots fine off v1
2023-02-16 19:53:19 <mbx_> I don't know
2023-02-16 19:54:24 <mbx_> minimal: thanks for letting me know about this, then I know what's wrong when it won't boot ahah
2023-02-16 19:55:06 <minimal> you show be able to run "cryptsetup isLuks --type /dev/mapper/crypt to see the version
2023-02-16 20:01:09 <mbx_> wasnt working, but luksDump worked, version 2, I'm screwed ahaha
2023-02-16 20:02:41 <mbx_> could I install the edge grub to test their allegations? ;)
2023-02-16 20:02:53 <minimal> basically your kernel and iniramfs are in /boot, if the filesystem containing /boot is LUKS encrypted then Grub needs to be able to decrypt it to get to them, if that fs is LUKS v1 then Grub will work, if it is (as is your case) it is v2 then Grub (in theory) should work but in reality won't due to several Grub issues which upstream has known about for 1.5+ years
2023-02-16 20:03:14 <minimal> edge Grub? you mean grub from Alpine Edge? it is the same release as in Alpine 3.17
2023-02-16 20:03:35 <mbx_> na, just meant their up2date version
2023-02-16 20:03:37 <mbx_> on git
2023-02-16 20:03:59 <minimal> I'm intending to test build it to check (I last tried about 3-4 months ago)
2023-02-16 20:04:44 <mbx_> well I think that is enough for me today, I need to focus on some work now, I will re-attempt all this on a different day, but I learned a lot today. Thanks for your help minimal.
2023-02-16 20:05:17 <Newbyte> lavaball: did you check why it isn't built for arm?
2023-02-16 20:06:01 <mbx_> minimal: next time I will re-partition to make a /boot partition and then I repeat everything, gonna do it again to learn the process I suppose :)
2023-02-16 20:06:40 <minimal> also even with the recent changes/fix you'd still have to careful running luksFormat as Argon is not supported as PBKDF and I don't remember what the cryptsetup default is if you don't specify one
2023-02-16 20:07:57 <minimal> so you could either using LUKS v1 for the rootfs (with no separate /boot partition) or else use LUKS v1 for /boot and use LUKS v2 for rootfs
2023-02-16 20:10:02 <minimal> mbx_: you'd also want to think about creating a keyfile to prevent being prompted for passphrase twice (or more)
2023-02-16 20:10:50 <mbx_> I prefer the prompt
2023-02-16 20:11:08 <minimal> mbx_: being prompted more than once?
2023-02-16 20:11:17 <mbx_> no problem
2023-02-16 20:11:34 <mbx_> is LUKS v1 even recommend?
2023-02-16 20:11:42 <mbx_> isn't that dated already?
2023-02-16 20:12:08 <minimal> I don't think I've ever seen anyone recommend NOT to use it
2023-02-16 20:12:23 <minimal> obviously v2 would be preferable
2023-02-16 20:13:11 <minimal> but you can't use one of the v2 benefits (Argon) with Grub as they haven't written code to support it yet (AFAIK)
2023-02-16 20:13:31 <mbx_> unfortunately i have to go now, but u say if I want to have an easy time, I will just luksFormat v1 and repeat everything I did, then I don't need to format?
2023-02-16 20:13:39 <mbx_> i mean, then I don't need to partition
2023-02-16 20:13:41 <minimal> so you can only use PBKDF2 instead
2023-02-16 20:14:18 <mbx_> ok, I will read up the topic about luks
2023-02-16 20:14:30 <minimal> with Grub and v1 then you can use a single partition (apart from ESP for UEFI obviously)
2023-02-16 20:15:20 <minimal> assuming you don't want a swap partition
2023-02-16 20:16:23 <mbx_> yep, no swap partition for me
2023-02-16 20:16:42 <mbx_> hm, cant umount /mnt/crypt ...
2023-02-16 20:16:45 <mbx_> resource busy
2023-02-16 20:17:42 <minimal> have you a shell currently inside its' mountpoint?
2023-02-16 20:17:50 <mbx_> no, all closed
2023-02-16 20:18:28 <minimal> check "ps"? maybe something hung that's still inside there
2023-02-16 20:20:06 <mbx_> could it be cryptd, kcryptdf_io, etc?
2023-02-16 20:21:07 <minimal> wouldn't think so (those ps names are enclosed in "[xx]" square brackets?)
2023-02-16 20:21:21 <lavaball> Newbyte, i wouldn't even know how to check. can you point me in the reight direction?
2023-02-16 20:21:23 <mbx_> yes they are
2023-02-16 20:21:48 <minimal> so they're kernel "threads". They wouldn't be an issue
2023-02-16 20:22:45 <Newbyte> lavaball: as far as I can tell nnn is built for arm
2023-02-16 20:23:00 <Newbyte> it's built for armhf, armv7, and aarch64
2023-02-16 20:23:01 <Newbyte> in Alpine
2023-02-16 20:23:03 <lavaball> hm?
2023-02-16 20:23:06 <Newbyte> https://pkgs.alpinelinux.org/packages?name=nnn&branch=edge&repo=&arch=&maintainer=
2023-02-16 20:23:07 <lavaball> what am i doing wrong then. 
2023-02-16 20:23:18 <lavaball> i can find all the other packages and install them and info and great and everything.
2023-02-16 20:23:19 <Newbyte> do you have the community repo enabled?
2023-02-16 20:23:22 <lavaball> !
2023-02-16 20:23:25 <lavaball> probably not.
2023-02-16 20:23:27 <lavaball> thanks.
2023-02-16 20:23:36 <Newbyte> :)
2023-02-16 20:24:14 <mbx_> minimal: these processes are without []: /sbin/init /bin/login /sbin/getty -ash /sbin/syslogd /sbin/udhcpc /sbin/acpid /jsr 
2023-02-16 20:24:25 <lavaball> there it is. thanks so much.
2023-02-16 20:25:00 <mbx_> /usr/sbin/crond ... sshd weechat -ash ps -a less, jsr was a typo
2023-02-16 20:25:18 <mbx_> I don't see anything that should block umount
2023-02-16 20:25:43 <minimal> normally I'd expect that error when some process is sitting inside the /mnt/crypt filesystem
2023-02-16 20:26:25 <mbx_> ok -f not working either, how do I force this please?
2023-02-16 20:29:09 <minimal> fuser -v /mnt/crypt should show you what is using it
2023-02-16 20:30:22 <mbx_> I relogged, I still can't umount /mnt/crypt ...
2023-02-16 20:31:52 <mbx_> I had to detach with -l
2023-02-16 20:31:56 <minimal> oops, "fuser -v -m /mnt/crypt"
2023-02-16 20:33:00 <mbx_> okay, now all worked, probably won't boot, but I know how to tackle the problem later, thanks to you minimal, bye ^_^
2023-02-16 22:13:04 <lavaball> do i have to use reboot and poweroff? i tried init 0 and init 6, but it tells me i have to be pid 1.
2023-02-16 22:14:50 <ikke> There is no sysvinit with runlevels
2023-02-16 22:15:19 <ikke> (not numeric anyway)
2023-02-16 22:15:29 <ikke> so yes, use reboot / poweroff
2023-02-16 22:15:32 <lavaball> so no other fancy commands? shutdown -p now isn't working either.
2023-02-16 22:15:38 <lavaball> well, okay then.
2023-02-16 22:15:41 <lavaball> thanks a bunch.
2023-02-16 22:19:20 <ikke> Reboot and poweroff are fancy enough for me. 
2023-02-16 22:31:46 <lavaball> everybody has those.
2023-02-16 23:13:22 <elly> that's a good thing :P
2023-02-17 00:00:51 <dwfreed> just don't use 'halt' on systemd and expect your computer to actually turn off
2023-02-17 02:14:10 <qaqland[m]> ikke: me too!
2023-02-17 03:50:24 <elly> I just noticed that fopencookie(3)'s man page is... double gzipped
2023-02-17 03:50:40 <elly> like it's literally /usr/share/man/fopencookie.3.gz.gz
2023-02-17 03:50:50 <elly> and `man fopencookie` dumps a gzip stream to my terminal
2023-02-17 03:50:52 <psykose> huh
2023-02-17 03:50:54 <psykose> not on my machine
2023-02-17 03:50:57 <elly> odd
2023-02-17 03:51:06 <psykose> oh
2023-02-17 03:51:08 <psykose> ...
2023-02-17 03:51:13 <elly> /usr/share/man/man3/fopencookie.3.gz.gz is owned by man-pages-6.03-r0
2023-02-17 03:51:23 <psykose> there's both
2023-02-17 03:51:33 <elly> I only have the double gzipped one
2023-02-17 03:51:37 <elly> which package owns the single gz one?
2023-02-17 03:52:36 <elly> heh, man-pages APKBUILD does this:
2023-02-17 03:52:37 <elly>         find "$pkgdir"/usr/share/man -type f -exec gzip -q -9 {} \+
2023-02-17 03:52:41 <elly> that seems a little ambitious
2023-02-17 03:54:56 <psykose> it was
2023-02-17 03:54:58 <psykose> :p
2023-02-17 03:55:23 <psykose> if anything it's silly it doesn't work
2023-02-17 03:55:25 <psykose> gr
2023-02-17 03:56:26 <elly> hmmm, I don't quite see how the files are getting gzipped twice
2023-02-17 03:56:34 <elly> the makefile *can* make gzipped files
2023-02-17 03:56:42 <psykose> i assume the output of exec gets find'd
2023-02-17 03:58:13 <elly> oh, that makes sense
2023-02-17 03:58:24 <elly> unfortunate
2023-02-17 03:58:27 <psykose> also don't think it can
2023-02-17 03:58:33 <psykose> i can only see outputs for a whole dist
2023-02-17 03:58:43 <elly> yeah, it can make a .tar.gz
2023-02-17 03:58:45 <elly> I confused myself
2023-02-17 03:58:56 <psykose> anyway i just typed in the names lazily and it works
2023-02-17 03:58:59 <psykose> not a fan of this however
2023-02-17 04:00:11 <elly> I think you can just tell man to ignore files with .gz in their name already?
2023-02-17 04:00:14 <elly> sort of horrible though
2023-02-17 04:00:41 <psykose> you mean find? yeah that seems a bit better
2023-02-17 04:00:47 <psykose> ! -name "*.gz"
2023-02-17 04:01:16 <psykose> it actually does work
2023-02-17 04:01:56 <elly> er yeah find
2023-02-17 04:02:05 <psykose> thanks for the suggestiong
2023-02-17 04:02:09 <psykose> implemented
2023-02-17 04:02:15 <elly> I'm glad it works for you, it doesn't seem to work for me
2023-02-17 04:02:18 <elly> maybe I'm holding it right
2023-02-17 04:02:35 <psykose> 66c2872e8c295d4042604d3baa5ac14d577f1bfc
2023-02-17 04:02:37 <elly> or not holding it right rather
2023-02-17 04:03:02 <psykose> can they release new cgit already
2023-02-17 04:03:05 <psykose> the new one has a dark theme
2023-02-17 04:03:45 <elly> oh I am indeed not holding it right
2023-02-17 04:03:48 <abby> looks pretty dark to me http://0x0.st/Hrdg.png
2023-02-17 04:04:03 <psykose> huu
2023-02-17 04:04:09 <elly> thanks for fixing it psykose :)
2023-02-17 04:04:13 <psykose> why do i get the crusty light theme
2023-02-17 04:04:14 <abby> https://darkreader.org
2023-02-17 04:04:17 <psykose> smh
2023-02-17 04:04:19 <psykose> that does not count
2023-02-17 04:04:37 <abby> :3
2023-02-17 04:04:44 <psykose> don't :3 me you smelly
2023-02-17 04:07:03 <elly> fmemopen or fopencookie be in a C standard challenge [impossible]
2023-02-17 04:10:05 <user2> t
2023-02-17 04:10:06 <user2> t
2023-02-17 04:10:07 <user2> t
2023-02-17 04:10:10 <user2> test
2023-02-17 04:16:02 <psykose> impossible indeed
2023-02-17 04:17:26 <elly> although really, it would be much nicer if opendir/readdir/closedir were...
2023-02-17 04:18:02 <psykose> just use them anyway
2023-02-17 04:18:02 <psykose> ez
2023-02-17 04:19:37 <elly> I guess :\
2023-02-17 04:19:43 <elly> C99 programmers defeated
2023-02-17 05:10:57 <user2> test
2023-02-17 05:32:34 <psykose> blech
2023-02-17 05:32:58 <psykose> losing my mind at why this abseil spaghetti doesn't actually install scoped_mock_log.so
2023-02-17 05:33:36 <valerius> you forgot to pour on the olive oil...
2023-02-17 05:34:03 <psykose> so true
2023-02-17 05:35:30 <psykose> ah
2023-02-17 05:37:47 <psykose> marked as testonly
2023-02-17 05:38:09 <elly> psykose: what are you trying to figure out?
2023-02-17 05:38:20 <psykose> and then a different google project just yolos it and gets it from a vendored install so every target is there by magic
2023-02-17 05:38:36 <psykose> duct tape applied successfully
2023-02-17 05:39:13 <psykose> elly: new protobuf uses abseil for logging/stuff and links to scoped_mock_log for tests :p
2023-02-17 05:39:23 <psykose> abseil doesn't contain that however, it's TESTONLY and never installed
2023-02-17 05:40:07 <psykose> i guess it kinda makes sense
2023-02-17 05:47:32 <psykose> next funny part: all the tests pass, and the test driver itself segfaults on exit in some atexit handler
2023-02-17 05:47:38 <elly> ah I see
2023-02-17 05:50:50 <psykose> https://img.ayaya.dev/kM991ntVbhS2 sadly finding where to look sounds hard
2023-02-17 05:54:13 <elly> is absl calling free from an atexit handler? but why
2023-02-17 05:54:30 <psykose> the tests are protobuf
2023-02-17 05:54:48 <psykose> but yes, don't know who sets one where
2023-02-17 05:55:27 <psykose> hmm
2023-02-17 05:55:32 <psykose> if i just change the gtest version
2023-02-17 05:55:33 <psykose> it works
2023-02-17 11:18:19 <donoban> due an emergency I had to suspend my laptop in the middle of apk upgrade, should it recover from here? https://tpaste.us/b1gB
2023-02-17 11:20:51 <psykose> get a quick strace to see what it's stuck on :D
2023-02-17 11:20:54 <psykose> idk, probably not
2023-02-17 11:21:02 <psykose> but if you abort that and run an upgrade again it will be fine
2023-02-17 11:21:20 <psykose> if you get an strace and see the stuck call you can probably open an issue for it in apk
2023-02-17 11:21:50 <donoban> uhM, do you mean a strace of the running process?
2023-02-17 11:22:04 <donoban> or open a newer one and repeat the suspend/resume?
2023-02-17 11:22:59 <donoban> I suppose that you mean   -p PID, --attach=PID
2023-02-17 11:23:26 <donoban> strace: Process 13829 attached
2023-02-17 11:23:28 <donoban> read(7, 
2023-02-17 11:25:30 <donoban> I tried to stop and resume it, but it stills stuck
2023-02-17 11:26:02 <donoban> https://tpaste.us/7MN1
2023-02-17 11:26:20 <Newbyte> donoban: shot in the dark, but maybe try to `rm /var/cache/apk/*`
2023-02-17 11:27:29 <donoban> it didn't affect
2023-02-17 11:32:56 <psykose> ah ok
2023-02-17 11:33:03 <psykose> yeah forgot you get zero history haha
2023-02-17 11:33:24 <Newbyte> get zero history?
2023-02-17 11:35:06 <donoban> I suppose that he refers to strace
2023-02-17 13:15:41 <donoban> uhhmm wops, s/he/she
2023-02-17 15:48:24 <WhyNotHugo> Is there a file where I can APPEND environment variables on top of the ones in /etc/abuild.conf? E.g.: something like ~/.config/abuild.conf ?
2023-02-17 15:48:42 <WhyNotHugo> (abuild and abuild.conf are really missing a man page)
2023-02-17 15:51:25 <psykose> ~/.abuild/abuild.conf
2023-02-17 15:51:29 <psykose> ah
2023-02-17 15:51:39 <psykose> yeah, it appents
2023-02-17 15:51:42 <psykose> appends*
2023-02-17 15:51:46 <psykose> just checked
2023-02-17 16:00:29 <WhyNotHugo> No ~/.config ? :c
2023-02-17 16:00:32 <WhyNotHugo> That works tho.
2023-02-17 16:00:38 <WhyNotHugo> rootbld is very hady too :)
2023-02-17 16:09:01 <psykose> https://gitlab.alpinelinux.org/alpine/abuild/-/issues/10090
2023-02-17 16:09:15 <psykose> i use rootbld for almost everything, yeah it works pretty good
2023-02-17 16:18:12 <lavaball> a 128mb ram system. better to not go diskless, right?
2023-02-17 16:18:30 <lavaball> just as router. wifi to ethernet. + firewall.
2023-02-17 16:18:31 <lavaball> nothing more.
2023-02-17 16:22:02 <psykose> idk if you can even install on 128
2023-02-17 16:22:16 <psykose> that sounds more like some openwrt stuff if it's a router
2023-02-17 16:25:27 <lavaball> why wouldn't i be able to install it there? even openbsd runs on it. 
2023-02-17 16:26:18 <lavaball> oh, it says 256 to install.
2023-02-17 16:26:22 <lavaball> but 128 to run?
2023-02-17 16:26:28 <lavaball> how does that work?
2023-02-17 16:30:36 <psykose> the iso itself is unpacked into ram
2023-02-17 16:30:50 <psykose> i don't think you will be able to finish installing on 128 :D
2023-02-17 16:30:55 <psykose> the booting part after sure, i guess, but
2023-02-17 16:31:21 <psykose> there's some workarounds i guess but
2023-02-17 16:32:30 <elly> I bet you can boot in less than 128 even
2023-02-17 16:34:22 <lavaball> i'll just install it via chroot then.
2023-02-17 16:56:59 <minimal> elly: I just took a VM I had handy (which defaults to 512Mb ram), changed it to 128Mb and it booted fine. I tried 64Mb and some error briefly appeared on screen (not sure if Grub or kernel) before it continued a bit further and stopped. This obviously is using linux-virt which is smaller than linux-lts
2023-02-17 16:57:18 <minimal> "htop" showed 46.7M used out of 101M
2023-02-17 16:58:08 <elly> 64 still seems like a lot, I wonder what the issue is
2023-02-17 16:58:17 <elly> the kernel's like, 8 or 9 that has to be resident
2023-02-17 16:58:21 <elly> then the initramfs is another... 15?
2023-02-17 16:59:30 <minimal> lavaball: chroot using apk.static is the approach I take
2023-02-17 17:00:32 <minimal> elly: not sure, I'll investigate more. I've gotten Sys-mode images down to 50-60Mb *disk space* easily for VMs by tweaking initramfs contents, etc
2023-02-17 17:01:23 <elly> yeah
2023-02-17 17:01:27 <elly> I'm kind of curious
2023-02-17 17:03:06 <minimal> checked it again, its after the initramfs is loaded. Console stops with "[    1.249290] Out of memory and no killable processes..." and then a kernel panic
2023-02-17 17:03:25 <minimal> I'll have to try via serial console to be able to scroll back to see the full output
2023-02-17 17:03:54 <elly> I wonder if it just has like, the compressed initramfs, the uncompressed initramfs, and it runs out of space
2023-02-17 17:04:44 <psykose> prolly
2023-02-17 17:05:03 <minimal> yeah, that would also depend on the compression algo used (gzip, zstd, etc). Tried 96Mb and it booted fine
2023-02-17 17:05:25 <elly> this is kind of a crazy idea but what if you *don't* compress the initramfs
2023-02-17 17:05:34 <elly> since you're gonna use the uncompressed amount of space regardless
2023-02-17 17:06:30 <lavaball> thanks, minimal. 
2023-02-17 17:07:47 <elly> if you did no modules at all + no initramfs at all, just run straight off disk, I bet you could get it into 16MB
2023-02-17 17:07:54 <minimal> true, worth an investigation. First though I should try one of my really slimmed down VM images first (this one has VirtualBox agent, rsyslog rather than Busybox, chrony rather than Busybox, etc)
2023-02-17 17:07:59 <psykose> hmm
2023-02-17 17:08:08 <psykose> are you sure the kernel keeps compressed in memory at the same time
2023-02-17 17:08:16 <psykose> it decompresses it for you to begin with
2023-02-17 17:08:34 <elly> I am not, but it wouldn't shock me if it did, at some point, have the whole compressed image and the whole uncompressed image in memory simultaneously
2023-02-17 17:08:41 <minimal> elly: yeah, I'm already working on that for Firecracker, building a firecracker/qemu-microvm kernel with no initramfs, virtio drivers compiled in, etc
2023-02-17 17:08:58 <elly> having it unload or replace compressed data after it's done with it would be more complicated and I'm not sure they would have it work that way
2023-02-17 17:09:13 <psykose> sounds like it warrants a check
2023-02-17 17:09:13 <elly> (incrementally as it uncompresses, I mean)
2023-02-17 17:09:17 <minimal> and no bootloader either, just kernel and rootfs file
2023-02-17 17:09:37 <psykose> and also sounds like it most likely already is implemented (no double) because idk, kinda nice?
2023-02-17 17:09:50 <elly> I don't even know where to look really
2023-02-17 17:10:10 <psykose> the initrd decompression code prolly
2023-02-17 17:10:26 <grawity> isn't the initramfs unpacked file-by-file into a ramfs/tmpfs whether it was compressed or not?
2023-02-17 17:10:50 <elly> hm, maybe if it's rw mounted?
2023-02-17 17:10:56 <elly> if it's ro I don't see why you would do that
2023-02-17 17:11:30 <elly> I'm guessing though
2023-02-17 17:12:23 <grawity> because the initramfs usually isn't a filesystem image, it's a cpio archive
2023-02-17 17:12:32 <grawity> so it's not mounted at all
2023-02-17 17:13:45 <elly> I forgot about cpio!
2023-02-17 17:15:05 <grawity> (I mean it used to be a fs image but the 'ramdisk' part of 'initrd' is mostly a relic from pre-2.6, I guess you can still do that but I don't think any distro does it that way...)
2023-02-17 17:15:28 <grawity> so I basically assume the kernel just unpacks it as a stream, like `tar x[z]f initrd`, even if it's compressed
2023-02-17 17:18:06 <minimal> is you do "dmesg | grep init" you'll see kernel message such as: "Unpacking initramfs..." and "Freeing initrd memory: 5272K" before the initramfs's init is run
2023-02-17 17:18:41 <minimal> so there's approx 5Mb being used for initramfs already
2023-02-17 17:19:22 <minimal> that's for a initramfs-virt file that "ls -lh" says is 5.2M in size
2023-02-17 17:19:38 <psykose> hmmm
2023-02-17 17:19:43 <psykose> that sounds suspiciously similar
2023-02-17 17:19:44 <psykose> :D
2023-02-17 17:20:11 <minimal> and I didn't change mkinitfs setting and so that file is gzip compressed
2023-02-17 17:20:38 <minimal> looks like it's about 10Mb uncompressed
2023-02-17 17:26:57 <xMass> Hello, anyone here currently uses the in.tftpd on the lastest alpine? i've tried to use on mine directly from APK but i'm having some trouble 
2023-02-17 17:46:39 <lavaball> i think i have 256mb ram after all. i'm just having a hard time identifying these old pc engine boards.
2023-02-17 17:52:24 <minimal> lavaball: ALIX?
2023-02-17 17:53:34 <lavaball> alix.2d
2023-02-17 17:53:56 <lavaball> one has 1350 above the alix.2d line on the board, the other has 1117.
2023-02-17 17:54:02 <lavaball> no idea if that means anything.
2023-02-17 17:54:12 <lavaball> i just konw 2d19 is bad. as in 128.
2023-02-17 17:55:17 <minimal> https://www.pcengines.ch/alix2d13.htm ?
2023-02-17 17:55:45 <minimal> that mentions 2d13 and 2d19
2023-02-17 17:57:38 <lavaball> but not how to distinquish between them.
2023-02-17 17:57:42 <lavaball> thanks for checking though.
2023-02-17 17:57:47 <lavaball> i think one is a 13.
2023-02-17 17:58:20 <lavaball> actually, both might be. if i read this correctly 2d19 doesn't even have usb.
2023-02-17 17:58:23 <lavaball> mine do.
2023-02-17 18:01:18 <grawity> my first time using alpine was booting it on an old HP server we found in the attic... 256 MB system, I think? iirc it was 2 years ago, it booted fine and had well over half the memory free
2023-02-17 18:02:14 <lavaball> grawity, you are here as well?
2023-02-17 18:02:33 <lavaball> are you the same guys who answers my dumb questions over at arch, netfilter, bash ... you name it.
2023-02-17 18:02:45 <grawity> no, I'm just loud enough that you're hearing my voice over from #arch
2023-02-17 18:03:45 <lavaball> i'm so glad i left them. like i had a hardened kernel panic at that was the final straw. i wanted to get away from systemd anyway. and then i finally installed alpine on bare metal. it was bitch to get it running on the bananapi, but now. wow. i loaf it.
2023-02-17 18:04:04 <psykose> don't think kernel panics are very indicative of the distro
2023-02-17 18:04:29 <lavaball> i only had them with arch hardened. and before that arch grsec.
2023-02-17 18:04:48 <lavaball> not that i mind. it just happens, but you never have a reason to switch distros since eveyrone runs fine forever usually.
2023-02-17 18:04:56 <psykose> sure, downstream kernels are their responsibility :D
2023-02-17 18:04:56 <abby> sounds more like an issue with the patches then
2023-02-17 18:05:35 <lavaball> i don't want to throw shade over there either. i still have it on my bigger box to play games. and as i said, it jsut runs ...
2023-02-17 18:05:45 <lavaball> also firejail is working fine. unlike over at void.
2023-02-17 18:06:04 <lavaball> luckily though i switched my main boxes to openbsd. the only sane desktop system. can't recommend it for servers or anything else for that matter.
2023-02-17 18:06:06 <psykose> you hear that abby
2023-02-17 18:06:14 <psykose> $software broken in the $voidenlinux
2023-02-17 18:06:18 <psykose> you better fix it doublefast
2023-02-17 18:06:24 <lavaball> oh, i told them. they didn't care. 
2023-02-17 18:06:37 <lavaball> said it has root exploits anyway.
2023-02-17 18:06:49 <lavaball> but then i was talking to one of the woke people. they are weird anyway.
2023-02-17 18:06:49 <abby> where did you tell us?
2023-02-17 18:07:09 <lavaball> there.
2023-02-17 18:07:45 <lavaball> though again i'm not complaining. openbsd is way better for desktops, and i don't need firejail on my bigger servers where i greatly enjoy this void.
2023-02-17 18:08:12 <psykose> everythin' be goin' all backwards-like suddenly
2023-02-17 18:08:16 <psykose> y'all, which way is up?
2023-02-17 18:08:21 <abby> down
2023-02-17 18:08:25 <psykose> fuck!
2023-02-17 18:08:45 <lavaball> depends on your current position and orientation and is relative to your point of reference.
2023-02-17 18:09:59 <abby> guess void is just too woke for firejail to work properly
2023-02-17 18:10:16 <lavaball> i don't think void is woke. how can a thing ... okay, media, i give you that.
2023-02-17 18:10:52 <lavaball> but like i said, when i brought it up i was told that it has privilege escalation exploits every other day basically.
2023-02-17 18:10:59 <lavaball> i mean it's a hack above a hack. what do you expecT?
2023-02-17 18:11:21 <lavaball> but so is all of this linux stuff. and don't get me started on openbsd and their "iBlock" - "this isn't the apple store".
2023-02-17 18:11:50 <lavaball> anyway, maybe firejail works now. firefox crashed when you stared it with firejail after a while.
2023-02-17 18:14:05 <psykose> you said woke people :p
2023-02-17 18:14:31 <ikke> We rejected firejail iirc
2023-02-17 18:14:33 <lavaball> is that funny here?
2023-02-17 18:14:40 <lavaball> maybe i'll buy land. 
2023-02-17 18:15:45 <donoban> lavaball: if you like firejail you could try bubblejail
2023-02-17 18:16:01 <lavaball> i'm beyond that. didn't you hear me mention openbsd?
2023-02-17 18:16:15 <lavaball> they have all that built in. 
2023-02-17 18:16:27 <donoban> ah great
2023-02-17 18:16:41 <lavaball> i do appreciate you trying to help out.
2023-02-17 18:16:52 <donoban> :)
2023-02-17 18:16:55 <lavaball> or was that a joke?
2023-02-17 18:16:59 <lavaball> you were pulling my leg, right?
2023-02-17 18:17:08 <donoban> no, in fact I maintain it for alpine
2023-02-17 18:17:12 <lavaball> ohhh
2023-02-17 18:17:22 <lavaball> maybe i will take a look at it, my good sir.
2023-02-17 18:17:31 <donoban> well, if psykose let someone to maintaint something :D
2023-02-17 18:18:08 <lavaball> though right now i don't really have a use case. i don't know alpine well enough to use it on my bigger boxes yet. also the users are retarded, so i have to have glibc and openssl available anyway.
2023-02-17 18:18:51 <lavaball> then i cna use lxc and don't need any restriction software inside the container i mean.
2023-02-17 18:20:44 <lavaball> oh, since everybody is listening. i do appreciate all you work. this alpine stuff is really nice and small and cool. except for "reboot" and "poweroff", there i prefer void and its init 6 and 0.
2023-02-17 18:21:49 <elly> why?
2023-02-17 18:22:05 <lavaball> i don't know. i'm weird.
2023-02-17 18:22:12 <ikke> make a shell script wrapper :P
2023-02-17 18:22:16 <lavaball> i stared with arch when it still had the init stuff.
2023-02-17 18:22:22 <lavaball> so it reminds of the good times!
2023-02-17 18:22:31 <lavaball> and no. alias or script is not the same.
2023-02-17 18:22:34 <lavaball> i like default.
2023-02-17 18:22:53 <lavaball> so i rather use the default i don't like than some script catered to my needs. because on the next box i might not have my script.
2023-02-17 18:23:11 <ikke> Then start getting reboot and poweroff in your muscle memory ;-)
2023-02-17 18:23:52 <lavaball> already have. arch switched to systemd years ago. and i still have even one local box. also ... let me think ... two servers and a virtual machine. i'm not gonna get away from it.
2023-02-17 18:26:27 <elly> make an alias? it is very straightforward
2023-02-17 18:26:41 <lavaball>  7:22:31 PM <lavaball> and no. alias or script is not the same.
2023-02-17 18:27:00 <psykose> just make an alias? it is very straightforward
2023-02-17 18:27:36 <lavaball> you are funny guy, aren't you?
2023-02-17 18:27:42 <ikke> gal
2023-02-17 18:28:01 <lavaball> psykose, is a female companion person?
2023-02-17 18:28:27 <lavaball> oh, there shouldn't be komma there. well, i'm sure you konw what i meant.
2023-02-17 18:28:43 <elly> it is seriously strange, to me, to be like "I dislike this superficial detail of the system but I do not want to use either of the straightforward ways to adapt the system to what I want"
2023-02-17 18:28:48 <abby> are you ferengi
2023-02-17 18:28:49 <lavaball> not that is matters. every girl on the net is a guy, and every child is the fbi. 
2023-02-17 18:29:01 <elly> don't be gross
2023-02-17 18:29:11 <elly> there are more than a few women here
2023-02-17 18:29:13 <lavaball> those are the old rules i was brought up with.
2023-02-17 18:29:24 <lavaball> well, i hope they like jokes better than you.
2023-02-17 18:29:25 <elly> we all have a moral obligation to become better people than we were raised to be
2023-02-17 18:29:53 <lavaball> you wanna go down the morals road/
2023-02-17 18:29:54 <lavaball> ?
2023-02-17 18:29:57 <psykose> ferengi rule of acquisition 112: Never have sex with the boss' sister.
2023-02-17 18:30:07 <elly> "you don't belong in this space" is not actually that funny of a "joke" when you hear it for the hundredth time :P
2023-02-17 18:30:10 <lavaball> i prefer you can't free a fish from water.
2023-02-17 18:30:14 <lavaball> i do'nt have the numbers in my head though.
2023-02-17 18:30:48 <lavaball> "you don't belong?" careful. i have been trained by the best. all that injecting statements into what i said doesn't work with me anymore.
2023-02-17 18:31:04 <elly> I see
2023-02-17 18:31:13 <elly> thanks psykose, that was weird
2023-02-17 18:31:18 <ikke> yeah
2023-02-17 18:31:29 <abby> yeah ^
2023-02-17 18:31:36 <psykose> look i'm a fan of the ole schizoposting so i give em a chance
2023-02-17 18:31:40 <psykose> that is bad schizoposting however
2023-02-17 18:32:25 <psykose> for good schizoposts plz like subscribe my youtube chanel
2023-02-17 18:32:34 <ikke> And hit the bell icon
2023-02-17 18:35:00 <Misthios> and dont forget to leave a comment
2023-02-17 18:43:53 <xMass> Hello, anyone here currently uses the in.tftpd on the lastest alpine? i've tried to use on mine directly from APK but i'm having some trouble
2023-02-17 18:44:16 <psykose> what issue do you have
2023-02-17 18:48:37 <xMass> I'm seding to the alpine server a file from a device, it creates the file in the directory but does not write it, i give permission to the tftp user and the permission on the file is correct, i get the WRQ from the device, the file is created, but no data is written on it 
2023-02-17 19:02:54 <psykose> even after sync? or
2023-02-17 19:02:58 <psykose> how do you know it's created
2023-02-17 19:14:15 <xMass> If i delete the file, and send the command on the device to make the upload, it creates the file
2023-02-17 19:14:28 <psykose> and it's empty
2023-02-17 19:14:28 <psykose> hah
2023-02-17 19:14:32 <psykose> interesting
2023-02-17 19:14:37 <xMass> Yes
2023-02-17 19:14:46 <xMass> The only thing i get on log is this : daemon.notice in.tftpd[2853]: WRQ from 172.31.245.6 filename teste.cfg
2023-02-17 19:14:52 <xMass> Then the upload fails on the device 
2023-02-17 19:15:05 <psykose> do you see anything else or just.. that's that
2023-02-17 19:16:09 <xMass> Just that 
2023-02-17 19:17:26 <xMass> The error from the device is the following: TFTP error string is tftp transfer failed: error 0x4b0007, i've searched and it's a timeout error
2023-02-17 19:17:48 <xMass> Wich makes not that much sense for me
2023-02-17 19:30:15 <minimal> xMass: perhaps try editing /etc/conf.d/in.tftpd.conf and add "-c" to the INTFTP_OPTS setting?
2023-02-17 19:30:50 <xMass> I already have it on
2023-02-17 19:31:32 <xMass> my current config: INTFTPD_OPTS="-s -c -p -v -u backup "/home/backup/""
2023-02-17 19:31:43 <minimal> ok, you didn't mention that ;-)
2023-02-17 19:33:03 <minimal> have you also tried adding "-v" multiple times to increase verboseness?
2023-02-17 19:34:35 <minimal> also "-B 1468", I haven't used TFTPd in a while but think I had to use that in the past
2023-02-17 19:34:38 <psykose> wonder if there are any other tftp servers
2023-02-17 19:34:47 <minimal> psykose: dnsmasq ;-)
2023-02-17 19:34:50 <psykose> haha
2023-02-17 19:35:01 <psykose> well that one is certainly less abandoned
2023-02-17 19:35:10 <minimal> yes, I used ot use dnsmasq as a combined dns/dhcp/tftp server for SIP desk phones :-)
2023-02-17 19:35:45 <ikke> I've used it as a tftp server as well
2023-02-17 19:36:48 <minimal> last few times I've used a tftp server has only to serve up ipxe program to then switch to HTTP/HTTPS :-)
2023-02-17 19:37:14 <xMass> Yes, i've tried to increase the verbosity, still no other message popsup 
2023-02-17 19:37:22 <ikke> I used it for updating firmware of access points
2023-02-17 19:37:24 <minimal> xMass: and "-B"?
2023-02-17 19:37:38 <xMass> Going to try it, just a minute 
2023-02-17 19:37:52 <minimal> some tftp clients don't negotiate/handle blocksize correctly
2023-02-17 19:39:56 <xMass> Still has the same comportament 
2023-02-17 19:40:56 <xMass> I'm only using it because the device does not support any other transfer method, i have a tftpd running in a CentOS vm that runs with the same arguments D: 
2023-02-17 19:41:04 <xMass> And it works
2023-02-17 19:41:09 <xMass> Kinda strange
2023-02-17 21:04:42 <charmango> Trying to find php81-mcrypt and php81-xmlrpc on 3.17 but they don't seem to exist.
2023-02-17 21:06:34 <psykose> php81-pecl-mcrypt php81-pecl-xmlrpc are the names, but they're in testing so edge only
2023-02-17 21:08:22 <charmango> ok thank you
2023-02-17 21:44:57 <boombim[m]> Hello everyone.
2023-02-17 21:44:57 <boombim[m]> May zfs be the one of possible fs for alpine in installation script?
2023-02-17 23:57:46 <mbx_> is there a difference between the EFI boot partition and the /boot directory on the file system?
2023-02-17 23:58:26 <mbx_> iirc, the EFI boot partition is the one that needs to be mounted into /boot, or are these two different things?
2023-02-17 23:59:55 <psykose> esp is a fat32 partition with a certain layout
2023-02-18 00:00:08 <psykose> boot directory is.. boot directory
2023-02-18 00:00:16 <psykose> maybe you just have the esp mounted on /boot and use it as one too
2023-02-18 00:00:19 <psykose> maybe it's in /boot/efi
2023-02-18 00:00:57 <abby> you might be conflating "efi system partition" and "/boot partition"
2023-02-18 00:01:09 <mbx_> ok right, there was the confusion. I need to mount the EFI system to /boot/efi after mounting my boot partition to /boot
2023-02-18 00:02:02 <mbx_> still tackling the luks2 challenge here :) I try to boot from grub, this time I made a seperate boot partition :))
2023-02-18 00:30:09 <mbx_> the docs confuse me about mkinitfs. It says that I need to edit stuff in /etc of my file system to help it to boot from grub, but how can grub read these configurations, when these configurations are encrypted. I would assume that I have to edit something in /boot (which is unencrypted), so that grub knows it needs to luksOpen and mount another system
2023-02-18 00:31:07 <mbx_> oh maybe mkinitfs will take care of that, but will it also work this way for grub2? The instructions state this is for extlinux
2023-02-18 00:32:26 <psykose> don't think either grub2 or extlinux will work with luks2
2023-02-18 00:32:34 <psykose> but yes the grub "compiled" a config into /boot
2023-02-18 00:32:49 <mbx_> you are right, with full encryption it will not
2023-02-18 00:32:55 <mbx_> but my /boot is not encrypted
2023-02-18 00:34:09 <mbx_> I think I just have to setup my initfs to luksOpen and mount properly, but I don't know how, yet
2023-02-18 00:45:32 <mbx_> mkinitfs fails cuz it misses the module /mnt/rootfs/lib/modules/5.15.93-0-lts. What package do I need to install and how do I install it on a different fs?
2023-02-18 00:46:18 <psykose> that doesn't sound correct
2023-02-18 00:46:38 <psykose> sounds more like you upgraded the kernel but didn't reboot
2023-02-18 00:47:36 <mbx_> I installed alpine to a mounted device via setup-disk ...
2023-02-18 00:48:06 <mbx_> and now I do mkinitfs -c /mnt/rootfs/etc/mkinitfs/mkinitfs.conf -b /mnt/rootfs
2023-02-18 00:48:28 <mbx_> so it updates the contents of /boot hopefully
2023-02-18 00:49:26 <mbx_> setup-disk also printed "=> initramfs: creating /boot/initramfs-lts", could it be that I don't have to do this step anymore?
2023-02-18 01:07:08 <mbx_> I just tested if I could boot, grub loads, but mounting UUID=XXXX... on /sysroot failed: No such file or directory, Mounting root failed.
2023-02-18 01:43:02 <mbx_> psykose: I got initramfs to work
2023-02-18 01:43:29 <mbx_> ls /mnt/rootfs/lib/modules would reveal that a wrong Kernel version is installed
2023-02-18 01:43:35 <mbx_> I had to symlink to fix it
2023-02-18 01:43:55 <mbx_> apparently, the installed initramfs depends on the latest kernel, but I installed stable
2023-02-18 01:46:01 <mbx_> bit.ly/3YFiRCN <- this dude reported the same problem a while ago on the Alpine Forums. He can use extlinux, I can't. I will see if this will work for me in the end.
2023-02-18 01:49:09 <minimal> mbx_: /etc/default/grub is where you change change Grub config settings, then run grub-mkconfig which uses those to create /boot/grub/grub.cfg
2023-02-18 07:05:26 <xaotong> f
2023-02-18 07:11:05 <xaotong> f
2023-02-18 09:26:29 <frag> anyone else have  alot of trouble with yt-dlp/mpv and youtube links recently?
2023-02-18 09:30:11 <abby> i think something changed on yt's end
2023-02-18 09:30:17 <abby> latest yt-dlp works fine though
2023-02-18 09:31:57 <frag> abby: latest yt-dlp in stable repo?
2023-02-18 09:32:07 <abby> latest upstream version
2023-02-18 09:32:40 <abby> which looks to be in edge
2023-02-18 09:39:29 <mercenary> youtube changed something in the last couple of days, had to update yt-dlp
2023-02-18 10:07:35 <ikke> mercenary: that's not unexpected
2023-02-18 10:10:47 <lopid> frag: https://github.com/yt-dlp/yt-dlp/issues/6247#issuecomment-1433096554
2023-02-18 10:26:18 <frag> ill just wait for the repo update i guess :]
2023-02-18 10:32:33 <lopid> easy enough to modify that file locally
2023-02-18 10:32:44 <lopid> you don't have to recompile anything
2023-02-18 10:33:08 <ikke> It's a single line change
2023-02-18 10:33:51 <frag> oh rly, hmm
2023-02-18 10:34:16 <frag> pyhton mhmhm
2023-02-18 10:35:42 <frag> curious, would it take like 2-5 min for the package maintainer to fix and get it into repo?
2023-02-18 10:36:21 <lopid> there may be other changes in the new release
2023-02-18 10:36:42 <ikke> it would take ~15m for it to appear in the 1rst tier repositories
2023-02-18 10:37:00 <ikke> other mirror may take longer
2023-02-18 10:47:24 <frag> okok :]
2023-02-18 15:28:53 <purple_boost> hello! i have a question about seamless switching between wired and wireless.. currently when i switch the default gateway is not changed so my connection stops working until i change it to the correct interface. do i need to configure ifupdown scripts to have this done automatically? i would also like the wired connections to be preffered over wireless when both are connected. am i on the 
2023-02-18 15:28:59 <purple_boost> right track with ifupdown or should i use a different method for alpine?
2023-02-18 15:29:41 <JohannesJ[m]> PostMarketOS is THE goto place if you want alpine on the phone, yes?
2023-02-18 15:29:49 <JohannesJ[m]> Or am i missing something else?
2023-02-18 15:29:52 <ikke> JohannesJ[m]: yes 
2023-02-18 15:30:48 <JohannesJ[m]> Goodie
2023-02-18 15:31:05 <ikke> purple_boost: If you have both wireless and wired connected, you should have a default route from both
2023-02-18 15:31:13 <ikke> then the metric should determine which is preferred
2023-02-18 15:31:35 <ikke> and if one drops, traffic should automatically go over the other route (not sure what you mean with changing it to the correct interface)
2023-02-18 15:32:35 <purple_boost> oh right! both have metric 0 now... what i meant is that when i unplug the enternet cable nothing works until i change the gw to use wlan0.. i guess i need to fix the metric first! thanks!
2023-02-18 15:33:40 <purple_boost> i had also looked into connman but it seems not needed when using iwd.. but maybe i misunderstood
2023-02-18 15:34:13 <ikke> purple_boost: you can have more then one default gateway
2023-02-18 15:34:25 <ikke> so you need to make sure you have one for each interface
2023-02-18 15:38:06 <purple_boost> hmm, just checked the routing table before and after unplugging ehternet and it doesn't remove the gw on eth0.. i guess thats the crux? if some helper/script can add/remove the gw when wired/wifi connects and disconnects it would be seamless? or am i makign it too complicated? 
2023-02-18 15:44:01 <ikke> Hmm, the kernel should automatically retract routes from interfaces that are down
2023-02-18 15:49:26 <ikke> purple_boost: what does cat /proc/sys/net/ipv4/conf/eth0/ignore_routes_with_linkdown return?
2023-02-18 15:52:22 <purple_boost> 0
2023-02-18 15:52:40 <purple_boost> i also noticed when both interfaces are up i can't ssh to the wifi ip, only wired
2023-02-18 15:52:59 <ikke> purple_boost: can you try setting it to 1?
2023-02-18 15:53:04 <purple_boost> of course!
2023-02-18 15:57:04 <purple_boost> with it set to 1 as soon as the cable is removed the routing and ip info is now removed!
2023-02-18 15:57:24 <ikke> good
2023-02-18 15:57:41 <ikke> You could add it to /etc/sysctl.d/*
2023-02-18 15:57:56 <ikke> net.ipv4.conf.eth0.ignore_routes_with_linkdown=1
2023-02-18 15:58:01 <ikke> same for wlan0 
2023-02-18 15:58:19 <purple_boost> just added those to sysctl indeed, thank you!
2023-02-18 16:00:50 <purple_boost> does it always default to 0 or did i mess up my installation somewhere?
2023-02-18 16:00:56 <purple_boost> just for my informaton :D
2023-02-18 16:01:25 <ikke> defaults to 0 apparently
2023-02-18 16:01:48 <ikke> https://www.kernel.org/doc/html/latest/networking/ip-sysctl.html does not mention a default
2023-02-18 16:01:52 <ikke> so wonder how that is determined
2023-02-18 16:02:10 <dwfreed> you'd have to check the source
2023-02-18 16:04:00 <ikke> https://github.com/torvalds/linux/commit/0eeb075fad736fb92620af995c47c204bbb5e829
2023-02-18 16:20:50 <purple_boost> everything now functions as expected! as always i was making it too complex.. thanks for helping me ikke! the solution was simple but i wouldn't have figured this out before midnight on my own haha
2023-02-18 16:24:58 <purple_boost> i must say that alpine as desktop os on a decade macbook pro is super snappy and only uses 150-250mb of ram running stock wayland/sway and some terminals.. freqency scaling and power management/acpi all work great.. im impressed!
2023-02-18 16:25:23 <purple_boost> perfectly usable now, some e-waste saved
2023-02-18 16:48:13 <ikke> And net even hopelessly out of date :)
2023-02-18 16:58:28 <minimal> elly: following on from yesterday chat, I booted a smaller VM with 80MB ram successfully, 64MB has kernel panic due to OOM (*before* initramfs' init is run)
2023-02-18 17:44:24 <purple_boost> ikke: thats right! macos catalina was quite slow on it despite having 16gb ram and a 860 evo ssd.. it is now perfect for managing remote servers, browsing the web and watching some youtube
2023-02-18 17:45:45 <purple_boost> feels like a new laptop and thats with full disk encryption (not sure if it adds extra load but still)
2023-02-18 17:46:46 <minimal> purple_boost: if disk encryption did feel slow you could try LUKS with Adiantum as it is more "lightweight" on CPU
2023-02-18 18:10:39 <purple_boost> minimal: oh thanks for the tip! haven't noticed any issues at all, its using aes-xts which is the fastest on this 3rd gen intel cpu according to the cryptsetup benchmark (https://0bin.net/paste/Sh3nHhN0#9VEBqy0rj2SkfeaSWNtEzsoTuisdQFlHRnfLsfQkIQf)
2023-02-18 18:12:18 <elly> minimal: wow, 64MB for just the kernel? oof
2023-02-18 18:12:23 <elly> oh well
2023-02-18 18:44:42 <minimal> purple_boost: Adiantum is aimed at low end CPUs (MIPS, Arm, etc) especially those without HW accel for AES etc, e.g. Android uses/used it
2023-02-18 18:47:53 <minimal> elly: well I think it is a combination of factors, perhaps bootloader used (this was with Grub, haven't tested Syslinux yet) as its code may still be in memory, the size of initramfs (I used a stripped down 4.4MB one, and as VM there's no firmware in there), size of kernel (I'm using linux-virt which is 7.4MB but linux-lts for physical machines will be larger), but there's also the factor of memory used by the kernel - once booted (with 
2023-02-18 18:47:53 <minimal> 80MB ram) I see the following in dmesg output:
2023-02-18 18:48:10 <minimal> [     0.547066] Unpacking initramfs...
2023-02-18 18:48:10 <minimal> [     0.658713] Freeing initrd memory: 4524K
2023-02-18 18:48:10 <minimal> [     0.018224] Memory: 44672K/81464K available (12294K kernel code, 1335K rwdata, 3032K rodata, 1764K init, 1752K bss, 36532K reserved, 0K cma-reserved)
2023-02-18 18:48:10 <minimal> [     0.716628] Freeing unused kernel image (initmem) memory: 1764K
2023-02-18 18:48:10 <minimal> [     0.719266] Write protecting the kernel read-only data: 18432k
2023-02-18 18:48:11 <minimal> [     0.721603] Freeing unused kernel image (text/rodata gap) memory: 2040K
2023-02-18 18:48:13 <minimal> [     0.723423] Freeing unused kernel image (rodata/data dap) memory: 1064K
2023-02-18 18:48:15 <minimal> [     0.725267] Run /init as init process
2023-02-18 18:55:34 <elly> "36532K reserved" is probably the culprit
2023-02-18 18:55:49 <elly> oh well
2023-02-18 21:18:50 <noname> is anyone familiar with botnets
2023-02-18 21:19:52 <noname> if so i need some help with building one if you don't mind
2023-02-18 21:22:32 <pj> 🤨
2023-02-18 21:41:33 <ikke> noname: find help somewhere else, that's not something we facilitate
2023-02-18 21:50:37 <Saijin_Naib> <purple_boost> "perfectly usable now, some e-..." <- Wonderful work! 
2023-02-18 21:51:05 <NepNepdmsalwaysopen[m]> How long does it usually take for edge packages to be tested and released on stable? Specifically talking about synapse right now.
2023-02-18 21:51:37 <NepNepdmsalwaysopen[m]> <purple_boost> "i must say that alpine as..." <- I'm using it as a server with an old athlon and 3GB of ram, works really well.
2023-02-18 21:51:56 <ikke> NepNepdmsalwaysopen[m]: No preset time and mostly depends on the maintainer
2023-02-18 21:52:35 <ikke> NepNepdmsalwaysopen[m]: from edge to stable would be just the next release
2023-02-18 21:52:52 <ikke> in other words, whatever is in edge when the release is made, will be in that release
2023-02-18 21:53:16 <ikke> NepNepdmsalwaysopen[m]: are you referring to synapse 1.77.0?
2023-02-18 21:54:21 <NepNepdmsalwaysopen[m]> ikke: Ah I see, is alpine edge similar regarding stability to other rolling release distros (for example arch linux)? If that's the case honestly the stability risk would probably be acceptable for me.
2023-02-18 21:54:32 <NepNepdmsalwaysopen[m]> ikke: Yes.
2023-02-18 21:54:50 <ikke> NepNepdmsalwaysopen[m]: maybe slightly less stable. edge would be more similar to arch testing
2023-02-18 21:55:15 <ikke> but there are plenty of people (including me) who run on edge
2023-02-18 21:55:49 <NepNepdmsalwaysopen[m]> hmm
2023-02-18 21:55:59 <ikke> Most issues happen during larger rebuilds (say, a new major python version is released)
2023-02-18 21:57:48 <NepNepdmsalwaysopen[m]> thank you for all the info ^^
2023-02-18 21:59:46 <ikke> np
2023-02-18 22:03:07 <Saijin_Naib> Can concur with ikke. Loved using edge, but had to fall back to stable because it broke an essential work tool too often 
2023-02-18 22:03:34 <Saijin_Naib> General stability was never a problem for me across three machines, though 
2023-02-18 22:04:10 <Saijin_Naib> If something bad happened, it was usually resolved in a few days tops, and I was never stuck 
2023-02-18 22:05:13 <Saijin_Naib> apk and Alpine feel resilient in a way I've never experienced before. 
2023-02-18 22:05:24 <Saijin_Naib> Worse case, mv my latest-stable apk repositories 
2023-02-18 22:06:39 <Saijin_Naib> apk fix */ (I believe it's the syntax, I have an alias for it) and done 
2023-02-18 22:44:07 <latex> Hey when will yt-dlp be updated on 3.17?
2023-02-18 22:44:39 <latex> at the moment you cannot download youtube videos with the current version of yt-dlp in the repos
2023-02-18 23:07:47 <Apachez> latex:   sudo ytp-dl -U
2023-02-18 23:09:38 <latex> ERROR: You installed yt-dlp with a package manager or setup.py; Use that to update
2023-02-19 04:46:32 <baalajimaestro> Is there any known issue on linux-lts that causes network timeouts after probably few days of uptime? I am unsure where this is from, since i haven't touched anywhere else on the entire stack
2023-02-19 05:33:15 <Saijin_Naib> Nep Nep (dms always open): this is what I have for my alias ```alias alpine-repair='sudo apk fix \* --directory-permissions --upgrade --xattr'```
2023-02-19 05:33:33 <ikke> baalajimaestro: no reports yet, nor experiencing it ourselves 
2023-02-19 06:00:34 <baalajimaestro> Hmmm, any pointers on what i can look into? I still haven't looked into dmesg, but i suspect much that it cannot be a panic, since the requests are timing out rather than being dropped. I doubt something in networking stack or the specific NIC is messing up, but i need more solid thoughts before asking the provider to replace it
2023-02-19 06:14:25 <psykose> a request timing out usually implies the host is unreachable
2023-02-19 06:14:53 <psykose> whether the kernel panicked or not is usually not related to this, but if it did they would indeed time out
2023-02-19 06:23:41 <psykose> or one of any other million things depending on how everything is set up :D
2023-02-19 07:03:33 <baalajimaestro> From what i perceive it should rather say host unreachable. But this just attempts to load forever. Even with ssh, it attempts to connect forever rather than a host unreachable instantly, if it died.
2023-02-19 07:04:49 <baalajimaestro> It is just a normal alpine on bare metal on hetzner with podman for all containers and nothing installed on host except that and ssh
2023-02-19 07:05:17 <baalajimaestro> I am just unsure if it's failing hw or failing kernel
2023-02-19 07:05:41 <abby> do you have kvm access
2023-02-19 07:07:43 <baalajimaestro> Yes i can ask them if needed to debug it
2023-02-19 11:12:16 <wyk72> hello ppl
2023-02-19 11:12:42 <wyk72> I propose to get rid of BusyBox "ip" command and make iproute2 as default
2023-02-19 11:13:25 <wyk72> issue is: the Busybox IP command does not work for most of the cool stuff like VXLANS etc, BUT it does not spew any errors if you use it
2023-02-19 11:13:26 <Apachez> as long as netstat remains
2023-02-19 11:13:46 <wyk72> This insanity makes people go crazy (like me ) :)
2023-02-19 11:13:52 <wyk72> I mean
2023-02-19 11:14:16 <psykose> most people never use said "cool stuff"
2023-02-19 11:14:24 <psykose> you need it and you install iproute2
2023-02-19 11:14:25 <wyk72> I think the days of having 128Kb of RAM are gone
2023-02-19 11:14:26 <psykose> doesn't sound very hard
2023-02-19 11:14:32 <wyk72> it's ok
2023-02-19 11:14:34 <wyk72> but please
2023-02-19 11:14:37 <wyk72> someone somebody
2023-02-19 11:14:49 <wyk72> TELL ip of BUSYBOX to SPEW AN ERROR
2023-02-19 11:14:57 <wyk72> pleaSE :)
2023-02-19 11:15:06 <quinq> There's a mailing list
2023-02-19 11:15:12 <wyk72> it's impossible to have a command that DOES NOT REPORT any error
2023-02-19 11:15:22 <wyk72> and the command simply does nothing
2023-02-19 11:15:26 <quinq> sounds reasonable
2023-02-19 11:15:27 <wyk72> I mean it's crazy
2023-02-19 11:16:13 <quinq> As psykose hinted, maybe it's a codepath that's not actually well tested/used, so maybe you're the first one annoyed enough about it to start something moving there
2023-02-19 11:16:19 <wyk72> ok, but SOMETIMES it's a little hard to realize that a command has not done what is supposed to do
2023-02-19 11:16:30 <Apachez> exit 0 would be a hint ;)
2023-02-19 11:16:31 <wyk72> and reports no error
2023-02-19 11:16:36 <quinq> I mean it sounds reasonable to return errors on errors
2023-02-19 11:16:39 <quinq> not the opposite :)
2023-02-19 11:16:45 <wyk72> example: VXLANs
2023-02-19 11:16:58 <wyk72> if you set them up with busybox "ip" command
2023-02-19 11:16:58 <Apachez> but its like complaining iptables doesnt do vxlan
2023-02-19 11:17:01 <wyk72> they get set up
2023-02-19 11:17:03 <wyk72> the interface is up
2023-02-19 11:17:03 <Apachez> iptables does iptables
2023-02-19 11:17:07 <wyk72> but the parameters ARE NOT
2023-02-19 11:17:15 <Apachez> but I agree the "ip" should be upgraded to whatever revision is needed
2023-02-19 11:17:20 <wyk72> and also
2023-02-19 11:17:22 <Apachez> and if you need iproute2 you should add it yourself
2023-02-19 11:17:32 <Apachez> because doing vxlan on the host isnt a common thing as you seem to be doing
2023-02-19 11:17:39 <wyk72> the "ip" command that is supposed to DO THE DEBUGGING does not show this
2023-02-19 11:17:45 <wyk72> ok
2023-02-19 11:17:52 <Apachez> "normally" vxlaning are handled by the switches/routers in the network and not hosts
2023-02-19 11:17:57 <psykose> i just ran some vxlan command with busybox
2023-02-19 11:18:03 <psykose> it created a vxlan in the device list
2023-02-19 11:18:15 <wyk72> but what's the point of having a command that has the same name, same syntax but does not work without telling anyone?
2023-02-19 11:18:25 <wyk72> tome it's not normal
2023-02-19 11:18:32 <Apachez> is it really the same syntax?
2023-02-19 11:18:34 <wyk72> yeah try to use it
2023-02-19 11:18:36 <wyk72> it does not work
2023-02-19 11:18:40 <ikke> wyk72: most bb commands are limited versions of their counterparts
2023-02-19 11:18:42 <quinq> Yeah, I think that the point has been made
2023-02-19 11:18:42 <psykose> even worked to set it up
2023-02-19 11:18:45 <Apachez> I have seen other things break when you opt for "cool stuff"
2023-02-19 11:18:51 <Apachez> like when netstat suddently vanishes
2023-02-19 11:18:53 <quinq> Start a discussion on busybox ML, live happy ever after
2023-02-19 11:19:03 <psykose> if you have a specific thing that doesn't fail but doesn't do anything, report it to the busybox ML
2023-02-19 11:19:09 <Apachez> and you must google on how to do "netstat -an" or "netstat -atunp" or "netstat -rn"
2023-02-19 11:19:19 <Apachez> since netstat is gone and replaced by ipaddr2
2023-02-19 11:19:23 <Apachez> iproute2
2023-02-19 11:19:41 <psykose> ip-tools is still there
2023-02-19 11:19:43 <Apachez> but having commands that behave differently isnt a new thing
2023-02-19 11:19:44 <wyk72> Apachez: so it's normal for you that a command that sets up network does not work and does not report any error, let's say works, but 50%
2023-02-19 11:19:50 <wyk72> it's plain dumb
2023-02-19 11:19:51 <Apachez> ever tried doing awk on various distributions? ;)
2023-02-19 11:19:57 <Apachez> turns out that awk isnt awk
2023-02-19 11:20:05 <wyk72> ok but
2023-02-19 11:20:12 <Apachez> there is gawk and gnu awk and awk and some fourth flavour and they all behaves differently :D
2023-02-19 11:20:13 <ikke> gawk, mawk, etc
2023-02-19 11:20:14 <Apachez> the fun...
2023-02-19 11:20:30 <wyk72> if a normal dude tried to set up a network interface, that gets set up but without any parameter
2023-02-19 11:20:33 <Apachez> so you do awk on one box and that fails miserable on another box
2023-02-19 11:20:33 <wyk72> to you this is normal ?
2023-02-19 11:20:39 <Apachez> similar with vi
2023-02-19 11:20:51 <wyk72> how much is the cost of having iproute2 as default ?
2023-02-19 11:20:52 <wyk72> 300kb ?
2023-02-19 11:20:59 <Apachez> well to me I program and verify which binary is installed
2023-02-19 11:20:59 <wyk72> 100Kb of ram ?
2023-02-19 11:21:12 <Apachez> you have zero clue if some admin have aliased that "ip" command into something else
2023-02-19 11:21:18 <wyk72> busybox is great for a router with 256Kb of ram maybe
2023-02-19 11:21:28 <Apachez> then why dont you just install iproute2 since you need that and call it a day?
2023-02-19 11:21:45 <wyk72> but it's 2023 routers have 1Gb of ram and we have a fat 8GB flash most of the time
2023-02-19 11:21:46 <paper_> do you think there are no roiters running alpine? :)
2023-02-19 11:21:48 <Apachez> also file a complain, I mean a feature request on alpine support pages that iproute should be replaced by iproute2 and why
2023-02-19 11:21:59 <paper_> if you found a bug, report it
2023-02-19 11:22:07 <psykose> it will not be replaced
2023-02-19 11:22:14 <psykose> but as i said already and everyone told you like three times
2023-02-19 11:22:17 <wyk72> it does not look reasonable to save 50kb to have a command that is not working and make people insane because it's SO MINIMAL that does not tell it
2023-02-19 11:22:20 <psykose> if you found a specific bug, report it to busybox
2023-02-19 11:22:21 <paper_> wyk72: nope, regular routers definitely don't have that much resources nowadays
2023-02-19 11:22:45 <Apachez> ALOT of stuff out there are VERY limited when it comes to ram
2023-02-19 11:22:53 <Apachez> which is why YOU have the option to add whatever package you want
2023-02-19 11:22:54 <psykose> if i could reproduce it myself i would do it, but i cannot, so it's not just "tried to make a vxlan", it's more complicated
2023-02-19 11:22:56 <Apachez> apk add iproute2
2023-02-19 11:22:57 <Apachez> done!
2023-02-19 11:23:04 <wyk72> I did add iproute2
2023-02-19 11:23:09 <wyk72> and problem solved magically
2023-02-19 11:23:15 <wyk72> ok
2023-02-19 11:23:19 <Apachez> so all the whining is ?
2023-02-19 11:23:21 <wyk72> I'll tell busybox guys
2023-02-19 11:23:26 <wyk72> if anyone hears me out
2023-02-19 11:23:30 <Apachez> again file a feature request to alpine support pages
2023-02-19 11:23:38 <psykose> there's no feature to request
2023-02-19 11:23:40 <Apachez> include a test case so others can verify your findings
2023-02-19 11:23:57 <wyk72> to AT LEAST make a WARNING on the labeli.e. "it may not work"
2023-02-19 11:24:01 <wyk72> ok
2023-02-19 11:24:02 <Apachez> saying "it doesnt work" and then psykose just tested it itself and found out "it does work" could mean that you did something wrong?
2023-02-19 11:24:05 <wyk72> test case
2023-02-19 11:24:09 <quinq> Better go the busybox way first, and then bruteforce a solution on alpine if there's no sane outcome on the busybox side
2023-02-19 11:24:21 <wyk72> VXLAN with unicast: one-liner
2023-02-19 11:24:31 <Apachez> yeah go to the source would help others using busybox aswell
2023-02-19 11:24:41 <Apachez> take your time and create a test case which shows what works and what doesnt
2023-02-19 11:24:47 <Apachez> file that as a featur request to busybox
2023-02-19 11:25:02 <Apachez> perhaps also a feature request to alpine so it can be tracked on the outcome of busybox
2023-02-19 11:25:06 <quinq> $ busybox ip one-liner
2023-02-19 11:25:06 <quinq> Usage: ip [OPTIONS] address|route|link|tunnel|neigh|rule [ARGS]
2023-02-19 11:25:10 <quinq> Doesn't seem to be working
2023-02-19 11:26:10 <Apachez> I filed some months ago a RFE regarding ntpsec which was malfunctioning and that got resolved within days
2023-02-19 11:26:21 <wyk72> I hve woked last night on the case
2023-02-19 11:26:25 <wyk72> until 4am :D
2023-02-19 11:26:30 <Apachez> then take a nap :)
2023-02-19 11:26:37 <wyk72> because it was insane to me that was not working
2023-02-19 11:26:42 <wyk72> just a sec
2023-02-19 11:26:46 <wyk72> I'll tell you
2023-02-19 11:26:50 <Apachez> then file a RFE towards busybox and one towards alpine (which links to your RFE at busybox) and get a good cup of tea :)
2023-02-19 11:27:20 <Apachez> in the meantime add "apk add iproute2" to your documentation on how to get shit going :)
2023-02-19 11:27:44 <wyk72> yeah :)
2023-02-19 11:27:46 <wyk72> I did it
2023-02-19 11:27:48 <Apachez> I have myself a list of stuff to do before I get an alpine image ready for my vm environments
2023-02-19 11:27:49 <wyk72> example:
2023-02-19 11:28:05 <Apachez> and once I have that I use that as a template for new vm's to be created
2023-02-19 11:28:38 <wyk72> this command creates a VXLAN link with unicast addresses: "ip link add name vxlan7 type vxlan id 7 dev eth0 remote 192.168.111.40 local 192.168.111.44 dstport 4789"
2023-02-19 11:28:53 <wyk72> in busybox "ip" the interface is created
2023-02-19 11:29:04 <wyk72> but the parameters ARE NOT PASSED DOWN to the kernel
2023-02-19 11:29:10 <wyk72> so basically it's not working
2023-02-19 11:29:19 <wyk72> and it's written NOWHERE
2023-02-19 11:29:52 <wyk72> iproute2 works
2023-02-19 11:29:58 <wyk72> as expected
2023-02-19 11:30:05 <quinq> ok
2023-02-19 11:30:14 <quinq> We trust you (from the start)
2023-02-19 11:30:29 <paper_> if it's a bug, then I wouldn't expect that to be documented anywhere
2023-02-19 11:30:30 <wyk72> I'll try to write the busybox guys
2023-02-19 11:30:38 <wyk72> maybe just to stick a warning would be nice
2023-02-19 11:31:39 <wyk72> same thing happens with OpenWRT
2023-02-19 11:32:01 <Apachez> wont you need some vni definitions before that vxlan link works?
2023-02-19 11:32:01 <wyk72> but at least thay say "you need tha package IPROUTE-FULL" somewhere
2023-02-19 11:32:08 <wyk72> networking is so hard :)
2023-02-19 11:32:18 <Apachez> like vlan vni xxx (for l2vpn) or vrf vni xxx (for l3vpn) ?
2023-02-19 11:32:20 <quinq> No, that's besides the point
2023-02-19 11:35:33 <wyk72> Apachez: it's a simple "unicast" vxlan, link declaration, basically a lvl3 UDP  tunnel that mocks  "standard" lvl2 interface
2023-02-19 11:38:06 <wyk72> bcause it's a complex world and a LOT of implementations still rely on lvl2 stuff (even if it's pointless), so you stretch your L2 domain endlessly (which is BAD THING imho btw)
2023-02-19 11:39:10 <quinq> lvl = layer?
2023-02-19 11:39:25 <wyk72> yeah layer2 
2023-02-19 11:39:29 <wyk72> sorry I0m tired
2023-02-19 11:46:53 <Apachez> you still need to define vni for the vxlan to work
2023-02-19 11:46:59 <Apachez> its like part of the vxlan standard :)
2023-02-19 11:47:49 <wyk72> Oh I read into the vxlan kernel module : "The management of vxlan is done in a manner similar to its two closest neighbors GRE and VLAN. Configuring VXLAN requires the version of iproute2 that matches the kernel release where VXLAN was first merged upstream."
2023-02-19 11:48:09 <wyk72> so it's a requirement, busybox IP won't dop
2023-02-19 11:48:18 <Apachez> so you mean that you didnt read that fine manual before you started? ;)
2023-02-19 11:48:55 <wyk72> I use VXLANS a lot aven in past years
2023-02-19 11:49:07 <psykose> i mean, it certainly doesn't "require" it in the general software sense
2023-02-19 11:49:15 <psykose> nothing stops someone from writing to spec anyway :p
2023-02-19 11:49:56 <Apachez> https://xkcd.com/927/  ;-)
2023-02-19 11:49:57 <alpine-meetbot> [xkcd] Standards | Alt-text: Fortunately, the charging one has been solved now that we've all standardized on mini-USB. Or is it micro-USB? Shit.
2023-02-19 11:49:57 <wyk72> I think a script that echoes parameters so /sys/net something will do probably
2023-02-19 11:50:25 <wyk72> but i think you have to go deep into the code
2023-02-19 11:50:34 <wyk72> of the module *I think*
2023-02-19 16:34:36 <mbx> is /etc/crypttab a thing in Alpine? It's the systemd-cryptsetup-generator, but systemd is not used. Is it used by grub or does openrc has something on its own?
2023-02-19 16:37:38 <minimal> mbx: you dropped off just before I could respond to you the other day
2023-02-19 16:38:23 <minimal> crypttab is used for mounting encrypted filessystems OTHER THAN the rootfs
2023-02-19 16:39:11 <mbx> oh yes, sorry, I was loosing my mind so I dropped off aha, ok thx, been reading so much stuff on the internet and so much is inconsistent
2023-02-19 16:40:00 <minimal> mbx: your previous comments were confusing, you references a problem someone else had but it didn't make sense. Are you using Grub still or did you switch to Syslinux?
2023-02-19 16:40:04 <mbx> like there is a guy who claims that editing /etc/fstab and /etc/crypttab is all that is needed to get it going, without re-generating a grub.cfg, which simply can't work, since the fs is yet to be mounted ...
2023-02-19 16:40:43 <minimal> mbx: the other day I pointed out you need to configure /etc/default/grub......but you dropped offline just as I posted that
2023-02-19 16:40:55 <mbx> I tried to switch to Extlinux, but setup-disk will always pick grub, I think because USE_EFI=1
2023-02-19 16:41:20 <mbx> my goal now is to /boot not encrypted and to use luks with argon2id
2023-02-19 16:41:25 <mbx> with grub2
2023-02-19 16:41:27 <minimal> yes, whilst Syslinux can be used with UEFI its really not properly supported, so setup-disk uses Grub for UEFI systems
2023-02-19 16:41:54 <mbx> u mean Grub for EFI systems
2023-02-19 16:41:57 <mbx> I thin
2023-02-19 16:42:13 <mbx> so let me please elaborate where I'm at
2023-02-19 16:42:17 <minimal> that's what I just said
2023-02-19 16:42:46 <minimal> so you definately want an unencrypted /boot partition? (separate from ESP partition) ?
2023-02-19 16:47:51 <invoked> technically you shouldn't need an intermediate bootloader at all, but i'll leave that as an exercise for the reader
2023-02-19 16:48:59 <mbx> as my understanding is that grub needs to mount my root system, I regenerated my grub.cfg. I put 'GRUB_CMDLINE_LINUX="cryptdevice=/dev/nvme0n1p4:crypt root=/dev/mapper/crypt"' into my /etc/default/grub after chroot'ing into my rootfs, then I called 'grub-mkconfig -O /boot/grub/grub.cfg'. It was unable to find the linux kernel, so I had to symlink to the newer Kernel version it
2023-02-19 16:49:02 <mbx> expected, so there is still a bug, but it worked and my grub.cfg got written properly, I double-checked. However, when I try to boot I'm stuck at initramfs without any prompt or error.
2023-02-19 16:49:54 <minimal> invoked: intermediate bootloader? as opposed to a bootloader? I've guessing you're referring to EFISTUB etc, but there there is no bootloader at all
2023-02-19 16:50:01 <mbx> minimal: yes, I want the benefits of argon2id. physical tampering will be avoided by third parties
2023-02-19 16:50:26 <minimal> invoked: but that may not fit in with what he is trying to achieve (if he wants kernel/iniramfs fs also encrypted)
2023-02-19 16:50:36 <mbx> it's all about my data being protected on a loss of the device or something like that
2023-02-19 16:51:30 <mbx> I don't want initramfs to be encrypted nor /boot to be encrypted, because grub2 doesn't support luks2/argon2id
2023-02-19 16:53:07 <minimal> mbx: why are you ":" separating entries in GRUB_CMDLINE_LINUX ?
2023-02-19 16:53:22 <mbx> if my understanding is correct, I neither need GRUB_ENABLE_CRYPTODISK=y nor GRUB_PRELOAD_MODULES="luks cryptodisk ...", since this is only necessary for full-encrypted disks, am I right?
2023-02-19 16:53:55 <mbx> minimal: ":" tells GRUB how to name the cryptdevice
2023-02-19 16:54:05 <minimal> "GRUB_ENABLE_CRYPTDISK=y" is only used when you have the fs containing the kernel and iniramfs encrypted - which you just said you you do not
2023-02-19 16:54:21 <mbx> thx for confirming
2023-02-19 16:54:32 <minimal> mbx: those entries are NOT ":" separated, cmdline entries are space separated
2023-02-19 16:54:52 <mbx> sec
2023-02-19 16:55:28 <minimal> also you want to configure GRUB_CMDLINE_LINUX_DEFAULT, not GRUB_CMDLINE_LINUX
2023-02-19 16:59:57 <mbx> GRUB_CMDLINE_LINUX is always effective, GRUB_CMDLINE_LINUX_DEFAULT is only effective during normal boot, not recovery. Dozens of samples show cryptdevice in GRUB_CMDLINE_LINUX and I can confirm the grub.cfg is updated with it, I have to find something to show you. ^^
2023-02-19 17:01:56 <mbx> minimal: https://unix.stackexchange.com/questions/223214/grub-cryptdevice-with-changing-device-names
2023-02-19 17:03:02 <mbx> I wonder how grub knows the name crypt, when not specified. A spacebar would introduce a new command in the GRUB parameter, so it seems to be ":", I need to double check
2023-02-19 17:03:14 <minimal> you want to set its value to "rootfstype=ext4 cryptroot=UUID=<the UUID of the partition> cryptdm=lukspart" (assuming you want luks to be /dev/mapper/lukspart)
2023-02-19 17:03:50 <invoked> minimal: yeah that's what i meant. but i thought they'd dropped the encryption requirement
2023-02-19 17:04:18 <minimal> invoked: only for the fs containing kernel/initramfs, not for rootfs
2023-02-19 17:05:16 <mbx> minimal: thanks copied. Can you link me to a resource that has this documented. It's unlike any other example I've seen so far.
2023-02-19 17:05:36 <minimal> mbx: it is what I use for my script Alpine installs
2023-02-19 17:07:54 <minimal> plus a fstab entry for "/dev/mapper/lukspart / ext4 ..."
2023-02-19 17:08:07 <mbx> it's weird how this works to me. There is a guy on the internet who did "cryptdevice=UUID=6d5b5f47-58e8-4a9c-89c8-4f503f35ff3f:cryptroot root=/dev/mapper/cryptroot" and his solution to his problem was to update /etc/crypttab (his OS is Ubuntu). I wish to have a manual about this ...
2023-02-19 17:08:57 <minimal> mbx: do you want to know how to get this working on Alpine (the way Alpine's initramfs' init works) or how other distros handle this?
2023-02-19 17:08:58 <mbx> maybe ":cryptroot" is the same as " cryptdm=cryptroot"
2023-02-19 17:09:22 <mbx> of course I want to get this to work on Alpine, but I also want to understand it
2023-02-19 17:09:52 <minimal> Alpine's (default) initramfs' init works differently to that on other distros
2023-02-19 17:10:12 <minimal> so there's no point reading articles on how to setup other distros for that
2023-02-19 17:11:24 <mbx> ohhh, so the mkinifs is different. Is my understanding correct that the grub.cfg parameters get passed to the initramfs?
2023-02-19 17:11:49 <minimal> the cmdline parameters get passed to the initramfs' init
2023-02-19 17:12:03 <minimal> which on Alpine, by default, is that of mkinitfs
2023-02-19 17:12:37 <mbx> thank you so much, then you are right that I should not bother with other distros. For some reason I thought grub does something with it and you can imagine the level of confusion it gave me looking at so many different examples ...
2023-02-19 17:12:40 <minimal> on other distros it may be dracut or booster or something else, which may work in different ways
2023-02-19 17:13:22 <mbx> yes, I saw dracut as well, and was even more confused
2023-02-19 17:14:02 <minimal> grub itself will do "something" if you have the fs containing kernel/initramfs encrypted as Grub then needs to load its luks/luks2 module and decrypt that fs, that however is a separate activity from the initramfs' init decrypting the rootfs
2023-02-19 17:15:02 <mbx> yes, this would be a thing on full disk encryption that we conclude is not possible with argon2id and grub as of now
2023-02-19 17:15:35 <mbx> I mkfs.ext4 -O ^has_journal on the /boot partition as adviced on the docs, this was right?
2023-02-19 17:17:15 <minimal> advised in which docs?
2023-02-19 17:18:11 <mbx> also I would like to understand one final thing: when mkinitfs already decrypts and mounts the rootfs (otherwise it could not start), why do I need an entry for the rootfs on fstab?
2023-02-19 17:18:12 <minimal> why would you want to disable the ext4 journal?
2023-02-19 17:18:17 <mbx> minimal: adviced in alpine docs
2023-02-19 17:18:32 <minimal> mbx: which doc?
2023-02-19 17:18:40 <mbx> sec
2023-02-19 17:19:33 <mbx> https://wiki.alpinelinux.org/wiki/Setting_up_disks_manually#Manual_partitioning
2023-02-19 17:19:47 <mbx> quote: "The following creates an ext4 partition with disabled journaling, to reduce write operations and allow the disk to spin down after the .apkovl and the packages have been read from the partition during the boot.
2023-02-19 17:21:48 <mbx> I keep reading that journaling is bad, some folks also pick ext2 for /boot for the same reason. Journaling makes no sense for ro partitions
2023-02-19 17:21:56 <minimal> that appears to be orientated at "to reduce write operations" which might make sense for a SDcard or flash-based device. The section where it is mention is for diskless and data mode installs", you're doing a Sys mode install...
2023-02-19 17:22:34 <minimal> shouldn't you be looking at the next section, For "sys" disk mode installs ?
2023-02-19 17:22:44 <mbx> you are right, is it a problem that I did it this way? What FS should I pick ideally?
2023-02-19 17:23:18 <minimal> mbx: the fs that (a) you want to use, and also (b) that is supported by setup-disk
2023-02-19 17:25:41 <mbx> The sys section won't specifcy, so I figured picking the same should be supported, I hope that is not a mistake
2023-02-19 17:26:29 <mbx> lastly, I wonder why I need to touch fstab when mkinitfs already mounts the rootfs, I think I miss some fundamentals on this one
2023-02-19 17:27:01 <minimal> the /etc/fstab file is included into the initramfs when it is built/rebuilt
2023-02-19 17:27:39 <minimal> are you looking to get a functioning system or are you looking for someone to teach you how exactly it works?
2023-02-19 17:31:12 <mbx> I'm looking for a functional system that I know why it functions, not exactly how in detail. Thank you for your patience, these were the missing pieces. I don't like asking things that I can figure out through documentation, but these were the few things that put me off road continuesly and I couldn't get my head around it.
2023-02-19 17:31:46 <mbx> I will note things down and re-try it ^^
2023-02-19 17:33:07 <minimal> mbx: as I said before I have such stuff scripts (with *some* comments that may explain why some things are done), you could look at the output of my script (though it doesn't use setup-disk)
2023-02-19 17:33:16 <minimal> s/scripts/scripted/
2023-02-19 17:33:16 <alpine-meetbot> minimal meant to say: mbx: as I said before I have such stuff scripted (with *some* comments that may explain why some things are done), you could look at the output of my script (though it doesn't use setup-disk)
2023-02-19 18:03:20 <cddr> o/
2023-02-19 18:05:33 <cddr> I am successfully running alpine s390x in qemu, edge branch. It crashes after “cargo update” run of some rust project. Sigh. 
2023-02-19 18:07:07 <psykose> gonna need way more details than that
2023-02-19 18:25:17 <mbx> cddr: maybe you can do "cargo update >cargo.log 2>&1" to inspect the last thing it does before the crash
2023-02-19 18:31:52 <psykose> i assumed it was the `cargo update` that crashes instead :D
2023-02-19 18:32:02 <psykose> and not the whole vm
2023-02-19 18:51:33 <cddr> I am a tester of pijul vcs
2023-02-19 19:09:59 <invoked> pijul seems cool. i'm not enthused if, for instance, it stimulates people to do a bunch more backporting, but the darcs model always seemed better, but haskell always put me off it.
2023-02-19 19:10:20 <invoked> wicked comma splice
2023-02-19 19:11:52 <elly> neat!
2023-02-19 19:11:57 <elly> I've been experimenting with fossil a bit
2023-02-19 20:11:47 <nando_> my xorg session doesnt run with non root users. xorg log says im missing driver i965 and xf86OpenSerial cannot open device /dev/input/event3 permission denied
2023-02-19 20:12:00 <nando_> theres no log file generated when i run startx as root
2023-02-19 20:12:07 <psykose> are you in the video/input groups
2023-02-19 20:12:27 <nando_> probably not. ill check right now
2023-02-19 20:14:57 <mbx> minimal: I tested the configuration as you told me, but it did not work. I took some photos and will upload them
2023-02-19 20:20:56 <nando_> psykose: my user was not in the input group. thanks
2023-02-19 20:56:43 <mbx> minimal: https://bit.ly/3EkK0mk <- here is what I did and the problem with the initfs not able to boot my encrypted partition on a MBP.
2023-02-19 20:58:10 <mbx> my /etc/fstab looks clean, my /etc/default/grub like you instructed, although I kept the modules, I think this needs to be the case
2023-02-19 20:58:27 <mbx> and it's stuck on Loading initial ramdisk ...
2023-02-19 21:02:55 <mbx> pleasen reload, image was wrong, now its ok^^
2023-02-19 21:04:57 <minimal> mbx: what is "cryptsetup luksFormat -M luks2" doing? to specify luks2 you'd use "--type luks2", what does "-M" do?
2023-02-19 21:07:03 <mbx> oh dang, -M is not shorthand for --type, but since it
2023-02-19 21:07:19 <mbx> since its v2, it should use luks2 by default
2023-02-19 21:07:36 <minimal> also is your NVME a 4K sector device?
2023-02-19 21:07:46 <mbx> yes
2023-02-19 21:07:55 <mbx> blocksize is 4kb
2023-02-19 21:09:45 <minimal> can you "cat" /dev/block/nvme0n1/queue/logical_block_size and physical_block_size ?
2023-02-19 21:10:56 <mbx> sure, but I have to reboot into the ISO and setup weechat again, give me a minute
2023-02-19 21:17:12 <mbx_> minimal: I'm in the system, do you want me to luksOpen and mount first?
2023-02-19 21:18:47 <mbx_> ls /dev/block/ returns 259:0 259:1 ... so 259 has 5 entries, but cat 259:0/queue doesn't exist
2023-02-19 21:19:28 <mbx_> these are just symlinks to ../nvme0n1...
2023-02-19 21:29:26 <minimal> oops, sorry /sys/block/... not /dev/block
2023-02-19 21:30:41 <mbx_> minimal: both cat's print 4096
2023-02-19 21:31:19 <minimal> ok, so it is indeed "true" 4K
2023-02-19 21:35:07 <minimal> whenever you run grub-mkconfig you should check the resultant /boot/grub/grub.cfg before rebooting to make sure it looks correct
2023-02-19 21:35:40 <mbx_> Okay, I will mount it now and check it out
2023-02-19 21:36:01 <mbx_> is there a way to also tell initramfs to print verbose messages?
2023-02-19 21:36:56 <minimal> it does take a "debug_init" option (on cmdline), though I don't remember exactly what that does
2023-02-19 21:36:57 <ikke> mbx_: add debug to the cmdline
2023-02-19 21:37:10 <ikke> oh yes debug_init
2023-02-19 21:37:16 <ikke> it enables set +x
2023-02-19 21:37:22 <ikke> https://gitlab.alpinelinux.org/alpine/mkinitfs/-/blob/master/initramfs-init.in#L379
2023-02-19 21:37:53 <minimal> ikke: yeah, was just looking at the code
2023-02-19 21:39:07 <minimal> mbx_: also I put "GRUB_DISABLE_LINUX_UUID=true" in /etc/default/grub when I'm using encryption (though think that's more relevant when having boot files inside encrypted partition)
2023-02-19 21:40:32 <minimal> your screenshots didn't show the full GRUB_CMDLINE_LINUX_DEFAULT value...
2023-02-19 21:40:45 <mbx_> minimal: my grub.cfg is https://termbin.com/bivh
2023-02-19 21:43:20 <mbx_> ikke: could you please shorten that link for me?
2023-02-19 21:45:13 <ikke> https://tinyurl.com/39yk7whk
2023-02-19 21:45:32 <mbx_> thx
2023-02-19 21:54:42 <mbx_> ok will try that, is there anything else I can do or test? Maybe I can tell grub to give me more logs? Also I have seen a recovery mode in grub. Anything in particular what I should execute and take photos of?
2023-02-19 21:55:16 <minimal> mbx_: was still waiting to see your GRUB_CMDLINE_LINUX_DEFAULT value
2023-02-19 21:56:19 <mbx_> minimal: https://termbin.com/xkanp
2023-02-19 21:56:21 <minimal> I don't think you're specifying "cryptdm=cryptroot" as part of its value
2023-02-19 21:56:35 <mbx_> also, have you seen my grub.cfg I posted above?
2023-02-19 21:57:08 <mbx_> dang it, you are right, how could I mess up that
2023-02-19 21:57:48 <mbx_> I was setting it before when I tried to use /dev/nvme0n1p4 on cryptroot instead
2023-02-19 21:57:58 <mbx_> when I changed it to UUID, I accidently removed it
2023-02-19 21:58:29 <minimal> the cryptdm value is used to name the /dev/mapper device, i.e. cryptdm=cryptroot is result in /dev/mapper/cryptroot being created
2023-02-19 21:58:49 <mbx_> https://termbin.com/hc1fx is my grub file now
2023-02-19 21:59:17 <mbx_> yes, cryptroot is right, because my fstab links to /dev/mapper/cryptroot
2023-02-19 21:59:21 <mbx_> points*
2023-02-19 21:59:58 <mbx_> here my fstab again: https://termbin.com/77ab
2023-02-19 22:00:27 <mbx_> does it look alright? can I regenerate my grub.cfg now?
2023-02-19 22:00:39 <minimal> yeah try grub-mkconfig now
2023-02-19 22:00:48 <minimal> and let's see what the result is
2023-02-19 22:01:33 <mbx_> the warning about os-prober is not anything to be concerned about, right?
2023-02-19 22:01:46 <mbx_> when I grub-mkconfig
2023-02-19 22:02:28 <minimal> you can ignore it, it is used to auto-add Grub menu entries for other installed distros/OSes
2023-02-19 22:03:04 <mbx_> alright, thx! here is my new grub.cfg https://termbin.com/llfn
2023-02-19 22:04:16 <minimal> try booting that now?
2023-02-19 22:05:31 <mbx_> I umount and luksClose first, then I boot, make photos and I'm soon back, thank you
2023-02-19 22:10:32 <mbx> minimal: nope, that didn't work at all
2023-02-19 22:10:38 <mbx> not even verbose logs
2023-02-19 22:10:49 <mbx> the very same as before
2023-02-19 22:12:18 <mbx> so I assume it might be a problem before initramfs is loaded, but I know that someone else was able to install Alpine on his MBP with Intel T2 chipset as I did, just without encryption
2023-02-19 22:17:08 <mbx> I approach a point in which I probably have simplify everything to nail down the problem, i.e. install it in one partition without encryption, if it works, then install with different /boot partition and if that works re-try this
2023-02-19 22:17:50 <mbx> I just have to say, digging into bootloaders and efi for the first time so deep really exhausted me the past days
2023-02-19 22:18:58 <mbx> I might give your script a shot, minimal, but it would not help me to understand what is going wrong here, unless the script fails as well
2023-02-19 22:19:51 <mbx> I used to have Windows as BOOTCAMP on this partition before, then I installed debian on it, which worked, then nobara/fedora, which I never completed
2023-02-19 22:20:43 <mbx> for some reason, since I installed debian in the partition of Windows, the diskutil of macOS is incapable of resizing the APFS, even when removing the other partitions. The hardware is so unforgiven it feels like
2023-02-19 22:21:28 <mbx> I still hope I can get that spinning, so the device doesn't have to be trashed in 6 years, cuz of missing macOS software support
2023-02-19 22:22:02 <mbx> but not today, thanks for all the help once again. :)
2023-02-19 22:23:10 <minimal> your earlier screenshot showed "Loading Linux lts ..." and "Loading initial ramdisk ..." - that's Grub doing that, the next point should be the kernel itself initialising
2023-02-19 22:29:59 <minimal> I don't know anything really about Mac hardware but if Debian worked then Alpine should work
2023-02-19 22:31:49 <minimal> hmm, wondering if you should also pass cryptsetup the "--sector-size 4096" option
2023-02-19 22:34:05 <mbx> minimal: I can edit the start script in grup, so I would debug the Kernel in there? I.e. Ubuntu says gfxmode text can print any errors, but that's something from the initramfs again, I think. If you have any resources to look into, I would appreciate it. In regards of the sector-size, shouldn't the initramfs report anything?
2023-02-19 22:34:31 <mbx> cryptsetup would report "Device size is not aligned to requested sector size."
2023-02-19 22:34:49 <mbx> if it wouldn't match, so I assume it determines the size automatically
2023-02-19 22:36:18 <minimal> from the manpage, "Set sector size for use with disk encryption. It must be a power of two and in range 512 - 4096 bytes. The default is 512 bytes sectors. This option is only available in the LUKS2 mode."
2023-02-19 22:38:06 <mbx> according to https://gitlab.com/cryptsetup/cryptsetup/-/issues/585 I should get a warning if it is not compatible, but I will set the sector-size explicitly on my next full attempt
2023-02-19 22:38:39 <mbx> good to know it defaults to 512
2023-02-19 22:40:47 <minimal> definately try and install without encryption - that should rule out something like a missing kernel module
2023-02-20 02:15:52 <mxdyeah> hello
2023-02-20 08:33:49 <alpinoobe_> Gnome only works on wayland for me (on liveusb) even if I do a setup-xorg-base before setup-desktop gnome. Also cannot get awesomewm to get running with x11
2023-02-20 08:34:11 <alpinoobe_> Not sure how I can copy the logs from the tty to post here
2023-02-20 08:34:55 <alpinoobe_> Om 3.17.2 standard usb iso
2023-02-20 08:47:08 <alpinoobe_> I'll be back with the exact setup I used, will share a git repo to reproduce
2023-02-20 09:12:02 <donoban> anyone knows if USB ports on my monitor (connected via HDMI-usb-c "converter") should work? HDMI at 4k@60hz+audio works perfectly but I don't see any info related to the usb ports
2023-02-20 09:19:48 <psykose> that does not sound like it would work at all
2023-02-20 09:20:47 <psykose> don't think there's any standard to carry extra stuff over hdmi except CEC and nobody uses it for that kind of thing
2023-02-20 09:21:14 <psykose> those ports are either only usable with the C-C connection (if there is one) or just have a usb input so you directly connect extra cables to get more usb ports via the monitor
2023-02-20 09:23:00 <donoban> I think that it has C-C connection but the usb-c port on the laptop is also used for charguing, so I need to use this split/conversor
2023-02-20 09:24:00 <donoban> maybe I should look for another split that maintains usb-c/usb-c but adds the power input
2023-02-20 09:26:07 <psykose> sounds like a mess
2023-02-20 09:26:22 <psykose> i would just not use the monitor ports and get some other splitter from another port
2023-02-20 09:27:38 <donoban> I would like to have the fewer things connected to the laptop directly, but meh, is not something life-changuing :)
2023-02-20 09:45:34 <telmich> <donoban> "I think that it has C-C connecti..." <- Depending on the monitor, usb-c can be used to charge the notebook. This is what I use for all devices here
2023-02-20 09:46:05 <psykose> oh yeah
2023-02-20 09:46:15 <psykose> that is indeed a thing if it passes through enough power
2023-02-20 09:47:09 <donoban> with mine the laptop complains about not getting enough power and it discharges slowly
2023-02-20 09:47:37 <donoban> I will check that the next time I buy a monitor
2023-02-20 11:36:00 <minimal> elly: update on booting with small RAM, tried with serial console to get messages that briefly appear before kernel output with small RAM setting: "Invalid physical address chosen!" and "Physical KASLR disabled: no suitable memory region!"
2023-02-20 14:01:03 <while> hi, I'm trying to connect to my linux iscsi target with linux iscsiadm as initiator. on same host using command `iscsiadm -m node --targetname "iqn.2023-02.org.wieldfield.com:iscsivolumea" --portal 172.24.0.1 --login` it gives message "Login to [iface: default, target: iqn.2023-02.org.wieldfield.com:iscsivolumea, portal: 172.24.0.1,3260] successful." but no block device is created, and on different host I get errors "iscsiadm: 
2023-02-20 14:01:04 <while> Could not login to [iface: default, target: iqn.2023-02.org.wieldfield.com:iscsivolumea, portal: 127.0.0.1,3260].  iscsiadm: initiator reported error (24 - iSCSI login failed due to authorization failure) iscsiadm: Could not log into all portals". the documentation on this issue appears somewhat spotty
2023-02-20 14:05:31 <WhyNotHugo> If I understand correctly, udev mostly listens to kernel events for new devices and applies rules (mostly: chmod/chown), right?
2023-02-20 14:08:06 <psykose> mostly yeah
2023-02-20 14:12:36 <WhyNotHugo> It somehow doesn't sound like such a complicated thing.
2023-02-20 14:13:04 <Guest5400> who here is good with iscsi?
2023-02-20 14:17:46 <minimal> WhyNotHugo: understatement of the day? ;-)
2023-02-20 14:59:28 <WhyNotHugo> Does `TAG+="uaccess"` do nothing with eudev?
2023-02-20 15:05:04 <psykose> it's an elogind thing to permit access to the tagged device for users at a seat
2023-02-20 15:05:21 <psykose> without elogind it kinda does nothing i guess
2023-02-20 15:09:03 <psykose> there is also i guess https://git.sr.ht/~kennylevinsen/pam_uaccess but it's not exactly complete for the same
2023-02-20 15:09:43 <WhyNotHugo> I'm just a member of the `input` group (which owns the file in /dev/). I guess that should grant the same access, right?
2023-02-20 15:09:54 <psykose> yes
2023-02-20 15:10:38 <psykose> the current way of merely assigning a group and having the perms as XX0 to give group access works for any user in said group
2023-02-20 15:10:53 <WhyNotHugo> Yup, that's exactly what I'm doing.
2023-02-20 15:10:55 <psykose> uaccess instead only works for users at a "seat" that have an authorised session, and is dynamic to the session lifetime
2023-02-20 15:11:14 <WhyNotHugo> Ah, I thought that was it, but it's good to be certain of it :)
2023-02-20 15:11:22 <psykose> that above module lets you pick between losing access in all sessions once one logs out, or never losing it after grant, and also has no hotplug, because it's incomplete
2023-02-20 15:11:27 <psykose> elogind has the full version i guess
2023-02-20 15:11:41 <psykose> no hotplug means, if you plug in something.. you won't get access
2023-02-20 15:11:59 <WhyNotHugo> Yup, yup. But dev file ownership is fine for me.
2023-02-20 15:12:01 <psykose> would have to relogin as it's triggered once
2023-02-20 15:12:03 <psykose> yeah, no worries
2023-02-20 15:12:07 <psykose> just describing what i do now
2023-02-20 15:12:10 <psykose> know*
2023-02-20 15:12:34 <psykose> chimera is working on a more better elogind replacement i guess, https://github.com/chimera-linux/turnstile
2023-02-20 15:12:52 <psykose> depending on how generic it is once it's stable i might look into seeing how realistic it is for integration into a desktopish scenario
2023-02-20 15:12:57 <WhyNotHugo> The idea of dynamically giving access to users as they log-in/out is sound... but somehow implementations so far seem too complex.
2023-02-20 15:13:14 <psykose> because the idea of a "session" (logging in/out) is very poorly defined
2023-02-20 15:13:35 <psykose> the same premise is why "systemd user services" are not that simple and not well thought out either; it's kind of the same list of issues
2023-02-20 15:14:22 <psykose> it of course sounds simple in theory, but then you get to the implementing it and the 20 layers through pam and friends, possible multiple sessions whether at a seat or forwarded, ssh stuff, ...
2023-02-20 15:14:35 <psykose> there's no standard for what anything means, people winged it and what you get is what you get :)
2023-02-20 15:15:27 <WhyNotHugo> And ttys make it all the weirder. Systemd explicitly doesn't support the same user starting two sessions.
2023-02-20 15:15:38 <psykose> yep
2023-02-20 15:15:48 <WhyNotHugo> And it shows in things like DBus, where it runs one per user, so the session on the second tty will send notifications to the daemon running on the first one.
2023-02-20 15:16:17 <psykose> seat management might also sound confusing, but the premise of it is that without things locked to a ""seat"" arbitrary programs that get spawned under a session can just read everything on your screen or keyboard
2023-02-20 15:16:59 <psykose> under the session granting, when you spawn a compositor, it gets exclusive acccess to the video/input, and nothing else can read it anymore
2023-02-20 15:17:14 <psykose> and then it kinda goes from there
2023-02-20 15:17:25 <psykose> tbh i'm not too versed in the weird desktop stuff, it's a bit all over the place
2023-02-20 15:17:46 <psykose> i wish i could write code and implement something more in the middle that actually works well defined on a desktop
2023-02-20 15:18:07 <psykose> these days everything is either "systemd lol" or "something from suckless that just ignores the problem"
2023-02-20 15:18:16 <WhyNotHugo> lol, very true.
2023-02-20 15:19:50 <WhyNotHugo> seatd is a nice one tho.
2023-02-20 15:19:55 <WhyNotHugo> And I like the design of greetd.
2023-02-20 15:20:36 <psykose> seatd just does the seat part :D
2023-02-20 15:20:45 <WhyNotHugo> Does one thing and does it well :P
2023-02-20 15:20:47 <psykose> afaict it is not doing the rest of what elogind does, though of course they can pair together
2023-02-20 15:20:48 <psykose> yeah
2023-02-20 15:20:58 <WhyNotHugo> So back to my real issue: my gamepad works on Firefox or by just reading /dev/input/js0, but no on Steam in Flatpak.
2023-02-20 15:21:05 <psykose> i use greetd too with the bare tty spawner because i thought i'd upgrade from bare tty logins
2023-02-20 15:21:13 <psykose> i had to fixup some pam rules so both worked at once
2023-02-20 15:21:26 <psykose> i think before like a month ago if you just used the greetd-tty greeter you would never get a pam login
2023-02-20 15:21:27 <WhyNotHugo> greetd+agreety is an interesting replacement for login(1)
2023-02-20 15:21:52 <psykose> hmm
2023-02-20 15:22:12 <psykose> gamepad, steam. flatpak, i assume you did install the steam-udev-devices thing for extra rules
2023-02-20 15:22:18 <psykose> steam-devices*
2023-02-20 15:23:31 <psykose> aside from that afaict the flatpak rules for steam do just pass all the devices, so..
2023-02-20 15:23:59 <WhyNotHugo> steam-devices mosty adds TAG+="uaccess".
2023-02-20 15:24:09 <WhyNotHugo> I did install it, but I feel it's kinda irrelevant here.
2023-02-20 15:28:05 <psykose> aha, yeah
2023-02-20 15:28:10 <psykose> if you already had the devices, indeed
2023-02-20 15:58:44 <JohannesJ[m]> i wonder if appimages could be made working on alpine, does anyone know?
2023-02-20 16:00:53 <ptrc> https://github.com/AppImage/AppImageKit/issues/877
2023-02-20 16:19:37 <umk3[m]> @psykose steam-devices is only necessary if you wanna use uinput
2023-02-20 16:20:30 <umk3[m]> if steam-devices is not installed but uinput is loaded: steam does not pop up an information dialog telling to install udev rules
2023-02-20 16:22:05 <umk3[m]> does pop up*
2023-02-20 16:41:12 <gmalsack> I installed ossec-hids-agent and ossec-hids on alpine 3.15 using apk and it won't start. fails saying: Unable to access queue: '/var/ossec/queue/ossec/queue'. Giving up..
2023-02-20 16:52:19 <xordspar0> JohannesJ[m]: +1 to ptrc's link. I have also experiemented with building appimages that can run on anywhere using Alpine and witchery and had success: https://github.com/AppImageCrafters/appimage-builder/issues/10#issuecomment-939379591
2023-02-20 16:52:56 <xordspar0> But my experiments don't help if you want to download appimages from the Internet have have them just work.
2023-02-20 16:54:20 <JohannesJ[m]> well, that would be the end goal ;-)
2023-02-20 17:28:58 <pj> if only the technology wasn't trash fire and maintainers were not insane
2023-02-20 17:59:15 <WhyNotHugo> umk3[m]: Steam isn't showing the popup for missing rules... but it also can't detect any controllers.
2023-02-20 18:01:06 <umk3[m]> hmmm Steam flatpak couldn't detect one of my HORI controllers, despite having steam-devices and uinput enabled once
2023-02-20 18:01:14 <umk3[m]> which model is it?
2023-02-21 00:52:48 <spiralw>  this might not be the right place to ask, but why can't i connect to dl-cdn (or anything else fastly) from an alpine docker container, while i can access it from my host?
2023-02-21 00:57:13 <ptrc> can you connect to anything at all?
2023-02-21 00:58:20 <spiralw> yes, issue is only with fastly, at least as far as i can tell
2023-02-21 01:00:26 <psykose> weird
2023-02-21 01:02:34 <spiralw> the docker iptables rules on my host look sane
2023-02-21 01:02:58 <psykose> if you do something like
2023-02-21 01:03:31 <psykose> apk add s6-dns; s6-dnsqr -D2 A dl-cdn.alpinelinux.org
2023-02-21 01:03:33 <psykose> whadya get
2023-02-21 01:04:58 <spiralw> https://paste.sr.ht/blob/cb25fc39d9d3abdf2ecdecd119b4a6d1ab5a7de9
2023-02-21 01:05:21 <psykose> well, does resolve
2023-02-21 01:05:33 <spiralw> i can also `nc` to these ips and manually type http requests
2023-02-21 01:05:35 <spiralw> that works
2023-02-21 01:05:46 <psykose> what mirror did you install that from? since if it was dl-cdn..
2023-02-21 01:05:56 <spiralw> oh, no, i changed the mirror to uwaterloo
2023-02-21 01:06:00 <psykose> ah
2023-02-21 01:06:07 <spiralw> but that's very inconvenient to do in my dockerfiles
2023-02-21 01:06:11 <psykose> which error do you get with the dl-cdn mirror specifically again
2023-02-21 01:06:35 <spiralw> no error, apk hangs
2023-02-21 01:07:00 <spiralw> wget also hangs after "writing to stdout"
2023-02-21 01:07:58 <psykose> that is so weird
2023-02-21 01:08:39 <spiralw> yeah. it's only happening on my laptop, whichever network i connect it to
2023-02-21 01:09:17 <psykose> sounds like some really bizarre network issue
2023-02-21 01:09:21 <psykose> since well, usually it's dns
2023-02-21 01:09:27 <psykose> this seems to not be :D
2023-02-21 01:10:03 <spiralw> yeah, i'm unsure what to even look at tbh
2023-02-21 01:10:12 <spiralw> just tried it with curl and it gets as far as to print the headers
2023-02-21 01:10:13 <spiralw> then hangs
2023-02-21 01:10:23 <elagost> dl-cdn works on my computer, and on a vm, but when I bring up a wireguard interface on the host, dl-cdn doesn't work on the vm.
2023-02-21 01:10:27 <spiralw> https://paste.sr.ht/~spiral/535307f8b1fa4a126aed4b76d006e0bed7f94cc0
2023-02-21 01:10:34 <spiralw> oh
2023-02-21 01:10:37 <spiralw> lemme try without wireguard
2023-02-21 01:10:58 <spiralw> ???
2023-02-21 01:10:59 <spiralw> that worked
2023-02-21 01:11:08 <psykose> ahh
2023-02-21 01:11:14 <psykose> this sounds weirdly familiar
2023-02-21 01:11:35 <psykose> ptrc: do you remember some weird wireguard routing where the stuff gets half truncated
2023-02-21 01:11:43 <psykose> something something mtu
2023-02-21 01:11:59 <ptrc> hmmmm
2023-02-21 01:12:01 <ptrc> not sure
2023-02-21 01:12:57 <ptrc> there was *something* with the mtu, but iirc it was a bit different than what's happening here
2023-02-21 01:14:56 <spiralw> ok, for now i added static routes to dl-cdn that don't go through my vpn
2023-02-21 01:14:58 <spiralw> that seems to work
2023-02-21 01:16:55 <psykose> gotta say getting all the headers and none of the body is quite the networking :D
2023-02-21 01:17:04 <dwfreed> if you control both sides of the wireguard tunnel, I would tcpdump the wireguard traffic
2023-02-21 01:21:21 <spiralw> good idea, will do that at some point
2023-02-21 01:21:34 <spiralw> i've been having really weird issues on and off with this wireguard tunnel
2023-02-21 01:58:44 <minimal> psykose: with wireguard a smaller mtu needs to be used (1420?)
2023-02-21 01:58:49 <psykose> ye
2023-02-21 05:34:55 <ikke> fastly is not sending fragmentation needed packets
2023-02-21 09:52:32 <caskd> Since ifupdown-ng switched to dhclient what's the correct way to disable modification of resolv.conf?
2023-02-21 09:52:57 <caskd> there used to be /etc/udhcpc/udhcpc.conf with RESOLV_CONF=no
2023-02-21 09:57:37 <psykose> it did?
2023-02-21 09:57:45 <psykose> doesn't it only use it if it's installed
2023-02-21 09:57:49 <caskd> oh
2023-02-21 09:57:57 <psykose> and even then, you can pick it
2023-02-21 09:57:58 <psykose> somewhere
2023-02-21 09:58:01 <caskd> damn
2023-02-21 09:58:04 <caskd> okay sec
2023-02-21 09:58:15 <caskd> i didnt explicitly install it
2023-02-21 09:58:16 <psykose> there's some priority order
2023-02-21 09:58:18 <psykose> idk in which script
2023-02-21 09:59:14 <caskd> uh the executor takes IF_DHCP_PROGRAM
2023-02-21 09:59:18 <caskd> but idk where to set that
2023-02-21 10:00:09 <Ermine> /usr/libexec/ifupdown-ng/dhcp
2023-02-21 10:00:26 <caskd> yeah that executor
2023-02-21 10:00:34 <caskd> but i dont wanna change the executor script itself
2023-02-21 10:00:40 <caskd> there must be a way from outside
2023-02-21 10:01:06 <psykose> conf.d of networking?
2023-02-21 10:01:14 <Ermine> it is controlled by env variable
2023-02-21 10:01:16 <psykose> export IF_DHCP_PROGRAM=
2023-02-21 10:01:36 <caskd> in /etc/conf.d/networking?
2023-02-21 10:01:40 <caskd> oh
2023-02-21 10:01:42 <Ermine> I've submitted a merge request to add an option to interfaces file. It wasn't reviewed yet
2023-02-21 10:01:44 <caskd> yeah
2023-02-21 10:03:17 <caskd> oh but if i call ifup or ifdown manually it doesn't apply
2023-02-21 10:03:22 <caskd> i gotta source the conf.d
2023-02-21 10:03:30 <caskd> :/
2023-02-21 10:03:42 <Ermine> Yes. OpenRC sources relevant confd files
2023-02-21 10:04:01 <Ermine> you can specify env variables manually
2023-02-21 10:04:42 <psykose> yeah not great
2023-02-21 10:06:44 <caskd> uh
2023-02-21 10:06:55 <caskd> the conf.d method seems to not work?
2023-02-21 10:08:06 <Ermine> psykose: what if I submit that patch to alpine? (https://github.com/ifupdown-ng/ifupdown-ng/pull/201)
2023-02-21 10:08:19 <psykose> if it exports it would work
2023-02-21 10:08:50 <psykose> Ermine: i'd rather it was upstream first (yes yes i know)
2023-02-21 10:08:57 <psykose> also you have spaces instead of a tab on one of the indents there
2023-02-21 10:09:16 <Ermine> Mew
2023-02-21 10:09:31 <caskd> export IF_DHCP_PROGRAM=udhcpc
2023-02-21 10:09:39 <caskd> i have exactly this in conf.d/networking
2023-02-21 10:09:48 <caskd> and it still spawns dhclient
2023-02-21 10:09:51 <caskd> O_o
2023-02-21 10:10:04 <caskd> oh wait i need the path
2023-02-21 10:10:36 <Ermine> It's rfc patch anyway. I wanted to get opinions on whether this is needed and whether proposed implementation is good. If ok, I need to work on documenting it and probably adding some tests
2023-02-21 10:11:00 <caskd> no i don't
2023-02-21 10:11:01 <caskd> uhh
2023-02-21 10:13:25 <caskd> seems like the var is swallowed somewhere
2023-02-21 10:13:38 <caskd> dang
2023-02-21 10:14:03 <caskd> even calling ifup directly doesn't help
2023-02-21 10:15:59 <psykose> weird
2023-02-21 13:04:07 <minimal> caskd: why not simply "apk del dhclient" ?
2023-02-21 13:08:34 <caskd> as i said, was pulled in as dep, not explicitly installed
2023-02-21 13:26:39 <WhyNotHugo> After upgrading to linux-edge 6.2.0, my computer won't reboot. I see POST, then a black screen, and then the system immediately reboots.
2023-02-21 13:26:46 <WhyNotHugo> I'm using an UKI as described here -> https://wiki.alpinelinux.org/wiki/UEFI_Secure_Boot#Generating_Unified_Kernel_Image
2023-02-21 13:26:52 <WhyNotHugo> But I'm NOT using secureboot.
2023-02-21 13:27:05 <WhyNotHugo> I reverted to kernel 6.1.11, and things work fine again.
2023-02-21 13:27:12 <WhyNotHugo> But I have no idea what's broken, nor where to look.
2023-02-21 13:27:56 <minimal> WhyNotHugo: compare the kernel settings of linux-lts and linux-edge to see what differs?
2023-02-21 13:28:08 <WhyNotHugo> Both are edge.
2023-02-21 13:28:08 <psykose> diffing the configs probably
2023-02-21 13:28:20 <WhyNotHugo> I upgraded package linux-edge 6.1.11 to 6.20.0.
2023-02-21 13:28:34 <WhyNotHugo> *6.2.0
2023-02-21 13:29:20 <psykose> https://cateee.net/lkddb/web-lkddb/EFI_GENERIC_STUB_INITRD_CMDLINE_LOADER.html was removed
2023-02-21 13:29:21 <psykose> so...
2023-02-21 13:29:46 <psykose> or am i looking at the wrong thing
2023-02-21 13:29:47 <psykose> hm
2023-02-21 13:30:29 <WhyNotHugo> That was indeed removed in aports between 6.1.11 and 6.2.0. I guess I need to understand if that's what's causing the breakage.
2023-02-21 13:30:58 <psykose> probably
2023-02-21 13:31:36 <psykose> everything else is random stuff
2023-02-21 13:33:39 <ikke> "This method is deprecated."
2023-02-21 13:33:42 <minimal> but is that option relevant for a UKI?
2023-02-21 13:34:45 <psykose> then it's not that
2023-02-21 13:35:06 <psykose> yeah just gone in 6.2 naturally
2023-02-21 13:38:27 <WhyNotHugo> My /proc/cmdline doesn't show initrd=.
2023-02-21 13:40:31 <WhyNotHugo> So I don't think that option is being used...?
2023-02-21 13:41:06 <WhyNotHugo> Hmm.. but sounds closely related enough.
2023-02-21 13:41:25 <psykose> it's gone entirely i.e. not an option anymore
2023-02-21 13:41:27 <psykose> i just got confused
2023-02-21 13:41:32 <psykose> the actual issue is something else
2023-02-21 13:41:46 <psykose> either 6.2 just broken on your hardware or some other option, but they all seem benign
2023-02-21 13:42:07 <WhyNotHugo> Well, it was deleted in 6.2.0, but if I don't see it in /proc/cmdline when botting on 6.1.11 it implies that it's not relevant to my setup... right?
2023-02-21 13:55:14 <WhyNotHugo> Okay, pinned linux-edge=6.1.11 and upgraded all other packages. System boots fine. So pretty sure the issue is the kernel.
2023-02-21 13:56:17 <psykose> yeah
2023-02-21 14:07:40 <minimal> WhyNotHugo: I suspect a different issue, UEFI linux-edge (not UKI) is not booting for me, double-checking linux-lts currently
2023-02-21 14:08:32 <psykose> i never had any uefi booting regressions through all the 6.2 rc's
2023-02-21 14:09:11 <minimal> psykose: I haven't tried linux-edge in a while
2023-02-21 14:09:25 <psykose> me either
2023-02-21 14:12:59 <minimal> linux-lts booted fine, linux-edge failed in initramfs' init with "Can't open blockdev" when trying to mount rootfs, am digging into it
2023-02-21 14:41:40 <minimal> looks like nlplug-findfs didn't load the scsi kernel modules (this is using virtio-scsi) for some reason
2023-02-21 14:42:14 <psykose> hm
2023-02-21 14:44:12 <minimal> even after using modprobe to load sd_mod, virtio_scsi etc I don't see a /dev/sda device
2023-02-21 14:47:49 <minimal> indeed I'm not even seeing "scsi host0: Virtio SCSI HBA" in dmesg output
2023-02-21 14:53:06 <minimal> WhyNotHugo: so the issue I'm seeing with linux-edge is not a UEFI issue, it is something else
2023-02-21 14:58:41 <WhyNotHugo> My main storage in /dev/nvme0, so it seems that it's not limited to SCSI.
2023-02-21 15:01:45 <minimal> yeah, I'm just trying to diagnose what's in front on me. Its complicated as there's no longer a linux-edge4virt package so I'm comparing linux-virt config with linux-edge and obviously there are so many differences in general :-(
2023-02-21 15:02:03 <minimal> hard to see the wood for the trees...
2023-02-21 15:03:34 <minimal> WhyNotHugo: you said you'd tried linux-edge 6.1.1, what about 6.1.12? it came out 7 days ago
2023-02-21 15:09:27 <WhyNotHugo> Lemme try it
2023-02-21 15:11:22 <WhyNotHugo> apk add linux-edge=6.1.12  -> linux-edge-6.1.11-r0: breaks: world[linux-edge=6.1.12]
2023-02-21 15:11:36 <WhyNotHugo> Tried changing the pin in /etc/world/apk to 6.1.12, but apk fix won't resolv either.
2023-02-21 15:11:54 <WhyNotHugo> This is a user error, but not sure what :/
2023-02-21 15:12:52 <minimal> well the problem is that unless you have a mirror that keeps old versions of packages around then 6.1.12 won't be there as 6.2.0 replaced it in the past day
2023-02-21 15:13:32 <psykose> it can't add linux-edge=6.1.12 to world because it doesn't exist in any mirror
2023-02-21 15:13:51 <psykose> and all i can select is linux-edge-6.1.11-r0, but that breaks your command of .12
2023-02-21 15:14:15 <WhyNotHugo> Lemme build 6.1.12 from aports and try.
2023-02-21 15:14:29 <WhyNotHugo> I have a btrfs snapshot to roll back to when things break (which is what I did today )
2023-02-21 15:18:48 <ikke> WhyNotHugo: apk add linux-edge~6.1.12
2023-02-21 15:19:25 <ikke> Or if it's clearer: apk add linux-edge=~6.1.12
2023-02-21 15:20:14 <WhyNotHugo> ikke: I think the issue is that it's gone from mirrors.
2023-02-21 15:21:02 <ikke> We only keep a single version around 
2023-02-21 15:24:15 <ikke> But wanted to point out that if you use ~, you don't need to specify the revision
2023-02-21 15:24:31 <WhyNotHugo> Ah, I missed that, thanks.
2023-02-21 15:24:33 <WhyNotHugo> 6.1.12-0-edge boots fine.
2023-02-21 15:45:20 <minimal> WhyNotHugo: linux-edge 6.2.0 UEFI using SATA is also not booting for me
2023-02-21 16:13:31 <WhyNotHugo> How did you get dmesg output?
2023-02-21 16:13:40 <WhyNotHugo> A VM?
2023-02-21 16:13:59 <WhyNotHugo> My setup reboots after POST.
2023-02-21 16:51:13 <minimal> hyNotHugo: I'm not seeing any reboots, I'm seeing an issue with the initramfs' init mounting the rootfs, I'm seeing it with both UEFI and BIOS. So it may be unrelated issue
2023-02-21 21:49:04 <edd> so, i used dd to write the install iso to a usb stick, booted it and got "mounting boot media failed", what gives?
2023-02-21 21:49:45 <edd> (have tried different usb sticks also)
2023-02-21 21:51:32 <edd> ok, that's funny, it boots if there is no keyboard present... wat?
2023-02-21 22:19:26 <elly> the keyboard enumerates as a mass storage device for some reason?
2023-02-22 09:51:24 <markand> I heard in the past that some LLVM versions were planned to add support for sanitizers on musl
2023-02-22 09:51:47 <markand> what are the possibilities right now? it looks like neither clang/gcc have for 3.17/edge
2023-02-22 09:59:53 <edd> morning
2023-02-22 10:00:10 <edd> i tried alpine for the first time yesterday, and have a few questions
2023-02-22 10:00:21 <edd> is there a way to always install the man pages?
2023-02-22 10:00:59 <caskd> apk install docs
2023-02-22 10:01:03 <caskd> add*
2023-02-22 10:01:06 <caskd> wups
2023-02-22 10:03:38 <edd> that install docs for all installed packages?
2023-02-22 10:07:45 <markand> yes
2023-02-22 10:08:00 <markand> install mandoc if you just want the man command (but you will not be able to man many things)
2023-02-22 10:08:11 <markand> or then you can install -doc package manually for what you want
2023-02-22 10:16:32 <edd> thanks
2023-02-22 10:16:48 <edd> also, is there any local mail delivery system by default?
2023-02-22 10:16:57 <edd> I notice mail(1) isn't present
2023-02-22 10:38:17 <edd> whats the name of the kernel in the installer iso, so i can append the arguments to use a serial line?
2023-02-22 10:39:30 <ikke> lts
2023-02-22 10:40:44 <edd> thank you!
2023-02-22 10:41:59 <edd> So I'm an openbsd user, and i have to say, this installer is very similar to the openbsd installer
2023-02-22 10:42:10 <edd> i assume that's no coincidence?
2023-02-22 10:50:23 <edd> hum, "enter ssh key" can't find a key in my `.pub` file
2023-02-22 11:38:42 <minimal> WhyNotHugo: did further testing, linux-edge is working for UEFI booting in general (there was an unrelated issue at my end) so I guess your problem may be UKI specific. I'll try UKI booting and see what happens
2023-02-22 12:04:24 <WhyNotHugo> minimal: Oh, the drives not picking up was an unrelated issue?
2023-02-22 12:05:01 <WhyNotHugo> I've been planninig on bisecting it. The main issue is that after every failed boot, I'd need to boot into recovery media to restore a working UKI.
2023-02-22 12:15:22 <minimal> WhyNotHugo: yeah, it had been a while since I'd last tested with linux-edge and so my script's config had bitrotted a bit with regard to linux-edge
2023-02-22 12:25:16 <WhyNotHugo> Are there many differences between lts and edge (aside from the usptream source)?
2023-02-22 12:25:26 <ikke> different config options
2023-02-22 12:25:44 <WhyNotHugo> I did notice that one enabled binderfs and the other doesn't. Is this intentional, or it's just that someone enabled it in lts and nobody did on edge?
2023-02-22 12:28:02 <minimal> yes there is quite a difference in the way they are configured
2023-02-22 12:28:47 <minimal> the configuration of linux-lts and linux-edge are not "aligned", they are maintained by different people
2023-02-22 12:40:56 <while> I'm having some issues with the auth on an iscsi installment I'm working on: https://termbin.com/0edmp
2023-02-22 12:42:41 <umk3[m]> WhyNotHugo: did you fix the Steam controller thing?
2023-02-22 13:49:59 <JohannesJ[m]> i have a raspberry 4B, i loaded alpine-3.17-2-aarch64 on it. it boots, but it doesnt seem to reckognize the network? it asks to configure eth0 but then it says fail to bring up network. does anyone know of this issue? 
2023-02-22 13:51:20 <JohannesJ[m]> bah, now it works :|
2023-02-22 13:53:11 <psykose> markand: sanitizers have worked since like llvm12 or so
2023-02-22 13:54:01 <psykose> apk add clang compiler-rt; clang -fsanitize=address,undefined
2023-02-22 13:54:13 <markand> let me check
2023-02-22 13:55:16 <markand> /usr/lib/gcc/aarch64-alpine-linux-musl/12.2.1/../../../../aarch64-alpine-linux-musl/bin/ld: cannot find -lasan: No such file or directory
2023-02-22 13:55:19 <markand> I get this
2023-02-22 13:55:35 <markand> ah nvm
2023-02-22 13:55:43 <markand> CC=clang and it works
2023-02-22 13:56:01 <markand> psykose, thanks!
2023-02-22 13:56:33 <markand> on aarch64 it reduces drastically the performances tho :P
2023-02-22 13:56:45 <psykose> what does
2023-02-22 13:57:29 <psykose> did you expect sanitizers to make your code faster? :p
2023-02-22 13:58:10 <psykose> they're pretty much just for testing and development, gwp-asan/hwasan/ubsan-subsets aside
2023-02-22 14:01:01 <markand> no of course, but on my laptop it does not slow down that much :P
2023-02-22 14:24:05 <JohannesJ[m]> downsizing is fun :)
2023-02-22 14:26:02 <minimal> rather than rightsizing? ;-)
2023-02-22 14:28:35 <JohannesJ[m]> hm?
2023-02-22 14:34:19 <markand> 
ACTION throws a rod on JohannesJ[m] 

2023-02-22 14:34:32 <JohannesJ[m]> hm?
2023-02-22 14:47:23 <xMass> Hi guys, does alpine has something similar to what rhel has with the hosts.allow config file? 
2023-02-22 14:49:26 <xMass> I'm using TFTP and i get a request on the alpine server from the port 69 and them when it's time to send something back to the device it returns that the port is unreacheable, looking in the packet capture it tries to check by ICMP the port, there is no firewall whatsoever between this guys and the port is avaliable 
2023-02-22 15:16:18 <minimal> xMass: you'd need to provide more detailed information
2023-02-22 15:44:48 <itwrx> trying to set up alpine mirror, getting: rsync error: syntax or usage error... line 32: rsync://rsync.alpinelinux.org/alpine/: not found
2023-02-22 15:45:40 <ikke> What is line 32?
2023-02-22 15:46:15 <psykose> just post the script
2023-02-22 15:54:31 <oleovit> hi hard workers. short & maybe dumb question: is it possible to perfoem a full disk encryption via setup-apine -e _after_  a complete install? i guess not..
2023-02-22 15:57:47 <itwrx> removed commented lines so it won't be line 32 anymore. i'm just trying to use the script from the wiki right now. it's the src line for the upstream alpine mirror
2023-02-22 15:57:52 <itwrx> #!/usr/bin/env sh
2023-02-22 15:57:53 <itwrx> # make sure we never run 2 rsync at the same time
2023-02-22 15:57:53 <itwrx>   exec env flock=1 flock -n $lockfile "$0" "$@"
2023-02-22 15:57:53 <itwrx> lockfile="/tmp/alpine-mirror.lock"
2023-02-22 15:57:53 <itwrx> if [ -z "$flock" ] ; then
2023-02-22 15:57:55 <itwrx> fi
2023-02-22 15:57:59 <itwrx> src=rsync://rsync.alpinelinux.org/alpine/
2023-02-22 15:58:01 <itwrx> dest=/var/www/alpine-mirror/
2023-02-22 15:58:03 <itwrx> # uncomment this to exclude old v2.x branches
2023-02-22 15:58:05 <itwrx> exclude="--exclude v2.*"
2023-02-22 15:58:07 <itwrx> mkdir -p "$dest"
2023-02-22 15:58:09 <itwrx>         --archive \
2023-02-22 15:58:11 <itwrx>         --update \
2023-02-22 15:58:13 <itwrx>         --hard-links \
2023-02-22 15:58:15 <itwrx>         --delete \
2023-02-22 15:58:17 <itwrx>         --delete-after \
2023-02-22 15:58:19 <itwrx>         --delay-updates \
2023-02-22 15:58:21 <itwrx>         --timeout=600 \
2023-02-22 15:58:23 <itwrx>         $exclude \
2023-02-22 15:58:25 <itwrx>         "$src" "$dest"
2023-02-22 16:00:26 <minimal> oleovit: setup-disk is for doing an install.........and so it shouldn't be run after the install is done
2023-02-22 16:01:16 <psykose> that looks like you deleted the /usr/bin/rsync line
2023-02-22 16:01:22 <minimal> as to whether it is possible a existing unencrypted filesystem, that's a different question ;-)
2023-02-22 16:01:37 <markand> itwrx, for paste longer than 3-5 lines it's still better to use a pastebin please
2023-02-22 16:01:40 <minimal> s/ a existing/to encrypt an existing/
2023-02-22 16:01:40 <alpine-meetbot> minimal meant to say: as to whether it is possibleto encrypt an existing unencrypted filesystem, that's a different question ;-)
2023-02-22 16:02:44 <oleovit> yeah, i thought so, thx minimal. maybe i should take a closer look at https://wiki.archlinux.org/title/Dm-crypt/Device_encryption
2023-02-22 16:04:25 <oleovit> .. or reinstall lovely alpine. thx minimal
2023-02-22 16:05:56 <minimal> oleovit: there is "cryptset reencrypt --encrypt" to do in-situ encryption in theory, I've never looked into it
2023-02-22 16:06:10 <minimal> s/cryptset/cryptsetup/
2023-02-22 16:06:10 <alpine-meetbot> minimal meant to say: oleovit: there is "cryptsetup reencrypt --encrypt" to do in-situ encryption in theory, I've never looked into it
2023-02-22 16:08:50 <itwrx> sorry for all the trouble. /usr/bin/rsync is there. pls see here: https://pastebin.com/bXXJVdTs
2023-02-22 16:09:26 <oleovit> that's interesting, have to study https://man7.org/linux/man-pages/man8/cryptsetup-reencrypt.8.html
2023-02-22 16:09:47 <oleovit> thx again minimal for the hint
2023-02-22 16:10:26 <itwrx> i'm just trying to use the script from the wiki and it's giving me the "not found" error, which i interpret to mean it can't find the alpine rsync domain.
2023-02-22 16:13:05 <itwrx> nvrmind. sorry... orig is running right now. something in my bastardized version is wrong...
2023-02-22 17:00:36 <minimal> Just happened to notice in this page: https://postgrest.org/en/stable/install.html
2023-02-22 17:00:37 <minimal> "We discourage building and using PostgREST on Alpine Linux because of a reported GHC memory leak on that platform."
2023-02-22 17:02:27 <ikke> https://github.com/PostgREST/postgrest/issues/867
2023-02-22 17:02:36 <ikke> from 2017
2023-02-22 17:02:45 <ikke> Wonder how accurate that still is
2023-02-22 17:03:34 <ikke> https://github.com/PostgREST/postgrest-docs/issues/399
2023-02-22 17:04:53 <psykose> does ghc's "memory management" still involve malloc(20GB)
2023-02-22 17:08:04 <ikke> That last issue mentioned there might not even have been concrete evidence there actually was a memory leak
2023-02-22 17:09:55 <psykose> yes
2023-02-22 21:34:59 <el[m]> there seems to be a possible issue with the fcron package: ERROR: fcron-3.3.1-r2.post-install: script exited with error 1 (if it's not just me)
2023-02-22 21:36:02 <ikke> It executes /usr/bin/fcrontab -z -u systab
2023-02-22 21:36:25 <ikke> so that appears to error out
2023-02-22 21:38:14 <el[m]> seems like it: https://paste.centos.org/view/raw/87a5dc42
2023-02-22 21:38:42 <ikke> did you execute it as root?
2023-02-22 21:39:02 <el[m]> yup
2023-02-22 21:39:36 <el[m]> it's a pretty standard alpine install although it went through a few upgrades (of alpine)
2023-02-22 21:44:51 <ikke> installs fine in a docker container
2023-02-22 21:45:20 <el[m]> yeah i'm guessing it's probably something related to the previous upgrades
2023-02-22 21:45:29 <ikke> hmm
2023-02-22 21:45:36 <ikke> If I run the command manually: fcron is not running :
2023-02-22 21:45:36 <el[m]> but i have no clue what, i actually also just use docker on this (alpine host with docker) so i didn't change much on the host system
2023-02-22 21:45:51 <ikke> el[m]: what happens if you stop fcron?
2023-02-22 21:46:04 <el[m]> this happens on install so i'm not sure it even starts
2023-02-22 21:46:19 <ikke> hmm, ok
2023-02-22 21:46:21 <el[m]> oh interesting it's running
2023-02-22 21:46:26 <el[m]> mmm
2023-02-22 21:46:50 <el[m]> that's interesting because i uninstalled it (i got this error on upgrade and was like hmm maybe i should just reinstall it)
2023-02-22 21:46:55 <el[m]> 😂
2023-02-22 21:47:08 <ikke> right, but uninstalling it does not stop it
2023-02-22 21:47:13 <ikke> nor does installing it start it
2023-02-22 21:47:35 <el[m]> ok i stopped it now, uninstalled again, reinstalled, same error in the post-install
2023-02-22 21:47:54 <el[m]> and it's indeed not running so the install didn't start it
2023-02-22 21:47:56 <ikke> and same error if you run it manually?
2023-02-22 21:48:12 <ikke> the command in the post-install script I mean
2023-02-22 21:49:25 <ikke> el[m]: what does `id systab` return?
2023-02-22 21:49:29 <el[m]> i ran that /usr/bin/fcrontab -z -u systab and that gives the same error yes
2023-02-22 21:49:43 <el[m]> i can start it via rc-service fcron start though and it keeps running, very interesting
2023-02-22 21:49:56 <ikke> yeah, I would not imagine this preventing it from working
2023-02-22 21:50:00 <el[m]> # id systab
2023-02-22 21:50:00 <el[m]> id: ‘systab’: no such user: Invalid argument
2023-02-22 21:50:07 <el[m]>  * ````... (full message at <https://matrix.org/_matrix/media/v3/download/matrix.org/sbfegTvOwWJtFjlgOeSDGvKV>)
2023-02-22 21:50:27 <ikke> ok, similar for me
2023-02-22 21:50:39 <ikke> id: unknown user systab
2023-02-22 21:51:05 <ikke> maybe time to strace it?
2023-02-22 21:51:08 <el[m]> to be honest, i don't actually know if this error hasn't showed for a year already or so. i don't always watch apk output as well as i should 😬 i never noticed it though
2023-02-22 21:52:01 <ikke> apk might keep showing that a package has an error
2023-02-23 06:27:12 <mbx> good morning. How to suppress errors from apk? "apk add unknown-package >&2" will still print that error message. I think apk sends errors to the wrong output.
2023-02-23 06:27:41 <mbx> nvm, ofc it will, I'm so smart ...
2023-02-23 06:27:56 <mbx> apk add unknown-package 2>/dev/null is what I searched for
2023-02-23 06:28:01 <caskd> yes
2023-02-23 06:28:05 <aron> indeed
2023-02-23 06:28:10 <caskd> i know those moments
2023-02-23 06:28:15 <mbx> man it's really too late for me ...
2023-02-23 06:28:24 <caskd> hitting them bed helps often
2023-02-23 06:28:37 <mbx> almost done :)
2023-02-23 06:35:10 <mbx> wait a sec, "apk add unknown-pkg 2>/dev/null" still prints the error for me, it really seems to print to the wrong output ...
2023-02-23 06:35:50 <mbx> I need to silent those errors without silencing success messages ...
2023-02-23 06:37:15 <ikke> The ERROR message is sent to stderr, but not the diagnostic information
2023-02-23 06:38:35 <caskd> mbx: it doesn't for me
2023-02-23 06:38:37 <caskd> huh
2023-02-23 06:38:57 <caskd> and in general for everyone
2023-02-23 06:39:14 <mbx> ikke: shouldn't diagnostics be part of the error as well?
2023-02-23 06:40:24 <mbx> I get "unknown-pkg (no such package): required by: world[unknown-pkg]"
2023-02-23 06:41:20 <caskd> oh wait
2023-02-23 06:41:22 <caskd> yes
2023-02-23 06:41:25 <caskd> ur right
2023-02-23 06:42:33 <caskd> might be some fancy formatting function which doesn't get redirected
2023-02-23 06:44:49 <ikke> https://gitlab.alpinelinux.org/alpine/apk-tools/-/blob/master/src/commit.c#L540
2023-02-23 06:49:23 <mbx> caskd: so is this a bug that will be fixed in the future or is this a feature? :D
2023-02-23 06:51:38 <mbx> ikke: I get that, I still think it makes more sense to have this as part of the stderr, because it enables to shut apk up on failed attempts
2023-02-23 06:51:42 <caskd> well, i cannot say anything about when or if this will be fixed but i assume that apkv3 will fix it?
2023-02-23 06:51:56 <caskd> when it eventually becomes usable...
2023-02-23 06:51:59 <ikke> Looks like quite deliberate
2023-02-23 06:52:31 <caskd> still looks and feels wrong
2023-02-23 06:53:03 <ikke> mbx: https://gitlab.alpinelinux.org/alpine/apk-tools/-/issues/7107
2023-02-23 06:53:16 <ikke> That's purely about errors
2023-02-23 06:54:49 <ikke> Best to open an issue
2023-02-23 06:56:50 <mbx> so there was an IO error in regards of printing errors and instead of fixing the IO error the commit was reversed
2023-02-23 07:03:30 <mbx> I re-opened: https://gitlab.alpinelinux.org/alpine/apk-tools/-/issues/10879
2023-02-23 09:27:19 <mbx> is there a way to install nodejs with the correct npm version right away, so i don't have to look the right version up?
2023-02-23 09:28:17 <ikke> What is the correct npm version? 
2023-02-23 09:36:12 <mbx> it can be found in this JSON: https://nodejs.org/dist/index.json
2023-02-23 09:36:38 <mbx> installing the latest NPM for an older node version often causes problems
2023-02-23 09:36:55 <mbx> I thought about parsing that file, but in POSIX this won't be easy
2023-02-23 09:37:44 <mbx> nvm (node version manager) does it right by packaging npm with node right away, I doubt there is anyone who uses node without npm
2023-02-23 09:38:24 <mbx> but I can't use nvm, cuz the official binaries aren't compatible with Alpine
2023-02-23 09:38:49 <mbx> any suggestion how to automate this?
2023-02-23 09:42:16 <lucidiot> is installing `jq` first to go through the JSON possible? `curl -s https://nodejs.org/dist/index.json | jq -r --arg version "$(node --version)" '.[] | select(.version == $version).npm'`
2023-02-23 09:44:10 <lucidiot> i guess you could also write some JS and use the node you just installed to make that JSON parsing
2023-02-23 09:50:07 <mbx> lucidiot: thanks, that's a good idea, I can use that node right away to get this done, also thx for jq, seems to be a useful tool :)
2023-02-23 09:50:34 <mbx> I will try that later, need a break for now, bye ^^
2023-02-23 14:04:38 <baalajimaestro> how would you guys consider debugging a system that hangs up (possibly), i am not sure whats going wrong, and it just times out, and doesnt issue a 521, clearly mentioning the server is up, but not responding to any request. This is even for ssh ports. It doesnt have a system journal or a previous boot dmesg, it is alpine 3.17, running LTS kernel
2023-02-23 14:04:38 <baalajimaestro> i am unable to deduce anything from metrics i collect either, it doesnt mention any anomaly at the time of death
2023-02-23 14:14:30 <minimal> baalajimaestro: "hangs up"? you mean the OS crashes/locks up? "doesnt have a system journal" - Alpine doesn't run systemd and so no journal, however you could/should run a syslogd, if you use one configured to write a "mark" message at intervals then after a system lockup and reboot you may have a better idea of when the lookup occurred. As for previous boot dmesg, edit /etc/conf.d/bootmisc and set previous_dmesg to "yes"
2023-02-23 14:16:21 <baalajimaestro> minimal: thanks for the pointers, It does feel like a lock-up. Since every network request i send to it, just ends up timing out, even ICMP pings. While there is no firewall limiting any of these.
2023-02-23 14:17:13 <baalajimaestro> What is this mark message at intervals? How can i accomplish it?
2023-02-23 14:19:08 <baalajimaestro> syslog does appear to be running, but i am unable to read any logs from previous boot
2023-02-23 14:20:18 <minimal> it is a feature of some syslog daemons
2023-02-23 14:20:28 <minimal> e.g. for rsyslog it is enable via the immark module
2023-02-23 14:21:10 <baalajimaestro> hmmm okay
2023-02-23 14:21:12 <minimal> also does console access work when you have network problems?
2023-02-23 14:22:38 <baalajimaestro> minimal: this is a remote server, so i am not sure about that. the server appears fully unresponsive over the network atleast. while my first suspicion was a DoS, there is no metric to prove it.
2023-02-23 14:27:28 <baalajimaestro> syslogd binary points to busybox?
2023-02-23 14:27:35 <baalajimaestro> s/?/, thats it for me/
2023-02-23 14:27:35 <alpine-meetbot> baalajimaestro meant to say: syslogd binary points to busybox, thats it for me
2023-02-23 14:39:20 <minimal> baalajimaestro: busybox syslogd does not appear to support writing MARK messages
2023-02-23 14:40:28 <minimal> if the server has a serial port then you could enable a serial console and access that from another machine to see if access is still possible that way
2023-02-23 14:43:25 <WhyNotHugo> umk3[m]: Never figured out what was wrong with it.
2023-02-23 14:52:41 <baalajimaestro> <minimal> "baalajimaestro: busybox syslogd..." <- So, how do i switch it to another syslogd? Like rsyslogd, for example.
2023-02-23 14:53:38 <minimal> baalajimaestro: you install it and configure it, e.g. "apk add rsyslog"
2023-02-23 15:03:34 <minimal> Whu
2023-02-23 15:04:46 <minimal> WhyNotHugo: I tried UKI with Edge kernel on Virtualbox and Virtualbox gives a critical error when the UKI is loaded, however it works ok with linux-lts kernel
2023-02-23 15:05:09 <minimal> not sure if this is a VirtualBox bug or related to the problem you are having
2023-02-23 15:05:20 <minimal> I can try QEMU next
2023-02-23 17:11:41 <oleovit> q: is linux-pam really a completely necessary dependency for swaylock? their git "says", pam is optional. (if no pam, one has to set the suid-bit for swaylock). 
2023-02-23 17:12:49 <psykose> yes
2023-02-23 17:13:51 <oleovit> for security reasons?
2023-02-23 17:15:19 <ikke> we want to avoid suid binaries
2023-02-23 17:16:00 <oleovit> thx psykose/ikke!
2023-02-23 17:16:46 <grawity> would be neat if something like seatd but for local auth existed
2023-02-23 17:19:12 <elly> like, a daemon that you could run as root and authenticate to?
2023-02-23 17:22:25 <ayushnix> ikke: are user namespaces preferable to suid binaries?
2023-02-23 17:24:36 <ikke> Alpine does not have unprivileged user namespaces
2023-02-23 17:25:06 <psykose> you sure
2023-02-23 17:26:43 <ikke> didn't that require a patch?
2023-02-23 17:26:54 <psykose> no, the patch did the opposite
2023-02-23 17:26:59 <psykose> it allowed disabling them
2023-02-23 17:27:03 <psykose> if you mean this https://github.com/archlinux/linux/commit/8ebb376b74c77cc4659e421f48f09ca848b5911c
2023-02-23 17:27:09 <psykose> but the default is allow
2023-02-23 17:27:30 <psykose> and with no patches it's allow pretty sure
2023-02-23 17:38:03 <ikke> Ok, misunderstood it then
2023-02-23 21:12:35 <scorpion2185[m]1> is there something like srm? secure-delete
2023-02-23 21:16:16 <psykose> no
2023-02-23 21:16:40 <XFaCiEer> But what if we imagine it really really hard?
2023-02-23 21:16:51 <psykose> i don't think flash storage fundamentally even lets you do that
2023-02-23 21:17:07 <psykose> aside from secure erasing the whole drive
2023-02-23 21:17:33 <psykose> if you want to hide something then encrypt your drive and it ~mostly hides it anyway
2023-02-23 21:18:46 <scorpion2185[m]1> but if at some point put something on a file and want to secure delete it?
2023-02-23 21:18:58 <scorpion2185[m]1> *you put
2023-02-23 21:19:28 <psykose> already answered that above
2023-02-23 21:19:32 <psykose> you cannot do that on flash storage
2023-02-23 21:19:43 <psykose> on an hdd probably maybe
2023-02-23 21:20:07 <psykose> and if you encrypt the whole drive then it shouldn't really matter anyway
2023-02-23 21:20:44 <scorpion2185[m]1> http://srm.sourceforge.net/
2023-02-23 21:20:55 <scorpion2185[m]1> app that usually distros have
2023-02-23 21:21:19 <scorpion2185[m]1> >A secure replacement for rm(1) that overwrites data before unlinking
2023-02-23 21:22:03 <psykose> that's nice
2023-02-23 21:22:19 <psykose> it doesn't actually do anything on flash storage, but that's nice
2023-02-23 21:24:40 <scorpion2185[m]1> i am looking flash storage. is it used by some SSDs too?
2023-02-23 21:27:03 <scorpion2185[m]1> NAND is flash storage then right?
2023-02-23 21:27:17 <psykose> nand flash is one type yes
2023-02-23 21:27:53 <psykose> every ssd, flash drive, sd card, emmc, ..
2023-02-23 21:31:43 <XFaCiEer> What size of hammer do you recommend for flash storage?
2023-02-23 21:34:41 <abby> ball-peen
2023-02-23 21:39:07 <psykose> is ball-peen a large size or a small size
2023-02-23 21:40:22 <abby> various sizes
2023-02-23 22:04:24 <lucidiot> should you be hammered while using it
2023-02-24 00:49:40 <qaqland[m]> Bad ones taken from nvme or sata are made for nano flash.
2023-02-24 00:49:53 <elly> you just have to use disk crypto and arrange to be able to secure erase the keys... I think psykose is right and you basically cannot reliably erase from flash
2023-02-24 01:23:21 <mystique> hey there.. dnsdist package.. has no openrc files.. is there an alpine way to deploy it? or just write my own openrc files? 
2023-02-24 01:51:52 <mystique> the s6 docs for alpine are sparse.. why is there nothing complete on the wiki :/ 
2023-02-24 02:05:07 <XFaCiEer> mystique: you ARE the wiki
2023-02-24 02:07:52 <mystique> agreed.. what is anything times zero? is that the wiki? seems like it. 
2023-02-24 03:39:13 <mbx> quick question: when I install a specific version of a package, will it be locked, so that the version won't change when I call apk upgrade?
2023-02-24 03:53:32 <psykose> sure
2023-02-24 04:05:20 <mbx> psykose: great, how does that work? If I decide to install nodejs~=18, will it upgrade only minor versions?
2023-02-24 04:05:32 <mbx> This would be a killer feature! :)
2023-02-24 04:05:47 <psykose> sure, 18.*
2023-02-24 04:05:54 <mbx> that's so good
2023-02-24 04:06:04 <mbx> thx
2023-02-24 04:06:44 <psykose> it will stay around forever though, so eventually when nodejs is 20 it will just not upgrade nodejs, forever
2023-02-24 04:07:21 <mbx> yes, that's what it should do then
2023-02-24 04:07:56 <mbx> it's all about getting security patches without nuking the app
2023-02-24 04:08:05 <mbx> service*
2023-02-24 04:08:37 <psykose> if nodejs became 20, 18 wouldn't exist anymore so it wouldn't get any updates
2023-02-24 04:08:53 <psykose> there's no multiple node versions (well, there's also -current but that doesn't matter here)
2023-02-24 04:09:52 <psykose> i.e. this isn't really very useful, because in stable the nodejs won't change major versions anyway in the same release, so it doesn't mean anything
2023-02-24 04:10:20 <psykose> and on edge you would just have... outdated nodejs that holds back every other thing on the system due to library linking, forever, until you remove the ~18
2023-02-24 04:10:52 <psykose> it doesn't save you from anything except some time in X months when nodejs upgrades to 20 on edge and you used edge and didn't read what upgraded
2023-02-24 04:11:33 <psykose> and even in that case you would want to know and manually upgrade because you don't want to keep the 18 version around, because it doesn't exist anymore
2023-02-24 04:15:04 <mbx> psykose: ahh, you are right, oh, that's not so brilliant then. While the newer LTS is active, the older is still on maintenance. 
2023-02-24 04:15:19 <psykose> yes, we don't ship multiple nodejs versions (well, -current aside)
2023-02-24 04:16:39 <mbx> it's very unfortunate. If 18 is still maintained, I want to get security patches for them. I thought you would update new minor versions of older packages sometimes too
2023-02-24 04:16:52 <psykose> it does get patches
2023-02-24 04:17:21 <mbx> I was giving nvm a shot, but the binaries from nodejs.org/dist aren't compatible with Alpine Linux ..
2023-02-24 04:17:39 <psykose> i.e. there is one version of nodejs, but it does get updated in stable branches
2023-02-24 04:18:03 <mbx> psykose: ok confused. If version 20 is out and version 18 gets a new release, will you add a new version 18 to apk or not?
2023-02-24 04:18:26 <psykose> each release branch is distinct and has nothing to do with the other ones
2023-02-24 04:18:39 <psykose> the version in 3.17 is 18, it gets updated, it stays on 18
2023-02-24 04:19:06 <psykose> the version in edge is whatever the latest lts is, so when 20 comes out it's 20
2023-02-24 04:19:33 <psykose> when a release gets made whatever version it was is what version stable gets
2023-02-24 04:22:10 <mbx> alright, so now it makes sense. If I want to preserve my LTS version, I have to make sure to stay on the last branch that this version is used on.
2023-02-24 04:29:15 <mbx> psykose: ok I falsely thought I can install older versions of nodejs in the edge branch. so apk add nodejs~=18 makes no sense, it's better to pick the current branch and just install nodejs.
2023-02-24 04:29:41 <psykose> indeed
2023-02-24 04:29:42 <mbx> kinda wish nvm would work out of the box here
2023-02-24 04:30:47 <mbx> or can I forcefully install from and older branch without being on that branch exclusively?
2023-02-24 04:32:50 <psykose> no
2023-02-24 04:32:51 <mbx> oh I could just add older branches as repositories, so that I can access older versions of the same packages
2023-02-24 04:32:54 <mbx> no?
2023-02-24 04:34:08 <psykose> can't mix branches
2023-02-24 04:37:46 <mbx> ye, it conflics, this is not nixos
2023-02-24 04:39:04 <mbx> i guess official node binaries fail due to musl, really need to get rid of ancient dependencies
2023-02-24 11:30:56 <ktprograms> I'd like to use PiHole to forward .local DNS queries to resolving using mDNS from devices which don't support mDNS (PiHole just because I already have it used as my DNS server). However, it seems that musl doesn't support nsswitch, so resolving mDNS using Avahi isn't possible. Is there an alternative to this, or would I need to manually implement this (perhaps as a PiHole plugin
2023-02-24 11:30:59 <ktprograms> so neither side needs to interact with the system DNS resolvers)?
2023-02-24 11:33:44 <psykose> nope, you'd have to implement a local dns that works with it
2023-02-24 11:34:04 <psykose> i.e. you put 127.0.0.1 in resolv.conf and then the thing you run that listens on 53 handles both dns and .local to make it work
2023-02-24 12:01:04 <ktprograms> Would using dnsmasq work? Based on https://superuser.com/a/1614564 it seems like I can get it to use the system for everything else but resolve .local using a localhost:53 dns server?
2023-02-24 12:04:12 <psykose> whadya mean the system for everything else
2023-02-24 12:11:47 <ktprograms> oh hmm... with PiHole when you set the upstream dns does it use resolv.conf?
2023-02-24 12:12:14 <ktprograms> Anyway, thanks, I'll mess around with it. at least now I know there isn't an easier way.
2023-02-24 12:12:23 <psykose> no
2023-02-24 12:12:40 <psykose> well
2023-02-24 12:12:44 <psykose> idk how they implement it
2023-02-24 12:12:48 <psykose> 99% no, wouldn't make sense
2023-02-24 12:13:20 <ktprograms> So I might have to set the upstream dns to localhost in PiHole, then in my local server forward unknown requests to 8.8.8.8 & friends right?
2023-02-24 12:13:36 <psykose> isn't pihole the dns
2023-02-24 12:13:49 <psykose> why would it connect to localhost to another dns
2023-02-24 12:14:35 <ktprograms> pihole isn't recursive it just filters stuff then forwards it to 8.8.8.8
2023-02-24 12:15:13 <ktprograms> at least afaik, could be completely wrong here
2023-02-24 12:41:49 <boombim[m]> 
ACTION uploaded an image: (150KiB) < https://matrix.org/_matrix/media/v3/download/matrix.org/dOOGIYYkHXDpXsgCjmVuFBSK/image_picker_F9814EFD-6B45-44DF-8915-BA47A6445CA2-4383-00000210A7DAE749.jpg >

2023-02-24 12:41:55 <boombim[m]> Do you guys know how to fix it?
2023-02-24 12:43:09 <psykose> didn't you have this same issue before and it was mixing edge with stable or something
2023-02-24 12:43:21 <psykose> anyway post apk policy dbus dbus-x11 dbus-libs
2023-02-24 12:43:45 <boombim> I haven't. I've just installed dbus-x11
2023-02-24 12:44:26 <psykose> you should also use dbus-run-session and not dbus-launch anyway, but that's unrelated
2023-02-24 12:45:29 <boombim[m]> https://0x0.st/Hsxa.txt
2023-02-24 12:45:42 <psykose> well
2023-02-24 12:45:49 <psykose> you have some completely random 1.15.0 dbus-libs installed
2023-02-24 12:46:18 <psykose> apk upgrade -Ua
2023-02-24 12:46:42 <boombim> psykose: could you please explain somehow why dbus-run-session over dbus-launch and why?
2023-02-24 12:46:56 <boombim> and what are the differences*
2023-02-24 12:48:39 <psykose> it doesn't exit when sway exits and stays around forever like that
2023-02-24 12:49:03 <psykose> there's --exit-with-session but it doesn't really work
2023-02-24 12:49:08 <psykose> dbus-run-session just does what you want
2023-02-24 12:49:54 <boombim> got it. I've done command to upgrade and instad of upgrade I get downgraded dbus to 1.14 :)
2023-02-24 12:54:32 <psykose> if you are on edge you have to use -a here and there because downgrades happen
2023-02-24 12:56:21 <boombim[m]> I've deleted dbus-x11
2023-02-24 12:56:22 <boombim[m]> How to upgrade dbus to 1.16 again?
2023-02-24 12:57:26 <psykose> there is no dbus 1.16
2023-02-24 12:58:02 <boombim[m]> Oh, sorry. I got you wrong
2023-02-24 13:12:43 <boombim> is it ok that I have 5 dbus instances after dbus-run-session sway?
2023-02-24 13:12:45 <boombim> https://0x0.st/Hsxf.txt
2023-02-24 13:13:31 <WhyNotHugo> boombim:Try ps aux | grep dbus.
2023-02-24 13:13:36 <WhyNotHugo> You'll see that one is the system one.
2023-02-24 13:13:43 <psykose> one of them is a relic of dbus-launch
2023-02-24 13:14:04 <psykose> probably
2023-02-24 13:14:08 <psykose> by default you'd have 4
2023-02-24 13:14:11 <WhyNotHugo> dbus-run-session spawns a dbus-daemon which itself has a dbus-daemon child.
2023-02-24 13:14:43 <boombim> so is it totally normal?
2023-02-24 13:14:44 <boombim> https://0x0.st/HsxW.txt
2023-02-24 13:15:37 <psykose> yea
2023-02-24 13:15:54 <WhyNotHugo> I think 3375 is a leftover one.
2023-02-24 13:16:50 <psykose> reboot and find out
2023-02-24 13:16:51 <psykose> hehe
2023-02-24 13:17:46 <boombim> let me try. I'm trying to start sway with emptty service
2023-02-24 13:19:58 <boombim> I have pretty same after reboot
2023-02-24 13:19:59 <boombim> https://0x0.st/Hsxy.txt
2023-02-24 13:29:05 <WhyNotHugo> You have one more than me: https://paste.sr.ht/blob/408ecd0ff65936a284231317b34c19f78ee61e05
2023-02-24 13:32:07 <boombim> Got it. I don't know is it ok or not...
2023-02-24 13:32:26 <boombim> Do you guys use emptty to run your WM?
2023-02-24 13:33:19 <boombim> It's strange becase sway works fine from common tty but its keybinds doesnt work with emptty...
2023-02-24 13:34:12 <WhyNotHugo> https://github.com/tvrzna/emptty indicates that it starts a dbus session by default.
2023-02-24 13:35:23 <boombim> I've removed dbux-x11 so it cannot run dbus-launch I guess
2023-02-24 13:36:11 <WhyNotHugo> Use dbus-run-session instead.
2023-02-24 13:36:18 <WhyNotHugo> It makes sure to exit when the child dies.
2023-02-24 13:36:29 <boombim> I do. I have this in config
2023-02-24 13:37:15 <boombim> https://0x0.st/Hsxx.txt
2023-02-24 16:31:12 <qaqland[m]> https://github.com/KDAB/codebrowser/issues/107 There is difference between alpine and other distro in llvm-dev that normal users can't follow readme to compile it. Could alpine package this software?
2023-02-24 16:31:30 <qaqland[m]> * compile it easily. Could
2023-02-24 16:36:06 <psykose> -DCLANG_BUILTIN_HEADERS_DIR=/usr/include or /usr/include/clang or whatever
2023-02-24 16:36:24 <psykose> this define does not exist, the cmake the project is using is wrong
2023-02-24 16:36:50 <psykose> they look up clang and then don't use the actual defined include dirs but some random shit named that instead, and if it's not defined (it doesn't exist) it falls back to llvm include
2023-02-24 16:38:07 <psykose> ah it was the other place
2023-02-24 16:40:37 <psykose> that's just the default includedir
2023-02-24 16:41:05 <psykose> it's /usr/lib/clang/$version/include/, but if you compile with clang that is just implicit
2023-02-24 16:41:10 <psykose> what is the point of even checking for that
2023-02-24 16:48:20 <psykose> annoying, indeed clang cmake doesn't define anything
2023-02-24 16:48:40 <psykose> have fun tho https://img.ayaya.dev/KfdT7MrVcyl3
2023-02-24 16:48:52 <psykose> well
2023-02-24 16:48:55 <psykose> remove the log
2023-02-24 16:49:16 <psykose> the first part is required and correct, the second is a hack
2023-02-24 19:17:56 <boombim[m]> 
ACTION uploaded an image: (115KiB) < https://matrix.org/_matrix/media/v3/download/matrix.org/qoXvrzLHBfJuLgjciLTrOCbM/image_picker_B0FC2BD1-0673-4C61-8515-753D83D5F85E-6148-0000026A42311133.jpg >

2023-02-24 19:18:06 <boombim[m]> It's greetd with tuigreet
2023-02-24 19:18:20 <boombim[m]> Should I open bug report?
2023-02-24 19:19:09 <psykose> which vt did you configure it to
2023-02-24 19:20:23 <boombim[m]> I guess default. How to check it? I just add tuigreet as cmf
2023-02-24 19:20:29 <boombim[m]> Cmd*
2023-02-24 19:20:54 <psykose> vt = in /etc/greetd/config.toml
2023-02-24 19:21:46 <boombim[m]> vt = 1
2023-02-24 19:22:02 <psykose> well
2023-02-24 19:22:13 <psykose> don't put it on the same vt as where there is a bunch of bootup logs
2023-02-24 19:22:41 <psykose> that said, maybe it is a bug anyway :p
2023-02-24 19:22:42 <psykose> do test
2023-02-24 19:22:56 <boombim[m]> Oh. Got it. I never know about this thing. I'll try 2 :D
2023-02-24 19:23:03 <psykose> the output looks like uhh
2023-02-24 19:23:09 <psykose> greetd starts
2023-02-24 19:23:12 <psykose> tuigreet starts
2023-02-24 19:23:17 <psykose> it cleared the vt1
2023-02-24 19:23:22 <psykose> and then, stuff started after
2023-02-24 19:23:28 <psykose> and still.. printed to vt1 (obv)
2023-02-24 19:23:31 <psykose> so it looks like that
2023-02-24 19:23:34 <psykose> any other vt should look fine
2023-02-24 19:24:55 <boombim[m]> You are absolutely right
2023-02-24 19:25:10 <boombim[m]> Thanks a lot
2023-02-24 19:25:38 <psykose> hm
2023-02-24 19:25:45 <psykose> i guess we should default to not-1 perhaps anyway
2023-02-24 19:25:53 <psykose> or is that a bad idea
2023-02-24 19:26:25 <boombim[m]> I've set it to "next"
2023-02-24 19:26:31 <boombim[m]> And it looks ok
2023-02-24 20:12:11 <bleb> haven't been able to get sound on my new alpine installation
2023-02-24 20:12:37 <bleb> alsamixer shows ALSA lib confmisc.c:855:(parse_card) cannot find card '0'
2023-02-24 20:12:50 <bleb> followed by some more errors
2023-02-24 20:13:28 <bleb> lspci|grep -i audio shows 00:1f.3 Multimedia audio controller: Intel Corporation Alder Lake PCH-P High Definition Audio Controller (rev 01)
2023-02-24 20:15:30 <psykose> alder lake makes me guess you might need sof-firmware
2023-02-24 20:16:05 <bleb> dmesg has about a dozen lines beginning with sof-audio-pci-intel-tgl but none of them are warnings/errors
2023-02-24 20:16:20 <bleb> I think I already had sof-firmware
2023-02-24 20:16:36 <psykose> ah
2023-02-24 20:16:42 <psykose> only the new version apparently added alder
2023-02-24 20:16:51 <psykose> or something
2023-02-24 20:16:56 <bleb> new version?
2023-02-24 20:17:02 <psykose> of sof-firmware
2023-02-24 20:17:25 <psykose> or maybe not
2023-02-24 20:17:26 <psykose> idk
2023-02-24 20:17:32 <psykose> just going by https://github.com/thesofproject/sof-bin/releases
2023-02-24 20:17:48 <psykose> should be in edge in a few hours, if you want to upgrade and try again then
2023-02-24 20:26:24 <bleb> can I use sof-firmware from edge but keep everything else stable?
2023-02-24 20:26:57 <bleb> I put @edge http://mirrors.edge.kernel.org/alpine/edge/main in /etc/apk/repositories, then apk add sof-firmware@edge
2023-02-24 20:27:05 <bleb> should that do what I want?
2023-02-24 20:27:24 <psykose> it's in community not main
2023-02-24 20:27:37 <psykose> uhh generally that is obviously not supported, but in this case firmware has no dependencies
2023-02-24 20:27:41 <psykose> it should work fine, so go for it
2023-02-24 20:27:54 <psykose> (when it builds in a few hours)
2023-02-24 20:28:09 <bleb> hmmm then why did apk add sof-firmware@edge complete without error
2023-02-24 20:28:22 <psykose> supported in a support sense
2023-02-24 20:28:25 <psykose> not a what can apk do sense
2023-02-24 20:29:04 <bleb> but wouldn't apk complain if sof-firmware@edge was not valid
2023-02-24 20:29:14 <psykose> it's not invalid to apk
2023-02-24 20:29:17 <bleb> i.e. if sof-firmware was not in edge/main
2023-02-24 20:29:47 <psykose> ah
2023-02-24 20:29:49 <bleb> can apk tell if sof-firmware is in edge/main?  I would think yes?
2023-02-24 20:30:53 <bleb> well apk add sof-firmware@edgecommunity also said OK
2023-02-24 20:31:04 <bleb> so I guess I don't know how to interpret when apk says OK
2023-02-24 20:31:27 <bleb> how can sof-firmware@edge and sof-firmware@edgecommunity both be OK?
2023-02-24 20:31:38 <psykose> the latter because it is
2023-02-24 20:31:41 <psykose> the former not sure
2023-02-24 20:31:59 <bleb> is there an apk command that would actually tell me whether the given package was installed from the specified repo?
2023-02-24 20:34:12 <bleb> both edge and v3.17 have version 2.2.2-r0 so no way to tell based on that
2023-02-24 20:35:05 <psykose> ah ok i see
2023-02-24 20:35:12 <psykose> the tag is not a requirement that it has to come from there
2023-02-24 20:35:19 <psykose> it's just preference over the usual greatest version
2023-02-24 20:35:24 <bleb> yeah... apk add sof-firmware@edge and apk add sof-firmware@edgecommunity both print the same thing: OK: 1644 MiB in 479 packages
2023-02-24 20:37:10 <bleb> psykose: and without the tag, it won't search the tagged repos for newer versions right
2023-02-24 20:37:25 <psykose> or older versions, yeah
2023-02-24 20:37:28 <psykose> they're just ignored
2023-02-24 20:37:53 <psykose> when you add the tag it then prefers whatever version is there, and also allows pulling in depends for the thing you added from the same tagged repo (but still prefers non-tagged for the depends)
2023-02-24 20:38:00 <psykose> doesn't really matter but just more knowledge
2023-02-24 20:38:29 <bleb> ok
2023-02-24 20:39:00 <bleb> so sof-firmware v2.2.4 will be in edge/community in a few hours?
2023-02-24 20:39:37 <psykose> eh
2023-02-24 20:39:50 <psykose> it will be in 3.17 in a few minutes
2023-02-24 20:39:53 <psykose> no need for tags
2023-02-24 20:40:10 <bleb> oh nice
2023-02-24 20:40:31 <psykose> read it over a bunch and looks fine to backport
2023-02-24 20:46:47 <bleb> cool
2023-02-24 20:47:35 <psykose> if you use dl-cdn it's there
2023-02-24 20:56:38 <bleb> got it, but still getting the same sound card error when I run alsamixer
2023-02-24 20:56:51 <psykose> you have to reboot/reload the modules/etc
2023-02-24 20:56:55 <psykose> easiest to just reboot
2023-02-24 20:57:03 <psykose> it won't load the new firmware otherwise
2023-02-24 20:57:06 <psykose> aside from that, no idea
2023-02-24 20:57:08 <bleb> fwiw I do see this at boot: alsactl: unrecognized options: -
2023-02-24 20:57:09 <psykose> could be anything else
2023-02-24 20:57:25 <bleb> followed by usage help
2023-02-24 20:57:42 <bleb> yeah I did reboot
2023-02-24 21:02:18 <bleb> what's the difference between sof-bin and sof-firmware?
2023-02-24 21:03:29 <chereskata> Hello, i can not play an mp4 video with totem on alpine linux edge:Failed to load grl-lua-factory plugin: Plugin “grl-lua-factory” not available 
2023-02-24 21:04:06 <chereskata> Does anyone know how to fix the issue
2023-02-24 21:04:09 <chereskata> ?
2023-02-24 21:04:41 <psykose> bleb: not sure
2023-02-24 21:04:57 <chereskata> Missing plugin: gstreamer|1.0|totem|H.264 (Main Profile) decoder|decoder-video/x-h264, level=(string)3.1, profile=(string)main, coded-picture-structure=(string)frame, chroma-format=(string)4:2:0, bit-depth-luma=(uint)8, bit-depth-chroma=(uint)8, colorimetry=(string)bt709 (H.264 (Main Profile) decoder)
2023-02-24 21:09:30 <psykose> should be fixed
2023-02-24 21:09:38 <psykose> soon :tm"
2023-02-24 21:14:01 <chereskata> psykose: thanks, is this for me :)?
2023-02-24 21:14:09 <psykose> ye
2023-02-24 21:14:17 <chereskata> ty again :)
2023-02-24 22:11:07 <chereskata> Thank you a lot for fixing the thing so fast!
2023-02-24 22:11:19 <chereskata> Appreciate all your work @world
2023-02-24 22:24:50 <invoked> another satisfied customer
2023-02-24 22:24:58 <invoked> 
ACTION rings a bell

2023-02-24 22:25:01 <chereskata> :D
2023-02-24 22:25:16 <chereskata> actually i add a lot of mappings to repology
2023-02-24 22:25:23 <chereskata> *release-monitoring
2023-02-24 22:25:42 <chereskata> Every package i install is added to it shortly
2023-02-24 22:32:21 <derfman> Good After morning all.
2023-02-24 22:34:19 <boombim[m]> Hello. Recommend me please good font for console
2023-02-24 22:34:53 <boombim[m]> I've installed and set terminus but looks like it has artefacts with ncurses
2023-02-24 22:37:03 <umk3[m]> font-noto is fine for me, notosansmono
2023-02-24 22:38:46 <boombim[m]> Does it work in tty?
2023-02-24 22:38:59 <boombim[m]> Vt
2023-02-24 22:41:53 <psykose> what artifacts
2023-02-24 22:52:26 <boombim[m]> Let me show it
2023-02-24 22:54:17 <boombim[m]> 
ACTION uploaded an image: (88KiB) < https://matrix.org/_matrix/media/v3/download/matrix.org/TrIlFpxeZsyDrSnQJenwxmvH/image_picker_EFF93378-8360-4021-BC61-CDF25E2F9FA9-7456-00000297ED6FF1AB.jpg >

2023-02-24 22:54:26 <boombim[m]> I get this instead of lines
2023-02-24 22:55:19 <boombim[m]> I've added consolefont service to boot runlevel
2023-02-24 22:55:39 <boombim[m]> I get a lot of artefacts even with fzf widget
2023-02-24 22:56:30 <boombim[m]> However if I don't use consolefont service and set font manually with setfont I don't see artefacts
2023-02-24 23:43:57 <derfman2> I'm trying to get around an issue with XRDP setup (its giving me an "error connecting to user session') I have danced with not quite right vnc setups in the past and not enjoyed it.
2023-02-24 23:43:58 <derfman2> Anyway I thought I could doge the issue by using chromium and the chrome remote desktop tool.   chromium is a pacgage listed in https://pkgs.alpinelinux.org/  but apk will not see it for some reason. Can anyone help me understand why please? I have tried edge and also testing repos just to be sure but the pkgs pages says edge community. ! 
2023-02-24 23:44:41 <elly> what does `apk info chromium` say?
2023-02-24 23:47:45 <derfman2>  # apk add chromium
2023-02-24 23:47:45 <derfman2> ERROR: unable to select packages:
2023-02-24 23:47:45 <derfman2>   chromium (no such package):
2023-02-24 23:47:45 <derfman2>     required by: world[chromium]
2023-02-24 23:48:01 <derfman2> sry that did not past well 
2023-02-24 23:49:12 <derfman2> #/media/cdrom/apks
2023-02-24 23:49:12 <derfman2> #http://eu.edge.kernel.org/alpine/v3.17/main
2023-02-24 23:49:13 <derfman2> #http://eu.edge.kernel.org/alpine/v3.17/community
2023-02-24 23:49:13 <derfman2> http://eu.edge.kernel.org/alpine/edge/main
2023-02-24 23:49:13 <derfman2> http://eu.edge.kernel.org/alpine/edge/community
2023-02-24 23:50:25 <derfman2> I think my irc client is reading carriage returns and new lines. sry for the mess 
2023-02-24 23:50:37 <elly> hm
2023-02-24 23:50:44 <elly> what arch are you on?
2023-02-24 23:51:03 <derfman2> That is in VirtBox 
2023-02-24 23:51:04 <elly> I see a chromium package in that edge/community repo
2023-02-24 23:51:19 <elly> right but if you do `uname -m` what do you get
2023-02-24 23:51:52 <derfman2> -m 
2023-02-24 23:51:52 <derfman2> i686
2023-02-24 23:52:33 <elly> aha
2023-02-24 23:52:41 <elly> chromium is only packaged for x86-64, not regular x86
2023-02-24 23:52:53 <elly> you are SOL unfortunately, chromium does not run on 32-bit x86
2023-02-24 23:52:55 <derfman2> intersting. I was thingking I was on x86_amd64. If im not wring i686 is 32 bit
2023-02-24 23:54:19 <elly> it is, yes
2023-02-24 23:54:32 <elly> it may just be that your VM is 32-bit, not sure
2023-02-24 23:57:48 <derf_> TY @elly.. appreciate you showing me the error of my ways. :-) 
2023-02-25 00:11:28 <elly> sure thing
2023-02-25 00:15:03 <derfman2> Here is another avenue I looked at for remote desktop. RustDesk. It looks really interesting but as yet I cant see a client for Alpine. Anyone happen to have any experience of ideas on running RustDesk client ( not the server) ? 
2023-02-25 04:10:23 <uuidNuni1> derfman2: compile yourself
2023-02-25 04:11:28 <uuidNuni1> using cargo-zigbuuld ^
2023-02-25 04:52:55 <Saijin_Naib> <derfman2> "Here is another avenue I..." <- Works great in FlatPak. They have the FlatPak binaries as build objects on their GitHub
2023-02-25 04:53:33 <Saijin_Naib> Even self-hosting, I've not found it to be any faster or nicer than VNC, which was a massive disappointment. RDP really seems to be the best thing going by far
2023-02-25 05:01:15 <qaqland[m]> psykose: thx! can I use the diff to comment in issue
2023-02-25 09:15:52 <derfman2> Saijin_Naib and uuidNunil  Thanks for the ideas. Compiling is not something I have done much but thats no reason to not take a look. Flatpack sounds like another good option.   I would prefer xRDP per the alpine wiki but I get a vnc error and thought I might just look for something less old and tangled with pain. I use it very comfortable with another desktop but its a debian style distro.   If the performance is bad with RustDesk then humm  !! 
2023-02-25 09:15:52 <derfman2> Thanks for the input
2023-02-25 09:21:08 <paper_> Rust desk looks shady, there was a package request to Void and someone looked into it a bit: https://github.com/void-linux/void-packages/issues/41107
2023-02-25 09:40:43 <Apachez> what do you mean by an opensource project is shady? ;-)
2023-02-25 09:42:09 <Apachez> so main question is if a bouncer/proxy is being used and if so if that will peak into the stream or not
2023-02-25 09:42:22 <Apachez> if you have direct connections then you can protect the stream yourself using ssh
2023-02-25 09:43:14 <Apachez> for example like so: https://www.questioncomputer.com/ssh-ip-vpn-tunnel-on-ubuntu-20-04/
2023-02-25 09:50:40 <paper_> one day all connections suddenly don't have encryption and they don't even see it as a vulnerability
2023-02-25 09:51:05 <paper_> and one of the selling points of rustdesk is that you don't have to have a p2p connection for it to work
2023-02-25 09:51:13 <paper_> opensource projects can be also shady
2023-02-25 09:57:21 <lopid> looks to me like the main selling point is the language it is written in, which says a lot
2023-02-25 12:11:09 <boombim[m]> https://wiki.alpinelinux.org/wiki/Fonts#Changing_the_console_font
2023-02-25 12:11:09 <boombim[m]> I get a lot of artefacts with this method unfortunately.
2023-02-25 12:11:09 <boombim[m]> Is it still relevant guide?
2023-02-25 14:31:55 <psykose> i get the same ones, forgot how to fix it
2023-02-25 15:42:20 <derfman> Good after morning. Is anyone here familiar with the xrdp and vino setup? Im getting an error. I get the xrdp password prompt but then a no connection response. Possibly someone knows of an alternative "how to" instead of the main wiki one.
2023-02-25 16:07:18 <yann-kaelig> Hello. Where can I find a documention about how check option in apkbuild is working. For example in automake options="!check" but how the test is disable, that what I'm looking for
2023-02-25 16:08:58 <minimal> yann-kaelig: that means it doesn't run any "check" function defined in the APKBUILD file
2023-02-25 16:10:23 <WhyNotHugo> Is there some tool that prints the info from /proc/X/stat in tabular format (e.g.: with labels for each field)?
2023-02-25 16:11:31 <leah2> prtstat from psmisc maybe
2023-02-25 16:11:59 <minimal> yann-kaelig: https://wiki.alpinelinux.org/wiki/APKBUILD_Reference#options
2023-02-25 16:20:07 <WhyNotHugo> leah2: Thanks
2023-02-25 16:20:48 <leah2> not sure it's complete
2023-02-25 16:23:54 <WhyNotHugo> I'm not sure the CPU times are correct.
2023-02-25 16:23:58 <wej> on alpine 3.17, what do i need to do to install xorg? when i try to install xorg-server, apk complains about missing packages glamor-egl and mesa-dri-intel. this wasn't an issue on 3.16
2023-02-25 16:25:21 <psykose> can you paste the console output
2023-02-25 16:25:28 <minimal> WhyNotHugo: did you get any further with your linux-edge UKI issue?
2023-02-25 16:26:36 <WhyNotHugo> minimal: Haven't had the time to dig further.
2023-02-25 16:27:10 <wej> it says: ERROR: unable to select packages: glamor-egl (no such package) required by: world[glamor-egl]
2023-02-25 16:27:16 <WhyNotHugo> Well, I did set up gummiboot so I can quickly build UKIs and then reboot again into a previous one if it fails. But no real debugging done yet.
2023-02-25 16:27:23 <psykose> then don't add glamor-egl?
2023-02-25 16:27:33 <psykose> i asked for the whole console, not 1 line
2023-02-25 16:28:34 <wej> all i did was run "apk add xorg-server", the only output was what i just wrote and the same error for mesa-dri-intel
2023-02-25 16:28:50 <wej> sorry, can't easily paste the error since it is on a different physical machine
2023-02-25 16:28:51 <minimal> WhyNotHugo: ok, am still investigating myself, I have an idea to try...
2023-02-25 16:29:24 <psykose> `xorg-server` in 3.17 doesn't depend on glamor-egl and installs fine, so your repositories/world are probably broken in some other way
2023-02-25 16:29:27 <psykose> pastebin both
2023-02-25 16:29:58 <psykose> nothing in any release depends on `glamor-egl` either
2023-02-25 16:30:04 <wej> huh, that's bizarre. i'll have to check what's going on there then i guess
2023-02-25 16:38:17 <wej> aha, i had those packages still in my world file and didn't notice it. that explains why apk wanted to install them. sorry for bothering you with this
2023-02-25 16:57:38 <benth0r> Which method would folks recommend to remote mount a directory from an alpine host on an archlinux client? Folks have been warning me of NFS, but what else is there besides SSHFS or Samba? (People keep suggesting samba, but I hesitate to use anything that was at least partially dreamt up in redmont)
2023-02-25 16:58:27 <benth0r> The context is my home network btw, traffic doesn't need to go over the internet
2023-02-25 16:58:43 <quinq> There's 9p
2023-02-25 16:59:50 <benth0r> oooh
2023-02-25 17:00:24 <benth0r> <3
2023-02-25 17:02:11 <psykose> either sshfs or nfs
2023-02-25 17:02:51 <quinq> That's not something else than nfs, sshfs or samba though
2023-02-25 17:02:54 <benth0r> I'll go with plan9port for now
2023-02-25 17:07:37 <psykose> indeed it is not something else
2023-02-25 17:08:07 <psykose> it also doesn't involve using poorly ported shitty plan9 utilities, so there's that
2023-02-25 17:08:09 <psykose> big upside imo
2023-02-25 17:14:15 <benth0r> client side, v9fs is working. Now to have something on the server side...
2023-02-25 17:29:19 <benth0r> quinq: any smart idea for a decent 9p file server implementation
2023-02-25 17:29:39 <JohannesJ[m]> 9p hmmz? Need to look into that
2023-02-25 17:36:42 <JohannesJ[m]> On the other hand, maybe i’ll stick to NFS
2023-02-25 17:36:48 <quinq> Nope, I just gave an alternative, as requested
2023-02-25 17:36:54 <quinq> Personnaly I'm mostly using sshfs
2023-02-25 17:37:38 <JohannesJ[m]> Yeah, i suppose i should look into that :)
2023-02-25 17:42:40 <JohannesJ[m]> Hmmz, quinq how stable is it? In your experience?
2023-02-25 17:43:37 <quinq> Never had any problem
2023-02-25 17:45:59 <invoked> https://wiki.qemu.org/Documentation/9psetup
2023-02-25 17:46:21 <invoked> i'm not going to help you with it, but that will get you started
2023-02-25 17:48:23 <minimal> isn't there also virtiofs/virtiofsd ?
2023-02-25 17:49:13 <minimal> with Alpine package qemu-virtiofsd for the server
2023-02-25 17:49:43 <JohannesJ[m]> <quinq> "Never had any problem" <- And if i may ask one last question, hows the speed? Since its all encrypted and all, do you know how that performs in comparison to samba etc? 
2023-02-25 17:51:38 <quinq> It's good enough for me
2023-02-25 18:55:49 <boombim[m]> <psykose> "i get the same ones, forgot..." <- Please ping me if you recall how to fix this consolefont issue
2023-02-25 18:56:06 <psykose> i tested some stuff and like any font has the same issue :D
2023-02-25 18:56:13 <psykose> idk what causes it
2023-02-25 18:57:53 <boombim[m]> You can set it manually with setfont and it works fine
2023-02-25 18:58:07 <boombim[m]> But with service it's not
2023-02-25 18:59:06 <psykose> if i setfont the same font or any font it doesn't change anything
2023-02-25 19:16:20 <boombim[m]> <psykose> "if i setfont the same font or..." <- Do you mean bug still reproducible ?
2023-02-25 19:16:34 <psykose> yeah, i haven't found a way to fix it
2023-02-25 19:16:55 <psykose> and i tried like every terminus variant (no idea what the difference is) and unifont and whatever the default is
2023-02-25 19:17:12 <psykose> they all print garbage for the `apk` status or for the special ttylogin chars
2023-02-25 19:17:22 <psykose> i'm starting to think it's not related to the font
2023-02-25 19:17:29 <psykose> but some other console property
2023-02-25 19:49:39 <boombim[m]> Yeah. I think so
2023-02-26 01:38:55 <Saijin_Naib> <paper_> "Rust desk looks shady, there was..." <- Oh, fantastic.
2023-02-26 01:39:13 <Saijin_Naib> Back to banging my head against the desk trying to get XRDP to work under Alpine again
2023-02-26 04:58:30 <dewdude_> i love how i can change the MAC address in an attempt to start using the dhcp reservation but the damn thing just requests the old IP
2023-02-26 06:03:49 <invoked> dewdude_: kill -USR2 on your running 'udhcpc' process should release the current ip
2023-02-26 09:56:19 <frag> is xmodmap better than xkb? seems eg i can use altgr+h for arrow key with xmodmap but not with xkb (eg in terminal)
2023-02-26 09:57:40 <lopid> at console?
2023-02-26 09:58:06 <frag> lopid: im in xorg
2023-02-26 09:58:28 <lopid> i think xmodmap is legacy now
2023-02-26 10:00:06 <frag> yeah, too bad xkb doesnt work that well, it is a lot faster tho
2023-02-26 10:00:17 <lopid> it worked fine for me
2023-02-26 10:01:49 <frag> lopid: oh yeah? how did you do it?
2023-02-26 10:02:11 <frag> i use: key <AC07>{[ j, J,down,next]};
2023-02-26 10:02:44 <lopid> initially with a config under ~/.xkb (iirc), but i think udev reset it when a new keyboard was plugged in, so i now use /usr/share/X11/xkb
2023-02-26 10:05:11 <lopid> /usr/share/X11/xkb/symbols/mine: https://privatebin.net/?166bef1890068d4d#67Lw3YRuWBLZmmFZYdnQJMqZd7sA7oiWMnsinJ74xfRX  then add ｢include "mine"｣ at the end of the "basic" section in /usr/share/X11/xkb/symbols/gb. gb being my keyboard layout
2023-02-26 10:06:28 <frag> omg, maybe just need to capitalize first char in left/right etc
2023-02-26 10:07:31 <lopid> keysymdef.h has the symbols
2023-02-26 10:08:01 <lopid> uparrow, downarrow, etc
2023-02-26 10:08:26 <frag> xkb so complex :X
2023-02-26 10:09:10 <frag> well seems to work now, will see after reboot
2023-02-26 10:09:48 <lopid> just remember to redo that file when a package updates it
2023-02-26 10:10:26 <lopid> or use ~/.xkb 
2023-02-26 10:35:54 <frag> lopid: hm? you can cp /usr/share/X11/xkb/symbols/us to  ~/.xkb ??
2023-02-26 10:36:58 <lopid> not exactly, but you can have a user config instead of modifying the system files
2023-02-26 10:37:30 <lopid> like i said, it can reset back to default if a usb keyboard if plugged out and in
2023-02-26 10:37:55 <lopid> it should be fine if you don't do that
2023-02-26 10:40:19 <frag> yes, i remove us and put a symlink there instead, but yeah, annoying that you have to fix it after update
2023-02-26 10:40:53 <lopid> that might still be replaced
2023-02-26 10:41:01 <frag> maybe i should upgrade once a q instaed of once a month hehe
2023-02-26 10:47:15 <psykose> or just put it in the correct file
2023-02-26 10:54:56 <frag> psykose: which is?
2023-02-26 10:56:09 <psykose> in a file in ~/.xkb
2023-02-26 10:57:52 <frag> ookok
2023-02-26 12:09:03 <[][l0][][m]> how can i fix broken packages ,it cause failure during installation
2023-02-26 12:10:35 <ikke> [][l0][][m]: what does `apk fix` return?
2023-02-26 12:14:00 <[][l0][][m]> ikke: Error : unable to select packages
2023-02-26 12:14:11 <ikke> complete output please
2023-02-26 12:32:28 <[][l0][][m]> idk y image unable to upload
2023-02-26 12:39:55 <[][l0][][m]> 
ACTION uploaded an image: (212KiB) < https://matrix.org/_matrix/media/v3/download/matrix.org/quQePgdeCjJziJLSXLDKacWV/break.1.png >

2023-02-26 12:39:56 <[][l0][][m]> ERROR: unable to select packages:... (full message at <https://matrix.org/_matrix/media/v3/download/matrix.org/PBdfXFXzBISMpYrtoIyXVMiX>)
2023-02-26 12:42:19 <psykose> what is your /etc/apk/world
2023-02-26 12:43:41 <[][l0][][m]> i dont have world directory
2023-02-26 12:43:54 <ikke> it's a file
2023-02-26 12:44:35 <[][l0][][m]> m sorry yup its a file
2023-02-26 12:45:56 <[][l0][][m]> on world ,there is list of packages name
2023-02-26 12:46:41 <ikke> We are aware
2023-02-26 12:46:53 <psykose> pastebin the file
2023-02-26 12:46:55 <ikke> The specific packages that are there is what matters
2023-02-26 12:50:49 <[][l0][][m]> can i send that as txt file ,right now i dont have any pastebin acc
2023-02-26 13:09:55 <[][l0][][m]> https://pastebin.com/QJNWZZay
2023-02-26 13:40:19 <psykose> does an `apk upgrade -a` fix it
2023-02-26 14:26:47 <[][l0][][m]> it does work but again during installing newer application it throw same error
2023-02-26 14:29:01 <psykose> so it doesn't fix it
2023-02-26 14:29:21 <[][l0][][m]> it doesnt fix
2023-02-26 14:30:10 <psykose> what do you have in /etc/apk/repositories
2023-02-26 14:31:24 <psykose> oh
2023-02-26 14:31:28 <psykose> `apk del mesa-utils`
2023-02-26 14:31:42 <psykose> ah no
2023-02-26 14:31:45 <psykose> misread
2023-02-26 14:31:48 <psykose> thought that didn't exist
2023-02-26 14:33:24 <[][l0][][m]> on repo apline linux cdn
2023-02-26 14:34:17 <ikke> [][l0][][m]: again, details matter
2023-02-26 14:34:38 <ikke> psykose wants to know what specific repositories you have enabled
2023-02-26 14:36:02 <[][l0][][m]> anyways i appreiciate ur time n energy ,i gonna try myself if doesnt work i will in touch 👍️
2023-02-26 14:36:15 <psykose> just paste the file
2023-02-26 14:38:56 <[][l0][][m]> https://pastebin.com/hDbzw88Q
2023-02-26 14:40:15 <psykose> weird
2023-02-26 14:41:16 <psykose> what exactly are you typing to get the error output
2023-02-26 14:41:19 <psykose> since you never posted that
2023-02-26 14:48:05 <[][l0][][m]> sudo apk add 
2023-02-26 14:48:49 <[][l0][][m]> the error persist even on package deletation process so m stuck
2023-02-26 14:49:04 <psykose> can you paste the output of `apk add 'gcc>12.1.1_git20220630-r2'`
2023-02-26 14:55:35 <[][l0][][m]> upgrading gcc in process
2023-02-26 15:04:01 <psykose> and now if you do whatever from before
2023-02-26 15:05:23 <[][l0][][m]> on executing
2023-02-26 15:06:07 <[][l0][][m]> its woking!! yay
2023-02-26 15:06:16 <[][l0][][m]>  * its working!! yay
2023-02-26 15:06:30 <psykose> weird
2023-02-26 15:06:35 <psykose> not sure why it couldn't resolve that
2023-02-26 15:07:01 <[][l0][][m]> what was the problm? with gcc
2023-02-26 15:08:19 <psykose> nothing i don't think
2023-02-26 15:10:30 <[][l0][][m]> thank you so much
2023-02-26 15:10:52 <dewdude_> invoked, I just set the IP as static
2023-02-26 15:11:26 <dewdude_> in the end it was going to effectively be static but letting it get a static lease from DHCP would have avoided some issues
2023-02-26 15:12:09 <dewdude_> i built the VM on one host then copied it to the others
2023-02-26 15:14:14 <dewdude_> i also found out after I did that my DHCP lease entry for that IP was wrong anyway...likely from when I moved the VM around back in October
2023-02-26 15:17:44 <dewdude> i also moved all my alpine vms to edge
2023-02-26 15:17:48 <dewdude> because...why not
2023-02-26 15:18:06 <dewdude> my asterisk isn't but that's because i compile that
2023-02-26 15:58:20 <JohannesJ[m]> edge is like a rolling release on other distro's, no?
2023-02-26 16:03:25 <dewdude> yeah. but that can sometimes bite you
2023-02-26 16:03:33 <qaqland[m]> Johannes J.: Maybe all version are rolling in fact
2023-02-26 16:03:52 <dewdude> yeah. edge is just bleeding edge stuff
2023-02-26 16:04:14 <JohannesJ[m]> hmz
2023-02-26 16:04:17 <dewdude> like last i looked 3.17.2 was still on v5 of kernel; edge is on 6
2023-02-26 16:04:37 <paper_> qaqland[m]: how?
2023-02-26 16:04:39 <dewdude> or maybe not
2023-02-26 16:04:56 <JohannesJ[m]> i always thought the v* versions where like debian stable, and the edge versions where like arch
2023-02-26 16:05:20 <dewdude> oh, it is. i didn't move this vm to edge
2023-02-26 16:06:10 <JohannesJ[m]> im still learning about alpine :)
2023-02-26 16:08:16 <dewdude> there's only been one VM I haven't been able to do under Alpine
2023-02-26 16:08:23 <dewdude> and that's because Jitsi just won't work with it
2023-02-26 16:08:40 <JohannesJ[m]> you have jitsi working on alpine
2023-02-26 16:08:41 <JohannesJ[m]> ?
2023-02-26 16:08:44 <invoked> don't complain if the bleeding edge makes you bleed
2023-02-26 16:08:44 <dewdude> no
2023-02-26 16:08:57 <dewdude> my jitsi is the only machine i can't alpine
2023-02-26 16:09:26 <dewdude> but i have two pi-holes, irssi, rutorrent, and asterisk running under alpine
2023-02-26 16:09:32 <dewdude> and i compile my own asterisk
2023-02-26 16:09:50 <JohannesJ[m]> you package asterisk for alpine?
2023-02-26 16:09:51 <dewdude> i also have a xen-orchestra VM I managed to get working
2023-02-26 16:09:52 <dewdude> no
2023-02-26 16:09:57 <dewdude> i just compile from source
2023-02-26 16:10:10 <JohannesJ[m]> i got asked to look into 3cx, but ive read they are going down hill rather quickly
2023-02-26 16:10:24 <JohannesJ[m]> nice for me.. i really try to keep everything on Alpine :D
2023-02-26 16:10:38 <dewdude> i like the fact i can cut the ram down
2023-02-26 16:10:46 <dewdude> like my asterisk VM only has 512MB
2023-02-26 16:10:51 <dewdude> and it's never hit the limit
2023-02-26 16:10:53 <JohannesJ[m]> yeah, and that its small and more secure
2023-02-26 16:11:09 <dewdude> and i do everything on Xen
2023-02-26 16:11:19 <JohannesJ[m]> i also like that it runs on phones :) Alpinizing my whole life
2023-02-26 16:11:24 <dewdude> i tried doing Xen as a dom0 but had some issues; not to mention I need xapi
2023-02-26 16:11:44 <JohannesJ[m]> ive replaced virtualbox with virtmanager :) 
2023-02-26 16:12:58 <JohannesJ[m]> not sure if it has a webinterface so i can replace proxmox, that would be awesome :)
2023-02-26 16:13:09 <dewdude> someday i'll get alpine as a xen dom0 and see how it does
2023-02-26 16:13:44 <dewdude> but there are things i need with Xen I lose if i drop xcp-ng
2023-02-26 16:14:21 <JohannesJ[m]> :)
2023-02-26 16:15:10 <dewdude> but i mostly got sick of the bloat of ubuntu and debian for a basic VM
2023-02-26 16:15:34 <dewdude> i don't need cloud-init
2023-02-26 16:15:59 <JohannesJ[m]> well, same here :) before Alpine ive used Debian since 1998
2023-02-26 16:18:43 <dewdude> https://img.pickmy.org/g2e422kb.png
2023-02-26 16:20:06 <minimal> JohannesJ[m]: I've used Debian since 1995 ;-)
2023-02-26 16:20:27 <JohannesJ[m]> i was still in school in 1995 :) 
2023-02-26 16:21:37 <psykose> imagine being alive in 1995
2023-02-26 16:22:09 <paper_> I can not
2023-02-26 16:22:25 <psykose> i think paper has been in use for a really long time though
2023-02-26 16:22:28 <psykose> ur lying
2023-02-26 16:22:45 <dewdude> i got my first PC in 1995
2023-02-26 16:23:33 <paper_> I think I got my first PC in 2016 lol
2023-02-26 16:24:01 <JohannesJ[m]> i got one much earlier, it was a tulip, metal case, an 8086 :D ran MS-DOS with a green fluoresent screen :D
2023-02-26 16:24:34 <minimal> got my first in 1992 (or 93?) - 386DX, Adaptec SCSI controller, 5.25" *full height* (that's twice the height of a "standard" 5.25" drive bay/optical drive) HDD, VL-bus graphics card, don't remember how much ram
2023-02-26 16:24:37 <JohannesJ[m]> with a star printer, the whole table would rock back and forth when i was printing :D
2023-02-26 16:24:53 <JohannesJ[m]> amazing days 
2023-02-26 16:25:21 <JohannesJ[m]> no internet, no tiktok
2023-02-26 16:25:24 <minimal> had computers before then, they just were not "PCs", i.e. Spectrum, C64, BBC Micro, etc
2023-02-26 16:26:08 <JohannesJ[m]> before that i only had game consoles
2023-02-26 16:26:24 <minimal> JohannesJ[m]: "no internet" - depends which year you're referring to, I setup an ISP in 1994 ;-)
2023-02-26 16:27:28 <JohannesJ[m]> no before that :) in 1994 we had Windows 3.11 already, i refer to the ms-dos time :)
2023-02-26 16:27:30 <JohannesJ[m]> there was internet, but for us it was unpayable
2023-02-26 16:28:20 <dewdude> i spent 1995 - 1998 jumping from free trial to free trial
2023-02-26 16:29:34 <dewdude> i first put a LAN in this house in 1999
2023-02-26 16:30:16 <dewdude> it was literally two PCs and a cheap hub
2023-02-26 16:30:28 <JohannesJ[m]> ohh yeah i rememver that!
2023-02-26 16:30:48 <dewdude> and primarily done for dial-up sharing because I put the second computer right where the phone line came in the house
2023-02-26 16:30:48 <JohannesJ[m]> using winpopup to send messages to my brother, awesome times
2023-02-26 16:30:49 <minimal> I spent 1994-1997 adding/switching modems from 14.4Kbps to 19.2K, 22.8K, 33.6K :-(
2023-02-26 16:31:25 <JohannesJ[m]> sometimes i long back for those days
2023-02-26 16:31:26 <dewdude> I remember keeping an ISA modem around to do dialup on linux
2023-02-26 16:31:56 <dewdude> heh
2023-02-26 16:32:05 <dewdude> the internet seemed extra magical when you had to connect to it
2023-02-26 16:32:08 <minimal> JohannesJ[m]: "it was unpayable", yeah it cost us approx 47,000 UK pounds per year for a 64Kbps transit Internet connection in 1994 lol
2023-02-26 16:32:13 <dewdude> I got DSL in 2001 and it's just been a part of life since
2023-02-26 16:33:01 <JohannesJ[m]> shame alpine wasnt around back then :D
2023-02-26 16:33:23 <dewdude> it was after that we got more computers and i had to start expanding the network
2023-02-26 16:33:56 <dewdude> i still remember going down to compusa and spending like, $300 on a 802.11b PCMCIA card and WAP
2023-02-26 16:33:58 <minimal> Over the years I've done MCC -> SLS -> Slackware -> Debian -> LFS and homebuilt derivatives -> Alpine
2023-02-26 16:34:12 <dewdude> because it was like "yes...now i can have internet in the bathroom
2023-02-26 16:34:51 <dewdude> now i got a small stack of stuff on a shelf with like, 20 ethernet lines running to it and a gigabit fiber connection
2023-02-26 16:36:35 <minimal> dewdude: I remember the pre-Wifi days when there were 2 competing technologies (DSSS and FHSS) which interfered with each other. Ran a couple of inter-building wireless link over several miles back then
2023-02-26 16:37:12 <dewdude> lol
2023-02-26 16:37:26 <dewdude> "you need to connect those buildings? buy this $300 box and give me a pringles can"
2023-02-26 16:37:42 <dewdude> they thought i was nuts
2023-02-26 16:38:04 <dewdude> but i got them a 5 mile 10mbps link for a fraction of what everyone else wanted to charge them
2023-02-26 16:38:36 <minimal> dewdude: nope, used by flat (A4 sized) directional antennas and also "cattle grid" antennas for further distances
2023-02-26 16:38:56 <dewdude> i literally built a waveguide out of a pringles can
2023-02-26 16:39:02 <minimal> no Pringle cans were hurt in the process lol
2023-02-26 16:39:10 <dewdude> they wanted cheap
2023-02-26 16:39:30 <dewdude> and it was easier/less involved than building a 2.4ghz yagi
2023-02-26 16:40:50 <JohannesJ[m]> sigh that moment when endusers start to check have i been pwned because their neighbour's smartass nephew told them to
2023-02-26 16:41:07 <dewdude> "cattle grid" antenna sounds like a parabolic reflector 
2023-02-26 16:42:42 <minimal> dewdude: yeah something like this: https://www.wifi-antennas.co.uk/24dbi-2-4ghz-parabolic-grid-antenna-n-type
2023-02-26 16:43:01 <minimal> we used it for a 7 mile 64Kbps link
2023-02-26 16:43:11 <dewdude> yeah. 
2023-02-26 16:43:20 <dewdude> parabolas provide a LOT of "gain"
2023-02-26 16:43:28 <dewdude> hence why they're commonly called satellite dishes
2023-02-26 16:44:01 <minimal> well we were in Europe where max "power" allowed was 100mW, from memory USA was/is 1W max
2023-02-26 16:44:35 <dewdude> it's been a while since i looked at the rules; I don't remember what the PEP limit here is for wifi
2023-02-26 16:44:49 <dewdude> PTP links could use higher PEP
2023-02-26 16:45:06 <dewdude> like we could run 10 or 20 watts if it was point-to-point
2023-02-26 16:45:28 <dewdude> i can legally run 1500W on channel 1...as long as it's Part 97 usage
2023-02-26 16:46:06 <minimal> 1500W? are you sure? for what, to cook some birds? lol
2023-02-26 16:46:10 <dewdude> heh
2023-02-26 16:46:12 <dewdude> ham radio
2023-02-26 16:46:28 <minimal> ham radio on 2.4GHz/5GHz?
2023-02-26 16:46:31 <dewdude> 2.4
2023-02-26 16:46:37 <dewdude> part of our allocation overlaps wifi
2023-02-26 16:46:47 <minimal> didn't know ham radio was allowed on those freqs
2023-02-26 16:46:51 <dewdude> yup
2023-02-26 16:47:03 <dewdude> we have primary allocation and we overlap channel 1 of common wifi
2023-02-26 16:47:26 <dewdude> there's not a lot of logical reasons to run that much power
2023-02-26 16:47:28 <dewdude> but...we can
2023-02-26 16:47:56 <dewdude> because as long as it's for ham use, ham rules apply
2023-02-26 16:48:06 <dewdude> and the maximum PEP on VHF and up is 1500w
2023-02-26 16:48:19 <minimal> that's almost twice the power of a typical domestic microwave oven
2023-02-26 16:48:20 <dewdude> actually..it's 1500W in 99% of bands
2023-02-26 16:48:27 <dewdude> lol
2023-02-26 16:48:28 <JohannesJ[m]> ubiquiti is running on a modified debian base no? basicly their apps and switches are linux boxes with a web gui :| one could do the same with alpine
2023-02-26 16:48:37 <JohannesJ[m]> well, thats what i was thinking anyway
2023-02-26 16:48:49 <dewdude> it's 300W higher than my microwave
2023-02-26 16:48:53 <dewdude> i buy 1200W models
2023-02-26 16:49:22 <dewdude> but usually you're using directional antennas or parabolic dishes...so you don't need 1500W
2023-02-26 16:49:47 <JohannesJ[m]> i really have too many ideas to ever get any of them done
2023-02-26 16:50:34 <psykose> sell them at the ideas store
2023-02-26 16:51:32 <JohannesJ[m]> thought of that, but the world is not ready for that :D
2023-02-26 16:54:16 <minimal> psykose: that's a great idea, you should sell that to someone ;-)
2023-02-26 16:55:20 <psykose> ask paper
2023-02-26 19:10:10 <Forza> Hi. I just did an apk update. Now I have an error starting gdm. This is a KVM guest. dmesg shows: udevd[1584]: failed to execute '/usr/bin/touch' '/usr/bin/touch /run/udev/gdm-machine-has-virtual-gpu': No such file or directory
2023-02-26 19:10:38 <Forza> `which touch" gives /bin/touch, so I guess something is pointing to the wrong place?
2023-02-26 19:11:22 <psykose> hmm
2023-02-26 19:11:44 <psykose> yeah
2023-02-26 19:11:47 <psykose> should patch that
2023-02-26 19:12:01 <Forza> :)
2023-02-26 19:13:17 <Forza> which file calls that line?
2023-02-26 19:14:04 <psykose> fixed
2023-02-26 19:14:07 <psykose> just upgrade in a bit i guess
2023-02-26 19:14:28 <psykose> 4fdf34068049567f861490a772a855648dc83d98
2023-02-26 19:16:01 <minimal> perhaps time to write a script that looks at all the Busybox softlinks and scan other packages for the same binaries and shows which are located at a differing patch than the equivalent BB tool? ;-)
2023-02-26 19:16:08 <minimal> s/patch/path/
2023-02-26 19:16:08 <alpine-meetbot> minimal meant to say: perhaps time to write a script that looks at all the Busybox softlinks and scan other packages for the same binaries and shows which are located at a differing path than the equivalent BB tool? ;-)
2023-02-26 19:17:14 <Forza> :)
2023-02-26 19:17:20 <Forza> thanks psykose
2023-02-26 19:28:18 <Forza> How can I force a reinstall of an apk?
2023-02-26 19:29:13 <psykose> apk fix name
2023-02-26 19:32:19 <Forza> ah thats it. thanks
2023-02-26 19:35:32 <Forza> The problem wasnt actually the package (linux-virt), but QEMU
2023-02-26 19:36:43 <Forza> I am runnig an Alpine VM with direct kernel boot via libvirtd/kvm. Even though the kernel image was updated, qemu seems to cache the kernel image when doing normal reboots. I had to do poweroff and then boot again for qemu to read the updated image file
2023-02-26 19:38:07 <Forza> I am experimenting with virtiofs as rootfs. Works pretty well I must say. One caveat is this issue with upgrading kernels.
2023-02-26 19:39:33 <psykose> hmm, haven't heard of that before
2023-02-26 19:39:51 <psykose> maybe there's some way to signal to qemu to read the new one
2023-02-26 19:40:00 <psykose> sounds not that outlandish
2023-02-26 19:40:21 <Forza> It seems to reduce ram usage a lot when the fsdrivers aren't needed inside thevguest.
2023-02-26 19:40:41 <Forza> Probably a qemu/libvrirt bug. maybe I should report it
2023-02-26 19:41:48 <Forza> basic alpine install with virtiofs as rootfs is using 75MB ram after a few days of uptime =)
2023-02-26 19:42:51 <Forza> and I don't have to allocate a (virtual) block device, so it saves lots of storage on the host. I can even do simple deduplication on the host between several guests as I am usin btrfs on the host
2023-02-26 20:22:32 <u0_a280> quit
2023-02-26 21:41:12 <WhyNotHugo> For the xdg-desktop-portal-wlr package, can we add install_if="xdg-desktop-portal sway" ?
2023-02-26 21:41:20 <WhyNotHugo> or maybe s/sway/wlroots/
2023-02-26 21:41:36 <ikke> WhyNotHugo: feel free to make an MR as a suggestion
2023-02-26 21:41:45 <WhyNotHugo> 
ACTION thumbs up

2023-02-26 21:42:08 <ikke> don't forget to pkgrel bump
2023-02-26 21:42:26 <psykose> i don't think that makes sense
2023-02-26 21:42:45 <psykose> install_if="xdg-desktop-portal wlroots" is probably a little more apt
2023-02-26 21:43:04 <ikke> We don't want apt :P
2023-02-26 21:43:22 <psykose> funnily i did add it to the repos just for libapt
2023-02-26 21:43:28 <ikke> heh
2023-02-26 21:43:44 <psykose> just dnf now :p
2023-02-26 21:43:55 <ikke> inbefore someone reporting advanced persistent threats in the repo
2023-02-26 21:46:23 <umk3[m]> labwc also makes use of xdg-desktop-portal-wlr
2023-02-26 21:47:23 <WhyNotHugo> labwc also uses wlroots
2023-02-26 21:47:45 <umk3[m]> yes
2023-02-26 21:49:01 <psykose> i'll add the wlroots one
2023-02-26 21:49:10 <WhyNotHugo> https://gitlab.alpinelinux.org/alpine/aports/-/merge_requests/44576
2023-02-26 21:49:15 <umk3[m]> i don't know if it's safe to assume that everyone who uses sway/labwc want xdg-wlr also installed
2023-02-26 21:49:31 <umk3[m]> same with gsettings-desktop-schemas
2023-02-26 21:50:12 <psykose> if they have xdg-desktop-portal installed they do
2023-02-26 21:50:54 <psykose> the classic of "i had even more changes" :p
2023-02-26 21:51:31 <umk3[m]> for me it's ok I have no problem
2023-02-27 03:45:51 <Asmadeus> Hmm, alpine doesn't have anything like kmod-static-nodes that creates e.g. /dev/loop-control so opening it automagics loop module load? ( https://github.com/systemd/systemd/blob/main/units/kmod-static-nodes.service.in )
2023-02-27 03:46:39 <Asmadeus> mkinitfs will 'modprobe loop' among others so I guess that's not a problem for most people, but wondering where I should create that node, or if I should just keep loop builtin...
2023-02-27 03:57:28 <Asmadeus> I guess just looping over /lib/modules/$(uname -r)/modules.devname like `kmod static-node` does is probably best given there's a couple of others in my config
2023-02-27 04:37:54 <Asmadeus> ah, spoke too fast, modules.devname is also created by kmod's depmod and not busybox's; I guess that'll require a bit more work to get the automagic dev nodes as it doesn't show up in modules.alias either
2023-02-27 04:38:41 <Asmadeus> ... or I could just hardcode it, it's not like modules.devname knows about permissions fixed in https://github.com/systemd/systemd/blob/main/tmpfiles.d/static-nodes-permissions.conf.in ...
2023-02-27 14:35:56 <benth0r> quinq, psykose, JohannesJ[m] regarding my 9p experiments: I tried https://github.com/chaos/diod on the server side. Feels rock solid in basic trials but copying a file from a mounted directory is about 5 times slower than scp-ing the same file. (Haven't tried pitting it against sshfs yet though, so it's a bit of an apples vs oranges comparison.
2023-02-27 14:36:14 <benth0r> )
2023-02-27 14:36:33 <JohannesJ[m]> nice :)
2023-02-27 14:36:35 <psykose> sounds about right
2023-02-27 14:36:40 <psykose> unusably slow
2023-02-27 14:36:54 <psykose> sshfs is generally the same as scp in my experience
2023-02-27 14:36:56 <JohannesJ[m]> i have issues with Samba on Alpine, very slow copy speeds even though it works great on Arch :( 
2023-02-27 14:37:08 <JohannesJ[m]> so im looking for other protocols :) maybe dav
2023-02-27 14:42:24 <benth0r> btw, I heard a rumor that folks are investigating alternatives to openrc for alpine. 
2023-02-27 14:42:43 <benth0r> What other options would there be?
2023-02-27 14:43:51 <ptrc> there were discussions about using skarnet's s6-rc instead
2023-02-27 14:44:28 <benth0r> but that died down again?
2023-02-27 14:44:42 <nickersonm> JohannesJ[m]: samba on Alpine to a Windows client works well for me
2023-02-27 14:45:29 <ptrc> well, currently the frontend part of s6 is a bit lacking iirc
2023-02-27 14:46:25 <psykose> it would require that to even be made and a bunch of other stuff and then somebody to do a bunch of work
2023-02-27 14:46:27 <psykose> maybe in 5 years
2023-02-27 14:46:55 <JohannesJ[m]> nickersonm: uploading files to an alpine based samba server results in very slow speeds. could you share your smb.conf perhaps?
2023-02-27 14:47:38 <JohannesJ[m]> at first i thought the problem was with windows 11, but when i copy to a synology nas (which uses samba under the hood) i dont have this problem
2023-02-27 14:50:20 <nickersonm> JohannesJ[m]: certainly. https://nickersonm.com/misc/smb.conf
2023-02-27 14:50:46 <nickersonm> JohannesJ[m]: https://nickersonm.com/misc/samba-btrfs-alpine-2022.png
2023-02-27 14:51:23 <JohannesJ[m]> i have BTRFS raid too :)
2023-02-27 14:51:43 <JohannesJ[m]> thank you very much! i'll tinker with it and see how it goes :)
2023-02-27 14:52:27 <nickersonm> JohannesJ[m]: so convenient for drive addition or removal :) You're welcome! Note that small files are still very slow for me, but that seems to be expected due to the overhead.
2023-02-27 14:53:24 <JohannesJ[m]> i suspect small files will always remain a thing :(
2023-02-27 14:54:46 <nickersonm> JohannesJ[m]: I also tried NFS and SSHFS, both were worse. But I will be trying NFS again some time just in case I tuned it poorly before. SSHFS had too much CPU overhead, though.
2023-02-27 14:55:36 <JohannesJ[m]> yeah, i figured that much :) so its either nfs or smb or dav for windows
2023-02-27 16:09:45 <oleovit> sry short  q: i want to try out zram on alpine - should i completely disable the swap partion in fst
2023-02-27 16:09:48 <oleovit> ab, using the zram-init service, or is it enough to set the priority to low (by default it's -2 anyway)
2023-02-27 16:16:50 <oleovit> i found an answer. i better let this go. should hang out less on r/inux. thy again like aöways for your work
2023-02-27 16:22:21 <psykose> swap and zram can be used at the same time indeed :)
2023-02-27 16:22:30 <JohannesJ[m]> heh
2023-02-27 16:35:10 <JohannesJ[m]> hmz, on my own alpine server i get speeds of 900MB/s 
2023-02-27 22:17:29 <boombim[m]> I've found cool thing in this article... (full message at <https://matrix.org/_matrix/media/v3/download/matrix.org/xbgsnBimbgjPLNJoarmsxtBq>)
2023-02-28 00:54:23 <winny__> i found alpine 3.8 being used in production :^)
2023-02-28 01:35:13 <Saijin_Naib> Was it internet-facing and hosting customer data like the unpatched Server 2003 at one of my prior jobs (within the last 4 years 😬) 
2023-02-28 03:32:23 <winny__> thankfully it was for internal customer use
2023-02-28 04:38:55 <dewdude> you think that's bad? a couple weeks ago I decomissioned an email server that was running Windows XP and had something like an 800 day uptime
2023-02-28 04:51:30 <Livy_> Hello. I'm new to Alpine.
2023-02-28 04:51:41 <dewdude> hi
2023-02-28 04:51:49 <Livy_> And new to IRC also.
2023-02-28 04:52:29 <dewdude> fantastic
2023-02-28 04:52:42 <dewdude> and people say irc is dead
2023-02-28 04:52:53 <heyman94> people who dont use irc: irc is dead
2023-02-28 04:53:18 <dewdude> same thing with ham radio. people who don't ham say ham radio is dead
2023-02-28 04:53:45 <heyman> lol
2023-02-28 04:53:51 <Livy_> Alpine is qite similar to OpenWrt -- very minimal. Luckily I am familiar with OpenWrt, so I figure out Alpine really fast.
2023-02-28 04:54:29 <heyman> i just found out about alpine an hour ago and so far ive gathered that u can install desktop environments as well as web browsers
2023-02-28 04:54:52 <dewdude> i've never used openwrt but i feel like it's more an appliance linux
2023-02-28 04:55:02 <heyman> but how do i install alpine easily without typing out all the commands to setup
2023-02-28 04:55:10 <Livy_> Right now I am running my IRC client as a Tmux session on my OpenWrt router.
2023-02-28 04:55:16 <Livy_> So I can check back message anytime later.
2023-02-28 04:55:36 <Livy_> Alpine is not good for desktop usage. I've tried to do that.
2023-02-28 04:55:41 <heyman> why?
2023-02-28 04:55:55 <Livy_> VScode, Google Chrome... and many software is not available on its repository.
2023-02-28 04:56:16 <heyman> uhh u could just use
2023-02-28 04:56:19 <heyman> chromium
2023-02-28 04:56:19 <Livy_> Mostly because these software are not designed for musl. Most distros use glibc.
2023-02-28 04:56:21 <heyman> and vscodium
2023-02-28 04:56:25 <heyman> no difference really
2023-02-28 04:56:30 <Livy_> To you them, you must use Flatpak, which is slow like hell.
2023-02-28 04:56:33 <heyman> theres a chromium package
2023-02-28 04:56:52 <dewdude> half the fun is patching source to work with musl then compiling
2023-02-28 04:56:52 <Livy_> Chromium cannot synchronize bookmarks to Google server anymore.
2023-02-28 04:57:06 <heyman> u dont wanna be using anything google anyways
2023-02-28 04:57:29 <Livy_> It sucks to work with a browser without bookmarks synchronization.
2023-02-28 04:57:56 <Livy_> I thought if switching to Firefox or Epiphany, but then all browser plugins I need are not available.
2023-02-28 04:57:59 <Livy_> lol
2023-02-28 04:58:14 <heyman> i never used that feature for years
2023-02-28 04:58:30 <dewdude> i just don't desktop on linux
2023-02-28 04:58:43 <heyman> and i have like 4 extensions 
2023-02-28 04:58:51 <heyman> every browser i use comes with ublock origin
2023-02-28 04:59:13 <Livy_> Anyway, I am trying to move my Ubuntu servers to Alpine.
2023-02-28 04:59:18 <heyman> really i could do without other extensions but if i happen to need them at a time ill just install them
2023-02-28 04:59:21 <heyman> on the spot
2023-02-28 04:59:23 <Livy_> And I face difficulties with QEMU.
2023-02-28 04:59:46 <heyman> maybe u want a hypervisor like proxmox
2023-02-28 04:59:58 <Livy_> Just like OpenWrt, Alpine separate a single package into multiple smaller packages.
2023-02-28 05:00:18 <Livy_> I had to guess a lot to make install all features I need.
2023-02-28 05:00:41 <Livy_> But still, I cannot make USB redirection from client to guest working.
2023-02-28 05:00:56 <Livy_> -device usb-redir,id="usb_redir0",bus="xhci0.0",chardev="spicevmc-usb0"
2023-02-28 05:01:15 <Livy_> And I get this error: qemu-system-x86_64: -device usb-redir,id=usb_redir0,bus=xhci0.0,chardev=spicevmc-usb0: 'usb-redir' is not a valid device model name
2023-02-28 05:01:32 <heyman> 'usb-redir' is not a valid device model name
2023-02-28 05:01:42 <heyman> means u need to put something else other than usb-redir which is valid
2023-02-28 05:01:55 <heyman> go search up "model name usb qemu config" maybe?
2023-02-28 05:02:06 <Livy_> Nah, it is a Spice device, which should be available in QEMU.
2023-02-28 05:02:33 <psykose> did you install qemu-hw-usb-redirect
2023-02-28 05:02:40 <Livy_> It is likely because the maintainer of QEMU package (Natanael Copa) splits it into a separate one.
2023-02-28 05:02:58 <Livy_> Before that I get error with 'qxl-vga' device for my Windoze guest.
2023-02-28 05:03:08 <heyman> psykose is there a fast way to install alpine
2023-02-28 05:03:18 <Livy_> And solved it by installing "qemu-hw-display-qxl".
2023-02-28 05:03:41 <Livy_> I'll try the one you suggest now. (qemu-hw-usb-redirect).
2023-02-28 05:04:51 <Livy_> OMG it works!!!
2023-02-28 05:05:19 <Livy_> Thank you very much @psykose.
2023-02-28 05:05:30 <elly> the fastest way is to run `setup-alpine`
2023-02-28 05:05:58 <Livy_> I knew this distro was like OpenWrt, which split a big package into multiple smaller ones.
2023-02-28 05:06:08 <Livy_> But I just do not know all package names I need.
2023-02-28 05:10:00 <dewdude> again..i say openwrt is designed to be an embedded os where as alpine is just minimal
2023-02-28 05:10:05 <dewdude> at least until you build it up
2023-02-28 05:10:10 <dewdude> the overall goals are seperate
2023-02-28 05:11:28 <heyman> what do i select for us keyboard layout
2023-02-28 05:11:31 <heyman> theres so many
2023-02-28 05:11:40 <dewdude> us
2023-02-28 05:11:46 <heyman> yes u select that
2023-02-28 05:11:50 <heyman> and it asks for specifics
2023-02-28 05:11:54 <dewdude> yes.
2023-02-28 05:11:56 <dewdude> us again
2023-02-28 05:12:15 <heyman> holy fuck that is misleading
2023-02-28 05:12:28 <heyman> thanks
2023-02-28 05:12:28 <dewdude> not...really?
2023-02-28 05:12:35 <dewdude> made sense to me the first time i installed
2023-02-28 05:12:52 <heyman> it shows so many variants
2023-02-28 05:12:58 <dewdude> yes
2023-02-28 05:13:00 <heyman> and just us isnt one of them
2023-02-28 05:13:09 <dewdude> yes it is?
2023-02-28 05:13:12 <heyman> so of course i think it wants me to choose one variant
2023-02-28 05:13:14 <heyman> no its not
2023-02-28 05:13:47 <dewdude> hold on. i'm booting up the installer
2023-02-28 05:15:16 <dewdude> https://img.pickmy.org/ghnaymni.png
2023-02-28 05:16:22 <dewdude> it's the very last option
2023-02-28 05:21:59 <heyman> god FUCK THAT
2023-02-28 05:22:08 <heyman> my brain completely ignored it
2023-02-28 05:54:37 <heyman> also guys
2023-02-28 05:54:47 <heyman> is there a package for realtek wireless firmware
2023-02-28 05:55:03 <heyman> i wanna make a wifi adapter with a rtl8188eu chipset work
2023-02-28 05:59:12 <psykose> all i can find is https://github.com/ivanovborislav/rtl8188eu, i.e. the whole module is out of tree
2023-02-28 06:08:41 <psykose> boombim[m]: change the LANG from 'C.UTF-8' to just 'C' or whatever, in /etc/profile.c/locale.sh
2023-02-28 06:08:44 <psykose> that's what breaks the symbols
2023-02-28 06:08:52 <psykose> the utf-8 suffix
2023-02-28 06:11:48 <heyman> https://git.kernel.org/pub/scm/linux/kernel/git/firmware/linux-firmware.git/tree/rtlwifi/rtl8188eufw.bin
2023-02-28 06:11:52 <heyman> psykose this?
2023-02-28 06:12:19 <psykose> that is certainly firmware
2023-02-28 06:12:32 <psykose> linux-firmware-rtlwifi
2023-02-28 06:12:39 <heyman> i dont know if thats it or not
2023-02-28 06:13:04 <heyman> eh only one way to find out
2023-02-28 06:13:36 <heyman> i managed to get youtube running on firefox-esr and xfce4 lol
2023-02-28 06:30:00 <psykose> boombim[m]: or printf '\033%%G' in the tty
2023-02-28 06:37:09 <humm> there’s a service to put the tty in UTF-8 mode
2023-02-28 06:40:23 <psykose> there is indeed
2023-02-28 06:42:14 <psykose> it also does not work, the classic
2023-02-28 07:34:11 <frag> hmm booted up my rpi4b as usual today ... its very slow today :S boot was fast, from login to starting xorg was slow (like 1-ish min, usually takes 3ish sec ), starting qutebrowser takes like couple of mins :S, anyone know what it _could_ be? havent seen any errors
2023-02-28 07:38:19 <frag> booted on another usb, seemed fine
2023-02-28 07:50:29 <frag> ... wondering if its speeding up gradually :S oh well ..
2023-02-28 07:59:44 <boombim> psykose: thanks. Is it workaround for consolefont, right?
2023-02-28 10:45:59 <psykose> figured out an even better one
2023-02-28 10:46:12 <psykose> put UNICODE=true in /etc/conf.d/termencoding
2023-02-28 10:46:17 <psykose> why that is not default i will never know
2023-02-28 11:42:22 <boombim> got it. Thank you very much. I'll try.
2023-02-28 11:43:34 <boombim> hm. I don't have such file.
2023-02-28 11:43:46 <boombim> should I create termencoding manually?
2023-02-28 12:21:17 <psykose> yes
2023-02-28 12:25:29 <JohannesJ[m]> so.. on Debian i am now running prometheus/grafana which errors because it says it has too many files open.. wil Alpine run into the same issue? from a bit of googling it seems a systemd thingy?
2023-02-28 12:41:48 <JohannesJ[m]> hmz.. docker seems to not have this problem :)
2023-02-28 12:58:11 <ikke> JohannesJ[m]: limit by default is not that high, you'd need to increase it. 
2023-02-28 12:58:45 <JohannesJ[m]> yes, thats what google told me too :) but that doesnt seem to work, or i'm stupid enough not to be able to do it
2023-02-28 12:59:06 <JohannesJ[m]> its all along the lines of this: https://linuxhint.com/linux-too-many-open-files-error/
2023-02-28 12:59:46 <psykose> its not global
2023-02-28 13:00:21 <psykose> rc_ulimit="-n 100000" in the service conf.d is one example
2023-02-28 13:02:39 <Fxzxmic> Failed to compute left/right minimum bearings for "System,System,Microsoft YaHei UI"
2023-02-28 13:02:43 <Fxzxmic> strange
2023-02-28 13:03:16 <JohannesJ[m]> psykose: yes but i would assume this is for Alpine ;)
2023-02-28 13:03:30 <JohannesJ[m]> but at the very least this confirms my question, that alpine has limits too :)
2023-02-28 13:23:06 <ikke> It's a Linux kernel thing 
2023-02-28 13:23:12 <ikke> Any Linux distro has it
2023-02-28 13:23:26 <ikke> (but the exact limits may differ) 
2023-02-28 13:57:21 <boombim> psykose: thanks for your research. Consolefont works now.
2023-02-28 13:57:29 <JohannesJ[m]> yeah, i figured :) 
2023-02-28 14:12:46 <linex[m]> hello, does anyone know if one can restrict ssh certificate to be short-lived ?
2023-02-28 14:13:15 <linex[m]> i.e. the ssh server will not accept keys that are valid for more than 3 days
2023-02-28 14:14:11 <minimal> linex[m]: keys or certificates? you have mentioned both
2023-02-28 14:15:03 <minimal> keys don't have a lifetime (expiry date) whereas certificates do
2023-02-28 14:15:29 <linex[m]> certificates, sorry
2023-02-28 14:17:15 <linex[m]> need coffee :p
2023-02-28 14:18:26 <minimal> https://en.wikibooks.org/wiki/OpenSSH/Cookbook/Certificate-based_Authentication#Time_Limitations_for_User_Certificates
2023-02-28 14:22:23 <linex[m]> the rfc does allow for extensions, maybe there exists one that enforces time-constraints
2023-02-28 14:25:34 <minimal> linex[m]: are you wanting to do "delta" certificate expiry rather than certs that have an explicit expiry date/time?
2023-02-28 14:29:22 <linex[m]> yes,
2023-02-28 14:30:16 <minimal> linex[m]: that sounds like where you'd have to use something like Hashicorp Vault then so issue/re-issue certs on a regular basis
2023-02-28 14:38:24 <minimal> linex[m]: e.g. https://www.hashicorp.com/blog/managing-ssh-access-at-scale-with-hashicorp-vault
2023-02-28 14:39:21 <JohannesJ[m]> we use small-step ca to issue short lived certificates
2023-02-28 14:39:43 <JohannesJ[m]> * for ssh
2023-02-28 14:40:48 <minimal> JohannesJ[m]: I maintain step-cli and step-ca for Alpine ;-)
2023-02-28 14:40:56 <JohannesJ[m]> nice :)
2023-02-28 14:41:00 <JohannesJ[m]> thank you :)
2023-02-28 15:25:39 <linex[m]> nice, thanks, yeah I think I'll end up using step-ca as I am already using it for x.509 certs. I was just wondering if one could enforce delta-times on certificates on the server side 
2023-02-28 15:29:15 <linex[m]> It seems a bit weird that it does not exist ? maybe not enough use-cases
2023-02-28 16:06:22 <NfixEstrada[m]> Hello! I need a pacakge that's on the testing repository, but it has some dependencies that on the testing repository are on a version and on my "normal" repository are on another. Is there a way to `apk add package@testing` and download all the dependencies also from the testing repository? Thanks! :)
2023-02-28 16:07:55 <lopid> sounds like a catastrophe in the making
2023-02-28 16:08:37 <NfixEstrada[m]> The catastrophe is already done, it's in production and now the repos have updated and it has broken all the pipelines xD
2023-02-28 16:11:51 <psykose> no, you have to upgrade to edge to use it
2023-02-28 16:14:21 <NfixEstrada[m]> And isn't there a way to "rollback" to the previous commit?
2023-02-28 16:14:26 <NfixEstrada[m]> https://git.alpinelinux.org/aports/commit/?id=2fdb1f964912be21c3ac055a383fc2e1637d2b79
2023-02-28 16:15:07 <NfixEstrada[m]> This doesn't seem to do anything special
2023-02-28 16:16:42 <ptrc> nope, and it does
2023-02-28 16:17:06 <NfixEstrada[m]> Or build it from the APKBUILD using the "normal" repositories?
2023-02-28 16:17:57 <minimal> NfixEstrade[m]: you have "testing" in production? hmm...doesn't the name "testing" give a hint it is not intended for production use?
2023-02-28 16:18:26 <NfixEstrada[m]> minimal: yes xD but it's the only way to get the package we need for production xD
2023-02-28 16:19:11 <ptrc> there are only two sane options here: you can either build it yourself, or upgrade your production to alpine edge
2023-02-28 16:19:21 <psykose> building it looks like 5 minutes of work
2023-02-28 16:19:33 <psykose> maybe i build too many things 
2023-02-28 16:19:38 <ikke> Or get it moved to community 
2023-02-28 16:19:54 <ptrc> fair, but that's only gonna land in the next stable release :p
2023-02-28 16:20:21 <NfixEstrada[m]> So, building it is "okay"?
2023-02-28 16:20:23 <kwadroke[m]> or run it in a container
2023-02-28 16:20:33 <kwadroke[m]> using edge
2023-02-28 16:20:57 <psykose> it's not illegal to compile code yourself, yeah
2023-02-28 16:21:30 <NfixEstrada[m]> kwadroke[m]: It's inside a container that has lots of other packages, and in order to update it it takes us like a week of testing that everything works and nothing breaks
2023-02-28 16:22:57 <minimal> NfixEstrade[m]: a package in Alpine "testing" requires testing? Who would have thought? ;-)
2023-02-28 16:24:05 <kwadroke[m]> depending on the container system used and system resources available, you might be able to run them side by side
2023-02-28 17:07:12 <j_v> anyone here have any experience with using System76 Pangolin with alpinelinux?
2023-02-28 20:11:42 <bleb> after installing ubuntu on a spare partition, grub can't boot alpine and instead prints "error: kernel doesn't support EFI handover.  error: you need to load the kernel first."
2023-02-28 20:13:04 <bleb> is there a command I can run from an alpine live cd to reinstall grub as it was before
2023-02-28 20:16:35 <dwfreed> what kernel are you running?
2023-02-28 20:19:38 <bleb> on alpine I had linux-edge installed
2023-02-28 20:19:54 <bleb> now I can only boot ubuntu 22.10
2023-02-28 20:20:27 <bleb> I have a live cd which is alpine 3.15, kernel 5.15
2023-02-28 20:25:18 <dwfreed> sounds like the ubuntu grub depends on the handover protocol still, and that was deprecated in 6.2
2023-02-28 20:25:41 <dwfreed> reinstalling grub from the alpine live cd would probably fix it
2023-02-28 20:28:54 <bleb> like just grub-install /path/to/esp/ ?
2023-02-28 20:29:12 <bleb> or is there another command I need to run to generate the menu entries etc
2023-02-28 20:29:32 <dwfreed> you'll probably have to give some special flags to grub; see the grub manpage for specifying target, efi dir, and boot dir
2023-02-28 20:30:12 <dwfreed> the menu entries most likely don't need changing, because ubuntu hopefully just copied the existing ones when rewriting grub.cfg
2023-02-28 20:32:50 <bleb> well it put ubuntu first, making it the default
2023-02-28 20:34:15 <dwfreed> that you can just edit manually
2023-02-28 20:34:49 <dwfreed> next time you boot ubuntu you'll want to adjust its config to not do that
2023-02-28 20:35:13 <dwfreed> (because there'll probably be a kernel update, and it'll probably trigger regenerating the grub config)
2023-02-28 20:36:00 <bleb> I am guessing grub-mkconfig is what generated the config originally, I will probably just do that if I can figure out what command line would replicate what was done in the past
2023-02-28 20:36:31 <dwfreed> grub-mkconfig -o /path/to/boot/grub/grub.cf
2023-02-28 20:36:34 <dwfreed> grub-mkconfig -o /path/to/boot/grub/grub.cfg
2023-02-28 20:36:45 <dwfreed> second one, not first one, missed the g before hitting enter
2023-02-28 21:13:48 <bleb> dwfreed: thanks
2023-02-28 21:15:09 <deckard> im trying to install py3-feedgen from testing but the feedgen module does not get added to site-packages (and therefore imports fail in python)
2023-02-28 21:15:16 <bleb> from the alpine media, grub-install --target=x86_64-efi --efi-directory=/mnt --boot-directory=/mnt1/boot seemed to work
2023-02-28 21:15:24 <deckard> however i cloned aports and built the package locally and installed that, and it works fine
2023-02-28 21:15:29 <deckard> why might this be?
2023-02-28 21:15:38 <bleb> with the esp on /mnt and the installed alpine filesystem on /mnt1
2023-02-28 21:17:12 <bleb> but now ubuntu is not showing up in the boot menu
2023-02-28 21:17:57 <bleb> booted to the alpine installation again, I ran grub-mkconfig -o /boot/grub/grub.cfg
2023-02-28 21:19:38 <bleb> I saw the warning about os-prober not being run, so I installed os-prober and added GRUB_DISABLE_OS_PROBER=false to /etc/default/grub
2023-02-28 21:21:42 <bleb> but grub-mkconfig still outputs nothing about ubuntu
2023-02-28 21:28:02 <deckard> nvm i figured it out, testing was building against python3.11 while my local build prefered python3.10 from alpine 3.17's main repo
2023-02-28 21:28:11 <bleb> cool
2023-02-28 21:31:58 <bleb> anyone know what os-prober looks for?  has no man page and just exists 0 when run.
2023-02-28 21:33:02 <bleb> or is there an alternative to os-prober which would be able to detect an ubuntu installation?
2023-02-28 21:41:04 <dwfreed> bleb: os-prober logs to syslog
2023-02-28 21:41:52 <dwfreed> so maybe check your syslog to see if it logged anything useful
2023-02-28 21:41:54 <bleb> where would I see its output in a default alpine installation?
2023-02-28 21:42:17 <dwfreed> presumably somewhere in /var/log :)
2023-02-28 21:42:26 <dwfreed> I don't actually use alpine, so I don't know for sure
2023-02-28 21:44:09 <bleb> I see some stuff in /var/log/messages
2023-02-28 21:47:33 <bleb> pretty cryptic though
2023-02-28 21:47:46 <bleb> forget it, sound doesn't work in ubuntu either
2023-02-28 21:48:10 <bleb> so I will focus on alpine
2023-02-28 21:56:52 <ikke> deckard: combining several releases is not supported, for the reasons you experienced 
2023-02-28 21:57:42 <deckard> right that makes sense
2023-02-28 22:12:13 <bleb> after messing with grub, my bios boot menu has three entries despite only one physical device:
2023-02-28 22:12:49 <bleb> UEFI OS (SHGP31-2000GM), ubuntu (SHGP31-2000GM), grub (SHGP31-2000GM)
2023-02-28 22:13:17 <bleb> before I just had UEFI OS
2023-02-28 22:13:31 <bleb> what command creates a new entry there, and how can I remove the extra ones now?
2023-02-28 22:14:01 <bleb> looks like each entry in that menu leads to a different grub menu; the grub menu created by the ubuntu installer is still accessible by selecting the ubuntu option
2023-02-28 22:45:20 <Habbie> bleb, i recall 'efibootmgr'
2023-02-28 23:24:06 <minimal> bleb: there are 2 different aspects, (1) which ever directories inside the ESP partition(s) EFI/ directory exist and contain EFI files, and (b) which Boot UEFI variables exist (and their boot order), which efibootmgr can show
2023-02-28 23:24:45 <minimal> actually there's a 3rd, which is whether EFI/BOOT/BOOTX64.EFI exists in ESP partition(s) also
2023-02-28 23:42:36 <bleb> minimal: so the bios boot menu entries are UEFI OS, ubuntu, grub
2023-02-28 23:42:49 <bleb> /boot/efi/EFI has directories alpine, boot, grub, ubuntu
2023-02-28 23:43:21 <bleb> all contain at least some .efi file
2023-02-28 23:44:17 <minimal> bleb: you haven't mentioned UEFI boot variables though...
2023-02-28 23:44:25 <bleb> I also see /boot/efi/EFI/boot/bootx64.efi, so that exists
2023-02-28 23:44:59 <bleb> minimal: because I can't find them out until I read the efibootmgr man page
2023-02-28 23:46:22 <minimal> I believe it you just run egibootmgr with no options it shows the current variables
2023-02-28 23:46:54 <minimal> so look for variable names beginning with "Boot"
2023-02-28 23:47:42 <bleb> yup
2023-02-28 23:47:59 <bleb> that shows the boot entries I see in the bios boot menu
2023-02-28 23:48:57 <bleb> so I guess I can remove the unneeded ones with efibootmgr
2023-02-28 23:49:38 <minimal> so "BootOrder" obviously defines the sequence of which to try booting
2023-02-28 23:50:05 <bleb> yup I see that
2023-02-28 23:50:33 <bleb> I wonder what the EFI/alpine directory is for
2023-02-28 23:50:42 <bleb> containing only grubx64.efi
2023-02-28 23:50:55 <minimal> it is for booting Alpine of course
2023-02-28 23:50:55 <bleb> it is not referenced in any of the efibootmgr variables
2023-02-28 23:51:25 <bleb> UEFI OS points to EFI\BOOT\BOOTX64.EFI
2023-02-28 23:51:27 <minimal> that is where (from memory) setup-disk tells grub-install to put the Alpine EFI file
2023-02-28 23:51:37 <bleb> grub points to \EFI\GRUB\GRUBX64.EFI
2023-02-28 23:51:41 <bleb> ubuntu points to ubuntu obviously
2023-02-28 23:52:30 <minimal> EFI\BOOT|BOOTX64.EFI is the UEFI "fallback" location if no (valid) UEFI boot variables can be found - it is *also* written by setup-disk and may perhaps also be written by the Ubuntu installer
2023-02-28 23:55:24 <bleb> setup-disk passes --bootloader-id=alpine to grub-install
2023-02-28 23:56:25 <bleb> and --no-nvram
2023-02-28 23:56:51 <bleb> iirc the --no-nvram doesn't work for multiboot