2023-04-01 00:01:24 yeah, also getting a black screen as soon as I get out of the bootloader 2023-04-01 00:01:56 that's why I'd assuming it was missing kms (for DRM driver) 2023-04-01 00:01:56 My keyboard and screen works up until I get out of the bootloader 2023-04-01 00:02:30 perhaps also missing firmware for your graphics? what is it? i915? amd? 2023-04-01 00:03:56 It's a thinkcenter with an AMD cpu with integrated amd graphics (afaik) 2023-04-01 00:05:17 so "apk add linux-firmware-amdgpu" perhaps, or "linux-firmware-radeon" if its an older generation 2023-04-01 00:07:18 then regenerate the initramfs as the automatic trigger doesn't seem to kick-in for firmware being installed 2023-04-01 00:16:00 I think I've had this issue with another OS as well. Had to do with IOMMU.. Setting amd_iommu=off makes the keyboard lit up after the bootloader but still no screen. I'll check how to install linux-firmware-amdgpu during alpine's installation 2023-04-01 00:38:44 ok in my case it seems like I need to install linux-firmware-radeon since it's an older card 2023-04-01 00:39:12 Is that possible to do during the install? From the disk image. I couldn't find a wiki entry about that 2023-04-01 00:40:37 the default is all the firmware pretty sure 2023-04-01 00:44:05 actually I see that the installer uses the amdgpu driver, not radeon. So now I'm confused. Guess amdgpu is fine.. Unsure why it doesn't work though.. 2023-04-01 00:45:24 the installer probably has both 2023-04-01 00:56:04 I reinstalled but without encrypting the root partition and it works fine. hum. 2023-04-01 00:59:07 Oh, I don't see kms in the features list. It was there during the install but I guess it generates another one. 2023-04-01 01:11:53 Yeah that worked. Thanks minimal and psykose . Had to edit /sbin/alpine-setup to add the kms feature (didn't know how else to do it). 2023-04-01 01:12:15 you edit /etc/mkinitfs/mkinitfs.conf 2023-04-01 01:14:36 Is that after or before setup-alpine? because I could see it in the list of /etc/mkinitfs/mkinitfs.conf, but setup-alpine uses a hardcoded list. 2023-04-01 01:18:08 after 2023-04-01 02:20:29 psykose: wondering why these sort of issues have become common recently 2023-04-01 02:20:52 the initramfs has simpledrm present so that should be enough to get a display working in initramfs 2023-04-01 02:22:36 wasn't it just the generic simpledrm issue that if you don't have the firmware there it's broken 2023-04-01 02:27:52 eh? simpledrm doesn't use any firmware and it should give a display on any device, and if/when "kms" is missing from mkinitfs.conf then the initramfs doesn't contain any i915/amd etc DRM modules so firmware isn't an issue 2023-04-01 02:29:20 out-of-the-box initramfs' init loads the simpledrm, if will *also* load a graphics card specific module if that's present in the initramfs 2023-04-01 02:35:00 my experience has been the opposite for some reason 2023-04-01 02:35:15 you either have all the modules and firmware in there, or you get a broken display :p 2023-04-01 02:36:54 also interestingly without enabling some of the legacy fbdev stuff in kernel i also get a broken display 2023-04-01 02:37:20 CONFIG_DRM_FBDEV_EMULATION or whatever it was 2023-04-01 02:37:57 doing purely drm drivers + simpledrm gives me a broken display too 2023-04-01 02:38:54 hmm, certainly for VMs here (Qemu & VirtualBox) only simpledrm is required to see boot (though I tend to only add bochs or vmwgfx to initramfs) 2023-04-01 02:40:36 and (without "quiet" on the cmdline) I see the dmesg of vmwgfx loading as obviously simpledrm has loaded first so that I can see that dmesg output 2023-04-01 02:41:00 simpledrm itself doesn't write anything to dmesg so its hard to see exactly when it loads 2023-04-01 02:42:09 I'll have to find some time to play with a physical PC to figure this out 2023-04-01 03:02:44 @psykose: Turns out my mouse doesn't work because the kernel is ignoring the key events because the mouse firmware says they are constants. I need to write a custom kernel driver or eBPF 2023-04-01 03:07:04 Saijin_Naib[m], what about blacklisting the bad driver/fw and letting a generic driver handle the mouse? 2023-04-01 03:13:22 I'm not sure what kernel module it is using... I'm assuming the basic HID handler  2023-04-01 03:14:03 there's a whole quirk system to make it handle stuff differently 2023-04-01 03:15:57 I filed an issue with libinput folks and they said this needs to be handled in the kernel since the events are being dropped by the kernel long before anything even makes it up to libinput 2023-04-01 03:16:13 But yeah, obviously not an Alpine problem, just closing the loop on it since I just found out what is going on.  2023-04-01 03:16:43 i told you as much quite fast didn't i :p 2023-04-01 03:17:17 Well, yeah, haha. But now I know exactly where and how I'm fucked, so that's nice 2023-04-01 03:17:26 libratbag also is packaged 2023-04-01 03:17:45 No chance I'm figuring out how to write a kernel driver and getting it upstream 🤣 2023-04-01 03:17:59 psykose: Stable? Am I stupid?  2023-04-01 03:18:12 testing/ratbag :p 2023-04-01 03:18:44 Dude... I can't go back to Edge, it breaks GNOME Builder 😭 2023-04-01 03:19:18 gnome builder 😳 2023-04-01 03:20:24 ... I need to write copious amounts of rST documentation for work, and it previews it best 🤷 2023-04-01 03:21:08 1:1 with how it looks built by Sphinx and published. Really neat. And has a rST linter now, too.  2023-04-01 03:22:23 yeah 2023-04-01 03:22:37 what breaks? i'd imagine it would just be broken in 3.18 if it's broken now 2023-04-01 03:22:42 I was trying to package and update/fixup Formiko, which works brilliantly as well (hence putting m2r2 in testing) but I got over my head fast 2023-04-01 03:23:35 psykose: I never figured it out. It even broke rST preview in the Flatpak of Builder, across both my machines. Dropping back to stable fixed whatever was going on 🤷 2023-04-01 03:23:44 weird 2023-04-01 03:24:24 Incredibly weird. I'm also like, the only person using Builder for this specific thing, I'm sure, so searching and asking around was fruitless m 2023-04-01 03:57:57 is there any workaround for this? https://paste.debian.net/plain/1275974 I'm just running bootstrap on abuild master HEAD 2023-04-01 04:03:21 that looks like you are running bootstrap from master while being on 3.17 or something 2023-04-01 04:03:49 for which the fix is obvious but whatever you are doing is probably not supported so i'd recommend to not do it :p 2023-04-01 04:04:45 I mean I just did git clone https://gitlab.alpinelinux.org/alpine/aports.git && sh aports/scripts/bootstrap.sh aarch64 2023-04-01 04:07:26 and i already answered why it didn't work 2023-04-01 04:08:20 oh I see, you're saying the host OS 2023-04-01 04:08:27 the host OS version must match the target? 2023-04-01 04:09:14 i could tell you but unless you are bootstrapping aports on a new architecture (as the name implies) this is not what it's for, so i'd recommend to not use it and forget it exists 2023-04-01 04:09:28 the aarch64 gives it away :p 2023-04-01 04:46:02 psykose: if I forget about bootstrap.sh, attempting to cross-compile any package for aarch64 fails with "Missing dependencies (use -r to autoinstall them): build-base-aarch64". Attempting to build build-base with "CHOST=aarch64 abuild" also fails the same way (Missing dependencies (use -r to autoinstall them): build-base-aarch64) 2023-04-01 04:46:17 there is no cross compiling in aports 2023-04-01 04:47:25 you would be right that trying to do it and the CHOST= thing is based on boostrap.sh, sure 2023-04-01 04:47:38 it's not used for anything and none of the builders use it and it's not meant to be used 2023-04-01 04:49:12 what is the supported way to build linux-lts for aarch64? can I cross-compile it from an x86-64 machine? 2023-04-01 04:49:34 you have to do it natively or not do it in aports 2023-04-01 04:50:16 "not do it in aports" you mean basically copy the alpine config over and build it myself? 2023-04-01 04:50:17 well, there's also the CHOST=aarch64 abuild rootbld thing 2023-04-01 04:50:24 maybe that works 2023-04-01 04:51:13 CBUILD rather 2023-04-01 04:51:20 are the officially provided aarch64 builds being cross-compiled or are they built on actual aarch64 hardware? 2023-04-01 04:51:21 which then uses qemu binfmt and it's native 2023-04-01 04:51:43 there's no cross compilation for any packages 2023-04-01 04:51:44 native 2023-04-01 04:52:25 I want to be really clear here, this package https://dl-cdn.alpinelinux.org/alpine/v3.17/releases/aarch64/alpine-uboot-3.17.2-aarch64.tar.gz is being built natively on a Pi or something? 2023-04-01 04:52:37 not on a pi, but natively yes 2023-04-01 04:53:22 but yeah CBUILD=aarch64 abuild rootbld should work i guess 2023-04-01 04:53:25 i always forget about that one 2023-04-01 05:02:26 I really wanted to set up Alpine on my R2S but this is a bit much, I think I'll just explore extracting the rootfs and providing my own u-boot and kernel 2023-04-01 08:21:04 Hi. Sorry if this is the wrong channel to ask this question: Currently using diskless install on Raspberry Pi 1 (armhf), Alpine 3.17.3 seems to boot with Linux 5.15.102-r0, and linux-rpi-dev is currently at 5.15.105-r0 . Should I be rebuilding alpine-rpi-3.17.3-armhf.tar.gz to use linux-rpi-dev-5.15.105-r0 if I need to compile additional kernel modules, or am I doing something wrong here? :( 2023-04-01 08:24:14 you have to also update the kernel to 5.15.105 by however that is done on diskless 2023-04-01 08:24:22 update-kernel 2023-04-01 08:24:28 or flash a new image 2023-04-01 08:24:48 (update-kernel apparently requires quite some memory) 2023-04-01 08:26:28 ah, ok thanks for the help :) 2023-04-01 16:11:13 I don't understand what happens here, I have VM booting linux 6.1.21, so can't load modules (I suppose that it loads some contained on initramfs) 2023-04-01 16:11:36 but /boot/vmlinuz-lts looks 6.1.22, it has same checksum than host whicn is running 6.1.22 2023-04-01 16:12:20 I tried, apk fix linux-lts, there is an error from syslinux 2023-04-01 16:12:22 uhMm.. 2023-04-01 16:12:29 maybe the boot is on another partition 2023-04-01 16:13:10 yeah... boot is in sda1, root on sda3 2023-04-01 16:13:24 no mention on /etc/fstab 2023-04-01 16:13:43 it's a straightforward install from setup-alpine 2023-04-01 16:14:39 oh great, finally running 6.1.22 2023-04-01 16:27:11 In WSL 2, tried rc-service rsyncd start, got "WARNING: rsyncd is already starting" 2023-04-01 16:27:20 No pid file in /run though 2023-04-01 16:28:49 donoban: you've installed in Sys mode (run-from-disk) on a UEFI or BIOS system? 2023-04-01 16:30:25 Same for s6 2023-04-01 17:12:36 minimal: I'm not sure, since it's a VM for just things I didn't care too much, it's currently using syslinux so it's probably BIOS? 2023-04-01 17:35:43 Why do my ttys switch on alt+left/right ?? 2023-04-01 17:35:52 s/??/?/ 2023-04-01 17:35:59 s/\?\?/?/ 2023-04-01 17:56:57 donoban: did you reboot after kernel package was updated? if not then the running kernel version will not match the kernel version and kernel module versions on disk 2023-04-01 18:38:29 linux-lts abuild errors with "comparison between two arrays" in mm/page-alloc.c 2023-04-01 18:38:34 config is custom, arch is x86_64 2023-04-01 18:40:41 cockliuser[m]: you'd need to provide more info, like a diff in kernel config file, in order for anyone to help 2023-04-01 18:42:08 Oh sorry 2023-04-01 18:42:26 http://dpaste.com/AT4PH5VCH 2023-04-01 18:44:23 what did you patch the kernel with 2023-04-01 18:45:10 Patches are whatever is in the aports, haven't modified them 2023-04-01 18:45:23 Config is from make menuconfig 2023-04-01 18:46:09 you're not diff'ing like-with-like - the linux-lts package does not come with a "full" config file, it is a diff from the "upstream" kernel defaults, so either you'd need to convert your modified config file into a similar "diff" before comparing, or else generate a full config from the linux-lts (e.g. cp lts.x86_64.config .config; make oldconfig) one before comparing 2023-04-01 18:46:12 i guess you created a magic config that doesn't build them then 2023-04-01 18:46:13 shrug 2023-04-01 18:47:31 minimal: ah ok 2023-04-01 18:47:57 cockliuser[m]: also that's kernel 6.3.0-rc4 you're trying to build 2023-04-01 18:48:25 i.e. an *RC* so there's a fair chance of stuff having problems 2023-04-01 18:48:46 didn't even notice 2023-04-01 18:49:23 it's probably easy to fix if you look at the actual error and look for the upstream fix 2023-04-01 18:49:34 posting 'abuild fail i run custom config' is going to get you 0 results however 2023-04-01 18:50:03 Sorry lol 2023-04-01 18:50:05 or at least look at linux-edge which is 6.2 rather than linux-lts which is 6.1.x 2023-04-01 18:52:54 i run rcs and it builds fine, so it's related to the specific file and what you enabled 2023-04-01 18:53:03 from there you can look at git or search the ml if you know how 2023-04-01 18:53:28 I'm thinking of setting up a media centre on an rpi with alpine. Any obvious red flags? 2023-04-01 18:53:50 WhynotHugo: the Chinese one? ;-) 2023-04-01 18:55:11 I have a real one. It runs "home assistant", but I'm not really giving it sufficient use, so I're repuprposing it. 2023-04-01 18:55:15 repurposing 2023-04-01 18:55:33 it will take a long time to transscode mp3 to flac 2023-04-01 18:56:04 I want to hook it up to the TV with a bluetooth remote-keyboard and use it for ... video playback basically. 2023-04-01 18:56:09 I don't think I'd be transcoding. 2023-04-01 18:57:32 WhyNotHugo: seems my joke went over your head, you: "Any obvious red flags?", me: "the Chinese one? ;-)" 2023-04-01 18:57:50 lol, it totally did 2023-04-01 18:58:50 userdocs_: why would it take a long time to convert (not "transcode") MP3 to FLAC? 2023-04-01 18:59:15 no, you transcode to improve the quality. 2023-04-01 18:59:39 I doubt it would take much longer to convert than the decode (i.e. for playing) the MP3 2023-04-01 19:00:15 userdocs_: you can you improve the quality at all? the quality is what it is (especially MP3 is lossy) 2023-04-01 19:00:32 s/you can/how can/ 2023-04-01 19:01:00 it only works if you have beats by dre headphones 2023-04-01 19:01:42 transcoding between 2 audio formats will either (a) keep the quality the same, or (b) decrease the quality, it can't increase the quality 2023-04-01 19:03:04 unless you have oxygen-free cables and then, yes, the tone/timbre of the audio will improve ;-) 2023-04-01 19:09:26 you can't improve audio by transcoding 2023-04-01 19:09:36 WhyNotHugo: the kodi stuff seems to work but didn't give it a full test 2023-04-01 19:09:40 it's up to date tho so gl 2023-04-01 19:10:20 that's like saying red rgb does not give you more fps in games 2023-04-01 20:00:57 red gives you better amin. UV lights give you more fps. 2023-04-01 20:15:34 i will try and confirm 2023-04-01 20:19:01 What's the point of /boot/config-edge ? 2023-04-01 20:19:38 I understand what a kernel config is used for. What's the point of that file being installed into /boot ? 2023-04-01 20:23:07 i think the base make install of the kernel does that, just old tradition to put the config next to the kernel 2023-04-01 20:25:47 yupe, the config file and System.map typically are put in the same directory as the kernel they are associated with 2023-04-01 20:27:46 if you have the "configs" module loaded you can get the same info from "zcat /proc/config.gz" 2023-04-01 20:32:03 sometimes it's not the same :p 2023-04-01 20:34:06 psykose: ok, splitting hairs, if you've updated kernel package and not yet rebooted.... ;-) 2023-04-01 21:51:35 Is there a simple way to check if any of the installed packages is only existing in the testing repository? I want to check this before disabling it. 2023-04-01 23:13:32 Hello, I'm thinking of installing alpine Linux on my laptop in data mode so that the system runs in RAM and I can easily undo changes with lbu but now I've seen these changes are saved only as .apkvol file and so on the file system and that does not sound like it is intended for laptop. Is it possible to incrementally save the changes not as .apkvol so that you can still undo it? 2023-04-02 01:08:59 So guys I got the kernel to nearly compile (thanks minimal it was a version mismatch lol), but then suddenly it starts building nls codepage modules even though those those are disabled 😭 2023-04-02 01:09:31 Ideally in my case everything would be compiled into the kernel instead of having to deal with module stuff 2023-04-02 01:52:08 It also compiled all filesystems as modules even though they were set to =y 2023-04-02 02:18:57 Should I just go for a standard kernel build instead of a package 2023-04-02 02:29:56 cockliuser[m]: did you use menuconfig? or did you just hand edit .config? 2023-04-02 02:30:38 menuconfig on the 6.1 tar.xz 2023-04-02 02:31:33 and after that, did you copy your saved .config back to APKBUILD dir? 2023-04-02 02:32:04 with a name like config-x86_64 or something 2023-04-02 02:32:10 Yeah 2023-04-02 02:33:16 are you building -edge or -lts 2023-04-02 02:33:23 -lts 2023-04-02 02:33:43 ok. -lts uses a different method to generate config 2023-04-02 02:33:53 a cleaned version of config. 2023-04-02 02:35:36 cockliuser[m]: https://git.alpinelinux.org/aports/tree/main/linux-lts/APKBUILD?h=3.16-stable#n173 2023-04-02 02:35:51 they use make savedefconfig 2023-04-02 02:36:58 and cp "$_builddir"/defconfig "$startdir"/$_config 2023-04-02 02:37:11 Ah ok 2023-04-02 02:37:33 I think that package is supposed to have modules so that might be the reason why it's building modules even though the config disables them 2023-04-02 02:39:22 linux-edge is different : https://git.alpinelinux.org/aports/tree/community/linux-edge/APKBUILD?h=3.16-stable#n97 2023-04-02 03:17:14 Ermine: openrc is not running 2023-04-02 06:28:50 busybox has runit ... 2023-04-02 07:12:10 pj: cool 2023-04-02 07:13:00 cool 2023-04-02 07:14:17 cool 2023-04-02 07:15:08 it was not cool 2023-04-02 07:16:57 yes it was ;-; 2023-04-02 08:43:21 I found a small bug: when both wpa_supplicant and networking both are added to default, networking is started before wpa_supplicant. Because of this udhcpd tries to get an address from something not connected. 2023-04-02 08:44:39 wpa_supplicant should be run before networking. 2023-04-02 08:45:37 it's supposed to 2023-04-02 08:46:56 please pastebin /run/openrc/deptree 2023-04-02 08:48:23 https://pastebin.com/pF0QzbWk 2023-04-02 08:48:43 that's not the whole file 2023-04-02 08:48:48 that's not even a useful part of the file 2023-04-02 08:48:57 pastebin *the whole file* 2023-04-02 08:49:00 Ok! 2023-04-02 08:51:37 https://pastebin.com/9BJqjSx4 2023-04-02 08:59:44 dwfreed: ^ 2023-04-02 09:14:43 I'm looking at it 2023-04-02 09:25:11 your deptree looks fine: https://i.imgur.com/YLW3UNL.png 2023-04-02 09:25:37 wpa_supplicant has 'before net' 2023-04-02 09:25:47 and networking has 'provide net' 2023-04-02 09:29:43 dwfreed: Still, during boot wpa_supplicant is started after networking. 2023-04-02 09:30:34 enable rc_logger in /etc/rc.conf, reboot, and then pastebin /var/log/rc.log ? 2023-04-02 09:30:34 It means that the boot process is slowed down while udhcpd tries to get a dhcp from wlan0, which is not connected because wpa_supplicant gets started after. 2023-04-02 09:31:27 I would guess wpa_supplicant is actually starting first, but is not connected before udhcpc tries to get a lease 2023-04-02 09:32:01 connecting to a wifi network 1) is not instant, and 2) does not finish before the wpa_supplicant service considers itself started 2023-04-02 09:33:43 you can use wpa_cli service to run udhcpc automatically once the wifi network actually connects instead of using networking for this 2023-04-02 09:34:19 ah, actually looks like you need both, it does not start udhcpc itself 2023-04-02 09:37:13 https://pastebin.com/sjiYsiY5 2023-04-02 09:37:34 dwfreed: There you have the rc.log file :) 2023-04-02 09:37:39 you're starting wpa_supplicant in boot runlevel 2023-04-02 09:37:55 and you're starting networking in sysinit 2023-04-02 09:37:58 both of which are wrong 2023-04-02 09:38:04 before only works in the same runlevel 2023-04-02 09:40:13 (you said you put them in default runlevel; neither of them are in the default runlevel) 2023-04-02 09:42:11 dwfreed: You are right. I have now removed from sysinit, both are in boot now. 2023-04-02 09:43:00 they should both be in 'default' runlevel, not 'boot' 2023-04-02 09:43:09 there's a difference, and it's important 2023-04-02 09:47:53 Ok, then I move them there 2023-04-02 10:44:47 udhcpc does not do carrier detection. dhcpcd does. 2023-04-02 10:45:02 when carrier lost, udhcpcd just do nothing 2023-04-02 10:47:37 Hi, I'm thinking of installing Alpine Linux on my laptop in data mode so that the system runs in RAM and I can easily undo changes with lbu, but now I saw that these changes are saved only as .apkvol file and so not on the file system. Is it possible to save the changes so that you can still see them for example in /etc on the filesystem when the machine is not started, like in btrfs as snapshots? 2023-04-02 11:21:42 "Hi, I'm thinking of installing..." <- no 2023-04-02 11:25:39 pj: What is the advantage of saving the changes as apkvol. 2023-04-02 11:26:06 clius[m]: they are saved on persistent medium? 2023-04-02 11:26:36 Ah ok 2023-04-02 12:18:38 hello. a short q: pip check output: yt-dlp 2023.3.4 requires pycryptodomex, which is not installed . is this dependency really necessary, or leave it as it is? 2023-04-02 12:27:45 oleovit: in case you want to install it: https://pkgs.alpinelinux.org/packages?name=py3-pycryptodome*&branch=edge&repo=&arch=x86_64&maintainer= 2023-04-02 12:35:01 @ikke thx, i see its available py3-pycryptodomex-3.17.0-r0 . pycryptodomex does not "interfere" with the old pycrypto i hope. is there a reason why it's not a dependency as yt-dlp alpine package? 2023-04-02 12:39:16 oleovit: I don't know 2023-04-02 12:40:04 pycryptodomex is installed as Cryptodome, so it should not interfere with Crypto 2023-04-02 12:42:06 thx ikke, like always. should i open an issue, regarding this dependency? 2023-04-02 12:42:37 if it's complaining that it's missing, then yes 2023-04-02 12:44:38 a little cryptography can almost never be that wrong. let's see.thx and have a nice day 2023-04-02 13:40:45 according to the readme, pycryptodomex is optional as it always was: https://github.com/yt-dlp/yt-dlp#dependencies 2023-04-02 15:11:12 Hi 2023-04-02 15:11:15 Me again. 2023-04-02 15:12:24 Got another problem with Kodi. The addon psykose updated recently fails to work, this time with another error. I suspect that this version's simply not supported by Kodi. 2023-04-02 15:12:47 (couldn't find much about the error I got) 2023-04-02 15:19:55 i guess we just have to read your mind to figure out what the error is 2023-04-02 15:23:44 psykose: just remember to leave a bookmark in it when you've finished so you don't have to read their mind right from the start next time ;-) 2023-04-02 15:23:59 It's quite vague. Kodi just says the addon couldn't be loaded. 2023-04-02 15:24:16 But I guess it's time to post some logs! 2023-04-02 15:24:21 (if any) 2023-04-02 15:36:48 Ok......Will get logs later 2023-04-02 15:37:04 It's not that much of a priority... 2023-04-02 18:37:10 Hi guys 2023-04-02 18:37:45 I initialized EC2 instance with alpine-3.17.3-aarch64-uefi-cloudinit-r0 but I don't know the default root password 2023-04-02 18:38:59 AlpineRootPassword: There is a user called alpine with doas enabled 2023-04-02 18:39:05 https://www.alpinelinux.org/cloud/ 2023-04-02 18:39:42 hmm, I will try it 2023-04-02 18:40:57 the password is a random uuid :p 2023-04-02 18:41:05 and the account is locked as well pretty sure 2023-04-02 18:41:14 psykose: you mean root? 2023-04-02 18:41:24 yes 2023-04-02 18:41:33 yeah, you are supposed to login to a non-root user 2023-04-02 18:42:17 OK, now understand the situation :) 2023-04-02 18:43:55 thanks 2023-04-02 18:54:35 yeah it's a AWS Marketplace requirement for the "root" account of an AMI to be locked 2023-04-02 18:58:11 AlpineRootPassword: also for the non-root account (i.e. "alpine") password access is also (AFAIK) disabled and you'd need to provide one (or more) SSH public key in the user-data you passed when you created the instance to permit key-based SSH login to work 2023-04-02 20:05:23 whats a nice POSIX simple PS1? (is there a default?) 2023-04-02 20:05:41 *whats a simple nice POSIX PS1? (is there a default?) 2023-04-02 20:07:11 PS1='$' 2023-04-02 20:07:47 \h \w \$ 2023-04-02 20:08:10 When I log in with lightdm it takes a long time to see xfce, the same when logging out and changing the background. Has anyone had the same problem and can help me? 2023-04-02 20:08:17 default for alpine at least is '\w $' 2023-04-02 20:08:37 abby: :D 2023-04-02 20:08:52 ikke: is \w posix? it doesnt work in dash 2023-04-02 20:12:21 POSIX seems to define very little for PS1 2023-04-02 20:12:35 https://pubs.opengroup.org/onlinepubs/9699919799/utilities/V3_chap02.html 2023-04-02 20:15:06 hmmmmmhm 2023-04-02 20:18:39 posix defines practically nothing for interactive shells 2023-04-02 20:18:50 and dash isn't really meant to be one so i bet it doesn't support anything there either 2023-04-02 20:19:53 abby wasn't joking :p 2023-04-02 20:24:31 Rather '$ ', as specified 2023-04-02 20:24:57 '$' is more minimal than '$ ' 2023-04-02 20:26:09 ACTION minimises abby 2023-04-02 20:26:38 '' is even more minimal, but that's not very nice 2023-04-02 20:30:16 What? you guys are not using '\e[1;31ma\e[1;33ml\e[1;32mp\e[1;34mi\e[1;36mn\e[1;35me\e[0m\033[30;37;5m😊\e[0m\w '???? 2023-04-02 20:34:58 :) 2023-04-02 20:52:37 Every time I run a CPU intensive process, my fans immediately speed up. Like, I start Firefox and fans blast to max for a moment and then wind down. 2023-04-02 20:52:52 My impression is that as soon as CPU usage/freq goes up, fans go up too. 2023-04-02 20:53:03 But honestly, fans should work proportionally to temperature, not usage/freq. 2023-04-02 20:53:30 Am I missing some kmod that handles this properly...? I've no idea where to start here. 2023-04-02 20:54:04 temperature goes up with cpu usage 2023-04-02 20:54:56 quinq: Right, but if CPU spikes to 100% for 100ms, I'd rather fans not blast to max too. 2023-04-02 20:55:16 Probably a kernel bug 2023-04-02 20:58:21 i doubt the fans are even controlled from userspace in that case 2023-04-02 20:58:29 or even kernelspace :p 2023-04-02 20:58:42 WhyNotHugo: is it a laptop? 2023-04-02 20:59:11 No, a desktop. Intel-based. 2023-04-02 20:59:16 not sure then 2023-04-02 20:59:22 Do you dualboot windows? 2023-04-02 20:59:34 on laptops specifically it does take like 1 second to reach 90C without much effort with turbo boosting 2023-04-02 20:59:38 Power management is also terrible, it uses 50W idle, and I also have to figure that out at some point. 2023-04-02 20:59:51 Ermine: nope, just Alpine 2023-04-02 21:00:14 what fan management settings do you have in bios 2023-04-02 21:01:00 psykose: I haven't honestly tinkered with them much, pretty much defaults. What should I be looking for / choosing? 2023-04-02 21:01:22 i dunno! it just sounds relevant because it's usually where all the configuration is 2023-04-02 21:01:36 i don't know of any fan management from a booted system that ever applies by default 2023-04-02 21:01:48 meaning if you didn't configure anything this is probably a bios thing and nothing to do with what you're running 2023-04-02 21:02:19 in my experience a lot of bios stuff defaults to some-high-fan-speed at like 60C which you can hit instantly on turbo cpus, even desktop ones 2023-04-02 21:02:45 usually there is a curve-like thing you can configure, if not literally a curve then a fan % at certain C 2023-04-02 21:02:53 and a few of them 2023-04-02 21:06:06 I'll have a look at fan curves in the BIOS 2023-04-02 21:43:24 "they are saved on persistent..." <- What is the advantage of using a read only filesystem? 2023-04-03 07:37:16 i gave up trying to run alpine on a laptop :( 2023-04-03 07:38:01 Why? 2023-04-03 07:39:10 because i couldnt get nm-plasma to work, and that in itself isnt a real problem as there is ofcourse nm-tui and all, but then after an update it failed to boot again :) 2023-04-03 07:49:07 am looking for cephadm and I don't see it in the different ceph17 package ? 2023-04-03 07:49:15 plasma-nm you mean? Works fine for me here 2023-04-03 07:49:26 You sure you have networkmanager running? 2023-04-03 07:49:35 Although booting at all seems like a priority lol 2023-04-03 08:03:20 blt[m]: where do you expect to find it? `cephadm` as a command is not a thing in general? 2023-04-03 08:04:12 hmm 2023-04-03 08:04:21 psykose: I was looking for it like in ceph17-utils more like an utility 2023-04-03 08:04:52 there is perhaps a good reason that we don't package it, but as it is largerly use for ceph I am just wondering why 2023-04-03 08:07:21 the build system doesn't install it 2023-04-03 08:07:56 has to be manually copied i guess 2023-04-03 08:23:13 would you think it deserve me to speak with package maintainer to see his opinion if we can think of adding it ? 2023-04-03 08:23:32 go for it 2023-04-03 10:09:07 Does each package from offical aports has its id? 2023-04-03 10:09:29 it has a name 2023-04-03 10:12:10 name is the same when different branch and arch 2023-04-03 10:14:01 There is no unique id for a package 2023-04-03 10:15:53 there are multiple fields which you can use to uniquely identify package 2023-04-03 10:16:07 but there isn't such field in itself 2023-04-03 10:16:48 technically you could use datahash field 2023-04-03 10:18:18 eventually a checksum field in index 2023-04-03 11:16:49 thx for the yt-dlp change: https://git.alpinelinux.org/aports/commit/?id=ed97a9bc51235fba51a51c9c3329c1268b1a2757 . one dependency is atomicparsley. but this dependency remians in "testing". i installed it via my tagged/pinned testing repo. but the yt-dlp pkg remains in community. so,maybe it's better to remove atomicparsley as dependency, or moving atomicparsley to edge/cummunity? 2023-04-03 11:17:43 ..my english.. 2023-04-03 11:34:27 https://github.com/yt-dlp/yt-dlp/blob/master/requirements.txt 2023-04-03 11:37:24 nvm that link 2023-04-03 11:56:34 thx userdocs_ 2023-04-03 11:57:09 no, i didn't meant to post it, but 2023-04-03 11:58:26 if it was not a depend, you can get ti from here https://github.com/wez/atomicparsley/releases/download/20221229.172126.d813aa6/AtomicParsleyAlpine.zip 2023-04-03 11:58:51 lol 2023-04-03 11:59:24 i was messing around building it. 2023-04-03 12:00:27 ;) i understand 2023-04-03 12:03:01 > Thumbnail embedding in mp4 is done with mutagen if possible. Use --compat-options embed-thumbnail-atomicparsley to force the use of AtomicParsley instead 2023-04-03 12:03:36 https://github.com/yt-dlp/yt-dlp#metadata 2023-04-03 12:11:09 thx again userdocs_ . however, i'm only considering tinkering around here. mainly i use pipe-viewer, with no fallback to yt-dlp. i hardly have any cases where i use yt-dlp. i just wanted to briefly share that since the split/update, one depenency can only be found in testing. 2023-04-03 12:18:01 hello, my initramfs is unable to mount automatically the root filesystem to /sysroot 2023-04-03 12:18:39 mount: mounting /dev/md127 on /sysroot failed: No such file or directory 2023-04-03 12:19:28 then in falls back on emergency initramfs shell, where I ca remroduce the error by typing `mount /dev/md127 /sysroot` 2023-04-03 12:20:10 I have to type `mount -t ext4 /dev/md127 /sysroot` then it works and it can boot 2023-04-03 12:29:12 raspbeguy: do you have mdadm on initramfs? 2023-04-03 12:29:19 yes 2023-04-03 12:29:43 mdadm.conf as well? 2023-04-03 12:30:14 yes. but I don't think the error comes from mdadm 2023-04-03 12:30:38 indeed just adding the type of the filesystem is enough to mount 2023-04-03 12:30:39 what is your fstab line for / ? 2023-04-03 12:31:00 tried with UUID=... and with LABEL=... 2023-04-03 12:32:07 bb mount doesn't really guess and will always fail 2023-04-03 12:32:20 you can rectify the -t to be there always with rootfstype=ext4 2023-04-03 12:33:12 oh ok, I did look for such kernel option but didn't find it, thanks 2023-04-03 12:33:20 yes, I have rootfstype=ext4 on cmdline as well 2023-04-03 12:34:06 afaik bb will try a predefined list of filesystem types 2023-04-03 12:34:08 let me reboot the server in rescue mode and try it (can takes a few minutes) 2023-04-03 12:34:21 it's not a kernel option, mkinitfs reads it 2023-04-03 12:34:29 ikke: only what is in /etc/filesystems iirc 2023-04-03 12:34:33 Yes 2023-04-03 12:34:35 which is not present at all in that case 2023-04-03 12:34:50 and is also terrible since it just probes every modue 2023-04-03 12:35:05 psykose, where do you put this setting then? 2023-04-03 12:35:12 on the cmdline 2023-04-03 12:35:22 same place you were going to put it 2023-04-03 12:35:26 oh ok 2023-04-03 12:35:41 ah 2023-04-03 12:35:42 no 2023-04-03 12:35:44 kernel does read it 2023-04-03 12:35:51 hehe 2023-04-03 12:36:14 it's quite useless afaik, since the kernel probes anyway, though this is all superceded by initrd and not relevant 2023-04-03 12:36:29 just put it to cmdline :) 2023-04-03 12:39:55 psykose, that's working, thanks 2023-04-03 12:41:32 man apk-add not showing full pages, just half 2023-04-03 12:44:25 ohh, it is not a bug. the man pages are separated 2023-04-03 15:09:14 Is there any proper way to replace a APK installed binary on Alpine (maybe something similar to alternative on Ubuntu)? 2023-04-03 15:09:47 There is no alternatives system 2023-04-03 15:10:01 It depends on what you want to achieve 2023-04-03 15:10:16 But putting it higher in the PATH hierarchy might work 2023-04-03 15:17:54 well, there is some ad-hoc support for that, e.g. for busybox 2023-04-03 15:19:30 there isn't 2023-04-03 15:22:39 ovf: Busybox is a "special case" where if the "full fat" version of a command is not installed then Busybox provides it via a symlink. But that's not a "true" alternatives system, it is more of a "fallback" mechanism as the busybox package is (almost) always installed 2023-04-03 15:24:30 minimal: i count that as an ad-hoc alternative, in that i can choose between a separately packaged command and the one that comes with busybox 2023-04-03 15:24:53 that's not alternative at all 2023-04-03 15:26:11 ovf: you can call it whatever you want 2023-04-03 15:28:41 certainly in Debian terms "alternatives" mean you can have multiple programs installed and chose which is the default, e.g. multiple vi clones and decide which one the "vi" command runs 2023-04-03 15:30:36 whereas the example of "login" (provided by busybox, util-linux-login and shadow-login) doesn't work that way 2023-04-03 15:40:58 ln -sf /usr/bin/vim /usr/local/bin/vi Androw 2023-04-03 15:41:55 alternative is basically symlinks. 2023-04-03 15:42:23 you can create a similar script to achieve that 2023-04-03 15:46:25 thanks all 2023-04-03 15:47:06 uniqNuuid, I might go this way. But if something reference /usr/bin/vim it won't work 2023-04-03 15:56:45 list all installed pkgs from community: apk policy $(apk info) | grep -B3 community | awk '/policy/ {print $1}' 2023-04-03 15:56:54 is there a better way? 2023-04-03 19:13:55 With biber, got the following error message: ERROR - Error: Found biblatex control file version 3.9, expected version 3.10. This means that your biber (2.19) and biblatex (3.18) versions are incompatible. 2023-04-03 19:41:14 nvmd, fixed. 2023-04-03 19:52:30 As a source of vulnerabilites, ioctl is kind of like a pandora's box, right? 2023-04-03 19:53:51 As a source of vulnerabilities, syscalls are kind of like a pandora’d box, right? 2023-04-03 22:49:07 I'm having a problem with my alpine VMs - /sys/class permissions are set to 700, and so are all the subdirectories (as far as I can tell), and this is preventing prometheus node exporter from being able to collect metrics 2023-04-03 22:49:35 I checked non-alpine systems and they are not configured this way, they are 755 2023-04-03 22:49:50 what is the correct way to deal with this issue? I don't want to run node exporter as root 2023-04-03 22:53:29 this is a change from older versions of alpine, too - I have 3.13 systems that have been up for a while that do not have this issue 2023-04-03 23:23:45 qman: find out why the permissions are set that way, I'm not seeing that on VMs here, it's 755 2023-04-03 23:28:44 I don't know where this could be configured, /sys being dynamic gets created at boot time, so it must be some setting somewhere, I haven't changed anything related to this, at least not intentionally 2023-04-03 23:29:17 I did find a different alpine VM where this isn't happening, so yeah, it's not strictly an upgrade issue 2023-04-03 23:33:46 qman: /sys in general is handled at boot by /etc/init.d/sysfs 2023-04-03 23:34:52 qman: so reboot your VM and check the permissions again 2023-04-03 23:35:12 this is happening on a fresh reboot 2023-04-03 23:35:24 I noticed it because I was upgrading systems to 3.17 and after reboot, node exporter crashed 2023-04-03 23:39:29 that niit.d script comes from the openrc package which hasn't been updated in v3.17 for 4 months 2023-04-03 23:40:07 do you have any 3.17 systems where the permissions are correct? 2023-04-03 23:40:19 yes, I have one that it is correct 2023-04-03 23:40:44 I checked /etc/init.d/sysfs and /etc/init.d/sysfsconf md5sums and they match exactly with one of the broken systems 2023-04-03 23:41:04 ok, so that would appear to confirm that there is nothing fundamentally wrong with basic Alpine and that something about the problematic machines is causing the issue 2023-04-03 23:52:04 ok, I think I might have found the issue, seems to only be the case when using the linux-hardened or linux-grsec (older versions) kernels 2023-04-03 23:52:17 the one that works is using linux-lts instead 2023-04-03 23:54:51 qman: how are you using linux-hardened or linux-grsec kernels with Alpine v3.17????? 2023-04-03 23:55:23 not sure, I've just been upgrading 2023-04-03 23:55:27 3.17 does not come with those kernel packages 2023-04-03 23:55:43 well, it apparently didn't replace them 2023-04-03 23:57:46 I don't see linux-hardened since Alpine v3.7 amd linux-grsec since v3.5 2023-04-03 23:58:42 I assume Releases Notes in the past mentioned their removal - you DO bother to read the Release Notes before upgrading to the next Alpine Release? ;-) 2023-04-04 00:00:53 https://alpinelinux.org/posts/Alpine-3.6.0-released.html 2023-04-04 00:01:02 "The -grsec kernel-related packages have been renamed to -hardened" 2023-04-04 00:01:06 using a kernel not updated in 5+ years is just peak security :p 2023-04-04 00:01:09 https://alpinelinux.org/posts/Alpine-3.8.0-released.html 2023-04-04 00:01:23 "End of support for hardened kernel (unofficial Grsecurity)" 2023-04-04 00:02:55 the Alpine 3.8 the kernel package was linux-vanilla, it wasn't renamed to linux-lts until more recently 2023-04-04 00:04:19 all I can assume is that you upgrade directly from an old release to another release multiple versions ahead and so any linux-grsec->linux-hardened->linux-vanilla->linux-lts checks were "defeated" by you upgrading in such a big hop in versions 2023-04-04 00:05:56 ah, makes sense 2023-04-04 00:06:14 I definitely skipped some versions this time and probably did in the past 2023-04-04 00:06:22 which would explain the broken chain 2023-04-04 00:06:44 having said that, I don't see any logic in v3.8 linux-vanilla to cope with linux-hardened 2023-04-04 00:07:51 you never noticed you were running 4.9 kernels currently? lol 2023-04-04 00:09:23 not really, no 2023-04-04 00:10:21 I have updates run automatically and every couple years or so I upgrade the major release, otherwise, if it ain't broke, I don't fix it 2023-04-04 00:11:04 if node exporter didn't break, I probably still would not have noticed 2023-04-04 00:11:47 if it ain't broke? the linux-hardened kernel in Alpine 3.7 hasn't been supported/updated (including security fixes) since Nov 2019 lol 2023-04-04 00:12:09 sure, but it kept working 2023-04-04 00:12:51 you said "if it ain't broke", to most people lack of important security fixes == broke 2023-04-04 00:13:17 which is actually one of the strengths of alpine, I've got a couple packages that I am deliberately running very old versions of because they were never fixed properly in newer releases, that continue to work 2023-04-04 00:13:57 I just didn't notice before now that they weren't being updated 2023-04-04 00:13:57 never thought of helping to get them fixed instead? 2023-04-04 00:14:12 I attempted, nothing came of it, and I don't have the time to be a package maintainer 2023-04-04 00:14:36 I submitted a bug report that got zero attention 2023-04-04 00:14:49 and they're not linked again libs that *have* been updated on your machines? 2023-04-04 00:15:01 seems not 2023-04-04 00:15:11 which packages? 2023-04-04 00:15:36 backuppc, I'll have to do some digging to find the version and my old bug report 2023-04-04 00:15:44 it might have even been dropped entirely by now 2023-04-04 00:15:45 this was ages ago 2023-04-04 00:16:34 removed before Alpine v3.14 2023-04-04 00:17:10 yeah, the version I'm running is from v3.4 2023-04-04 00:17:21 it was broken from then on 2023-04-04 00:18:02 that system's pretty out of date, on 3.10 2023-04-04 00:18:57 I must have forgot about it when I did the 3.13 upgrades on everything else 2023-04-04 00:19:42 seems removed as no active maintainer 2023-04-04 00:21:32 I don't see any other ones I'm pinning to old versions right now but I know I did in the past, I think one was my mail server when postfix had a major overhaul and needed complete reconfiguration 2023-04-04 00:22:01 qman: so you're now on 3.13 for your machines? 2023-04-04 00:22:21 most of them, I'm in process of upgrading those to 3.17 2023-04-04 00:22:34 since it dropped off support a few months ago 2023-04-04 00:22:54 ok, was going to point out 3.13 went out of support in Nov last year 2023-04-04 00:22:58 yeah 2023-04-04 00:23:12 and v3.14 goes out of support in 1 months time 2023-04-04 00:27:19 ever though of running RHEL instead? Seems like it's more your speed, I think they still support the 2.6.x kernel ;-) 2023-04-04 00:27:26 s/though/thought/ 2023-04-04 00:29:17 so, another upgrade related issue 2023-04-04 00:29:39 apparently my wireguard VPN is using the wireguard-virt package which disappeared in 3.13, not sure what changes I need to make to replace it 2023-04-04 00:29:52 or if that's just no longer needed 2023-04-04 00:30:37 I see wireguard-go 2023-04-04 00:31:15 but that might be old too 2023-04-04 00:31:35 because wireguard is now in the kernel 2023-04-04 00:31:49 ok, so I should just be able to remove that 2023-04-04 00:31:53 and boot the new kernel 2023-04-04 00:32:32 wireguard-virt was the out-of-kernel driver for wireguard 2023-04-04 00:33:51 https://git.alpinelinux.org/aports/commit/community/wireguard-lts/APKBUILD?h=3.13-stable&id=42bcbfb31c32767344702bd6b6ae921491488b42 2023-04-04 00:33:54 cool, that worked perfectly 2023-04-04 03:04:35 hello 2023-04-04 12:12:20 is anyone using edge for production cases? 2023-04-04 12:13:38 i don't know… 2023-04-04 13:16:35 Does anyone here have *.m4a files? And what do you use to tag them? 2023-04-04 13:20:24 I noticed there is a riscv64 arch for edge. What boards does it support? 2023-04-04 13:23:25 I'm trying to get an rpi to work over serial port. I'm using a FT232 TTL-USB adapter, but get no serial output from the device. 2023-04-04 13:23:38 I added enable_uart=1 to usercfg.txt and console=serial0,115200 to cmdline.txt 2023-04-04 13:38:33 try with console=ttyAMA0,11500 2023-04-04 13:41:55 mercenary: was 11500 a typo for 115200? 2023-04-04 13:42:32 somercet1, i'd use easytag and/or kid3 2023-04-04 13:42:52 WhyNotHugo: it was, sorry 2023-04-04 13:50:56 I love how everyone is either "I just upgraded from Alpine v3.2 to v3.17" or "Is Edge safe for production" and nothing in between 2023-04-04 13:53:15 pj: those in between have fewer issues, so you don't hear from them 2023-04-04 13:54:14 ACTION uses edge in prod 2023-04-04 13:56:37 ACTION uses $current in prod, and $current-1 for things that must not fail like desktop and infra to create prod 2023-04-04 14:17:48 the alpine ISO's root passwd is empty but is able to login, how was it made? i'd like to create such user and is able to `su` that user. 2023-04-04 14:18:45 a. i found it 2023-04-04 14:19:16 adduser -s /bin/sh user, then enter twice. it will be empty passwd 2023-04-04 14:22:40 the vim performance on alpine is kinda terrible. vim --clean some_700lines_of_shell_script 2023-04-04 14:29:07 Also getting no output with console=ttyAMA0,115200 2023-04-04 14:31:54 known-good FT232? you can test by shorting rx/tx. And what happens to me 50% of the time is that I have rx/tx swapped on the Pi 2023-04-04 14:35:57 uniqNuuid, how is neo-vim? 2023-04-04 14:38:55 "vim -Nu NONE" for an actual clean, default vim 2023-04-04 14:42:51 WhyNotHugo: which model of RPI is this? 2023-04-04 14:45:27 lopid: -Nu NONE is fast 2023-04-04 14:45:36 theme performance? 2023-04-04 14:45:52 the so-called defaults.vim 2023-04-04 14:47:15 no syntax hl 2023-04-04 14:47:30 i disabled syntax in my .vimrc, it is fast too. 2023-04-04 14:53:38 ok. syntax off. but everything else remain functional, just now color when browsing the code. 2023-04-04 14:54:03 search shows color though, it's much faster 2023-04-04 14:54:22 your terminal can affect how fast it's redrawn 2023-04-04 14:54:45 terminal? tmux-256color 2023-04-04 14:54:59 that's $TERM 2023-04-04 14:55:17 my terminal is iTerm2 2023-04-04 14:55:17 i meant how efficient your terminal emulator program is 2023-04-04 14:55:46 alpine is on aarch64 on LAN via ssh 2023-04-04 14:56:44 minimal: rpi4 2023-04-04 14:57:17 mercenary: It's a new FT232. I don't have another device on which I can test. 2023-04-04 14:58:25 You mean shorting them with a wire and the running minicom? I should get a lot of rubbiush output, right? 2023-04-04 14:58:42 WhyNotHugo: according to official Rasp Foundation docs, on RPI3 & 4 running Raspbery Pi OS the mini UART is presented as /dev/ttyS0 (whereas the PL011 UART) is /dev/ttyAMA0. Not sure if Alpine behaves the same 2023-04-04 14:58:55 so perhaps you should be using /dev/ttyS0? 2023-04-04 15:02:44 lopid: the performane of terminal is not an issue, it runs on Apple M1. 2023-04-04 15:03:06 well get this guy with his fancy Apple M1. 2023-04-04 15:03:07 it is the syntax hl that effect the performance 2023-04-04 15:03:51 is it any filetype? 2023-04-04 15:04:07 vim --version: Compilation: gcc -c -I. -Iproto -DHAVE_CONFIG_H -Os -Os -flto=auto -D_REENTRANT -U_FORTIFY_SOURCE -D_FORTIFY_SOURCE=1 2023-04-04 15:04:11 -Os? 2023-04-04 15:05:12 you can also check syn-sync-maxlines and -minlines 2023-04-04 15:10:23 WhyNotHugo: disconnect the FT232 from the Pi, connect rx and tx wires together. What you type in minicom should echo back. 2023-04-04 15:37:38 heyo all, I have alpine 3.17 running and I am trying to install `aws-cli` (v2), but `apk update && apk add aws-cli` is giving me v1 still. Not sure why, as I can see `2.11.9-r0` on the alpine packages webpage and I've ensured that the community repo is enabled. 2023-04-04 15:38:15 thelounge57: that's on edge what you see on pkgs.a.o 2023-04-04 15:38:21 unless you select v3.17 2023-04-04 15:38:54 I don't understand. What? 2023-04-04 15:39:21 https://pkgs.alpinelinux.org/packages?name=aws-cli&branch=v3.17&repo=&arch=&maintainer= 2023-04-04 15:39:43 Oh, I see. 2023-04-04 15:39:57 Can I use edge with apk to get the v2 installed? 2023-04-04 15:40:17 that's not supported / likely to break 2023-04-04 15:40:44 I guess I will just curl and install myself then? 2023-04-04 15:41:27 I mean, apk itself can install it just fine, it's just the dependencies that might not match 2023-04-04 15:41:53 Yeah, v3.17 has python3.10 and edge has python3.11 2023-04-04 15:42:27 so the edge package won't work on a v3.17 system 2023-04-04 15:45:47 Okay, good info, thanks :) 2023-04-04 16:39:22 mercenary: I've wired rx to tx and when I type a key in minicom nothing happens. 2023-04-04 16:39:52 Just wiring those two should work, right? I don't need to worry about POWER or GROUND for this kind of test, do I? 2023-04-04 16:41:15 Odd, minicom is frozen. Ctrl-A Z won't show help. 2023-04-04 16:42:05 Oh, never mind, user error. 2023-04-04 16:44:25 just rx/tx works on the ft232's I have 2023-04-04 16:45:11 WhyNotHugo: did you check /dev/ttyS0 like I suggested? 2023-04-04 16:47:33 Dmesg says "FTDI USB Serial Device converter now attached to ttyUSB0"... 2023-04-04 16:47:39 I tried `minicom -b 115200 -D /dev/ttyS0` anyway, same result 2023-04-04 16:49:13 WhyNotHugo: dmesg from where says that? on the RPI? 2023-04-04 16:49:30 dmesg on the host; I'm shorting tx to rx now. 2023-04-04 16:50:17 my point is that the RPI4 has both a UART and a miniUART which on the RPI will be different device names, so are you sure that ttyAMA0 is the correct name for the miniUART? 2023-04-04 16:51:26 WhyNotHugo: I wasn't asking about the device name you use on the host, I was referring to the device name used in the "console=" cmdline entry of the RPI 2023-04-04 16:52:16 minimal: I tried both console=serial0 and console=ttyAMA0 2023-04-04 16:52:29 when you boot the RPI and do a "ls /dev/tty*" what do you see there? 2023-04-04 16:52:57 minimal: I can't run any code on the rpi; I'm not getting serial output. 2023-04-04 16:53:28 lopid: this is the thing makes vim slow : FUNCTION 19_Highlight_Matching_Pair() 2023-04-04 16:53:48 on a RPI running Raspberry Pi OS ttyAMA0 is the UART, whereas ttyS0 is the miniUART 2023-04-04 16:54:20 and serial0 is the primary UART (miniUART) and serial0 is the secondary UART (UART) 2023-04-04 16:54:33 nice 2023-04-04 16:54:33 oops, serial1 for the secondary 2023-04-04 16:55:08 WhyNotHugo: don't you have a keyboard/display for short-term testing before you go headless? 2023-04-04 16:56:29 you could try a different syntax matcher, eg vim-matchup 2023-04-04 16:56:37 pair matcher 2023-04-04 16:58:07 minimal: Do I have to restore console=tty1 for that to work? 2023-04-04 16:58:28 I only have the one monitor and keyboard, so it's kind of a pain, but I can try that quickly. 2023-04-04 16:59:05 you can have multpiple "console=" entries, I think it's the last one that if used by default 2023-04-04 17:13:45 The rpi lists /dev/ttyAMA0,ttyS0,ttyprintk 2023-04-04 17:14:07 so sounds the same as Raspberry Pi OS 2023-04-04 17:14:55 and so ttyS0 is the miniUART that you want (ttyAMA0 is the PL011 UART used for Bluetooth) 2023-04-04 17:15:16 tail g 2023-04-04 17:15:26 https://www.raspberrypi.com/documentation/computers/configuration.html#raspberry-pi-zero-1-2-and-3 2023-04-04 17:15:52 lok at the 2nd table in the "Primary and Secondary UART" section 2023-04-04 17:17:02 Yup, definitely looks like /dev/ttyS0 is correct 2023-04-04 17:25:01 But no input with that. I have a strong impression that the FT232 is not working though; shorting tx and rx does nothing. 2023-04-04 17:27:39 So I'm shorting tx and rx, running `minicom -b 115200 -D /dev/ttyUSB0`, and whatever I types does nothing. Doesdn't sound like i'm doing anything wrong software-wise, right 2023-04-04 17:33:05 Can't say for minicom, but with screen you have to be in a specific group for access to the tty device 2023-04-04 17:33:33 (my general goto is 'screen /dev/ttyUSB0 115200') 2023-04-04 17:36:49 I'm in the dialout group, but I'll try screen before considering the hardware faulty. 2023-04-04 17:43:48 Huh, so shorting rx and tx with screen works. 2023-04-04 17:43:58 So minicom wasn't working for some reason ... ? 2023-04-04 17:44:22 When I press "d", it renders on screen, if I remote the short/wire, it doesn't. 2023-04-04 17:45:48 s/remote/remove 2023-04-04 17:49:06 that sounds promising. Time to connect Pi again, and try with screen 2023-04-04 17:49:29 Yeah, that's what I've tried, but it's not working. 2023-04-04 17:50:56 With what on the Pi? I think we ended up with ttyS0 ? 2023-04-04 17:52:44 console=ttyS0,115200 in cmdline.txt and enable_uart=1 in usercfg.txt 2023-04-04 17:52:50 Pretty confident taht ttyS0 is correct 2023-04-04 17:53:41 Does the ft232 have LEDs to indicate traffic? That helps a lot debugging these things 2023-04-04 17:53:59 It does not. 2023-04-04 17:55:07 alas. then you will have to trust screen. 2023-04-04 17:59:37 Only other thing I can think of is connect monitor/kbd to Pi, 'stty 115200 /dev/ttyS9', and the same for any other /dev/ttyXXX0 you find 2023-04-04 17:59:56 Oh, plugged the power supply to the rpi works... but the ft232 has 5V, that's not enough!? 2023-04-04 18:00:47 the ft232 only has a power line so the output can match 3.3V or 5V levels 2023-04-04 18:01:05 Aaaah. 2023-04-04 18:01:14 It is not meant for running anything off 2023-04-04 18:03:13 Well, minicom still doesn't work, so that was also an issue. But I guess screen is good enough. 2023-04-04 18:03:31 I though screen was a terminal multiplexer like tmux, not a serial console thing. 2023-04-04 18:05:50 what's the usual cause of "IO ERROR" in apk? i don't have time to debug this 2023-04-04 19:16:44 minimal, mercenary: thanks for all the help today! 2023-04-04 21:56:53 is nftables fully supported in 3.17.3? i have forwarding enabled in sysctl, and nftables is using nftables' own router example and dhcp and dns works, but packets don't seem to be forwarded from lan to wan. 2023-04-04 21:57:39 i never have gotten nftables logging working... 2023-04-05 02:12:17 Hi all :) I am trying to use a yubikey to authenticate on a website. On debian i could plug my key in and it just worked magically but here (on firefox or qutebrowser) it did not. does anyone have experience with that? I am a little bit clueless as to what I need to get this to work. 2023-04-05 02:20:17 I love AUR helper drama 2023-04-05 02:20:47 shit, wrong window 2023-04-05 02:49:07 bpalu: I don't believe mdev sets up yubikey/u2f device entries 2023-04-05 02:52:22 indeed eudev doesn't appear to have rules for yubikey/u2f either whereas as systemd-udev does 2023-04-05 03:17:18 the rules are in the yubikey-whatever package 2023-04-05 03:18:59 or somewhere that i forgot 2023-04-05 03:19:20 libfido2 2023-04-05 03:40:07 psykose: ah, right, libfido2, but that would only be for eudev 2023-04-05 04:08:35 continue the vim performance issue, i find it is only alpine that has the issue, void/arch vim is fine, same hardware. 2023-04-05 04:09:32 did you compare vim -v output 2023-04-05 04:09:53 alpine: vim --version: Compilation: gcc -c -I. -Iproto -DHAVE_CONFIG_H -Os -Os -flto=auto -D_REENTRANT -U_FORTIFY_SOURCE -D_FORTIFY_SOURCE=1 || void: Compilation: aarch64-linux-gnu-gcc -c -I. -Iproto -DHAVE_CONFIG_H -fstack-clash-protection -O2 -pipe -march=armv8-a -g -I/usr/aarch64-linux-gnu/usr/include -ffile-prefix-map=/builddir/vim-9.0.1321=. -D_REENTRANT -U_FORTIFY_SOURCE 2023-04-05 04:09:55 -D_FORTIFY_SOURCE=1 || arch: Compilation: gcc -c -I. -Iproto -DHAVE_CONFIG_H -march=armv8-a -O2 -pipe -fstack-protector-strong -fno-plt -fexceptions -Wformat -Werror=format-security -fstack-clash-protection -D_REENTRANT -U_FORTIFY_SOURCE -D_FORTIFY_SOURCE=1 2023-04-05 04:11:06 is it due to -O and -march args? 2023-04-05 04:11:52 the march is the same 2023-04-05 04:12:06 rebuild it with O2 and find out 2023-04-05 04:12:33 abby: if i turn syntax off on vim, it is very fast browsing or editing code (700 lines sh script) 2023-04-05 04:15:32 psykose: i don't have an abuild environment 2023-04-05 04:15:42 very unfortunate for you 2023-04-05 04:17:05 if i create a vm and setup abuild as root, will it build? 2023-04-05 04:26:43 i guess i have to change make flags in abuild's conf file? 2023-04-05 04:27:49 you can just change the export CFLAGS= line in the apkbuild 2023-04-05 04:45:41 psykose: Compilation: gcc -c -I. -Iproto -DHAVE_CONFIG_H -s -flto=auto -O2 -D_REENTRANT -U_FORTIFY_SOURCE -D_FORTIFY_SOURCE=1 2023-04-05 04:45:46 does not make a difference 2023-04-05 04:46:13 let me turn off lto 2023-04-05 04:47:46 still the same :) 2023-04-05 04:50:26 what other flags can i use? 2023-04-05 04:50:40 have you compared the build options that are enabled 2023-04-05 04:50:48 not just the cflags 2023-04-05 04:52:38 good idea 2023-04-05 04:53:55 uniqNuuid: abuild refuses to build as root by default (there is an option to bypass this check, but I recall someone still having issues then) 2023-04-05 04:55:07 i am not using abuild 2023-04-05 04:55:21 but i am using alpine vim's ./configure args 2023-04-05 04:55:41 and i am building on alpine edge in a vm 2023-04-05 04:55:53 -static 2023-04-05 05:44:28 the difference between void and alpine's vim is : the former enabled +gettext 2023-04-05 05:44:56 does this have something to do with syntax hl performance? 2023-04-05 05:44:58 that'll just be localisation 2023-04-05 05:53:40 you should probably learn to profile how much time is spent in where and compare instead of shooting at random things in the dark 2023-04-05 05:53:44 then it is musl vs glibc 2023-04-05 05:53:54 probably 2023-04-05 05:54:01 it did profile.log 2023-04-05 05:54:07 it is regex 2023-04-05 05:57:54 uniqNuuid | lopid: this is the thing makes vim slow : FUNCTION 19_Highlight_Matching_Pair() 2023-04-05 05:59:59 people still use regex highlighting? 2023-04-05 06:00:36 this thing 2023-04-05 06:00:37 https://github.com/vim/vim/blob/master/runtime/plugin/matchparen.vim 2023-04-05 06:01:27 i did n't enable any third party plugins, this is from vim. vim --clean is slow as well. 2023-04-05 06:02:05 it means, as long as there is `syntax on` it is slow. 2023-04-05 06:02:53 it is like browing dmesg with no framebuffer on a console 2023-04-05 06:06:44 autocmd! CursorMoved,CursorMovedI,WinEnter,BufWinEnter,WinScrolled * call s:Highlight_Matching_Pair() 2023-04-05 06:07:26 exactly, when cursor move or scroll with pageup/down , it is slow. especially cursor move up in full screen 2023-04-05 06:08:45 Vim plugin for showing matching parens // it refers { } ? 2023-04-05 06:09:47 oh , not just { }, but also () [] 2023-04-05 09:08:09 bpmedley: You need to install some extra udev rules. 2023-04-05 09:17:53 The rules come with packaeg libfido2 2023-04-05 14:13:35 is eudev required for proper xorg functionality? trying to get it to work with just mdev and keyboard / mouse are not working 2023-04-05 14:21:33 trunc88: install libudev-zero and dont't forget to add yourself to input, audio and video groups 2023-04-05 14:22:47 oh yeah, forgot about libudev-zero, thanks. lets try it 2023-04-05 14:23:02 Also afaik you need to write some helper to rebroadcast events to nlgroup 4 2023-04-05 14:24:35 mah man, it works after installing libudev-zero 2023-04-05 14:31:31 Ermine: can i safely ignore the helper if it seems to just be working now? 2023-04-05 14:43:26 trunc88: write it just in case 2023-04-05 14:58:14 Ermine: i dont actually know anything about that stuff tbh 2023-04-05 15:51:38 trunc88: well, you didn't have any problems so far. In case something doesn't work, consider switching to mdevd (which does that rebroadcasting, among other good things). And if that doesn't work out, you'd need to switch to eudev 2023-04-05 16:02:33 has anyone managed to set up an sshd under initramfs? 2023-04-05 16:02:40 I already asked in the libera channel 2023-04-05 16:03:08 Ermine: oh cool, i was going to switch to mdevd anyway, was just trying to get it working first. didnt realize there was that much of a difference. thanks for the help 2023-04-05 16:22:45 spagnod: yes, and I answered you there ;-) 2023-04-05 16:40:33 ohai 2023-04-05 17:50:06 I use Gentoo musl-llvm  2023-04-05 17:50:07 Hello  2023-04-05 17:51:07 dani-g5x[m]: so? -> #gentoo 2023-04-05 17:51:20 uniqNuuid: maybe let people continue? 2023-04-05 17:51:45 they'll say : sorry wrong channel 2023-04-05 17:52:00 maybe it was a prelude to a question 2023-04-05 17:52:32 ok, then it was a fast guess 2023-04-05 17:52:34 or just a brag, and they are finished. 2023-04-05 17:53:14 userdocs_: like btw i use arch? 2023-04-05 17:53:33 I didn't said that to brag.  userdocs_: I just started using musl-llvm and I like it  2023-04-05 17:54:07 dani-g5x[m]: sry about my wrong guess 2023-04-05 17:54:48 Alpine uses busybox right? I like the concept of it 2023-04-05 17:54:53 #gentoo is also on libera. 2023-04-05 17:55:00 (not oftc) 2023-04-05 17:55:00 dani-g5x[m]: alpine is pure musl, and you don't have to watch the compilation :) 2023-04-05 17:55:05 I know, I'm also there  2023-04-05 17:56:13 musl is good until you encounter performance issue :P 2023-04-05 17:56:43 uniqNuuid: you don't have to alao on Gentoo :D... (full message at ) 2023-04-05 17:56:45 also* 2023-04-05 17:57:20 What performance issues? 2023-04-05 17:57:22 yeah, i am on there too, because of openrc 2023-04-05 17:59:04 dani-g5x[m]: only obviouse on low end hardware, like vim 2023-04-05 17:59:46 s/like vim/ vim on aarch64 for example/ 2023-04-05 17:59:47 My software is being compiled with clang  2023-04-05 18:00:01 So I don't know if this applies to me 2023-04-05 18:00:06 uniqNuuid: Oh I'm on amd64 2023-04-05 18:00:37 most of the perf issues with musl that i've seen are mostly due to malloc, which you can just use another malloc for something that needs it. 2023-04-05 18:00:59 musl has malloc-ng 2023-04-05 18:00:59 easily resolved 2023-04-05 18:01:44 I'm using musl-llvm on a desktop  2023-04-05 18:02:47 invoked: like this? https://github.com/microsoft/mimalloc 2023-04-05 18:02:57 i'll give it a try right now, seconds 2023-04-05 18:06:00 that's one you could try 2023-04-05 18:07:14 invoked: oh my god, awesome LD_PRELOAD=/usr/lib/libmimalloc-secure.so.2.0 vim my_700_lines_bash_code 2023-04-05 18:07:20 it's super fast 2023-04-05 18:07:32 without this , vim is laggy 2023-04-05 18:09:09 i've been talking about vim performance issue all these days, and this mimalloc solved the issue! 2023-04-05 18:09:28 usually people will use another malloc for something like mysql but if it helps. :) 2023-04-05 18:13:50 invoked: now, i'd like to set this system wide 2023-04-05 18:14:56 well, well, grep: memory exhausted 2023-04-05 18:15:45 i wouldn't recommend setting it system wide 2023-04-05 18:16:00 that's just like, my opinion, man 2023-04-05 18:22:33 i've tried also jemalloc, a little bit slower than mimalloc2 2023-04-05 18:22:47 but still much much faster than musl's default one 2023-04-05 18:26:15 scudo-malloc is good too. 2023-04-05 19:00:26 Hi there. Is there really no way to get CUDA in Alpine? I think I already know the answer, but I had to ask anyway, just in case. 2023-04-05 19:00:55 The alternative is the ubuntu docker container, which is enormous. 2023-04-05 19:01:05 If CUDA requires the proprietary nvidia drivers, then no 2023-04-05 19:01:28 Is there a non-proprietary driver then? 2023-04-05 19:01:58 There is nouveau, but I doubt it would support CUDA 2023-04-05 19:02:06 Ah, alright. Well, thanks anyway. 2023-04-05 19:04:14 moonsheep: or get Nvidia to provide driver compiled for musl as well? ;-) 2023-04-05 19:04:38 Man, how I wish I could do that. 2023-04-05 19:05:26 You would win a Nobel peace price 2023-04-05 19:05:40 I currently have an aging server at home (an old HP MicroServer with old spinning rust sata disks). It runs openbsd. I am afraid the hardware will not last for very long. Because of this I am planning a migration to Alpine. 2023-04-05 19:06:28 is jemalloc suitable for use on a whole account? I'd like to try that. 2023-04-05 19:06:45 My plan is to use Alpine running on a cheap SOC with home mounted from Onedrive. This means hardware could be easly replaced while data remains in the cloud. 2023-04-05 19:07:28 What SOC would you recommend? There will not be any heavy usage, mostly irssi, and some family photos etc. 2023-04-05 19:08:52 engblom: if you don't have specific needs then rpi probably makes sense. (not a recommendation, as such) 2023-04-05 19:09:31 engblom_: "spinning rust"? get your Dremel out to remove the rust ;-) 2023-04-05 19:11:17 I will have to migrate away from openbsd because the openbsd port of rclone does not support mounting. 2023-04-05 19:11:23 why run a NAS at home when the NAS doesn't actually store any data, it's "in the cloud"? 2023-04-05 19:11:52 minimal: I still want a shell I can access from anywhere. 2023-04-05 19:12:31 And I want the computer in my own network. rclone support encrypted mounting, meaning that the content in onedrive can only be accessed from the home device. 2023-04-05 19:18:14 hello ppl 2023-04-05 19:18:40 I have a lvm2 question thay maybe you kind people can help me with 2023-04-05 19:18:44 mellow 2023-04-05 19:19:05 I have made thin-provisioned logical volumes on my disk 2023-04-05 19:19:33 I used them with qemu and installed Windows 11 on them 2023-04-05 19:19:47 I'd like to mount them into my Alpine Linux machine 2023-04-05 19:20:00 but I can't find the devices for the life of me 2023-04-05 19:20:07 I can't find the partitions 2023-04-05 19:20:21 usually in /dev/mapper I see devices w partitions 2023-04-05 19:20:42 but I can only see the logical device 2023-04-05 19:20:47 but no partitions 2023-04-05 19:20:53 and I don't know how to mount them 2023-04-05 19:21:02 comically, fdisk "sees" the NTFS partitions 2023-04-05 19:21:26 but I don't have a /dev/device to mount them 2023-04-05 19:21:34 I can't find this information anywhere 2023-04-05 19:21:39 truly puzzled 2023-04-05 19:21:57 i.e.: 2023-04-05 19:21:59 Disk /dev/Baldr/wyk-win11: 300 GiB, 322122547200 bytes, 629145600 sectors Units: sectors of 1 * 512 = 512 bytes Sector size (logical/physical): 512 bytes / 512 bytes I/O size (minimum/optimal): 262144 bytes / 262144 bytes Disklabel type: gpt Disk identifier: 0918C4DD-78F6-483A-A82C-0B26676C42B5 Device Start End Sectors Size Type /dev/Baldr/wyk-win11p1 34 262177 262144 128M Microsoft reserved /dev/Baldr/wyk-win11p 2023-04-05 19:22:31 now I usually do "mount /dev/Baldr/wyk-win11p2 /mnt" 2023-04-05 19:22:39 but /dev/Baldr/wyk-win11p2 does not exist 2023-04-05 19:22:48 wyk72: with LVM you have PVs and LVs 2023-04-05 19:22:53 and I havee no idea how to make it appear in /dev 2023-04-05 19:23:07 Is this lvm? 2023-04-05 19:23:12 lvm2 2023-04-05 19:23:21 vgchange -a y 2023-04-05 19:23:24 it seems you used Qemu to pass a LV into the VM and that LV was treaded as a disk by Windows 2023-04-05 19:23:32 yes exactly 2023-04-05 19:23:44 already did vgchange 2023-04-05 19:23:55 nothing in /dev/mapper 2023-04-05 19:24:05 the partitions are seen by fdisk 2023-04-05 19:24:09 so to the Alpine host the LV is a LV and the /dev/mapper entries reflect that 2023-04-05 19:24:16 yup 2023-04-05 19:24:21 the disk works into qemu 2023-04-05 19:24:26 perfectly 2023-04-05 19:24:32 fdisk "sees" the partitions 2023-04-05 19:24:39 what does lsblk return? 2023-04-05 19:24:41 but nothing in /dev/mapper 2023-04-05 19:24:56 nvme500:/dev/mapper# lsblk NAME MAJ:MIN RM SIZE RO TYPE MOUNTPOINTS sda 8:0 0 223.6G 0 disk ├─sda1 8:1 0 128M 0 part ├─sda2 8:2 0 100M 0 part └─sda3 8:3 0 223.3G 0 part sdb 8:16 1 7.3G 0 disk └─sdb4 8:20 1 7.3G 0 part nvme0n1 259 2023-04-05 19:24:57 so why would you expect to see partition entries in /dev/mapper for a "partitioned LV"? 2023-04-05 19:25:28 minimal: I'd like to mount them via ntfs-3g into native Linux 2023-04-05 19:25:42 there is a way ? 2023-04-05 19:25:51 partprobe returns nothing 2023-04-05 19:26:54 maybe I should mount a loop device with -p parameter ? 2023-04-05 19:27:07 /dev/mapper reflects devicemapper devices (e.g. LVM and LUKS), it doesn't reflect any partitions you may have created on such devices 2023-04-05 19:27:20 minimal: I see 2023-04-05 19:27:59 so there is no way to directly mount a ntfs partition inside ? 2023-04-05 19:28:47 how can fdisk see them (and name them) but the system does not ? 2023-04-05 19:28:53 seems crazy to me 2023-04-05 19:29:22 fdisk looks at the 1st part of whatever you tell it to look at where it then sees a partitioning "label" 2023-04-05 19:29:29 fdisk -l /dev/Baldr/wyk-win11 returns the partitions 2023-04-05 19:29:58 "1st part" == "1st sector(s)" 2023-04-05 19:30:04 like "/dev/Baldr/wyk-win11p3 468992 629143551 628674560 299.8G Microsoft basic data" 2023-04-05 19:30:05 somercet1: what do you mean `whole account` 2023-04-05 19:30:28 where is this "/dev/Baldr/wyk-win11p3" device ? 2023-04-05 19:30:35 such as setting that LD_PRELOAD nii .profile, and running sway under it. 2023-04-05 19:30:41 engblom: you can run alpine in RAM 2023-04-05 19:31:51 wky72: it is at /dev/Baldr/wyk-win11p3 apparently ;-) 2023-04-05 19:32:09 what is /dev/Baldr/ ? 2023-04-05 19:32:21 I see 2023-04-05 19:32:23 I solved it 2023-04-05 19:32:30 using a loop device w partition scan 2023-04-05 19:32:37 i.e.: "losetup -f -P /dev/Baldr/wyk-win11" 2023-04-05 19:32:50 the at loop0p3 I get the NTFS partition 2023-04-05 19:32:58 somercet1: what is what i am doing. i set it in .bashrc 2023-04-05 19:33:11 kind of odd is written nowhere 2023-04-05 19:33:15 wyk72: yes that's how I mount disk images when I'm working on them 2023-04-05 19:33:30 wyk72: what is "written nowhere"? 2023-04-05 19:34:01 that you can't access partitions inside a LVM logical volume 2023-04-05 19:34:19 wyk72: you can, you just confirmed you were able to using losetup 2023-04-05 19:34:24 somercet1: i don't know it will pass LD_PRELOAD to it if you start sway as a rc-service 2023-04-05 19:34:28 usually in /dev/mapper I get all the partitions 2023-04-05 19:34:36 you may cat /proc/pid/environ 2023-04-05 19:34:42 I don't, I run it from the command line 2023-04-05 19:35:10 wky72: nope, in /dev/mapper you get all the devicemapper *devices* 2023-04-05 19:35:20 worth tring somercet1 2023-04-05 19:35:26 an LV is not a partition 2023-04-05 19:35:54 I see 2023-04-05 19:36:11 I've learned something new :) 2023-04-05 19:36:38 to get partitions inside a LV you need a loop device then 2023-04-05 19:37:02 I mean to get access to them 2023-04-05 19:37:09 i.e. "mount" 2023-04-05 19:37:10 there may be other ways but using a loop device is how I would do it 2023-04-05 19:37:21 somercet1: https://github.com/daanx/mimalloc-bench/blob/master/README.md 2023-04-05 19:37:22 That's because the entire LV represents a disk right? 2023-04-05 19:37:34 For most usages, an LV is a partition 2023-04-05 19:37:55 I mean, in the case of wyk72 2023-04-05 19:38:25 ikke: well an LV is a "device", in the same way that a disk is a device, but an LV isn't a disk 2023-04-05 19:38:49 I mean, if you present it to qemu as one 2023-04-05 19:39:05 With a partition table 2023-04-05 19:39:53 somercet1: with my simple test, scudo from llvm.org is a good one 2023-04-05 19:40:11 ikke: depends on the options you pass qemu I guess, you could probably pass it as something other than a disk. I'm assuming Windows then partitioned the device 2023-04-05 19:40:32 I passed the LV to qemu directly as s RAW disk, so it just writes it as if it were a disk 2023-04-05 19:40:48 and indeed it made the usual windows partitions inside 2023-04-05 19:40:49 ikke: e.g. you could write an ISO to the LV and then present it to qemu as a CDROM I presume ;-) 2023-04-05 19:41:29 So linux does not look into the LV to discover more disks 2023-04-05 19:42:13 I guess it could but there's nothing setup to do that automatically AFAIK 2023-04-05 19:43:02 e.g. perhaps you could have mdev/eudev rules to spot something and add device entries (though they wouldn't go in /dev/mapper) 2023-04-05 19:43:37 I think the loop device w part scan -P option is the right way 2023-04-05 19:44:17 generally with LVM you don't bother with partitioning as you create LVs instead 2023-04-05 19:44:32 minimal: exactly 2023-04-05 19:45:34 but this case is a different way of using LVM to handle VM disk images rather than using QCOW2/VMDK/etc files instead 2023-04-05 19:46:21 in both cases you're giving a "blob" of storage to the VM for it to use as it sees fit 2023-04-05 20:09:22 uniqNuuid, I will try it out, ty 2023-04-06 00:05:00 hello, i've had a strange problem with alpine (pmos specifically): at some point, the files provided by pulseaudio-alsa disappeared. I tried apk add and apk fix but neither restored the missing files. Eventually, I just extracted the contents of the apk and put them in the proper locations, but is there a better way to resolve this sort of problem? 2023-04-06 00:06:00 apk fix pulseaudio-alsa would put them back, unless that's not what you had installed and had pipewire-alsa installed instead 2023-04-06 00:18:56 psykose: thanks, that seems to be the case. how can I install pulseaudio-alsa in place of pipewire-alsa? 2023-04-06 00:19:08 i can't use pipewire, since it's not supported by phosh apparently 2023-04-06 00:19:35 can you share /etc/apk/world and `apk policy pulseaudio-alsa` 2023-04-06 00:26:56 here: https://bpa.st/QKK5Q 2023-04-06 00:29:40 perhaps start with removing pipewire-alsa from world 2023-04-06 00:34:37 i've done that. now i get the following error when i run `apk add pulseaudio-alsa`: https://bpa.st/M6DWE 2023-04-06 00:37:40 apk add pulseaudio-alsa !pipewire-alsa 2023-04-06 00:38:43 thanks so much, that's what i was missing 2023-04-06 00:39:24 where can i read more about apk so i don't have to bother people on irc? 2023-04-06 00:40:23 apk --help; apk --help add; etc. ? 2023-04-06 00:43:55 don't think the ! is documented anywhere 2023-04-06 02:05:48 wayland-libs-egl updated to 1.22.0 and now firefox 111.0.1 is crashing on any kind of movie or canvas. 2023-04-06 02:06:57 "Failed GL context creation for hardware WebRender" 2023-04-06 02:29:08 Hi, I messed with my kernel Append line by disabling kms, now I don't have any graphical display. I tried to boot into a live USB environment to change the file, but when I do that the /boot folder of my hard-drive is empty (all other files, such as in my home directory, are still there). Any ideas how I can change or reset my kernel append line? 2023-04-06 02:29:33 I know I'm still booting because when I type "doas poweroff" it does so. I just can't see it 2023-04-06 02:31:04 I tried nano'ing back into the file to remove the changes- I have another laptop on Alpine so I thought I could just replicate the keystrokes. But I must've messed something up, because when I finished and typed "doas reboot" nothing happened. Obviously if I'm SoL I can back up my files and reflash the OS, but if there's a quicker way I'd sooner do that 2023-04-06 02:31:07 is /boot on a separate filesystem that needs to be mounted? 2023-04-06 02:32:08 if you make any changes you have to reupdate the initrd with `apk fix mkinitfs` or whatever, and that needs you to also mount /boot or whatnot (sounds like it's another partition) 2023-04-06 02:34:03 I didn't think of mounting /boot, thank you, I'll go back into the live-USB and try again 2023-04-06 02:38:38 Quillith: try adding "nomodeset" to your kernel "Append" line 2023-04-06 02:40:48 That's what I put in: nomodeset i915.modeset=0 video=eDP-1:1920x1080 2023-04-06 02:41:08 on a BIOS or a UEFI system? 2023-04-06 02:41:16 UEFI 2023-04-06 02:41:23 coreboot seabios 2023-04-06 02:41:40 uh sorry that may not be UEFI, it's coreboot seabios 2023-04-06 02:43:18 I was trying to change my default resolution, since for some reason my laptop keeps defaulting to 4k when it really shouldn't. Normally I just set the res in my Sway config, but I'm trying out tmux and thought I'd be more efficient by changing the res at the source 2023-04-06 02:45:02 ohhh mounting the /boot worked perfectly, I've fixed the error and am back in. Still trying to work out this res, but at least I can use my screen again lol. Thanks for the help! 2023-04-06 07:05:35 hi all, Alpine novice here. I've been using Jekyll on Alpine for a while now, but I have messed something up due to recent upgrades or something. How should I use Jekyll on v3.17; the jekyll package, or via ruby gems? I build a site as a non-root user, but going the gem/bundler route gives me permissions issues. 2023-04-06 07:16:10 is there a way to prevent extracting files upon installing a pkg? like, i don't need all man pages other than English. 2023-04-06 07:17:25 no 2023-04-06 07:22:55 Okay, I think I resolved it: using gem install --user-install, and setting GEM_HOME 2023-04-06 07:23:41 (GEM_HOME w/ bundle install; vs --path) 2023-04-06 08:16:22 install mawk does not trigger busybox? 2023-04-06 08:26:33 it does 2023-04-06 08:28:00 yes it does, after trigger: /usr/bin/awk -> /bin/busybox 2023-04-06 08:28:12 mawk doesn't replace awk 2023-04-06 08:28:16 it's just mawk 2023-04-06 08:28:49 but gawk provide awk? 2023-04-06 08:29:10 yep 2023-04-06 08:41:01 i removed /usr/share/man/^[^m] , then apk fix *-doc, it says no such dir, wonder what package created those dirs 2023-04-06 08:41:52 the same packages that have -doc, apk fix just can't remake the directory tree 2023-04-06 08:42:44 by default apk fix reinstall pkgs, right? 2023-04-06 08:43:14 maybe i should try `apk del docs` and apk add docs? 2023-04-06 15:29:12 apk does not do any logging of what package was installed when? 2023-04-06 15:32:53 somercet1: no 2023-04-06 15:33:30 gonna have to script that to fix it 2023-04-06 15:35:03 as it is, firefox would be unusable without noscript, it keeps crashing on any kind of video content, incl. ads. 2023-04-06 15:39:44 somercet1: apkv3 will have it 2023-04-06 15:40:43 when will that be out? 2023-04-06 15:42:47 no ETA yet 2023-04-06 15:49:32 somercet1: browse your *sh history 2023-04-06 15:49:53 for what? 2023-04-06 15:50:28 logs 2023-04-06 15:51:08 how does searching for "apk -iU upgrade" tell me what was changed? 2023-04-06 15:53:34 hmm. apk does not do a lot of things, which i like. 2023-04-06 15:55:29 "not keeping track of what was changed" on a *rolling distro* is not one of the things you can do without. 2023-04-06 15:56:07 somercet1: note that Alpine Linux is not a rolling distro 2023-04-06 15:56:21 edge isn't? 2023-04-06 15:56:49 depends on how you config the repo 2023-04-06 15:56:58 edge/latest-stable are rolling 2023-04-06 16:33:13 it's a rolling edge :D 2023-04-06 16:39:44 if it gets too close to the edge then it'll start rolling ;-) 2023-04-06 17:29:07 I have 2x alpine liunx vms to which I added to NIC cards one connects over NAT to the host the other NIC is bound to an isolated network. DHCP gives both nics a different ip on startup when i change the state of the second nic to down the first nic which is connected to internet should work but when i ping a website i get nothing 2023-04-06 17:29:26 *to which i added 2 NIC's 2023-04-06 17:29:39 how do I make networking work in this situation 2023-04-06 17:31:23 when the second interface is down it still has the ip dhcp originally gave it 2023-04-06 17:31:26 aiena: how does the route table on the vm looks like 2023-04-06 17:31:34 how do I check 2023-04-06 17:31:37 ip route 2023-04-06 17:34:06 it says `192.168.122.0/24 dev eth0 scope link src 192.168.122.107` 2023-04-06 17:35:02 No default route? 2023-04-06 17:35:06 it looks sort of ok as in its matching the default config in virsh net-list 2023-04-06 17:35:24 i can ping the host so it may not be the vm trying to think 2023-04-06 17:35:53 aiena: just to be sure, did internet work before, or has it never worked? 2023-04-06 17:36:03 it worked before 2023-04-06 17:36:28 Before you brought the 2nd NIC down? 2023-04-06 17:36:42 no it never worked when both the nic's were up 2023-04-06 17:36:46 ok 2023-04-06 17:37:05 let me try disabling the second nic in the config and restart networking 2023-04-06 17:38:29 You need a default route 2023-04-06 17:38:58 ok by disabling second nic it worked 2023-04-06 17:39:15 i guess thats because the default route because the only obvious one 2023-04-06 17:39:44 aiena: what does your /etc/network/interfaces file look like? 2023-04-06 17:39:45 how do I keep both ifaces up and then switch between what is used 2023-04-06 17:41:40 Can you bring both interfaces up, restart networking, and then show the complete output of `ip route`? 2023-04-06 17:43:05 https://paste.opensuse.org/pastes/be52b30f2e95 2023-04-06 17:43:30 i had commented out the eth1 lines and restarted networking to test if it work with only 1 interface configured 2023-04-06 17:43:57 one sec let me keep both up with above config and inform you 2023-04-06 17:45:49 hmm very confused 2023-04-06 17:46:22 if both the interfaces are configured the default is via eth0 2023-04-06 17:46:37 if only eth0 is configured default is also via eth0 2023-04-06 17:47:13 ah but a default route is not set for eth1 2023-04-06 17:47:27 eth1 is your isolated network? 2023-04-06 17:47:31 yes 2023-04-06 17:47:33 You only need one default route 2023-04-06 17:47:43 that sounds like it should work 2023-04-06 17:48:05 yes there is no other difference in the config 2023-04-06 17:48:36 but ping fails with both nics up and if i turn of the second nic without disabling the auto eth1 and related dhcp line 2023-04-06 17:49:06 is each nic in a different subnet? 2023-04-06 17:49:21 yes they are different networks totally 2023-04-06 17:49:38 if both nics are up, can you ping the default gateway? 2023-04-06 17:50:23 yes i can ping the default gateway with both nic's up 2023-04-06 17:50:42 can you ping something like 8.8.8.8? 2023-04-06 17:51:02 yes that works so it could be dns 2023-04-06 17:51:14 aiena: what is the content of /etc/resolv.conf 2023-04-06 17:51:48 it has nameserver as the isolated network gateway 2023-04-06 17:51:56 it has nameserver as the isolated network gateway 2023-04-06 17:52:03 ok 2023-04-06 17:52:04 is this file created automatically? 2023-04-06 17:52:14 it's updated by the dhcp client 2023-04-06 17:52:26 but as each network has dhcp, it's updated twice 2023-04-06 17:52:36 ok so the second nic going up later updated the file 2023-04-06 17:53:34 aiena: do you want dns only from eth0? 2023-04-06 17:54:04 yes 2023-04-06 17:54:13 you can install openresolvconf 2023-04-06 17:54:24 sorry, openresolv 2023-04-06 17:54:46 then edit /etc/resolvconf.conf to add `allow_interfaces=eth0` 2023-04-06 17:54:47 i just reran udhcpc -i eth0 and it fixed resolv.conf 2023-04-06 17:55:00 aiena: until dhcp for eth1 renews 2023-04-06 17:55:12 oh 2023-04-06 17:57:35 aiena: the alternative is to change the dhcp server for your private network to not provide dns 2023-04-06 17:57:37 ikke, there is a like there by default resolv_conf=/etc/resolv.conf 2023-04-06 17:57:49 aiena: that's fine 2023-04-06 17:57:51 *line 2023-04-06 17:57:59 just add it below there 2023-04-06 17:58:42 ok i added it below do I restart networking? 2023-04-06 17:59:04 yes 2023-04-06 18:00:09 ikke, the default route is for going outside the lan network right 2023-04-06 18:00:26 yes, it's the route used in case no other route matches 2023-04-06 18:02:32 ok after adding that line resolv.conf still shows the other gateway 2023-04-06 18:02:43 adding the line and restarting networking 2023-04-06 18:03:11 i.e. `allow_interfaces=eth0` is not working 2023-04-06 18:03:19 hmm 2023-04-06 18:04:33 Does it contain '# Generated by resolvconf'? 2023-04-06 18:04:38 i will try disabling dns at virsh level for the isolated net 2023-04-06 18:04:57 will make it easy for all vms then I guess 2023-04-06 18:05:04 one sec 2023-04-06 18:05:10 yup 2023-04-06 18:05:35 but for learning this is good 2023-04-06 18:06:55 hmm '# Generated by resolvconf' line is not there in resolv.conf 2023-04-06 18:06:57 looks like you can use udhcpc_opts -o 2023-04-06 18:07:13 uniqNuuid: where do you set that? 2023-04-06 18:07:19 -o,--no-default-options Do not request any options 2023-04-06 18:07:27 in /etc/networking/interfaces 2023-04-06 18:07:36 https://wiki.alpinelinux.org/wiki/Configure_Networking#IPv4_DHCP_Configuration 2023-04-06 18:08:42 also read this : https://wiki.alpinelinux.org/wiki/Udhcpc 2023-04-06 18:10:02 I think that udhcpc does not have hooks for openresolv 2023-04-06 18:10:20 ikke, that is the conclusion i came to 2023-04-06 18:10:32 there is a udhcpc.conf though 2023-04-06 18:10:56 On my reference system I manually executed resolvconf and disabled udhcpc updating /etc/resolv.conf 2023-04-06 18:11:00 so I was confused 2023-04-06 18:11:22 maybe in interfaces eth1, use this : udhcpc_opts -o -r 2023-04-06 18:11:37 request only ip 2023-04-06 18:11:40 what is the -r 2023-04-06 18:11:50 o no. 2023-04-06 18:11:54 -r needs an arg 2023-04-06 18:12:35 you may use: udhcpc_opts --no-default-options --request-option=ip 2023-04-06 18:12:36 -r is to request an ip address via dhcp 2023-04-06 18:12:58 lets try with -o alone 2023-04-06 18:13:16 yeah, try -o alone on eth1 , see what happens 2023-04-06 18:15:13 that makes no difference resolv.conf still gets eth1's gateway 2023-04-06 18:17:30 its pretty interesting situation to troubleshoot. 2023-04-06 18:18:10 Any other ideas I'll eventually disable dns att virsh level but this problem may be intresting to solve without it for troubleshooting experience 2023-04-06 18:19:54 aiena: install dhcpcd 2023-04-06 18:20:00 which should work with resolvconf 2023-04-06 18:20:08 and ifupdown-ng will automatically use it 2023-04-06 18:20:35 "dhcpcd may not be the only daemon running that wants to configure DNS on the host, so it uses openresolv to ensure they can co-exist." 2023-04-06 18:20:57 ikke, i removed openresolv 2023-04-06 18:21:07 and purged its config to rule that out 2023-04-06 18:21:26 Well, that's the alternative 2023-04-06 18:21:34 but better to do it on the network level 2023-04-06 18:22:04 how is dhcpd different from udhcpc 2023-04-06 18:22:38 it's different software 2023-04-06 18:22:48 udhcpcd is provided by busybox 2023-04-06 18:24:16 ok 2023-04-06 18:26:52 after reading the script in /usr/share/udhcpc/default.script, i think you can use : NO_DNS="eth1" in /etc/udhcpc/udhcpc.conf 2023-04-06 18:27:11 L110 2023-04-06 18:27:43 the syntax is NO_DNS="eth1 eth2 ..." 2023-04-06 18:43:37 hmm strangely that didnt work 2023-04-06 18:44:06 i noticed that the default.script does not export the NO_DNS variable 2023-04-06 18:48:39 aiena: you sure it is set in /etc/udhcpc/udhcpc.conf 2023-04-06 18:49:52 yes I added it at the end of the file it should work as per the alpine documentation 2023-04-06 18:51:12 wait 2023-04-06 18:51:26 I have my udhcpc.conf directly in /etc 2023-04-06 18:51:44 its not in /etc/udhcpc/ 2023-04-06 18:51:53 then that's the issue 2023-04-06 18:52:07 you need to create /etc/udhcpc/ and move the file in there 2023-04-06 18:52:22 i think udhcpcd.conf is something else 2023-04-06 18:52:27 one sec 2023-04-06 18:52:31 you meed mkdir 2023-04-06 18:52:40 and use /etc/udhcpc/udhcpc.conf 2023-04-06 18:52:47 mkdir 2023-04-06 18:55:19 that worked 2023-04-06 18:55:41 thanks a ton 2023-04-06 18:56:32 ok now I know how to get this working. Now testing how to disble dns for the isolated network at virsh level 2023-04-06 19:17:51 ok seems like at virsh level its not working 2023-04-06 19:18:32 I added to the isolated network definition but alpine still specifies the isolated networks gateway 2023-04-06 19:25:59 guess that may not work https://wiki.libvirt.org/VirtualNetworking.html#Basic_command_line_usage_for_virtual_networks 2023-04-06 19:33:10 back to the alpine method :) 2023-04-06 20:00:34 how do i configure apache2 modules in alpine 2023-04-06 20:00:41 on ubuntu there is a2enmod etc. 2023-04-06 20:02:03 in alpine, it's subpackages you install 2023-04-06 20:02:20 ok so if i put apache2-ssl 2023-04-06 20:02:27 then the ssl module is enabled 2023-04-06 20:02:32 uninstall its disabled 2023-04-06 20:02:33 https://pkgs.alpinelinux.org/packages?name=apache2-*&branch=edge&repo=&arch=x86_64&maintainer= 2023-04-06 20:02:53 See https://pkgs.alpinelinux.org/contents?branch=edge&name=apache2-ssl&arch=x86_64&repo=main 2023-04-07 00:28:05 How the hell does mate-terminal have my LANG variable, but bash gets reset to C.UTF-8? 2023-04-07 00:33:49 oh, the system bashrc imports profile.d... improperly. 2023-04-07 06:19:18 psykose: i think i have fixed slow openrc stop problem. at least on supervise-daemon 2023-04-07 06:20:11 after a strace, i found this call takes 1 second : wait4(-1, [{WIFEXITED(s) && WEXITSTATUS(s) == 0}], 0, NULL) 2023-04-07 06:20:21 somercet1: are you sure it's improper and not just that the default in locale.sh is C.UTF-8 2023-04-07 06:20:37 then go back to openrc source code, it is in a https://github.com/OpenRC/openrc/blob/b9241988a7c32f00ffd6d87f1666d7babfd705e5/src/supervise-daemon/supervise-daemon.c#L674 2023-04-07 06:20:47 if (do_healthcheck) function 2023-04-07 06:21:06 mhm 2023-04-07 06:21:28 so, i set supervise_daemon_args="--healthcheck-delay 0" 2023-04-07 06:21:40 this fixed the issue. 2023-04-07 06:21:48 no more 1 second on each stop 2023-04-07 06:22:23 :) 2023-04-07 06:23:40 gentoo by default, does not have this issue, maybe it is a default setting or they modifed the deley to 0? 2023-04-07 06:26:52 psykose, in the default configuration, bash loads either .profile or .bashrc. profile can load bashrc, but it should not be the other way around. 2023-04-07 06:28:02 aren't those both in the home folder 2023-04-07 06:28:11 bash always lods /etc/profile except under some conditions 2023-04-07 06:30:17 and /etc/profile always sources /etc/profile.d, etc 2023-04-07 06:30:56 wel wel.. supervise-daemon: invalid health check delay 0 2023-04-07 06:31:08 i did not see this error while stracing 2023-04-07 06:32:13 looks like the delay is hardcoded. at least 1 second 2023-04-07 06:33:38 so, it has nothing to do with this option, if i set it to -A 5. still 1 second each stopping 2023-04-07 06:37:34 i straced the wrong thing..... 2023-04-07 06:37:39 bash never loads /etc/profile unless it is a login shell. After I log in, no other shell should touch any profile file. 2023-04-07 06:38:54 my profile has nothing in it except for setting up system variables the way I want them, plus a few more one-time tasks. then it loads my bashrc. 2023-04-07 06:40:09 this : ppoll([{fd=4, events=POLLIN}], 1, NULL, NULL, 8) = ? ERESTARTNOHAND (To be restarted if no handler) 2023-04-07 06:46:36 somercet1: yeah, so on login it loads LANG=C.UTF-8. do you set another LANG in bashrc? 2023-04-07 06:46:41 if not, that's why it's that 2023-04-07 06:47:39 No, ~/.profile loads *after* /etc/profile*, and so bashrc sets no system variables at all. LANG is set in ~/.profile. 2023-04-07 06:50:40 yes 2023-04-07 06:55:47 e.g., aliases and shell functions are set in bashrc, because how could sway or startx pass those along? 2023-04-07 06:56:04 But environment variables are, so they go in ~/.profile, and they load only once. 2023-04-07 06:59:13 yes, we're both talking about the same thing 2023-04-07 06:59:23 i don't know what you have in your .bashrc or .profile 2023-04-07 06:59:40 if i set a different LANG in ~/.profile and open a bash login shell it has the lang set there fine 2023-04-07 06:59:59 then any further children/regular shells also have it 2023-04-07 07:00:09 so i'm not sure what you're asserting is wrong 2023-04-07 07:14:01 in /etc/bash/bashrc, the last stanza loads all the files in /etc/profile.d 2023-04-07 07:14:38 ouch 2023-04-07 07:24:28 good point 2023-04-07 07:24:29 removed 2023-04-07 07:24:31 thanks :) 2023-04-07 07:26:33 np 2023-04-07 12:05:22 Does `swaynag -h` (from edge) segfault for anyone else here? 2023-04-07 12:06:36 no, but i see it has malloc/free issues via other means 2023-04-07 12:06:49 ah 2023-04-07 12:06:50 it does 2023-04-07 12:06:54 right after printing the help 2023-04-07 12:06:58 it does still work funnily 2023-04-07 12:07:38 looks like https://github.com/swaywm/sway/commit/5e73acb431258d81a01709fbd28da2e151bd3bd3 2023-04-07 12:08:39 fixed 2023-04-07 12:09:39 thanks :) 2023-04-07 12:15:56 when to replace openrc? 2023-04-07 12:16:58 any minute now, it's but 2 lines of code away 2023-04-07 12:18:30 Is Alpine actually doing away with OpenRC? 2023-04-07 12:19:14 ok, i ended up using `&` on rc-service stop, which is a temporary workaround. 2023-04-07 12:20:11 Newbyte: I cannot tell irony from reality and also need answers to this question 2023-04-07 12:20:34 I've heard murmurs about it before but I don't understand if it's actually happening 2023-04-07 12:21:03 and if it is happening, it would be good if we (postmarketOS) can be notified of it so we can do our part before it happens 2023-04-07 12:21:49 there's no secret meetings, you're not missing anything 2023-04-07 12:22:14 there is a "wish to", a million prerequisites to do so, and thousands of hours of work to do it 2023-04-07 12:22:35 so it's not happening any time soon? 2023-04-07 12:23:02 when it would be happening it would be visible and not really overnight in 5 seconds 2023-04-07 12:23:33 personally i have my doubts it would ever really happen, because it's an unupgradable breaking change 2023-04-07 12:23:39 well, nothing is impossible, but 2023-04-07 12:23:53 it's a very funny amount of work to do 2023-04-07 12:26:54 yeah, it is a huge work to do. 2023-04-07 12:27:28 some distro like artix they have like 4 choices. 2023-04-07 12:28:40 we alpine does't have to replace openrc, we build a alternative init+ service manager 2023-04-07 12:29:43 runit init+supervise+busybox+musl :) how about that 2023-04-07 12:30:37 or s6 init + service manager + busybox + musl 2023-04-07 12:31:12 that would make alpine the fastest and smallest distro on earth 2023-04-07 12:39:14 Ugh, foot updated it's default theme to a low-contrast one. 2023-04-07 12:39:41 I guess it never occurrs to some people that someone might be using a terminal because it's easy to read by default, and making the new default hard-to-read leave users in a really tough spot. 2023-04-07 12:48:49 terminals are so personal i just configure every option so the defaults never matter, something something 2023-04-07 12:49:08 i'm surprised someone used only the default the whole time :D 2023-04-07 12:49:14 then again, it was a good one.. 2023-04-07 12:49:34 high contrast default is definitely my preference too 2023-04-07 12:51:43 "or s6 init + service manager..." <- Not s6-rc?  2023-04-07 14:56:10 ACTION uses base16 and a crate called 'flavours' to set a consistent color scheme across stuff automagically 2023-04-07 14:56:24 i'm not a huge fan of theming, or rust for that matter, but it works 2023-04-07 14:57:11 if i ever decide to make a small change, it's easy to make that change everywhere with that tool. 2023-04-07 14:59:22 Hi! Just turned my chromebook into Alpine, I am amazed! Still working on the functionality. 2023-04-07 15:11:02 psykose: the openrc rc-service stop issue is really only seen on alpine. 2023-04-07 15:11:19 i just tested on gentoo/musl rootfs, it is ok. 2023-04-07 15:12:00 yeah i know the exact issue, pay me $100,000,000 to fix it 2023-04-07 15:12:13 :D 2023-04-07 15:12:44 i am just reporting what i have done 2023-04-07 15:12:57 (You might not like the solution though) 2023-04-07 15:13:27 i copied alpine's /bin/busybox to gentoo rootfs, and run 5 busybox nc instance, and stop them.. 2023-04-07 15:18:56 https://0x0.st/HXBQ.txt 2023-04-07 15:20:34 ^^ (ls /bin/busybox ; ldd /bin/busybox; echo ; rc-service --version ; echo ; cat /etc/os-release ; echo ; ./sd.sh 2>&1) > log 2023-04-07 15:30:21 if you spent as much effort fixing the code as copypasting this same reproduction script like 5 times by now, maybe it would be fixed 2023-04-07 15:33:01 i don't understand what you mean. 2023-04-07 15:34:12 i mean we already know and you've mentioned it 10 times 2023-04-07 15:34:21 what do you want to be done about it :D 2023-04-07 15:34:40 generally nobody cares about an extra 4 second on shutdown, so nobody else is going to fix it 2023-04-07 15:35:09 i thought it was due to musl. the test i have done sometimes ago was on gentoo/glibc and alpine. today i did on gentoo/musl 2023-04-07 15:36:08 psykose: i got time today, so i tried again :) 2023-04-07 15:36:45 never mind, i don't know what is going on actually. it's not musl, it's not openrc upstream, i don't know what it is. 2023-04-07 15:50:02 i don't care about shuting down either, when i run shutdown, i'm already gone. but it is not a good experience when i need to manage bunch of services via scripts. 2023-04-07 15:57:27 WhyNotHugo: which theme do you use? 2023-04-07 15:57:38 (in foot) 2023-04-07 16:37:06 Ermine: I was using the default one, "zenburn". 2023-04-07 16:38:03 What window manager do you guys use? Currently trying out xfce. 2023-04-07 16:47:16 sway here 2023-04-07 16:48:43 moved to sway a week ago 2023-04-07 16:49:07 did you use tiled desktop before? 2023-04-07 16:51:52 I used Plasma. But sway turned out to perfectly suit my use-case. 2023-04-07 16:53:18 I was using plasma too like 1 year ago, now I feel that tiled desktops suit all use-case :D 2023-04-07 16:55:44 1) Requires a bit of learning (I haven't understood its model fully so far) 2) Some broken apps don't fit well, e.g. zoom 2023-04-07 16:58:30 I would use lxqt but im not entirely sure how to setup. Xfce seems to fit my needs. 2023-04-07 17:01:16 one day my mouse was empty of battery and I continued working with only keyboard for three or four days, I feel that I got a better flow experience using only the keyboard than mixing with the mouse 2023-04-07 17:02:05 but unfourtanely I was forced to use a mouse for a windows app on a VM, I tried to replace the mouse needed with autohotkey but it was too time consuming :D 2023-04-07 17:02:35 funnily windows is generally pretty ok for keyboard-only in itself 2023-04-07 17:02:49 pretty much literally everything has a hotkey 2023-04-07 17:03:18 they're usually even more polished than half the linux apps.. but then you get the applications that need a mouse and it falls apart for sure 2023-04-07 17:03:30 yep, but this app didn't have the common alt+key defautl hotkets 2023-04-07 17:03:35 it was totally unusable without a mouse :\ 2023-04-07 17:04:52 I wonder if firefox fails to use video decoding acceleration even in windows 2023-04-07 17:06:34 Sounds unconfortable, im sure its a fast experience. 2023-04-07 17:07:07 it kinda works on linux 2023-04-07 17:07:13 it used to crash quite a bit but now it's ""ok"" 2023-04-07 17:07:23 with whatever radeonsi vaapi i have anyway 2023-04-07 17:09:15 I was using it today with intel integrated graphics, and cpu usage was like 60%. Need to test it on linux though 2023-04-07 17:09:43 a3e3a[m]: do you mean tiled desktops/sway? 2023-04-07 17:09:44 generally you should see less cpu usage 2023-04-07 17:09:51 by default it's off ofc 2023-04-07 17:09:59 you can see in about:config if you search vaapi 2023-04-07 17:10:11 there's also a force-copy thing which probably does something next to it 2023-04-07 17:10:45 Does anyone use alpine for gaming? 2023-04-07 17:12:03 sure 2023-04-07 17:12:06 hm.. you can install steam with flatpack 2023-04-07 17:12:43 pretty much 2023-04-07 17:12:54 i guess i play ps2 games without flatpak 2023-04-07 17:13:03 Seems like no 0ad in repos 2023-04-07 17:13:03 > a3e3a: do you mean tiled desktops/sway? 2023-04-07 17:13:03 Yea, tiled desktops must be fast. I've tried a few but I can't seem to fully understand how to use them. 2023-04-07 17:14:15 Retroarch for me 2023-04-07 17:15:06 never got the appeal of retroarch personally 2023-04-07 17:15:06 Got the flatpak version, able to play Dreamcast on a Chromebook. 2023-04-07 17:15:15 it's just a scuffed version of the actual emulator 2023-04-07 17:15:28 Also firefox still can't get their sharing indicator right. >< 2023-04-07 17:19:41 > it's just a scuffed version of the actual emulator 2023-04-07 17:19:41 It is. I enjoy the convinience of the all-in-one form. I also recommend ludo https://ludo.libretro.com/ 2023-04-07 17:23:26 oh my god psykose , you don't let me time for calling your attention :D 2023-04-07 17:31:35 :) 2023-04-07 17:40:03 How would I disable Wconversion ? 2023-04-07 17:40:59 -wconversion 2023-04-07 17:41:04 remove this https://github.com/neovim/neovim/blob/d675bd01b1e78b93e559320b262bdae40b3b54b2/src/nvim/CMakeLists.txt#L159 2023-04-07 17:41:09 ah 2023-04-07 17:41:10 i see 2023-04-07 17:41:16 -DCI_BUILD=OFF 2023-04-07 17:41:18 just pass that 2023-04-07 17:41:25 if it works to pass it via cmake 2023-04-07 17:41:38 That disables other checks... but this is mostly a temporary hack I guess. 2023-04-07 17:42:09 no, it removes werror 2023-04-07 17:42:12 which should just never be set 2023-04-07 17:42:42 these are some weird ass options 2023-04-07 17:42:42 https://github.com/neovim/neovim/blob/d675bd01b1e78b93e559320b262bdae40b3b54b2/src/nvim/CMakeLists.txt#L164 2023-04-07 17:42:43 haha 2023-04-07 17:47:07 ikke: i was wondering if we had a reason to not unset CI before invoking the builds 2023-04-07 17:47:28 there are quite a few things that do CI-magic 2023-04-07 17:48:34 psykose: Don't think so 2023-04-07 17:48:57 i forget if `unset` works or not for sub-invoked-programs 2023-04-07 17:49:33 should 2023-04-07 17:51:32 I can't think of anything downstream of the CI pipeline relies on CI being set 2023-04-07 17:53:18 don't think there is 2023-04-07 18:19:43 Can I find which package provides a file via apk? Assuming the package is not locally installed. 2023-04-07 18:19:56 nope 2023-04-07 18:20:02 :c 2023-04-07 18:20:06 there's just the pkgs.a.o site 2023-04-07 18:20:18 I need a cli that curls that. 2023-04-07 18:20:20 and then there's a script that just calls the site i guess 2023-04-07 18:20:23 yeah it's uhh 2023-04-07 18:20:25 lol 2023-04-07 18:20:33 apk-file 2023-04-07 18:20:43 Any thoughs on including a separate db that apk can download with this info? 2023-04-07 18:20:55 Basically, the same db used by pkgs.alpinelinux.org/contents ? 2023-04-07 18:21:04 it is much bigger than you think it is 2023-04-07 18:21:15 :p 2023-04-07 18:21:28 that also is very silly to just expose because you cannot just diff such a db by default 2023-04-07 18:21:36 e.g. to download a daily update you would download the entire db again 2023-04-07 18:22:00 Meh, usually I'd be fine with a week old one. 2023-04-07 18:22:02 void linux actually had a discussion about this today about how to improve their xlocate which is the same kind of thing 2023-04-07 18:22:11 to make it better and more incremental 2023-04-07 18:22:25 week old is still the same issue, you're downloading a gigantic ass db 2023-04-07 18:22:31 Like "oh, what provides fastboot": "android-tools". It doesn't really need to be fresh. 2023-04-07 18:22:36 Huh 2023-04-07 18:22:41 And did they find a solution? 2023-04-07 18:24:12 they already had one, it's improving it i guess 2023-04-07 18:24:20 ours isn't very good, could use some improvement i guess 2023-04-07 18:24:35 99% of the issue is that we have releases so there are a million releases and lots of duplication like that.. 2023-04-07 18:25:00 Yeah, that's keep one for edge, no big deal. 2023-04-07 18:25:16 apk-file fastboot works, but it's not as exact as typing it on the website 2023-04-07 18:25:29 so it's slow as hell because it does *fastboot* or something under the hood 2023-04-07 18:25:37 which is not a very fast operation 2023-04-07 18:26:00 It does fastboot? 2023-04-07 18:26:07 someone suggested fts5 2023-04-07 18:26:10 for sqlite 2023-04-07 18:26:41 i don't think fts5 is very useful for this 2023-04-07 18:26:54 unless they secretly did magic 2023-04-07 18:26:58 well 2023-04-07 18:27:03 it's probably better than what is there 2023-04-07 18:28:14 Wouldn't decompressed text result in better deltas? 2023-04-07 18:28:41 in what sense 2023-04-07 18:28:55 I mean plain text. Like csv. You distribute it compressed, but can generate diffs based on the decompressed csv. 2023-04-07 18:29:17 definitely not 2023-04-07 18:32:29 but currently we don't have any of this at all, so :D 2023-04-07 18:32:56 seeing if fts5 is an improvement for pkgs.a.o in terms of query speed would be nice in that sense i guess 2023-04-07 19:02:46 How do i run a .sh file in alpine? 2023-04-07 19:12:50 the same way as literally everywhere else 2023-04-07 19:13:17 ./file.sh if it's executable, which invokes it with what is in the #!shebang in the file 2023-04-07 19:13:21 or `sh file.sh` 2023-04-07 20:28:41 "Does anyone use alpine for..." <- Lutris and Steam in Flatpak, which covers pretty much every game I own legally. Been great. Hoping to get an Arc A380 and play a bit nicer  2023-04-07 20:31:11 a3e3a yes, steam fightcade and other emulators using flatpak mostly 2023-04-07 20:32:13 did you ever get to the bottom of that fightcade url issue 2023-04-07 20:32:16 i sadly never figured it out 2023-04-07 20:33:46 "I would use lxqt but im not..." <- Been using XFCE on Alpine for over a year now. Excellent experience. For Flatpak functionality, make sure you have xdg-desktop-portal-gtk installed  2023-04-07 20:34:06 yes psykose it's fixed since a long time now 2023-04-07 20:34:08 Otherwise browser integration and other things won't work right and will silently fail 2023-04-07 20:34:37 :D that's good 2023-04-07 20:34:40 i remember it was a support.dbus setting in previous wireplumber versions 2023-04-07 20:44:39 a3e3a: I only use alpine on my laptop cause my desktop has nvidia gpu but I played games like rainworld, cuphead and stuff like that effortlessly, just used the lutris flatpak and it just worked 2023-04-07 20:55:21 I gotta play rainworld 😑 backlog...  2023-04-08 07:30:15 Does the ntfs-3g driver work on windows compressed NTFS drives? 2023-04-08 07:31:10 I found this https://github.com/tuxera/ntfs-3g/blob/71ecccf279d3660812b18803a75f368d2f76eea6/README#L21 -- is that what is being referred to there? 2023-04-08 07:32:06 the plugin is probably https://github.com/ebiggers/ntfs-3g-system-compression 2023-04-08 07:35:18 Is that what they are talking about there though? The compression used when you check the "Compress this drive to save disk space" checkbox in windows explorer? 2023-04-08 07:35:21 afaik, ntfs-3g natively supports the standard per-file compression (the kind that you enable/disable via userspace), while the plugin is needed for a different type of compression that Windows' own system files sometimes use (e.g. c:/windows/fonts) 2023-04-08 07:36:18 (from what I remember seeing in git log, kernel ntfs3 supports both natively) 2023-04-08 07:36:27 Good then either way I should be able to get it to work 2023-04-08 07:36:38 Ok I consider this safe then lol 2023-04-08 13:05:17 @umk3:matrix.org @saijin_naib:matrix.org @reggie:matrix.im 2023-04-08 13:05:17 I have a chromebook with amd graphics and 4gb of ram. Added flatpak, retroarch, fightcade. So far everything works like a charm! Compared to chromeOS this is blazing fast! 2023-04-08 13:05:43 cool 2023-04-08 13:06:18 Anyone wanna play fightcade sometime please let me know. 2023-04-08 13:09:21 > <@saijin_naib:matrix.org> > <@a3e3a:matrix.org> I would use lxqt but im not entirely sure how to setup. Xfce seems to fit my needs.... (full message at ) 2023-04-08 13:12:01 I was trying out fractal for matrix but ram is limited. Think ill just run matrix through hydrogen web. 2023-04-08 13:12:25 https://hydrogen.element.io/ 2023-04-08 13:13:26 Today is my bday anyone reading this please have a blessed day! 2023-04-08 13:22:38 a3e3a[m]: If you really need Matrix, neochat is pretty lightweight 2023-04-08 13:25:20 Thanks ill have a look! 2023-04-08 14:14:46 neochat repo indicates that it still doesn't support e2ee 2023-04-08 14:17:35 nheko is another QML-based client which does support e2ee, and its latest version is in alpine (i'm running it on another distro and it worksforme) 2023-04-08 18:36:46 my Firefox problem of crashing on watching any video is cured with the latest mesa updates. 2023-04-08 18:52:48 Aber alles andere was Du hast geht signal und so 2023-04-08 18:53:13 Sorry, wrong Channel! 2023-04-08 23:08:01 Hi all! 2023-04-08 23:18:22 hi meanly 2023-04-08 23:18:36 hi ptrc 2023-04-08 23:27:10 okaaay i'm stupid shrimp and i cant find what are default credentials for installation 2023-04-08 23:28:33 root 2023-04-08 23:31:42 okay thanx 2023-04-09 00:05:19 hi a polite feature request to the wiki's admins, could you please create a script or sth that would create case insensitive redirects? (e.g.) redirect Apparmor to AppArmor, XFCE to Xfce etc. 2023-04-09 00:15:54 also I've been thinking about creating a script called setup-security which would set up stuff like sysctl security tweaks, change the default umask, possibly get some malware/tracking hosts blacklist and create a cronjob to auto-update it, setup stuff like polkit, apparmor, hardened malloc, awall; generally stuff listed out here https://madaidans-insecurities.github.io/guides/linux-hardening.html, with options to choose various 2023-04-09 00:15:54 security levels with some elementary explanations relying on typical threat model determination rules. How likely would it be that it'd eventually become included in the OS by default after testing in the edge/testing repo? 2023-04-09 00:17:33 you'd need to make it first before anybody could even give you an opinion about it 2023-04-09 00:19:38 the way these things usually work out best in open source (in my experience) is... you solve your own problem first, and then make what you did available 2023-04-09 00:23:11 well it depends 2023-04-09 00:27:46 if there's something that alpine doesn't do, that you want it to do, solve your needs first. if you want to see it be merged, it would have to align with how alpine does things, probably 2023-04-09 00:28:20 for fresh projects it might be more true, but changes to existing projects are something that requires more communication. Since I believe it might be beneficial if such script eventually became the part of the OS by default (with some patches maybe even applied to the standard installation, after all security is in Alpine's motto), I asked here for an opinion. Although such scripts aren't usually too time consuming to write and even if 2023-04-09 00:28:20 it won't receive too much interest it can still be helpful to anyone interested to use it right away. 2023-04-09 00:28:54 in any case my suggestion is -- build whatever you want to solve your porblems, then maybe file a bug to get feedback from the devs on whether it makes sense to have what you did merged 2023-04-09 00:29:04 s/porblems/problems/ 2023-04-09 00:31:20 stuff like apparmor, what malloc to use, sysctl changes from upstream defaults, etc... those are all opinions, in a way 2023-04-09 00:31:57 well same can be said about not using systemd, coreutils and glibc 2023-04-09 00:32:15 alpine does have opinions, like for instance, when firejail got pulled. but not all opinions will win out, or probably, not all of yours will 2023-04-09 00:33:11 yeah I see, just it 2023-04-09 00:33:38 it's actually a fact that linux's security is an illusion to a notable degree 2023-04-09 00:34:05 sure, but you can't get everyone to agree on what knobs to turn 2023-04-09 00:34:36 eg, you could turn a lot of knobs and then make it impossible to debug anything. this is not acceptable to a lot of people 2023-04-09 00:35:03 it's a set of choices that every sysadmin makes 2023-04-09 00:35:13 i'm not sure it's up to alpine to make all of those decisions for people 2023-04-09 00:37:11 yeah I get it, for me also too opinionated defaults would be against the core philosophy of free software, that is giving people a choice. If someone is blinded by hubris and believes that they won't break their computer or idk, wants to break it in some fit of rage, then I got no right to prevent them from doing that. 2023-04-09 00:38:28 i'm just one person though, i'm not speaking for anybody. who knows, They might love everything you want to do. 2023-04-09 00:38:40 i'm not a betting man 2023-04-09 00:41:37 Does the extended edition *have* to run within RAM? 2023-04-09 00:54:43 still there are a lot of people out there that are on some level not fully aware of the actual implications of the warranty waiver that comes with free software and I think that the end-users deserve some convincing encouragement to read the docs in case of doubt (possibly with some tip that they can also use tools like tealdeer to skim through particularly lengthy docs – it was only after like 8 years of using Linux when I discovered 2023-04-09 00:54:43 it, and given this particular case, some advice on how to minimize the odds of having to deal with the consequences of this lack of warranty, like some link to a wiki page with a guide on making things more stable and secure or precisely such scripts as the one I've outlined). Still there are more passive, systemic ways besides fixing CVEs to improve the security, such as rewriting stuff in memory safe languages. As long as it doesn't 2023-04-09 00:54:43 hinder usability outside really rare edge cases – why not? For instance there are surely still some people that use telnet for whatever reason, but do we prompt for setting up telnet and not just ssh during installation? 2023-04-09 00:57:19 "Does the extended edition *have*..." <- no 2023-04-09 00:58:52 novorol[m]: How would you configure that? 2023-04-09 00:59:16 https://docs.alpinelinux.org/user-handbook/0.1a/index.html 2023-04-09 01:02:46 disclaimer: it's been a while since I last installed alpine, but iso images can be inspected with most archive utilities, if you see setup-alpine missing from it you can just put it in the iso before putting it on an installation medium 2023-04-09 01:11:48 novorol[m]: That page has zero answers about how to do this 2023-04-09 01:11:54 Does not even contain the word RAM 2023-04-09 01:13:06 But hold on, I bet it points to something that does 2023-04-09 01:14:11 did you take a look at the index on the left? 2023-04-09 01:14:22 That's what I'm talking about 2023-04-09 01:14:59 you can customize your iso 2023-04-09 01:15:16 drop setup-alpine into it and you should be good 2023-04-09 01:15:21 https://wiki.alpinelinux.org/wiki/How_to_make_a_custom_ISO_image_with_mkimage 2023-04-09 01:16:52 or if you're already running it 2023-04-09 01:20:35 once you have set up networking, install git, clone https://git.alpinelinux.org/alpine-conf with the --branch flag for the version you're targetting (3.15.1 is for 3.17 iibc, for edge just use master) and run setup-alpine 2023-04-09 01:21:04 Nope I'm not running it already 2023-04-09 01:22:07 But as long as those user manuals cover specifically how to configure whether alpine is restricted to RAM, then I should be good 2023-04-09 01:24:08 I mean... why would it do so in the first place. https://docs.alpinelinux.org/user-handbook/0.1a/Installing/medium.html#_image_type Here it says Often used for non-disk installations 2023-04-09 01:24:28 I don't know how that would be decided 2023-04-09 01:24:34 often, thereby not always 2023-04-09 01:24:55 I just don't want to end up with a non-disk installation on a machine with a disk 2023-04-09 01:26:06 novorol[m]: It is only one or the other no matter what 2023-04-09 01:26:10 Something has to choose 2023-04-09 01:26:16 And I would like it to be me 2023-04-09 01:27:58 no you won't 2023-04-09 01:28:16 alpine-conf is included in this iso 2023-04-09 01:28:56 so you just boot it, log in, run setup-alpine and voila 2023-04-09 04:43:32 hello, I'm trying to set up a light weight container to run a zabbix proxy for 6.x and I can't seem to find any zabbix pxy packages for alpine. does anyone know of a repo or aarch64 binary for this? at least with posters support by ideally sqlite3 2023-04-09 04:46:28 *postgres 2023-04-09 07:06:34 ugh, another mesa update. This won't bork my Firefox? 2023-04-09 09:35:08 Good day! The mesa-dri-gallium package provides openGL drivers, right? 2023-04-09 09:35:42 yes 2023-04-09 09:35:53 I am asking because Alacritty doesn't work without this package on my machine, but the dri-gallium package is not in Alacritty's dependencies 2023-04-09 09:36:11 sounds correct to me 2023-04-09 09:36:55 Ah, there are other implementations of openGL in Alpine? 2023-04-09 09:37:21 Well, in the apk repository 2023-04-09 09:40:14 not really 2023-04-09 09:40:24 but it's one of hundreds of packages that require a gl driver 2023-04-09 09:40:33 and without a gl driver you cannot even start a wayland compositor 2023-04-09 09:40:40 and setup-xorg-base already installs mesa-dri-gallium 2023-04-09 09:43:25 @psykose: I see, thank You! It was not a part of my dependency tree, I just had it left over from some flatpak application 2023-04-09 12:05:02 is there some established way to detect docking with busybox acpid, or should i just do something like [ `grep ^connected /sys/class/drm/card0-*/status | wc -l` -gt 1 ] || echo mem > /sys/power/state? 2023-04-09 12:06:20 does docking print an event? 2023-04-09 12:07:01 i forgot what handler script it calls, but if you log every message in it you can see if it looks unique 2023-04-09 12:07:40 and if the 'bigger community/acpid' that doesn't use the same script (iirc) logs it too, in case that would work as an alternative 2023-04-09 12:08:02 hm, no 2023-04-09 12:08:05 docking is a level signal, not edge 2023-04-09 12:09:19 i guess, but it would be nice if there was a trigger anyway 2023-04-09 12:09:39 i guess the answer to that then is "no" so i'm not sure how you're meant to handle a docked event there 2023-04-09 12:09:51 on lid close, you don't want to go to sleep if there is stuff connected to you (most typically, a display). what i wrote above works for me, but i wonder if there's something already there. of course e.g. logind will handle this for you automagically 2023-04-09 12:10:05 ah so you want that kind of handling 2023-04-09 12:10:22 ok, then it's checking the level inside the existing handling as you wrote, and that would work 2023-04-09 12:10:37 i was thinking more "do something on dock", which requires something to emit an edge 2023-04-09 12:10:48 yep, sorry, i wasn't clear 2023-04-09 12:11:03 yeah there's nothing magic in the handling script i don't think 2023-04-09 12:11:09 you'll just have to check exactly what you want 2023-04-09 12:11:30 unless, the emitted 'lid close' event, differs when docked 2023-04-09 12:11:35 which would be weird :p 2023-04-09 12:12:54 i suspect 'docked' is an imaginary event anyway on my laptop which doesn't have a real docking connector (pretty sure acpi isn't concerned with external display) 2023-04-09 12:14:32 :) 2023-04-09 12:14:33 definitely 2023-04-09 13:04:27 I'm trying to switch from alpine on sway to alpine on tty using just tmux. So far it's working really well, but I can't seem to figure out mouse support. I've put 'set -g mouse on' in my config, and I can change tmux with other parts of the config, but it doesn't make my mouse start. When I switch to sway the mouse is there. This makes me wonder if there's some service or variable that sway 2023-04-09 13:04:33 normally handles that I need to set manually. Anyone have any ideas? Thanks 2023-04-09 13:04:54 you mean using the mouse in a tty? 2023-04-09 13:05:27 Yes 2023-04-09 13:05:38 a tty does not have a mouse 2023-04-09 13:05:43 you need `gpm` for that 2023-04-09 13:05:59 but it still won't give you "a mouse" 2023-04-09 13:06:04 you just can't do that 2023-04-09 13:06:09 and what you are doing is imo a mistake 2023-04-09 13:06:16 I installed gpm and got a cursor that can highlight, but it didn't actually do anything. Apparently there's some compatibility issue with tmux over their TERM variable? 2023-04-09 13:06:20 you can replicate 99.99% of it by starting something terminal-only 2023-04-09 13:06:31 i.e. 2023-04-09 13:06:35 you can just do 2023-04-09 13:06:39 `cage foot` 2023-04-09 13:06:46 to start a terminal in an empty wayland session 2023-04-09 13:06:47 cage? 2023-04-09 13:06:50 with nothing else at all 2023-04-09 13:06:57 it's a wayland compositor that just holds one app 2023-04-09 13:07:03 and then in that one terminal, you can run tmux 2023-04-09 13:07:11 and there you go- minimal one-screen tmux thing 2023-04-09 13:07:18 doing things in a bare tty is really silly 2023-04-09 13:07:29 it doesn't support anything like decent fonts for unicode, ... 2023-04-09 13:07:40 all the resizing/font stuff is completely fucked 2023-04-09 13:07:55 if you want something that doesn't start wayland/X with one terminal, there's explicitly a tty terminal thing 2023-04-09 13:08:03 that starts an opengl terminal or whatever 2023-04-09 13:08:05 what was the name... 2023-04-09 13:08:40 kmscon 2023-04-09 13:08:44 but idk if you will get a mouse in it 2023-04-09 13:08:57 with cage+terminal you would 2023-04-09 13:09:30 if you still want the "feel" you can use a pixel font in the terminal or whatnot so it looks the same :D 2023-04-09 13:09:34 the font thing is really true, my laptop's default resolution is (for some ungodly reason) 4k and it makes my tty font super small. Even with changing to terminus-32 it's still tiny, and with coreboot I can't figure out how to change my default resolution. So if I do cage foot I would get a terminal in an empty wayland session without the rest of the dependencies? 2023-04-09 13:09:53 changing the "resolution" is wrong 2023-04-09 13:09:57 the screen is 4k so it's 4k 2023-04-09 13:10:05 oh 2023-04-09 13:10:20 it's mostly a hack to make it 'not 4k' and is worse than just doing scaling 2023-04-09 13:10:33 e.g. doing scaling with 2x in sway or similar in a regular setup 2023-04-09 13:10:36 so when it draws the terminus font that's why the resulting console font is so tiny, because the pixels are 4k 2023-04-09 13:10:45 in the tty that doesn't exist, you just get the bare pixel font 2023-04-09 13:10:58 That makes sense 2023-04-09 13:11:05 but if you use kmscon/cage/whatever you can then do scale and a vector font normally 2023-04-09 13:11:24 i use ter-v32b on my 4k laptop 2023-04-09 13:11:43 one issue with consolefont is that the service has a slight race condition i never figured out 2023-04-09 13:12:00 there's a chance what you try and set never gets applied, so make sure you test fonts with setfont manually first to see if they even are what you expect 2023-04-09 13:12:05 I only use my laptop to write novels, so I really don't need more than the most basic of nano functions. The gui of sway lets me do firefox, which lets me do reddit, which is a distraction lol. I liked that in a tty I can use lynx for research but not for fun 2023-04-09 13:12:17 yeah, just cage foot would work 2023-04-09 13:12:25 ter-v32 doesn't render non-ascii correctly for me, but i never bothered to figure out why 2023-04-09 13:12:28 or sway with a config that turns everything off 2023-04-09 13:12:37 hm okay yeah, cage foot! I'll give that a shot! 2023-04-09 13:12:49 you can launch sway with a custom config file that has all your mappings disabled and just starts a terminal 2023-04-09 13:13:27 (so you can't e.g. open the program launcher to open firefox, etc) 2023-04-09 13:13:46 I usually just keybind firefox to super+i and call it a day lol 2023-04-09 13:13:59 tbh now that you mention it 2023-04-09 13:14:04 i get sniped by firefox all the time 2023-04-09 13:14:10 i should probably keep that closed :p 2023-04-09 13:15:00 lol it does tend to suck time away. Even on legit research! I started by researching technology of the late Renaissance period and ended up reading about territorial disputes between the vatican and italian monarchy. Fun, but not relevant haha 2023-04-09 13:15:11 :) 2023-04-09 13:15:20 well, do try, let me know how it goes 2023-04-09 13:15:37 there's three options to go through, the kmscon being one i'm not sure about but that also should work well by giving you a real font and a terminal 2023-04-09 13:15:48 dunno about the mouse part 2023-04-09 13:17:21 Will do, thanks for your help! 2023-04-09 13:56:51 What does "masked in: cache" mean? 2023-04-09 13:57:22 what's the whole output 2023-04-09 13:57:34 psykose: https://gitlab.com/postmarketOS/pmaports/-/issues/2059#note_1346113861 2023-04-09 13:58:18 "fabled Basically it's package not present in any index" 2023-04-09 13:59:27 Ah, another day of "everything is broken". 2023-04-09 13:59:47 something went wrong : ncurses-terminfo , there is no files in /usr/share/terminfo after installation 2023-04-09 14:00:08 psykose: 2023-04-09 14:00:40 after upgrade, my tmux crash on mouse selection. 2023-04-09 14:01:20 fd -t f tmux-256color / : /etc/terminfo/t/tmux-256color 2023-04-09 14:03:27 oh, it is symlinked? 2023-04-09 14:07:53 WhyNotHugo: basically just a spurious message that the apk add -t .virt failed, the .virt itself doesn't exist anywhere so it's masked in cache instead of some repo or whatever 2023-04-09 14:08:03 it doesn't mean anything there and isn't the cause of the issue or relevant 2023-04-09 14:08:57 uuidNuniq: there are files in /usr/share/terminfo from ncurses-terminfo 2023-04-09 14:14:09 downgrade to ncurses-terminfo-base-6.4_p20230401-r1 fixed the issue 2023-04-09 14:15:05 thanks for the downgrade commit 2023-04-09 14:19:44 there was no issue with terminfo 2023-04-09 14:23:03 ok. it was just a guess. my first reaction was to check tmux-256color using fd 2023-04-09 14:26:15 well, the issue could be the terminfo changes, just not whatever you thought happened 2023-04-09 14:28:43 psykose: maybe take a look at it's changelog? https://invisible-mirror.net/ncurses/NEWS.html#index-t20230408 2023-04-09 14:28:58 i know what is in the changelog 2023-04-09 14:30:33 so what was really happenned 2023-04-09 14:30:51 don't know and don't care 2023-04-09 14:30:58 :) 2023-04-09 14:31:01 but usr/share/terminfo was never empty 2023-04-09 14:31:27 that's awesome, they'll fix it probably when people report tmux issue. 2023-04-09 14:31:32 yeah 2023-04-09 14:31:48 it's probably something invalid in the tmux terminfo that trips the new checks 2023-04-09 17:55:27 Hello chat, ive been thinking about siwtching to alpine. i intend to use it for desktop 2023-04-09 17:57:25 + what are some things i should know? 2023-04-09 17:57:58 it is musl based, so glibc based things won't work 2023-04-09 17:58:42 what are the advantages of musl? 2023-04-09 17:59:21 http://musl.libc.org/about.html 2023-04-09 18:00:43 Also will the installation take a lot of data? im using mobile data as internet 2023-04-09 18:01:56 unlikely 2023-04-09 18:02:31 im gonna do a minimal install with i3wm 2023-04-09 18:03:28 Also, does alpine supported advanced security features such as SELinux? 2023-04-09 18:03:48 there is nothing integrated 2023-04-09 18:04:15 i mean is it usable? 2023-04-09 18:04:24 Or perhaps AppArmor? 2023-04-09 18:05:42 did they move the files to /etc/terminfo? I see them there. 2023-04-09 18:06:19 there is both /etc/terminfo and /usr/share/terminfo 2023-04-09 18:06:30 damn 153MB iso? that things is small as hell 2023-04-09 18:08:28 ali: https://alpinelinux.org/ 2023-04-09 18:10:07 glibc-bound things can work with a shim or via other means, eg flatpak 2023-04-09 18:10:28 I am downloading an ISO, i will burn it to flash drive and try alpine out 2023-04-09 18:11:21 ali: things like apparmor / selinux only work well if the entire system is setup to use it 2023-04-09 18:11:40 which is not the case for AlpineLinux (it's a huge maintenance burden) 2023-04-09 18:12:30 I mean i mostly use a browser 2023-04-09 18:14:33 web or file 2023-04-09 18:15:35 i dont get what you mean, i use librewolf 2023-04-09 18:16:22 doesn't look like there is a package for it 2023-04-09 18:17:42 ali, you can use qutebrowser + jmatrix 2023-04-09 18:18:01 https://gitlab.com/jgkamat/jmatrix 2023-04-09 18:18:25 i mean firefox is ok too 2023-04-09 18:20:51 bruh i dont know vim stuff for qutebrowser :skull: 2023-04-09 18:34:19 "i dont get what you mean, i..." <- Hey me too! :) You can download it with flatpak. 2023-04-09 18:35:39 I recommend you use flatseal as well for permissions. 2023-04-09 18:37:38 Can someone help me? I'm trying to install a host file https://gh.odyssey346.dev/Ultimate-Hosts-Blacklist/Ultimate.Hosts.Blacklist im not sure why i can't get the .sh file to work 2023-04-09 18:39:24 what is the error 2023-04-09 18:40:28 i just don't understand why i cant execute a .sh file in alpine. Could you walk me through it? 2023-04-09 18:42:23 a3e3a[m]: Can you show what you are trying to execute and what is not working 2023-04-09 18:42:26 a3e3a[m]: can you execute a .sh file in another Linux distro? then do the same in Alpine. Unless the .sh file is written for Bash, in which case install the Bash package for Alpine before trying... 2023-04-09 18:42:43 a3e3a[m]: it helps to give details 2023-04-09 18:43:33 it's most likely missing bash 2023-04-09 18:44:04 ok, downloaded .sh file in terminal chmod +x .sh, then ./.sh file file not found 2023-04-09 18:44:32 a3e3a[m]: head -n1 .sh 2023-04-09 18:46:01 head -1 ./file.sh? 2023-04-09 18:46:19 a3e3a[m]: you say you execute ./.sh 2023-04-09 18:46:39 the first line in the file is something like #!/bin/bash 2023-04-09 18:46:44 that line is called the shebang 2023-04-09 18:46:54 when you ./file on an executable file, the file is executed with that line 2023-04-09 18:47:04 it says not found if that thing doesn't actually exist 2023-04-09 18:47:52 in this case, the line is exactly that (bash), so if you don't install bash (`apk add bash`), then it's not found 2023-04-09 18:48:12 the script also calls `sudo` so you have more issues after 2023-04-09 18:48:29 for bonus fun, the script first checks if you are root 2023-04-09 18:48:30 psykose: correct opened file in mousepad, fist line is, #!/bin/bash 2023-04-09 18:48:35 then exits if you are not root 2023-04-09 18:48:41 and tells you to run the script with sudo 2023-04-09 18:48:54 and then the command after just does `sudo wget -O` 2023-04-09 18:48:55 sweet 2023-04-09 18:49:18 the entire thing is just `wget https://hosts.ubuntu101.co.za/hosts -O /etc/hosts` 2023-04-09 18:49:20 you can just do that 2023-04-09 18:50:22 ok so i need to dl bash then? 2023-04-09 18:50:36 feel free to read what i wrote multiple times until you get it :) 2023-04-09 18:51:05 yeah its going over my head 2023-04-09 18:51:52 tl;dr: don't run random scripts from internet 2023-04-09 18:52:57 psykose: so it wont apply to alpine? 2023-04-09 18:53:09 i answered every question you asked 2023-04-09 18:54:12 Sure, but its not clear to me. 2023-04-09 18:54:58 "you can just do that" <- I can do just what? 2023-04-09 18:55:02 the solution to that is to read it a few more times and have a think 2023-04-09 18:56:04 "the entire thing is just `wget..." <- open /etc/hosts? 2023-04-09 18:58:08 it's a shell command 2023-04-09 19:01:43 can i just unhelpfully add to psykose's dissatisfaction with the script that the unix user is one whose uid is 0, not whose account name is 'root' 2023-04-09 19:01:55 er, *unix root user 2023-04-09 19:02:43 sometimes i'm tempted to make a small personal distro where root is not called root for fun 2023-04-09 19:02:56 then again i realise that gobolinux probably already fixed most of that in the wild 2023-04-09 19:03:21 Administrator 2023-04-09 19:03:45 i'd probably go for reallylongusername 2023-04-09 19:03:46 Or.. NT Authority/System 2023-04-09 19:06:29 or 'nobody' :P 2023-04-09 19:06:46 that script first tests if the user is called "root", then calls sudo anyway... wtf 2023-04-09 19:08:40 I'd just `if [ "$(id -u)" -ne 0 ]; then printf "You must be root"; exit 1; fi` and drop the calls to sudo 2023-04-09 19:09:57 also they seem to be replacing /etc/hosts with their version entirely. for all i know maybe systemd does something to rectify this automagically, but generally it's a pretty bad idea not to have your hostname resolve to something sensible 2023-04-09 19:10:06 also completely unnecessary `exit 0` in the end... x) 2023-04-09 19:10:16 Is there another host file that can be recommended? 2023-04-09 19:10:22 in all, i probably wouldn't let people who make these kinds of mistakes manage configuration in my /etc 2023-04-09 19:11:58 a3e3a[m]: There's https://github.com/StevenBlack/hosts 2023-04-09 19:11:59 strajder: I wouldn't expect random scripts from the internet to have high quality 2023-04-09 19:14:02 a3e3a[m]: definitely create a backup of /etc/hosts somewhere and append instead of replace the contents of /etc/hosts 2023-04-09 19:14:03 I wouldn't try to apply "some hosts blocklist" not knowing how to run bash scripts or not verifying random scripts from internet 2023-04-09 19:14:32 pj: that too 2023-04-09 19:17:03 i probably would also have a real dns server (busybox dnsd will do if all that's needed is a static blocklist) just for a little more flexibility/reliability 2023-04-09 19:23:47 that argument in open PR on github is great 2023-04-09 19:24:35 psykose: do you think we can merge !42665? it's been on hold for some time now 2023-04-09 19:25:11 »/11 2023-04-09 19:28:31 pj: arguably the "else" branch in that PR is also unnecessary, the decision to escalate privileges should be more transparent for the user; and if the user doesn't know how to do that it probably shouldn't be administering the system by changing /etc/hosts 2023-04-09 19:32:14 I'm just amazed at how much arguing it generated when it could just drop `sudo` in script 2023-04-09 19:34:39 I wouldn't run any script which has 'sudo' in it. 2023-04-09 19:36:01 the "easy" protection against that is to use doas :P 2023-04-09 19:36:06 s/ which has 'sudo in it// ftfy 2023-04-09 19:37:00 strajder, that would just be hilarious. Maybe I'd smirk confused. 2023-04-09 19:40:09 In 3.17, vim doesn't know about foot terminal somewhy 2023-04-09 19:40:58 Vim knows as much about foot as $TERM is in terminfo 2023-04-09 19:41:21 Ermine: i have term=xterm in foot.ini, because i'm not going to copy foot's terminfo to every box to which i ssh 2023-04-09 19:41:22 ncurses-terminfo has it in 3.17 2023-04-09 19:41:32 ovf: it makes more sense to change it for ssh 2023-04-09 19:41:46 Host * SetEnv TERM=xterm-256color 2023-04-09 19:42:36 psykose: thank you 2023-04-09 19:42:47 but what features am i really missing on locally, to justify occasionally having the kind of problems Ermine is dealing with (weird software which ignores or mishandles terminfo) 2023-04-09 19:52:47 not having 256color 2023-04-09 19:56:23 some stuff subtly breaks without terminfo=foot 2023-04-09 19:56:36 by some i mean few, and if you don't notice then it doesn't matter 2023-04-09 19:56:46 also the ncurses foot terminfo is also subtly broken 2023-04-09 19:57:07 the only correct one is 'foot-extra' from foot-extra-terminfo itself 2023-04-09 19:57:17 (it's normally just foot but it's renamed to not conflict) 2023-04-09 19:57:38 maybe name it foot-correct? :P 2023-04-09 19:57:42 haha 2023-04-09 19:57:50 if i was more aggressive i would remove the ncurses one 2023-04-09 19:57:56 like jirutka did for kitty 2023-04-09 19:57:58 maybe i should 2023-04-09 19:58:15 idk why dickey is so weird about having correct terminfo files 2023-04-09 19:59:02 So, foot-extra-terminfo is also needed? 2023-04-09 19:59:50 only if you actually need it and care about the semantics 2023-04-09 19:59:55 it's broken inside like.. tmux 2023-04-09 19:59:57 that's about it 2023-04-09 20:00:09 it might have even been fixed, haven't checked in a year :p 2023-04-09 20:00:58 i somehow still feel my solution is slightly superior 2023-04-09 20:01:07 it works 2023-04-09 20:01:13 i also used the same thing for years 2023-04-09 20:01:27 until i found something broken with it, and then it is what it is 2023-04-09 20:03:30 ACTION thinks about moving to alacritty with (hopefully) fine terminfo stuff 2023-04-09 20:04:13 iirc same thing 2023-04-09 20:04:23 but legit i'm just doomsaying, if you don't notice issues there are none 2023-04-09 20:04:58 i wonder if that things that broke with term=xterm could have been fixed by foot becoming ever slightly more xterm-like 2023-04-09 20:07:17 ACTION thinks about making not broken terminal 2023-04-09 20:07:30 This is perfectionism though 2023-04-09 20:10:25 i wish things were less xterm like 2023-04-09 20:10:32 that would be a good future 2023-04-09 20:11:06 Ermine: foot isn't very broken 2023-04-09 20:16:25 the best solution is to not use alacritty 2023-04-09 21:37:34 About the script earlier... have you noticed how the "then" is indented after the "if", like it's a command... 2023-04-09 22:27:10 half of the script is just copyright notices 2023-04-09 22:27:32 as if it was something copyrightable and also worth copyrighting 2023-04-09 22:48:05 can you copyright a copyright message? ;-) 2023-04-09 22:48:27 yes 2023-04-09 22:51:21 and copyright the copyright message using that copyright message? ;-) 2023-04-10 02:22:56 to install an .apk on local disk, what is the command for that? 2023-04-10 02:23:34 apk add 2023-04-10 02:25:20 ohh, apk add ./ 2023-04-10 02:25:52 or the tab completion would search for pkgs in repository 2023-04-10 02:27:02 and then, the world will change, nice. 2023-04-10 02:32:41 psykose: have you considered using version from here? https://invisible-mirror.net/archives/ncurses/ncurses-${pkgver}.tar.gz 2023-04-10 02:32:58 why would i use that 2023-04-10 02:33:07 these are `stable` release 2023-04-10 02:33:32 >thomas dickey 2023-04-10 02:33:34 >stable release 2023-04-10 02:33:35 hahaha 2023-04-10 02:33:58 lmao 2023-04-10 02:34:07 the /current/ are dev release 2023-04-10 02:35:11 there's nothing stable about the stable releases 2023-04-10 02:36:06 ok. 2023-04-10 02:36:26 currently it would be mostly the same, just with more broken terminfo 2023-04-10 02:36:41 in 1.5 years when there hasn't been another "stable release" by then though.. 2023-04-10 02:37:08 mawk is especially funny in this case 2023-04-10 02:37:14 because it does the same thing 2023-04-10 02:37:18 except the stable release was in 2009 2023-04-10 02:38:40 i remember few months ago, it probably was due to ncurses update, vim auto insert random stuff when openning any file 2023-04-10 02:39:54 you're not gonna believe what caused that 2023-04-10 02:39:56 are you ready 2023-04-10 02:40:05 it's the completely broken terminfo for foot shipped with ncurses 2023-04-10 02:40:26 xD 2023-04-10 02:42:01 what is foot? i don't use that 2023-04-10 02:42:40 i never used foot, most of the time i am on tmux and iterm2 2023-04-10 04:35:00 doas is funny: permit nopass 1000 cmd /sbin/apk 2023-04-10 04:35:13 then `doas apk` still needs password. 2023-04-10 04:35:24 but `doas /sbin/apk` does not need password 2023-04-10 04:48:03 the former depends on path and could just be any command 2023-04-10 04:48:18 sudo has some whitelisted configurable paths for such a thing 2023-04-10 04:55:14 i'll just do : .... cmd apk 2023-04-10 04:55:24 for the sake of simplity 2023-04-10 10:05:46 https://www.unixsheikh.com/articles/the-heavy-responsibility-of-the-package-maintainer.html 2023-04-10 13:28:49 uniqNuuid, thank you, this blog is inspiring. Someone I probably could be friends with. 2023-04-10 15:15:54 Hi, anyone know of 'askpass' programs that work with Alpine? 2023-04-10 15:22:00 ringring: you mean like openssh-askpass which is packaged for Alpine? 2023-04-10 15:25:07 minimal: I installed openssh-askpass on a Pinephone (sxmo sway) but couldn't find the program with 'wich' of 'find'. 2023-04-10 15:25:33 which 2023-04-10 15:27:45 it seems to install gtk-ssh-askpass program 2023-04-10 15:30:42 minimal: gtk-ssh-askpass was installed - not sure if it is compatiable with my environment? 2023-04-10 15:33:33 is that a question or a statement? 2023-04-10 15:38:44 minimal: Good point - sorry :) I'm trying to invoke a sway keyboard bindsym exec with -A optiont to run a srcipt with sudo permission. I'm wondering if an askpass program can automatically supply the expected password. 2023-04-10 15:41:04 then I assume the above is not suitable as it's intended for use with passphrase-protected SSH keys 2023-04-10 15:44:26 minimal: OK, thanks - time to consider alternatives. 2023-04-10 16:35:01 ah, right, still joined here. 2023-04-10 16:35:53 Is there a way for me to request Alpine on x86 to have a PAE-enabled kernel? 2023-04-10 16:43:41 You can open an issue here: https://gitlab.alpinelinux.org/alpine/aports/-/issues 2023-04-10 16:48:57 Thanks 2023-04-10 17:15:31 q 2023-04-10 18:22:20 hiya, i can't figure out how to install thunderbird at the moment because i get an apk error: https://paste.sr.ht/~autumnull/d4768ac7881d34d3c3dc110b8e1ee3cdbbe4ad60 2023-04-10 18:23:03 it seems like its dependencies are messed up somehow 2023-04-10 18:23:44 no issues installing thunderbird on edge or v3.17 2023-04-10 18:23:51 did you mix stable and edge? 2023-04-10 18:23:58 .. maybe 2023-04-10 18:23:59 i'll check 2023-04-10 18:24:22 yes 2023-04-10 18:24:56 ah yeah that fixed it, thanks ikke <3 2023-04-10 18:25:01 np 2023-04-10 19:58:39 Hi, I'm using alpine edge in a amd64 system. Tried to setup xen as per the wiki, but each time I choose one of the Xen options at the grub boot manager, the system just restarts. 2023-04-10 19:59:55 https://wiki.alpinelinux.org/wiki/Xen_Dom0 2023-04-10 20:02:36 In another try, dowloaded the latest alpine-xen-3.17.3-x86_64.iso image. After the grub boot console, several messages are shown on the screen and the last one says 'Xen is relinquishing VGA console' or something similar. After that, the screen is black, nothing more is displayed. My screen is a 4K one. 2023-04-10 20:03:19 I'd prefer to keep my installed system and just be able to start using Xen instead of installing the system again. 2023-04-10 20:03:32 Any tips on how to solve this? 2023-04-10 20:11:24 Richard Simmons gets dateraped by Regina native Edna Skilton. Read about the drama involving hockey, roofies, and flabby elbows! https://pastes.io/3zb8ipyqbv 2023-04-10 20:12:53 what the heck did i just read 2023-04-10 20:13:25 can we get a +b in the chat 2023-04-10 20:13:37 It is a story of steamy passsion, hockey, daterape, and richard simmons unknowingly being a father 2023-04-10 20:13:51 after Edna Skilton slips something in his coffee, takes him to her hotel 2023-04-10 20:13:57 and milks that semen right out of him 2023-04-10 20:14:06 but accidentally gets some of it in her dripping wet Canadian beaver 2023-04-10 20:14:24 why 2023-04-10 20:14:41 It all starts at the North Charleston Coliseum 2023-04-10 20:15:00 this is top-tier Alpine discourse 2023-04-10 20:15:25 ikr 2023-04-10 20:15:51 hey guys how to patch KDE 2 under FreeBSD 2023-04-10 20:16:04 first, install alpine 2023-04-10 20:16:22 gerhard's Alpine Chat can only dream of reaching this level of Alpineness 2023-04-10 20:16:44 LMAO 2023-04-10 20:17:13 Then can I masturbate imagining the puñeta skiltoniana? 2023-04-10 20:17:15 Alpine/kFreeBSD wen 2023-04-10 20:21:52 This is the first time I'm seeing horny spam 2023-04-10 20:25:30 pure art 2023-04-10 20:32:03 Once they arrived at her hotel room, Edna quickly undressed Richard and placed him in her bed. She then stripped down to her bra and panties, revealing her massive, supple elbows, which were tattooed with the areolas and nipples. She placed each elbow next to Richard's bratwurst and began to give him a Puñeta Skiltoniana, moving her elbows up and down to stimulate him. As she did so, she grinded her ***, wet Canadian beaver into his face and 2023-04-10 20:32:03 Edna continued to stroke Richard's hair with her muffler and screamed and mooed as she neared climax. Richard was barely conscious, but he managed to moan in pleasure as Edna pleasured him with her unique technique. Edna continued to ride Richard's face, feeling more and more aroused by the minute. 2023-04-10 20:32:03 forehead, feeling electrified as the kinky hairs of Simmons's afro tickled her labia and beaver tongue. 2023-04-10 20:32:05 As she reached the height of her climax, Edna let out a loud scream and collapsed onto the bed next to Richard. She lay there for a few minutes, catching her breath and basking in the afterglow of her orgasm. When she looked over at Richard, she realized that he was still unconscious, but she wanted to make him jizz all over the place. 2023-04-10 20:32:21 ncopa: ikke: 2023-04-10 20:32:27 She then lubed up her big, bulging and supple fat elbows with coconut oil and began to give another Puñeta Skiltoniana and move them up and down on both sides of Richard's erect penis. As she worked her magic, Richard moaned in pleasure, his breaths becoming quicker and more ragged. Mrs. Skilton was relentless, her strokes getting faster and more intense as Richard got closer and closer to orgasm. Finally, even in unconciousness he couldn't 2023-04-10 20:32:27 hold back any longer and shot loads of cream, staining his dolphin shorts and his sequined Black Lives Matter tank top. This was when he began stirring. 2023-04-10 20:32:29 ..any other ops in this channel? 2023-04-10 20:36:29 is there a new set of packages required for ceph to function? when running an `apk upgrade` I get things likd so:libboost_iostreams.so.1.81.0 (no such package): 2023-04-10 20:37:24 wasutton3: are you by chance combining stable and edge? 2023-04-10 20:38:10 ikke, i've got both the 3.17 community and edge community enabled 2023-04-10 20:38:18 Well, that the issue then 2023-04-10 20:38:27 gotcha 2023-04-10 20:38:33 I'm working from this guide: https://wiki.alpinelinux.org/wiki/K8s 2023-04-10 20:38:38 and it has both enabled 2023-04-10 20:38:38 you are getting ceph from edge, while boost is from v3.17 2023-04-10 20:39:54 ok, so if i disable 3.17, I get those errors 2023-04-10 20:39:58 is boost not in edge? 2023-04-10 20:40:06 apk upgrade -Ua 2023-04-10 20:40:26 no dice 2023-04-10 20:41:41 http://paste.debian.net/plain/1276962 2023-04-10 20:45:55 looks like quite a few things changed in the 16->17 migration 2023-04-10 21:01:49 yep on a fresh install something is wrong 2023-04-10 21:02:20 need to use edge for main as well 2023-04-10 21:07:40 https://pkgs.alpinelinux.org/package/edge/main/x86/boost-dev 2023-04-10 21:07:44 edge is 1.81 2023-04-11 01:45:09 Hello, would anyone know how to enable something like dropbear in initramfs so I can decrypt my root partition remotely? 2023-04-11 01:46:03 I've found tutorials and resources for remotely decrypting my root partition for systemd-based distros like Debian and Arch but not one for Alpine or any OpenRC-based distros really 2023-04-11 02:08:29 novorol[m]: You said alpine-sonf is included in this iso, which one are you talking about? The one I downloaded yields no results when searching it for alpine-setup 2023-04-11 02:08:41 s/alpine-sonf/alpine-conf 2023-04-11 02:18:24 Also (anyone), what determines which of syslinux or grub is used to boot the iso/installed distro? I see both on the iso so I don't know which is the default 2023-04-11 02:26:07 How would I set permanent environment variables in alpine? 2023-04-11 02:27:32 I failed to find something synonymous to `/etc/environ` 2023-04-11 04:31:14 asking: you can put a .sh file in /etc/profile.d/ with all of your exports in the shell form like: export VARIABLE=VALUE 2023-04-11 04:32:19 it will be sourced by any login shell, which will cover most cases, but won't be visible to stuff started by openrc 2023-04-11 04:33:23 i think /etc/environment comes from linux-pam ? 2023-04-11 09:20:39 an3223 thanks for the info. I li 2023-04-11 09:23:42 *I saw `profile.d` but didn’t like that is wasn’t readable by all certain programs (like openrc), so I was search for a better solution 2023-04-11 09:24:10 I find some dependencies are very hard to find. Like, nemo will render webp thumbnails only if `webp-pixbuf-loader` is installed. But I know this due to following some thread on the topic months ago, it's super hard for anyone to figure this out when installing `nemo`. 2023-04-11 09:24:32 I wish there were some way to make this information easier to discover. 2023-04-11 09:33:33 like suggests/recommends in Debian 2023-04-11 09:42:21 that one is more the fault of gdk-pixbuf not supporting anything in itself 2023-04-11 09:49:55 Habbie: As long as they're never automatically installed by default (e.g.: merely informative). 2023-04-11 09:50:13 sure 2023-04-11 09:50:23 i think debian defaults to recommends but not suggested 2023-04-11 10:02:38 I have encountered an issue though that when I am trying to boot into my os I am getting a mounting uuid=... on /sysroot failed: no such file or directory. The install steps I took are as follows. Plug 2 usbs, one which held the live image the other which I used to select as my sys partition udring the setup-alpine steps. After it finished installing, I rebooted and tried to boot off the usb I installed the alpine linux on however, I got that error 2023-04-11 10:05:37 I did swap the ports the usbs are plugged, I just tried plugging it into the original spots and it seems to boot 2023-04-11 10:07:25 it seems pretty weird :\, so it works on one port and fails on the other? 2023-04-11 10:24:57 donoban: exactly, I wonder if the UUID changes when changing ports and therefore it can not find the correct path 2023-04-11 10:26:50 could you paste your /etc/default/grub ? 2023-04-11 10:39:29 "I find some dependencies are..." <- still better than wine on arch linux as far as I remember 2023-04-11 10:40:30 https://github.com/lutris/docs/blob/master/WineDependencies.md#archendeavourosmanjaroother-arch-derivatives 2023-04-11 10:41:11 there is a guide now which makes things easier but back when I started using linux I had to figure this shit out myself or via random forum posts 2023-04-11 12:10:38 i've decided to do some web dev for the first time in about a decade and i can't figure out what's the proper way to run nginx as a transient user process, not a daemon. the alpine packaged version seems intent on writing under /var/lib/nginx for pidfile, logs etc. am i missing something, or should i individually change all the defaults in my nginx.conf? 2023-04-11 12:17:03 ovf: You'll need a custom conf for that. 2023-04-11 12:17:28 I tend to use containers for this kind of scenarios (e.g.: an ephemeral nginx isntance running as my user). 2023-04-11 12:17:41 of course i have a custom nginx.conf as i don't want to use anything system-wide 2023-04-11 12:17:55 a container sounds like a bit of an overkill for running a single process 2023-04-11 12:18:59 running nginx with a custom conf shouldn't have it write anything anywhere, since the conf configure the logs and pid stuff, etc 2023-04-11 12:19:43 i would probably just not use nginx though? python3 -m http.server 9999 by default will just server your cwd at 0.0.0.0:9999 2023-04-11 12:19:58 there's easier things to use than nginx with a whole config for small fast stuff 2023-04-11 12:21:37 i don't know what nginx does by default mind you, it's configured with --prefix=/var/whatever so that affects that, but i imagine it's not many config lines to just 'write nothing, log stdout' 2023-04-11 12:22:00 eventually i guess my conf will do that, i'm just finding myself playing a whack-a-mole with all the paths that default to /var/lib/nginx (e.g. at the moment i'm past client_body_temp_path and at proxy_temp_path -- each i have to deduce from EACCESS path and look up) 2023-04-11 12:22:14 aha, so i guess it's not that easy :D 2023-04-11 12:22:23 yeah i would just pick something else if it's that insistent 2023-04-11 12:22:35 nginx has -p which is supposed to change the prefix, but stuff still seems to default to /var/lib/nginx 2023-04-11 12:23:36 i want to do php stuff (because i heard it's the new hot tech that all the kids use), so http.server won't do 2023-04-11 12:24:33 ah, so fcgi 2023-04-11 12:24:37 sure, more complex 2023-04-11 12:27:28 i'd have expected everything to be `-p`-relative 2023-04-11 12:29:30 ovf: the entire usecase of docker containers is running single processes 2023-04-11 12:30:30 ikke: not exactly single processes but single service 2023-04-11 12:57:09 Yeah 2023-04-11 13:03:59 i have either been living without bash-completion since forever or something recently changed that i now need an explicit `. /usr/share/bash-completion/bash_completion` in my ~/.bashrc (though i don't know how it could have worked before) 2023-04-11 13:05:28 nothing ever sourced it by default 2023-04-11 13:05:54 the recent change is to not re-source profile.d from the system bashrc but that would not affect this unless you put your completion sourcing thing in profile.d 2023-04-11 13:07:11 oh, nvm 2023-04-11 13:07:19 bash-completion does indeed put something in profile.d 2023-04-11 13:07:24 so yes, it was that change 2023-04-11 13:07:29 but that is wrong, so shrug 2023-04-11 13:09:42 oh, i forgot about the system-wide bashrc. re-sourcing profile.d is mostly wrong (e.g. breaks PATH=foo:$PATH), but the bash-completion bit still feels like a regression 2023-04-11 13:10:10 when i try to install alpine on my sd card, i get /dev/mmcblk0 read only filesystem 2023-04-11 13:10:16 how to fix 2023-04-11 13:11:36 heyman: slide the write protection switch on your sd card/adapter to the open position 2023-04-11 13:11:58 isnt that just to lock the thing in place? 2023-04-11 13:11:58 wtf? 2023-04-11 13:12:19 wait, no way lol 2023-04-11 13:12:20 ovf: i'll fix it by just changing around the source to make bashrc source it when present i guess 2023-04-11 13:12:45 which is the same behaviour as before, always sourced when installed 2023-04-11 13:12:55 nobody complained about completions always being loaded so i guess it's fine 2023-04-11 13:13:18 i always thought installing bash-completion was the 'opt-in' 2023-04-11 13:13:26 yeah 2023-04-11 13:13:36 is there a standard dir for system bash.d? for people to drop things into 2023-04-11 13:14:10 that is not /etc/bash/ itself 2023-04-11 13:16:51 don't think so. btw i see archlinux /etc/bash.bashrc sourcing bash_completion 2023-04-11 13:17:39 i already did that now, but i wonder what to do about that old comment that says 'dont edit this, use profile.d' :p 2023-04-11 13:18:26 Somethign that bothers/worries me about pipewire, is that fact that it's becoming so ubiquotous, but there's (as far as I can tell) no documentation on the wire protocol it uses. 2023-04-11 13:19:21 well, even debian, king of management complexity, doesn't seem to have invented bash.d yet 2023-04-11 13:22:44 upgrade and should work again 2023-04-11 13:22:54 apologies for that one 2023-04-11 13:23:01 when it builds that is 2023-04-11 13:30:57 ovf it worked 2023-04-11 13:31:00 thanks a bunch 2023-04-11 13:31:34 psykose: thanks! 2023-04-11 13:31:43 :) 2023-04-11 13:34:41 but unfortunately, my sd card wont detect the sd card as a boot disk 2023-04-11 13:34:47 my laptop* 2023-04-11 13:35:02 just installed alpine in less than 3 mins lol 2023-04-11 13:51:41 our lawyers and payment department will be getting in touch shortly 2023-04-11 14:02:00 Hey, nobody answered my question about the extended iso -- I'll paste it again quick if nobody saw it 2023-04-11 14:37:08 gilbert: this was discussed several days ago - I developed a solution for mkinitfs some time ago but it has never been merged, alternatively it could be done using dracut instead of mkinitfs but I have not (to-date) tried that 2023-04-11 15:04:32 nice! i can reliably crash firefox 112 on wayland when displaying an element's title (little pop-up). also, what's up with libwayland-client calling itself through... ffi? http://0x0.st/HXUH.txt 2023-04-11 15:07:09 that's just how wayland works 2023-04-11 15:07:39 it crashing sounds not good because there aren't any funny things to fix this time.. but if you tell me the exact reproduction i can try something 2023-04-11 15:16:24 ovf, element's title? in the debug window? 2023-04-11 15:19:41 psykose: somercet1: https://0x0.st/HXUN.html (this might well crash your browser...) 2023-04-11 15:20:06 psykose: this one sounds much more like an upstream bug 2023-04-11 15:20:19 hmm, nope 2023-04-11 15:20:40 ah, I'm on 111 2023-04-11 15:21:10 crashes for me if i point to the first thing, wait for the popup to show, then point to the second thing 2023-04-11 15:23:35 doesn't for me 2023-04-11 15:23:39 where's firefox-dbg when it's needed 2023-04-11 15:24:02 they have a weird setup for that that is not very standard and requires setting up some server stuff 2023-04-11 15:24:05 are you running it in wayland (i.e. not xwayland)? 2023-04-11 15:24:15 and having literal -dbg would be like gb's or something lol 2023-04-11 15:24:17 yes, wayland 2023-04-11 15:25:45 ff this version with wayland has some bugs waiting fix 2023-04-11 15:27:36 https://img.ayaya.dev/KZuo4bTxQwMX 2023-04-11 15:28:22 it still crashes for me even if i reset the gtk theme to Adwaita (since the popup is styled, it could be relevant) 2023-04-11 15:28:42 mine is adw-gtk3:dark 2023-04-11 15:31:49 can't think of what else is different on my system 2023-04-11 15:32:16 it could just be also gpu driver related by a kilometer or screen dimensions or whateverthefuck 2023-04-11 15:32:31 i use WLR_RENDERER=vulkan for instance 2023-04-11 15:45:04 well, it's not that. :-) 2023-04-11 15:56:29 at least i'm pretty sure (but don't feel like building 111) this came in with 112 -- i got firefox crashing twice today by accident, and that never happened before 2023-04-11 15:56:53 not unlikely 2023-04-11 16:07:36 novorol[m]: You said alpine-conf is included in this iso, which one are you talking about? The one I downloaded yields no results when searching it for alpine-setup 2023-04-11 16:07:45 Also (anyone), what determines which of syslinux or grub is used to boot the iso/installed distro? I see both on the iso so I don't know which is the default 2023-04-11 16:07:57 Liver_K: setup-alpine 2023-04-11 16:08:24 Either one, they are not there 2023-04-11 16:09:36 I searched for "setup" and "setup-alpine" among other things, can confirm it is not there 2023-04-11 16:12:54 there is no /sbin/setup-* ? 2023-04-11 16:13:04 which iso did you use 2023-04-11 16:13:41 psykose: https://dl-cdn.alpinelinux.org/alpine/v3.17/releases/x86_64/alpine-extended-3.17.3-x86_64.iso 2023-04-11 16:13:58 are you sure you didn't run out of memory using it or something that made it fail to unpack fully 2023-04-11 16:14:05 because /sbin/setup-* is always present with that 2023-04-11 16:33:54 Liver_K: setup-alpine installs/uses Syslinux for BIOS booting and uses Grub for UEFI booting 2023-04-11 16:34:55 that assumes you're installing to a disk, rather than running from RAM 2023-04-11 16:44:14 why use extended 2023-04-11 17:02:26 minimal: would you happen to have a link to the solution? 2023-04-11 17:08:12 minimal: Thanks 2023-04-11 17:08:18 psykose: I didn't unpack it at all 2023-04-11 17:08:26 I am searching through the mounted iso 2023-04-11 17:08:30 lol 2023-04-11 17:08:39 well yes, that's not how it works :) 2023-04-11 17:08:42 it installs packages on boot 2023-04-11 17:08:47 Ok cool :) 2023-04-11 17:08:59 Thanks I think I am all set up to try this out then 2023-04-11 17:10:00 Actually minimal, I thought you installed to disk after booting the installer image 2023-04-11 17:10:08 ok, good(?) news is that i can reproduce it with official releases. mozregression + my educated guess point to https://hg.mozilla.org/mozilla-central/rev/d57b1b2adb5d35e1b686df035e17bede08ad77f0 . not sure where to go from here. :-) 2023-04-11 17:25:37 https://gitlab.alpinelinux.org/alpine/aports/-/merge_requests/45895#note_300749 2023-04-11 17:25:48 are here packagers who do package rust stuff? 2023-04-11 17:25:54 and do have the same/simmilar issue 2023-04-11 17:26:00 * and do have/had the same/similar issue 2023-04-11 17:26:02 * and do have/had the same/similar issue? 2023-04-11 17:26:11 please don't edit on matrix 2023-04-11 17:26:17 aha, and this was broken too 2023-04-11 17:26:18 hmm 2023-04-11 17:28:25 psykose: too ? 2023-04-11 17:28:55 ah you commented ... 2023-04-11 17:29:40 well I have no idea about the "lfs64 change" ... If you have a clue, I'll patient wait for what you will come up with 😅 psykose 2023-04-11 17:29:53 yeah kinda have to fix a million things :p 2023-04-11 17:43:52 psykose: couldn't figure out how to symbolize stuff locally, but at least The Cloud did it: https://crash-stats.mozilla.org/report/index/c9c43332-59a7-4e6f-8bdb-88d4a0230411 . someone on mozilla's #firefox told me (but i didn't believe then) that wayland 1.22 likes to complain about stuff, and firefox deliberately crashes on wayland warnings (which seems to match the backtrace) 2023-04-11 17:45:04 > wl_log_set_handler_client(WlCrashHandler); 2023-04-11 17:47:16 minimal: I mean, it still has to boot when running from RAM right? 2023-04-11 17:47:25 Is it just the other way around in that case? 2023-04-11 17:51:34 why is alpine taking so long to install on my usb stick 2023-04-11 17:52:22 have you tried asking the usb stick 2023-04-11 17:52:33 yes 2023-04-11 17:52:56 ive no idea at this point 2023-04-11 17:54:44 its taking minutes when i tried with an sd card it took less than 2 seconds 2023-04-11 17:54:56 i tried on different environments, same thing 2023-04-11 17:57:01 now the progress bar is stuck cause its moving way too slow 2023-04-11 17:57:06 SD cards generally have good write performance, unlike the cheap USB sticks you get these days 2023-04-11 17:57:28 its a 3.0 kingston 2023-04-11 17:57:39 that doesn't mean anything :P 2023-04-11 17:57:45 ovf: ah yes that makes sense 2023-04-11 17:57:54 i thought i got rid of those with the mesa patches but guess not entirely 2023-04-11 17:57:56 SD cards performance is based on its class 2023-04-11 17:58:05 i can boot liveusbs fine but installing alpine is so slow 2023-04-11 17:58:31 USB flash tends to have decent read speeds 2023-04-11 17:58:41 T_T 2023-04-11 17:58:43 it's the write speeds that are pure garbage 2023-04-11 17:58:45 Oh hey I am also using a Kingston datatraveler 3.0 2023-04-11 17:58:54 At least that is what I will be writing the iso to 2023-04-11 17:59:07 holy fuck 2023-04-11 17:59:26 psykose: it all seems to be fixed somewhere around 113, but i suspect highly incidentally (as part of some popup refactor), so if those warning something you can influence it'd still be worth a shot 2023-04-11 18:00:19 there's nothing i know backportable 2023-04-11 18:00:37 the specific wl_log crash seems to be fixed by the mesa patches and i can't reproduce it 2023-04-11 18:00:53 dunno which popup refactor you're talking about but if it's not big i could just apply it anyway 2023-04-11 18:02:04 mozregression points me to https://phabricator.services.mozilla.com/D170368 , but i'm not sure i understand enough of what i'm doing with it to be sure that's the fix 2023-04-11 18:03:08 i've been getting Xwayland cores lately. haven't had time to look beyond that 2023-04-11 18:04:04 that doesn't sound related to this 2023-04-11 18:04:28 probably isn't 2023-04-11 18:04:40 i suspect their build archive doesn't have enough resolution, but i can't seem to get it to show the full range of revisions 2023-04-11 18:04:41 ovf: do you know how to build firefox yourself to test that since i can't reproduce 2023-04-11 18:04:55 i can give you a patch if it makes it more convenient 2023-04-11 18:05:56 i know how to type 'abuild rootbld'. i need to go get some groceries, so i can leave the heater^W build on 2023-04-11 18:06:11 sure, gimme a few minutes to get you a patch 2023-04-11 18:06:27 though you would have to uhh work around that in a weird way for a few hours anyway, to rebuild musl first 2023-04-11 18:06:41 because the builders have 800 things in the queue and without rebuilding it will fail rust :p 2023-04-11 18:07:07 and rootbld won't pull local main/musl unless you also add /home/you/packages/main to community/.rootbld-repositories 2023-04-11 18:08:40 gilbert: https://github.com/dermotbradley/alpine-remote-LUKS-unlock contains my patches for it, https://gitlab.alpinelinux.org/alpine/mkinitfs/-/merge_requests/86 was my MR that never was merged 2023-04-11 18:08:47 oh, looks like a chose a bad time to do this stuff 2023-04-11 18:09:23 Liver_K: I was referring to setup-alpine (which runs setup-disk) being used to create a bootable disk/SSD 2023-04-11 18:11:14 minimal: Ever thought of adding tpm integration to that or something alike? For like, the SSH host keys? 2023-04-11 18:14:11 WhyNotHugo: well I haven't done anything with it since I originally submitted the MR. I do see any point expanding one functionality if it's not going to get merged. At some point I'll look at doing an equivalent MR for dracut 2023-04-11 18:14:47 man: No entry for init.d in the manual. 2023-04-11 18:14:53 Where can I override stuff for /etc/init.d/dhcpcd ? 2023-04-11 18:15:04 Or is it retained on package upates? 2023-04-11 18:15:21 override what exactly? 2023-04-11 18:16:56 speaking of which a new dhcpcd release just came out today. Anyone working on an Alpine package update yet? 2023-04-11 18:17:54 ovf: https://img.ayaya.dev/aqzqoBC93XwV 2023-04-11 18:18:04 ofc you need that workaround first, to build musl and add .rootbld repo 2023-04-11 18:18:08 rest should be ok afaik 2023-04-11 18:18:46 psykose: new dhcpcd seems to have musl-specific privsep changes/fix - yuo disable privsep in Feb as it didn't work ;-) 2023-04-11 18:19:03 i tested master at the time and it also didn't work 2023-04-11 18:19:19 nor do i bet anyone tested every architecture with that or anything 2023-04-11 18:19:32 minimal: I want to add -b 2023-04-11 18:19:41 psykose: thanks. on it 2023-04-11 18:19:56 don't forget to git pull 2023-04-11 18:22:02 WhyNotHugo: ah, ok, currently dhcpcd doesn't have a /etc/conf.d/ file, so the typical way would be to add on and have the init.d source it and then you could add "-b" to the options variable in the conf.d file 2023-04-11 18:27:30 minimal: I can just create one, or something is missing elsewhere? 2023-04-11 18:27:43 any libre browsers available in alpine packages? 2023-04-11 18:28:18 No closed-source browsers here 2023-04-11 18:28:26 firefox 2023-04-11 18:28:29 chromium 2023-04-11 18:28:41 chromium has google code in it by default though? 2023-04-11 18:28:49 qutebrowser 2023-04-11 18:28:55 + jmatrix 2023-04-11 18:29:15 Given that chromium is developed by google... 2023-04-11 18:31:51 there is iridium and ungoogled-chromium but i cant find them packaged for my version and platform 2023-04-11 18:32:00 I'm surprised that `apk search librewolf` yields no results. 2023-04-11 18:33:38 WhyNotHugo: /etc/init.d/dhcpcd appears to use command_args variable if defined, so create /etc/conf.d/dhcpcd with contents of "command_args="-b" 2023-04-11 18:34:19 heyman: I mean, chromium is still 100% libre 2023-04-11 18:34:48 are u sure? last time i checked, it was communicating with google servers 2023-04-11 18:35:08 that does not change anything 2023-04-11 18:35:25 doesnt seem very libre to me 2023-04-11 18:35:34 heyman: License is still open source. You can even read the code to check how much they're sending out. 2023-04-11 18:35:37 your definition of libre is wrong 2023-04-11 18:36:55 that wasn't very libre of you 2023-04-11 18:37:06 open source != libre 2023-04-11 18:37:38 i didnt read the code, i tested it with wireshark 2023-04-11 18:38:25 heyman: What do you understand by "libre"? 2023-04-11 18:39:07 you are conflating privacy concerns with licensing 2023-04-11 18:44:43 I prefer ungoogled-chromium as well, but newer versions of chromium will be crippled by Manifest V3 so the choice is either to stay on the last version before that or switch to another browser; additionally in Alpine, there is no package for it, but I settled on qutebrowser with jmatrix (uMatrix for qb) 2023-04-11 18:45:04 manifest v3 is delayed 2023-04-11 18:45:16 n-th time, unsurprising 2023-04-11 18:46:07 meh, only a matter of time 2023-04-11 18:46:08 okay aside from that, i think setup-alpine installed the bootloader on my usb stick instead of my hard drive 2023-04-11 18:47:05 heyman: which device did you tell it to install to? ;-) 2023-04-11 18:47:20 qutebrowser is packaged for alpine, I don't know if this jmatrix thing required you to build it yourself though 2023-04-11 18:47:21 https://pkgs.alpinelinux.org/packages?name=qutebrowser&branch=edge&repo=&arch=&maintainer= 2023-04-11 18:47:23 strajder: yes, but it will not be implemented as is since other browser vendors are very not ok with it 2023-04-11 18:47:36 well, when i first booted into the newly installed system, df said root was on my hard disk 2023-04-11 18:48:12 has been since 3.12 2023-04-11 18:48:20 or actually, maybe my computer is just retarded 2023-04-11 18:48:30 qman: it is installed from git 2023-04-11 18:48:34 https://gitlab.com/jgkamat/jmatrix 2023-04-11 19:02:47 What group do I need to be in to use `fastboot` without being root? 2023-04-11 19:09:33 probably plugdev 2023-04-11 19:09:43 and you need the android-udev-rules to assign to that group for the devices 2023-04-11 19:09:52 otherwise the android device just stays root only 2023-04-11 19:20:25 What's the criteria for logs going to /var/log/messages vs /var/log/$SERVICE 2023-04-11 19:22:25 psykose: lo and behold, that seemingly fixed it 2023-04-11 19:22:59 psykose: i'll give it a bit more usage, but so far it went from 100% to 0% reproducible 2023-04-11 19:23:00 WhyNotHugo: I guess stuff which uses syslog goes to messages, and stuff which does logging itself goes to $SERVICE 2023-04-11 19:23:14 ovf: lovely, i'll apply it then i think 2023-04-11 19:24:26 i don't particularly like bug fixes i don't understand, but i really want a non-crashing firefox 2023-04-11 19:38:11 minimal: Oh, ok, is the installer any different? 2023-04-11 19:38:36 In how it boots vs what it installs 2023-04-11 19:39:05 Without any configuration (by default) 2023-04-11 19:39:30 guys how do i fix 2023-04-11 19:39:32 Liver_K: you mean the ISO? Its's a ISO9660 filesystem, so it uses isolinux (part of syslinux) to boot AFAIK 2023-04-11 19:39:32 no such package 2023-04-11 19:40:12 nvm lol repos were commented out 2023-04-11 19:41:44 minimal: Oh I did not know that 2023-04-11 19:41:58 Can't grub boot isos? 2023-04-11 19:42:04 kinda 2023-04-11 19:42:19 abby: Elaborate 2023-04-11 19:42:55 (Please, lol) 2023-04-11 19:43:54 Liver_K: why do you care? the ISO uses what it uses 2023-04-11 19:43:55 you can mount an iso to a loopback and boot the initramfs directly 2023-04-11 19:44:27 minimal: Well it sounded from your answer like ISO9660s always use syslinux 2023-04-11 19:44:58 Liver_K: no, my answer was that the Alpine release ISOs use isolinux 2023-04-11 19:45:08 not that all ISOs must use isolinux 2023-04-11 19:45:46 "It's a ISO9660 so[...]" I guess you just meant that is the reason why it uses isolinux (the part of syslinux) specifically :) 2023-04-11 19:47:55 Syslinux (the software project) provides several tools for different types of booting: extlinux, isolinux, pxelinux 2023-04-11 19:50:33 How does grub do it? 2023-04-11 19:50:43 how do i setup networkmanager on xfce 2023-04-11 19:51:07 Liver_K: actually the ISOs do appear to be using Grub EFI as well (to handle UEFI booting) as the ISOs are hybrid bios/uefi booting images 2023-04-11 19:51:31 Yeah I do see grub efi files in the iso 2023-04-11 19:51:38 https://gitlab.alpinelinux.org/alpine/aports/-/blob/master/scripts/mkimg.base.sh#L241 2023-04-11 19:51:52 there's were it sets up booting for the CD 2023-04-11 19:52:16 anywhat what actual *problem* are you having? 2023-04-11 19:52:34 No problem, just want to know what the hell I'm doing before I do it 2023-04-11 19:52:51 what you're doing? or what the ISO is doing? 2023-04-11 19:52:55 That's about the last piece of info I need to start my install attempt I think 2023-04-11 19:53:03 minimal: iNSTALLING ALPINE TO DISK 2023-04-11 19:53:06 Sorry caps lock 2023-04-11 19:53:16 why do you need to know that before you can boot the ISO? you just boot it 2023-04-11 19:53:58 help plz 2023-04-11 19:54:17 how the ISO boots is unrelated to how the installer (setup-alpine) will set up an installed Alpine for booting 2023-04-11 19:54:24 If I have problems, they will be a shitload easier to troubleshoot if I know what I am using 2023-04-11 19:54:48 which package is the networkmanager applet for xfce 2023-04-11 19:54:56 Liver_K: what the ISO uses to boot is NOT related to what the installer sets up 2023-04-11 19:55:11 minimal: I mean breaking before boot 2023-04-11 19:55:28 Whatever, I want to know because this is when it is relevant to me so why not do it now 2023-04-11 19:56:01 But you already answered so it doesn't really matter 2023-04-11 19:56:12 an installation won't be using a ISO9660 filesystem 2023-04-11 19:56:21 I know 2023-04-11 19:56:49 so how the booting of a ISO9660 occurs isn't relevant as to how the booting of a ext4/xfs/whatever works 2023-04-11 19:57:23 I said, I was talking about if the booting of the ISO does not work, this will help 2023-04-11 19:58:12 After I have installed alpine I am pretty sure I will be fine, or I can ask about that then 2023-04-11 19:58:57 heyman: should be network-manager-applet (programe name: nm-applet) 2023-04-11 19:59:06 What is the best way to get OpenRC to mount a NFS-Share on boot? The nfs-utils services? And if so should I prefer nfsclient or nfsmount? 2023-04-11 19:59:31 why not actually try booting the ISO first rather than deeply investigating how it works for the unlikely situation that the ISO doesn't boot? 2023-04-11 19:59:48 Liver_K: just boot the iso, it is easy to do and if it doesn't work it is easy to just reboot back 2023-04-11 19:59:58 ^ " Whatever, I want to know because this is when it is relevant to me so why not do it now" 2023-04-11 20:00:16 herkulessi[m]: use the netmount init.d service 2023-04-11 20:02:11 while at it, it must be said that alpine installation is very straightforward and smooth 2023-04-11 20:05:18 minimal: That seems to depend on nfsclient... But from what I can tell it does the same thing as nfsmount for my usecase so it doesn't matter. Thanks anyway though 2023-04-11 20:06:43 kwadroke[m] is that package for xfce? 2023-04-11 20:07:06 gtk, not xfce specific 2023-04-11 20:07:38 i installed it, hope something appears after reboot 2023-04-11 20:08:10 I used it in the past, but it's been a while 2023-04-11 20:08:49 don't remember if you have to tell XFCE to start it for you or not 2023-04-11 20:09:22 well it appears, but now im seeing device not ready 2023-04-11 20:10:32 make sure its using the right device 2023-04-11 20:10:41 how? 2023-04-11 20:11:01 btw it's started up by /etc/xdg/autostart/nm-applet.desktop 2023-04-11 20:11:18 yeah, i didnt have to do anything else and it started 2023-04-11 20:11:30 click on the icon in the tray and see what interface it has listed 2023-04-11 20:11:43 literally nothing 2023-04-11 20:11:45 T_T 2023-04-11 20:11:55 right-click then? 2023-04-11 20:12:23 nothing, ill try editing networkmanager.conf 2023-04-11 20:12:28 you've configured the system to use NetworkManager, correct? 2023-04-11 20:13:16 the limited time I tried to get NM to work on Alpine was a failure. Haven't tried since 2023-04-11 20:14:00 cat /var/log/messages says failed to d_bus activate wpa_supplicant service 2023-04-11 20:16:03 did you follow the guide at https://wiki.alpinelinux.org/wiki/NetworkManager ? 2023-04-11 20:24:09 seems like i just didnt install wpa_supplicant 2023-04-11 20:24:15 apk add wpa_supplicant 2023-04-11 20:24:16 lol 2023-04-11 20:25:35 now i gotta fix browsers not wanting to start 2023-04-11 20:25:40 cause of display error 2023-04-11 20:25:48 when i do firefox, it says 2023-04-11 20:25:58 authorization required, but no authorization protocol specified 2023-04-11 20:27:53 you logged in as root? 2023-04-11 20:28:27 no 2023-04-11 20:28:52 running firefox via SUDO/DOAS? 2023-04-11 20:28:57 i did setup-desktop xfce as root though? 2023-04-11 20:29:04 no just normal user 2023-04-11 20:29:15 setup is fine as root 2023-04-11 20:29:18 hmm.... 2023-04-11 20:37:53 you're not trying to run it remotely (SSH/RDP) or via Docker, are you? 2023-04-11 20:38:22 looks like setup-xorg-base 2023-04-11 20:38:25 fixed it 2023-04-11 20:38:28 call me hackerman 2023-04-11 20:38:36 only other thing I can find is a reference to a bad hostname 2023-04-11 20:40:06 thought setup-xorg-base was ran via setup-desktop 2023-04-11 20:41:34 i have no fucking clue my man 2023-04-11 20:41:36 yeah, it is... 2023-04-11 20:41:53 not quite sure why running it again would fix it, but glad it did :) 2023-04-11 20:42:46 yay 2023-04-11 20:43:18 you able to get to your network device via nm-applet? 2023-04-11 20:43:32 after installing wpa_supplicant 2023-04-11 20:45:01 yep 2023-04-11 20:45:06 cool 2023-04-11 20:51:00 Is there anything I have to configure to get crond to send job output to sendmail? Pretty sure I have (my replacement for) sendmail configured correctly, but looking at syslog it doesn't look like crond is even calling sendmail 2023-04-11 20:51:44 crond itself or cron jobs? 2023-04-11 20:52:08 either? 2023-04-11 20:52:15 for the latter you need to add MAILTO=… in the crontab 2023-04-11 20:52:26 … being the user 2023-04-11 20:52:57 I was under the impression that crond would add that automatically (defaulting to the user's crontab it's running) 2023-04-11 20:53:38 don't think bb crond does, not sure about cronie but iirc it don't either 2023-04-11 20:53:38 i don't think i've encountered a cron that does it automatically 2023-04-11 20:53:54 interesting 2023-04-11 20:54:00 it would be weird because it would autofail a lot of the time 2023-04-11 20:57:12 I guess some distros set it to something by default? Not in Alpine, but I know I've seen systems where root's dead.letter file was huge because of said failures 2023-04-11 20:58:17 I'll know in a couple of minutes (/15 minute cron) if that fixes it 2023-04-11 21:01:02 That did it, thanks lopid and psykose 2023-04-11 21:01:09 :) 2023-04-11 21:01:32 now I wonder if there's a way to set that as a default instead of having to make sure all the crontabs have it 2023-04-11 21:02:15 lopid, vixie cron (the usual cron on slightly heavier distros) definitely does it 2023-04-11 21:02:56 it's been so long 2023-04-11 21:03:54 throwback 2023-04-11 21:04:12 'nam flashbacks 2023-04-11 21:17:16 I do not know if it is alpine related but firefox in edge behaves strange: when closing several tabs, firefox resizes the tabs after each tab closing as soon as a mouse button is pressed. 2023-04-11 21:17:34 This means one need to move the mouse for every tab closing. 2023-04-11 21:35:27 engblom: if by 'pressed' you mean 'clicked', that sounds normal to me (but i have been using a tab sidebar instead for years so i haven't seen the native tab bar in a while). if you mean 'pressed down', that's really weird 2023-04-11 21:51:58 $ sh dbvis_linux_23_1.sh 2023-04-11 21:51:58 No suitable Java Virtual Machine could be found on your system. 2023-04-11 21:52:20 $ java -version 2023-04-11 21:52:20 openjdk version "1.8.0_362" 2023-04-11 21:52:20 OpenJDK Runtime Environment (IcedTea 3.26.0) (Alpine 8.362.09-r1) 2023-04-11 21:52:20 OpenJDK 64-Bit Server VM (build 25.362-b09, mixed mode) 2023-04-11 21:52:22 ??? 2023-04-11 21:52:54 maybe it expects a newer JVM or it needs one that can load binaries linked against glibc 2023-04-11 22:01:47 works4me with openjre 20 2023-04-11 22:02:48 psykose are u trolling? 2023-04-11 22:03:08 https://img.ayaya.dev/wCQXJbpiGA8t 2023-04-11 22:03:12 raven523 ur right, i checked the install script, it says it expects JVM 17 2023-04-11 22:03:23 but it didnt echo that line out lol 2023-04-11 22:03:44 it asked for an activation email for a trial 2023-04-11 22:03:47 i put `a` 2023-04-11 22:03:49 it accepted it 2023-04-11 22:03:51 nice! 2023-04-11 22:04:09 I usually end up using 'nobody@nowhere.com' when asked for an email 2023-04-11 22:04:29 try a single letter next time 2023-04-11 22:04:31 apparently 2023-04-11 22:04:32 heh 2023-04-11 22:04:41 most things actually look for at least an @ 2023-04-11 22:04:43 there's a free version 2023-04-11 22:04:45 and sometimes also a . 2023-04-11 22:04:47 im using that 2023-04-11 22:05:10 or larry@oracle.com 2023-04-11 22:05:11 maybe next time I'll try root@localhost.localdomain 2023-04-11 22:05:29 heyman: tbh usually these kinds of random internet apps don't work at all :D 2023-04-11 22:05:32 i'm surprised this does 2023-04-11 22:05:38 just pure java i guess no funny stuff 2023-04-11 22:05:42 yeah 2023-04-11 22:05:45 it does, i was using it since last month 2023-04-11 22:06:00 i went through several other options, they were all paid except this one 2023-04-11 22:06:06 that fulfils my needs 2023-04-11 22:06:11 supermicro's IPMIview would work for basic things, but there's a lot of things that wouldn't because they use C plugins 2023-04-11 22:06:15 in the sense they all usually have a bunch of glibc-linked deps they ship in a dir instead of a static bin, and on alpine that doesn't run 2023-04-11 22:06:35 sometimes you can get away with gcompat but eh 2023-04-11 22:07:00 sometimes it's better to just have a glibc chroot/container 2023-04-11 22:11:18 Our helpdesk IM chat on the support side used a Java app in an RHEL VM. I cracked open the VM, chrooted into it because I got sick of having to click twice on the java app, once for VM, second for the app field. 2023-04-11 23:53:23 Two questions 2023-04-11 23:53:40 three answers 2023-04-11 23:53:54 Firstly, which package is the correct one for Python interpreter? 2023-04-11 23:53:57 python3 2023-04-11 23:54:20 Thank you, and is dotnet still available on Alpine? 2023-04-11 23:54:26 yes 2023-04-11 23:54:52 `dotnet7-sdk` didn’t return an apk. I’m currently on edge 2023-04-11 23:55:03 'return'? 2023-04-11 23:55:20 As is a package wasn’t installed 2023-04-11 23:55:27 *As in 2023-04-11 23:55:29 https://img.ayaya.dev/pysGtnfLUQ9a 2023-04-11 23:55:38 which architecture 2023-04-11 23:55:56 x86 2023-04-11 23:56:09 did you pick the wrong iso 2023-04-11 23:56:13 99% you got that wrong 2023-04-11 23:56:18 and you want x86_64 2023-04-11 23:56:24 but yes it doesn't exist on x86 2023-04-11 23:56:32 I'll try it again with the 7 this time 2023-04-11 23:56:37 Ahhhh 2023-04-11 23:56:43 Rip 2023-04-11 23:57:26 I'll check my architecture one more time to be safe, but thanks for the answers 2023-04-11 23:58:33 etc/apk/arch says it 2023-04-11 23:58:43 if it's x86 then it's probably wrong, dunno what cpu you have 2023-04-12 00:35:25 psykose: I live in the worst timeline, I’m x86_32 😨 2023-04-12 00:35:31 what is your cpu 2023-04-12 00:36:37 my cpu is M1, I’m emulating alpine 2023-04-12 00:37:09 well 2023-04-12 00:37:12 then don't emulate it 2023-04-12 00:37:26 you could also be emulating.. x86_64 2023-04-12 00:37:30 you're inventing problems there 2023-04-12 00:37:43 or just use aarch64 2023-04-12 00:40:48 Yeah, I’m going to swap my setup. I was emulating through an application I really like, and it’s limitation is x86_32 2023-04-12 00:41:15 But I’ll just opt for the aarch64 since Dotnet won’t work 2023-04-12 03:41:32 cool 2023-04-12 03:41:40 I love the new regression in 6.3-rc6 2023-04-12 03:42:19 it can't consistently reset the firmware on my Arc card, so it drops me to a mostly hung up efi framebuffer 2023-04-12 03:52:37 is that sarcasm? what else doesa one expect from rc kernels 2023-04-12 03:55:18 I mean, at the same time, somebody's gotta test them or they'll never get fixed 2023-04-12 03:56:40 yeah, that's partly what i mean 2023-04-12 03:57:07 can't complain about finding what is good to find in that case 2023-04-12 05:47:13 sorry 2023-04-12 05:47:22 I didn't realize I said that to the alpine channel 2023-04-12 05:47:36 somehow this client doesn't make it obvious to me any more when I'm looking at one channel or another 2023-04-12 05:47:51 my response goes the same in basically every channel :p 2023-04-12 05:48:56 and yes, sarcasm 2023-04-12 05:49:17 https://gitlab.freedesktop.org/drm/intel/-/issues/8325 this may or may not be the same issue I was experiencing 2023-04-12 06:23:28 I do not know if it is alpine related but firefox in edge behaves strange: when closing several tabs, one after the other, the first one closes normally (and no resizing of the tab space is happening unless I move the mouse out from the tab field). The X for the following tab is automatically under the mouse cursor but when I press, the mouse button the tab moves away and nothing got closed. Then I 2023-04-12 06:23:35 need to move the mouse to the new location and try again and this time it does close the tab. 2023-04-12 06:31:25 that just sounds like a random firefox change in a new update 2023-04-12 06:32:51 although i can't reproduce that 2023-04-12 06:47:52 engblom: like a cat chasing a mouse 2023-04-12 09:39:52 If I enable dhcpcd (via `service start dhcpcd`), then my /etc/resolv.conf gets emptied. 2023-04-12 09:39:59 I can't quite figure out why this happens, any hints? 2023-04-12 09:44:41 it manages it by default and has a config to not 2023-04-12 09:45:33 nohook resolv.conf 2023-04-12 09:45:40 somewhere in the docs is more options 2023-04-12 09:52:33 But my dhcp does return DNS servers. 2023-04-12 09:52:59 Oh, although `networking` is also running a dhcp; I wonder if the mix of both messes things up. 2023-04-12 09:53:20 networking normally 'starts dhcpcd' for you 2023-04-12 09:53:25 so if you have uh both idk how that works 2023-04-12 09:53:41 i'm not sure you can use dhcpcd service at all with networking enabled at the same time unless you make them mutually exclusive etc 2023-04-12 09:53:48 it's a bit of a everything at once scenario 2023-04-12 09:54:10 and by starts for you i think some codepath in ifupdown-ng calls dhcpcd to manage one interface or something 2023-04-12 09:54:28 Ah, but `networking` starts its own dhcpcd instance, not the dhcpcd service itself. 2023-04-12 09:54:41 'use dhcp' 2023-04-12 09:54:49 WhyNotHugo: if you have dhcpcd it will start it 2023-04-12 09:55:16 Right, it executes /sbin/dhcpcd, it doesn´t go through dhcpcd-openrc. 2023-04-12 09:55:20 not starting the dhcpcd service, but the command 2023-04-12 09:55:51 My reason for wanting dhcpcd is that, in reality, i want dhcp automatically on any interface I connect (e.g.: usb0). 2023-04-12 09:56:26 then stop/disable networking 2023-04-12 09:57:04 networking will only start the interface configured 2023-04-12 09:57:48 networking still does the wireless bit for me (I think?) 2023-04-12 09:58:07 Maybe I should just use networking for usb0. The problem is, usb0 MAY be absent at startup 2023-04-12 09:58:36 in your case perhaps full dhcpcd and networking only for wlan0 might work 2023-04-12 09:58:40 absent is ok. 2023-04-12 09:59:00 dhcpcd will detect carrier 2023-04-12 09:59:08 psykose: full dhcpcd breaks resolv.con 2023-04-12 09:59:24 you can configure it to do what you want can't you 2023-04-12 10:00:05 stuff like no_dns 2023-04-12 10:00:11 I think the issue is that full dhcpcd messes up wlan0 2023-04-12 10:00:13 read dhcpcd.conf? 2023-04-12 10:01:33 WhyNotHugo: which interface do you want to receive dns from 2023-04-12 10:02:08 wlan0 2023-04-12 10:02:55 I think I can use networking for usb0 tho, I'll give that a try first. 2023-04-12 10:03:02 Mostly, I want to see how hotplugging works. 2023-04-12 10:04:56 So if `usb0` is absent at startup, `networking` doesn´t handle hotplugging it. 2023-04-12 10:05:16 dhcpcd handles it 2023-04-12 10:05:41 networking will start dhcpcd -i usb0 and dhcpcd is listening for usb0's carrier in background 2023-04-12 10:06:12 usb0: interface not found\ndhcpcd exited 2023-04-12 10:06:17 It doesn't. 2023-04-12 10:06:50 usb0 isn't "unplugged" when it's not plugged, it doesn't exist at all. 2023-04-12 10:07:04 ok, then, i think you may want to duplicate /etc/init.d/dhcpcd for each interface 2023-04-12 10:07:11 and start them 2023-04-12 10:07:39 I think a single dhcpcd configured to monitor all interfaces should work -- but I need to disable dhcpcd for networking. 2023-04-12 10:07:54 for dns? 2023-04-12 10:08:57 dhcpcd-openrc handling all interfaces overlaps with networking-dhcpcd handling wlan0. They both handle wlan0 and I suspect that's what messes up resolv.conf 2023-04-12 10:10:16 Ah, yes, that works. Removed `iface wlan0 inet dhcp` from networking, and enabled dhcpcd. 2023-04-12 10:10:27 *dhcpcd-openrc 2023-04-12 10:10:56 then .. how is wlan0 connected to the AP 2023-04-12 10:11:31 I just removed `iface wlan0 inet dhcp`. 2023-04-12 10:11:35 Oh, that's done by iwd. 2023-04-12 10:11:47 .. 2023-04-12 10:11:52 networking doesn't even detect my AP, and I never figured out why (same issue with the installation media). 2023-04-12 10:12:44 networking does not care if it is accociated. as long as the hardware is present 2023-04-12 10:13:24 Sorry, I was referring to `setup-interfaces`. It doesn´t detect my AP (but does detect a few others). 2023-04-12 10:15:41 since you have iwd, then, it is easy. start iwd for wlan0, start rest of all with dhcpcd-$i, where $i is eth0 usb0 balaba, edit dhcpcd.conf: nohook resolv.conf 2023-04-12 10:17:42 This approach seems to have worked. Just run dhcpcd-openrc (this version runs a single daemon that automatically operates on all interfaces), and remove dhcpcd handling from networking. 2023-04-12 10:18:16 what about dns :( 2023-04-12 10:19:38 dhcpcd handles that by default. 2023-04-12 10:20:00 but you said only want dns from wlan0 . 2023-04-12 10:20:15 usb0 doesn't advertise dns (it's a pmOS phone) 2023-04-12 10:20:40 it's news 2023-04-12 10:31:20 Shit. After 10 minutes something emptied by /etc/resolv.conf 2023-04-12 10:31:33 still running dhcp process? 2023-04-12 10:31:51 I'll check next time that it happens (I restarted it now). 2023-04-12 10:34:24 git push (without any changes) to my fork of aports on gitlab.a.o takes between uncomfortably long and forever. am i doing something wrong or is it just how life is in the git??b land? 2023-04-12 10:34:45 it is a thing that happens to some people with our instance for some reason 2023-04-12 10:34:55 in a lot of cases git pull takes a while for me for some reason when behind a bit 2023-04-12 10:44:19 so, the latter 2023-04-12 10:56:28 not surprising when you understand that gitlab is a steaming pile of ruby garbage 2023-04-12 10:57:00 ``` 2023-04-12 10:57:52 "git push (without any changes..." <- you can try syncing alpine/aports:master to your own master and then try to push, for me that usually fixes the issue 2023-04-12 10:58:15 dwfreed: it's mostly go, not ruby 2023-04-12 10:58:59 pj: you might be thinking of gitea? 2023-04-12 10:59:14 gitlab is 66.5% ruby according to their own metrics 2023-04-12 10:59:19 no, I'm not thinking of gitea 2023-04-12 10:59:20 gitaly is go 2023-04-12 10:59:25 which handles the git repos 2023-04-12 10:59:36 but it does call the webapp for authentication 2023-04-12 11:00:01 I doubt that any part of ruby code is actually responsible for that issue 2023-04-12 11:00:22 well, maybe 0.6% of gitaly ruby code is 2023-04-12 11:00:51 ikke: do you use gitlab sshd on alpine instance? 2023-04-12 11:00:55 no 2023-04-12 11:01:06 openssh sshd 2023-04-12 11:01:08 any plans? 2023-04-12 11:01:23 need to evaluate it 2023-04-12 11:01:28 pj: my master is up-to-date with upstream, but i still get almost hanging push every so often 2023-04-12 11:01:32 pj: do you see improvements? 2023-04-12 11:01:35 same, that's why I'm asking 2023-04-12 11:01:40 right 2023-04-12 11:02:04 it sounds like a terrible idea because it means being liable to gitlab code quality for an sshd implementation 2023-04-12 11:02:09 ovf: unfortunately then, you're special 2023-04-12 11:02:28 psykose: I would be more worried about Go SSH code quality 2023-04-12 11:02:28 psykose: you do rely on gitlab code for https authentication 2023-04-12 11:02:42 yeah 2023-04-12 11:04:07 still have to upgrade gitlab in prod, not a fan of that 2023-04-12 11:04:25 I'm upgrading gitlab once / twice per month 2023-04-12 11:05:34 we have 2 years old instance that wasn't touched :) 2023-04-12 11:05:36 fun fact: bitbucket uses the Go SSH server implementation 2023-04-12 11:05:59 (source: I know the guy that wrote their ssh server) 2023-04-12 11:06:02 fun fact: gitlab rewrites ruby for go to improve performance 2023-04-12 11:06:07 s/gitlab/github 2023-04-12 11:06:41 gitlab does too though 2023-04-12 11:06:52 like, for the past 4 years? 2023-04-12 11:06:57 they are removing all ruby code in the gitaly server 2023-04-12 11:07:04 one componenet left 2023-04-12 11:07:28 but also, a bunch of stuff was moved from the main codebase to gitaly 2023-04-12 11:07:32 true 2023-04-12 11:07:38 which is really fun when you're trying to find where some bug was introduced :') 2023-04-12 11:08:45 oh, lovely, instead of doing the sane thing and linking to an svg of their massive component diagram, they just *inline it* right into the html 2023-04-12 11:08:47 and after they're done with the go rewrite, it'll be time for rust? always something to do. :-) 2023-04-12 11:08:56 https://docs.gitlab.com/ee/development/architecture.html#component-diagram 2023-04-12 11:09:31 dwfreed: i'm not sure which is worse, that, or the png inserted in the gitaly readme: https://gitlab.com/gitlab-org/gitaly/ 2023-04-12 11:38:11 It happened again, `/etc/resolv.conf` was emptied. `dhcpcd: wlan0` was still running. 2023-04-12 11:41:18 Huh, I see the logs mention "dhcp6_openudp: Address in use\ndhcp6_start1: Address in use" around the time that this happens. 2023-04-12 12:30:44 I am planning to migrate my openbsd server to alpine. On openbsd I have used this script for blocking cyber attacks: https://www.geoghegan.ca/pub/pf-badhost/latest/pf-badhost.sh 2023-04-12 12:31:04 Is anything similar existing for alpine? 2023-04-12 12:31:48 That script is pulling down lists of IPs from various sources and then blocking pf blocks them. 2023-04-12 12:32:02 s/blocking // 2023-04-12 12:58:04 Darn, /etc/resolv.conf gets cleaned after every like 10-15 minutes. 2023-04-12 12:58:17 dhcpcd doesn't die though, apparently. 2023-04-12 13:23:21 WhyNotHugo: is 10-15 minutes suspiciously close to the dhcp lease time you get? 2023-04-12 13:25:09 No, it's 43200 seconds (12 hours) 2023-04-12 13:28:09 Any other interfaces that dhcpc might be running on? 2023-04-12 13:34:29 mercenary: I think not; I believe the other one had no carrier. I've left it unplugged to confirm. 2023-04-12 13:34:49 mercenary: But the second interface only connects to a dhcp server which serves no DNS. Would dhcpcd keep the existing one, or delete it? 2023-04-12 13:35:57 If it asked for DNS, and gets a lease but no DNS, it might well wipe it 2023-04-12 13:37:05 Is this proper dhclient or the udhcpc thing? 2023-04-12 13:39:38 Okay, I've timed it properly this time. /etc/resolv.conf is cleared at the same time that /var/log/messages shows: dhcpcd[4176]: dhcp6_openudp: Address in use dhcpcd[4176]: dhcp6_start1: Address in use 2023-04-12 13:39:58 mercenary: dhcpcd-9.4.1-r3 2023-04-12 13:43:46 And it happened again. This time I'm sure this was the only active interface 2023-04-12 13:44:03 It it aligns perrfectly with that error log. 2023-04-12 13:46:52 hmm. that implies dhcpcd is indeed wiping it. That comes with a 20-resolv.conf script, maybe put some debug info in that script to see when/how/why it is called? 2023-04-12 13:54:46 WhyNotHugo: in your /etc/network/interfaces files does it have the entry "auto eth0"? If so then remove it and ifupdown/ifupdown-ng will not run dhcpcd and then you can use the dhcpcd service for it 2023-04-12 13:55:26 (or whichever interface you want dhcp on) 2023-04-12 14:01:14 WhyNotHugo: those log messages obviously are DHCPv6 (IPv6) related 2023-04-12 14:34:33 minimal: "auto wlan0" runs dhcp? I thought it was "iface wlan0 inet dhcp" that run dhcp. 2023-04-12 14:38:17 WhyNotHugo: no, "auto wlan0" means that when the machine boots then Alpine with use ifupdown to "manage" the interface. If the /e/n/i file also defines "iface wlan0 inet dhcp" then ifupdown will run a dhcp client (i.e. dhcpcd) 2023-04-12 14:39:13 you earlier had the problem of a dhcpcd being run by ifupdown AND another dhcpcd being run by dhcpcd init.d service, removing "auto wlan0" from /e/n/i would then result in only a single dhcpcd being run by its init.d service 2023-04-12 14:40:25 I don't follow, does "auto wlan0" mean "automatically enable the interface" or "start a dhcp client on this interface"? 2023-04-12 14:40:59 I removed "iface wlan0 inet dhcp" with the expectation that this would not run a dhcp client on the interface. 2023-04-12 14:41:05 it means that the interface will be managed during boot, how it is managed is governed by the "iface wlan0 inet" line 2023-04-12 14:41:48 minimal: What does "managed" mean here? up/down? Or dhcp? 2023-04-12 14:42:02 EVERY interface with an auto line will be "managed" during boot as the networking init.d does a "ifup -a" (the "-a" means select every interface with an "auto" line) 2023-04-12 14:42:36 "managed" as in brought up, regardless of whether it uses dhcp, static ip, etc 2023-04-12 14:42:43 Okay, and how does that trigger a second dhcp? 2023-04-12 14:43:07 you said you had /etc/init.dhcpcd enable - that starts a dhcpcd 2023-04-12 14:43:34 you said you had "auto wlan0" and "iface wlan0 inet dhcp" in /e/n/i, that starts a dhcpcd 2023-04-12 14:43:59 I *removed* "iface wlan0 inet dhcp" from /e/n/i 2023-04-12 14:44:38 https://paste.sr.ht/blob/82ec9075367115a5231d2586d68eddd398339991 2023-04-12 14:44:48 Well, I left it commented now, but it was removed before, same idea 2023-04-12 14:44:57 ok, what referring to your original situation, you could have simply removed "auto wlan0" as the /etc/init.d/networking would ignore wlan0 comnpletely 2023-04-12 14:45:25 But I do want wlan0 to be brought up. 2023-04-12 14:45:42 by /etc/init.d/networking? 2023-04-12 14:47:35 but your /e/n/i file doesn't define how to configure wlan0 (static ip/dhcp/etc) so why have "auto wlan0" in /e/n/i ? 2023-04-12 14:48:27 Because I want it up. I have dhcpcd running and it operates on **all** interfaces. 2023-04-12 14:48:39 But the issue here is that my /etc/resolv.conf gets cleared every once in a while. 2023-04-12 14:49:02 right, so have /etc/init.d/dhcpcd bring it up then, not /etc/init.d/networking via /e/n/i 2023-04-12 14:50:02 If a list of interfaces are given on the command line, then dhcpcd only works with those 2023-04-12 14:50:36 dhcpcd --OPTION eth0 usb0. 2023-04-12 14:50:55 but don't put wlan0 here. instead, use networking to configure wlan0 2023-04-12 14:51:12 and then, in /etc/dhcpcd.conf: disable DNS 2023-04-12 14:51:52 WhyNotHugo: read : man dhcpcd : Multiple interfaces 2023-04-12 14:52:16 and then, you clone /etc/init.d/dhcp to /etc/init/dhcp.nowlan0 2023-04-12 14:52:28 WhyNotHugo: this confusing, which interfaces do you want the dhcpcd init.d service to manage? 2023-04-12 14:52:29 and do some edit. and start /etc/init/dhcp.nowlan0 2023-04-12 14:52:54 /etc/init/dhcpcd.nowlan0 2023-04-12 14:52:58 typo. 2023-04-12 14:53:21 they want all interface to be DHCP, but they only want DNS from wlan0 2023-04-12 14:53:37 minimal: wlan0 and usb0 (which is not always available!) 2023-04-12 14:53:42 minimal: that's what they told me many hours ago 2023-04-12 14:53:44 uniqNuuid: That won't work, usb0 is not always present. 2023-04-12 14:53:58 uniqNuuid: I didn't ask about DNS, I asked which interfaces should the dhcpcd service manage 2023-04-12 14:54:32 What I'm trying to figure out is who's emptying my /etc/resolv.conf and why. 2023-04-12 14:54:34 WhyNotHugo: then you may consider udev/mdevd 2023-04-12 14:54:47 those are rule based 2023-04-12 14:55:05 I don't want to invent my own mechanism and then maintain it forever, I want to understand what's broken now. 2023-04-12 14:55:11 WhyNotHugo: start by commenting out the "auto wlan0" line in /etc/network/interfaces 2023-04-12 14:55:36 WhyNotHugo: i told you, dhcpcd by default setup your dns, it set it up based on last interface it configures. 2023-04-12 14:55:55 if you use dhcpcd on all interfaces, the last one () defines your dns 2023-04-12 14:55:59 (random) 2023-04-12 14:56:11 uniqNuuid: And only wlan0 is connected here. usb0 is unplugged (e.g.: doesn't exist right now). 2023-04-12 14:56:29 and your dns is empty? 2023-04-12 14:57:07 uniqNuuid: Not this moment, because I just restarted dhcpcd. 2023-04-12 14:57:32 iwd? networking? 2023-04-12 14:57:45 you have so many services with same functionality configured 2023-04-12 14:58:10 WhyNotHugo: did you comment out the "auto" line? 2023-04-12 14:58:48 /e/i/n does not touch your resolv.conf, it lets dhcp helper do the job. 2023-04-12 14:59:04 iwd may touch resolv.conf 2023-04-12 14:59:24 dhcpcd 100% touch your resolv.conf by default 2023-04-12 14:59:52 uniqNuuid: I didn't say that, I simply asked if he's followed my initial suggestion 2023-04-12 15:02:06 he's not gonna follow your suggestion before he understands how those services/programs work all together 2023-04-12 15:02:31 ok, then I'll stop trying to help then 2023-04-12 15:03:27 he seems does not know what program does what, and he does not want to read manual 2023-04-12 15:07:30 Removed all references from wlan0 in /etc/networking/interfaces a while ago. Now again, resolv.conf is empty. 2023-04-12 15:07:39 This error is in the logs again: https://paste.sr.ht/~whynothugo/783519b02ed99f926bcdc36469d81f818d34ecc6 2023-04-12 15:07:47 It ALWAYS matches with resolv.conf getting emptied. 2023-04-12 15:08:09 The pid matches the instance of dhcpcd that had gotten the lease. 2023-04-12 15:09:01 minimal: I did follow your advise and remove all references ot wlan0 on /e/n/i. 2023-04-12 15:09:11 It's just that I was waiting -- resolv.conf gets emptied after a few minutes. 2023-04-12 15:13:38 Oh shit, I figure it out. I though dhcpcd had a "master" instance and then forked a new instance per interface (hence, why I was seeing two dhcpcd processes, one "manager" and one "wlan0") 2023-04-12 15:14:02 The wlan0 one is leftover by networking. If I remove it from networking and restart networking, it leaves behind the running dhcpcd child. 2023-04-12 15:19:51 Ok, to make sure, I can remove the boot media as soon as alpine is booted from RAM right? 2023-04-12 15:21:24 Liver_K: as long as it is not mounted 2023-04-12 15:22:10 And I can unmount it safely immediately? 2023-04-12 15:25:19 have you tried? 2023-04-12 15:27:40 Well I'm not there right now, but apparently I should have heh 2023-04-12 15:53:14 Liver_K: it's general Unix/Linux - if you run "umount" on a fs and it is being used then the umount will fail 2023-04-12 15:54:20 My worry was more that it might be needed for use later after unmounting 2023-04-12 15:55:02 well if you want to save something to it then obviously you'd need it there 2023-04-12 15:55:11 i.e. "lbu commit" 2023-04-12 15:55:29 Cool, I'm reading up on how to do that now 2023-04-12 16:04:21 Now I am closer to knowing the cause of my strange Firefox tab problem: I am using spectrwm (same setup as with other distros) and this tab problem only appears with the spectrwm in Alpine. 2023-04-12 16:05:42 Is there any other decent tiling wm for Alpine? 2023-04-12 16:21:28 engblom: wayland or xorg? 2023-04-12 16:22:16 https://wiki.archlinux.org/title/wayland 2023-04-12 16:23:02 sway dwm are good ones, pretty stable, 2023-04-12 16:27:04 I have been using dwm in the past, it means a lot of patching. 2023-04-12 16:27:27 Isn't sway an i3 clone? 2023-04-12 16:28:22 engblom: yes, it's i3 for Wayland 2023-04-12 16:28:25 more or less 2023-04-12 16:28:32 What I do not like about i3 is that every window got its own title bar instead of showing the title in the status bar. 2023-04-12 16:29:03 what would be the point of title bars if if it didn't do that 2023-04-12 16:30:05 grawity: With spectrwm, xmonad, dwm you have no title bar at all 2023-04-12 16:30:25 title can be disabled 2023-04-12 16:31:07 ^^ i mean on swat 2023-04-12 16:31:37 uniqNuuid: But I still nead the title, is it possible to put it in the status bar? 2023-04-12 16:32:11 title is on the status bar on the top as well as on window. 2023-04-12 16:32:22 you can disable the title on window 2023-04-12 16:34:01 Then maybe I should give sway a try. I am a bit against wayland (x11 is good). 2023-04-12 16:45:21 engblom: sway does the same regarding titlebars, so you probabý won't like it. 2023-04-12 16:45:37 engblom: https://github.com/swaywm/sway/issues/7409 2023-04-12 16:47:57 WhyNotHugo: Thanks! 2023-04-12 16:53:52 If I wanted a package to trigger mkinitfs, what's the approach I would take? I'd also want to avoid mkinitfs running if multiple of such packages are updated. 2023-04-12 16:55:01 packages can subscribe to changes in files on the filesystem by package upgrades / installations 2023-04-12 16:55:11 Not sure if the latter is possible though 2023-04-12 16:56:50 ikke: But that would require altering mkinitfs itself, right? 2023-04-12 16:57:51 When do you want to trigger mkinitfs? 2023-04-12 17:14:09 ikke: Trying to se if I can fit pbsplash into one. 2023-04-12 17:14:24 Mostly a learning thing TBH. 2023-04-12 17:14:52 In a post-install / post-upgrade script you can trigger mkinitfs yourself 2023-04-12 17:14:57 if that's what you want 2023-04-12 17:15:32 But the issue remains that it can trigger multiple times (mkinitfs is triggered if the kernel is upgraded for example) 2023-04-12 17:19:58 Yeah, that's what I'd like to avoid. 2023-04-12 17:21:06 I've been looking at pmOS and their approach is to have an mkinitfs-triggers directory. mkinitfs has an apk trigger whenever anything in that file changes, and packages all put a dummy file in there (the dummy file being touched on package build). 2023-04-12 17:21:12 It works, but I feel that it's somewhat hacky. 2023-04-12 17:21:46 apk itself does not have an alternative 2023-04-12 17:23:18 pacman (archlinux) has PostTransaction hooks 2023-04-12 17:24:28 Not sure that's either what you want 2023-04-12 17:24:59 this would involve mknitfs defining something to trigger which other packages can then execute 2023-04-12 17:25:00 pacman's PostTransaction hooks are pretty close to apk's triggers tbh. 2023-04-12 17:25:04 yeah 2023-04-12 17:25:28 And then logic to make sure that trigger is executed only once 2023-04-12 17:25:54 apk also runs a trigger once if multiple files match, right? 2023-04-12 17:26:00 If I update cryptsetup, does the initfs get rebuilt? 2023-04-12 17:26:02 yes 2023-04-12 17:26:11 But the direction is in reverse of what you want 2023-04-12 17:26:35 package A executes something if some files are changed 2023-04-12 17:27:15 (implied is that they are owned by other packages) 2023-04-12 17:27:18 So maybe pmOS's approach isn't bad at all. 2023-04-12 17:27:45 Yeah, indeed 2023-04-12 17:27:51 Although I keep feeling like there's SOMETHING wrong there. 2023-04-12 17:28:29 Like, packages basically register themselves saying "I want an initfs rebuild"... but then mkintifs knows to include that package via separate channels? 2023-04-12 17:28:56 and what defines what a "mkinitfs rebuild" involves? 2023-04-12 17:30:25 What triggers the update when cryptsetup updates? 2023-04-12 17:30:43 s/triggers the update/triggers the rebuild/ 2023-04-12 17:31:37 I don't think anything? 2023-04-12 17:32:16 no, mkinitfs' APKBUILD defines its trigger script is called when changes occur inside /usr/share/kernel/* 2023-04-12 17:32:34 yeah 2023-04-12 17:32:52 so if ANY package add/update modifies files in there then apk runs mkinitfs' trigger script 2023-04-12 17:32:54 Yeah, but cryptsetup doesn't have any files in that location. 2023-04-12 17:33:00 So if there's a bug in cryptsetup tomorrow, and I update it, the initfs keeps the old binary I guess. 2023-04-12 17:33:00 So it's not triggered 2023-04-12 17:33:19 I guess that answers my question :P 2023-04-12 17:33:19 so then the trigger is not called 2023-04-12 17:33:46 initfs are weird. They're a locally-built artifact, that needs to be updated on certain pacakge events. 2023-04-12 17:34:09 not necessarily 2023-04-12 17:34:29 I mean, built locally on Alpine and other distros, I know it can be done differently. 2023-04-12 17:35:01 isn't the whole "signed UKI" stuff that Systemd and some distros moving to where the initramfs is NOT updated locally, it is supplied as part of the distro provided UKI file 2023-04-12 17:35:10 It would involve: 1) a package defining some hook that can be triggered. 2) packages that can declare the hook should be triggered. 3) apk executing that hook once at the end if any of the subscribed packages change 2023-04-12 17:36:20 minimal: I used an UKI for a while (it was called UEFI bundle back then). The system adminsitrator has to sign it, since you need the key that's installed into the firmware to sign it. 2023-04-12 17:36:21 doesn't that mean they would need to enable every feature that someone needs? 2023-04-12 17:36:27 So it's either the local admin or the admin of a fleet, but not upstream distro. 2023-04-12 17:36:55 also some alternative initram tools like booster can be configured to autogenerate initramfs' tailored to the machine they're running on 2023-04-12 17:37:42 minimal: mkinitcpio also does that. It stripped drivers not in use in the system. 2023-04-12 17:38:16 ikke: I haven't kept up with what's going on with that exactly so don't know the answer, as it's an initramfs it wouldn't need every module, only modules *required* to boot into the distro (where then other modules can be loaded) 2023-04-12 17:39:05 so basically "just" fs modules, storage drivers, simpledrm, crypto (for LUKS) etc 2023-04-12 17:39:10 yes, but it would require all of them to be included 2023-04-12 17:40:45 https://fedoraproject.org/wiki/Changes/Unified_Kernel_Support_Phase_1 2023-04-12 17:40:53 The goal is to move away from initrd images being generated on the installed machine where possible." 2023-04-12 17:41:43 Phase 1, VMs, seems "straightforward" if you're assuming basically virtio-only 2023-04-12 17:42:48 also see that they're not removing non-UKI kernels, so I guess use UKI for *some* usecases 2023-04-12 17:45:07 talking about UKIs and mkinitfs triggers, I was tinkering on one of my UKI Alpine VMs the other day making mkinitfs feature file changes and regenerating the initramfs to test some stuff and couldn't understand why the changes I'd made have no effect.........until I realised that whilst the initramfs was being rebuild that the UKI was not lol 2023-04-12 18:26:52 yay! apk 2.14 actually ignores case for `search -d`, as has been documented since forever 2023-04-12 18:31:55 ovf: case sensitive makes trouble on search -d 2023-04-12 18:34:39 uniqNuuid: yes, and that got fixed 2023-04-12 18:41:24 Hi, is it possible to run appimage in Alpine? 2023-04-12 18:43:18 no 2023-04-12 18:54:08 now. musl handles tcp dns too. when to fix the poor malloc 2023-04-12 18:55:58 When you sumbit the patches which fix supposed inefficiency in malloc 2023-04-12 19:00:03 I believe that malloc is unfixable 2023-04-12 19:00:33 I just came in, "that" malloc? 2023-04-12 19:01:16 You can also come out 2023-04-12 19:01:28 long ago ;-) 2023-04-12 19:01:38 🎉 2023-04-12 19:03:31 I've been running my sway wayland sessions with libscudo. It seems fast, but all procs *seem* to take more RAM. 2023-04-12 19:45:46 "When you sumbit the patches..." <- What type of workloads would hit that (supposed) behavior?  2023-04-12 19:46:29 Saijin_Naib[m]: idk, it's not me who claims that this malloc is broken/poor/unfixable/etc 2023-04-12 19:50:11 Fair enough. I run repeat runs of a specific data processing Task, so I have a long log of times. Curious if running the same under a glibc distro would show this behavior or not, or if it'd be statistically insignificant variance  2023-04-12 19:50:55 Guessing it's the second  2023-04-12 19:57:14 Oh I see what you think I meant minimal, the boot medium being needed for writing at a later time -- I meant for reading, but I just read that it unloads the entire image into memory on boot so I guess I really don't need it at all after then 2023-04-12 19:58:44 Saijin_Naib[m]: mathematical statistics is a thing 2023-04-12 20:01:47 any of you have a symbol not found after updating some packages in edge ? (like copyq) 2023-04-12 20:02:47 Ermine: Indeed. 2023-04-12 20:02:48   2023-04-12 20:02:48 Know of a glibc distro that does -Os and does not use systemd? It will help focus the variables to what I want to actually look at  2023-04-12 20:05:48 Idk of the distros which do -Os 2023-04-12 20:35:24 Liver_K: what I thought you meant? nope, I only stated that you'd needed the media present *IF* you wanted to write to it (i.e. lbu commit), I never indicated you would need the media to read "anything" 2023-04-12 20:37:47 Saijin_Naib[m]: isn't the point that malloc-ng in musl places emphasis on "correctness" (of managing memory) over "speed"? Other malloc implementations may not have the same emphasis 2023-04-12 20:38:45 I'm sure dalias talked about the background of that in the past either in one of the Alpine channels or in the musl channel 2023-04-12 20:50:30 No idea! Curious if it'd even register as a significant variance or not in my workloads.  2023-04-12 20:50:53 I've certainly felt nothing to complain about with Alpine performance in general  2023-04-12 23:44:43 I've got a tor relay running on Alpine Linux. I'm trying to raise the file descriptor limit for the daemon user (tor), I've edited the `sysctl.conf` file to say `fs.file-max = 135000` and rebooted. I've confirmed that it's applied with `sysctl -a`. However, whenever I run `ulimit -Hn`, it still shows the default file descriptor limit (4096). 2023-04-12 23:57:03 /etc/security/limits.conf 2023-04-12 23:58:31 I edited that too: `tor soft nofile 135000` `tor hard nofile 135000`. Then I rebooted. 2023-04-13 00:00:32 holy cow, that's a huge upgrade. 498 MiB 2023-04-13 00:00:47 will firefox et al be working after this? 2023-04-13 00:01:46 are you running "ulimit -Hn" as the user "tor" 2023-04-13 00:08:19 Yes, the output from `ulimit -Hn` is coming from the tor user. 2023-04-13 00:11:13 the `/etc/security/` directory was not generated by default, I had to manually generate it as well as the `limits.conf` file. 2023-04-13 00:39:57 croc_: I believe you need to install the linux-pam package (which includes /etc/security/limits.conf) for this to work 2023-04-13 00:40:43 this comes with the pam_limits module 2023-04-13 00:44:04 hi all :-) i am having trouble setting sound on my alpine. i have this in my dmesg output: http://0x0.st/HX7i.log. i remember last time i set up sound i downloaded the missing driver files from the sof git directly but i would prefer installing them with an apk package but i can not find the right one. am i missing something? 2023-04-13 00:45:26 do you have the sof-firmware package installed? 2023-04-13 00:49:47 i did not... thank you 2023-04-13 01:30:23 Which do I unmount to remove the boot media? /dev/loop0 or /dev/sd*? 2023-04-13 01:41:07 minimal: I installed linux-pam and you're right, the rest of the files in `/etc/security/` appeared. However, I still don't think it's working, unless I configured it incorrectly. The steps I took was: Install `linux-pam` and `shadow-login`, alter the default `limits.conf` file with what I want. Then reboot. `ulimit -Hn` is still showing the default value. 2023-04-13 01:43:26 Liver_K: /dev/sd* 2023-04-13 01:48:41 minimal: Sorry I should have asked which one I am supposed to unmount, because when I tried /dev/sd* it wouldn't proceed and said "resource busy" 2023-04-13 01:49:06 what does "mount" show for /dev/sda? 2023-04-13 01:51:07 It's actually /dev/sdb, it is mounted at /media/sdb according to lsblk, `mount /dev/sdb' can't find /dev/sdb in fstab 2023-04-13 01:52:12 Liver_K: Make sure your current working directory isn't `/media/sdb/*` when you're trying to unmount the drive. 2023-04-13 01:52:26 It is not 2023-04-13 01:54:00 Liver_K: I didn't say to run "mount /dev/sdb", I said to run "mount" and look for /dev/sdb (i.e you could pipe that to "grep") 2023-04-13 01:54:38 "mount" shows EXISTING mounts, "mount /dev/sdb" tries to mount /dev/sdb 2023-04-13 01:55:19 Ah. mount also shows it is mounted at /media/sdb 2023-04-13 01:55:47 croc_ no idea. "ulimit" is a Busybox shell builtin, perhaps it doesn't work right, you could try installing bash and see if its ulimit acts differently? 2023-04-13 01:56:18 Liver_K: so "umount /media/sdb" says busy? 2023-04-13 01:56:29 Yes 2023-04-13 01:57:32 That is abnormal right? 2023-04-13 01:58:24 minimal: same result when I'm running bash. This is really weird, I do appreciate the help regardless. 2023-04-13 02:00:10 croc_: could be an issue with the pam config (assuming you're running Edge) - there have various changes in the past week or 2 2023-04-13 02:01:28 Nope, I haven't made any changes to the repos. 2023-04-13 02:06:13 Liver_K: right, had to boot a VM with a ISO to check as I don't use run-from-ram, do "umount /.modloop; umount /media/sdb" 2023-04-13 02:07:15 croc_: in general I'm suspicious of the current PAM config 2023-04-13 02:07:28 minimal: Ok yeah that's what I did before, doesn't that break anything? 2023-04-13 02:08:05 like what? 2023-04-13 02:08:33 You didn't say you were supposed to unmount the loop device first, is that for any reason at all? 2023-04-13 02:09:01 I forgot that it was being using for the modloop stuff 2023-04-13 02:09:20 when is on the boot media 2023-04-13 02:09:25 s/when/which/ 2023-04-13 02:28:31 Ok. After I have booted and unmounted/removed the boot medium, when I run setup-alpine, it comes up with a couple errors at the beginning and then proceeds to ask for a keyboard input. Is that to be expected? 2023-04-13 07:19:00 Continuing the upgrade transaction with new apk-tools 2023-04-13 07:19:15 245 pkgs got upgraded? 2023-04-13 07:23:10 ohh, this : main/*: rebuild for dt_relr change 2023-04-13 08:10:02 croc: the current pam config it's a pretty mess 2023-04-13 08:10:47 croc: uhm, it's the same croc_? 2023-04-13 08:16:50 croc_: you want to set rc_ulimit="-n 135000" in the conf.d of the service 2023-04-13 08:17:05 the pam stuff means nothing to init.d services 2023-04-13 08:17:15 i don't think it's ever used at all 2023-04-13 08:17:47 only if the service starts a pam session, but very few will do 2023-04-13 08:17:50 openrc doesn't do any pam stuff and it just changes user into the service one and no login or pam stuff is ever applied 2023-04-13 08:21:12 nofile is maxed at `sysctl -a | grep fs.nr_open` 2023-04-13 09:13:30 https://imgflip.com/i/7hzq34 2023-04-13 09:50:31 :D 2023-04-13 13:49:57 s/input/layout 2023-04-13 13:53:24 s/layout/input 2023-04-13 13:53:58 >:P 2023-04-13 13:54:06 Question still stands anyway 2023-04-13 13:54:47 MY question prefers to sit. 2023-04-13 13:56:01 Lazy XFaCiEer 2023-04-13 13:58:39 No, only my questions are lazy. 2023-04-13 14:02:19 *Silly XFaCiEer's questions 2023-04-13 15:11:24 What's DT_RELR? (context: aports@dcf926d126ee921df71282780cae6c160fe30783) 2023-04-13 15:14:46 https://maskray.me/blog/2021-10-31-relative-relocations-and-relr 2023-04-13 15:15:35 tl;dr: it's a fancy elf relocation compression technique 2023-04-13 15:16:00 average of 10% size reduction for elf objects 2023-04-13 15:16:14 bit more with pie, bit less without 2023-04-13 15:16:23 more for exec than sharedobject 2023-04-13 15:18:29 After I have booted and unmounted/removed the boot medium, when I run setup-alpine, it comes up with a couple errors at the beginning and then proceeds to ask for a keyboard layout. Is that to be expected? 2023-04-13 15:21:32 depends on the errors 2023-04-13 15:23:04 I upgraded more than 500 packages and don't noticed a big size reduction 2023-04-13 15:24:29 psykose: But are errors in general expected in this situation? 2023-04-13 15:24:45 donoban: which architecture 2023-04-13 15:24:50 Liver_K: can't you past that errors? 2023-04-13 15:24:50 Liver_K: depends on the errors 2023-04-13 15:24:57 x86_64 2023-04-13 15:24:59 s/past/paste 2023-04-13 15:25:16 did you actually check what apk said as the size difference with -i 2023-04-13 15:25:29 because idk how else you would "notice a reduction" for what is like 10% on average lol 2023-04-13 15:25:37 sorry friend, i didn't make it 95% 2023-04-13 15:26:11 it's also only elfs, so it's way below 10% when you account all the other files 2023-04-13 15:26:43 ok ok, just read some package reducing like 50% on alpine-dev 2023-04-13 15:27:11 * @psykose | >>> Size difference for libphonenumber: 17 MiB -> 6872 KiB 2023-04-13 15:27:24 that was a comedic exception :p 2023-04-13 15:27:31 hehe 2023-04-13 15:27:39 also only main/ is rebuilt more or less 2023-04-13 15:29:49 so DT_RELR is why we had that massive update? 2023-04-13 15:35:09 stuff was rebuilt, yes 2023-04-13 15:37:22 k 2023-04-13 17:39:30 psykose: Thanks. I admit I don't see a noticeable difference in total package size tho. 2023-04-13 17:39:41 But maybe binaries don't account for the majority of all the crap I have installed :P 2023-04-13 17:46:38 you'd be surprised how much is non-elf data 2023-04-13 18:00:31 gnome data 2023-04-13 19:19:31 Hi did the room changed? I see a lot of "user left the room". 2023-04-13 19:19:44 s/changed/change/ 2023-04-13 19:20:26 scorpion2185[m]1: the matrix bridge was disconnected 2023-04-13 19:21:58 Can I hibernate using a swap file on aarch64? On pmos actually I already asked there 2023-04-13 19:25:25 I made an entry in /etc/fstab, is that how you do it here? 2023-04-13 19:26:15 /swap-file none swap defaults 0 0 2023-04-13 19:27:55 i'm not sure why the architecture is relevant at all 2023-04-13 19:28:37 uboot instead of grub maybe? 2023-04-13 19:29:07 in case it is important I said it idk 2023-04-13 19:29:52 swapon -s no output 2023-04-13 19:35:14 that means you never mounted anything to swap in the first place 2023-04-13 19:36:09 I added an entry on /etc/fstab to do that, and I rebooted 2023-04-13 19:36:26 is that how you do it on Alpine? 2023-04-13 19:38:57 scorpion2185[m]1, you can just run 'swapon -a' 2023-04-13 19:39:19 all swap files/partitions in fstab will be mounted 2023-04-13 19:44:20 I added the entry day ago therefore I rebooted many times 2023-04-13 19:44:26 *days 2023-04-13 19:44:56 swapon -a turns it on 2023-04-13 19:45:23 I guess pmos is not using fstab at all 2023-04-13 19:46:55 hmm 2023-04-13 19:49:58 I need to enable localmount 2023-04-13 19:50:10 accordign to the link of pmos fstab 2023-04-13 19:51:29 When you use rc-update add localmount default, does it also start the service? 2023-04-13 19:51:37 no 2023-04-13 19:52:25 therefore it was already running since 2023-04-13 19:52:25 * WARNING: localmount has already been started` 2023-04-13 19:52:25 `service localmount start 2023-04-13 21:37:08 you shouldn't mess with localmount at all 2023-04-13 23:28:00 psykose: The errors are, loosely, that setup-alpine is looking in /media/sdb for the .setup-keymap-deps package 2023-04-13 23:31:34 hello 2023-04-13 23:55:40 So is there some practical way to move the package db on the install medium into memory? 2023-04-14 00:32:16 hi I didn't get my answer on go channel sadly, but go 1.20 release notes mention this: 2023-04-14 00:32:16 so how much and how does it affect the alpine docker images for Go (I guess there'd be some gains in performance?) 2023-04-14 00:32:16 >On Linux, the linker now selects the dynamic interpreter for glibc or musl at link time. 2023-04-14 00:37:13 why would it affect the performance 2023-04-14 00:39:57 How do I avoid that error? 2023-04-14 00:40:17 Something other than moving the package db into memory? 2023-04-14 00:44:23 you are X-Y problem-ing 2023-04-14 00:45:12 https://xyproblem.info/ 2023-04-14 02:41:36 Ack. 2023-04-14 02:42:44 dwfreed: I don't know if you saw the actual question from earlier, but I was just continuing it because it looked like everybody forgot about it 2023-04-14 02:44:55 dwfreed: If you look back in your scrollback you'll see that I already gave the problem, steps to reproduce, and an error message. That last thing was just trying to solicit an answer without re-asking everything over 2023-04-14 02:46:12 I mean, for one, you didn't actually pastebin the errors 2023-04-14 02:46:26 Well I haven't gotten networking setup on that machine yet 2023-04-14 02:46:44 you can write them to a removable disk 2023-04-14 02:46:58 Hrm fine. BRB. 2023-04-14 03:09:56 Gah I can't get every bit of the error message with tee or I'm doing it wrong 2023-04-14 03:10:28 It keeps cutting off the last bit after I SIGINT setup-alpine 2023-04-14 03:10:34 Even with tee -i 2023-04-14 03:25:38 Whatever, I just typed it in verbatim myself: https://0x0.st/H8KB.txt 2023-04-14 03:29:55 kbd-bkeymaps looks like a typo 2023-04-14 03:30:24 It's not 2023-04-14 03:30:34 At least not by me 2023-04-14 03:31:54 nevermind, it's a package https://pkgs.alpinelinux.org/package/edge/main/x86_64/kbd-bkeymaps 2023-04-14 03:45:02 a b infront of the thing is always intentional 2023-04-14 03:45:04 like babby 2023-04-14 03:46:47 I have managed to install dash-binsh using apk add --force --force-overwrite, but cant figure out how to now remove busybox-binsh 2023-04-14 03:47:18 perhaps the --force should've clued you to the fact you can't actually do it? 2023-04-14 03:47:31 psykose: So you have all the information now, is this problem avoidable easily? 2023-04-14 03:47:49 never seen it before 2023-04-14 03:49:08 you would have to log literally everything you ran up to that point and what you started with 2023-04-14 03:49:20 kbd-bkeymaps does exist afaik on the iso or whatever so it wouldn't give you that 2023-04-14 03:51:16 psykose: if i cant do it then why do those packages exist? 2023-04-14 03:51:26 because they exist 2023-04-14 03:51:38 if i cant do it then why are there options which claim to force but actually dont? 2023-04-14 03:51:55 that isnt an answer 2023-04-14 03:52:20 psykose: It works fine without any errors if I don't unmount the iso before running setup-alpine 2023-04-14 03:52:25 Strange huh 2023-04-14 03:52:32 Liver_K: you can't unmount the iso without breaking everything 2023-04-14 03:52:51 Well then, minimal has mislead me 2023-04-14 03:53:13 well, you kinda can, but not for that i guess 2023-04-14 03:53:17 i forget why you had to 2023-04-14 03:53:22 doesn't sound required for most stuff 2023-04-14 03:53:49 Well a RAM only image is kinda weird if you can't remove the boot medium 2023-04-14 03:54:07 Maybe there are some other benefits? 2023-04-14 03:55:26 Eh I guess just "small memory footprint" on its own is great 2023-04-14 03:55:36 it doesn't sound like you have any actual issues 2023-04-14 03:57:26 If "y=x+1, solve for x" is a "problem" I think it's safe to say this one is 2023-04-14 03:58:34 Goal is to have a functioning running diskless system, but actually without a disk 2023-04-14 05:10:28 btw is the adwaita gtk theme in the repos anywhere? i noticed gnome-themes-extra is no longer packaged 2023-04-14 05:10:56 adwaita is just implicitly from gtk 2023-04-14 05:11:00 afaik 2023-04-14 05:11:03 in gtk4 2023-04-14 05:11:29 and 3 2023-04-14 05:11:32 part of the tree https://gitlab.gnome.org/GNOME/gtk/-/tree/gtk-3-24/gtk/theme/Adwaita 2023-04-14 05:11:57 for 'libadwaita' (slightly different and gtk4 only), there's a version for it in gtk3 called adw-gtk3 2023-04-14 05:12:20 but yeah that's just default 2023-04-14 05:13:04 gnome-extra-themes was a port of adwaita to gtk2, and also a dark version, but gtk3 has the dark version by itself 2023-04-14 05:13:18 (now) 2023-04-14 05:13:25 and gtk2, is, well, gtk2 2023-04-14 05:16:22 thats kinda what i thought but the dark variant wasnt working. i just discovered that it works with Adwaita:dark as opposed to Adwaita-dark 2023-04-14 05:17:40 so thanks 2023-04-14 05:18:46 weird because Adwaita:dark seemed to not work on void im pretty sure 2023-04-14 05:18:55 i don't really recall how that worked tbh 2023-04-14 05:19:11 all i know with :dark is it also sets media-prefers in browser so you get the darktheme websites by default.. 2023-04-14 14:27:24 hellow, I'm seeing some Connection aborted when installing packages (especially rust & firefox), I wonder if apk has a builtin timeout or is my network just bad 2023-04-14 14:28:20 there is some long (kernel level) timeout which would be hit by the latter 2023-04-14 14:28:21 I'm already using the closest mirror, so I guessing I don't really have much options here 2023-04-14 14:28:22 or something 2023-04-14 16:09:44 Liver_K: I didn't mislead you, I answered the question you asked: "I can remove the boot media as soon as alpine is booted from RAM right?", you didn't clarify whether you were referring to booting the standard ISO or a system already installed, you didn't indicate that you intended to run "setup-alpine" after umounting the boot media. So I answered the question you asked not the one it appears you meant to ask ;-) 2023-04-14 16:22:50 minimal: I was talking about this "Liver_K> And I can unmount it safely immediately?" which you just told me to try 2023-04-14 16:24:19 Liver_K: well you could unmount it safely, doing so didn't affect the running system 2023-04-14 16:24:35 whether doing so would affect setup-alpine wasn't a question that you asked... 2023-04-14 16:25:16 I guess that means psykose saying unmounting it "breaks everything" isn't entirely true then 2023-04-14 16:25:21 indeed I'd assumed you were referring to an already installed system (which is why I'd mentioned "lbu commit" wouldn't work in you'd unmount/removed the storage media) 2023-04-14 16:26:17 Ok 2023-04-14 16:32:51 minimal: Also, weren't you gone when I sent that? 2023-04-14 16:34:17 "I'm already using the closest..." <- I made my own local apk cache and manually fetched the packages needed using a download manager, and kept fetching missing depends upgrades until I had what I needed. Inelegant, but worked until I was back in a good connection  2023-04-14 16:34:21 Liver_K: your point being? 2023-04-14 16:34:43 minimal: How did you see the message? 2023-04-14 16:34:50 2 words "irc logs" 2023-04-14 16:34:59 You have to be here to log 2023-04-14 16:35:08 Are there public logs? 2023-04-14 16:35:12 yes 2023-04-14 16:35:19 https://irclogs.alpinelinux.org 2023-04-14 16:35:23 Oh 2023-04-14 16:36:05 Why isn't that in the topic? 2023-04-14 16:36:42 :) 2023-04-14 16:40:14 what? omg 2023-04-14 16:40:17 ACTION pulls pants up 2023-04-14 16:40:31 Heh yup. 2023-04-14 19:17:49 Power management on my system is terrible, and there seem to be so many tools to deal with that, that I've no idea where to start. I'd prefer most hardware to aggresively power down when low on work. It uses 50W when idle, and 'powerctl standby' is not available for reasons that I can't figure out. 2023-04-14 19:18:10 Any suggestions? 2023-04-14 19:19:15 is it a laptop? 2023-04-14 19:19:55 if yes I would try tlp 2023-04-14 19:32:50 tweaking the bios and disabling some things maybe it helps? 2023-04-14 19:36:40 WhyNotHugo: what does 'cat /sys/power/state' tell you 2023-04-14 19:37:06 (regarding your standby issue) 2023-04-14 21:24:24 I'm trying to use multi-services (https://docs.alpinelinux.org/user-handbook/0.1a/Working/openrc.html#_multi_services) but it seems like start-stop-daemon refuses to start the second service running the same binary. How can I fix/workaround that? 2023-04-14 21:25:43 (first and second service is running a daemon, both using the same binaries, and there is an actual reason on why I'm running second service with the same binary) 2023-04-14 21:52:41 Hi, from a edge x86_64 installation, I'd like to add Xen. It was a couple of weeks ago, but I think I followed https://wiki.alpinelinux.org/wiki/Xen_Dom0, secion "Existing installation". 2023-04-14 21:54:09 When I choose one of the Xen boot options at the grub menu, the system restarts after the message 'Xen is relinquishing VGA console'. 2023-04-14 21:55:05 Also tried to run `setup-xen-dom0, but get several errors as 'modprobe: ERROR: could not insert 'xen_netback': No such device' 2023-04-14 21:55:25 How can this be fixed? 2023-04-14 22:06:48 Just looked at /boot/grub/grub.cfg and the lines regarding Xen are very different from the non-Xen, as they don't have all the modules: cryptsetup,kms,lvm and also cryptdm=lvmcrypt. The system is installed on a disk using cryptsetup and lvm. 2023-04-14 22:07:05 What a mess this grub.cfg file! 2023-04-14 22:08:04 Also load_video and insmod gzio were not inserted in the Xen entries. 2023-04-14 22:12:16 The line 'initrd /intel-ucode.img /initramfs-lts' is also not present in the grub's Xen entry, although there a 'initrd /intel-ucode.img /initramfs-lts'. 2023-04-15 03:54:14 Hey, could someone please help me with installing Docker on my Raspberry Pi 4, running Alpine Linux 3.17.3? I'm also running on a sys install, btw. So far, I've installed Docker, Docker-Compose, and performed the configuration for isolating containers with user namespace. When I run `docker info`, at the bottom of the output I see:... (full message at ) 2023-04-15 03:54:18 For some reason I went with the first solution on the documentation for it, ignoring the wrong version, but that didn't help, and after resorting to ChatGPT (mistake), I ended up needing to reinstall Alpine. I'm now ignoring the first solution, but the other two solutions also don't seem to apply to me. I don't see either /etc/default/grub or /etc/update-extlinux.conf within their directories, and update-extlinux doesn't seem 2023-04-15 03:54:18 to be a command either. 2023-04-15 06:40:14 <|cos|> good morning! i'm trying to get my first alpine machine up and running. 2023-04-15 06:40:24 <|cos|> the first thing i encountered was this bug: https://git.netizen.se/alpine-conf/commit/?h=fix/utf8_ssid 2023-04-15 06:41:39 |cos|: hey 2023-04-15 06:42:51 Do you mind making a merge request with that here? https://gitlab.alpinelinux.org/alpine/alpine-conf 2023-04-15 06:43:41 I see you found my previous MR fixing an issue in setup-interfaces :) 2023-04-15 06:43:55 / commit 2023-04-15 06:45:52 <|cos|> i'd preferably defer getting too entangled into administrative stuff prior to even having an alpine host running, but i guess the patch risks getting lost unless i follow process... 2023-04-15 06:49:22 <|cos|> i'm having trouble getting a graphical environment running. i did get xfce4 working back in july, so i believe it to be possible. my last attempt lacked luks-root though, so i scrapped it and now when reinstalling current alpine Xorg refuses to start. 2023-04-15 06:51:10 <|cos|> if having libva-intel-driver and xf86-video-intel i get an error about missing fcon, which supposedly was removed in linux 4.14. should one or both of those packages be purged from alpine too, or are they still usable? 2023-04-15 06:54:06 I still using xf86-video-intel on my laptop 2023-04-15 06:54:20 So it still usable 2023-04-15 06:57:22 <|cos|> i'm fumbling in the dark. am not really up to date with this stuff. those packages were not present in the installation that i scrapped. that's where my guess of obsoletion came from. 2023-04-15 07:05:17 Can you show the xorg log output? 2023-04-15 07:05:28 ~/.local/share/xorg/Xorg.0.log 2023-04-15 07:06:26 You do need some sort of x86-video-* package installed 2023-04-15 07:26:39 "For some reason I went with..." <- After some thinking, fiddling, and googling, the answer was simple and I am dumb lol. Raspberry Pi has its own bootloader, so all instructions on the Alpine Linux documentation wouldn't work normally. The answer was to instead set the cgroup options listed in `/boot/cmdline.txt`, all on the same line, and then reboot. 2023-04-15 07:36:10 <|cos|> https://tpaste.us/yYBp <- i'm missing the fbdev module, which seems better than missing the fbcon module. i did try adding kms to the initrd, without really understanding. 2023-04-15 07:38:08 <|cos|> oh. i'm confused. that's not the Xorg.0.log mentioning fbdev. apologies! 2023-04-15 07:42:09 <|cos|> what still confuses me is why Xorg reports i915.ko to fail loading when lsmod lists it as loaded. is it not the already present kernel module it attempts to load? what's going on?! 2023-04-15 07:57:48 <|cos|> https://gitlab.alpinelinux.org/alpine/alpine-conf/-/merge_requests/154 2023-04-15 07:59:23 Thank you 2023-04-15 10:02:29 <|cos|> There are no Release Notes for alpine somewhere, are there? 2023-04-15 10:03:35 https://alpinelinux.org/ … 2023-04-15 10:03:42 <|cos|> I'm asking after realizing https://wiki.alpinelinux.org/wiki/LVM_on_LUKS appears to be less crucial since 3.13 when support for installing to crypt was added in setup-disks (if I read git blame correctly) 2023-04-15 10:18:07 there are, e.g. https://alpinelinux.org/posts/Alpine-3.17.0-released.html with more stuff in an equivalent whatever-someone-wrote-in-the-wiki https://wiki.alpinelinux.org/wiki/Release_Notes_for_Alpine_3.17.0 2023-04-15 10:18:20 but they are not comprehensive, nor do they imply random wiki articles got updated for anything 2023-04-15 10:38:14 <|cos|> ah. there were! neither the post nor the wiki release notes mentions the addition of luks install though. I added a bolded **Note** to the wiki at https://wiki.alpinelinux.org/wiki/LVM_on_LUKS 2023-04-15 10:39:04 <|cos|> guess it would be good if someone with better insight in alpine could fact check my claim 2023-04-15 10:43:01 <|cos|> in addition to the graphics not working, i'm experiencing two other major issues. 2023-04-15 10:45:10 <|cos|> i seem to consistently get edge installed, even though selecing v3.17 when booting from netboot.xyz. does that make sense due to me misunderstanding, or should a bug report be filed somewhere? 2023-04-15 10:46:03 https://gitlab.alpinelinux.org/alpine/alpine-conf/-/issues/10554 2023-04-15 10:49:08 <|cos|> ah. thanks! that's well analyzed and work-around should get me going. 2023-04-15 10:51:14 <|cos|> the other issue is worse though. passwords don't seem to get set correctly. since it has happened repeatedly and i get past verification prompts i doubt it's me mistyping. also it seems to happen both for login passwords and luks passphrases. also well-known? 2023-04-15 10:54:50 that sounds like a keyboard layout issue 2023-04-15 11:27:29 <|cos|> I've posted https://tpaste.us/QRK0 containing my /var/log/Xorg.0.log after a fresh reinstall, fully using `setup-alpine` + `setup-xorg-base` with no attempts at loading modules manually or similar confusement. 2023-04-15 11:28:13 <|cos|> Did try setting MESA_LOADER_DRIVER_OVERRIDE as hinted as necessary on https://wiki.alpinelinux.org/wiki/Intel_Video but it appears to not make any difference. 2023-04-15 11:29:46 <|cos|> (The workaround for installing 3.17 didn't really work on my first attempt. Still got edge. Will try being more persisting wen reinstalling.) 2023-04-15 11:31:59 is there a /dev/dri/card* 2023-04-15 11:33:14 <|cos|> nope. /dev/dri is all empty. i'm suspecting some module doesn't get loaded, given that and that other errors i manage to transistion into is missing fbcon and fbdev modules 2023-04-15 11:33:46 You generally don't want to use fbcon / fbdev 2023-04-15 11:34:08 You can ignore errors like '(EE) Failed to load module "fbdev" (module does not exist, 0)' 2023-04-15 11:34:23 if it's all empty either your gpu is not detected entirely or udev never ran 2023-04-15 11:34:45 which would be weird for the latter as setup-devd udev is ran by setup-xorg-base 2023-04-15 11:35:40 <|cos|> i'm not sure there should be a gpu. this is a very old system. 2023-04-15 11:36:21 <|cos|> i read somewhere on the wiki that one should run old alpine releases on old hardware. is that the actual advise? 2023-04-15 11:36:24 There has to be a gpu 2023-04-15 11:36:34 what system do you have 2023-04-15 11:36:53 unless you connect via a serial console 2023-04-15 11:37:04 (or headless via network) 2023-04-15 11:38:07 <|cos|> a Dell Inspiron Mini 10 (PP19S) netbook, and this is why i really wish to use it (as a terminal)... http://blog.netizen.se/p/seasonal-outdoors-laptop/ 2023-04-15 11:40:48 you can "try" xf86-video-vesa or whatever they are to just do software rendering or whatever, but that doesn't sound very usable 2023-04-15 11:41:52 as for the gpu in question, gma500 isn't supported by mesa anymore afaik 2023-04-15 11:42:11 That's for 3d acceleration? 2023-04-15 11:42:16 there's a xf86-video-intel to directly use the xorg driver that maybe works or maybe doesn't 2023-04-15 11:42:21 or general hardware acceleration 2023-04-15 11:42:49 it's for "you get a graphical display" 2023-04-15 11:42:56 mesa? 2023-04-15 11:43:47 i dunno how xorg does this or what it allows you to software render yourself into 2023-04-15 11:52:23 <|cos|> heh. setting Driver to "vesa" effectively made Xorg an alias for reboot 2023-04-15 11:52:35 lol 2023-04-15 11:52:50 How much memory does your system have? 2023-04-15 11:53:28 <|cos|> 1GiB (: 2023-04-15 11:53:49 <|cos|> perfect for one rdesktop + ssh (: 2023-04-15 12:01:25 <|cos|> it's working! it seems vesa did the trick! and it didn't actually reboot, it just displayed the dell boot logo prior to blanking the screen and did god knows what with Ctrl-Alt+Fxx 2023-04-15 12:01:39 <|cos|> Thank you everyone who helped! 2023-04-15 12:09:33 <|cos|> having to explicitly specify vesa in the xorg.conf is a bit unexpected though. i read the intention that it should be within the fallback chain already. dunno if i'm prepared to fully open the can of worms it would be to put a question mark anywhere in this message... 2023-04-15 12:11:12 it is by default in the fallback chain 2023-04-15 12:11:16 well 2023-04-15 12:11:21 i also read that somewhere 2023-04-15 12:11:38 the log you posted picked modesetting 2023-04-15 12:11:43 which then failed with no gpu 2023-04-15 12:12:06 perhaps it's just silly and bad at fallbacks :D 2023-04-15 12:13:05 <|cos|> do we know how and or where detection happens? could a fix be as simple as matching pci id to driver somewhere? 2023-04-15 12:41:55 It's xorg itself that does the detection 2023-04-15 12:42:02 So you'd have to fix xorg 2023-04-15 12:43:59 hear that's not going so well these days 2023-04-15 19:36:06 uhM, sound broken after reboot, I'm not sure if I removed someting related 2023-04-15 19:36:22 https://tpaste.us/ogxk 2023-04-15 19:37:45 https://pkgs.alpinelinux.org/contents?file=pipewire-media-session&path=&name=&branch=edge uhM only riscv&&ppc64? 2023-04-15 19:39:26 maybe this: https://git.alpinelinux.org/aports/commit/?id=4dc6a181bf59 ? 2023-04-15 19:46:31 so I need to fix my config? 2023-04-15 19:46:49 switch to wireplumber yeah 2023-04-15 19:47:46 do you know what file should I edit? 2023-04-15 19:47:51 I already had wireplumber installed 2023-04-15 19:48:49 you start `wireplumber` instead of `pipewire-media-session` and that's it 2023-04-15 19:49:03 uh, i was starting just "pipewire" 2023-04-15 19:49:05 maybe you edited pipewire.conf to start the latter, maybe you started it manually 2023-04-15 19:49:06 change it with wireplumber? 2023-04-15 19:49:10 probably former then 2023-04-15 19:49:35 some people edit /etc/pipewire/pipewire.conf to start it for instance 2023-04-15 19:50:11 uhM 2023-04-15 19:50:22 there is also /usr/libexec/pipewire-launcher 2023-04-15 19:51:04 yeah that worked 2023-04-15 19:51:30 maybe the pipewire package could fix the conf, many people will lose sound this way 2023-04-15 19:51:46 have to go, thx guys see you 2023-04-15 19:52:05 cya 2023-04-15 19:52:07 the pipewire package doesn't start anything 2023-04-15 19:52:20 by default you get no sound without starting wireplumber 2023-04-15 19:52:32 or pipewire-media-session before that obviously 2023-04-15 19:52:48 I mean this psykose , on pipewire.conf 2023-04-15 19:52:56 yeah, that never did anything 2023-04-15 19:53:03 context.exec = [ 2023-04-15 19:53:04 #{ path = [ args = "" ] } 2023-04-15 19:53:06 `apk add pipewire; pipewire` -> this never gave you audio 2023-04-15 19:53:07 # 2023-04-15 19:53:08 # Execute the given program with arguments. 2023-04-15 19:53:10 yeah 2023-04-15 19:53:10 # 2023-04-15 19:53:12 # You can optionally start the session manager here, 2023-04-15 19:53:13 that never started anything 2023-04-15 19:53:14 # but it is better to start it as a systemd service. 2023-04-15 19:53:16 # Run the session manager with -h for options. 2023-04-15 19:53:18 # 2023-04-15 19:53:20 { path = "/usr/bin/pipewire-media-session" args = "" } 2023-04-15 19:53:22 oooh, sorry 2023-04-15 19:53:24 I get a tpaste url 2023-04-15 19:53:26 but missed copy&&pasting 2023-04-15 19:53:34 you edited that yourself 2023-04-15 19:53:52 hmm.. 2023-04-15 19:54:06 I will check 2023-04-15 19:54:32 i.e. by default there is no /etc/pipewire at all these days 2023-04-15 19:54:37 and even when it was shipped that was commented out 2023-04-15 19:55:00 but some long time ago it was recommended to add that there i guess 2023-04-15 20:00:32 ok ok 2023-04-16 05:48:57 need some help with initrd... (full message at ) 2023-04-16 05:54:02 /dev is populated on boot 2023-04-16 05:54:12 They are not files that are persisted 2023-04-16 05:54:51 If you are missing devices there, chance is you missing the kernel modules for them 2023-04-16 06:04:13 I mean, while booting block device couldn't be found and throwing me in rescue shell and there also block devices r not available... (full message at ) 2023-04-16 06:04:37 s/there//, s/```/`/, s/```/`/ 2023-04-16 06:07:13 XADE[m]: Would you be able to use a single line per message please, it's anoying for me to follow here where I need to follow a link to see what you are asking about 2023-04-16 06:07:27 (this channel is primarily irc) 2023-04-16 06:07:57 OK, sorry for this time 2023-04-16 06:08:33 What storage do you have? 2023-04-16 06:09:29 setup-alpine detects what you have and adds the proper entries to the feature list 2023-04-16 06:12:13 I've trying it on qemu, it's HDD i.e. /dev/sda 2023-04-16 06:13:03 How are you providing the disk to the vm? 2023-04-16 06:14:14 qemu-system-x86_64 -enable-kvm -boot menu=on -m 2g -cpu host -vga virtio disk.img 2023-04-16 06:15:02 ok, so ide 2023-04-16 06:15:10 so add `ide` to the feature list 2023-04-16 06:15:21 ok 2023-04-16 06:17:49 and make sure you regenerate the initramfs 2023-04-16 06:17:59 you might need something like ext4 as well 2023-04-16 06:19:09 sure, will keep that in mind. Also where can I find documentation for available features ? 2023-04-16 06:19:20 ls /etc/mkinitfs/features.d 2023-04-16 07:21:34 ata is the module for ide 2023-04-16 07:22:33 even though now block devices exist under /dev/, mounting them on /sysroot says no such file or directory :( 2023-04-16 07:25:35 What is your kernel cmdline 2023-04-16 07:27:02 you want rootfstype and root or cryptroot in there 2023-04-16 07:49:13 I don't have rootfstype in there, but still why am i not able to mount from rescue shell when those files exist 2023-04-16 07:51:41 did you do `mount -t $type x y` 2023-04-16 07:51:55 that's why you need rootfstype, since it sets the -t 2023-04-16 07:52:02 busybox mount by default doesn't really guess what the type is 2023-04-16 07:52:17 it can iterate and probe everything in /etc/filesystems to 'guess', but this sucks because 2023-04-16 07:52:23 - you could not have the fs type in there 2023-04-16 07:52:29 oh i'll try then 2023-04-16 07:52:30 - even if you do, it loads all the modules you probably don't want 2023-04-16 07:52:43 - that file is not present in the initrd that mounts your rootfs in this situations 2023-04-16 07:53:31 it would be better if that functionality was just removed entirely because idk who is happy with it 2023-04-16 07:54:05 aside from that util-linux-mount via `apk add mount` does the correct thing, not sure if it's copied to initrd via being present or not 2023-04-16 07:54:06 probably not 2023-04-16 11:11:41 http://lists.busybox.net/pipermail/busybox/2021-November/089343.html 2023-04-16 11:14:57 crickets 2023-04-16 11:16:40 I assume this is unmerged, and that busybox doesn't like doing anything useful except for existing in its current state 2023-04-16 11:22:54 haven't heard anything. maybe it should be added to alpine since this is a common complaint/confusion 2023-04-16 11:39:42 you should ping it 2023-04-16 11:50:54 do you experience mpv youttube and only get sound sometimes? 2023-04-16 14:13:22 frag, yes, if I don't set the format via ytdl-raw-options, sometimes the video is fetched, but not the audio track. 2023-04-16 14:14:05 somercet1: ah, i gues audio and not video .. 2023-04-16 14:14:12 *get 2023-04-16 14:15:04 I may have seen that, but usually it's no audio. 2023-04-17 17:45:14 prob been asked/answered. i cant get gpu acceleration working in wsl2/g. am i missing something? 2023-04-17 18:16:25 i think that mesa is compiled without d3d12 2023-04-17 18:17:54 indeed it's not 2023-04-17 18:18:00 not interested in supporting that either 2023-04-17 18:24:08 would I just need to build my own mesa-gl with gallium? 2023-04-17 18:24:24 or is there a limitation with musl atm 2023-04-17 18:24:37 you mean mesa-gallium with d3d12 2023-04-17 18:24:42 but it's basically just that yeah 2023-04-17 18:26:10 ill get to work then, really dont want to use ubuntu 2023-04-17 18:28:19 i'm not interested in adding it and i refuse 2023-04-17 18:28:23 microsoft can pay me to add their garbage 2023-04-17 18:28:35 xd 2023-04-17 18:30:51 is alpine wsl unofficial then? the agowa338 one 2023-04-17 18:31:21 yes 2023-04-17 18:48:14 is it just mesa-dri-gallium? 2023-04-17 18:49:27 yes 2023-04-17 18:53:19 i already got it to work (or rather, build), it's not very indepth 2023-04-17 18:53:35 it is however a useless driver with another 5% size to dri and 10% to va 2023-04-17 18:54:31 my recommendation to using alpine is to not use windows, shrug 2023-04-17 18:54:38 or just use ubuntu as wsl itself intended 2023-04-17 18:54:46 unless they recommend cbl-mariner these days 2023-04-17 19:19:34 cbl-mariner is not recommended at all and won't be 2023-04-17 19:33:28 i get your position, but i dislike ubuntu and like alpine. so if i can get alpine working nicely with wsl2/g thats a win 2023-04-17 19:35:49 you can get it to work locally sure, here you go https://img.ayaya.dev/GO8LkastT08x 2023-04-17 19:43:46 the patch has +case "$CARCH" in +x86_64) but the source is case "$CARCH" in x86*), just curious how it came out that way 2023-04-17 19:44:30 there's both, read again 2023-04-17 19:47:31 oh sorry 2023-04-17 19:53:31 am i applying the patch to this? https://git.alpinelinux.org/aports/plain/main/mesa/APKBUILD 2023-04-17 19:54:10 oh derp, its a diff 2023-04-17 19:54:33 aports 2023-04-17 20:02:58 sorry, im pretty noob but im catching up 2023-04-17 20:06:29 do you have main/directx-headers/APKBUILD you can provide 2023-04-17 20:14:56 nm, i think i can recreate it from the diff 2023-04-17 20:16:01 idk why the command didnt respect new file mode 100644 2023-04-17 20:25:49 you probably ran something other than patch -Np1 2023-04-17 20:35:22 indeed, ran -p1, noob 2023-04-17 20:52:07 well its all installed, but still Accelerated: no and its not detecting Vendor: Microsoft Corporation (0xffffffff) Device: D3D12 (Radeon RX Vega) (0xffffffff), more digging, but fun learning 2023-04-17 20:53:31 are you sure you installed the new mesa-dri-gallium 2023-04-17 20:53:48 and then reopened the whatever everything of stuff 2023-04-17 21:02:28 i did a wsl --shutdown alpine. mesa-dri-gallium-23.0.2-r4 x86_64 {mesa} (MIT SGI-B-2.0 BSL-1.0) [installed] 2023-04-17 21:03:53 installed all of these packages doas apk add mesa-gl-23.0.2-r4.apk mesa-gbm-23.0.2-r4.apk mesa-egl-23.0.2-r4.apk mesa-glapi-23.0.2-r4.a pk mesa-23.0.2-r4.apk mesa-dri-gallium-23.0.2-r4.apk directx-headers-1.610.0-r0.apk 2023-04-17 21:21:37 Greetings! 2023-04-17 21:21:47 Anybody here interested in making schilytools packages for Alpine? 2023-04-17 21:22:15 Schilytools (https://codeberg.org/schilytools/schilytools) is a collection of libraries and tools including the famous cdrtools, star, smake, sccs, and others. 2023-04-17 21:22:33 I'm the maintainer of this project and would assist you in your efforts. 2023-04-17 22:01:52 FUZxxl, making them? Like, to see if they work or do you want someone to take over the package? 2023-04-17 22:03:24 somercet1: as in to package it at all, i.e. a packaging request 2023-04-17 22:04:11 ah 2023-04-17 22:04:33 somercet1: yes correct, this is a packaging request. 2023-04-17 22:04:51 I am not an Alpine Linux user. I maintain the project and the FreeBSD ports, but do not have capacities for additional ports. 2023-04-17 22:06:24 why would someone use this? 2023-04-17 22:06:50 (that question is not in good faith so don't get sniped into answering it) 2023-04-17 22:07:07 why is it not in good faith? 2023-04-17 22:08:24 you have a habit of those :p 2023-04-17 22:08:49 (as do i) 2023-04-17 22:09:01 I never heard of schillytools and "Schily-Tools are a set of tools developed or maintained by Jörg Schilling. " doesn't tell me why would someone use it 2023-04-17 22:09:12 cdrtools bruv 2023-04-17 22:09:23 history 2023-04-17 22:09:28 thank you for this rather useless answer 2023-04-17 22:09:51 bruv 2023-04-17 22:09:53 https://en.wikipedia.org/wiki/Cdrtools 2023-04-17 22:10:03 pj: schilytools contains cdrtools, the standard UNIX disc burning suite 2023-04-17 22:10:12 it also contains star, a very good tar implementation 2023-04-17 22:10:18 thanks 2023-04-17 22:10:42 as well as other tools, such as an implementation of SCCS, a screen editor (ved), a modernised Bourne shell (bosh), and assorted other tools. 2023-04-17 22:11:03 bosh is a cute name 2023-04-17 22:11:47 oh and bosh comes with two variants: obosh (which has only the original Bourne shell features) and pbosh (which has exactly the features specified by POSIX) 2023-04-17 22:12:48 > Starting with version 2.01.01a09 in May 2006, most code from cdrtools has been relicensed under the CDDL 2023-04-17 22:13:18 and looks like cdrkit is packaged in alpine 2023-04-17 22:14:18 that looks like a prehistory version 2023-04-17 22:15:42 ah, it's that one 2023-04-17 22:17:27 so many different burning apps, cdrdao, dvd+rw-tools, &c 2023-04-17 22:18:21 pj: cdrkit is a spite fork from the Debian people 2023-04-17 22:18:26 development ceased in 2006 2023-04-17 22:18:29 or 2007 maybe 2023-04-17 22:18:40 meanwhile cdrtools is still maintained 2023-04-17 22:19:13 it's an especially hilarious spite fork 2023-04-17 22:19:28 i'm not exactly sure why it matters if a licence in itself is gpl-compatible for arbitrary software 2023-04-17 22:19:59 knowing Jörg, it was 80% because they wanted an excuse not to have to work with him 2023-04-17 22:20:00 (calling it hilarious is how i sleep at night, because this stuff drives me up the wall) 2023-04-17 22:20:12 psykose: muh derivative work 2023-04-17 22:20:24 well.. you can still do derivative work on the cddl? 2023-04-17 22:20:26 the license conflict is somewhat obscure and pretty much irrelevant 2023-04-17 22:20:31 psykose: you can 2023-04-17 22:20:50 it's only an issue if you want to copy in some gpl code into it or vice versa or whatever, uh, theoretically 2023-04-17 22:20:56 the issue raised is that the build scripts are CDDL licensed open source whereas some parts of the software are GPL licensed open source 2023-04-17 22:21:06 anyway, maybe it would be good to drop cdrkit 2023-04-17 22:21:20 if it died in 2k7 why is it still packaged 2023-04-17 22:21:21 technically, this means it is not distributable as the build scripts cannot be redistributed in a GPL-compatible way 2023-04-17 22:21:34 sleeping emoji 2023-04-17 22:21:42 (i know why, it's rhetorical) 2023-04-17 22:21:53 but you know, even Suse lawyers in their audit found that it would be extremely hard to prove any damage from this and to win a court case 2023-04-17 22:21:57 so really, it's a none issue 2023-04-17 22:22:09 psykose: imagine bed 2023-04-17 22:22:37 imagining 2023-04-17 22:22:46 now imagine sleeping 2023-04-17 22:25:25 cdrtools now handles dvds and sao/dao/tao, right? 2023-04-17 22:26:03 yes 2023-04-17 22:26:11 It can also do Bluray 2023-04-17 22:26:30 ah. Does it still output Latin-1? 2023-04-17 22:26:50 yes :-) 2023-04-17 22:26:55 schily was a big fan of latin1 2023-04-17 22:27:04 yes he was!! 2023-04-17 22:27:11 did you know him? 2023-04-17 22:27:14 or he just hated unicode. 2023-04-17 22:27:26 he didn't 2023-04-17 22:27:26 no, no, I was just around for the Big Spat. 2023-04-17 22:27:32 it has just always been this way 2023-04-17 22:27:40 ISO-8859-1 is the standard encoding for German UNIX 2023-04-17 22:27:48 so why change anything for these newfangled UTF-8 people? 2023-04-17 22:27:54 that's his way of thinking 2023-04-17 22:27:56 echt deutsch 2023-04-17 22:28:03 frankly, we are looking into changing this, but it's hard. 2023-04-17 22:28:15 The codebase is huge and has character set shit everywhere 2023-04-17 22:28:20 who knows what the impact is? 2023-04-17 22:28:24 I believe it. 2023-04-17 22:29:36 I haven't even tried using abuild yet. I've been installing apps in /usr/local, with the tarballs to know what to remove later on. 2023-04-17 22:29:49 https://lwn.net/Articles/195169/ 2023-04-17 22:29:51 haha 2023-04-17 22:29:57 i had never read this stuff before 2023-04-17 22:36:54 oh, and the SCSI thing... Schilling refused to allow /dev/sdc or the like, you had to scan the drive adapters to find the SCSI LUN, like 0,1,1. 2023-04-17 22:37:40 "It all came flooding back to me like the hot kiss on the end of a wet fist." 2023-04-17 22:39:51 it's still like this. 2023-04-17 22:40:19 Though I'm willing to accept patches that change the situation, as long as they do not lead to the limitations Jörg outlined. 2023-04-17 22:40:30 Does Alpine have current CD/DVD-rw tools? 2023-04-17 22:41:34 for what purpose specifically 2023-04-17 22:42:00 cyanrip is good for.. cd ripping 2023-04-17 22:42:00 we have cdrkit, but it is apparently unmaintained 2023-04-17 22:42:07 no, burning. 2023-04-17 22:42:25 dvd+rwtools exists 2023-04-17 22:42:33 as someone that hasn't burnt a dvd since 2023-04-17 22:42:38 ACTION checks 2023-04-17 22:42:41 2010? 2023-04-17 22:42:45 i have no idea what is good or works 2023-04-17 22:43:33 (the dvdtools thing uses cdrkit) 2023-04-17 22:44:45 ... it can also use cdrtools! 2023-04-17 22:45:17 i need to learn apk packaging, i can help 2023-04-17 22:45:35 this is probably the worst place to start 2023-04-17 22:45:59 well the build system is a bit special 2023-04-17 22:46:05 but not too hard to wrap 2023-04-17 22:46:34 if you want an easier project, perhaps you could package gcli: https://gitlab.com/herrhotzenplotz/gcli 2023-04-17 22:48:45 if it builds its fine 2023-04-17 22:48:55 packaging isnt difficult as long as stuff builds 2023-04-17 22:49:02 main issue is that it needs to be split into a bunch of subpackages 2023-04-17 22:49:04 ive done debian and fedora packaging 2023-04-17 22:49:06 you would be very incorrect 2023-04-17 22:49:08 with interdependencies 2023-04-17 22:49:42 ive never had issues packaging, i struggle with linking and flags and such 2023-04-17 22:49:42 I don't know how you handle config file templates 2023-04-17 22:49:47 but there are a bunch of these in there too 2023-04-17 22:49:55 also setuid binaries and stuff like that 2023-04-17 22:51:29 stuff building is usually the least amount of work that needs to be done for package 2023-04-17 22:51:52 randvmb, if you need another pair of eyes, let me know 2023-04-17 22:52:13 i guess it depends on what you know 2023-04-17 22:52:22 ttps://github.com/freebsd/freebsd-ports/tree/main/devel/schilybase 2023-04-17 22:52:25 https://github.com/freebsd/freebsd-ports/tree/main/devel/schilybase 2023-04-17 22:52:26 ive done a lot of deployment and scripting 2023-04-17 22:52:28 little building 2023-04-17 22:52:36 ^ these are the FreeBSD files for packaging schilytools 2023-04-17 22:52:50 there's a bunch of slave ports for the other components as we don't have subpackages on FreeBSD 2023-04-17 22:53:01 but this is the master Makefile included from the others 2023-04-17 22:53:57 randvmb, you should probably look at Void (musl) and Arch pkgbuilds (like abuild), too. 2023-04-17 22:56:00 i was learning abuild earlier trying to get d3d12 working for wslg 2023-04-17 22:56:58 I used cdrtools for a long time (on both Linux and Solaris, and perhaps on other Unix systems)...."those were the days" 2023-04-17 22:57:31 nice to hear! 2023-04-17 22:57:52 the whole cdrkit/wodim fork was "interesting" at the time lol 2023-04-17 22:58:39 Alpine currently uses xorrisofs to create its ISOs 2023-04-17 23:01:51 Rock Ridge? Rock Right? Splendid! 2023-04-17 23:02:03 ...........Mongo like candy ;-) 2023-04-18 00:37:12 looks like gallium d3d12 needs _gallium_drivers="r300,r600,radeonsi,nouveau,swrast,virgl,zink,d3d12" 2023-04-18 00:37:38 an aur pkgbuild does just swrast,d3d12, fine for solely sw and wslg i guess 2023-04-18 00:40:55 trying them now 2023-04-18 00:41:03 that's exactly what the patch did 2023-04-18 00:41:17 like, literally 2023-04-18 00:41:32 i see it in the patch now 2023-04-18 00:41:48 but i just opened the apkbuild in main/mesa it wasnt in there 2023-04-18 00:42:16 but it added all the other stuff 2023-04-18 00:42:28 ? 2023-04-18 00:42:48 it doesn't do literally literally that, the "other stuff" does literally that however 2023-04-18 00:42:50 oh it put it in extra gallium per arch 2023-04-18 00:42:52 ic ic 2023-04-18 00:43:10 _gallium_drivers= + _gallium_drivers="$_gallium_drivers,d3d12" 2023-04-18 00:43:44 im trying just swrast, gallium 2023-04-18 00:43:58 i can already tell you that won't do anything 2023-04-18 00:44:04 just so you can save yourself some time 2023-04-18 00:44:51 i was just reading a debian bug report, they waited to add it because some discrepancy with the dx headers vs what mesa expected 2023-04-18 00:45:23 you can try 1.608.2 i guess 2023-04-18 00:46:29 https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1014472 2023-04-18 00:47:51 that's about a build error only 2023-04-18 00:48:30 *grumbles* why am i even wasting my mental time thinking about this microsoft garbage 2023-04-18 00:48:30 where can I submit a support request regarding Nextcloud 24.0.9-r0 on Alpine 3.16 getting upgraded to Version 24.0.11 released March 24 2023? 2023-04-18 00:48:47 nextcloud on alpine 3.16 is not supported 2023-04-18 00:48:48 so, nowhere 2023-04-18 00:51:54 the mesa sources dx wrapper is for revision v1.606.4 2023-04-18 00:51:55 psykose, so I would upgrade to 3.17 then? 2023-04-18 00:51:57 would that matter? 2023-04-18 00:53:20 pretty much 2023-04-18 00:54:46 similar question: where could I submit a support request regarding Nextcloud 25.0.4-r0 on Alpine 3.17 getting upgraded to Version 25.0.5 released March 24 2023? 2023-04-18 00:58:50 https://gitlab.alpinelinux.org/alpine/aports/-/issues 2023-04-18 02:34:18 im hoping this is it, i overlooked it and as soon as i saw it i facepalmed 2023-04-18 02:34:20 https://devblogs.microsoft.com/commandline/d3d12-gpu-video-acceleration-in-the-windows-subsystem-for-linux-now-available/ 2023-04-18 02:34:29 -Dgallium-va=true 2023-04-18 02:37:18 omg im done, im walking away 2023-04-18 02:43:29 heh 2023-04-18 02:55:24 sorry for wasting your efforts @psykose 2023-04-18 02:55:31 it requires sytemd -_- 2023-04-18 02:57:04 LIBVA_DRIVER_NAME=d3d12 vainfo gives nothing? 2023-04-18 02:57:09 i don't see why systemd is relevant at all 2023-04-18 02:57:40 (that part isn't for gpu but rather va, but still a test) 2023-04-18 02:57:53 i poorly assume something to do with initializing the microsoft display driver, ill check in a moment, i started from the beginning and am building the packages again 2023-04-18 03:05:57 i need the libva-utils apkbuild 2023-04-18 03:06:37 you just install it 2023-04-18 03:06:56 for which you have to be on edge i guess 2023-04-18 03:07:33 libva-utils (no such package): 2023-04-18 03:07:34 i got it now 2023-04-18 03:07:55 libva error: /usr/lib/dri/d3d12_drv_video.so init failed 2023-04-18 03:07:56 libva info: va_openDriver() returns 18 2023-04-18 03:08:01 vaInitialize failed with error code 18 (invalid parameter),exit 2023-04-18 03:08:24 what gpu do you have 2023-04-18 03:08:30 rx vega 2023-04-18 03:08:59 https://img.ayaya.dev/MzgoZTuf0qVr 2023-04-18 03:09:11 pretty sure vega is older than rx5000 2023-04-18 03:09:11 xD 2023-04-18 03:09:24 one gen behind 2023-04-18 03:10:02 (that's video accel; not gpu itself) 2023-04-18 03:10:09 so vainfo will never work 2023-04-18 03:10:14 but idk where the gpu list is 2023-04-18 03:12:35 i get the same on ubuntu, but i also get this https://i.imgur.com/liY00Dj.png 2023-04-18 03:13:23 yeah that is 'working glxinfo' 2023-04-18 03:14:24 what do you get for GALLIUM_DRIVER=d3d12 glxinfo here 2023-04-18 03:15:27 in alpine? 2023-04-18 03:16:41 yeah 2023-04-18 03:18:42 https://i.imgur.com/LcoEMaE.png 2023-04-18 03:19:51 the whole thing 2023-04-18 03:20:03 preferably not as a picture of your screen 2023-04-18 03:20:48 its a screenshot lol, that was the whole output 2023-04-18 03:22:00 https://i.imgur.com/s8ur8cI.png 2023-04-18 03:30:31 that doesn't have the command you ran 2023-04-18 03:35:01 i ran clear; GALLIUM_DRIVER=d3d12 glxinfo 2023-04-18 03:35:47 and /usr/lib/xorg/modules/dri/d3d12_dri.so exists right 2023-04-18 03:39:16 ~ $ ls /usr/lib/xorg/modules/dri/ crocus_dri.so kms_swrast_dri.so radeonsi_dri.so zink_dri.so d3d12_dri.so nouveau_dri.so swrast_dri.so i915_dri.so r300_dri.so virtio_gpu_dri.so iris_dri.so r600_dri.so vmwgfx_dri.so 2023-04-18 08:28:42 Hmm I have a harddisk encrypted with dmcrypt and when unlocked it has a LVM partition on top of it. Even though the right services are enabled at boot time, I can't get it to properly initialize that partition on reboot. It does decrypt it but I still have to manually run `vgchange -ay && lvscan -ay` after which I can run `mount -a`. With dmcrypt and lvm services enabled on boot and , why doesn't this happen automatically? 2023-04-18 08:44:33 PureTryOut: is lvm enabled in initfs? 2023-04-18 08:46:21 It... is not. Although I was assuming the partition would be enabled after initfs once the regular system has started 2023-04-18 11:07:37 invoked: It lists "freeze mem disk". Like I said; no "standby". 2023-04-18 11:07:43 donoban: No, it's a desktop. 2023-04-18 11:11:47 you probably want mem 2023-04-18 11:11:57 what does cat /sys/power/mem_sleep have 2023-04-18 11:12:13 psykose: s2idle [deep] 2023-04-18 11:12:52 "deep" is "s2ram" which is "standby" 2023-04-18 11:13:13 they have 50 terms for this i guess 2023-04-18 11:13:14 https://www.kernel.org/doc/html/v4.18/admin-guide/pm/sleep-states.html 2023-04-18 11:13:19 that has all the info you want more or less 2023-04-18 11:18:32 "deep" and "s2idle" are variants of "s2ram". But standby should be more energy efficient that s2ram. 2023-04-18 11:20:28 if you mean s2ram in the 15-year-old sense of ACPI S1 2023-04-18 11:20:52 mem_sleep="deep" means the full ACPI S3 standby mode 2023-04-18 11:21:21 (mem_sleep="s2idle" means the S0ix "Modern Standby" which is a bit less efficient, but Windows Likes It™) 2023-04-18 11:22:58 What I'm trying to understand is why /sys/power/state has no "standby"; this variant is documented as being more energy-saving than "mem_sleep". 2023-04-18 11:23:12 Based on that kernel.org link or on powerctl's man page: https://git.sr.ht/~sircmpwn/powerctl/tree/master/item/powerctl.8.scd#L38 2023-04-18 11:23:43 Currently this uses 5W when suspended to memory, which is more than what my laptop used when NOT suspended. 2023-04-18 11:24:09 I don't think I've seen "standby" on any system I've used Linux on 2023-04-18 11:24:28 "mem" is the one that invokes full suspend mode, if mem_sleep=deep 2023-04-18 11:25:16 I see "standby" is listed in kernel sources 2023-04-18 11:25:23 apparently it is ACPI S1 2023-04-18 11:25:43 that's *less* efficient than S3 (deep), but also mostly obsolete 2023-04-18 11:26:10 So mem/deep should be the most power efficient one? 2023-04-18 11:26:15 yes 2023-04-18 11:27:02 my understanding is 2023-04-18 11:27:15 all manufacturers completely dropped S1 ("standby") support at some point 2023-04-18 11:27:27 and all desktops and tools got accustomed to there being only one suspend state, "mem" 2023-04-18 11:28:01 I wonder if there's something that I can do to improve this, or it's just hardware limitations. 2023-04-18 11:28:03 so when they added support for another new state (S0ix "modern standby" or "low power idle") 2023-04-18 11:28:13 But I'm using 5W suspended and 50W when idle, which is all ridiculous. 2023-04-18 11:28:20 they ended up bodging it onto "mem" for compat reasons instead of re-adding a whole new state 2023-04-18 11:28:45 make sure mem_sleep is indeed set to "deep" though, 5W does sound a lot 2023-04-18 11:28:59 is it above by cat [] 2023-04-18 11:29:09 is it a laptop or desktop? desktop 5w seems.. not rare 2023-04-18 11:29:35 desktop, apparently 2023-04-18 11:29:51 maybe it's powering a NIC for wake-on-lan, or something along those lines 2023-04-18 11:30:54 It says 3.2W today (the PSU is connected to a meter under the desk). It's a desktop. 2023-04-18 11:31:04 3.2 indeed sounds average as hell 2023-04-18 11:31:07 blech 2023-04-18 11:31:15 i wish systems had per-component power draw 2023-04-18 11:31:24 +1 2023-04-18 11:31:30 hmm even 3.2W for s3 sleep sounds a bit too much at first glance 2023-04-18 11:31:47 3.2 is atrocious in arm land 2023-04-18 11:31:49 it's just the RAM and maybe the NIC, no? 2023-04-18 11:32:12 OTOH, 3.2 is what a laptop might use when editing a text file. Including a display!! 2023-04-18 11:32:14 I mean my dell "micro PC" home server is 5W when powered-on and idle, and it's x64 2023-04-18 11:32:18 this reminds me maybe i should get a meter 2023-04-18 11:32:25 and optimise my desktop a bit 2023-04-18 11:32:27 WhyNotHugo: do you have like 8 sticks of RAM or something? 2023-04-18 11:32:53 2 sticks of ram. The video card is... well, for videogames. But I expect it to power off. 2023-04-18 11:32:56 i have more like 'an amd boost raising' instead and no power settings so i bet i have like 20w idle like a chump 2023-04-18 11:33:01 Maybe that's the expectation that's wrong. 2023-04-18 11:33:27 one way to find out: remove it 2023-04-18 11:33:54 lol 2023-04-18 11:33:56 The GPU fans are off most of the day (basically, any time except when gaming). 2023-04-18 12:21:05 Can I easily move from stable to edge and then back to stable once 3.18 comes out? 2023-04-18 12:21:13 Or if it's gonna be a hassle, I'll stay put till it will be out 2023-04-18 12:22:18 more or less 2023-04-18 13:28:26 i'd like to get into the tasmota supported devices for power monitoring, someday 2023-04-18 14:53:53 is it normal for the aarch64-virt image to not have a UEFI config that's compatible with UTM on macOS? 2023-04-18 14:54:28 sounds like https://gitlab.alpinelinux.org/alpine/aports/-/issues/14407 2023-04-18 14:55:12 the standard image does work though 2023-04-18 14:55:47 or wait, it just spinloops and dies, gonna see what the serial output says and comment that on gitlab 2023-04-18 14:56:47 drat, no serial output 2023-04-18 14:56:56 ah well, at least it's a known issue 2023-04-18 15:07:35 Xe: this is UTM using Apple Virt, rather than UTM using Qemu, right? 2023-04-18 15:07:52 minimal: yeah 2023-04-18 15:08:04 (apple virt is more power-efficient) 2023-04-18 15:09:45 Xe: and Grub config is missing "insmod linux" as mentioned in that Issue? 2023-04-18 15:52:08 how can i get the PID of the last started instance of eg mpv? 2023-04-18 15:52:41 that is still running or 2023-04-18 15:52:58 pgrep mpv 2023-04-18 15:53:11 if you mean started-multiple-get-latest 2023-04-18 15:53:12 hmm 2023-04-18 15:53:13 no idea 2023-04-18 15:53:47 a quick hack is -eo start,pid | sort -g. doesn't quite work if it's been more than a day, but it's a hack 2023-04-18 15:54:34 there's no good way to convert a procname to a pid to begin with 2023-04-18 15:54:41 so yeah, everything on top of it sucks 2023-04-18 15:54:53 gotta make do 2023-04-18 15:58:10 i need to kill the last instance of mpv thats still running yeah 2023-04-18 15:58:30 its running in the bg with only audion :p 2023-04-18 15:58:56 ah kill 'em all ;) 2023-04-18 15:59:09 barbaric! :o 2023-04-18 15:59:20 frag: do you remember the mpv args? ps -ef | grep '[m]pv' lists all 2023-04-18 16:00:05 uniqNuuid: oh this is helpful :) 2023-04-18 16:00:26 also you can use etime option of ps 2023-04-18 16:01:04 got that sucker! 2023-04-18 16:01:22 ps -eo pid,etime,args | grep mpv 2023-04-18 16:01:56 the etime is ELAPSED time 2023-04-18 16:02:58 maybe not with busybox ps :P 2023-04-18 16:04:28 yes also busybox :D 2023-04-18 16:04:43 what would we do without busybox 2023-04-18 16:11:31 frag: "apk add " ? 2023-04-18 16:51:00 ikr :] 2023-04-18 16:53:49 ikari warriors 2023-04-18 18:26:50 is there a meta package for shell completion like there is for -doc packages? 2023-04-18 18:27:36 meta package for -doc pkgs? 2023-04-18 18:30:00 if you apk add docs, every new package you install will /also/ install package-doc 2023-04-18 18:30:24 so i was hoping something like bash-completions would soft-depend (or whatever the mechanism is) on every -bash-completion package 2023-04-18 18:34:29 there is literally a bash-completion 2023-04-18 19:01:01 psykose, but does that pull in other completion packages? 2023-04-18 19:01:12 other? 2023-04-18 19:01:35 ye 2023-04-18 19:03:57 somercet1: Yes, it does the equivalent of what docs does for man pages. 2023-04-18 19:04:04 ah 2023-04-18 19:04:16 somercet1: There's also subpackages for zsh ;) 2023-04-18 19:04:38 zsh itself pulls completions funnily 2023-04-18 19:04:40 so you always get them 2023-04-18 19:04:42 sad! 2023-04-18 19:05:10 psykose: Zsh is an interactive shell. It's not made for scripting. I hope nobody pulls it for non-interactive usage. 2023-04-18 19:05:24 i never really use completions for anything 2023-04-18 19:06:00 of course, i got a bit tired of weird random stuff aside from that, so even i moved to zsh these daysa 2023-04-18 19:06:05 and enabled all the 'completions' 2023-04-18 19:06:08 maybe it's just the defaults some distros use, but zsh gets in my way on things like Arch install dvds... 2023-04-18 19:06:09 hit or miss on if they do anything 2023-04-18 19:06:13 i never use the -ones 2023-04-18 19:06:18 psykose: oh duh 2023-04-18 19:06:22 i thought i'd checked that 2023-04-18 19:06:29 and everythihng magic like 'git' or 'ssh' is 50/50 on if it works or breaks 2023-04-18 19:06:30 meh 2023-04-18 19:08:01 thanks psykose! 2023-04-18 19:08:05 np :) 2023-04-18 19:11:56 frag: pkill -n mpv 2023-04-18 19:12:08 somercet1: I didn't find the one in Arch's install media intuitive. OTOH, I use it in a sort of vim mode and with fzf shortcuts. 2023-04-18 19:12:42 then it's a lousy advertisement. 2023-04-18 19:12:43 2023-04-18 19:13:25 @ikke: you have good taste in House music 🤣 2023-04-18 19:13:38 ikke: share 2023-04-18 19:15:40 Hello71: wow, cool, ty! 2023-04-18 20:54:21 How does one become a package maintainer? I see InfluxDB is an old version. I understand it's possible to report old packages, but surely there's a way to becoming a maintainer to alleviate the toil of other maintainers? 2023-04-18 20:55:09 L1Cafe: You can mail the existing maintainer if they are okay that you take over maintainership 2023-04-18 20:55:19 Fair enough 2023-04-18 20:59:35 any idea why emacs-gtk3 cannot open the display when GDK_BACKEND=wayland? or is that another suport channel question? 2023-04-18 21:00:08 it doesn't use actual gtk and does its own custom thing 2023-04-18 21:00:16 emacs 29 uses actual gtk (not released) 2023-04-18 21:00:52 aha 2023-04-18 21:01:04 thanks! that explains it. time to build from source! :P 2023-04-18 21:18:03 Just realised there's no "current maintainer" https://pkgs.alpinelinux.org/package/edge/community/x86_64/influxdb, how to apply to become maintainer? 2023-04-18 21:18:57 Poof - all packages gone, 3.17 still has the old one though https://pkgs.alpinelinux.org/package/v3.17/community/x86_64/influxdb 2023-04-18 21:19:40 Anyone know how to reach an Alpine Wiki administrator? I'm trying to create a talk page about a change in conf options for rsnapshot and it's marking me as a spammer... 2023-04-18 21:20:29 any tips on building emacs w musl? 2023-04-18 21:21:04 acdw, the port probably has hints 2023-04-18 21:21:09 oh duh thanks 2023-04-18 21:21:19 you mean the APKBUILD ? 2023-04-18 21:21:27 yes 2023-04-18 21:21:50 thanks :) 2023-04-18 21:21:55 np :) 2023-04-18 21:22:45 jahway603: you reached one 2023-04-18 21:22:58 jahway603: what is your wiki username? 2023-04-18 21:24:07 jahway603, my first edit was so tagged, but it took it. It said I posted a URL, which I emphatically did not do. 2023-04-18 21:25:24 ikke: my wiki username is Jahway603 2023-04-18 21:25:46 somercet1: I posted a URL in my first edit to a page explaining the new retain options 2023-04-18 21:26:38 well, mine was a page edit, not creation, so URLs *were* on the page. 2023-04-18 21:26:46 so I dunno 2023-04-18 21:27:24 jahway603: can you try again? 2023-04-18 21:29:03 ikke: thanks, that worked 2023-04-18 21:35:25 build-base depends on gcc? how can i set up a minimal musl build env? 2023-04-18 21:35:28 libc-dev? 2023-04-18 21:40:37 i think i'm confused. musl is /just/ a library, correct? sorry i'm new at this... 2023-04-18 21:43:31 yes, it's indeed just a library 2023-04-18 21:43:45 one that is linked to almost any program though 2023-04-18 21:43:53 ahh 2023-04-18 21:44:07 okay so if i install gcc that's not installing glibc ... it's gcc linked against musl 2023-04-18 21:44:10 i /think/ 2023-04-18 21:49:11 gcc is the c/c++ compiler, libc is the c language std library 2023-04-18 21:49:21 acdw: yes it is linked against musl 2023-04-18 21:49:25 on most systems, it's glibc, here, it's musl 2023-04-18 21:49:39 glibc is not supported on alpine 2023-04-18 21:50:59 ahhhhh 2023-04-18 21:51:25 thanks for the clarification. i got confused. i compile my own stuff from source fairly often but i'm still basically a consumer of that software 2023-04-18 23:50:47 how is the latest firefox edge update working under Wayland? 2023-04-19 08:28:36 anyone have any ideas about why my gvim fonts would be screwed up after upgrading alpine? fc-list still sees the fonts, but gvim has reverted to only the microsoft ttf fonts 2023-04-19 08:29:14 I uninstalled fonts and reinstalled them for fun, and also cleared all the caches, nothing seems to change what gvim can see 2023-04-19 08:32:08 i'm going to guess ncurses 2023-04-19 08:43:09 Is it possible to install on a device without an ethernet port? 2023-04-19 08:43:22 I cannot connect to wifi following the wiki page about wifi because it instructs to install iwd or wpa_supplicant first. 2023-04-19 08:43:32 So it seems to implicitly assume an ethernet connection already exists. 2023-04-19 08:43:36 It doesn't though, because my only option to connect to the internet is wifi. 2023-04-19 08:44:03 you can download individual package files, and install them with apk 2023-04-19 08:44:22 download on a different machine, ofc 2023-04-19 08:45:24 then using i.e. an USB pendrive, move the files to your network-less machine 2023-04-19 08:48:17 Thank you for the hint, I will research how to do that. 2023-04-19 08:48:42 but consider package dependencies 2023-04-19 08:49:11 you can use this to check deps: https://pkgs.alpinelinux.org/packages 2023-04-19 08:54:47 ok, so it seems that something is wrong with the mix of fontconfig and terminus-font packages in 3.11, where fontconfig doesn't think terminus is monospaced....upgrading to terminus-font from 3.14 fixes the problem 2023-04-19 08:55:31 which I guess is when terminus started shipping with otb in addition to the original pcfs (which are still not seen as monospace) 2023-04-19 09:01:32 RZ_MJ[m]: you can use the extended image which has wpa_supplicant included 2023-04-19 09:03:37 Oh thanks, I will try that. 2023-04-19 10:04:40 hewwo, is xmonad usable on alpine? 2023-04-19 11:57:21 Hello 2023-04-19 11:57:28 Guys I keep having a very weird problem with alpine 2023-04-19 11:57:40 So I'm trying to install Akkoma atm 2023-04-19 11:57:54 I have a directory full of binary files 2023-04-19 11:58:37 https://0x0.st/H8YS.txt 2023-04-19 11:59:05 However, that's what happens when I try to run any of those files: 2023-04-19 11:59:12 https://0x0.st/H8YQ.txt 2023-04-19 11:59:18 I have no idea why it happens 2023-04-19 12:03:23 Some files actually seem to launch' 2023-04-19 12:03:35 ACTION sent a code block: https://matrix.org/_matrix/media/v3/download/matrix.org/cSbMgPkNhAGiucRouazgdaZV 2023-04-19 12:03:52 > <@maksim:wherelinux.xyz> ```... (full message at ) 2023-04-19 12:04:34 I've no idea why it happens 2023-04-19 12:06:27 I thought that the issue might be in my Proxmox lvm running out of space, but the same works on Debian, what's wrong? 2023-04-19 12:24:46 hi. i'm trying to configure an alpine router. Please see here: https://pastebin.com/KGkQv3JT 2023-04-19 12:24:58 what am i missing? anything else anyone needs to see? 2023-04-19 12:28:44 a description of the problem, here 2023-04-19 12:42:38 dhcp and dns work on client. ping shows "network is unreachable" for upstream router lan ip and internet ips and domains. it seems like traffic is not forwarded but i'm not sure why. 2023-04-19 12:43:34 Found a fix for my problem 2023-04-19 12:43:39 Blame glibc 2023-04-19 12:43:41 here's the fix 2023-04-19 12:43:42 https://stackoverflow.com/questions/66963068/docker-alpine-executable-binary-not-found-even-if-in-path/66974607#66974607 2023-04-19 12:43:54 It's for docker, but this also works (probably) https://github.com/sgerrand/alpine-pkg-glibc 2023-04-19 13:31:26 any one avail? 2023-04-19 13:32:25 602 ought to be 2023-04-19 13:32:32 i'm not 2023-04-19 14:04:09 I've been able to update this Alpine wiki article with new info, but couldn't fix the script in the article. Anyone here more versed in shell scripting that can fix the script located here? https://wiki.alpinelinux.org/wiki/Rsnapshot 2023-04-19 14:17:23 ${dirs[@]} is not a thing in sh 2023-04-19 14:19:53 arrays other than $@ are not a thing in sh 2023-04-19 14:23:57 not sure why there was a script at all 2023-04-19 14:54:11 I tried to add espeakup/espeak-ng/espeak_soft to initramfs but this is my first time here and not sure why it's not working or how to debug what is going on during these early times. I want espeakup/speech to let me know when the luks prompt comes up for passphrase for my drive. aka, I don't want to rely on a display. 2023-04-19 14:55:11 I made an espeakup.modules with kernel/drivers/accessibility/speakup/* in it 2023-04-19 14:55:33 espeakup.files with /usr/bin/espeak*, /usr/lib/libespeak*, and /usr/share/espeak-ng-data/* 2023-04-19 14:56:43 then added "espeakup" to "features" in the conf and ran mkinitfs but no speech for the luks passphrase on reboot. :) 2023-04-19 14:57:03 I did get one error once: lddtree: /usr/share/espeak-ng-data/voices/!v/Mr: file does not exist 2023-04-19 14:57:03 lddtree: serious: file does not exist 2023-04-19 14:57:25 So maybe mkinitfs or lddtree scripts don't handle spaces? Running the same mkinitfs command more times seems to succeed. 2023-04-19 14:58:49 Do I need to do something to cause the "espeakup" program to start? 2023-04-19 15:01:46 psykose: should I just remove that rsnapshot broken script from the wiki article? 2023-04-19 15:02:25 i would also be happy to receive a quote from any dev with online presence for the funding of an up to date, working router/firewall wiki page for alpine that uses nftables and unbound 2023-04-19 15:05:11 got some tips about my initramfs issue from z3ntu in #main:postmarketos.org so am good for now. Cheers. :) 2023-04-19 15:10:04 jahway603: already did 2023-04-19 15:10:19 it was clear enough already how to do it without it :p 2023-04-19 15:32:00 unrznbl[m]: some months ago I did help someone with getting espeak activated for an installed system, I didn't add it to the initramfs though 2023-04-19 15:34:11 unrznbl[m]: are you intending to use espeak via a sound card (rather than a speach synth card)? 2023-04-19 16:44:32 Where can I find the APKBUILD files for nextcloud plugins offered on pkgs.alpinelinux.org? I can only find the one for nextcloud when I click thru the git repo on any of them 2023-04-19 16:49:43 jahway603: which plugins? the nextcloud package has multiple subpackages, all handled by the nextcloud APKBUILD file 2023-04-19 16:51:33 minimal: For example, I see weather_status here https://git.alpinelinux.org/aports/tree/community/nextcloud/APKBUILD?h=3.16-stable#n91 but I do not see a section in the APKBUILD where it is installed and I can't find a separate APKBUILD for that pkg (or any of them) 2023-04-19 16:52:04 trying to understand how these additional nextcloud packages get installed as I'd like to update the calendar one 2023-04-19 16:52:54 jahway603: as I said it is a subpackage of nextcloud 2023-04-19 16:54:11 minimal: I'm not following - I search for weather in the nextcloud APKBUILD and it only appears once 2023-04-19 16:54:17 jahway603: https://git.alpinelinux.org/aports/tree/community/nextcloud/APKBUILD?h=3.16-stable#n91 2023-04-19 16:55:22 and scroll up 2023-04-19 16:56:39 now you're sharing the same link I shared *sigh* 2023-04-19 16:57:28 and did you scroll up? 2023-04-19 16:57:41 yep & I still have the same question 2023-04-19 16:57:51 and did you read line 53? *sigh* 2023-04-19 16:58:26 yes & still don't understand 2023-04-19 16:58:52 is there some general Alpine Linux docs on understanding subpackages in APKBUILD? 2023-04-19 16:59:14 whether you understand or not doesn't change the fact that this APKBUILD file is used to create the weather status subpackage 2023-04-19 16:59:29 you don't understand line 58? 2023-04-19 17:01:32 I just upgraded alpine from 3.16 to 3.17, which upgraded nextcloud from 24.x to 25.x & then proceeded to set the calendar plugin to an "untested app" with version 3.5.4... then we look at https://apps.nextcloud.com/apps/calendar and see that calendar for Nextcloud 25.x is v 4.3.3 .... 2023-04-19 17:01:56 I'm trying to understand what is happening in the APKBUILD so I can create a nextcloud-calendar "submodule" 2023-04-19 17:02:07 or subpackage or whatever it's called 2023-04-19 17:02:20 jahway603: you didn't answer my question about line 58... 2023-04-19 17:02:30 Is there a script that will install and setup network manager? 2023-04-19 17:02:59 AnyCreatureMorph[m]: when your msg pops up, my terminal shakes. 2023-04-19 17:03:32 setup-interface 2023-04-19 17:03:42 rc-service networking start 2023-04-19 17:04:03 minimal: is _apps a variable and I can just add calendar to the list to fix what I'm trying to fix? 2023-04-19 17:04:30 jahway603: I'm talking about your previous question regarding weather status 2023-04-19 17:04:39 so what does line 58 do? 2023-04-19 17:05:18 uniqNuuid: I want my wifi to work in gnome control center. 2023-04-19 17:05:22 i don't care about weather status, as it's working fine -- it was just an example for my initial question of how does this work 2023-04-19 17:05:44 jahway603: well my question is relevant to how it works 2023-04-19 17:06:16 minimal: I clearly am not understanding -- care to explain instead of asking me if I get it when I don't? 2023-04-19 17:06:22 AnyCreatureMorph[m]: install iwd 2023-04-19 17:06:32 jahway603: well what do you think line 58 is doing? 2023-04-19 17:06:45 what line 58 is doing is not relevant to anything 2023-04-19 17:07:03 psykose: its relevant to lines 93-95 ;-) 2023-04-19 17:07:10 uniqNuuid: When I install it will the wifi show in gnome settings? 2023-04-19 17:07:19 jahway603: those are the nextcloud modules shipped with nextcloud itself, there are no packaged external ones like calendar 2023-04-19 17:09:08 AnyCreatureMorph[m]: `ip -c link` do you see your wifi device? 2023-04-19 19:35:01 "AnyCreatureMorph: `ip -c link..." <- Yeah 2023-04-19 19:47:41 not to interrupt, but nextcloud is an EU AWS, with various built-in services (like Google) such as weather, calendaring, etc? 2023-04-19 19:59:56 uhm what 2023-04-19 20:00:08 what are you asking 2023-04-19 20:32:08   2023-04-19 20:32:08 I think I get them, and it's a private cloud storage provider with tons of apps and integrations. 2023-04-19 20:32:08 Have you looked at the site? It has good documentation  2023-04-19 20:32:37 You can even self-host 2023-04-19 21:14:07 "unrznbl: are you intending to..." <- Absolutely. So adding espeak_soft and include espeak-ng binaries and voices in initramfs. 2023-04-19 21:14:58 "unrznbl: some months ago I did..." <- Yes, I have it working just fine, comes up just around first login after bootup, but I really need feedback when the luks passphrase is requested. :) 2023-04-19 21:15:00 unrznbl[m]: and sound card drivers? it won't get far trying to use sound card for speech without those ;-) 2023-04-19 21:15:27 oh dang. so true. :p 🤦‍♂️ thank you. 2023-04-19 21:18:30 is there something I need to do to start espeakup the program which makes it all happen? maybe add it to the mkinitfs init script? 2023-04-19 21:22:18 or maybe add a script in /usr/share/mkinitfs/hooks ? like debug-shell does in postmarket? https://gitlab.com/postmarketOS/pmaports/-/blob/master/main/postmarketos-mkinitfs-hook-debug-shell/APKBUILD 2023-04-19 21:31:14 mkinitfs doesn't do hooks at all does it 2023-04-19 21:33:07 ah, right, postmarket custom maybe: https://gitlab.com/postmarketOS/pmaports/-/blob/master/main/postmarketos-mkinitfs/APKBUILD 2023-04-19 21:34:42 yeah, that's a go project, not the /sbin/mkinitfs script I have on my thinkpad w/ alpine :( so maybe do add a command to the initscript after all... at /usr/share/mkinitfs/initramfs-init? 2023-04-19 22:18:52 unrznbl[m]: yeah that's the other point, you need to run the espeak program in the initramfs' init script and also stop it before the switch_root 2023-04-20 01:30:49 Hey I am seeing man-db in the online package index, but I can't find it with apk when booted from the installer 2023-04-20 01:30:52 Why is that? 2023-04-20 01:37:49 forgot to add the actual http repos 2023-04-20 01:38:07 setup-apkrepos or whatever it was, which the installer runs 2023-04-20 02:00:18 psykose: I already configured the installer environment with setup-alpine, shouldn't that run the rest of the setup scripts too? 2023-04-20 02:00:25 it does 2023-04-20 02:00:33 so, what do you have in /etc/apk/repositories 2023-04-20 02:00:39 and does it have 'community' 2023-04-20 02:17:16 psykose: https://0x0.st/H8I7.txt 2023-04-20 02:17:23 well 2023-04-20 02:17:27 it does not have community 2023-04-20 02:17:34 uncomment line 3 2023-04-20 02:17:36 :) 2023-04-20 02:17:50 but, you probably don't want it 2023-04-20 02:17:59 you usually want `mandoc` for manpages, not man-db 2023-04-20 02:18:04 (unrelated to enabling community or not) 2023-04-20 02:18:05 I was gonna ask why those were commented lol 2023-04-20 02:18:12 they just are by default 2023-04-20 02:18:21 no real reason aside from it is that way i guess 2023-04-20 02:18:25 Ok I'll keep that in mind when I actually install :) 2023-04-20 02:18:27 don't uncomment the edge ones 2023-04-20 02:18:42 unless you want to use edge, in which case comment the first lines and uncomment only edge 2023-04-20 02:19:20 Cool thanks 2023-04-20 06:07:41 Is MGLRU enabled fully on linux-lts under latest-stable? 2023-04-20 06:07:42 https://gitlab.alpinelinux.org/alpine/aports/-/blob/master/main/linux-lts/lts.x86_64.config#L147 2023-04-20 06:08:18 This line seems to build the kernel with support for it, but I'm not seeing the line to enable it 2023-04-20 06:08:26 https://www.kernel.org/doc/html/latest/admin-guide/mm/multigen_lru.html#multi-gen-lru 2023-04-20 06:10:01 it is yeah 2023-04-20 06:10:19 ah 2023-04-20 06:10:20 hm 2023-04-20 06:11:18 zgrep CONFIG_LRU_GEN /proc/config.gz 2023-04-20 06:11:35 nah without _enabled it defaults off 2023-04-20 06:11:39 cat /sys/kernel/mm/lru_gen/enabled 2023-04-20 06:11:41 > 0 2023-04-20 06:19:44 Is that something that can be added within a stable release series 2023-04-20 06:19:51 I think it would help my work massively 2023-04-20 06:20:27 And we have a kernel point release coming up Thursday/Friday 2023-04-20 06:21:05 already did 2023-04-20 06:21:17 was probably overlooked 2023-04-20 06:22:04 Oh, wow. Sweet! You are the best 2023-04-20 06:22:37 Wish I didn't have to worry about memory pressure and resource contention, but my builds are not new and big enough for that :P 2023-04-20 06:23:25 don't get me wrong it's purely out of self interest 😎 2023-04-20 06:23:28 (just kidding) 2023-04-20 06:24:54 Haha, even if it were, a good thing done for the wrong reasons still got done, and that's what matters :P 2023-04-20 06:25:13 Who knows, maybe it will speed things up for lots of folks. That's the hope, right? 2023-04-20 06:26:07 magic 2023-04-20 06:27:09 Seriously, it seems like it is. All the testing I've seen indicates that it makes a huge difference. openwrt backported/enabled it late this past year due to how good it was shaping up 2023-04-20 06:27:35 yeah it's fantastic 2023-04-20 06:27:55 if you use linux-lts from edge you can just echo a value until it builds haha 2023-04-20 06:30:48 I wish I could :C I need linux-lts to match the rtl88xxx packages to keep internet, so no fun for me. Gottawait 2023-04-20 06:38:34 the -src ones? those just rebuild against the new one 2023-04-20 07:04:42 One is that, but seems to not rebuild, and the other is the rtl8821ce-lts 2023-04-20 07:05:20 sure, the -lts would have to match 2023-04-20 07:08:10 The src should work, even if the src is 5.x and I'm building against 6.x kernel from edge? If so, I need to break my laptop again and get logs for why it fails  2023-04-20 07:12:38 the src isn't 5.x, it's nothing 2023-04-20 07:12:45 well 2023-04-20 07:12:50 it could not build against new kernel or whatever 2023-04-20 07:15:55 fixed it https://gitlab.alpinelinux.org/alpine/aports/-/merge_requests/46240 2023-04-20 07:21:30 realtek is a pain :P 2023-04-20 14:11:49 You are awesome, thank you!  2023-04-20 14:27:53 Where do logs for modem-manager end up? I'm running it via openrc. 2023-04-20 14:28:40 if it's using syslog, /var/log/messages 2023-04-20 14:29:47 if it's not, maybe a file, or nowhere 2023-04-20 14:30:55 Apparently nowhere is the right answer, worked around it by just running it as 'doas ModemManager --debug' 2023-04-20 14:34:49 So it doesn't seem to detect my modem, but the same modem worked on another distro. Any idea what might be missing? 2023-04-20 14:34:58 what's in the logs 2023-04-20 14:35:25 if other distro is similar generic linux then it could be random config or missing packages or firmware etc 2023-04-20 14:35:34 if it was something mobile then they usually have quite a few patches to get stuff to work 2023-04-20 14:36:21 At least postmarketOS and thus Alpine doesn't patch modemmanager 2023-04-20 14:36:23 I tested on Arch and Asahi. 2023-04-20 14:36:59 definitely no meaningful patches 2023-04-20 14:37:03 post the logs i guess 2023-04-20 14:38:54 https://paste.sr.ht/~whynothugo/cba2682a410092ce9670a2a5be02b99019a12919 2023-04-20 14:39:35 I'll try a upgrade and reboot just in case. 2023-04-20 14:41:18 nothing jumps out at me 2023-04-20 14:44:10 Oh shit, it works now. It was a simple matter of "have you tried turning it off and on again". 2023-04-20 14:44:30 means that your kernel was upgraded and the modules could not be loaded 2023-04-20 14:44:56 I suspect so. I usually tie upgrades to reboots, but there might have been an exception. 2023-04-20 14:46:45 it's rolling 2023-04-20 14:48:05 the first thing i do after a kernel upgrade is ::REBOOT:: 2023-04-20 14:48:49 If you have the old kernel package cached, you can extract the modules from there as well 2023-04-20 14:49:13 do apk cache pkgs? 2023-04-20 14:49:18 not by default 2023-04-20 14:49:22 right 2023-04-20 14:49:23 but you can enable that 2023-04-20 14:49:56 uniqNuuid: I have a vague memory of needing to install something to urgently complete a task, and that required updatede dependencies so I ran a full upgrade but couldn't reboot at the time. 2023-04-20 14:50:02 It must have been during this power-cycle. 2023-04-20 14:50:02 so i set linux-edge=VER in /etc/apk/world 2023-04-20 14:50:29 you can pin the kernel so it does not get upgraded 2023-04-20 14:50:46 I do btrfs snapshots before upgrade, so they're usually safe -- as long as I reboot. 2023-04-20 14:50:50 and when you want to upgrade? just apk add linux-edge 2023-04-20 14:51:25 and pin it again with apk add linux-edge=NEW_VER 2023-04-20 14:51:44 most of the time, the kernel upgrade is non sense 2023-04-20 14:52:15 the upgrade probably has nothing to do with your hardware 2023-04-20 14:53:11 True, but the operation overhead of not having to pin anything makes life simpler. 2023-04-20 14:53:18 FWIW, would be nice to see https://gitlab.alpinelinux.org/alpine/aports/-/issues/13928 addressed. 2023-04-20 14:53:43 It's not easy to fix 2023-04-20 14:54:35 archlinux has the same issue, not something they have addressed yet either 2023-04-20 14:54:48 I just posted the link to solutions that work on arch. 2023-04-20 14:54:49 :P 2023-04-20 14:55:25 third-party solutions 2023-04-20 14:55:37 They're just triggers and hooks. 2023-04-20 14:55:49 One of them is in the official arch repos. 2023-04-20 14:56:12 it would be better if it was possible to keep at least 1 previous kernel package installed upon upgrade (as a fallback in case the new version has problems) 2023-04-20 14:56:17 ikke: https://github.com/archlinux-jerry/pkgbuilds/tree/master/kernel-modules-hook-hardlinks 2023-04-20 14:56:22 they have a hook 2023-04-20 14:57:40 minimal: you want the currently running kernel version kept 2023-04-20 14:57:44 previous does not mean a lot 2023-04-20 14:58:12 I think what minimal is saying is to allow two versions of the kernel to remain installed in case you need to boot into an old one. 2023-04-20 14:58:28 old one is useless... 2023-04-20 14:58:57 keep a piece of running kernel's module is nice 2023-04-20 14:59:10 anyway, pin is the best solution :P 2023-04-20 14:59:36 pin them all, pin the pkg that depends on a kernel version too. like zfs? 2023-04-20 14:59:41 My setup was designed with being able to roll back in mind. btrfs snapshots on upgrades, and a uefi bundle (initfs+kernel+cmdline) for older kernels that boot into their matching btrfs snapshot. Then UEFI stubs broke. 2023-04-20 14:59:44 ikke: that's what I meant by "previous", the one previous to the upgrade (in the case of not having rebooted yet then the "current" version, in the case where have rebooted then the "previous" running kernel) ;-) 2023-04-20 15:00:39 You could upgrade a kernel multiple times without rebooting, that's why previous is ambiguous :) 2023-04-20 15:01:14 WhyNotHugo: UEFI stubs didn't break, you were using the linux-edge kernel which stopped supporting that particular EFI stub (because of a newly introduced kernel config setting which could be changed) 2023-04-20 15:01:18 Right, keeping modules only needs to be done if the kernel being removed matches the currently running kernel. 2023-04-20 15:01:47 ikke: anyway, I'm talking about being able to have 2 versions of either linux-lts or linux-virt installed 2023-04-20 15:01:55 minimal: yeah, others pointed that out 2023-04-20 15:01:57 minimal: Yes, exactly that. 2023-04-20 15:03:05 You need some kind of 'back-up-files-and-restore-after-upgrade' mechanism 2023-04-20 15:03:32 BTW I saw that mps is giving up maintainership of all his packages, including linux-edge 2023-04-20 15:03:35 and also cleanup previous-1 2023-04-20 15:03:42 yes 2023-04-20 15:04:27 I'd guess this would be a good time to align the config of linux-edge with lts/virt 2023-04-20 15:06:05 minimal: merge them? that is good 2023-04-20 15:06:19 linux-edge has many good options enabled for aarch64 2023-04-20 15:08:45 minimal: where do you see that info, can you share the link? 2023-04-20 15:09:08 what info? 2023-04-20 15:09:11 that mps is giving up packages? 2023-04-20 15:09:15 He mentioned it in #-infra 2023-04-20 15:09:16 yeah 2023-04-20 15:09:42 :\ 2023-04-20 15:09:43 yes #alpine-infra yesterday 2023-04-20 15:10:29 mps wants to keep everythign minimal (meaning just what he needs) and feels frustrated because other people need other features 2023-04-20 15:12:19 mps made my amlogic s905d networking possbile by enabing one option 2023-04-20 15:13:18 death by a 1000 cuts 2023-04-20 15:13:51 but there are so many boxes out there, it is hard to meet everyone's expectation. 2023-04-20 15:14:10 alpine's linux-edge is the smallest working kernel in size 2023-04-20 15:14:24 i mean for my aarch64 hardware 2023-04-20 19:25:35 trying to get espeakup working in initramfs, I can't seem to get device nodes to be present? I first tried copying them with espeakup.files entries, then with mknod commands in /usr/share/mkinitfs/initramfs-init, but they don't seem to get created. In the init script if I do ls -l /dev I see only vga_arbiter and zero even though I'm pretty sure null, kmsg and ptmx should be present. I do "modprobe snd" in the init script but not all the 2023-04-20 19:25:35 related modules are loaded. Maybe I should be approaching this in a different way? 2023-04-20 19:28:00 unrznbl[m]: Any modules you need, should be enabled in /etc/mkinitfs/mkinitfs.conf 2023-04-20 19:29:10 ok, cool. What about device nodes? How should I deal with those? 2023-04-20 19:30:15 unrznbl[m]: they should be created when the kernel modules are loaded 2023-04-20 19:31:15 cool. will try it out! thanks! 2023-04-20 19:35:31 on my running system lsmod says 2023-04-20 19:35:32 and 2023-04-20 19:35:33 snd 90112 9 snd_hwdep,snd_opl3_lib,snd_timer,snd_mpu401_uart,snd_rawmidi,snd_seq_device,snd_es1938,snd_pcm 2023-04-20 19:35:33 soundcore 16384 1 snd 2023-04-20 19:35:39 do I need to include all of those or just 'snd' or 'soundcore' and it will resolve things? I have already added all of these to espeakup.modules and espeakup to the features="..." in mkinitfs.conf 2023-04-20 19:37:42 so right now I have features="ata base ide scsi usb virtio ext4 cryptsetup keymap espeakup" 2023-04-20 19:37:42 and espeakup.files and espeakup.modules and I have confirmed that the initramfs-lts has the bits I think I need. I did include the /dev nodes in the files listing, I guess I should take those out then. 2023-04-20 19:40:34 I think you need to include them all 2023-04-20 19:40:57 oh no 2023-04-20 19:41:00 it does it for you 2023-04-20 19:41:53 here? https://gitlab.alpinelinux.org/alpine/mkinitfs/-/blob/master/initramfs-init.in#L461-468 2023-04-20 19:42:52 https://gitlab.alpinelinux.org/alpine/mkinitfs/-/blob/master/mkinitfs.in#L68-L101 2023-04-20 19:43:28 indeed. nice! 2023-04-20 20:05:29 I added a big modprobe command to the init script and that's much better. I get info in /proc/asound/cards but when I try speaker-test I get "[ 12.288209] Playback open error: -2,No such file or directory" 2023-04-20 20:05:48 missing the whole alsa config files too i think 2023-04-20 20:06:22 I do see all the dev nodes I would expect... except one by-path entry: pci-0000:00:0b.0 -> ../controlC0 2023-04-20 20:06:41 right... of course :p amazing how many things need to be collected ;) 👍️ 2023-04-20 20:06:51 unrznbl[m]: you can also add the modules that need to be probed to the modules= kernel cmdline 2023-04-20 20:08:06 where? 2023-04-20 20:08:25 in your bootloader 2023-04-20 20:09:32 psykose: like /usr/share/alsa? I don't have anything in /etc/alsa. 2023-04-20 20:10:03 unrznbl[m]: the by-path links are normally created by eudev 2023-04-20 20:10:10 ikke: like /boot/extlinux.conf maybe, but this change is good "so-far". 2023-04-20 20:24:47 sweet. speaker-test is working now :) sadly I think it may never stop :O ha ha ha 2023-04-20 20:25:24 when it gets tiring just shoot the speakers 2023-04-20 20:26:10 excellent idea ;) I love this :O ;) 2023-04-20 20:31:33 so... I guess I'll need to take out the hard drive, luks decrypt it, and do some modifications then? no other way to get into the system right? Sounds relatively easy anyhow. Thanks for all the help! I think it was indeed /usr/share/alsa and modprobe -a (lots of modules) that fixed things up. 2023-04-20 20:32:06 or I could boot alpine on a usb stick, yeah. cool. 2023-04-21 12:09:02 So, how should I run daemons in WSL? 2023-04-21 12:09:51 Ermine: There is an option you can use to enable the use of an init system 2023-04-21 12:10:24 Ermine: I used https://wsl.dev/wsl2init/ 2023-04-21 12:10:26 as a guide 2023-04-21 12:10:36 which is specific for openrc 2023-04-21 12:11:53 ikke: thank you 2023-04-21 12:13:20 fyi, you don't need to do everything in that guide 2023-04-21 12:14:10 I think the configuration mentioned here should suffice: https://wsl.dev/wsl2init/#initial-configuration, though I forgot if I did other things 2023-04-21 12:14:20 (make sure openrc is installed ofcourse) 2023-04-21 14:09:40 Executing man-db-2.10.2-r3.trigger 2023-04-21 14:09:40 mandb: iconv_open ("UTF-8//IGNORE", "ISO-8859-1"): Invalid argument 2023-04-21 14:09:47 I run apk add man-db and get this 2023-04-21 14:10:52 you probably want mandoc instead 2023-04-21 14:20:28 thx, (I follow wiki and)it works 2023-04-21 14:55:29 qaqland[m]: that is due to lack of glibc 2023-04-21 14:55:49 there is no way to fix that. alpine linux does not have glibc 2023-04-21 14:56:33 psykose: but, there is a workaround 2023-04-21 14:56:40 mandb -q 2>/dev/null 2023-04-21 14:57:51 your analysis is wrong 2023-04-21 14:58:19 Can the desktop script install kde? 2023-04-21 14:58:50 open it and see :) 2023-04-21 14:59:19 iconv_open is a glibc function and mandb uses glibc for that. i digged it for a while, and that's my conclusion 2023-04-21 15:00:42 iconv_open is neither a glibc function nor does it 'use glibc' 2023-04-21 15:06:52 https://unix.stackexchange.com/questions/34133/why-man-cmd-shows-nothing 2023-04-21 15:07:39 lol 2023-04-21 15:11:14 lol what? it is annoying, and the fix is to redirect that error to /dev/null. 2023-04-21 15:11:40 not every one like to use makewhatis 2023-04-21 15:12:47 'lol' in that nothing on that page says anything you are claiming 2023-04-21 15:14:30 open("/usr/lib64/gconv/gconv-modules", O_RDONLY) = -1 ENOENT (No such file or directory) 2023-04-21 15:14:40 https://www.gnu.org/software/libc/manual/html_node/glibc-iconv-Implementation.html 2023-04-21 15:14:45 lmao 2023-04-21 15:16:11 i'll give you a hint, in that nothing on any of these pages are relevant to the actual issue 2023-04-21 15:16:48 i don't care actually. i've dropped any db 2023-04-21 15:17:03 i don't use man -k at all 2023-04-21 15:22:29 idk,but man -k works well when mandoc-apropos installed 2023-04-21 15:26:27 mandoc-apropos/makewhatis is slow because it recreate a db from zero when files in manpath changes. unlike mandb, it does that incrementally 2023-04-21 15:34:07 yep 2023-04-21 15:35:08 in my aarch64 case, it is 20 seconds! when apk add/del any -doc package. 2023-04-21 15:39:24 btw, the cold boot time on this box is 24 seconds. 2023-04-21 15:40:19 that is why i don't install mandoc-apropos. 2023-04-21 15:42:04 define "cold boot time" - including BIOS/UEFI initialisation? including bootloader starting? from turning on power to login prompt? ;-) 2023-04-21 15:42:45 minimal: ssh root@box ; ping mybox 2023-04-21 15:42:55 from noping to ping avalable 24seconds 2023-04-21 15:43:08 ssh root@box reboot 2023-04-21 15:43:39 reboot? that's not "cold" boot time then, that's a "warm" boot ;-) 2023-04-21 15:43:51 what? 2023-04-21 15:44:00 reboot= poweroff + poweron 2023-04-21 15:44:25 nope, reboot doesn't turn off the machine, it "resets" it 2023-04-21 15:45:08 unplug power cable. then plugin + ping at the same time, 24 seconds too 2023-04-21 15:45:16 Gotta love APIC/ACPI 2023-04-21 15:45:36 my arm64 box does not use ACPI 2023-04-21 15:46:28 also I don't think many people would agree that "able to ping" == "booted", as the machine will be able to ping once the interface is configured which typically occurs before a login prompt/sshd is available 2023-04-21 15:46:55 anyway it's not important, what i wanted to say is that makewhatis takes 20seconds to create db, it is close to a cold boot 2023-04-21 15:47:44 the moment i am able to ping is the moment i am able to ssh into it. 2023-04-21 15:48:16 well, not the case in openrc. it depends on how network was setup 2023-04-21 15:48:49 maybe on openrc, needs extra 4 seconds to ssh after ping available 2023-04-21 15:49:31 that was exactly my point - the networking service and the sshd service do not run at same time - there will be slow delay between them (however small) 2023-04-21 15:49:38 s/slow/some/ 2023-04-21 15:49:41 static ip is zero wait. dhcp is extra seconds. 2023-04-21 15:50:39 zero wait? depends on if you're counting in minutes, or seconds, or milliseconds ;-) 2023-04-21 15:51:37 just because you don't observe a delay between the networking service and the sshd service does not mean there is no delay (however small) 2023-04-21 15:52:50 minimal: it is about service dependency 2023-04-21 15:53:32 and parallel 2023-04-21 15:54:12 e.g. on runit, it is almost zero between ping avaiable and sshd available. 2023-04-21 15:54:45 "almost zero", so not zero then, so you're agreeing with me ;-) 2023-04-21 15:55:35 minimal: yes yes, i am honest. 2023-04-21 16:24:16 qaqland[m]: man-db hoook should be fixed on next upgrade for it 2023-04-21 16:46:44 psykose: thanks! finally, it is somehow related to glibc/musl, i was not correct but close to correct. 2023-04-21 16:47:11 not even in the same universe 2023-04-21 16:48:22 as a non programmer i can only guess based on comparison on musl based distro and glibc based 2023-04-21 17:00:46 psykose: thank you for your works!! 2023-04-21 19:32:49 I would like to install alpine on an odroid c4. Is any of the aarch64 images working with this board? 2023-04-21 19:39:21 engblom: generic arm should work I guess? 2023-04-21 19:49:34 i doubt it 2023-04-22 08:58:04 Hello there. I have a problem with 'doas' command. 2023-04-22 08:58:20 I want to run a command in the background as root. 2023-04-22 08:58:52 With 'sudo', I can use 'sudo --background '. It will ask for password, then runs in the background. 2023-04-22 08:59:37 I cannot do 'doas &' because it will ask for the password in the background and stuck there forever. 2023-04-22 09:29:42 livy: doas sh -c 'foo &' ? 2023-04-22 09:30:28 (or make a trivial wrapper that does this) 2023-04-22 12:30:28 doas "command &" maybe 2023-04-22 12:30:51 & outside of quotes will detach doas, & inside quotes will detach command 2023-04-22 12:30:57 I'm guessing 2023-04-22 12:31:15 and assuming doas invokes sh -c by itself automatically 2023-04-22 14:54:04 how am i expected to install simple-scan? in a way that is actually useful (as in, scanner actually works) 2023-04-22 14:54:19 it sadly seems to be packaged in a kind of useless state where no scanners work 2023-04-22 14:56:21 "doesn't work" 2023-04-22 14:59:06 lopid: it installs none of the sane packages so none are detectable 2023-04-22 14:59:18 basically the way it installs is in a way where it can just show you an error that no scanners were found 2023-04-22 14:59:32 yeah that's not uncommon 2023-04-22 14:59:55 right but it's a bit frustrating, i think it should either tell me in the UI what to do about it or just install all the scanner backends too 2023-04-22 15:00:10 technically you don't need any other packages to "use" this one, never mind that it's actually useless without one of those other packages 2023-04-22 15:00:16 i tried to install the right ones but it's like stabbing in the dark. i dont know which exact one i need, there seems to be no sane-backends-all or alike, so it's all pretty frustrating 2023-04-22 15:00:41 well i dont know what exactly i should be doing which is the sad part lol. like i assume i'm just missing something else but i cant figure out what it is 2023-04-22 15:00:52 so if anyone happens to know pls tell me. it's a canon lide 110 2023-04-22 15:01:07 on mobian it worked so on postmarketos/alpine it also should be able to 2023-04-22 15:01:29 is it at least in the wiki? 2023-04-22 15:02:23 also part of the problem is i keep finding more packages for sane with canon in them, i'm at like 3 already 😂it's very confusing 2023-04-22 15:03:05 maybe installing simple-scan should just install all the scanner stuff or at least suggest some sort of meta command/meta package (like sane-backends-all) that would do that, if i just want a sort of "just do whatever i need to be able to use the scanner i have" comman 2023-04-22 15:03:11 s/comman/command/ 2023-04-22 15:03:41 ACTION nods 2023-04-22 15:04:48 i even installed cupsd and saned and added both to auto launch and restarted lol 2023-04-22 15:04:53 so i tried all the random nonsense i could think o f 2023-04-22 15:04:59 s/o f/of/ 2023-04-22 15:05:05 (i assume i might need neither but who knows) 2023-04-22 15:05:16 i had the same problem on another distro recently 2023-04-22 15:06:11 apk should really have suggested packages, this would be a perfect use case 2023-04-22 15:06:35 like i get not everyone wants to just install all scanner drivers with simple-scan but it's kind of clunky that it doesn't suggest me an easy obvious way to do so either 2023-04-22 15:08:37 is it a printer, too? 2023-04-22 15:09:13 i ended up using it over wifi 2023-04-22 15:10:15 no just a usb scanner 2023-04-22 15:10:26 lopid: well it works on fedora so i blame alpine packaging 😛 2023-04-22 15:11:13 did you have to install any particular packages for it in fedora? 2023-04-22 15:19:30 I am asking one time more: is anyone having experience in running alpine on an odroid c4? 2023-04-22 15:21:45 engblom: didn't you get the answer for the first time? 2023-04-22 15:22:59 Ermine: You answered, but then psykose said he doubt. I want to know if someone has actually run it 2023-04-22 15:23:13 Ermine: Do you run it on c4? 2023-04-22 15:24:24 I wish to know if it is really working or not. 2023-04-22 15:24:46 No, I don't have this device, so I can just guess. 2023-04-22 15:25:36 All I find by searching the net is how people take ubuntu image and erase everything else than /lib/modules and /boot and then drop in alpine rootfs. But that would not give any kernel updates. 2023-04-22 15:26:17 And that is like a year old posts, so much can have happened since then. 2023-04-22 15:29:54 I can only suggest you trying out generic arm image yourself and see whether that works. 2023-04-22 15:30:46 engblom: https://wiki.odroid.com/odroid-c4/software/building_kernel implies you need a specific kernel 2023-04-22 15:34:06 pmaports only has hc2/xu4 references too 2023-04-22 15:34:19 so either it works by magic now (doubt) or you're out of the easy luck 2023-04-22 15:44:05 engblom: get the dtb, extract the info with dtc -I dtb, find all the `compatible` lines, find all the kernel configs to the compatible lines, search all the found configs against alpine linux-edge's config. 2023-04-22 15:44:33 that's how you know if it works for you or not. 2023-04-22 15:49:28 also, for alpine to work, you may need to create a file like the one in /etc/mkinitfs/feature.d/ and load the file in mkinitfs.conf and also add a kernel cmdline modules=a,b,c,d.... 2023-04-22 15:49:57 mostly mmc, usb, filesystem, ethernet,regulators 2023-04-22 16:04:04 whats making /tmp/tmp.R6aNwP like files? 2023-04-22 16:04:25 anything that calls the mktemp api either in shell or c 2023-04-22 16:04:29 that is like the default thing you get 2023-04-22 16:05:21 it's usually some one specific stray program that would be making them if you care (most things clean up), and some directory watch thing could probably tell you what it is 2023-04-22 16:09:43 thought it was one of my script and deleted them since they were filling up the diskless-ram ... 2023-04-22 17:11:59 is there a reason in particular why setup of networkmanager is required, and the config file recommended in the wiki doesn't ship by default? 2023-04-22 17:12:55 elagost: I am using alpine without networkmanager. Why do you need it? 2023-04-22 17:13:33 It's nice to have. Integrates with desktop environments and is very convenient for laptops. 2023-04-22 17:14:29 I don't suppose I need it, but I've gotten used to it with other distros and like using it. But that's beside the point. Installing networkmanager in alpine requires config before it's usable; just wondering why. 2023-04-22 17:23:21 there's no guarantee it would work with whatever ifupdown config at the same time 2023-04-22 17:23:38 and wifi would just be wrong 2023-04-22 17:23:45 that's the only thing in it, so idk 2023-04-22 21:27:31 hello 2023-04-22 21:28:27 I'm having an issue finding the root directory of firefox (flatpak). I don't know if it's due to alpine or flatpak, but firefox is reporting that the directory is at /home/user/.mozilla/firefox/* but it's not there. 2023-04-22 21:29:49 I'm talking about the root directory of the profile. 2023-04-22 21:30:44 you could do a search for a "firefox" dir 2023-04-22 21:31:28 How would I do that? 2023-04-22 21:31:29 flatpak dirs are 'mapped' so the path is not a real one 2023-04-22 21:31:35 I see 2023-04-22 21:32:01 How would I go about searching for the directory 2023-04-22 21:32:23 well, if it's not real there's no point 2023-04-22 21:32:58 find ~ -type d -name firefox 2023-04-22 21:33:04 it's somewhere in ~/.var/app/something/something/something 2023-04-22 21:33:14 probably 2023-04-22 21:34:40 I found it 2023-04-22 21:34:48 Thank you 2023-04-23 01:39:37 @Asmadeus I've tried your suggestion to use 'doas sh -c ' and it works. 2023-04-23 01:40:51 I mean 'doas sh -c &' 2023-04-23 01:41:09 However I have got a new problem: 'jobs' command no longer shows the status of that background command (probably because it spawns a subshell). 2023-04-23 01:41:38 By using 'sudo --background Is there anything I can do to control the status of the subshell processes spawned by 'doas sh -c'? 2023-04-23 02:02:33 livy: your commands exited probably, try 'doas sh -c "sleep 100" &' 2023-04-23 02:02:43 and run `jobs`, do you see it 2023-04-23 02:03:51 livy: by the way, put `&` outside of sh -c 2023-04-23 02:04:43 @uniqNuuid Your command is incorrect. It basically try to run 'sleep 100' as root using 'doas'. 2023-04-23 02:04:59 And the & put the 'doas' command to background. 2023-04-23 02:05:10 It ask for password in the background, and stuck there forever. 2023-04-23 02:08:39 The '&' symbol needs to be put inside the script, like this: doas sh -c "sleep 100 &" 2023-04-23 02:21:21 livy: it is not possible. 2023-04-23 02:21:44 maybe you want use stuff like tmux to do what you wanted 2023-04-23 02:27:51 Actually I used tmux already, with each session for a QEMU virtual machine. 2023-04-23 02:28:29 Inside each session, I need to run a 'qemu-system-x86_64' command and attach the QEMU monitor to the terminal. 2023-04-23 02:28:47 So that I can switch between multiple tmux sessions to control each VM. 2023-04-23 02:29:22 Each VM, however, may require additional background jobs such as virtiofsd daemons. 2023-04-23 02:29:53 And I need to control the status of these daemons, to make sure they no longer run after the VM is stopped. 2023-04-23 02:52:22 livy: maybe sh -c "commnad & echo $?" 2023-04-23 02:52:33 get the pid, so you can control the pid 2023-04-23 02:53:51 $? is return code 2023-04-23 02:53:57 $$ is pid 2023-04-23 02:54:28 oh right 2023-04-23 02:54:46 i mean $! 2023-04-23 02:56:11 sh -c "command & echo $!" 2023-04-23 03:01:47 Keeping track of the pid is a pain to deal with compared to a simple 'jobs' command. 2023-04-23 03:05:09 livy: but that is not possible to use jobs to control a backgrounded job in subshell 2023-04-23 03:15:58 I am thinking of using this trick: 2023-04-23 03:16:11 doas true; doas /usr/lib/qemu/virtiofsd --socket-path="${virtiofs0_sock}" -o source="${virtiofs0_source}" -o cache=always & 2023-04-23 03:16:45 The first command will ask for the password and caches it. 2023-04-23 03:17:17 Then the 2nd command can use the cached password and safely runs in the background. 2023-04-23 03:28:21 you can set nopass for /usr/lib/qemu/virtiofsd 2023-04-23 05:50:40 Mrs. Skilton gets her holes filled and satisified by Rush's Geddy Lee. Later, Geddy meets Jean Pierre Manikariza, her husband, and gets filled by his big black cock! Read all about it! https://pastebin.com/1ExdrDQA 2023-04-23 08:35:23 KDE Gear 23.04 is out! 2023-04-23 08:38:49 When will we keep up? 2023-04-23 09:23:39 when the time is right 2023-04-23 10:38:49 livy: for qemu as a use-case, you can just use qemu-system -serial mon:stdio and interact with the VM directly (stdio will default to serial console, ^a+c will switch to/from human monitor interface) 2023-04-23 10:40:24 (ah, didn't read all the way to virtiofsd, if you need to manage a separate daemon I guess caching credentials is probably just as good) 2023-04-23 12:30:29 i have a matrix stack running via docker on alpine but after doing update/upgrade caddy wouldn't serve http(s) anymore, even though its exposed (running for over a month). tried it with nginx, same issue, also turned off the fw, nothing. no config or record changes, just updates of the alpine host and the containers. another container that's exposed on a high port responds, 2023-04-23 12:30:31 generally no container/application seems to have an issue. the logs are pretty useless and im not sure where to look so i thought i ask for help or ideas 2023-04-23 12:30:55 oh and everything runs unprivileged 2023-04-23 12:32:28 cits1: What does it mean that it cannot serve https anymore? 2023-04-23 12:32:46 if you try to visit the site it's supposed to service it doesn't do it 2023-04-23 12:32:55 *supposed to serve 2023-04-23 12:33:47 the setup is pretty much a 1:1 copy of the alpine wiki if that's useful info 2023-04-23 12:34:20 It helps to be more specific 2023-04-23 12:34:39 "doesn't do it' is too vague 2023-04-23 12:35:38 oh, my bad. curl doesn't do it, telnet whines about connection closed 2023-04-23 12:35:51 What does curl say? 2023-04-23 12:35:56 nothing 2023-04-23 12:36:19 And what if you connect to it with browser? 2023-04-23 12:36:37 connection reset 2023-04-23 12:37:04 cits1: does `netstat -tlpn` show something listening on 443? 2023-04-23 12:38:06 0.0.0.0:443 0.0.0.0:* 2023-04-23 12:38:22 ufw is disabled 2023-04-23 12:38:22 Doesn't list a process? Or you are not executing it as root 2023-04-23 12:38:31 it doesn't run as root 2023-04-23 12:38:43 You need to run it on the host, not in a container 2023-04-23 12:38:57 ideally as root 2023-04-23 12:39:06 it worked like this before perfectly 2023-04-23 12:39:14 Just bear with us 2023-04-23 12:39:16 and i tried running nginx as root, same issue 2023-04-23 12:39:33 Not nginx 2023-04-23 12:39:36 the netstat command 2023-04-23 12:39:40 ah 2023-04-23 12:39:43 my bad :D 2023-04-23 12:40:22 https://pastebin.com/zTupCbnb 2023-04-23 12:40:38 ok, that's ok 2023-04-23 12:40:48 iptables -S INPUT 2023-04-23 12:41:06 or 2023-04-23 12:41:17 sorry, that's not relevant for docker 2023-04-23 12:41:27 iptables -S FORWARD 2023-04-23 12:41:34 https://pastebin.com/385iVzJF 2023-04-23 12:41:40 ah 2023-04-23 12:41:43 ufw 2023-04-23 12:41:51 but it's disabled 2023-04-23 12:41:57 can you show the FORWARD table as well? 2023-04-23 12:42:04 or rather chain 2023-04-23 12:42:22 so `iptables -S FORWARD` 2023-04-23 12:43:13 https://pastebin.com/B6TSjB04 2023-04-23 12:43:32 yeah, so ufw messed with the rules that docker adds 2023-04-23 12:44:05 yeah that's kind of intentional but i didn't expect it to break 2023-04-23 12:44:07 cits1: easiest way to fix is to restart the docker daemon 2023-04-23 12:44:15 cits1: it does 2023-04-23 12:44:19 i was going along with https://github.com/chaifeng/ufw-docker 2023-04-23 12:44:48 ikke: i tried that that didn't do it 2023-04-23 12:49:33 is there a better way of handling firewalling with docker? 2023-04-23 12:58:08 Hi. How I include extra firmwares with mkinitfs ? 2023-04-23 13:05:29 I tried to create a file in /etc/mkinitfs/features.d/ucode.files and put /lib/firmware/intel-ucode/* there, but it seems to not include anything 2023-04-23 13:10:07 did you add the ucode as a feature to /etc/mkinitfs/mkinitfs.conf? 2023-04-23 13:24:44 ikke yes 2023-04-23 13:25:10 did you regenerate the initfs? 2023-04-23 13:25:20 Seems it doesnt want to include the intel-ucode files. if I specify just /lib/firmware it includes lots of files, but not the intel-ucode 2023-04-23 14:14:54 I installed the intel-ucode package using "apk add intel-ucode", which added the /boot/intel-ucode.bin and added it to grub.cfg as "initrd /volume/root/boot/intel-ucode.img /volume/root/boot/initramfs-lts" but it seems it isn't loaded from there either. I have late microcode loading disabled in the kernel, perhaps this is why? 2023-04-23 14:34:28 Microcode updates must otherwise be enabled by adding /boot/amd-ucode.img or /boot/intel-ucode.img as the first initrd in the bootloader configuration file. This is before the normal initrd file. See below for instructions for common bootloaders. 2023-04-23 14:34:54 my grub file says, "initrd /boot/intel-ucode.img /boot/initramfs-lts" 2023-04-23 14:41:42 yea. here too. but "dmesg|grep microcode"only gives "[ 1.551256] microcode: Microcode Update Driver: v2.2." 2023-04-23 14:42:03 maybve BIOS already has the updated microcode for this CPU, so an update wasn't needed 2023-04-23 14:51:18 the cheapest (as in how much work) idea: git clone spectre-meltdown-checker somewhere and it might tell you if you miss out on a new firmware 2023-04-23 14:51:29 it's of limited use, but better than nothing 2023-04-23 15:43:14 hello from new Alpine Linux user.. where does one report issues with the testing branch? 2023-04-23 16:01:00 swb0z0: open an Issue on https://gitlab.alpinelinux.org/alpine/aports/issues 2023-04-23 17:14:30 uh, are you targeting some specific date for 3.18 release? 2023-04-23 17:45:32 lnl: officially May 1st, but whenever it's ready in practice 2023-04-23 20:19:46 hello, first-time alpine linux user.. running sway on edge, super+enter key not launching terminal.. do i have to create ~/.config/sway for that to work? 2023-04-23 20:21:21 swb0z0, do you have the foot terminal emulator installed? 2023-04-23 20:21:42 yes, which says /usr/bin/foot 2023-04-23 20:21:53 hmmm 2023-04-23 20:22:09 looked at /var/log/messages, no trace of sway.. does it log anywhere else? 2023-04-23 20:22:35 are you running sway from the command line? or a display manager? 2023-04-23 20:22:43 swb0z0: default sway config is in /etc. check the default emulator and default mod button 2023-04-23 20:23:40 also, does anything else work in sway? Like super+d, or switching workspaces... 2023-04-23 20:23:42 thx Ermine, will look in /etc.. launched sway from console via: dbus-run-session -- sway 2023-04-23 20:24:33 /etc/sway/config shows: 2023-04-23 20:24:34 set $term foot 2023-04-23 20:24:34 set $mod Mod4 2023-04-23 20:24:52 did u cpy the sway config to ur config folder? 2023-04-23 20:25:15 Misthios: no, i'll try that, thx 2023-04-23 20:25:32 the other possibility is foot itself doesn't launch 2023-04-23 20:25:47 for which to debug you would have to find a way to run it from.. another terminal i guess :D 2023-04-23 20:26:01 btw, how does one quit sway? i switched to /dev/tty2 to get another login console 2023-04-23 20:26:12 mod shift e 2023-04-23 20:27:20 psykose: ran ldd /usr/bin/foot to see if anything missing, all solibs check out 2023-04-23 20:29:16 hmm, Mod+Shift+e didn't quit sway but displayed a bar at top of screen, with placeholder chars? look like squares, not sure if it wasn't able to switch /dev/tty1 back to text mode 2023-04-23 20:30:08 did u install fonts? 2023-04-23 20:31:55 legendary missing font 2023-04-23 20:32:05 Misthios: shoot, i missed that step.. which fonts do i need? 2023-04-23 20:32:21 any 2023-04-23 20:32:27 ^ 2023-04-23 20:32:30 apk search font- 2023-04-23 20:32:38 and pick whatever you like i guess 2023-04-23 20:32:59 well not 'any', it's more like any regular ttf/otf/whatever font, which is 90% of them 2023-04-23 20:36:30 ok, installed font-noto pkg, will restart sway to see what happens, brb 2023-04-23 20:39:56 Misthios & psykose: 10x thank you!! :) installing fonts fixed it! i should write this up and post on wiki 2023-04-23 20:40:52 :) 2023-04-23 20:41:13 https://wiki.alpinelinux.org/wiki/Sway already says 'get a font' 2023-04-23 20:42:30 psykose: yeah, you're right.. i didn't see a single list of steps to get sway working from a base text-mode install 2023-04-23 20:43:53 another "easy" question.. do i create+edit ~/.sway/config to switch from sway to qtile? 2023-04-23 20:44:45 don't see why you would, it's something else entirely isn't it 2023-04-23 20:45:24 qtile is another compositor, so it should work, based on what i've read.. of course, i could be wrong ;) 2023-04-23 20:45:48 oops, i see... should be another config entirely, gotcha 2023-04-23 20:46:24 maybe dbus-run-session -- qtile start ? 2023-04-23 20:48:18 yeah, osmething like that 2023-04-23 20:49:00 pressing Mod+Shift+e now shows the proper text: You pressed the exit shortcut. Do you really want to exit sway? This will end your Wayland session. [Yes, exit sway] [X] 2023-04-23 20:49:48 what does xdg-open ytdl:// print on your systems? 2023-04-23 20:50:38 -ash: xdg-open: not found 2023-04-23 20:50:55 prolly missing yet another pkg :) 2023-04-23 20:51:26 in such cases, apk search cmd:command-you-are-missing 2023-04-23 20:51:46 xdg-tools 2023-04-23 20:51:47 ... will show the package you're looking for 2023-04-23 20:52:00 Ermine: thx again for the tip! :) 2023-04-23 20:52:34 oops, xdg-utils 2023-04-23 20:53:43 Newbyte: x-www-browser: not found :D 2023-04-23 20:53:48 Newbyte: output from xdg-open shows "not found" for all (www-browser, links2, elinks, lynx, w3m) 2023-04-23 20:53:52 and then opens it in firefox 2023-04-23 20:54:14 swb0z0, if you want a log from sway, I run it: 'exec dbus-run-session -- sway > .local/session.log 2>&1' 2023-04-23 20:54:32 hm, wonder what package I could be missing. I have yt-dlp 2023-04-23 20:54:39 psykose: hahaha.. you're right, guess i'm still learning the ropes on alpine linux :) 2023-04-23 20:54:47 for me it says the protocol isn't supported 2023-04-23 20:55:01 Newbyte: it's not related to having a package afaik 2023-04-23 20:55:05 somercet1: thx! 2023-04-23 20:55:13 it's some weird mime database thing and having a .desktop that handles ytdl:// 2023-04-23 20:55:56 the last thing that xdg-open printed was "xdg-open: no method available for opening 'ytdl://'" 2023-04-23 20:56:45 guess i'll install firefox, and then i should be set.. many, many thx to all of you for helping out 2023-04-23 20:56:51 psykose: what would configure this then? 2023-04-23 20:58:44 oh, regarding qtile.. i got: libqtile.utils.QtileError: No DISPLAY set.. sounds like it's trying to launch via Xorg instead of Wayland 2023-04-23 20:59:32 install xwayland 2023-04-23 20:59:54 no, it's just starting as the wrong mode 2023-04-23 21:00:01 hmm, but I thought Qtile was a Wayland compositor 2023-04-23 21:00:29 most def dont want to install Xorg, if i can avoid it 2023-04-23 21:00:51 wm for both x11 and wayland... that's kinky 2023-04-23 21:00:52 nothing against Xorg, but i want a stripped down Wayland install 2023-04-23 21:01:26 Ermine: hahaha.. well, hopefully i can get Qtile running as a compositor somehow 2023-04-23 21:04:56 hmm, wonder if firefox supports wayland or will require xwayland.. guess i'll find out soon 2023-04-23 21:05:10 it does 2023-04-23 21:06:10 MOZ_ENABLE_WAYLAND=1 2023-04-23 21:06:22 somercet1: thx! 2023-04-23 21:07:09 guess i can set+export MOZ_ENABLE_WAYLAND in ash before launching sway 2023-04-23 21:22:56 yup, I set it in ~/.profile 2023-04-23 22:35:12 https://www.youtube.com/watch?v=X4WZrpSPs3E 2023-04-23 22:35:14 oops 2023-04-23 23:43:59 apk uninstalled xf86 driver crocus, but libEGLS is still asking for it 2023-04-23 23:44:10 cannot restart sway 2023-04-23 23:55:27 mesa-dri-gallium was somehow uninstalled. :-( 2023-04-24 00:23:25 Hi, I have setup elogind and xorg on an Alpine Linux VM. Right after I run startx, I get org.freedesktop.DBus.Error.ServiceUnknown the name org.freedesktop.consolekit was not provided by any .service file. Any ideas? 2023-04-24 04:50:39 I'm trying to get par running, and it keeps erroring out. :-P 2023-04-24 14:00:29 ugh. Looks like Intel SNA video acceleration is broken on Edge. 2023-04-24 14:01:05 Xorg crashes and burns with ``sna_option_accel_blt:18146 assertion 'sna->kgem.gen < 0120' failed'' 2023-04-24 14:04:36 are you using xf86-video-intel by chance or a real driver 2023-04-24 14:06:17 answered that one myself 2023-04-24 14:06:25 yes, probably related to the new version then i guess 2023-04-24 14:07:07 Yeah, sorry, this system is laggy AF with UXA acceleration. (read: nearly unusable ) 2023-04-24 14:07:19 I'll pastebin some stuff in a minute 2023-04-24 14:07:24 uxa? 2023-04-24 14:07:44 The old acceleration option 2023-04-24 14:08:00 i've always just used the mesa drivers 2023-04-24 14:08:44 https://wiki.archlinux.org/title/intel_graphics#AccelMethod <-- I have to do this to get XOrg to even launch now. 2023-04-24 14:09:06 before it just worked, with no config whatsoever 2023-04-24 14:09:31 not great 2023-04-24 14:11:06 the only difference is the top 7 commits from here https://gitlab.freedesktop.org/xorg/driver/xf86-video-intel/-/commits/master 2023-04-24 14:14:40 which gpu do you have? 2023-04-24 14:16:17 whichever Iris Xe is embedded in the i5-1145G7 2023-04-24 14:16:45 lspci reports it as Intel Corporation TigerLake-LP GT2 [Iris Xe Graphics] (rev 01) 2023-04-24 14:18:05 you should definitely not be using this driver at all then instead of xf86-modesetting or whatever via mesa gallium 2023-04-24 14:29:15 Any suggestion on how to get Xorg to use the supposedly correct driver? with no config it wants the intel driver. 2023-04-24 14:30:04 Driver "modesetting" 2023-04-24 14:30:12 by default it does pick that pretty sure unless you added some config 2023-04-24 14:30:54 sweet, that worked. 2023-04-24 14:31:10 it seriously picks intel on this system by default. dunno what the deal there is 2023-04-24 14:31:17 hmm 2023-04-24 14:31:25 mind pasting the startup log of that 2023-04-24 14:32:07 sure, give me a moment. 2023-04-24 14:33:12 the default things don't ship any configs and afaik the internal default list wouldn't do that 2023-04-24 14:34:28 (i.e., having any .conf at all implies you would have put it there, which would be a reason, though i'm not 100% sure) 2023-04-24 14:34:48 that said, just fluff :) glad it works 2023-04-24 14:34:53 https://pastebin.com/jnyLtMwm 2023-04-24 14:34:53 or hope it does, maybe it's half broken :p 2023-04-24 14:34:56 test some stuff 2023-04-24 14:35:19 ah 2023-04-24 14:35:27 that loads both 2023-04-24 14:35:30 The only config I had before was to disable dontzap; I started it just then (for that log file) with no config whatsoever 2023-04-24 14:35:34 not having xf86-intel installed would have been fine 2023-04-24 14:35:48 heh. 2023-04-24 14:35:49 i guess the internal list picks it if available, but other than that it's whatever 2023-04-24 14:35:59 Let me try that and see if the results vary, just for shits and giggles. 2023-04-24 14:36:02 :) 2023-04-24 14:38:21 yeah, not having xf86-intel* installed and no config loaded just prints a warning about being unable to load intel; but it works just fine. 2023-04-24 14:38:40 yeah, usual meaningless warnings 2023-04-24 14:38:49 now see if glxgears works :p 2023-04-24 14:39:07 curious if anything improved at all 2023-04-24 14:39:14 I was going to launch Steam -> Fallout4 ;p 2023-04-24 14:39:49 insert generic new vegas was better comment here 2023-04-24 14:40:10 haha 2023-04-24 14:40:21 Performance is the same as before this fiasco. For the win 2023-04-24 14:40:27 30+ FPS on a shitty laptop is great. 2023-04-24 14:40:37 \o/ 2023-04-24 14:40:38 :) 2023-04-24 14:40:40 poggers 2023-04-24 14:41:06 appreciate the help, as always! 2023-04-24 14:42:02 ^_^ 2023-04-24 15:34:39 is there a date/time-format that most 'date -d' programs understand that specify both date and time without space in it? aaand thats human readable? :< 2023-04-24 15:37:58 don't think so 2023-04-24 15:38:34 if you mean to pass a -d date spec and get back something from that time 2023-04-24 15:38:42 date -d '2023-04-23T01:01:00' 2023-04-24 15:39:12 busybox rejects it 2023-04-24 15:41:02 hmm, strange there isnt a better standard imo .. i guess seconds since 1970 but not too human friendly 2023-04-24 15:42:29 the date command isn't exactly very posix specified hehe for this functionality 2023-04-24 15:43:51 posix doesnt even specify -d iinm 2023-04-24 15:44:55 surprised busybox date allows YYYYMMDDHHMM 2023-04-24 15:58:07 hello, first time Alpine Linux user.. greetd service fails to start: NoCompositorListening, src/app.rs:347:49.. does sway need to run as a service? 2023-04-24 16:26:26 got further by adding sway config for greetd.. now getting rpc error from seatd.c:66 - could not connect to socket /run/seatd.sock: Permission denied 2023-04-24 16:45:16 are you member of 'seat' group? 2023-04-24 16:46:04 and no, sway is not needed to run as service 2023-04-24 16:46:36 you can run it directly or with dbus-run-session, it only needs a properly XDG_RUNTIME_DIR env 2023-04-24 16:47:33 it's some greeter thing that runs sway for a greet screen 2023-04-24 16:47:40 don't remember which one or the config for it 2023-04-24 16:53:14 ah 2023-04-24 16:54:08 so add that user to 'seat' group? 2023-04-24 16:56:31 it is the seat user already 2023-04-24 16:56:45 the service is what starts it 2023-04-24 16:57:27 but the socket is root:seat 2023-04-24 16:58:41 ah, it's already in to seat group too 2023-04-24 23:06:46 In Alpine diskless how can I grow / a bit? 2023-04-24 23:07:57 I see that /run has ~180M available, so I would want to take some 100M from there. 2023-04-24 23:45:38 kinda nice script/alias ... alias l="ls -Ssr" 2023-04-25 05:50:08 hey, maybe someone know what is going on here? I try install symfony (alpine).... (full message at ) 2023-04-25 06:00:47 Cubixusin[m]: afaik, we have never provided symfony-cli 2023-04-25 07:40:25 ikke: that adds a 3rdparty repo with it 2023-04-25 08:32:58 psykose: I expected as much 2023-04-25 09:21:12 Can the standard version also be fully loaded in RAM like the Extended version? 2023-04-25 09:29:58 siger_revolution[m]: yes, it just ships with fewer packages (and misses ucode updates) 2023-04-25 09:32:50 (https://gitlab.alpinelinux.org/alpine/aports/-/blob/3.17-stable/scripts/mkimg.standard.sh#L42-61 contains the list of additional apks for the extended image) 2023-04-25 11:49:52 I have a patched version of imv, so what I've done is update the APKBUILD, set pkgrel=100 and I rebuild this and install it locally for use... 2023-04-25 11:50:09 ... The problem is that whenever I use `apk upgrade -a`, the package gets overwritten (as expected). 2023-04-25 11:50:23 What's a good workaround for this issue? 2023-04-25 11:50:43 The patch is: https://lists.sr.ht/~exec64/imv-devel/patches/39632. I expect it to be merged upstream some day, but that might be in many moons. 2023-04-25 12:06:25 keep the patched version in a local repository 2023-04-25 12:06:39 then it's just there and the newest version 2023-04-25 12:06:57 make it pkgver=999 instead so a real upgrade wouldn't replace it 2023-04-25 13:21:34 Hello! I’m trying to install Alpine on two separate disks (compact flash for boot and SATA hd, which is not available on boot, for the rest) 2023-04-25 13:35:10 psykose: I tried "apk index * > APKINDEX.tar.gz", but get "WARNING: No provider for the dependencies: so:libc.musl-x86_64.so.1 so:libcairo.so.2..." 2023-04-25 13:35:23 I guess I need to specify that my repo depends on main+community+testing..? 2023-04-25 13:35:27 nah 2023-04-25 13:35:33 apk index --no-warnings 2023-04-25 13:35:45 but also if you build stuff from a repo, abuild makes all that for you anyway 2023-04-25 13:36:11 if you make a ~/src/somegit/reponame/imv, then abuild it, you get ~/packages/reponame to add to /etc/apk/repositories 2023-04-25 13:36:44 alternatively just build it in aports, then ~/packages/community will keep it around 2023-04-25 13:36:51 nothing cleans the repo 2023-04-25 13:36:55 I keep the APKBUILD in my aports checkout, since I want to pull any updates that might happen to it. 2023-04-25 13:37:05 yeah, you could just keep it in a local branch 2023-04-25 13:37:13 then after you build it once it's what you'll have 2023-04-25 13:37:16 i do that for some stuff 2023-04-25 13:37:37 (assuming you put /home/you/packages/community in /etc/apk/repositories) 2023-04-25 13:40:23 Nice, yeah, that all makes sense. 2023-04-25 13:40:56 BTW, I use pkgrel 102 since I add 100 to the upstream version. Hopefully aports will never get to pkgrel=103 2023-04-25 13:42:06 just do pkgver=999 pkgrel=0 xD 2023-04-25 13:42:10 it will never be overtaken 2023-04-25 13:42:26 news tomorrow: imv 1000 has been released 2023-04-25 13:42:49 Nah, if imv is updated upstream I want to update my patched package. 2023-04-25 13:42:56 Althuogh hopefully the patch will make it into the next release. 2023-04-25 13:44:08 you would still update it :p 2023-04-25 13:44:10 doesn't stop ya 2023-04-25 13:46:24 What's the point of the split -pyc packages? 2023-04-25 13:59:02 It seems there is no boot param that lets me tell Alpine to use more than 50% of RAM for root? 2023-04-25 14:06:16 WhyNotHugo: allow people to opt-out from the python byte code files in case they care 2023-04-25 14:06:25 And some people do care 2023-04-25 14:46:53 Error loading shared library libjli.so: No such file or directory (needed by /usr/bin/java) 2023-04-25 14:47:05 It seems that there are some issues with OpenJDK. 2023-04-25 14:48:18 which openjdk 2023-04-25 14:59:04 openjdk17-jre-headless 2023-04-25 14:59:08 this one 2023-04-25 15:12:32 Trying to run 'smbpasswd' on an 'edge' based VM, and it's just segfaulting. Am I being stupid? 2023-04-25 15:32:49 Fxzxmic: works for me 2023-04-25 15:41:46 ikke: That mostly adds a tiny startup time but saves space, right? 2023-04-25 15:41:56 yes 2023-04-25 15:42:01 but the tiny startup time adds up 2023-04-25 15:42:36 But this gives people the choice 2023-04-25 15:49:22 it's something i would not pay for python scripts, i.e. you run python programs 2023-04-25 15:49:29 it's very noticable on non-desktop hardware :p 2023-04-25 15:49:35 but if it's long-running, it doesn't matter at all 2023-04-25 15:49:44 when you e.g. run a longrunning service in a container you don't care about this 2023-04-25 15:50:08 well, if your ops team cares about some request latency for cold requests maybe they will shoot you 2023-04-25 15:50:12 i ain't your lawyer 2023-04-25 15:51:30 I'm back on my bullshit, psykose. Hopefully I haven't gotten worse at MRs in the past year or so off. It'd be hard to have gotten worse, I think 2023-04-25 15:52:14 Oh, I totally get the point of having the 'pyc' files themselves. I just didn't think the space usage would be so substantial. 2023-04-25 15:52:51 it's roughly double, since it's the same thing per file 2023-04-25 15:52:57 https://gitlab.alpinelinux.org/alpine/aports/-/issues/11906 2023-04-25 15:53:44 not pictured: you run python in production but care about 20mb of disk 2023-04-25 15:53:47 :D 2023-04-25 15:54:34 well, "double" with default pyc, for some stuff in arch they also add -OO to you get pyc, opt-1 pyc, and opt-2 pyc, so it's quadruple the disk space 2023-04-25 15:55:27 ACTION wonder why you can't just run python with the 'pyc' files and skip the 'py' files altogether. 2023-04-25 15:55:36 psykose: strange 2023-04-25 15:55:46 WhyNotHugo: you can 2023-04-25 15:55:55 if you have .pyc and delete the .py it will work 2023-04-25 15:56:12 i forget what the limitations are but there are some 2023-04-25 15:56:59 Fxzxmic: do you have multiple javas installed or a JAVA_HOME set, etc 2023-04-25 15:58:18 no, installing openjdk for the first time. 2023-04-25 15:59:21 which architecture 2023-04-25 15:59:56 export LD_LIBRARY_PATH=/usr/lib/jvm/default-jvm/lib/ This works for me. 2023-04-25 16:00:08 x86_64 2023-04-25 16:04:10 readelf -d /usr/lib/jvm/java-17-openjdk/bin/java 2023-04-25 16:06:16 interestingly ldd (aka musl --list) doesn't seem to like $ORIGIN on a symlink 2023-04-25 16:06:42 i.e. ldd /usr/bin/java vs. ldd /usr/lib/jvm/default-jvm/bin/java 2023-04-25 16:07:20 sure, but loading java itself is fine 2023-04-25 16:07:24 as is running it 2023-04-25 16:07:50 https://imgloc.com/i/isu5EZ 2023-04-25 16:08:50 Fxzxmic: do you get the error when running ldd on java, or the executable itself? 2023-04-25 16:10:49 I am executing it. 2023-04-25 16:20:54 psykose: But Alpine doesn't let me opt out of the .py files, does it? 2023-04-25 16:21:00 indeed no 2023-04-25 16:21:04 That seems like the better approach to save space. 2023-04-25 16:26:58 sorry, internet was down 2023-04-25 16:28:32 it's not that simple or very possible 2023-04-25 16:32:45 So, that's it for now. 2023-04-25 16:33:26 time to bed. 2023-04-25 17:48:25 @psykose, thanks for the fast merge. That's only one part of what I'm after, but I need to spend time learning the build stuff for the various image formats before I open the other MRs 2023-04-25 17:48:43 what do those need 2023-04-25 17:48:44 L 2023-04-25 17:48:58 L 2023-04-25 17:49:08 + ratio 2023-04-25 17:50:57 🤣 2023-04-25 17:51:32 gdk-pixbuf build flags, if supported, if not I need to figure out how to package them  2023-04-25 17:52:47 This should enable them for Thunar, Ristretto, ShotWell, and anything else that understands gdk-pixbuf  2023-04-25 17:53:00 which formats 2023-04-25 17:53:05 just be specific and i'll look at it 2023-04-25 17:54:08 a lot of gdk-pixbuf stuff is external 2023-04-25 17:54:09 WebP, AVIF, JPEG-XL, HEIC/HEIF (basically what I targeted for the exiv2 update )  2023-04-25 17:54:13 I've not dug enough to find other pixbuf loaders we don't have but should  2023-04-25 17:54:29 Ahhh okay. 2023-04-25 17:54:56 webp/heif exist 2023-04-25 17:55:32 What package(s)?  2023-04-25 17:55:50 Does a meta for pixbuf loaders make sense or is that anti-Alpineness 2023-04-25 17:55:50 webp-pixbuf-loader/libheif-gdk-pixbuf 2023-04-25 17:56:23 Ahhh, awesome. Somehow I missed them.  2023-04-25 17:57:10 i'll set it up 2023-04-25 17:58:57 I don't pay you enough for all your help  2023-04-25 17:59:31 Thank you, though. This I think will be a really nice quality of life boost that hopefully nobody even notices 2023-04-25 17:59:38 That's the best kind  2023-04-25 18:07:24 and of course the jxl one requires a random git submodule from some google repo 2023-04-25 18:07:26 why wouldn't it 2023-04-25 18:12:57 😑 2023-04-25 18:13:41 And I thought pixbuf was supposed to make things easier and more standardized  2023-04-25 18:30:56 c81a156309e95f8c5f742ee12a967760b27f2582 2023-04-25 18:30:56 7740df70571e652f863b1c56f1e17da33d3c7a31 2023-04-25 18:31:01 9edc7d2c38556aa0d5fa47599ab2dec670ee8d85 2023-04-25 18:31:02 enjoy 2023-04-25 18:31:12 apk add gdk-pixbuf-loaders next release 2023-04-25 18:31:27 should get webp/jxl/avif/heif 2023-04-25 18:31:38 WhyNotHugo: i renamed the heif one, so you should update that 2023-04-25 18:33:57 hi. wondering if its a good idea to use alpine for bare metal servers. i enjoy its simplicity alot and it would be a good fit, but i've never used it for very long running stuff. 2023-04-25 18:45:04 "apk add gdk-pixbuf-loaders..." <- You are incredible! Thank you!  2023-04-25 18:47:27 Ooooh, those are commit hashes, not three accidently yubikeys! XD 2023-04-25 18:47:41 Right, that approach is perfect, because it leaves a clear path to "I just want thumbnails for everything". 2023-04-25 18:51:03 :) 2023-04-25 18:51:24 do you use the regular thumbnailers too or just ffmpegthumbnailer 2023-04-25 18:52:32 I use nemo, which uses _something_ that uses pixbuf. 2023-04-25 18:52:35 Not sure what's in the middle there. 2023-04-25 18:52:52 It actually works really well for thumbnails. 2023-04-25 18:53:14 it queries usr/share/thumbnailers/*.thumbnailer for MimeType=; lines 2023-04-25 18:53:30 each one specifies what generates the thumbnail and supported types 2023-04-25 18:53:38 Exec=/usr/bin/gdk-pixbuf-thumbnailer -s %s %u %o 2023-04-25 18:53:38 MimeType=image/avif; 2023-04-25 18:53:43 this is avif.thumbnailer, for instance 2023-04-25 18:54:00 heif uses.. its own tool instead of gdk-pixbuf-thumbnailer as the exec for some reason 2023-04-25 18:54:19 ffmpegthumbnailer has a file that has like everything in MimeType= ever, and wraps ffmpeg lib thinly to get the data 2023-04-25 18:55:13 I'm pretty sure that /usr/bin/gdk-pixbuf-thumbnailer can also do heif. 2023-04-25 18:55:13 i'm actually guessing all of that as i have no idea tbh :p 2023-04-25 18:55:18 yeah maybe 2023-04-25 18:55:22 but they ship a tiny bin for it 2023-04-25 18:55:31 FWIW, I don't have heif-thumbnailer in $PATH, but I do get thumbnails for those. 2023-04-25 18:56:12 Actually, that's a bug. I have /usr/share/thumbnailers/heif.thumbnailer, but no heif-thumbnailer in $PATH. 2023-04-25 18:56:18 something else might handle it, or nemo defaults to trying gdk-pixbuf as a fallback 2023-04-25 18:56:34 i already fixed that bug, 04348cd8baf6d76056343e05fbd455f5095cc760 2023-04-25 18:56:57 nemo defaults to trying gdk-pixbuf as a fallback <<-- based on figuring how to get this workiing about a year ago, i believe this is the case. 2023-04-25 18:57:20 what i wish i did know except all of this that i don't even use, is how the fuck mimetype fallback shit works 2023-04-25 18:57:20 Because I first had my own ".thumbnailer" file, and could eventually drop it since nemo worked without it. 2023-04-25 18:57:32 because every 3 months i get a fat 'click open in firefox, get .txt in winebrowser' 2023-04-25 18:58:39 WINEDLLOVERRIDES=winemenubuilder.exe=d ? 2023-04-25 18:59:16 sure, for that part, but how to set like anything and how 2023-04-25 18:59:27 Yeah, that's it: https://git.sr.ht/~whynothugo/dotfiles/commit/f6381accf0b10ec0adb87fdd86d3be876e2776b7 2023-04-25 18:59:40 i actually did forget that one so thanks :D 2023-04-25 19:00:13 i checked libheif for a second to see why they do their own loader 2023-04-25 19:00:17 https://git.sr.ht/~whynothugo/dotfiles/tree/main/item/home/.local/bin/xdg-make-default 2023-04-25 19:00:22 didn't find any reasons, did find like 10 bug reports for it though 2023-04-25 19:00:23 classic 2023-04-25 19:01:02 aha 2023-04-25 19:01:08 so it's just xdg-mime default handler type 2023-04-25 19:01:15 thanks 2023-04-25 19:01:35 Yeah, I use this wrapper for appliations that handle a bunch of types (e.g.: imv). 2023-04-25 19:02:11 i wonder if the app list is hardcoded of if it's XDG_DATA_DIRS/applications 2023-04-25 19:02:13 aep: I guess it depends on exactly what you want to use it for - Alpine does the job for me as a server 2023-04-25 19:03:17 psykose: The wut? 2023-04-25 19:03:37 you check two specific places for .desktop but it might be governed by anything in xdg dirs 2023-04-25 19:03:52 That's the usual order of presedence. 2023-04-25 19:04:07 Most appliations tend to check both in that order. Yeah, probably XDG_DATA_DIRS 2023-04-25 19:04:30 yee 2023-04-25 19:04:44 You can also override a system-wide one with this. 2023-04-25 19:04:58 Oh, /usr/local/share/applications doesn't usually work tho :( 2023-04-25 19:14:55 Trying apk upgrade gives me this set of errors, a lot of which are related to an upgrade from boost1.81 to boost1.82. What's the easiest way past this? http://sprunge.us/WO0xKW 2023-04-25 19:23:13 apk list -O 2023-04-25 19:26:43 psykose: http://sprunge.us/BLthNS 2023-04-25 19:26:55 "Right, that approach is perfect,..." <- And hopefully loading of everything in Ristretto and other image viewers and tools!  2023-04-25 19:28:34 I updated my workstation yesterday (a few hundread package updates) and now my cursor is tiny in firefox, foot and a few other programs. Qt applications OTOH now render a blurry cursor. 2023-04-25 19:28:39 bovis: apk del everything on that list 2023-04-25 19:32:31 Thanks 2023-04-25 19:35:36 WhyNotHugo: not quite sure which dependency that would be :D 2023-04-25 21:02:52 Hello. I'm looking for a way to use integrated/shared clipboard for my alpine guest VM using QEMU. I installed "spice-vdagent" and restarted the guest but the clipboard still isn't shared. Any advice? 2023-04-25 21:06:56 are packages reproducible builds? 2023-04-25 21:07:44 for example the package toxic, in aports i see it in aports it has a sha512 sum declared in the APKBUILD file. 2023-04-25 21:08:00 just wondering if things installed from package manger are reproducible builds are not 2023-04-25 21:08:18 the shasum is the shasum of the source files 2023-04-25 21:08:32 not of the binary output then? 2023-04-25 21:08:48 so im trusting the build process or maintaner then? how can i build packages from source? 2023-04-25 21:09:13 hey @anon4362, you can use git 2023-04-25 21:09:46 depedency trees, so for educational purposes how would i build toxic from source as an example? 2023-04-25 21:09:58 it is in aports and in package repo 2023-04-25 21:10:06 using postmarketOS 2023-04-25 21:10:07 is git not a solution? 2023-04-25 21:10:19 ha, massive dependency trees from multiple sources 2023-04-25 21:10:40 not as simple as git clone, make 2023-04-25 21:11:01 oh but that's what I'm saying, you can use git clone and make 2023-04-25 21:11:06 am i only one that builds entire dependency tree from source? 2023-04-25 21:11:15 what are aports? 2023-04-25 21:11:22 apologies if I'm missing something, I'm not a dev 2023-04-25 21:11:34 im familiar with debian new to alpine 2023-04-25 21:11:46 moving away from debian 2023-04-25 21:12:40 how would you build from source in debian 2023-04-25 21:12:41 i want to compile packages because i dont trust prebuilt binaries 2023-04-25 21:12:51 toxic from source in debian, it was a nightmare 2023-04-25 21:13:06 https://wiki.alpinelinux.org/wiki/Abuild_and_Helpers 2023-04-25 21:13:10 would PKGBUILD on arch be better 2023-04-25 21:13:10 apt-src all the dependencies and the dependencies depedencies, some things had to manually build from git with troublehsooting 2023-04-25 21:13:20 maybe you'd be interested in gentoo 2023-04-25 21:13:32 gentoo work on linux phone? 2023-04-25 21:13:37 true, alpine was inspired by gentoo after all 2023-04-25 21:13:46 why im going with postmarketOS, debian includes non-free firmware now and cannot get mobian to actually build 2023-04-25 21:14:05 I'm looking for a way to use integrated/shared clipboard for my alpine guest VM using QEMU. I installed "spice-vdagent" and restarted the guest but the clipboard still isn't shared. Any advice? 2023-04-25 21:14:20 im reading that link later thanks 2023-04-25 21:14:53 but as i understand, trust cannot be placed in prebuilt packages right? 2023-04-25 21:15:00 cant audit it came from source code unmodified 2023-04-25 21:15:51 in debian i used to install and build packages from source repo using apt-src, just wondering if was simliar way in alpine 2023-04-25 21:17:11 <|cos|> anon4362: debian and its derivatives are quite unique in including the full source of all its packages. 2023-04-25 21:18:08 <|cos|> anon4362: alpine, and "most other" dists, typically download src from all over. 2023-04-25 21:21:03 <|cos|> anon4362: what is your threat model? do you trust debian's content distribution network better than github? why? 2023-04-25 21:21:45 nothing to do with git vs debian 2023-04-25 21:22:11 my concern is supply chain security, and not placing trust in maintainer or build system of prebuilt binaries for sensitive applications 2023-04-25 21:22:21 applications used for security and encryptoin 2023-04-25 21:22:31 i think alpine hosts it packaged version of the what it builds? 2023-04-25 21:22:37 just simply want to build from source in a reasonable way 2023-04-25 21:23:01 open source is not open source if not from the source code 2023-04-25 21:23:12 only way can know from the source code is if reproducible build or if compile yourself 2023-04-25 21:23:32 userdocs_: quite a few distros host mirrors of distfiles but it's usually as a backup in case the real upstream goes away 2023-04-25 21:23:33 tor, toxic, gpg, i2p, very high targets 2023-04-25 21:23:39 of supply chain attacks 2023-04-25 21:24:49 abby: i guess most of the ones i looked at had dist files but maybe for more prominent packages, but not the first one looked at here https://git.alpinelinux.org/aports/tree/community/lua-brieflz/APKBUILD#n12 2023-04-25 21:30:14 dropping off of here. I should say.... postmarketOS has firefox as the default browser in their releases, if alpine does also then that should be reconsidered. mozilla firefox is spyware 2023-04-25 21:30:59 persistant UIDs that share your traffic history with 3rd parties, run "watch netstat -tunp" and launch firefox, see what happens. and packet captures. 2023-04-25 21:31:08 and be in for a suprise 2023-04-25 21:31:21 <|cos|> anon4362: my recommendation is that you read the links posted to you, and reflect on questions people ask you. 2023-04-25 21:31:33 yes, 2023-04-25 21:31:34 thanks 2023-04-25 21:31:40 reading later, just have to go now 2023-04-25 21:46:31 There is no default browser in Alpine, you need to pick what you want when you build up your environment  2023-04-25 21:47:12 I'm looking for a way to use integrated/shared clipboard for my alpine guest VM using QEMU. I installed "spice-vdagent" and restarted the guest but the clipboard still isn't shared. Any advice? 2023-04-25 21:53:18 abby: why put spyware my pc 2023-04-25 21:54:18 to attack your supply chain 2023-04-25 21:58:35 <|cos|> uncannily: i think you might need post-it notes and a good pen! either that, or expand on what's running the vm. qemu seems vague? 2023-04-25 22:02:01 apologies 2023-04-25 22:02:33 Virtual Manager, Fedora host. 2023-04-25 22:02:41 libvirt 2023-04-25 22:03:08 Gnome 2023-04-25 22:03:19 if you need more info let me know 2023-04-25 22:05:50 cos: any ideas? 2023-04-25 22:06:30 <|cos|> uncannily: that's good added! unfortunately i have no clue. just read the man page for spice-vdagent. did you attempt to start it with `-dd`? 2023-04-25 22:08:03 I found this: https://wiki.alpinelinux.org/wiki/How_to_run_Spice_Agent 2023-04-25 22:08:10 I'm gonna go through it. 2023-04-25 22:08:22 No, I started it with rc-service 2023-04-25 22:09:58 <|cos|> https://serverfault.com/q/874309/virt-manager-spice-copy-paste-doesnt-work might be a lead too, perhaps. 2023-04-25 22:10:28 <|cos|> i'm off to bed now though. good luck! 2023-04-25 22:10:58 okay thanks 2023-04-25 22:11:36 installing spice-vdagent and enabling it doesn't do anything 2023-04-25 22:13:00 <|cos|> have you tried turning it on and off again? that stackexchange says: "I had to reboot after installation of spice-vdagent but worked like a charm. Thanks." 2023-04-25 22:14:22 <|cos|> apart from that; looking at the xml definition, trying with another spice client, attempting to see logs on both sides? 2023-04-25 22:14:32 <|cos|> now i'm really off to bed. 2023-04-25 22:34:29 Yes I rebooted several times. 2023-04-25 22:34:43 I gave up. 2023-04-25 22:34:56 I'll revisit this another time. 2023-04-25 22:35:09 a shared clipboard is less secure anyway, and I'm using alpine for security 2023-04-25 22:35:18 goodnight 2023-04-26 09:22:43 psykose: gdk-pixbuf-loaders pulls in heif-thumbnailer instead of libheif-gdk-pixbuf? 2023-04-26 15:17:13 WhyNotHugo: both 2023-04-26 16:32:51 hey ^^ what is the best way to install an old nodejs version on the latest OS version? It sounds like I have to install the unofficial musl build, since the package versions are locked to the alpine version. I could also use a legacy branch of Alpine, but then everything is old and the risk is too high, because the node version I need is ancient ... 2023-04-26 16:33:27 "It sounds like I have to install the unofficial musl build" <- this one 2023-04-26 16:33:54 okay I guessed right, so manual installation is the way to go then. 2023-04-26 16:34:01 thx 2023-04-26 16:34:05 or create a container for it? 2023-04-26 16:34:49 minimal: wdym? It's already an LXC container, but I obviously don't like to run an ancient Alpine version. 2023-04-26 16:35:17 Docker? No thanks, I want to keep it as minimal as possible 2023-04-26 16:35:27 mbx_: assuming it's for use by an app I meant create a docker container for the app along with the required node version 2023-04-26 16:36:16 mbx_: you want to keep it as minimal [sic] as possible by running an OS container (i.e. including init daemon etc) rather than an application container 2023-04-26 16:37:45 minimal: speaking of best practices, this is true, but reality looks different. tight due dates, app was build with persistent storage in mind, I need to get this thing out and hope I can ship it on Alpine for the smallest footprint. Fixing the app for a better model? I will port it to the latest nodejs LTS anytime soon, but right now shipping it has highest priority. ;) 2023-04-26 17:32:12 I see that Alpine 3.15 is still maintained until autumn this year, so will it still receive security patches? 2023-04-26 17:34:07 for things that are known in the main/ branch sure 2023-04-26 17:34:58 oh dang, how long is support for the community branch? 2023-04-26 17:35:24 one release 2023-04-26 17:36:01 okay, very tight, but I get that it makes sense for a distro with that contribution size 2023-04-26 17:36:56 if anything it's too long 2023-04-26 17:37:26 okay, I see that is no problem though, main branch has the essentials that are required for security, so everythings fine, I guess I have a couple of months to bring this app up2date then. 2023-04-26 17:37:40 OS security* 2023-04-26 22:39:57 anyone running alpine on a Le Potato? https://libre.computer/products/aml-s905x-cc/ 2023-04-27 09:57:58 Hello. I seem to have issues pulling the "i686-mingw-w64-binutils" package. Says package not found 2023-04-27 09:58:11 https://pkgs.alpinelinux.org/package/edge/community/x86_64/i686-mingw-w64-binutils 2023-04-27 09:58:33 Running 3.17 2023-04-27 10:01:41 Ah, my bad, had to specifically refer to http://dl-cdn.alpinelinux.org/alpine/edge/community/ repository 2023-04-27 10:25:18 ImQ009: don't mix 3.17 with edge 2023-04-27 10:26:01 donoban, well, I need just that one package 2023-04-27 10:26:15 ImQ009: does not matter 2023-04-27 10:26:19 dependencies can be different 2023-04-27 10:26:25 Right 2023-04-27 10:27:07 Is the testing repo available on edge branch by default as well? 2023-04-27 10:27:30 it's not enabled by default 2023-04-27 10:41:27 Ouch 2023-04-27 10:41:37 Edge uses musl, that seems to be a deal breaker for me 2023-04-27 10:42:00 as well as other branches 2023-04-27 10:42:04 ImQ009: Alpine has been using musl on all releases since 3.0 2023-04-27 10:42:50 (And a different non-glibc libc before thaat) 2023-04-27 10:43:03 I see 2023-04-27 10:43:10 uclibc namely 2023-04-27 10:44:21 That's kind a nuisance, but let's see if I can manage to install glibc manually 2023-04-27 10:44:40 that's not how it works 2023-04-27 10:44:55 Just use glibc distro 2023-04-27 10:45:53 The wiki mentioned that it can be done? 2023-04-27 10:46:06 https://wiki.alpinelinux.org/wiki/Running_glibc_programs 2023-04-27 10:46:24 gcompat is not glibc, but it can work 2023-04-27 10:46:40 (it shims the glibc specific thigns) 2023-04-27 10:48:23 ImQ009: https://git.adelielinux.org/adelie/gcompat 2023-04-27 10:49:43 Yeah, I see it. Thanks a bunch 2023-04-27 14:39:49 hi. what's wrong with apk? it lists a few packages as upgradeable, yet it doesn't upgrade them. https://ttm.sh/WUv.txt 2023-04-27 14:40:40 i feel like i asked this long ago before, but can't find it 2023-04-27 14:45:34 I think they are pegged to a specific version 2023-04-27 14:46:17 it does the same here, "list" shows my pegged pkgs 2023-04-27 14:48:00 Piraty: what does `apk add fwupd=1.8.14-r2` do? 2023-04-27 14:59:09 On Alpine diskless, how would I "correctly" switch the lts kernel? 2023-04-27 14:59:23 to* 2023-04-27 14:59:27 Update the boot media 2023-04-27 14:59:35 there is update-kernel helper as well 2023-04-27 15:02:05 ikke: it says https://0x0.st/HKYS.txt 2023-04-27 16:17:55 weird, after purging my local repo it works, $HOME/packages is mentioned in /etc/apk/repositories 2023-04-27 16:18:14 oh no it doesn't 2023-04-27 16:18:37 fooled by my wrapper showing upgradeable before `apk upgrade` 2023-04-27 16:50:23 Piraty: you have a bunch of old fwupd-plugin packages that don't exist anymore 2023-04-27 16:50:31 it's just fwupd now 2023-04-27 18:10:30 is there some video-conference server in repo? 2023-04-27 18:10:40 for video chats 2023-04-27 18:20:15 jitsi-meet is in testing 2023-04-27 18:22:40 ikke: ty! 2023-04-27 18:32:33 grep '\bfoo\b' works for whole word, how do you use \w ? isnt that supposed to have the same function? 2023-04-27 18:32:57 psykose: The gdk-pixbuf-loaders are incredible, thank you. Such a huge UX boost 2023-04-27 18:33:50 Another big one would be taking my libopenraw aport from testing into community and building/bundling its pixbuf loader 2023-04-27 18:33:54 ahh \w is just for any word.. nvm.. 2023-04-27 18:34:03 Then pretty much every Camera RAW format would be supported 2023-04-27 18:34:12 frag: \w is [a-zA-Z] 2023-04-27 18:34:43 Oh sorry ' [a-zA-Z0-9_].' 2023-04-27 18:34:52 [a-zA-Z0-9_] 2023-04-27 18:36:14 ikke: oh right .. 2023-04-27 18:37:12 \d is digit but does not work with grep? 2023-04-27 18:37:21 regex is confusing :S 2023-04-27 18:38:21 frag: grep -E 2023-04-27 18:39:03 What most people consider regular expressions is PCRE (perl compattible regular expressions) 2023-04-27 18:39:09 echo foo bar|grep -E '\d' .. i get : grep: warning: stray \ before d 2023-04-27 18:39:15 grep by default uses basic regular expressions 2023-04-27 18:39:22 (BRE) 2023-04-27 18:39:31 grep -E is extended regular expressions (ERE) 2023-04-27 18:40:19 frag: works for me 2023-04-27 18:40:23 at least, no warning 2023-04-27 18:41:10 ah, gnu grep gives that warning 2023-04-27 18:41:38 frag: use grep -P to get PCRE 2023-04-27 18:42:31 ERE does not support \d either 2023-04-27 18:46:44 see what i mean :> 2023-04-27 18:47:02 ikke: which grep do you use? 2023-04-27 18:47:20 wish busybox grep supported colors 2023-04-27 18:47:28 or something similar 2023-04-27 18:47:38 bb grep would not complain, but \d would not work as expected either 2023-04-27 18:48:09 for bb grep, you'd have to use [0-9] instead 2023-04-27 18:48:13 (grep -E) 2023-04-27 18:49:41 >:/ 2023-04-27 18:50:02 frag: the nice thing about standards is that there are so many of them 2023-04-27 18:50:35 hehe seems so 2023-04-27 18:50:42 Saijin_Naib: already is 2023-04-27 18:57:15 do we have such an application in repo? https://vid.pr0gramm.com/2023/04/26/ea1463be37a89e80.mp4 2023-04-27 18:59:14 it says the name of the app ...sorry guys something im a bit lazy :> 2023-04-27 19:12:40 😱 2023-04-27 19:13:08 No wonder Thunar was thumbnailing my RAFs 2023-04-27 20:59:07 printf '1\t2\t34343434343436\tfoo\n' # if the cursor where at the end of the line and you move it to the left past o o f , how would you calculate how far the cursor should jump to with '6'? 2023-04-27 20:59:42 s/with/hit/ 2023-04-27 21:01:25 not so easy to calculate how much space a string with tabs take up eh(?) 2023-04-27 21:02:52 Isn't the end of a tab always a modules of the tab width? 2023-04-27 21:03:24 But you have to take the whole string into account 2023-04-27 21:03:32 you can, as you imply, not start from the right 2023-04-27 21:04:03 mhm 2023-04-27 21:13:07 seems wc -L can calculate width with expanded tabs 2023-04-27 21:24:40 MAGIC 2023-04-28 04:53:26 hewwo 2023-04-28 04:54:01 can you install glibc apps using nix on alpine? 2023-04-28 04:54:12 like vscode? 2023-04-28 05:19:07 monoid[m]: I don't know nix that well, but if it creates a completely self-contained set of libraries, it might work 2023-04-28 06:08:14 linux-edge 6.3 on arm64 : cacheinfo: Unable to detect cache hierarchy for CPU 0 2023-04-28 06:08:39 getconf -a | grep -i cache : outputs nothing 2023-04-28 06:08:58 hope this is a bug from 6.3 2023-04-28 08:36:52 Hello, is there a reason why ash forgets about what is in /etc/profile when running it as non login shell ? 2023-04-28 08:37:17 It's like it's actively unsetting env 2023-04-28 08:38:04 that is usually part of what non-login means 2023-04-28 08:40:14 Non login means that it won't source profile, provided that it will inherit from what launches it 2023-04-28 08:41:07 yes? 2023-04-28 08:42:34 But it doesn't inherit anything 2023-04-28 08:43:04 even exported variables? 2023-04-28 08:43:35 How do you execute this non-login shell? 2023-04-28 08:44:14 Oh right, just export the thing, that workes 2023-04-28 08:45:56 The file /etc/profile.d/color_prompt.sh shipped with alpine linux don't export anything to I was struggling to get my color prompt on non login shells 2023-04-28 08:47:40 you'll have to source login files yourself if you're not logging in 2023-04-28 08:50:14 prompt vars generally have no need to be exported; they're just pollution 2023-04-28 08:53:01 That's subjective. The file isn't named with ".disabled" at the end by default if you don't like prompt vars 2023-04-28 08:54:13 But for people willing to have colorful prompt there is no reason not to export prompt variable 2023-04-28 08:54:37 consider that most executions of the shell don't involve a prompt 2023-04-28 08:57:17 Of course. But it does sometimes 2023-04-28 09:01:07 nothing else uses the var besides the shell 2023-04-28 09:01:26 there's no need for the var to be set anywhere besides the shell 2023-04-28 09:05:20 Yes. But there is no other way than to export PS1 to have color prompt without sourcing manually the profile file on non login shells 2023-04-28 09:14:13 Bane of my life. That behaviour is different between various shells, and various OS'es. 2023-04-28 09:15:05 you can start ash with ENV set to a suitable file that contains whatever you want 2023-04-28 09:15:11 (in my case it is to switch *off* anything color, and make sure the current hostname is in the prompt) 2023-04-28 09:15:31 and ash will source this on startup, regardless of whether it's started with -l or not 2023-04-28 09:20:19 That mean exporting ENV instead of PS1. This isn't really better. 2023-04-28 09:21:26 it is, because ENV is used by lots of shells (it's POSIX), and the file can contain a lot more than just PS1 2023-04-28 09:24:04 Interresing 2023-04-28 09:24:26 Thanks for the tip 2023-04-28 09:48:24 i hate to upgrade kernel. the so called `stable` kernel is not stable at all 2023-04-28 09:48:40 now i don't have CPU level1/2 cache 2023-04-28 09:49:03 wat 2023-04-28 09:49:21 if you didn't have level1/level2 cache, your CPU would not work 2023-04-28 09:51:02 getconf -a | grep -i cache 2023-04-28 09:51:08 i don't see anything 2023-04-28 09:51:22 just because getconf does or does not return something doesn't mean it doesn't work 2023-04-28 09:51:29 dmesg : cacheinfo: Unable to detect cache hierarchy for CPU 0/1/2/3 2023-04-28 09:54:02 <_Mystified> if using a alpine-linux OS system on a 64bit PC, what would the most logical choice for a FS for DLNA server PC. 2023-04-28 09:54:21 also no such dir : /sys/devices/system/cpu/cpu0/cache/ 2023-04-28 09:54:29 /sys/devices/system/cpu/cpu0 is there though 2023-04-28 09:54:43 <_Mystified> most users will be using linux & android, one user will be using ios& macos 2023-04-28 09:54:54 <_Mystified> Thanks guys. 2023-04-28 09:57:05 _Mystified: ext4 2023-04-28 09:57:21 <_Mystified> Thanks guys. 2023-04-28 10:48:43 Anyone using alpine with samba as a domainserver? 2023-04-28 11:46:06 and now with firefox+wayland+intel the hw video decoder appears to crash, with the video window turning pink for a second and leaving behind gobs of zombie 'RDD Process'. happens rather occasionally on random h264 (i think) video, e.g. youtube 2023-04-28 12:11:24 uuidNuni1: "the so called `stable` kernel is not stable at all" - which kernel is called "stable" where? you're using linux-edge on edge which is not declared stable anywhere, secondly you're using the "edge" repos which are work-in-progress repos rather than a release, and 3rd you're using linux-edge rather than linux-lts, where linux-lts would be more likely considered "stable" than linux-edge as it is an LTS release 2023-04-28 12:15:27 I'm not sure that I fully understand how package installation works in diskless mode. 2023-04-28 12:15:32 Aren't all installed packages lost on reboot? 2023-04-28 12:15:41 How do I make "persistent pacakge installations"? 2023-04-28 12:19:32 using "lbu commit" 2023-04-28 12:22:11 together with apkcache 2023-04-28 13:52:21 If I would want to build the Generic Arm aarch64 image, but with edge instead of 3.17.3 and with linux-edge instead of linux-lts, how should I do? 2023-04-28 13:53:19 It looks like this combination might work on my odroid c4, as c4 support is added in linux-edge. 2023-04-28 14:01:29 engblom: create a "disk image" file, partition it and create filesystem(s), use loop0 to mount it, use apk.static to create a base install in a "root" directory mounted via loop, chroot into it and so the rest of the OS install and config. 2023-04-28 14:01:48 then write the resultant disk image to the odroid 2023-04-28 14:35:47 michael5050[m]: But lbu commit only saves /etc/, not installed packages, right? 2023-04-28 14:36:10 * sorry, this was meant for minimal 2023-04-28 14:40:44 WhyNotHugo: apkcache stores packages on boot media 2023-04-28 14:41:00 and you use lbu to store setting on boot media 2023-04-28 14:45:34 The list of installed packages is stored in /etc/apk/world, so is part of the lbu backup 2023-04-28 14:46:14 ikke: yes but also need apkcache to have the actual packages as well to reinstall on boot 2023-04-28 14:49:18 minimal: But that would require me having a working aarch64 system.... 2023-04-28 14:49:29 Otherwise it is not possible to chroot into it. 2023-04-28 14:49:57 engblom: nope, use binfmt together with qemu-system-aarch64 - that's how I create Alpine arm disk images on x86_64 2023-04-28 14:51:49 minimal: How do you use that? 2023-04-28 14:52:19 engblom: you install both the qemu aarch64 package and the binfmt 2023-04-28 14:52:24 package 2023-04-28 14:52:59 Or qemu-openrc 2023-04-28 14:53:05 then qemu-user-aarch64 (sorry user, not system) is run automatically when aarch64 binaries are executed 2023-04-28 14:59:11 There seems to be no package named binfmt 2023-04-28 14:59:26 Is it the same as the fmt package? 2023-04-28 15:01:09 as ikkie mentioned the qemu-openrc package comes with a qemu-binfmt init.d script 2023-04-28 15:02:09 you'd need to install this plus the qemu-aarch64 package 2023-04-28 15:02:33 and run the /etc/init.d/qemu-binfmt script to set up the binfmt mapping 2023-04-28 15:03:11 minimal: Got it! 2023-04-28 15:52:49 Did anybody experience issues with pass bash completion? 2023-04-28 16:12:48 no 2023-04-28 16:13:28 Well, it's not working for me 2023-04-28 16:15:47 sad to hear 2023-04-28 16:21:41 minimal: `stable` from kernel.org 2023-04-28 16:21:57 i did not complain about alpine, i am talking about kernerl.org 2023-04-28 16:23:58 uuidNuni1: ok, but the Alpine config for it could also be a factor... 2023-04-28 16:36:49 alpine linux-lts does not have the config my hardware needed 2023-04-28 16:41:40 uuidNuni1: you mean the *code* your hardware needs, rather than then *config*? If it is a config issue then raise a Git Issue to have the relevant config enabled 2023-04-28 16:53:44 minimal: the entire CONFIG_ARCH_MESON were not enabled at all, that is the config for amlogic SoC 2023-04-28 16:54:20 starting from linux 4.19, amlogic is well supported actually 2023-04-28 16:55:26 uuidNuni1: so raise a Git Issue to have that enabled then... 2023-04-28 16:56:19 i don't think it is possible. edge lts have different purpose, a merge would be nice. 2023-04-28 16:57:22 you don't think it is possible for you to raise a Issue to enable Meson related config in linux-lts? wow! 2023-04-28 17:00:10 i don't think they would enable it, it's not just one option, it's hundreds 2023-04-28 17:00:39 well if you don't ask then it's is unlikely to ever be enabled 2023-04-28 17:15:36 ok. i found it : https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?h=2e4a4e36285fa9739cb45a26206c958037d6b773 2023-04-28 17:24:04 so seems like it's a DeviceTree related issue for your SBC 2023-04-28 17:49:08 i fixed it : https://0x0.st/HKk6.txt 2023-04-28 17:49:16 added these to my dtb. 2023-04-28 17:50:46 minimal: So in diskless mode packages are all reinstalled from cache at startup? 2023-04-28 17:50:50 now cache shows: https://0x0.st/HKkl.txt 2023-04-28 17:50:56 Doesn't that make boot very slow? 2023-04-28 17:51:20 but musl's getconf does not report it, never mind, musl needs to fix getconf. dmesg warning fixed too 2023-04-28 17:51:59 uuidNuni1: so the problem was using a kernel before the device-specific DTB was updated to handle kernel change in behaviour 2023-04-28 17:52:16 minimal: no, i use dtbo. 2023-04-28 17:52:20 is the DTB part of upstream kernel or separate? 2023-04-28 17:52:43 no what? 2023-04-28 17:52:43 it was due to the upstream dtbs was not updated to work with kernel 6.3 2023-04-28 17:52:51 dtb overlay 2023-04-28 17:53:33 WhyNotHugo: doesn't what make boot very slow? 2023-04-28 17:53:59 Reinstalling all packages from cache. 2023-04-28 17:54:42 minimal: this is dtb overlay, it will load these on top of /boot/dtbs/module/xxx.dtb : https://0x0.st/HKkY.txt 2023-04-28 17:55:00 Right, diskless mode and data-disk mode never create ANY system partitin into which packages cna be installed. 2023-04-28 17:55:01 *model 2023-04-28 17:56:34 WhyNotHugo: how else would you expect diskless mode to work? 2023-04-28 17:56:54 everything happens in tmpfs 2023-04-28 18:06:55 Yeah, I was just being dumb. 2023-04-28 18:08:26 I mostly want to avoid wearing out the sdcard which has the OS, but also want packages installed on it. I guess I'm actually looking for a readonly-root, not diskless. 2023-04-28 18:14:50 WhyNotHugo: you want to avoid wearing out SDcard? how will diskless do that? you write Alpine to it, you install the packages in the apkcache, you configure things and do "lbu commit".......then no more writing occurs 2023-04-28 18:15:36 diskless is effectively a read-only root as the rootfs is READ from storage and unpacked in memory and run from memory 2023-04-28 18:41:19 I expected more magic, smh 2023-04-28 18:42:47 pj: you forgot to "apk add file"? ;-) 2023-04-28 18:57:52 (gnome-mines:15043): Gtk-WARNING **: 14:49:08.206: Found an icon but could not load it. Most likely gdk-pixbuf does not provide SVG support. 2023-04-28 19:06:52 usr/share/thumbnailers/librsvg.thumbnailer yup, missing 2023-04-28 19:07:16 whoa librsvg is missing everything 2023-04-28 19:07:44 usr/lib/gdk-pixbuf-2.0/2.10.0/loaders/libpixbufloader-svg.so 2023-04-28 19:14:32 apk upgrade 2023-04-28 19:22:58 ZZ saves & exits vi, how to quit without saving? 2023-04-28 19:23:17 :q 2023-04-28 19:23:24 or :q! if you have unsaved changes 2023-04-28 19:24:03 was thinking of quiting without the commandline 2023-04-28 19:25:36 ZX 2023-04-28 19:26:48 psykose, no libsrvg upgrade offered 2023-04-28 19:28:41 abby: hmm doesnt work with busybox vi 2023-04-28 19:28:54 neither does ZQ 2023-04-28 19:34:28 probably a vimism then 2023-04-28 19:45:35 Hi all, i have question. Is there any way to used abuid with git local folder? 2023-04-28 19:45:54 what does that mean? 2023-04-28 19:46:42 do you want to build from source that's in a folder somewhere locally? 2023-04-28 19:46:57 Yes 2023-04-28 19:52:53 Right now what i do creating archive of the source folder and create APKBUILD file. and call abuild. 2023-04-28 21:23:44 somercet1: yeah, you either need -U or apk update first to also update the indexes (and that also assumes the mirror is up to date), -r3 had it fixed 2023-04-28 21:24:19 maybe my mirror isn't as quick to update, I did. 2023-04-28 21:24:58 mirror.pit.teraswitch.com 2023-04-28 21:26:27 yeah, it still doesn't have it 2023-04-28 21:26:38 you probably want to use dl-cdn.alpinelinux.org anyway 2023-04-28 22:10:42 Hello, is ther a way to install AlpineLinux on a HDD directly from the iso file, not from a USB storage ? 2023-04-28 22:11:42 yes, write the ISO to a CD/DVD and boot from that ;-) 2023-04-28 22:12:04 not through the USB storage, nor a CD/DVD. 2023-04-28 22:12:18 well how do you expect to boot it? 2023-04-28 22:12:33 you can use grub to boot the iso, i suppose? 2023-04-28 22:12:37 with a grub menu. 2023-04-28 22:12:40 I guess you could use an existing bootloader's support for ISO images 2023-04-28 22:12:53 but if you don't already have a bootloader on the machine... 2023-04-28 22:13:04 or just.. write the iso to another partition and point grub there 2023-04-28 22:13:46 yes I have a boot loader. Actually, I want to test Alpine Linux on a recent x86_64 hardware, then to install it on my old Mac. 2023-04-28 22:14:23 well if you have Grub already installed I assume then you have some OS already installed on the machine - if you then run setup-alpine/setup-disk to install Alpine in Sys mode then it tends to assume it is the *only* OS on the disk and repartition/format the disk... 2023-04-28 22:16:41 I have a Linux installed. I dont see how to run setup-alpine tool. But I have a new brand SSD which I could use for AlpineLinux. 2023-04-28 22:18:28 you boot the ISO, and then as "root" run setup-alpine and it takes you through steps to install Alpine - however this "installer" is not designed with multi-OS installations in mind 2023-04-28 22:18:43 so if you want multi-boot then you would need to do some steps manually 2023-04-28 22:21:12 I need to learn more about grub to boot the iso. 2023-04-28 22:24:28 I was wondering to put in a /etc/grub.d/40_custom the contents of the 'grub.cfg' found in the /boot/grub/grub.cfg from Alpine would work 2023-04-29 00:59:19 There's no way to setup a monthly donation to Alpine through GitHub Sponsorship, right? 2023-04-29 00:59:37 Does Alpine have a LiberaPay presence? 2023-04-29 00:59:57 I looked through the wiki and I'm not confident the information about Supporting via donations is current/accurate 2023-04-29 02:02:38 don't think there are any donations 2023-04-29 02:12:45 I can do a little : 2023-04-29 02:36:26 Ah, rats. I'm in a place where I can support it (minimally) monthly, and I'd like to do so alongside a few other FOSS projects I do that for 2023-04-29 05:14:12 donate to alpine by buying me armored core vi when it comes out /j 2023-04-29 05:41:35 yea lmao 2023-04-29 05:41:43 usual japan figurine thing i guess 2023-04-29 05:41:53 pay hundreds of dollars for a plastic model 2023-04-29 05:45:44 Hey, beats spending about that much for a one-game-only controller like with Steel Batallion, right? 2023-04-29 05:46:30 telling on yourself? :D 2023-04-29 05:46:52 I wish ☹️ I got my OG Xbox in 2007, long long after that 2023-04-29 05:47:16 I have a MadCatz racing wheel 😭 2023-04-29 05:47:51 https://gitlab.alpinelinux.org/alpine/aports/-/merge_requests/46544 2023-04-29 05:48:11 If this looks okay, can someone smash merge? 2023-04-29 05:49:32 looks correct 2023-04-29 05:54:56 Weee 2023-04-29 05:55:32 If my life doesn't fall apart again this Summer, I might actually be able to maintain my packages in testing 2023-04-29 06:19:54 Holy shit, I had no idea LibreOffice had a GTK4 front-end. It's BEAUTIFUL. No more ugly fonts with the default (QT?) one 2023-04-29 06:20:25 default is some very old 'nothing' thing 2023-04-29 06:21:12 and yeah, the gtk3 one looked mostly the same too 2023-04-29 06:21:17 It's painful. It lags and everything is rendered oddly... Like it was downsampled using a really bad filter when drawn to screen 2023-04-29 06:21:19 used to just be called libreoffice-gnome i guess 2023-04-29 06:22:01 This is so much better integrated and just works/feels better. Wow. 2023-04-29 06:22:01 I assume the qt5/qt6 packages do the same for folks under Plasma? 2023-04-29 06:22:10 yeah pretty much 2023-04-29 06:23:37 `SAL_USE_VCLPLUGIN=gen libreoffice` if you want to burn your retinas again 2023-04-29 06:24:37 if you set SAL_USE_VCLPLUGIN=kde5 it uses a proper qt ui 2023-04-29 06:25:16 same as =qt5 isn't it 2023-04-29 06:25:25 maybe idk 2023-04-29 06:25:26 but they all look worse than the gtk4 one by my shitty tastes 2023-04-29 06:26:17 Nono, I'm good. I thought it had something to do with using noveau before, or my shit Intel HD 530 now... But really, it was just that I didn't install a package -_- 2023-04-29 06:28:10 psykose, you got me to put both machines back on Edge from latest-stable, haha 2023-04-29 06:28:29 Took a year-ish break, but I think I can deal again 2023-04-29 06:28:57 3.18 is in like two weeks or so anyway 2023-04-29 06:29:47 Yeah, I'm debating switching the work laptop back down after that is launched, but maybe staying with edge on the desktop. Who knows 2023-04-29 06:30:05 Is it better to maintain and do backports and such from edge or latest-stable, if I do get myself to that point this summer 2023-04-29 06:44:19 Saijin_Naib: when you switch between stable to edge, you get lots of .apk-new files 2023-04-29 06:45:26 is there a tool equivalent to needrestart on debian? 2023-04-29 06:49:57 nope 2023-04-29 09:08:26 is there nftables compatibility available? like the equivalent of iptables-nft in other distros that will allow iptables to have a limited view of nftables rules? 2023-04-29 09:13:15 iptables-nft is there 2023-04-29 09:13:23 iptables just doesn't point to it by default 2023-04-29 09:18:39 psykose: oh. is it in a separate package? i'll have to check again. 2023-04-29 09:20:19 nevermind. i'm blind. it's in the iptables package. >_< 2023-04-29 09:20:25 thanks. 2023-04-29 10:55:58 why abuild depends on tar, not busybox tar 2023-04-29 10:56:41 tar-1.34-r2 installed size: 488 KiB 2023-04-29 11:05:22 options and flags 2023-04-29 11:07:26 incidentally that makes up.. *checks notes* 2023-04-29 11:07:30 0.2% of build base 2023-04-29 11:08:53 https://gitlab.alpinelinux.org/alpine/abuild/-/blob/master/abuild-sign.in#L42 2023-04-29 11:08:53 these options all supported by busybox 2023-04-29 11:12:55 --owner is not 2023-04-29 11:13:00 --group is not 2023-04-29 11:17:03 just -o is probably similar 2023-04-29 11:23:02 but there is zero reason to use the bb version except have a worse tar by every metric, and save 0.2% of build-base 2023-04-29 11:23:05 idk what the point is 2023-04-29 11:23:35 up next: /usr/bin/cc is tcc 2023-04-29 11:30:31 I understand 2023-04-29 11:32:15 in general that is a nice thing to have for a base system, it's why busybox is nice in general 2023-04-29 11:32:21 you get all that stuff in one thing at no real cost 2023-04-29 11:32:38 past that once you set anything up, you usually always want real versions for anything you work with, etc 2023-04-29 15:22:43 uuidNuniq: .apk-new files are where *you* have modified a (config) file provided by a package and then a update of that package comes with a changed version (compared to the file in the previous version of the package) of the file - rather than overwrite your locally modified file the updated package file is stored with the .apk-new suffix 2023-04-29 19:35:11 is upgrading skipping minor versions allowed? e.g. 3.13 -> 3.17 2023-04-29 19:42:30 is anybody else using LXC on alpine? I'm having terrible trouble getting it to work correctly, and I was hoping somebody could try and replicate my issue 2023-04-29 19:42:50 Nulo[m]: yes 2023-04-29 19:43:03 Nulo[m]: though, you need to make sure you have the latest keys first 2023-04-29 19:43:28 3.13 -> 3.17 worked fine at least 🤷 2023-04-29 19:43:48 You probably had the newest keys already 2023-04-29 19:43:57 They have been backported 2023-04-29 19:49:16 cool! 2023-04-30 04:25:34 hi all :) my networking services (wpa_supplicant, networking, chronyd) always delay my boot although I do not usually need them immediately after logging in so I am looking for a way to start them asynchronously. 2023-04-30 04:26:45 I have added another runlevel which is stacked on top of the default runlevel and does not wait on the processes to exit. as described in the wiki 2023-04-30 04:27:55 but now the output from the networking services writes over my login prompt which is annoying 2023-04-30 04:30:17 adding a > /dev/null 2>&1 in the respective inittab line removes the " * starting ... [ ok ]" output but i still get output from for example dhcp broadcasting for a lease and i dont really know how to avoid that. does anyone know how that could be avoided? 2023-04-30 04:30:59 or does anyone know of a better way to solve my original problem? 2023-04-30 05:55:53 bpalu: rc_parallel in /etc/rc.conf might help with the startup time, but it can also cause other problems. YMMV. 2023-04-30 05:58:52 You could potentially have the kernel+init write its stuff to a different tty also. (on my machine I am using "console=tty2") 2023-04-30 09:37:18 i need a little help to understand something about alpine linux 2023-04-30 09:44:45 Best help 2023-04-30 11:34:23 psykose, it seems that webkit2gtk-4.1 needs to be rebuilt against libxml2 2023-04-30 11:35:48 When trying to link against its library, ld complains about symbols being for LIBXML2_2.4.30 2023-04-30 11:36:04 (that's on edge/x86_64) 2023-04-30 11:36:42 (where libxml2 is 2.11.0) 2023-04-30 11:40:25 oh, haha 2023-04-30 11:40:26 (10/12) Upgrading webkit2gtk-4.1-dev (2.40.1-r1 -> 2.40.1-r2) 2023-04-30 11:40:30 Allright. :) 2023-04-30 11:40:42 Right on time, thanks! 2023-04-30 13:55:17 zcrayfish: thank you! console=tty2 as a kernel parameter in your bootloader? 2023-04-30 15:02:15 good afternoon. How to unzip .rar? 2023-04-30 15:08:07 bsdtar xf file.rar 2023-04-30 15:08:23 libarchive-tools 2023-04-30 15:12:23 hi, can you have the cache dir on another disk than lbu? I tried, but cache disk is not mount automatically, I can mount it with fstab but than simply packages are not installed during boot 2023-04-30 15:38:40 it seems that creating a fake apkovl file on the disk make it automount, I'm not sure where on the alpine code this is defined 2023-04-30 15:40:34 godpt: nlpug-findfs (part of the initramfs) finds the apkovl file AFAIK 2023-04-30 15:44:01 godpt: https://gitlab.alpinelinux.org/alpine/mkinitfs/-/blob/master/initramfs-init.in#L624 2023-04-30 15:47:09 minimal: great thanks, what is the .boot_repository thing? 2023-04-30 15:48:04 godpt: it's a file isn't it? 2023-04-30 15:48:39 I tried creating it on the disk root but it does not make it automount 2023-04-30 15:52:34 Hmm, apparently Alpine does not let you know whether a USB device is connected :/ 2023-04-30 15:52:58 On virtually every distro I've tried whenever I insert a USB device I can see in the logs something like: kernel: usb 2-1.1: new high-speed USB device number 5 using ehci-pci 2023-04-30 15:53:12 That happens on Alpine Linux as well 2023-04-30 15:54:44 And that should appear in /var/log/messages right? 2023-04-30 15:55:06 no, dmesg 2023-04-30 15:56:33 It's the kernel that logs that 2023-04-30 15:57:15 But tail -F /var/log/dmesg is also silent. 2023-04-30 15:57:43 that would depend on whether you have a syslog daemon running and its configuration 2023-04-30 15:58:08 Guest10863: try dmesg as a command (as root) 2023-04-30 16:00:35 Oh wow indeed, doas dmesg does have the latest logs and apparently I don't have a syslog daemon running. Thought it was on by default (since it was on virtually all distros I tried). Truly a minimal system :p 2023-04-30 16:02:00 afaik, busybox syslog is enabled by default 2023-04-30 16:02:07 which is what provides /var/log/messages 2023-04-30 16:04:01 Thank's 2023-04-30 16:04:48 ikke: isn't it klogd (also part of Busybox) that handles kernel messages? 2023-04-30 16:06:27 in that klogd passes kernel messages via syslog and syslogd then writes them to a file 2023-04-30 16:08:09 This is what I have running: https://0x0.st/HPpu.jpg - I don't think I manually enabled any services. 2023-04-30 16:08:44 (sorry for blurtyness my phone sucks) 2023-04-30 16:08:55 so you have neither klogd nor syslogd running 2023-04-30 16:09:11 oh lol 2023-04-30 16:09:26 minimal: I got it, the .boot_repository file needs to be inside the cache dir, then the disk is automounted on boot, much cleaner than a fake apkovl. Thanks again 2023-04-30 16:15:37 damn, it looks like the virt kernel although having alsa enabled is missing any of the sound cards supported by qemu. Can you please add at least one? 2023-04-30 16:15:59 godpt: from memory it has virtio_snd support 2023-04-30 16:16:20 minimal: unfortunately it's not yet supported by qemu 2023-04-30 16:16:40 it'd be great if it was 2023-04-30 16:16:55 godpt: so use linux-lts kernel if you want to emulate a physical device? 2023-04-30 16:17:53 tried it, but lts had another problem, I think it doesn't have virtio gpu drivers 2023-04-30 16:18:53 still, I think it would make sense for linux-virt to have one of the cards supported by qemu, considering alsa is enabled 2023-04-30 16:18:53 well are you trying to have yuo VM pretend to be a "real" physical machine or just a virtual (basically virtio only) machine? 2023-04-30 16:19:15 virtio as much as possible 2023-04-30 16:19:49 exactly, my thought too, I'm surprised that Qemu doesn't support virtio-snd, I assumed it did 2023-04-30 16:20:32 yes, it has been in the making for a long time, who knows how long it will take 2023-04-30 16:21:17 Apparently root does have /sbin/syslog -t running, but it's not showing up in rc-status. Not sure if normal. 2023-04-30 16:21:39 errr, /sbin/syslogd -t 2023-04-30 16:22:13 It may have been started because of another service being dependant upon it 2023-04-30 16:22:24 then rc-status -a should show it 2023-04-30 16:23:14 godpt: yeah strange, an online search for "qemu virtio-snd" isn't picking up any mentions 2023-04-30 16:24:19 Indeed, under Runlevel: boot 2023-04-30 16:24:48 minimal: yeah, so you think it's an idea to add intel hda drivers to linux-virt? 2023-04-30 16:25:57 minimal: https://wiki.qemu.org/Internships/ProjectIdeas/VirtioSound 2023-04-30 16:28:18 godpt: This can be a gamechanger. Imagine virtualizing audio production software which is Windows-only. That is currently almost impossible. 2023-04-30 16:28:57 well, I'm not sure Windows has the virtio sound driver 2023-04-30 16:29:05 Guest10863: does "rc-status -a" also show "klogd"? 2023-04-30 16:29:27 It does not, so now I'm figuring out how to add it to services most effectively. 2023-04-30 16:29:33 godpt: I guess hda would be the most likely driver to add 2023-04-30 16:30:05 Guest10863: "rc-update add klogd boot" 2023-04-30 16:30:44 godpt: at least virtio-snd is on Qemu's radar, looks like it'll be probably 6+ months before a release ships with it though 2023-04-30 16:33:33 where can I propose the change for hda? It seems the package mantainer is Natanael Copa 2023-04-30 16:35:53 Awesome, I also did doas rc-service klogd start and I can see kernel messages inside /var/log/messages. Now I'm hopefully closer to figuring out why my unattended alpine laptop's DNS got borked and upon reboot couldn't even get a network running :p 2023-04-30 16:36:55 godpt: You open an Issue in Gitlab 2023-04-30 16:37:35 godpt: also I forgot the virtio-snd isn't currently enabled in linux-virt (it's on my list of things to enable), CONFIG_SOUND is not set for linux-virt at all 2023-04-30 16:37:58 nor should it be 2023-04-30 16:38:33 Why not? 2023-04-30 16:38:38 psykose: if virtio-snd is to be enabled then I believe CONFIG_SOUND would also need to be enabled 2023-04-30 16:39:13 >named -virt 2023-04-30 16:39:17 >enable literally every option 2023-04-30 16:39:23 we have that already it's called -lts 2023-04-30 16:39:46 virtio-snd is a VM-specific device for sound so it only makes sense to enable it in linux-virt 2023-04-30 16:39:53 ah, right, it should have the options you want and not the other options, right 2023-04-30 16:40:09 psykose: that's understandable, but in that case then shouldn't -lts have all the virtio drivers? 2023-04-30 16:40:36 probably yeah 2023-04-30 16:40:59 psykose: the problem is *currently* QEMU doesn't support virtio-snd and so godpt wants a single *physical* sound device enabled in linux-virt to do audio (until Qemu supports virtio-snd) 2023-04-30 16:41:01 virtio-snd would also actually make sense in the other too 2023-04-30 16:41:26 right now you can't use alpine as a desktop VM without a custom kernel 2023-04-30 16:41:48 i'm pretty sure -lts would work for that fine 2023-04-30 16:42:12 godpt: you can (using linux-lts), it just isn't "optimal" 2023-04-30 16:42:29 ok I'm trying again, I think it had problems 2023-04-30 16:44:35 in any case yes, what minimal said, until virtio snd is not on qemu, it might make sense to have drivers for one physical sound card supported by qemu 2023-04-30 16:47:58 linux-lts doesn't start at all on qemu 2023-04-30 16:48:47 it should. How far did it get? 2023-04-30 16:49:35 /init: line 581: nplug-findfs: not found 2023-04-30 16:49:53 before that there is a lot of no such file or directory 2023-04-30 16:50:21 sounds like problem with your initramfs 2023-04-30 16:51:02 I'm using the standard iso 2023-04-30 16:52:42 just qemu-system-x86_64 -cdrom alpine-standard.iso 2023-04-30 16:53:04 hmm 2023-04-30 16:53:57 you need to set the memory too 2023-04-30 16:54:43 ah of course, personally I never run qemu with some a minimal command line ;-) 2023-04-30 16:54:53 s/some/such/ 2023-04-30 16:54:54 ok, it was running out of ram 2023-04-30 16:57:17 yeah it defaults to 128MiB but the ISO "installs" and runs Alpine from memory 2023-04-30 16:59:41 maybe it's fine, it's just that my desktop vm will take more resources than otherwise. A bit sad 2023-04-30 17:02:45 godpt: I've suggest you open an Issue to add HDA support to linux-virt and mention it is a "temporary" solution until Qemu supports virtio-snd 2023-04-30 17:03:19 minimal: ok, I'll do it 2023-04-30 17:16:02 https://gitlab.alpinelinux.org/alpine/aports/-/issues/14850 let me know if I should change anything 2023-04-30 17:19:27 godpt: looks fine 2023-04-30 17:20:31 Ok, I think I've figured it out. Not sure why the DNS got messed up, but upon boot the system lost network connectivity because it was attempting to bring up the internal NIC (which isn't connected) and not the external NIC. It basically switched eth0 and eth1 around. I edited /etc/network/interfaces to change th0 into eth1 and now everything works again. Yay! 2023-04-30 17:21:14 minimal, ok thanks, with github I'm used to having to fill millions checkboxes for new issues 2023-04-30 17:22:21 Guest10863: I guess that was to do with the order network device drivers were loaded 2023-04-30 17:25:59 That could very well be. 2023-04-30 17:33:45 So suppose I wanted to only load external NIC and assign it eth0. How would I best do that? Via modprobe or netdev=? There's a bunch of hints here: https://unix.stackexchange.com/questions/10254/how-to-change-the-order-of-the-network-cards-eth1-eth0-on-linux - but I would like to know the best method suitable for Alpine and maybe even add it to the docs here: https://wiki.alpinelinux.org/wiki/Configure_Networking 2023-04-30 17:35:15 are you using mdev or eudev? eudev certainly supports having a file to define MAC address to device name mapping 2023-04-30 17:36:37 Not sure, I just plugged in the USB NIC and it worked. Got recognized in the output of ip a so I went with that. 2023-04-30 17:38:31 I do have /etc/mdev.conf and the mdev binary (which links to BB) 2023-04-30 17:39:06 is the "eudev" package installed? 2023-04-30 17:43:05 Apparently yes: eudev-libs-3.2.11-r0 - Dynamic library to access udev device information 2023-04-30 17:43:37 that's the libs, not the main package 2023-04-30 17:43:41 so you're using mdev then 2023-04-30 17:43:57 https://wiki.alpinelinux.org/wiki/Custom_network_interface_names 2023-04-30 17:54:05 Guest10863: setup-devd mdevd 2023-04-30 17:54:17 mdevd is much better than mdev 2023-04-30 17:55:34 uniqNuuid: that doesn't solve his issue of wanting to ensure 2 network devices are named in a particular order 2023-04-30 17:57:40 ok, the article is good, 2 files needed to modify 2023-04-30 17:58:19 but it is manual. 2023-04-30 17:59:06 manual name the interface in /etc/mactab ? 2023-04-30 18:35:51 hi, I installed alpine today on my laptop, I used setup-desktop to install Gnome, but for some reasons my shell locale is set to fr_FR even if I used export LC_ALL=fr_FR.UTF-8 in /etc/profile.d/locale.sh , if I log in XFCE the locale is correct... I'm out of idea 2023-04-30 18:37:30 solene: I just saw your thread on mastodon 2023-04-30 18:37:53 solene: would it be possible for you to open an issue on httpsgitlab.alpinelinux.org/alpine/aports? 2023-04-30 18:38:02 ikke, sure 2023-04-30 18:38:04 https://gitlab.alpinelinux.org/alpine/aports? 2023-04-30 18:38:52 We have some dedicated gnome contributors who might be able to help 2023-04-30 19:07:01 there is no difference between fr_FR and fr_FR.UTF-8 2023-04-30 19:07:08 or `fr` 2023-04-30 19:07:13 it's the same meaning for all of those 2023-04-30 19:12:07 psykose: they mentioned flatpak has issues with it 2023-04-30 19:12:29 https://fosstodon.org/@solene@bsd.network/110288780312831711 2023-04-30 19:14:47 glibc probably treats them differently so it matters there i guess 2023-04-30 19:16:56 psykose, indeed, I spotted the issue with flatpak programs, they take fr_FR as non utf-8, overriding the flatpak LC_ALL to fr_FR.UT-8 solves the problem, but it's a workaround 2023-04-30 20:03:32 I made an issue :) thanks for your support 2023-04-30 20:04:42 Shouldn't flatpak pass to its clients LC_ALL and related variables? 2023-04-30 20:05:33 afaik, fr_FR inside flatpak is actually a utf8 locale, even if it doesn't have .utf8 at the end 2023-04-30 20:05:58 (because the codepage is defined inside locale data, not *by* the .suffix, and programs that check for the .suffix instead of nl_langinfo() are broken) 2023-04-30 20:06:13 no idea, but the thing is it's not working :) 2023-04-30 20:06:18 `LC_ALL=fr_FR locale charmap` inside a flatpak shell should output utf8 2023-04-30 20:06:53 I could reproduce it with Lagrange when asking to view the Downloads directory, or qbittorrent when defining a directory 2023-04-30 20:07:52 funnily you don't need flatpak for either of those 2023-04-30 20:09:02 indeed, but that should work as expected 2023-04-30 20:10:35 weirdly i can't actually reproduce that 2023-04-30 20:10:47 just fr_FR set for flatpak shows Téléchargements or whatever fine 2023-04-30 20:11:39 psykose, can you share how to reproduce that? 2023-04-30 20:11:52 export LC_ALL=fr_FR 2023-04-30 20:11:55 flatpak run .. 2023-04-30 20:12:29 i tested lagrange directories (funnily flatpak lagrange doesn't even support wayland) 2023-04-30 20:12:45 or $ flatpak run --env=LC_ALL=fr_FR ... 2023-04-30 20:18:38 minimal: that guide doesn't work for me unfortunately. Tried renaming eth0 to eth1 but it's still the same after boot. Also the `doas nameif -s` exits with status 1 even if I rebuilt the ramdisk, 2023-04-30 20:18:56 s/,/./g 2023-04-30 20:19:55 psykose: and with flatpak --env="LC_ALL=fr_FR" run fi.skyjake.Lagrange ? 2023-04-30 20:20:08 for me, it doesn't work as expected until I use flatpak --env="LC_ALL=fr_FR.UTF-8" run fi.skyjake.Lagrange 2023-04-30 20:20:32 yeah, works fine 2023-04-30 20:20:46 interesting 2023-04-30 20:21:00 I'll try on a fresh user 2023-04-30 20:21:54 just env -i HOME="$HOME" USER="$USER" XDG_RUNTIME_DIR="$XDG_RUNTIME_DIR" DBUS_SESSION_BUS_ADDRESS="$DBUS_SESSION_BUS_ADDRESS" DISPLAY="$DISPLAY" flatpak --env="LC_ALL=fr_FR" run fi.skyjake.Lagrange works fine, eliminating anything i might've had set or not 2023-04-30 20:21:56 so no idea 2023-04-30 20:22:04 you can try that too just in case magic other stuff affects it 2023-04-30 20:22:46 solene: so with --env, it works fine? 2023-04-30 20:23:40 no, only with the utf-8 version for them 2023-04-30 20:25:00 ok, so my fresh user, upon gnome login, has a correct locale of fr_FR.UTF-8, which is a first big difference 2023-04-30 20:26:22 `flatpak run --command=/bin/sh org.this.that` and inspect `env` and `locale` and `locale charmap` inside 2023-04-30 20:26:49 if `locale charmap` says it's utf-8 then it's utf-8 2023-04-30 20:26:58 also grep your home folder for stuff that sets random things i guess in the profile scripts 2023-04-30 20:27:08 or just type `env` and pastebin it (from the broken one) 2023-04-30 20:32:45 grawity, locale charmap says ISO-8859-1 2023-04-30 20:33:20 I'll investigate, I want to understand where this comes from now 2023-04-30 20:33:32 I'll keep you updated, if not on IRC, on the gitlab issue 2023-04-30 20:40:10 ok, so I changed my shell from fish to bash, and the locale is correct, so it's a fish + gnome issue 2023-04-30 20:40:23 because fish + xfce works 2023-04-30 20:41:04 can u post `env` from fish + gnome, and fish + xfce 2023-04-30 20:41:18 also fish does not read /etc/profile* stuff 2023-04-30 20:41:49 but it shouldn't really matter as all you need is just `LANG=fr` generally and everything else works, etc 2023-04-30 20:42:46 my friend has read fr_FR and wants to share a funny meme with the class: "🇫🇷 🇫🇷 no cap" 2023-04-30 20:42:49 you're welcome 2023-04-30 20:48:16 GNOME env https://pbot.rmdir.de/5VJMGPxwAJAQ8SgqAVRLYg (logged in with gdm) 2023-04-30 20:48:35 xfce env https://pbot.rmdir.de/V9gYxoL4LifqryyRYclrNw (logged in with lightdm) 2023-04-30 20:49:09 I sorted the env output so a diff is easier 2023-04-30 20:52:51 i guess CHARSET=UTF-8 is in the latter but none of those matter in my testing 2023-04-30 20:55:59 uniqNuuid: Tried that, but does not work on reboot. 2023-04-30 20:56:09 (the /etc/mactab thing) 2023-04-30 20:59:05 Guest10863: "Also the `doas nameif -s` exits with status 1" - in that case it is obviously not working. 2023-04-30 20:59:42 were the interfaces "down" when you tested that? 2023-04-30 21:00:26 According to ip a yes. 2023-04-30 21:00:39 Also, I did a reboot after that. 2023-04-30 21:01:41 well if it doesn't work when you manually run it then it is unlikely to work on boot 2023-04-30 21:04:22 I tested fish on my test user, and everything is working fine for it... the diff between my user env and the new user env, both on fish https://pbot.rmdir.de/u/1FBlv9McyHwT4dldLigu8Q I can't explain 2023-04-30 21:04:50 maybe GNOME has some setting somewhere upon login that is stored in the user configuration 2023-04-30 22:38:03 that's puzzling, still on my locale issue. I tried to empty my $HOME, log in again and the locale was wrong 2023-04-30 22:38:22 I deleted my user, recreated it with the same username, chowned the $HOME to the new user, and the locale is right now 2023-04-30 22:44:09 i think gdm does some weird things https://github.com/void-linux/void-packages/pull/39264#issuecomment-1369171391 2023-04-30 22:44:35 abby, here it looked like gdm was doing something special depending on the user :s 2023-04-30 22:44:48 I have to go, I filled update in the gitlab issue 2023-04-30 23:04:52 perhaps there's a legendary non-env dconf attribute or whatever it was that impacts it 2023-04-30 23:05:39 abby: that wouldn't really affect it because it's only the LANG value and not really in itself the issue here 2023-04-30 23:22:35 IDENTIFY