2022-02-01 01:29:13 https://f.sakamoto.pl/8ed84506-2a2c-4fc0-90bf-ee5f0cf81fddfile_28.jpg https://f.sakamoto.pl/9ba8d4d2-faa7-42a3-9a3a-7c42c764fba6file_29.jpg 2022-02-01 01:29:39 T_T I added a version number to the edit summary field, why does it get matched as an IP address? 2022-02-01 01:32:41 you can just submit twice and it should go through 2022-02-01 01:32:43 for an edit 2022-02-01 02:00:25 psykose: changed the message a bit and it went through, just felt like I should report that 2022-02-01 02:00:45 the block is for any change at all- the spam filtering is quite.. hostile 2022-02-01 02:00:53 but you submit it twice and it should be fine 2022-02-01 02:01:09 and now that you have an edit in it should let you make new pages as well after ~24 hours 2022-02-01 02:42:08 How do I make Alpine mount my filesystem(s) with compress=zstd? Seemingly fstab is ignored 2022-02-01 02:44:03 what is the fs type? i don't think most filesystems support an arbitrary compress= param 2022-02-01 02:49:40 psykose: BTRFS 2022-02-01 02:49:51 It works fine if I `mount -o remount,compress=zstd / # or /home` 2022-02-01 02:51:19 hm 2022-02-01 02:51:41 can't think of why the localmount service wouldn't apply it 2022-02-01 02:58:53 I'm pretty sure / is a special case, and may or many not be remounted 2022-02-01 03:00:29 hello alpine, I'm running the latest armv7 version diskless. when I lbu_commit -e it complains of deprecated key derivation used, when I try to boot with an encrypted overlay I am prompted: enter aes-256-cbc decryption password but when I do I get error writing output file and get dumped into emergency shell. am I missing a step? 2022-02-01 04:06:28 atka: Do you have mdadm and cryptdev? 2022-02-01 04:06:43 Does anyone know if you can build a package against a dependency in Testing? 2022-02-01 04:08:38 "seatd is much cleaner and..." <- i found that needed elogind to set up network manager, even being in the plugdev group was insufficient to manage the settings 2022-02-01 04:08:41 i think cos it ties in with polkit 2022-02-01 04:10:33 Saijin_Naib[m]: only if the package is also in testing 2022-02-01 04:10:49 as for yourself, sure, if you're on edge you just add the -dev 2022-02-01 04:11:00 psykose: Ahh, makes sense. Barrier between core stuff and testing... 2022-02-01 04:11:05 mhm 2022-02-01 04:11:27 Yeah, I'm on Edge with repositories all enabled, but the APKBUILD didn't seem to pull in the dependency... 2022-02-01 04:11:32 Let me see if I did someting wrong 2022-02-01 04:11:47 you could paste it 2022-02-01 04:11:52 i'd have a look 2022-02-01 04:11:59 and post any errors 2022-02-01 04:12:07 ACTION sent a code block: https://matrix.org/_matrix/media/r0/download/matrix.org/KsIywgYPXLNHzCyzcCqXythr 2022-02-01 04:12:18 ACTION sent a code block: https://matrix.org/_matrix/media/r0/download/matrix.org/tvfeEnPKhtfJPxuStEfvmBGC 2022-02-01 04:12:44 Basically trying to test tumbler bult against libopenraw to see if it will properly thumbnail RAW image files 2022-02-01 04:13:03 libopenraw-dev exists 2022-02-01 04:13:08 hm 2022-02-01 04:13:10 Then possibly ask for these changes to be integrated, and someone knowledgeable to take over maintainership of my libopenraw so it can go Community 2022-02-01 04:13:10 what error do you get 2022-02-01 04:13:49 None, it just doesn't come in as one of the makedepends/installed packages, and the autoconfigure script doesn't detect or use it, so the RAW thumbnailer doesn't get built 2022-02-01 04:13:49 also that patch looks like it doesn't change the lines 2022-02-01 04:13:55 oh 2022-02-01 04:13:58 >_> 2022-02-01 04:13:59 not an abuild error then 2022-02-01 04:14:03 i think we went over this months ago 2022-02-01 04:14:14 Probably -_- 2022-02-01 04:14:17 the issue is the .pc of the libopenraw/libopenraw-gnome stuff requests a wrong version 2022-02-01 04:14:23 the .pc files try to look for something that doesn't exist 2022-02-01 04:14:35 like the -gnome thing looks for openraw 0.3 but the actual version is 1.0 2022-02-01 04:14:37 or something 2022-02-01 04:14:40 i forgot, but it was that 2022-02-01 04:14:50 the thing has to be patches on libopenraw itself iirc 2022-02-01 04:14:59 https://gitlab.xfce.org/xfce/tumbler/-/merge_requests/18 2022-02-01 04:14:59 Ah, this patch is supposed to sidestep that: 2022-02-01 04:15:15 According to the maintainer for tumbler 2022-02-01 04:15:49 this patch does look like it changes the ac checks but they were never the issue 2022-02-01 04:15:55 https://gitlab.xfce.org/xfce/tumbler/-/commit/ad680ef6c79a6430ca6bb0b9e48538eb9c64d202 2022-02-01 04:15:57 the issue is the .pc files of libopenraw itself 2022-02-01 04:16:21 Ah, okay. So their patch doesn't actually address it 2022-02-01 04:16:46 libopenraw-gnome-0.3.pc says it depends on libopenraw-0.1 2022-02-01 04:16:58 unless... that is fine 2022-02-01 04:17:03 and it picks the newer versions 2022-02-01 04:17:08 but i would expect that to have a >= or something 2022-02-01 04:17:14 shrug 2022-02-01 04:17:21 you can try hotpatch random numbers into the two files 2022-02-01 04:17:26 and also see if just the upstream patch works 2022-02-01 04:18:00 I think I'm using that patch locally in my build as the 10-dependencyversions.patch file 2022-02-01 04:18:06 Or am I not understanding you 2022-02-01 04:18:24 the patch you pasted is not the same patch 2022-02-01 04:19:19 It's a patch diff between release 4.16.0 acinclude.m4 and the one from the highlighted commit 2022-02-01 04:19:26 I did that wrong then 2022-02-01 04:20:11 i mean just the commit applies cleanly 2022-02-01 04:20:15 but it doesn't fix the issue 2022-02-01 04:20:45 then changing uhh 2022-02-01 04:21:16 Yeah, a few other files 2022-02-01 04:21:20 `Requires: libopenraw-0.1` to libopenraw-0.3 in libopenraw-gnome-0.3.pc fixes it and it builds 2022-02-01 04:21:33 you're so freaking fast 2022-02-01 04:22:20 yeah, you need both the upstream patch and the change to libopenraw 2022-02-01 04:22:26 not sure why the .pc is broken but a sed should fix it 2022-02-01 04:23:20 So I should fix my libopenraw APKBUILD with that change to the pc file, right, and push that? 2022-02-01 04:25:41 think just this https://img.ayaya.dev/9EMs2gwIzmBg 2022-02-01 04:25:49 but i can't tell you if the final thing will work 2022-02-01 04:25:58 at runtime 2022-02-01 04:26:02 after building everything 2022-02-01 04:26:46 That's incredible, thanks for the fix 2022-02-01 04:26:56 I should hopefully be able to verify soon 2022-02-01 04:28:31 it also needs a patch for new gcc.. 2022-02-01 04:30:58 >_> this stupid package... 2022-02-01 04:32:09 https://img.ayaya.dev/NN7oveO8NBNG 2022-02-01 04:32:10 this 2022-02-01 04:32:29 and this https://img.ayaya.dev/ipFrq5u6IK0R 2022-02-01 04:32:33 and it detects and builds fine 2022-02-01 04:32:36 i did not check the runtime though 2022-02-01 04:32:40 you can do that :) 2022-02-01 04:33:00 How did you know it needed a GCC patch? 2022-02-01 04:33:16 it was just missing an include that was seemingly fine before 2022-02-01 04:33:23 i just assume it's because of new gcc 2022-02-01 04:33:36 cause nothing else would make the same source not compile anymore with the same includes 2022-02-01 04:33:37 i guess 2022-02-01 04:33:45 unless a dependency also changed some headers 2022-02-01 04:34:02 either way fixed above 2022-02-01 04:38:45 Damn, awesome. Thanks so much for that! 2022-02-01 04:38:56 Also really interesting to see how people work around and on this stuff 2022-02-01 04:56:56 ACTION sent a code block: https://matrix.org/_matrix/media/r0/download/matrix.org/dabgantcsxqwmdcmHZEbrAeo 2022-02-01 04:57:00 Of course I broke it 😛 2022-02-01 04:57:17 that means you forgot the patch 2022-02-01 04:57:25 that's the error w/o it 2022-02-01 04:58:33 ACTION sent a code block: https://matrix.org/_matrix/media/r0/download/matrix.org/PmHnDfpAksCpBUDsvYXupdZt 2022-02-01 04:58:51 That's what I have as the fix-include.patch which I added to the APKBUILD 2022-02-01 04:58:59 ACTION sent a code block: https://matrix.org/_matrix/media/r0/download/matrix.org/MrrCVmBWJyOfVqNXZnsPalNy 2022-02-01 04:59:37 are you sure you re-ran prepare or abuild from scratch 2022-02-01 04:59:53 oh 2022-02-01 04:59:55 your patch is wrong 2022-02-01 04:59:59 abuild checksum && abuild -r 2022-02-01 05:00:12 you copied the top part into it 2022-02-01 05:00:25 it creates a testing/libopenraw.. 2022-02-01 05:00:29 instead of the actual file below 2022-02-01 05:00:41 delete @@ -0,0 and the parts above 2022-02-01 05:00:51 Oh... So just begin with line 4? 2022-02-01 05:00:56 well 2022-02-01 05:01:00 the entire thing is wrong actually 2022-02-01 05:01:04 cause you didn't apply it but just copied it 2022-02-01 05:01:24 go to aports root and `patch -p1 < thatpatch` 2022-02-01 05:01:30 then it will make the correct one 2022-02-01 05:03:44 Thanks, as ever. Have no idea why I thought I could just use the patch like that 2022-02-01 05:04:07 i should've just send the actual file instead :p 2022-02-01 05:04:26 Eh... That's a lot of hand-holding beyond you already doing a lot 2022-02-01 05:04:33 I should know what a patch means 2022-02-01 05:04:39 Getting there! 2022-02-01 05:05:17 So when I install my patched libopenraw, will the APKBUILD pull my locally installed one with the patches when I rebuild tumbler? 2022-02-01 05:05:38 I don't know how its solver works 2022-02-01 05:06:39 if you have your local repo added to /etc/apk/repositories 2022-02-01 05:06:43 the new one will have a higher pkgrel 2022-02-01 05:06:47 so it will be the one installed 2022-02-01 05:07:00 you can see by what it installes 2022-02-01 05:07:10 if it's 0.3.0-r1 then it can only be the new one you made 2022-02-01 05:07:25 you can also manually install it ahead of time if you wish to be sure 2022-02-01 05:08:23 Ahh, awesome. Thanks! 2022-02-01 05:35:18 psykose: built perfectly and the raw-thumbnailer.so got created/installed 2022-02-01 05:35:33 Doesn't work to thumbnail RAF files (which it should), but hey haha. Progress! Thanks! 2022-02-01 05:35:38 nice 2022-02-01 05:38:46 Worth submitting the changes for this if it isn't technically resolved fully? 2022-02-01 05:41:06 sure 2022-02-01 05:41:10 there are upstream commits you can use 2022-02-01 05:41:15 both my changes are fixed upstream 2022-02-01 05:41:16 https://gitlab.freedesktop.org/libopenraw/libopenraw/-/commit/21a37ac5487ed6138c3745d5bf33c15d9535147a 2022-02-01 05:41:18 https://gitlab.freedesktop.org/libopenraw/libopenraw/-/commit/43fac545b135bce981b6837aa6cb111ee853f090 2022-02-01 05:41:22 identically 2022-02-01 05:41:30 didn't even check 2022-02-01 05:41:32 whaaaat 2022-02-01 05:42:06 now you don't have to pretend the patches fell off the back of a truck 2022-02-01 05:42:07 :p 2022-02-01 05:42:32 Hah, nice 2022-02-01 05:42:39 just copy paste those in (you can add .patch to those urls to just pull the patch) and embed a Patch-Source: with the same url into them 2022-02-01 05:44:10 That's a pretty neat trick 2022-02-01 05:44:15 Alright, sweet. I'll try that out 2022-02-01 05:44:22 Thanks! 2022-02-01 06:17:32 hello 2022-02-01 06:17:54 I have a problem with pipewire 2022-02-01 06:18:26 I set it up using the wiki's instructions 2022-02-01 06:19:25 when i run `pw-cat -p test.flac` I get this error 2022-02-01 06:19:39 stream node 47 error: no node available remote error: id=3 seq:7 res:-2 (No such file or directory): no node available 2022-02-01 06:20:05 test.flac was replaced with a different flac 2022-02-01 08:50:24 anyone got a policyklit for alpine linux udisk2? 2022-02-01 08:50:41 like for example if i mount a disk with "udisksctl mount -b /dev/disk" 2022-02-01 08:50:46 and then want to unmount it in thunar 2022-02-01 08:51:31 i get the error "Failed to eject "6EA7-F153" unmount: /run/media/user/6EA7-F153: must be super user to unmount. 2022-02-01 08:51:58 iirc to fix that you need a policy in /etc/polkit-1/rules.d 2022-02-01 08:52:26 i've got dbus, udev-postmount, udev-settle, etc running and i installed lxqt-polkit 2022-02-01 08:52:47 and polkitd is running 2022-02-01 08:53:48 lxqt-policykit-agent is also running 2022-02-01 08:55:36 iirc you need an authorization rule for it https://wiki.archlinux.org/title/Polkit#Authorization_rules 2022-02-01 08:57:21 interesting those commands do not ask for a password at all https://wiki.archlinux.org/title/Udisks#Usage 2022-02-01 08:57:29 makes me think that thunar isn't using it 2022-02-01 09:22:24 AH figured it out 👍️ 2022-02-01 09:22:51 seems running thunar from the launcher doesn't run it with `dbus-run-session` 2022-02-01 09:23:20 you should be running *your entire desktop session* via `dbus-run-session` 2022-02-01 09:23:22 not individual apps 2022-02-01 09:23:36 like if you're using i3, make it `dbus-run-session i3` 2022-02-01 09:24:11 yeah that's true 2022-02-01 09:24:59 yeah i think i figured out where i messed up 2022-02-01 09:25:06 my exec script only had "exec sway" 2022-02-01 09:25:33 i know why this worked on archlinux, because before it i had: dbus-update-activation-environment --systemd DISPLAY 2022-02-01 09:25:39 which obviously does nothing on alpine linux lol 2022-02-01 09:26:40 kinda, it was also that on Arch systemd would automatically start a dbus session bus upon login, regardless of that 'update' command 2022-02-01 09:27:33 https://dbus.freedesktop.org/doc/dbus-update-activation-environment.1.html 2022-02-01 09:29:01 yeah I know what it does 2022-02-01 09:44:54 systemd has rotted my brain lol 2022-02-01 09:45:05 i forget about these things 2022-02-01 09:52:59 dngray[m]1: it happens 2022-02-01 09:53:44 the rot is real 2022-02-01 10:01:40 but of course it depends on how you look at it.... should i really be exercising brain power worrying about such things 2022-02-01 10:05:47 its a valid question? 2022-02-01 10:06:13 expend brain power on cute 2022-02-01 10:13:33 i do that already 2022-02-01 10:13:38 https://www.theguardian.com/world/2022/jan/31/polar-bears-move-into-abandoned-arctic-weather-station-photo-essay 2022-02-01 10:14:16 https://www.youtube.com/watch?v=9vgnXRypc4o 2022-02-01 10:20:01 they weigh less than 2 pounds inside the den 2022-02-01 11:27:01 Hi! Anyone know why the linux-rpi kernel version in Alpine lags so far behind the linux-lts one? In 3.15, linux-rpi is at 5.15.4-r0 from Nov 2021, while linux-lts is at 5.15.16-r0, which is from last week. It's a similar story in 3.14 and 3.13 too.... (full message at https://matrix.org/_matrix/media/r0/download/matrix.org/MYZNatuOxJSWxOiyzsXtzvWu) 2022-02-01 11:36:47 does anyone know how you define a portrange in awall? 2022-02-01 11:38:50 jhjacobs81: the port field of a service accepts a range 2022-02-01 11:39:35 alexhaydock: current ubuntu 21.10 are at 5.13.0-29-generic #32-Ubuntu SMP Fri Jan 28 11:51:06 UTC 2022 x86_64 x86_64 x86_64 GNU/Linux 2022-02-01 11:41:00 jhjacobs81: "service": [ {.., "port": "4000-4100"} ] 2022-02-01 11:42:04 Apachez - Well true, Alpine is certainly tracking the stable upstream faster than Ubuntu do, but rather than comparing release models or distributions I'm just comparing the versions of the two kernels (Pi one and standard one) within the same Alpine branch and noting they're quite different 2022-02-01 11:42:59 ahh okay thanks :) 2022-02-01 11:43:22 Patch version of 5.15.4 vs 5.15.16 suggests the Pi version of the kernel (which shouldn't really be all that different) is missing 12 upstream patches that are in the LTS kernel, if I'm understanding it right. 2022-02-01 11:43:38 alexhaydock: it depends on how quick https://github.com/raspberrypi/linux is updated 2022-02-01 11:43:54 Ahh there it is - that'll be what I was missing 2022-02-01 11:45:32 there ya go 2022-02-01 11:45:47 but yeah otherwise I agree with you would be nice with same linux major/minorversion for each arch 2022-02-01 11:45:49 So I guess that means the Pi still doesn't actually run on mainline Linux without patches, if Alpine is forced to use the Pi foundation's kernel upstream, rather than just using the stable upstream from Greg K-H? That's not what I thought and a little disappointing but I guess it makes sense. 2022-02-01 11:47:00 that fork adds extra features, but I believe most things work fine on stable upstream 2022-02-01 11:48:50 keep in mind that Alpine is probably the first distribution which using 5.15 kernel for rpi's which still isnt officialy "stable" one 2022-02-01 11:49:43 5.15 is marked 'longterm' on kernel.org 2022-02-01 11:49:46 alexhaydock: if it was mainstream, it would not be a separate -rpi kernel 2022-02-01 11:50:11 j`ey: Im talking about https://github.com/raspberrypi/linux 2022-02-01 11:50:35 MY-R: ah 2022-02-01 11:52:46 So what we do is we take the upstream kernel, and then apply a rolled-up rpi alpine patch 2022-02-01 11:54:44 That also makes sense. Thanks all! I guess my next thing to research will be what the criteria are for package promotion from edge to stable. It seems like the linux-rpi package does update fairly frequently in edge but then most of those updates don't make it to the stable branch (there have been 4 revisions published to edge in-between the current version in 3.15 and the current version in edge). 2022-02-01 11:54:47 ikke can i specify multiple ports as well? the wiki only mentions one port at a time :) 2022-02-01 11:54:55 jhjacobs81: yes, it accepts a list 2022-02-01 11:54:56 Alternatively, maybe I should think about just running my Pi deployments on edge 2022-02-01 11:55:15 alexhaydock: do you need the most recent kernels that often? 2022-02-01 11:55:28 nice :) 2022-02-01 11:56:03 jhjacobs81: an example: https://gitlab.alpinelinux.org/alpine/awall/-/blob/master/mandatory/services.json#L10 2022-02-01 11:57:32 Not particularly, j`ey, I'm more interested in just making sure the security patches from upstream (which is all patches, if you listen to Linus and Greg K-H) make it into whatever I'm running in a timely manner 2022-02-01 11:57:53 I care more about the patch level than the actual kernel version to be honest 2022-02-01 11:59:06 ikke i had this: https://pastebin.com/xck044Qj 2022-02-01 11:59:18 alexhaydock: fair 2022-02-01 11:59:49 It needs to be a json array 2022-02-01 12:01:39 ok, i think i got it :) 2022-02-01 12:07:02 ikke so this would be correct? https://pastebin.com/NiU4Bv1G 2022-02-01 12:08:43 well.. no.. 2022-02-01 12:08:53 it apears im getting to learn json as well :) 2022-02-01 12:19:43 interesting when did this change https://wiki.alpinelinux.org/wiki/Setting_up_disks_manually#For_.22diskless.22_and_.22data.22_disk_mode_installs_2 2022-02-01 12:20:00 i seem to remember alpine-linux being one volume FAT32 on mbr partition 2022-02-01 12:20:52 ie 2022-02-01 12:20:53 1 1049kB 7759MB 7758MB primary fat32 boot, esp 2022-02-01 12:21:15 might be why secureboot wouldn't work for me 2022-02-01 12:33:11 https://wiki.alpinelinux.org/wiki/Installation#Diskless_Mode oh might have said data mode 2022-02-01 12:45:27 well, i got the multi port thingy down.. now im trying to forward them 2022-02-01 13:23:39 hi guys! i want to auto-install bare metal machines using alpine, the goal is to make a minimal installation that will be configured further using ansible. 2022-02-01 13:23:39 i've written an answerfile for alpine-setup that seems to work, but i'm not sure what would be the best way to proceed from here. the most prominent problem is that setup-alpine still asks for a root password interactively 2022-02-01 13:25:15 but secondary questions linger around too, like what would be the best way to get the answerfile onto the system - wget it before running setup-alpine by hand or remaster the boot medium to include it, etc. 2022-02-01 13:26:24 i did this (without ansible for school) 1: make a iso with my scripts, 2: mount it and run it 2022-02-01 13:26:34 as for the root passwd i used this: yes | setup-alpine -f ./answer 2022-02-01 13:26:48 and then generate a random one in the script aft erhte setup part 2022-02-01 13:38:14 Misthios: so you remastered the release media to include your scripts? 2022-02-01 13:38:29 no, just added them to an iso which i mounted 2022-02-01 13:38:33 second iso 2022-02-01 13:41:36 i see 2022-02-01 13:42:02 thanks for your input :) 2022-02-01 13:42:06 np 2022-02-01 13:46:55 *sigh* 2022-02-01 14:03:06 mogus 2022-02-01 14:03:13 Whoops, wrong channel 2022-02-01 15:01:37 @e3t: I'm using wireplumber without modifying the pipewire conf to launch and it's been working well 2022-02-01 19:28:25 hello 2022-02-01 19:28:28 it feel good 2022-02-01 20:08:28 alexhaydock: RPI Foundation still only support kernel 5.10.x. They are apparently "intending" to move to 5.15.x in the next month or two 2022-02-01 20:09:42 Alpine is using Foundation's 5.15.x branch which is what Foundation still currently use as an "experimental" next version kernel and therefore there is a risk that 5.15.x could break in any particular version. 2022-02-01 20:10:12 Foundation have indicated they do not intend to move beyond 5.15.x (as it is LTS) for the forseeable future 2022-02-01 20:10:43 https://gitlab.alpinelinux.org/alpine/aports/-/issues/13420 2022-02-01 20:10:49 Would this change in login providers explain a standard user not seeing a local device like a microSD card on their XFCE desktop whereas logging in as root shows the expected volumes on the XFCE desktop? 2022-02-01 20:10:49 What would be a positive test I could perform to verify that this recent change in behavior is due to the login provider changing? 2022-02-01 20:11:02 I dont even know if they upstream their fixes? 2022-02-01 20:12:15 j`ey: RPI stuff does get added to upstream but I think some/most of it is not by Foundation themselves 2022-02-01 20:12:32 yeah I dont remember seeing stuff from them directly :/ 2022-02-01 20:12:53 I've seen in their own repo that they revert some upstream RPI changes and then reapply their "old way" of doing that functionality 2022-02-01 20:13:07 nice :/ 2022-02-01 20:13:35 so unfortunately I don't think the diff between Foundation and upstream is shrinking, I think its growing 2022-02-01 20:19:56 :\ 2022-02-01 20:20:54 minimal: you told me to check this CONFIG_EARLY_PRINTK but it is already "y" or should I disable it? 2022-02-01 20:22:06 MY-R: I don't think having that set has any actual effect unless an option (early-printf??) is passed via cmdline to enable it. I'd have to check 2022-02-01 20:23:07 then it should be some information about I guess 2022-02-01 20:25:30 MY-R: looks like "earlyprintk" is needed on cmdline to activate it. No sure if it will help or not, was just a suggestion. I did place around with another cmdline setting a few days ago to increase the size of the log_buf but it didn't make any apparent difference to missing earliy entries in "dmesg" 2022-02-01 20:25:34 minimal: the problem is I see some messages in console before see when screen switching to kms and those informations are already cut off from dmesg, for me it is kind of kernel bug or maybe some option missing in alpine 2022-02-01 20:26:08 sugestions are always welcome, is ok :) 2022-02-01 20:26:15 minimal meant to say: MY-R: looks like "earlyprintk" is needed on cmdline to activate it. No sure if it will help or not, was just a suggestion. I did play around with another cmdline setting a few days ago to increase the size of the log_buf but it didn't make any apparent difference to missing earliy entries in "dmesg" 2022-02-01 20:26:15 s/place/play/ 2022-02-01 20:26:52 because ye dmesg doesnt make sense if it already start from line: [ 0.336564] ... 2022-02-01 20:28:08 yupe. Sometimes I've notice the /var/log/dmesg file (when enabled) does contain a few earlier entries (as "dmesg" output the buffer which is a max size so over time earlier stuff gets dropped off it) 2022-02-01 20:30:40 minimal: do you think "earlyprintk" is enough or should I put some dunno earlyprintk=keep ? 2022-02-01 20:30:52 "Append ,keep in order not to disable the messages shown by this option when the real kernel console is initialized and takes over the system. " 2022-02-01 20:31:58 hello brothers and sisters 2022-02-01 20:32:42 MY-R: not sure, I guess you'd have to experiment :-) 2022-02-01 20:32:45 minimal: funny thing because I cant even see which kernel command line is ignored by kernel since I dont see it :D 2022-02-01 20:33:19 minimal: ye exactly, but ye will try few reboot later probably but thank you for the tips :) 2022-02-01 20:33:22 MY-R: for either Syslinux or Grub you can edit the cmdline at boot time 2022-02-01 20:34:12 minimal: ye I know but there are kernel command lines which are for kernel and another for user space and that information is at begining of dmesg :) 2022-02-01 20:34:29 would thee mind telling me which os art thee using ? 2022-02-01 20:34:52 minimal: it looks something like that: Unknown kernel command line parameters "BOOT_IMAGE=/boot/vmlinuz-5.15.16-gentoo-v0 nmi_watchdog=0", will be passed to user space. 2022-02-01 20:35:08 minimal: so then I can be sure that all rest command line pass correct to kernel :) 2022-02-01 20:39:58 MY-R: yeah I've seen such messages, they'd just informational - then is a later console/dmesg entry when initramfs is loaded that indicates which cmdline params are passed to it 2022-02-01 20:41:59 minimal: yeah, in my case are late but too early to see them :) 2022-02-01 20:43:30 in gentoo drmsimple stuff is almost disabled and since using nvidia I dont care about kms stuff, but in home router got Alpine with amd gpu and it is annoying :/ 2022-02-01 20:45:02 replaced rpi4 to some x86_64 board with pcie slot to put ther some normal intel nic 2022-02-02 00:24:23 I still can't figure out how to make Alpine mount root and home with my own flags 2022-02-02 00:24:54 It seems to be ignoring fstab which has the flag. But it _is_ using fstab, because that's the only place /home is defined 2022-02-02 00:29:06 Are there any *nice* ways of navigating logs (filtering by service, etc)? 2022-02-02 00:37:15 Nulo: for SSD-based machines I do add "commit-300" to the mount options in fstab and have seen, via "mount" that indeed the option is picked up 2022-02-02 00:37:37 s/-/=/ 2022-02-02 00:37:37 minimal meant to say: Nulo: for SSD=based machines I do add "commit-300" to the mount options in fstab and have seen, via "mount" that indeed the option is picked up 2022-02-02 00:37:57 argh, "commit=300" 2022-02-02 00:38:48 minimal: mount as in the command? 2022-02-02 00:38:55 I would like it to be done on boot 2022-02-02 00:40:08 Nulo: it *is* in my case done at boot for rootfs 2022-02-02 05:26:09 hey alpine, when using the setup-bootable script to make a writable + diskless install media, is there any limitation to what filesystem I can use? I notice that on non fat32 drives syslinux will fail by default but could I use a btrfs formatted usb stick and grub? 2022-02-02 16:20:01 Is there a mechanism to unflag flagged packages at pkgs.alpinelinux.org ? 2022-02-02 16:21:36 no 2022-02-02 16:21:42 except upgrading it 2022-02-02 16:25:49 Okay 2022-02-02 19:52:36 I've been trying to clean the flagged and Maintained by None list for the last days... 2022-02-02 19:52:46 but some should just be cleaned 2022-02-02 19:53:40 Androw: if you can give a list, I'll see if I can manually clear them 2022-02-02 19:55:48 https://pkgs.alpinelinux.org/flagged?origin=&maintainer=None (sorry for the obvious link): according to me, only the 2 most recents are really outdated 2022-02-02 19:58:32 I'm trying to use setup-disk for mdadm/crypt/sys install with two 4TB disks but it complains because disk are too big for dos partition table 2022-02-02 19:58:42 can I tell it to use gpt instead? 2022-02-02 20:01:02 donoban: have a look at setup-disk, its a shell script. From a quick glance looks like you can do DISKLABEL="gpt" setup-disk 2022-02-02 20:01:29 yeah I was reading it but it's pretty confusing and I'm prety tired hehe 2022-02-02 20:04:33 something failed, probably due my other attemps 2022-02-02 20:05:00 I will reboot and try from start, thanks minimal 2022-02-02 20:06:40 lol, donoban says he's going to reboot, but minimal leaves :D 2022-02-02 20:08:58 :P 2022-02-02 21:46:07 hey 2022-02-02 23:28:10 can someone suggest a simple CLI-only kvm virtual machine manager without a ton of dependencies. (not virsh) 2022-02-02 23:28:18 thanks 2022-02-02 23:39:49 qemu+openrc is quite usable 2022-02-02 23:40:09 as in, hand-written openrc scripts 2022-02-02 23:46:33 quemu's inteface is a bit too verbose, but that might be the plan B 2022-02-03 02:05:32 praise psykose , fastest feedback I've ever had on an MR :D 2022-02-03 02:05:58 :3 2022-02-03 02:56:40 Here's a fun one: https://0x0.st/oHUX.json 2022-02-03 02:56:59 My machine froze after suspending, I got a dmesg out of it. 2022-02-03 02:57:11 Should I report this to upstream Linux? How? 2022-02-03 03:01:57 you can try but i suspect with minimal results 2022-02-03 03:02:19 if you're using xf86-video-intel switch to xf86-video-modesetting 2022-02-03 03:08:56 Hello71, I have... neither. I use Wayland 2022-02-03 03:09:29 do you have i915 firmware loaded 2022-02-03 03:10:04 Yep.. I think, how do I check that? 2022-02-03 03:10:13 i915 module is loaded 2022-02-03 03:10:41 Ah, I forgot I do this terrible thing export MESA_LOADER_DRIVER_OVERRIDE=crocus 2022-02-03 03:12:34 It's to avoid using the i965 mesa driver 2022-02-03 03:22:42 Nulo: maybe it is triggering some kernel bug 2022-02-03 03:23:46 you can report it but i think intel graphics devs are not focused on fixing poorly reproducible bugs 2022-02-03 03:29:33 Is anyone using Parole+gstreamer for media playback? Can you play x264/h264/mp4 content at all? Parole refuses for me 2022-02-03 03:29:51 I feel like maybe I'm missing a demuxer but it's really quite mystifying 2022-02-03 03:29:57 what's the error 2022-02-03 03:30:10 Hello71, any clue how? I never reported a kernel bug before ^^ 2022-02-03 03:31:46 Error: GStreamer backend error. Internal data stream error. 2022-02-03 03:32:02 Nothing prints to console, likely because our build config for parole has all debugging off/minimal 2022-02-03 03:33:04 ACTION sent a code block: https://matrix.org/_matrix/media/r0/download/matrix.org/bkLTqmsqEkiGuilfoQlxTaIf 2022-02-03 03:33:24 mm 2022-02-03 03:33:25 no idea 2022-02-03 03:33:47 100% vanilla file, plays on Parole on Debian-deriveds and works flawlessly in shotcut and other things 2022-02-03 03:34:00 Yeah, this one has been kicking my butt for months now 2022-02-03 03:34:28 Dig throguh the gstreamer build logs to see if we're not building all the codecs? 2022-02-03 03:34:43 Parole just crashes for me because of invalid cast from 'GdkWaylandDisplay' to 'GdkX11Display' pfft 2022-02-03 03:35:07 I'm sorry to hear that. I'm XFCE, so purely X11 for now as far as I know 2022-02-03 03:35:15 It did work when I was using GNOME/GDM, though 2022-02-03 03:35:22 And that forced Wayland 2022-02-03 03:35:56 No clue, unsetting WAYLAND_DISPLAY to force it to use Xwayland works. Anyway, after installing gst-libav it works for me 2022-02-03 03:36:05 But maybe you're not interested in gst-libav.. 2022-02-03 03:36:19 Why not? 2022-02-03 03:36:25 I'm not sure what it does 2022-02-03 03:36:54 I flagged our parole package since upstream has 4.16.x builds out for a while now. Maybe it'll play nicer with Wayland? 2022-02-03 03:37:05 It uses libav as a backend, which supports a bunch of codecs (AFAIK it's a fork of FFMPEG? I thought it died? I honestly have no clue) 2022-02-03 03:37:45 ACTION sent a code block: https://matrix.org/_matrix/media/r0/download/matrix.org/wlmfEqYyWVnomSddJqpjKKHD 2022-02-03 03:37:47 Got it apparently 2022-02-03 03:38:58 Try running Parole with GST_PLUGIN_FEATURE_RANK=libav:MAX , I understand it should prioritize it over other 2022-02-03 03:43:41 Do you have a syntax for that? I can't seem to find docmentation on how to use that 2022-02-03 03:49:35 Saijin_Naib[m], https://gstreamer.freedesktop.org/documentation/gstreamer/running.html?gi-language=c#environment-variables 2022-02-03 03:50:59 Ahh, thanks 2022-02-03 03:54:52 Nulo: No dice, but thanks for the assistance anyway 2022-02-03 03:55:21 :( Maybe try with other gstreamer-based video players 2022-02-03 04:02:07 ACTION uploaded an image: (535KiB) < https://matrix.org/_matrix/media/r0/download/matrix.org/TsbwGXXdgZETjujoTGTZeMyU/image.png > 2022-02-03 04:02:23 Provided Totem 3.38 still builds against gstreamer, then yeah, it seems like just Parole is broken for some reason 2022-02-03 08:22:05 Sorry, quick question. My alpine initramfs doesn't want to mount /dev/nvme0n1p3 to /sysroot, claiming it doesn't exist. I can see the device with `ls` and `nvme` and `ext4` are in the /etc/mkinitfs/mkinitfs.conf file. Any clue what I'm missing? 2022-02-03 08:39:33 hmm i these udisks rules https://github.com/coldfix/udiskie/wiki/Permissions 2022-02-03 08:39:38 and polkit is running, as is lxpolkit etc 2022-02-03 08:39:51 though i am having trouble with udiskctl mount -b /dev/device 2022-02-03 08:40:07 Error looking up object for device /path/to/device 2022-02-03 08:40:17 works on my silverblue machine 2022-02-03 08:40:35 i would have thought 2022-02-03 08:40:36 "org.freedesktop.udisks2.filesystem-mount": YES, 2022-02-03 08:40:38 was sufficient 2022-02-03 08:41:10 i can obviously mount with "sudo mount /dev/device mountpoint" 2022-02-03 08:41:17 but that's not using polkit 2022-02-03 08:42:06 tried doing it with udisksctl monitor running and nothing much was revealed 2022-02-03 08:43:01 seems to be 2022-02-03 08:43:02 Operation: filesystem-mount 2022-02-03 08:43:07 that it is having trouble with 2022-02-03 08:43:17 maybe my polkit rules aren't being used 2022-02-03 08:43:34 perhaps dumping that in /etc/polkit-1/rules.d/50-udiskie.rules is not sufficient 2022-02-03 08:43:41 (using elogind) 2022-02-03 08:44:08 and polkitd user is running /usr/lib/polkitd --no-debug 2022-02-03 08:48:53 Huh, my probem is that the initramfs doesn't load ext4 module automatically. 2022-02-03 08:51:28 do you pass modules=ext4,... on the kernel cmdline 2022-02-03 08:51:30 seems to on my alpine linux machine 2022-02-03 08:52:46 and i would have thought ext4 was compiled into kernel 2022-02-03 08:54:22 though it might depend on which alpine linux image you're using 2022-02-03 08:55:45 does seem to be a module https://git.alpinelinux.org/aports/tree/main/linux-lts/config-lts.x86_64#n8411 2022-02-03 09:00:00 Pfft, I just needed the modules=ext4 line. Thanks 2022-02-03 09:00:22 I usually compile my own kernel with everything required to boot built-in and skip the entire initramfs so was a bit confused. Thanks. 2022-02-03 09:01:46 i'm too lazy to compile my own kernel lol 2022-02-03 09:01:57 i used to do it, a long time ago 2022-02-03 09:05:13 I got used to it, mainly running Gentoo. 2022-02-03 09:05:18 So the habit stuck 2022-02-03 09:46:18 Hello. Just here to say that I recently gave a try to alpine, I love it, and now I am working to deploy it in production 2022-02-03 10:17:00 so, has anyone else got encrypted usb stick working on alpine linux? 2022-02-03 10:17:09 Wdym? 2022-02-03 10:17:10 seems the decryption part works, but not mounting, ie in nautilus, thunar etc 2022-02-03 10:17:30 Do you mean a usb with alpine on it 2022-02-03 10:17:34 Or an encrypted usb 2022-02-03 10:17:38 for example when trying to mount with udiskctl -b /dev/mapper/luks-xxxxxxxxx 2022-02-03 10:17:46 i get "Error looking up object for device" 2022-02-03 10:17:51 but it works on my systemd machine 2022-02-03 10:18:16 What fs is it? 2022-02-03 10:19:08 ext4 2022-02-03 10:19:12 so it should work perfectly fine 2022-02-03 10:19:36 That’s weird then 2022-02-03 10:19:40 the second part there doesn't seem to work https://clbin.com/6iBl2 2022-02-03 10:19:46 Have u tried running it with root 2022-02-03 10:19:51 The mount command 2022-02-03 10:20:02 yes that works, and it works on my non-alpinelinux machine 2022-02-03 10:20:20 So the issues is you can’t mount it without root? 2022-02-03 10:20:26 udiskctl unlock works 2022-02-03 10:20:26 s/issues/issue/ 2022-02-03 10:20:26 guardant[m] meant to say: So the issue is you can’t mount it without root? 2022-02-03 10:20:40 guardant[m]: ccorrect 2022-02-03 10:21:07 Right 2022-02-03 10:21:47 Tbh I have no idea how to fix that 2022-02-03 10:21:53 Do non encrypted drives work? 2022-02-03 10:25:18 i should try that 2022-02-03 10:25:51 yup, so unencrypted devices mount 2022-02-03 10:26:23 Even weirder 2022-02-03 10:26:39 so basically if you have a USB stick, that's dmcrypted, you can udiskctl unlock it 2022-02-03 10:26:50 but then you can't udisksctl mount it 2022-02-03 10:27:04 but if you have no encryption on the partition you can straight udisksctl mount 2022-02-03 10:27:16 and its not a problem with the volume because sudo mount /dev/mapper/volume works 2022-02-03 10:27:53 How did you set up mounting drives as a user? 2022-02-03 10:28:38 well i installed elogind polkit-elogind 2022-02-03 10:28:50 sway is running with: dbus-run-session sway 2022-02-03 10:28:55 so dbus is working 2022-02-03 10:29:21 and my sway session runs /usr/bin/lxqt-policykit-agent 2022-02-03 10:29:33 Oh right 2022-02-03 10:29:38 additionally i added a storage group 2022-02-03 10:29:44 and added these polkit rules https://github.com/coldfix/udiskie/wiki/Permissions 2022-02-03 10:30:42 I’m really not sure what the issue is then 2022-02-03 11:06:30 Hi everybody, I can't seem to figure out how to set XDG_RUNTIME_DIR.. i installed and enabled elogind, dbus and dbus-openrc but it just doesnt show up. Loginctl says "No Sessions" 2022-02-03 11:07:08 You shouldn’t need to if you have installed and enabled elogind 2022-02-03 11:07:55 you probably want util-linux-login as well 2022-02-03 11:08:27 then the pam login will hook elogind to provide it 2022-02-03 11:08:43 you could also have no elogind, and also add pam-rundir, and manually add pam_rundir.so to base-session 2022-02-03 11:08:52 u can set it manually in ur .profile if u want? 2022-02-03 11:08:59 doesn't work the same 2022-02-03 11:09:20 whats the difference? 2022-02-03 11:09:37 does not clear on logout 2022-02-03 11:09:52 Have you looked at the wayland page 2022-02-03 11:09:57 It says how to set it 2022-02-03 11:17:00 "Hi everybody, I can't seem to..." <- need to set it as an environmental variable 2022-02-03 11:18:29 https://wiki.alpinelinux.org/wiki/Wayland 2022-02-03 11:18:32 Dngray did you ever fix your pipewire? 2022-02-03 11:18:40 yeah i fixed that 2022-02-03 11:18:53 What did you do to fix it? 2022-02-03 11:19:06 Got the same issue on my fresh install 2022-02-03 11:19:09 Followed the wiki 2022-02-03 11:19:11 i think it was i wasn't starting dbus correctly 2022-02-03 11:19:26 i did notice in the configuration tab it defaults to Analog audio 2022-02-03 11:19:37 and not "Pro Audio" 2022-02-03 11:19:46 Interesting 2022-02-03 11:20:02 I’ll have to try it later 2022-02-03 11:20:09 i think this udiskctl issue above might be a bug 2022-02-03 11:20:22 because the rules are being loaded, no reason it shouldn't work 2022-02-03 11:20:23 I think so 2022-02-03 11:20:32 Unless you need another rule 2022-02-03 11:20:34 guardant[m]: do you already fix it? I removed eloginctl and lost XDG_RUNTIME_DIR 2022-02-03 11:20:38 For encrypted drives 2022-02-03 11:20:50 ? 2022-02-03 11:20:54 yeah maybe 2022-02-03 11:21:06 sry i accidentally rebooted my own computer instead of ssh. but yeah thanks a lot util-linux-login was exactly what i wanted 2022-02-03 11:21:06 although on my silverblue machine i don't 2022-02-03 11:21:08 for now I just have a script in /etc/profile.d 2022-02-03 11:21:23 http://ix.io/3Owb 2022-02-03 11:21:26 donoban: i just do this https://wiki.alpinelinux.org/wiki/Wayland 2022-02-03 11:21:35 it doesnt work with waypipe 2022-02-03 11:21:52 ah lol, same solution 2022-02-03 11:22:08 Are you trying to set up sway? 2022-02-03 11:22:20 Make sure you follow this completely: 2022-02-03 11:22:22 those rules are pretty much https://src.fedoraproject.org/rpms/udiskie/blob/rawhide/f/50-udiskie.rules the same as the ones in https://github.com/coldfix/udiskie/wiki/Permissions 2022-02-03 11:22:23 https://wiki.alpinelinux.org/wiki/Sway 2022-02-03 11:22:34 so the fact it works on silverblue but not alpine makes me think there's not an issue with the rules 2022-02-03 11:22:38 I have already working 2022-02-03 11:22:46 What’s the problem then? 2022-02-03 11:22:48 I was using elogind and decided to remove it 2022-02-03 11:22:50 ah 2022-02-03 11:23:05 no problem here 2022-02-03 11:23:20 and its not a udiskie bug because it should work with udisksctl 2022-02-03 11:24:08 loginctl still says no sessions and no users though 2022-02-03 11:24:09 I wonder who should init that env var, the login manager? maybe seatd? (since it seems that it was done by elogind before) 2022-02-03 11:24:29 or just add it to profile 2022-02-03 11:45:53 Hi. How this possible. When getting an update for alpinelinux through squid proxy with c-icap and clamav module - clamav detects 'PUA.Win.Exploit.CVE_2012_1461-1(7c71d608edd54a547359da399bf1301f:1661728) FOUND' in APKINDEX.tar.gz file ? 2022-02-03 11:46:55 false alarm ? 2022-02-03 11:47:04 https://askubuntu.com/questions/1006237/clamav-detects-trojans 2022-02-03 11:48:25 There are no applications in APKINDEX files. They just contain metadata 2022-02-03 11:49:09 ikke, got it. 2022-02-03 11:57:05 ikke, thank you. correct reason. 2022-02-03 12:25:47 is there a way to change XDG_RUNTIME_DIR from /tmp/1000-runtime-dir to /run/user/UID when using elogind? 2022-02-03 12:27:51 add util-linux-login and it will probably work 2022-02-03 12:28:00 you disconnected before :p 2022-02-03 12:28:26 thats what i did 2022-02-03 12:28:32 and it set that variable to /tmp 2022-02-03 12:29:00 are you sure you don't export some magic stuff in your profile 2022-02-03 12:29:16 my profile is pretty empty, but i'll check it again... 2022-02-03 12:29:40 and in the /etc/profile* hierarchy potentially too 2022-02-03 12:29:49 sometimes one throws stuff in there and forgets.. been there 2022-02-03 12:31:04 oh yeah, thats absolutely correct, i forgot to remove script from alpine wiki 2022-02-03 12:31:28 i used to do that too but it does not really work correctly after you log out and back in again 2022-02-03 12:31:29 wait, its just not set now again 2022-02-03 12:31:34 mm 2022-02-03 12:31:44 did you relog 2022-02-03 12:31:47 yea 2022-02-03 12:31:49 i rebooted 2022-02-03 12:32:21 if you want the most definite way just add `pam-rundir` and `session required pam_rundir.so` to /etc/pam.d/base-session 2022-02-03 12:32:25 though it should work already 2022-02-03 12:33:03 also note if you forget to add that package `required` means you will not be able to log in ever without it :p 2022-02-03 12:33:09 so perhaps optional is better 2022-02-03 12:33:17 isnt that what the -prefix for though 2022-02-03 12:33:23 maybe 2022-02-03 12:35:46 i have added session required pam_rundir.so to /etc/pam.d/base-session, rebooted, and its still not defined 2022-02-03 12:36:31 are you sure you added util-linux-login.. cause i can't imagine that not working at all 2022-02-03 12:36:32 hm 2022-02-03 12:36:40 weirdly, all other XDG_ variables disappeared after i removed that script from alpike wiki 2022-02-03 13:04:37 something goes wrong with setup-disk and crypt/raid1, at least with gpt. After mounting cryptroot on /mnt, it fails to mount boot partition on /mnt/boot 2022-02-03 13:18:16 you know what is interesting, when i udisksctl unlock -b /dev/sdb1 on alpine linux 2022-02-03 13:18:29 in udisksctl monitor 2022-02-03 13:18:36 it says that the prefered device is /dev/dm-0 2022-02-03 13:18:43 not /dev/mapper/luks-xxxxx-xxx-xxx 2022-02-03 13:18:52 which it does on my silverblue machine 2022-02-03 13:20:31 it does create a symlink in /dev/mapper though 2022-02-03 13:20:58 well, that could be improved but at least start with a working solution 2022-02-03 13:25:05 Are there any resources for hardening alpine? In particular I'm interested in what people do with LSMs 2022-02-03 13:31:27 sm2n: there are apparmor packages in community 2022-02-03 13:36:06 minimal: Do people recommend it? 2022-02-03 13:37:00 "people"? 2022-02-03 13:37:36 anyone who has written about it 2022-02-03 13:38:00 I'm just trying to understand the area 2022-02-03 13:38:02 I don't know if anyone has written about it 2022-02-03 13:38:23 okay then 2022-02-03 13:38:36 the profiles haven't been adapted for Alpine though, they are just the default upstream profiles 2022-02-03 13:39:05 but LSMs are not the only way to harden a system 2022-02-03 13:40:55 I already do traditional user-based privsep, I'm curious what else is available though 2022-02-03 13:45:46 sm2n: that's not really an Alpine-specific question though, more of a generic Linux one 2022-02-03 13:49:33 Yeah, fair enough. 2022-02-03 13:52:20 sm2n: there's more to hardening that just LSMs - firewall rules, SSHd config, sysctl settings, etc 2022-02-03 14:00:11 i posted there https://github.com/coldfix/udiskie/issues/220#issuecomment-1029018750 2022-02-03 14:00:16 i wonder if someone else could reproduce this 2022-02-03 14:00:37 yeah, of course... I just realized that while I've done the other things, I've never really used LSMs, and I wanted to know what alpine had available 2022-02-03 14:00:41 get a usb stick, gpt, luks format, then format the single partition as ext4 2022-02-03 14:01:24 and try to unmount it with `udisksctl unlock -b /dev/device` and `/dev/mapper/xxx` 2022-02-03 14:02:01 you'll need polkit, and a polkit agent running, and dbus that should be it 2022-02-03 14:02:31 I tried other filesystems fat32 and same result 2022-02-03 14:03:01 bleh that should be `udisksctl mount -b /dev/mapper/xxxx` 2022-02-03 14:04:34 if someone wants to test this for me, who uses alpine linux as their regular desktop i'd much appreciate it 2022-02-03 14:11:36 you know what's also interesting there 2022-02-03 14:11:43 a lot of details are missing on alpinelinux 2022-02-03 14:12:01 for example 2022-02-03 14:12:10 Id: by-id-dm-name-luks-ce782db4-f9ee-4154-80e1-72c0d71b08da 2022-02-03 14:12:15 shows up on fedora 2022-02-03 14:12:47 if you compare 4. and 11. https://github.com/coldfix/udiskie/issues/220#issuecomment-1029018750 2022-02-03 14:15:21 dngray: are you using eudev or mdev? I have a PR to submit to mdev to add some missing /dev/disk/by-id entries (including for dm devices) 2022-02-03 14:16:06 i was using eudev 2022-02-03 14:17:23 i wonder if i should try with mdev 2022-02-03 14:17:34 what's generally preferred? 2022-02-03 14:18:45 https://www.linuxquestions.org/questions/slackware-14/what-is-going-to-happen-now-that-eudev-is-not-going-to-be-mantained-4175700098/ oh kinda sounds like eudev isn't maintained 2022-02-03 14:18:54 eudev is more functional/feature filled 2022-02-03 14:19:26 dngray: not true, several people including Ariadne maintain eudev 2022-02-03 14:19:42 oh. 2022-02-03 14:19:51 so that thread was probably wrong, figured as much 2022-02-03 14:20:25 the /dev/disk/ missing entries thing is with mdev, not eudev 2022-02-03 14:21:53 yeah so that's not the issue 2022-02-03 14:22:33 well the issue seems to be `udisksctl mount` doesn't seem to work with eudev 2022-02-03 14:22:40 wonder if it works with mdev 2022-02-03 14:23:00 that is assuming you have a device, that has a with a luks encrypted partition 2022-02-03 14:23:06 works if it's not luks encrypted 2022-02-03 14:24:40 dngray: well if it requires those /dev/disk/ entries then obviously it won't (yet) work with mdev 2022-02-03 14:24:51 yeah, maybe so 2022-02-03 14:24:59 should work with mdev though 2022-02-03 14:25:14 dngray? eh? ^^^ 2022-02-03 14:25:48 err sorry eudev 2022-02-03 14:25:52 is what i meant to write lol 2022-02-03 14:26:21 I've never used udiskie so not familiar with it in general 2022-02-03 14:26:31 its not udiskie 2022-02-03 14:26:42 it's udisksctl which is part of udisk 2022-02-03 14:27:12 incidentally mounting in nautilus, thunar, etc doesn't work either 2022-02-03 14:27:21 you click on it, you get the password prompt to decrypt 2022-02-03 14:27:31 and as soon as you enter password, the volume disappears 2022-02-03 14:29:37 i got a feeling its got something to do with the missing Id, IdType, IdUUID, IdUsage, IdVersion, Symlinks 2022-02-03 14:36:33 i wonder if i should test on artix 2022-02-03 14:36:36 to see if it works there 2022-02-03 14:36:53 that also uses eudev/elogind 2022-02-03 17:10:34 dngray: It sounds like you got hit with a similar recent change like I have. I can no longer see my microSDHC nor regular filesystem mounts in XFCE as my normal user (with being added to storage and a ton of other groups), but can operate like normal when logged in as root 2022-02-03 17:10:42 I was sure I saw a script thta was doing some bootstrapping of alpine, but I cant find it anymore 2022-02-03 17:10:58 This happened in the past two weeks. Thought maybe it was a side-effect of the recent util-linux change, but I was directed towards udisks intstead 2022-02-03 17:11:32 Also using eudev/xfce polkit, lightdm, XFCE+Thunar full DE 2022-02-03 17:13:39 j`ey: my script for creating Alpine server disk images? 2022-02-03 17:14:14 or one of the other chroot or vm image creation scripts? 2022-02-03 17:15:00 I think it was on https://git.alpinelinux.org/ somewhere 2022-02-03 17:15:29 https://gitlab.alpinelinux.org/alpine/aports/-/blob/master/scripts/bootstrap.sh\ 2022-02-03 17:15:31 ? 2022-02-03 17:15:48 ikke: thanks! 2022-02-03 17:16:28 I was looking in abuild and somewhere else, but missed this 2022-02-03 17:23:03 is `abuild -r` run from the dir of a pkg? 2022-02-03 17:23:32 yes 2022-02-03 17:24:54 thanks, I was trying to follow abuild.in but got a little confused 2022-02-03 17:26:42 That's referred to as STARTDIR 2022-02-03 17:50:44 https://git.alpinelinux.org/abuild/tree/abuild.in#n743 trying to understand where this targz() function is used 2022-02-03 18:27:42 so abuild -r also installs the package? 2022-02-03 18:27:57 No 2022-02-03 18:28:22 It places the package in ~/packages// 2022-02-03 18:29:20 ikke: did you know about the earlier question, wrt 'targz()'? 2022-02-03 18:30:02 Looks like something you can invoke manually 2022-02-03 18:31:12 hm, so what format are they in, in ~/packages//? 2022-02-03 18:31:26 j`ey: *.apk files 2022-02-03 18:31:29 apk packages 2022-02-03 18:31:40 ok, so thats done via create_apks() 2022-02-03 18:31:43 It's a proper repo you can add to /etc/apk/repositories (without the part) 2022-02-03 18:32:15 Yes, though in the future apk itself will handle that part 2022-02-03 18:33:59 ah REPODEST in abuild 2022-02-03 18:34:42 ikke: thanks for all the help btw 2022-02-03 18:35:06 j`ey: which is set in /etc/abuild.conf 2022-02-03 18:35:18 Im trying to understand this line now https://git.alpinelinux.org/aports/tree/scripts/bootstrap.sh#n97 2022-02-03 18:35:33 it builds a cross gcc, so I'm trying to understand how that gets used 2022-02-03 18:36:06 I must say, I have never actually used it myself 2022-02-03 18:36:20 but the goal is to bootstrap a native compile toolchain for a new arch 2022-02-03 18:36:46 It's not meant to create a generic cross-compile toolchain 2022-02-03 18:37:26 I see 2022-02-03 18:51:49 https://git.alpinelinux.org/abuild/tree/functions.sh.in#n160 is where it seems to set a cross compiler 2022-02-03 18:53:24 That just sets up the environment variables to use one 2022-02-03 18:54:31 yeah, I was trying to find where it actually set that, couldnt find it in abuild.n 2022-02-03 18:54:55 ah ok 2022-02-03 18:56:18 followin the maze, trying to understand :) 2022-02-03 19:03:31 If I want to update the email for my maintainer/contributor lines across the few APKBUILD I maintain, should that be a single MR/Commit? What would that commit message look like 2022-02-03 19:03:39 COMMITSTYLE.md is down for the Gitlab upgrade it seems haha 2022-02-03 19:05:57 Saijin_Naib[m]: one MR is fine, typically one commit per package 2022-02-03 19:06:18 ': update maintainer e-mail' or something like that 2022-02-03 19:06:24 Saijin_Naib[m]: and gitlab is up again 2022-02-03 19:06:40 ikke: Thank you. Perfect, just going to ask if I should just follow the generic aport change rules 2022-02-03 19:06:46 And looks like that answer is yes 2022-02-03 19:06:58 Saijin_Naib[m]: don't forget to bump pkgrel for maintainer changes 2022-02-03 19:07:36 Oh, whoops. That was not going to happen. Thanks for the save 2022-02-03 19:07:57 Someone would've told you :-) 2022-02-03 19:10:15 bump it thrice to make sure 2022-02-03 19:10:28 -r99 2022-02-03 19:16:59 Hello, where can I read policies on a) what goes from testing to main or community repo; b) what goes to the next release? 2022-02-03 19:18:17 ikke: Probably, but I'm trying to cut down on how many antacids people need for my MRs 😛 2022-02-03 19:18:40 Ermine: We are working on improving documentation about these kinds of thing 2022-02-03 19:18:42 thigns 2022-02-03 19:18:50 Ermine: some of it is documented on the wiki 2022-02-03 19:18:58 but most of it goes by worth-of-mouth atm 2022-02-03 19:19:28 thighs 2022-02-03 19:19:29 Ermine: The policy for main is that we support packages in main for 2 years (4 releases) 2022-02-03 19:19:36 *sticks irc logs on the wiki* solved 2022-02-03 19:19:42 very good plan 2022-02-03 19:20:21 Ermine: community is supported for 6 months (1 release) 2022-02-03 19:20:48 Ermine: the goal of the testing repository is / was to give a chance for people to actually test the built packages 2022-02-03 19:21:11 there is no set time or procedure to move things out of testing, though packages should generally be in shape 2022-02-03 19:22:09 Ermine: so anything in main or community would be part of the next release 2022-02-03 19:23:43 ikke: got it, thank you very much! 2022-02-03 19:26:25 https://gitlab.alpinelinux.org/alpine/aports/-/merge_requests/30433 2022-02-03 19:26:30 If anyone has a moment to verify I did this one right 2022-02-03 19:27:01 missed pkgrel on xiccd 2022-02-03 19:27:09 Wait, so would the pkgrel for that be 1_gitwhatever? 2022-02-03 19:27:21 no there is still a pkgrel= 2022-02-03 19:27:26 just like the others 2022-02-03 19:28:04 yep, sorry. Got confused apparently 2022-02-03 19:29:54 you have two commits for the last change 2022-02-03 19:30:02 Awh yiss. Even rebased it without nuking the branch and starting over 2022-02-03 19:30:04 Refresh again? 2022-02-03 19:30:40 now your xiccd change doesn't update the email 2022-02-03 19:31:09 ahhh... I'm so bad at this 2022-02-03 19:32:42 now you have two commits again 2022-02-03 19:33:07 and fixed 2022-02-03 19:33:08 Refresh 2022-02-03 19:33:23 Sweet 2022-02-03 19:33:31 now you missed the pkgrel again 2022-02-03 19:33:39 Not that anyone uses what I maintain, but it's probably important the email actually gets to me in case 😛 2022-02-03 19:33:48 oh my god, you're kidding me. I didn't 2022-02-03 19:33:51 🤭 2022-02-03 19:33:53 -_- 2022-02-03 19:34:11 I can't with me rihgt now 2022-02-03 19:34:12 I'm so sorry 2022-02-03 19:36:33 Okay I fundamentally do not understand pick/drop apparently 2022-02-03 19:36:47 you shouldn't use either here 2022-02-03 19:36:55 drop would dismiss the entire commit 2022-02-03 19:37:00 Oh 2022-02-03 19:37:28 either manual --amend of last commit, or --fixup of the commit and rebase -i, or rebase and edit the commit and rebase --continue, or... 2022-02-03 19:37:33 there's like 50 ways to skin this cat 2022-02-03 19:37:51 Or reset --mixed to master, stage appropratiely, commit, repeat (without reset) 2022-02-03 19:38:10 These are all things I've not used before 2022-02-03 19:38:25 Can I use squash on the last commit to merge the pgkrel and email changes 2022-02-03 19:38:44 fixup is what you would use to not change the message 2022-02-03 19:38:45 but yeah 2022-02-03 19:38:51 same mechanic 2022-02-03 19:39:05 If i happned to put the same message, it'd have the same result though? 2022-02-03 19:39:09 yes 2022-02-03 19:39:11 Alright, let me redo it for the 50th time now 2022-02-03 19:39:12 Thanks! 2022-02-03 19:39:19 squash is a fixup but just prompts you for the message info again 2022-02-03 19:39:24 fixup uses the top one 2022-02-03 19:40:47 pick c400f21eb5 testing/lsd: update contributor/maintainer emails 2022-02-03 19:40:47 pick 5fc7d97974 testing/xiccd: update maintainer email 2022-02-03 19:40:47 fixup 9346703b10 testing/xiccd: update maintainer email 2022-02-03 19:41:10 yes 2022-02-03 19:41:48 12 pipeline re-runs 😬 2022-02-03 19:42:05 thank you both so much haha 2022-02-03 19:52:25 Does anyone use sccache vs ccache? 2022-02-03 19:59:34 i use the former just for the rust caching 2022-02-03 19:59:39 aside from that it's... worse 2022-02-03 19:59:52 mostly cause there are no offline stored stats so you get no cool screenshot after running it for months 2022-02-03 19:59:53 :p 2022-02-03 20:00:10 lol 2022-02-03 20:00:22 don't tell me you don't like cool stats... 2022-02-03 20:00:36 tragedy.. 2022-02-03 20:00:39 Saijin_Naib: this seems relevant https://github.com/mozilla/sccache/issues/160 2022-02-03 20:00:50 Can they live side-by-side? Yeah, I was looking at that yesterday 2022-02-03 20:01:08 I want something to cache Rust builds and I was hoping I could just keep ccache around for C/CUDA/etc stuff 2022-02-03 20:01:36 I'm trying to build Popsicle USB imaging tool by System76 and it takes for-freaking-ever. A Rust cache I think might be nice for that one 2022-02-03 20:01:43 psykose: like https://i.imgur.com/gndwnHN.png? 2022-02-03 20:02:16 ooh 2022-02-03 20:02:17 pretty.. 2022-02-03 20:02:37 you can sure, the default sccache behaviour with RUSTC_WRAPPER exported will be rust only 2022-02-03 20:02:52 aside from the few packages that check for sccache explicitly and do weird shit 2022-02-03 20:08:12 That weird shit part is scary 2022-02-03 20:11:07 Does this apk world look about as minimal as one can expect for building stuff (minus my three nice-to-haves of afetch,ccache,lsd)? 2022-02-03 20:11:11 ACTION sent a code block: https://matrix.org/_matrix/media/r0/download/matrix.org/bzJaHCRbuVrCFQiuVZYPOXhI 2022-02-03 20:11:44 don't think it matters how minimal the world is 2022-02-03 20:11:54 I want to avoid having stuff hanging around that might be depends/makedepends for stuff I'm building so I don't miss stuff 2022-02-03 20:12:01 Oh, okay. Good to know 2022-02-03 20:12:04 mm 2022-02-03 20:12:05 abuild rootbuild 2022-02-03 20:12:13 you probably want `libffi-dev musl-dev` depending on what you're planning to build 2022-02-03 20:12:13 that is usually just -dev and some specific libs 2022-02-03 20:12:33 I'm cool if abuild pulls that in each time 2022-02-03 20:12:35 musl-dev is already covered by build-base 2022-02-03 20:12:45 and libffi-dev should be listed in makedepends if necessary 2022-02-03 20:12:59 I don't see build-base in the example world 2022-02-03 20:13:07 alpine-sdk is git + abuild + build-base 2022-02-03 20:13:13 setting up the apk cache is nice as well 2022-02-03 20:13:16 ah, ty for clarifying 2022-02-03 20:14:06 So this looks reasonably minimal to folks then? I missed some stuff in flatseal since my world apparently had too much at the time (was a full GUI DE environment at the time) 2022-02-03 20:15:19 you'll know if you're missing something when compilation fails :) 2022-02-03 20:15:39 the list looks fine to me 2022-02-03 20:15:41 That's the thing! Apparently it didn't fail, there was just something broken at runtime and it was fixed up by someone else since I had no idea 2022-02-03 20:15:56 why not try and use abuild rootbuild? 2022-02-03 20:16:17 I just barely am making it with just regular abuild. To be honest, I'm scared of adding another layer I don't understand well 2022-02-03 20:16:29 it makes a clean chroot to build in 2022-02-03 20:16:47 Is the Wiki documentation for it current/correct for me to read up on? 2022-02-03 20:18:04 ACTION uploaded an image: (878KiB) < https://matrix.org/_matrix/media/r0/download/matrix.org/KRMvDFLZVzzOQrLtTzjXFbYH/image.png > 2022-02-03 20:18:14 This is my build environment, so it doesn't need to do anything but build, cache, and update things easily 2022-02-03 20:18:30 oh its not a normal system 2022-02-03 20:18:51 No. Dedicated WSL2 image specifically for my package porting and maintaining 2022-02-03 20:19:06 My live Alpine system is a full XFCE DE on a laptop. I don't build on that since Celeron and 4GB RAM 😛 2022-02-03 20:19:18 y not :) 2022-02-03 20:19:55 In a pinch I have, but it is painful, haha. 6700k and 32GB RAM on a real SSD vs eMMC is far more pleasant 2022-02-03 20:19:59 but i think rootbuild is not usefull in yourt case then 2022-02-03 20:20:13 Since it's basically stripped anyway? 2022-02-03 20:20:20 ye 2022-02-03 20:20:33 Awesome. Thank you all. That's been the plan haha 2022-02-03 20:20:52 gl with packaging 2022-02-03 20:20:59 happy hacking! 2022-02-03 20:21:36 🙂 Thanks! It's been a challenge, but everyone here has been super helpful, so I think I'm getting better. Looking to add a few packages that have GUI/CLI, so I'm trying to learn how to split those things up neatly 2022-02-03 20:21:56 Basically, trying to get everything that I use via Flatpak native so it takes up like, 1/10th the space on disk or less 2022-02-03 20:22:16 as long as it does not use electron ur fine 2022-02-03 20:22:53 Sadly, at lot of things I might want do (which is a problem aside from porting anyway since I'm not the biggest electron fan from my experience with it) 2022-02-03 20:23:14 main problem is: no one got electron working on alpine or tried to 2022-02-03 20:23:17 Those I just use PWAs through my browser 2022-02-03 20:23:37 I read some chatter upstream and it seems like it might not be a trivial thing to get done. I don't understand why 2022-02-03 20:23:49 Not saying it should be, but literally did not follow what the blockers were 2022-02-03 20:24:12 lots of patches required and every app ships another version of electron if i recall correctly 2022-02-03 20:25:18 Yeah, that sounds like headaches. 2022-02-03 20:25:52 I saw some talk that we might get MUSL wheels in pypi, which is exciting 2022-02-03 20:25:58 yup 2022-02-03 20:53:07 I'm trying to setup a luks/btrfs raid with grub/mbr, does alpine use /etc/crypttab ? 2022-02-03 21:24:52 hi 2022-02-03 21:35:58 ACTION uploaded an image: (64KiB) < https://matrix.org/_matrix/media/r0/download/matrix.org/KUnXSlYMRKMkYlbIOyhbCvNi/image.png > 2022-02-03 21:35:59 My project list haha 2022-02-03 21:42:56 who is fiona? i think that dir is in the wrong place ;) 2022-02-03 21:43:30 why would it be in the wrong place 2022-02-03 21:43:38 random geo program 2022-02-03 21:44:52 what is the text editor in Alpine? 2022-02-03 21:45:03 whatever text editor you install and want to use 2022-02-03 21:45:14 on the iso? 2022-02-03 21:45:23 probably just vi 2022-02-03 21:45:24 vi 2022-02-03 21:45:25 vi is pretty much ubiquitous 2022-02-03 21:45:35 ok! 2022-02-03 21:45:38 Thank you 2022-02-03 21:45:45 and ed 2022-02-03 21:45:48 both from busybox 2022-02-03 21:45:54 Busybox VI, yeah. I might just be too used to it by now, but it feels less sticky than regular VI to me now 2022-02-03 21:46:42 Yeah, fiona should be a mostly easy port, since the person who maintains all the Alpine geosci stack has done a ton of heavy lifiting with dependencies already 2022-02-03 21:47:13 Speaking of the ISO, does anyone know if the Alpine setup-* scripts live in an apk I can fetch? The minirootfs that I used to make my WSL2 image doesn't have them 2022-02-03 21:47:22 And I want to setup-apkcache as per psykose 2022-02-03 21:47:27 alpine-conf 2022-02-03 21:47:36 beautiful, thank you 2022-02-03 21:49:07 ACTION sent a code block: https://matrix.org/_matrix/media/r0/download/matrix.org/WEYXQOZEAbOmWXDombYcRGXs 2022-02-03 21:49:13 Is this a test failure, or a build failure? 2022-02-03 21:49:25 package() 2022-02-03 21:49:53 ACTION sent a code block: https://matrix.org/_matrix/media/r0/download/matrix.org/HFtHlbHncuxzdnNFiSvdWnnh 2022-02-03 21:50:10 that install probably does not install anything 2022-02-03 21:51:30 mm... This one might be a challenge. This is also a split CLI/GUI one, which I've never seen with a Rust package 2022-02-03 21:52:07 https://github.com/pop-os/popsicle#installation-instructions 2022-02-03 21:52:10 switch package from cargo install to make install? 2022-02-03 21:53:04 you can just install the single file 2022-02-03 21:53:20 but since it has extra files 2022-02-03 21:53:21 sure 2022-02-03 21:53:42 `make` then make install-cli install-gtk separately 2022-02-03 21:54:10 and put that cli one into the separate cli() section, right? 2022-02-03 21:54:23 Can I model this off of any *-cli APKBUILD I find? 2022-02-03 21:59:31 setup-apkrepos (but no mirrors found?) 2022-02-03 22:00:27 ahh... nm 2022-02-03 22:27:06 ACTION sent a code block: https://matrix.org/_matrix/media/r0/download/matrix.org/SKiUouxtglAtvpnYYvtrHEVw 2022-02-03 22:27:12 ACTION sent a code block: https://matrix.org/_matrix/media/r0/download/matrix.org/BcVhIjXMLTEHkQnjutZAuRIL 2022-02-03 22:27:24 remove the rm 2022-02-03 22:27:37 I don't need to worry about those sticking around and being packaged? 2022-02-03 22:27:43 you can make the pkgdesc="$pkgdesc (cli only)" 2022-02-03 22:27:45 no 2022-02-03 22:27:49 they're not there 2022-02-03 22:28:00 also need a PREFIX= probably 2022-02-03 22:28:10 Does that go into the prepare? 2022-02-03 22:28:28 installs 2022-02-03 22:30:08 can I use it without ./configure? 2022-02-03 22:30:18 That's all the examples I'm seeing for using --prefix=/usr 2022-02-03 22:30:41 the makefile has a PREFIX= in it 2022-02-03 22:30:48 set to usr/local by default 2022-02-03 22:31:00 also you probably want opt-level="s" 2022-02-03 22:31:16 and you don't need to run the build if you run make 2022-02-03 22:33:39 so use sed or a patch to change the makefile? 2022-02-03 22:33:55 passing make PREFIX=/usr overrides it in the makefile 2022-02-03 22:34:01 that's how they are always used 2022-02-03 23:04:08 ACTION sent a code block: https://matrix.org/_matrix/media/r0/download/matrix.org/VAytnmRPDtdWwWBZldneuObU 2022-02-03 23:04:20 ACTION sent a code block: https://matrix.org/_matrix/media/r0/download/matrix.org/rvaaehpyNGMKyyWxYdVbJhSP 2022-02-03 23:05:46 Does the cli() need another make statement? make cli then make install-cli 2022-02-03 23:11:29 when you are in a subpackage function the cwd is pkgdir not builddir 2022-02-03 23:11:35 you need to add a -C "$builddir" 2022-02-03 23:11:57 it's running make on a directory with no makefile 2022-02-03 23:11:59 hence no rule 2022-02-03 23:12:12 anyone have experience getting broadcom wifi to work? im on a 2012 macbook pro 2022-02-03 23:12:21 ive followed the instructions at https://wiki.alpinelinux.org/wiki/Wi-Fi#Broadcom_Wi-Fi_Chipset_Users to no successs 2022-02-03 23:22:54 how can I specify multiple luks devices on boot? 2022-02-03 23:23:04 rd.luks.name=device-UUID=cryptroot root=/dev/mapper/cryptroot 2022-02-03 23:23:32 is this expected to work on alpine? 2022-02-03 23:25:15 it seems redhat/dracut stuff 2022-02-03 23:27:59 ACTION sent a code block: https://matrix.org/_matrix/media/r0/download/matrix.org/dQhARwVcjPaCgCbvgiittPox 2022-02-03 23:28:09 ACTION sent a code block: https://matrix.org/_matrix/media/r0/download/matrix.org/blZdnBOJYipWhFebCREwlLSm 2022-02-03 23:29:15 deckard: any error messages in dmesg? 2022-02-03 23:32:08 yes, `dmesg | grep firmware` gives the following 2022-02-03 23:32:10 > b43-phy0 ERROR: You must go to https://wireless.wiki.kernel.org/en/users/Drivers/b43#devicefirmware and download the correct firmware for this driver version. Please carefully read all instructions on this website. 2022-02-03 23:32:30 donoban: it looks like alpine only supports cryptroot in initramfs, you can add any additional partitions to fstab though 2022-02-03 23:32:34 but i've built and installed the b43 firmware as the wiki prescribes 2022-02-03 23:33:13 ptrc: my problem seems that the btrfs raid fails to mount if the second disk is unlocked 2022-02-03 23:33:25 so I have to manual open it, mount /sysroot and continue 2022-02-03 23:33:43 oh, i see... you might just edit /usr/share/mkinitfs/initramfs-init and run mkinitfs, but that seems quite hacky 2022-02-03 23:34:03 ok let me check 2022-02-03 23:34:22 I'm not sure what should be the proper solution, maybe add crypttab to the initramfs (but I don't know if it reads it) 2022-02-03 23:34:31 or add other parameter to bootline 2022-02-03 23:40:27 donoban: there's a 2 year old issue open in mkinitfs about that 2022-02-03 23:40:28 https://gitlab.alpinelinux.org/alpine/mkinitfs/-/merge_requests/54 2022-02-03 23:40:44 merge request* 2022-02-03 23:41:08 "I use btrfs on multiple LUKS-encrypted disks." hehe same situation here 2022-02-03 23:42:49 https://gitlab.alpinelinux.org/alpine/mkinitfs/-/issues/16 2022-02-03 23:43:16 well 2022-02-03 23:43:24 I will continue tomorrow 2022-02-03 23:43:26 thanks ptrc 2022-02-04 01:22:32 why would running "apk add alpine-desktop" return ERROR: unable to select packages: alpine-desktop (no such package) 2022-02-04 01:23:01 required by: world[alpine-desktop] 2022-02-04 01:32:07 mountjoy: Where did you see to install alpine-desktop? If it's from an old wiki page, do note that there's been lots of changes over the years, and I think now you should follow the wikit page for the specific de/wm you want 2022-02-04 01:34:33 ktprograms I was using the Alpine Linux "docs" link 2022-02-04 01:34:50 I found the wiki now hopfully it's the right one 2022-02-04 01:34:58 https://wiki.alpinelinux.org/wiki/Xfce 2022-02-04 01:35:33 mountjoy: Yes that's good if you want Xfce. 2022-02-04 01:36:04 I do! 2022-02-04 01:36:22 Or I think I do 2022-02-04 01:36:31 https://git.alpinelinux.org/aports/commit?id=59ed4dcbaef9dec859eb014a003cb7a4c89780b0 2022-02-04 01:37:17 ktprograms : should I try somthing else? 2022-02-04 01:38:08 mountjoy: Up to you. Depends what DE you like 2022-02-04 01:39:10 xfce seems to work fine for me 2022-02-04 01:41:03 maybe KDE would be better? 2022-02-04 01:43:53 mountjoy: Up to you, you can always install a different DE if you don't like the one you're currently on 2022-02-04 01:47:19 ok :) 2022-02-04 01:56:31 XFCE over Alpine here. Quite enjoy it. Miss some of the GNOME extensions, but otherwise incredibly light/fluid and very usable 2022-02-04 01:57:07 mountjoy: If you want QT/KDE stack but something lighter, you may enjoy LXQT (QT/KDE update of LXDE) 2022-02-04 02:08:51 Thank you too all the Apine Linux people! I have been trying for a very long time to get my Talos II (PPC) running using gentoo and void. 2022-02-04 02:09:33 your documentation is far superior 2022-02-04 02:09:46 Thank you 2022-02-04 02:11:31 Another great resource for almost-always modificaiton free documentation/tutorials is the Arch Linux Wiki 2022-02-04 02:50:58 ACTION sent a code block: https://matrix.org/_matrix/media/r0/download/matrix.org/jGOkggWieGMjnanSPXVWpzoS 2022-02-04 02:51:02 What obvious thing am I missing here? 2022-02-04 02:51:16 It can find and build the popsicle-cli target, which is great 2022-02-04 02:52:06 It looks exactly like the regular part of the build/package run 2022-02-04 02:52:24 But it fails to cd for some reason? 2022-02-04 02:56:29 Commented out the $subpackages line and got this 2022-02-04 02:56:32 ACTION sent a code block: https://matrix.org/_matrix/media/r0/download/matrix.org/uNXvfOABfUXUCogOtHYmdyRI 2022-02-04 02:57:01 Does this mean I messed up my destdir and prefix lines for package steps? 2022-02-04 02:57:10 ACTION sent a code block: https://matrix.org/_matrix/media/r0/download/matrix.org/xNkqPwHOVjeDclTMLfRxozAv 2022-02-04 04:13:29 "dngray: It sounds like you got..." <- not sure that it is the same thing 2022-02-04 04:13:46 in this case udisksctl works if the volume is not encrypted 2022-02-04 04:13:57 it only has an issue if the volume is in an encrypted partition 2022-02-04 04:14:18 the thing i noticed in `udisksctl monitor` on my fedora system was: 2022-02-04 04:14:31 ACTION sent a code block: https://matrix.org/_matrix/media/r0/download/matrix.org/WLSWmebNKtHkWTNoCPnxFKpD 2022-02-04 04:14:43 * ```... (full message at https://matrix.org/_matrix/media/r0/download/matrix.org/mHZeceSBKMIRblhbzNuWNDBJ) 2022-02-04 04:14:47 whereas on alpinelinux 2022-02-04 04:15:13 * ```... (full message at https://matrix.org/_matrix/media/r0/download/matrix.org/hEWvrQjmhzSGFxgFBSssaZwo) 2022-02-04 04:15:32 ACTION sent a code block: https://matrix.org/_matrix/media/r0/download/matrix.org/XDkIHbcQsJhaUhuhKFeevTub 2022-02-04 04:15:52 Oh wow yeah, nothing 2022-02-04 04:15:59 on fedora it shows: 2022-02-04 04:15:59 PreferredDevice: /dev/mapper/luks-ce782db4-f9ee-4154-80e1-72c0d71b08da 2022-02-04 04:16:08 on alpine linux: 2022-02-04 04:16:09 PreferredDevice: /dev/dm-0 2022-02-04 04:16:21 so i think this is an eudev bug 2022-02-04 04:16:38 it can decrypt the partition, open a device mapper 2022-02-04 04:16:43 but can't mount the device mapper object 2022-02-04 04:17:20 anyway i think i'll report this on the alpine linux bug tracker 2022-02-04 04:17:24 cos clearly something is broken 2022-02-04 04:17:37 trying to verify it on another eudev distribution 2022-02-04 04:17:44 seems artix doesn't use eudev anymore 2022-02-04 04:18:43 maybe Devuan has a livecd or something 2022-02-04 04:27:36 dngray: Alpine eudev does not have the files /lib/udev/rules.d/55-dm.rules and /lib/udev/rules.d/60-persistent-storage-dm.rules that I see on a Debian box 2022-02-04 04:27:49 these relate to device mapper 2022-02-04 04:29:03 on fedora i have 10-dm.rules, 11-dm-parts.rules, 13-dm-disk.rules 2022-02-04 04:29:58 https://clbin.com/631lA, https://clbin.com/2YfOd, https://clbin.com/unGEW 2022-02-04 04:30:38 i think it might be 11-dm-parts.rules 2022-02-04 04:33:07 dngray: each distro I think does their own thing with some of these udev device rules 2022-02-04 04:37:23 anyway testing from artix isn't going to help, they don't use eudev anymore 2022-02-04 04:49:56 minimal: i wonder if i can copy those polkit rules and see if it works 2022-02-04 04:54:32 hmm, nope, the answer is no to that. 2022-02-04 08:12:32 if i've got a diskless usb stick, is putting alpine_dev=usbdisk:vfat still the correct way to prevent alpine linux from mounting to /mnt/sdX 2022-02-04 08:21:47 hmm that worked for syslinux but i think newer alpine linux uses grub now 2022-02-04 08:51:01 https://unix.stackexchange.com/questions/502610/alpine-persistent-disk-mounted-to-media-sda-even-though-fstab-has-media-usb i think that only worked with syslinux 2022-02-04 09:14:23 that's a pretty nifty guide https://github.com/IronOxidizer/alpine-persistent-usb 2022-02-04 10:04:53 btw i solved my problem re: broadcom drivers 2022-02-04 10:05:28 turns out the b43-firmware package on alpine was missing a couple individual files my chipset needs, so i found them manually 2022-02-04 10:13:29 hey... (full message at https://matrix.org/_matrix/media/r0/download/matrix.org/nzTcumAUmUpuTyiqUtTrwemz) 2022-02-04 10:14:41 we're not worthy of a summary even? 2022-02-04 10:24:28 Gotta love multi-line messages D: 2022-02-04 10:25:35 mustn't have been important 2022-02-04 10:25:52 Heh 2022-02-04 10:26:00 is the formatting broken? 2022-02-04 10:26:07 It was about a kernel panic trying to boot with btrfs / 2022-02-04 10:26:14 hey... (full message at https://matrix.org/_matrix/media/r0/download/matrix.org/nzTcumAUmUpuTyiqUtTrwemz) 2022-02-04 10:26:19 That's all we saw 2022-02-04 10:26:24 it's a matrix thing ig 2022-02-04 10:26:31 yeah right 2022-02-04 10:26:39 DanielYrovas[m]: http://vpaste.net/NTir0 2022-02-04 10:26:52 ahhah 2022-02-04 10:27:33 wouldn't b e surprised if your matrix client retrieved that content and showed it as if it isn't a problem 2022-02-04 10:27:52 hey, I am getting a kernel panic on boot of a new install, I used custom partitioning and btrfs with setup-disk /mnt, the bootloader is extlinux and I installed that manually following the guide here: https://wiki.alpinelinux.org/wiki/Bootloaders, anyone know anything I can do to fix this install? Thanks for your help 2022-02-04 10:27:57 is that better? 2022-02-04 10:28:10 Yes 2022-02-04 10:28:14 in that the message is actually there, yes 2022-02-04 10:28:27 hehe cool 2022-02-04 10:28:32 it was fine before 2022-02-04 10:28:36 people can just click the link 2022-02-04 10:28:42 well if you have time I would love debug it 2022-02-04 10:28:43 you should post the actual boot output 2022-02-04 10:28:47 imagine if every message was a url 2022-02-04 10:28:51 fine by me 2022-02-04 10:28:55 ehh don't wanna make people clicky 2022-02-04 10:28:57 of course 2022-02-04 10:29:36 is there anything important to do when you first set up an install using manual partitioning and btrfs? 2022-02-04 10:29:59 I can try to get the boot message if that would help 2022-02-04 10:30:16 but its kind wack on the resolution and can't really see it all 2022-02-04 10:33:08 maybe try to eliminate certain elements of the installation 2022-02-04 10:34:29 as in start deleting random parts of / ? 2022-02-04 10:34:53 try not btrfs, or not customer partitioning 2022-02-04 10:35:01 -er 2022-02-04 10:35:36 prove that alpine will work in the simple case 2022-02-04 10:35:41 I have ran alpine successfully using a normal install with setup-disk 2022-02-04 10:35:53 without the custom partitioning 2022-02-04 10:36:31 but I wanted the flat toplevel structure 2022-02-04 10:38:35 does it panic only with btrfs? 2022-02-04 10:39:03 I haven't tried custom partitioning with ext4 2022-02-04 10:46:32 this is the kernel panic, I don't think it is very enlightening. 2022-02-04 10:46:33 ACTION uploaded an image: (417KiB) < https://matrix.org/_matrix/media/r0/download/matrix.org/wINshHVClAKPurLVuQEFFUPz/20220204_214535.jpg > 2022-02-04 10:47:15 hope IRC can see the image 2022-02-04 10:48:05 We can see a link to the image 2022-02-04 10:48:27 Love how much of the stacktrace is the panic code D: 2022-02-04 10:49:12 yeah 2022-02-04 10:49:17 not a whole lot of use 2022-02-04 10:49:20 IIRC the actual error message is above all what you showed though 2022-02-04 10:49:35 and I can't ctrl-a esc to scroll up or anything 2022-02-04 10:51:25 I do have a btrfs-scan script in init.rd, I assume btrfs-progs was installed into the system on creation 2022-02-04 10:51:39 maybe I need to chroot install btrfs-progs 2022-02-04 10:53:49 Can you change your monitor resolution? 2022-02-04 10:53:58 To see more output 2022-02-04 10:54:59 4:3 doesn't help 2022-02-04 10:55:33 DanielYrovas[m]: increase the resolution (not change the aspect ratio)? 2022-02-04 10:57:13 yeah, you can usually tell the kernel to boot in a smaller text mode 2022-02-04 10:58:01 or, record a video 2022-02-04 10:59:34 can't see anymore in any monitor settings unfortunately 2022-02-04 11:00:19 DanielYrovas[m]: Just record a video then? 2022-02-04 11:00:39 with the kernel? 2022-02-04 11:00:58 I don't think I could get it in time with my phone 2022-02-04 11:01:19 https://gitlab.alpinelinux.org/alpine/aports/-/issues/13501 2022-02-04 11:01:33 anything else i could provide to be helpful? 2022-02-04 11:01:46 DanielYrovas[m]: You got a tripod? 2022-02-04 11:01:54 Or some phone mount? 2022-02-04 11:02:07 does ctrl-s pause the screen? 2022-02-04 11:02:31 lopid: Probably not during a panic 2022-02-04 11:03:29 I am currently in the panicked state 2022-02-04 11:03:33 try portrait mode if possible… 2022-02-04 11:03:37 and ctrl-s doesn't do anythiing 2022-02-04 11:04:45 go closer to the screen and look up. maybe you can see the top of the text 2022-02-04 11:05:13 haha true 2022-02-04 11:06:51 don't think I can change to portrait in display settings 2022-02-04 11:07:38 maybe a higher resolution display, like a 4k tv… 2022-02-04 11:07:48 ahaha 2022-02-04 11:07:55 not my luxury unfortunately 2022-02-04 11:09:12 could it be extlinux not configured correctly? 2022-02-04 11:09:42 can I install grub in a chroot? 2022-02-04 11:16:00 DanielYrovas[m]: Try adding fbcon=font:6x8 to the kernel parameters (6x8 is the smallest font) 2022-02-04 11:31:03 should I reinstall following a chroot install guide? 2022-02-04 13:55:24 "turns out the b43-firmware..." <- Do you think you could open an issue or MR to fix the Alpine package? 2022-02-04 13:58:08 Saijin_Naib[m]: sure i can look into that later today 2022-02-04 13:58:18 i got those drivers from the source described here http://linuxwireless.sipsolutions.net/en/users/Drivers/b43/#Other_distributions_not_mentioned_above 2022-02-04 14:50:59 When I boot into xfce my user I created is not listed. It just says Linux User. How can I add my user to the login screen? 2022-02-04 14:51:45 that is your user probably, it's just displaying the gecos field which is by default 'Linux User' 2022-02-04 14:53:12 Oh 2022-02-04 14:54:49 Yes 2022-02-04 14:57:16 do you know how I can change it to show the user name? 2022-02-04 14:57:52 you can edit /etc/passwd by hand 2022-02-04 14:58:02 ok 2022-02-04 15:03:51 psykose: that worked! 2022-02-04 15:03:58 :) 2022-02-04 17:24:08 deckard: Thanks so much! Be a shame to lose all that work and figuring out you did 🙂 2022-02-04 17:26:12 mountjoy: We also have mugshot which will integrate into XFCE as "About Me" and lets you set stuff graphically 2022-02-04 20:41:14 I assume -devel, -linux, -docs and -offtopic were locked to reggednick only due to spam? 2022-02-04 20:44:47 yes 2022-02-04 20:48:23 and also -commits now. would be good to add -cloud to that 2022-02-04 20:58:21 Done 2022-02-04 21:07:16 Looks like some of this channels are not mentioned on alpinelinux.org/community and wiki.alpinelinux.org/wiki/Alpine_Linux:IRC . Should I add these to wiki? 2022-02-04 22:19:55 also if someone knows how to auth on OFTC with Matrix IRC bridge, I would appreciate help 2022-02-04 22:29:06 Ermine: yes 2022-02-05 03:34:00 Does anyone know what would cause apk upgrade -l to error out like this 2022-02-05 03:34:17 ERROR: unable to select packages: Huh? Error reporter did not find the broken constraints. 2022-02-05 03:34:49 no tagged nor local packages in world 2022-02-05 03:34:55 (already ran apk upgrade -a) 2022-02-05 08:37:46 What -commits channel is for? 2022-02-05 09:04:55 Ermine: algitbot anounces information about commits and builds 2022-02-05 09:38:54 ikke: thank you 2022-02-05 10:42:19 Is USER env variable present in CI? 2022-02-05 10:43:28 Don't know from the top of my head 2022-02-05 11:10:29 Hmmm... seems like no. One of the tests fail because of missing USER 2022-02-05 11:11:16 Ah, this test should have been skipped 2022-02-05 15:31:03 Are MINOR versions "guaranteed" to not have breaking changes? For example, if I run apk upgrade, can I be sure that I won't need to upgrade config files? 2022-02-05 15:32:12 pgrep from busybox seems nice ... 2022-02-05 15:33:09 Nulo: we rely on upstream to not make breaking changes on patch upgrades 2022-02-05 15:33:59 help text for busybox programs also very neat 2022-02-05 15:34:34 Nulo: so we do our best to prevent it, but we cannot give an absolute guarantee 2022-02-05 15:34:50 ikke, okay, thanks 2022-02-05 15:35:20 And I assume you were refering to stable branches 2022-02-05 15:40:03 Yes, of course. 3.15.x 2022-02-05 19:58:45 How can I make Alpine use a SquashFS as root? The image is in the same partition as the kernel and bootloader 2022-02-05 20:04:18 not sure 2022-02-05 21:07:51 hey guys, ran into some funky behavior in apk-fetch: https://pastebin.com/9fdJjmQu 2022-02-05 21:08:23 apk add does it just fine, but apk fetch seems to struggle as long as the cdrom repo is mentioned in the repos file 2022-02-05 21:09:48 nt: maybe report it here? https://gitlab.alpinelinux.org/alpine/apk-tools 2022-02-05 21:10:55 aight! 2022-02-05 21:21:04 wasn't sure if it was actually a bug or just pebkac :p 2022-02-05 22:51:19 Ermine: if you feel like packaging something else, Gajim (XMPP client) is missing. I haven't tried to build it yet, not sure if there's a reason it's not packaged. 2022-02-05 23:06:20 it was, but it's in unmaintained now 2022-02-05 23:06:36 https://gitlab.alpinelinux.org/alpine/aports/-/blob/master/unmaintained/gajim/APKBUILD 2022-02-06 09:02:08 moin - someone here successfully played full-hd content in a browser on a Pi4 with alpine/aarch64? with both: chromium&firefox I was out of luck (trying out youtube) 2022-02-06 09:04:53 of course I did the stuff from the wiki like increasing gpu_mem and adding 3D support and installating corresponding kms stuff 2022-02-06 11:00:59 hello, can i ask a question? 2022-02-06 11:01:06 installed and ran nyx, but an error appears as follows. Is there a solution? Traceback (most recent call last): File "/usr/bin/nyx", line 33, in sys.exit(load_entry_point('nyx==2.1.0', 'console_scripts', 'nyx')()) File "/usr/lib/python3.9/site-packages/nyx/__init__.py", line 176, in main nyx.starter.main() File "/usr/lib/python3.9/site-packages/stem/util/conf.py", line 289, in wrapped return func(*args, config = c 2022-02-06 11:02:56 I miss the actual error message 2022-02-06 11:03:11 maybe paste it somewhere 2022-02-06 11:03:58 Traceback (most recent call last): File "/usr/bin/nyx", line 33, in sys.exit(load_entry_point('nyx==2.1.0', 'console_scripts', 'nyx')()) File "/usr/lib/python3.9/site-packages/nyx/__init__.py", line 176, in main nyx.starter.main() File "/usr/lib/python3.9/site-packages/stem/util/conf.py", line 289, in wrapped return func(*args, config = config, **kwargs) File "/usr/lib/python3.9/site-packages/nyx/starter.py", line 1 2022-02-06 11:04:50 hmm just a sec 2022-02-06 11:05:24 Traceback (most recent call last): File "/usr/bin/nyx", line 33, in sys.exit(load_entry_point('nyx==2.1.0', 'console_scripts', 'nyx')()) File "/usr/lib/python3.9/site-packages/nyx/__init__.py", line 176, in main nyx.starter.main() File "/usr/lib/python3.9/site-packages/stem/util/conf.py", line 289, in wrapped return func(*args, config = config, **kwargs) 2022-02-06 11:05:34 File "/usr/lib/python3.9/site-packages/nyx/starter.py", line 106, in main _set_process_name() File "/usr/lib/python3.9/site-packages/nyx/starter.py", line 222, in _set_process_name stem.util.system.set_process_name('nyx\0%s' % '\0'.join(sys.argv[1:])) 2022-02-06 11:05:40 File "/usr/lib/python3.9/site-packages/stem/util/system.py", line 1432, in set_process_name _set_argv(process_name) 2022-02-06 11:05:44 File "/usr/lib/python3.9/site-packages/stem/util/system.py", line 1459, in _set_argv raise IOError("Can't rename process to something longer than our initial name (this would overwrite memory used for the env)") OSError: Can't rename process to something longer than our initial name (this would overwrite memory used for the env) 2022-02-06 11:06:08 I don't get it 2022-02-06 11:12:22 is this caused by 3.10 being 1 char longer than 3.9... 2022-02-06 11:12:24 lol 2022-02-06 11:12:35 ah no 2022-02-06 11:12:45 sounds like just a broken package i guess 2022-02-06 11:15:12 I installed the package without anything special. 2022-02-06 11:15:59 fresh out installing os, add edge/testing repo. that's all 2022-02-06 11:21:45 @psykose thx for answering. It was solved by setting all repository; main to edge. 2022-02-06 11:24:24 ah 2022-02-06 11:24:28 so it was caused by mixing thing 2022-02-06 11:24:35 I had a supsicion 2022-02-06 11:25:45 I made a simple mistake. 2022-02-06 11:26:49 Next time, I will set all the repos to the edge and installing the package. 2022-02-06 11:47:27 does anyone know what the trick is to running a tmux session via openrc? 2022-02-06 11:47:48 I've found a few things online but nothing seems to work properly, just crashes 2022-02-06 12:52:16 gjabell what is the command you are using 2022-02-06 13:10:34 pj: I've tried something like this: https://gist.github.com/kirelagin/099098c7418ddc66f3f3 2022-02-06 13:55:52 crookedyelp: That's the traceback, was there an error message higher up? 2022-02-06 13:56:04 oops sorry, didn't scroll to the end 2022-02-06 14:45:16 I am trying to setup a local alpine linux PXE server in an airgapped network. What I would like to achieve is the ability to only serve the packages that are in my local apk cache. The local cache seems to be setup as a catch all for all packages, independent of upstream repository. Is there a way to have main and community in separate folders? 2022-02-06 14:47:28 well, the cache is only a cache and doesn't mirror the directory structure on the mirrors 2022-02-06 14:47:56 so without some mapping between packages -> repo in the web server that'll be difficult 2022-02-06 14:48:43 Lets say, it is easy to symlink everything. That cannot be achieved for the APKINDEX, and would require a webserver trick. 2022-02-06 14:49:23 But if it would be possible to directly get the files in the right folder, matching the upstream name, that is obviously easier. 2022-02-06 14:53:35 skinkie: set up a local mirror of the Alpine repos and serve that locally via HTTP 2022-02-06 14:58:14 minimal: that would cost me > 200GB of files. With less than 300MB of files in my cache, this is not what I want to do. 2022-02-06 14:59:31 minimal: skinkie: If you just have the packages you want to cache in your local http repo, and put it as higher priority than the main mirrors, apk should download from the cache right? 2022-02-06 15:02:24 skinkie: In appreciate that, but if you do not have a working solution currently then isn't a working solution better than no solution? 2022-02-06 15:03:09 alternatively if all machines are using the same set of packages why not make a "installed" disk image and write that to each machine's disk... 2022-02-06 15:03:28 200 GB sound like a lot, if you need just one architecture, mirroring main and community should be a lot less than that 2022-02-06 15:04:02 minimal: the working solution i am going to make, is setting up my webserver, and serving the cache files. Then serve every other request from the cache folder. My question here was: is there an existing solution. If there is none, that is cool. 2022-02-06 15:04:19 ktprograms: how would that priority thing works? 2022-02-06 15:05:18 skinkie: higher up in /etc/apk/repositories? I haven't tested this, it's just an idea. 2022-02-06 15:06:28 ktprograms: I think it has the same issue; you would have to find a way to get APKINDEX from the right folder. 2022-02-06 15:08:28 wej: yeah ballpark ~20Gb for 3.15 x86_64 main & community 2022-02-06 15:08:50 yep, sound about right 2022-02-06 15:09:26 one could also explicitly exclue certain packages, like all debug packages or something, we you know you don't need them 2022-02-06 15:09:31 *exclude 2022-02-06 15:10:12 if the "work" machines can access the mirror and the mirror can access the internet then it's not really "airgapped" is it 2022-02-06 15:10:47 Hello71: well he already has got stuff into the cache dir on the "airgapped" server somehow in the first place 2022-02-06 15:10:56 maybe you want a caching http proxy 2022-02-06 15:11:10 so could use the same approach to create a local "airgapped" mirror 2022-02-06 15:15:08 Hello71: it is really airgapped, just a VM on a stick 2022-02-06 15:29:10 i use alpinelinux on a usb stick for airgapping 2022-02-06 15:29:33 https://gitlab.alpinelinux.org/alpine/aports/-/issues/13501 was the only problem i had hehe 2022-02-06 15:32:50 dngray[m]: relating to your udisksctl/eudev issue, I noticed some more (non-disk) udev rules in systemd that are missing from upstream eudev - eudev do not appear to be tracking such upstream-upstream rule additions. 2022-02-06 15:33:24 yeah i'm not really sure which ones are missing but that seems to be the issue 2022-02-06 15:33:49 is eudev really maintained? tbh i think its a bit silly all these udevd implementations 2022-02-06 15:33:58 there's eudev, there's mdev, there's zero-udevd 2022-02-06 15:34:14 if i can't use systemd-udev which one should i use 2022-02-06 15:36:02 gentoo dropped eudev a bit ago, not maintained anymore 2022-02-06 15:37:33 https://www.gentoo.org/support/news-items/2021-08-24-eudev-retirement.html < statement on it 2022-02-06 15:37:45 joecool: eudev is maintained, after gentoo drop it a group of people including Ariadne stepping to keep maintaining it 2022-02-06 15:38:05 s/stepping/stepped in/ 2022-02-06 15:38:05 minimal meant to say: joecool: eudev is maintained, after gentoo drop it a group of people including Ariadne stepped in to keep maintaining it 2022-02-06 15:38:45 interesting, did they keep the name or fork it to something different? 2022-02-06 15:38:57 dngray[m]: eudev is a fork of systemd's udev. mdev is something completely different - an alternative that does not work the same way 2022-02-06 15:39:19 joecool: https://github.com/eudev-project/eudev 2022-02-06 15:39:24 i've been using mdev on my server 2022-02-06 15:39:42 "a new project was established on 2021-09-14 by Alpine, Devuan and Gentoo contributors" 2022-02-06 15:40:00 dngray[m]: yes, mdev is the default on Alpine 2022-02-06 15:40:18 cool 2022-02-06 15:41:07 iirc eudev was more fully fledged or something and better for desktop use 2022-02-06 16:12:55 where does the default black cursor theme comes from in sway? 2022-02-06 16:13:32 there are no entries in /usr/share/icons, yet there's this little nice black arrow 2022-02-06 16:15:10 Is there a tool in aline that does all the tricks to create a modloop file? Or is it always a manual process? 2022-02-06 16:21:20 check the scripts/ dir in the aports repo 2022-02-06 16:28:59 isn't modloop from update-kernel 2022-02-06 16:29:18 https://gitlab.alpinelinux.org/alpine/alpine-conf/-/blob/master/update-kernel.in 2022-02-06 16:31:07 Hello71: that's a way to update it for a running system 2022-02-06 16:31:12 but the iso comes with one as well 2022-02-06 16:32:48 Oh, that uses update-kernel 2022-02-06 16:32:54 https://gitlab.alpinelinux.org/alpine/aports/-/blob/master/scripts/mkimg.base.sh#L6 2022-02-06 16:32:59 Hello71: so yes, you are right 2022-02-06 16:34:33 Hello71: thanks, yes I just used that. 2022-02-06 16:55:12 So my plan using the cache folder has an interesting twist. The cache folder adds some sort of a checksum to the file. Basically making it unusable to use it as a static folder. 2022-02-06 18:20:42 Can I make udhcpc 'sending discovers' in background so it does not slow down the boot process? 2022-02-06 18:23:00 gera: do you have any init.d services that need network? 2022-02-06 18:23:53 ntpd 2022-02-06 18:24:22 so if you but udhcpc into background then ntpd service may fail as network may not be up then 2022-02-06 18:24:27 s/but/put/ 2022-02-06 18:24:27 minimal meant to say: so if you put udhcpc into background then ntpd service may fail as network may not be up then 2022-02-06 18:25:58 But can they proceed loading but give me tty access or the display manager as soon as possible? 2022-02-06 18:26:20 ok, and what about your network services like ntpd? you do not care if they fail? 2022-02-06 18:28:06 idk, user handbook recommended enabling it 2022-02-06 18:30:08 yes ntpd is used to ensure your computer's clock is accurate and so it is general recommended 2022-02-06 18:31:02 and other network services enabled? sshd? webserver? as, again, if you background udhpc then they may fail to start 2022-02-06 18:31:40 yes, sshd as well 2022-02-06 18:32:37 But it just feels strange that i need to wait until it connects to wifi before I can start using the system. Especially when there is no wify around. 2022-02-06 18:34:14 gera: I wouldn't expect it to run udhcpc until after it has connected to a Wifi access point 2022-02-06 18:34:29 is the udhcpc using the Wifi interface or the ethernet interface? 2022-02-06 18:36:23 /etc/network/interfaces: http://ix.io/3OSs Wify is connected via wpa_supplicant 2022-02-06 18:39:15 it's using eth0 2022-02-06 18:41:24 ok, so is there an ethernet cable actually connected? 2022-02-06 18:41:59 sry, it's from /etc/udhcpcd.conf . udhcpcd != udhcpc, right? 2022-02-06 18:42:25 minimal: well i may be using ethernet as well as wify 2022-02-06 18:42:41 usually only wify 2022-02-06 18:45:08 gera: I'm making the point of why run udhcpc on an interface (whether wifi or ethernet) that is not actually available? (no cable or no AP connection) 2022-02-06 18:45:36 udhcpc is a DHCP client, to get an IP, udhcpcd is a DHCP server, not give IPs to devices 2022-02-06 18:45:42 s/not/to/ 2022-02-06 18:45:42 minimal meant to say: udhcpc is a DHCP client, to get an IP, udhcpcd is a DHCP server, to give IPs to devices 2022-02-06 18:51:40 I seem to have solved it by installing dhcpcd 2022-02-06 18:52:18 I am aware that apkovl exists, and pkgs, is there anything that can for example state make sure that a script executes after this installation? 2022-02-06 18:53:03 there's no such thing as udhcpcd.conf 2022-02-06 19:11:53 You can assign static IPs and then dhcp client won't even start 2022-02-06 20:31:40 Is there any eason why modloop would not get executed in combination with apkovl? 2022-02-06 20:32:11 skinkie: is the modloop service enabled? 2022-02-06 20:34:20 ikke: if I boot without the ovl the modloop works, so I expect it is enabled at that point 2022-02-06 20:34:33 but I might misunderstand how apkovl works 2022-02-06 20:38:11 ikke: when I am using apkovl, does it overwrite the existing files, keeping the rest? Or is there more to it, for example deletes and so? 2022-02-06 21:02:00 Is there a reason why stuff like this https://gitlab.alpinelinux.org/alpine/aports/-/issues/10796 does not get merged? 2022-02-06 21:17:41 skinkie: its mkinitfs, very little changes to it get merged 2022-02-06 21:18:19 mainly due to a lack of a test framework and therefore the risk that new changes might break existing functionality 2022-02-06 21:20:46 skinkie: plus I don't see any issue raised for it against mkinitfs, rather than in aports 2022-02-06 21:21:17 I can move it to the mkinitfs project 2022-02-06 21:21:27 and Issues don't get get merged, MRs get merged 2022-02-06 21:40:23 minimal: lets say I am facing the same issue, I also want to have my community pkgs added 2022-02-06 21:41:09 regarding the modloop issue, there is an issue, certainly because under the openrc part it assumes something it can loopmount 2022-02-06 21:44:21 So te be very specific. When I don't specify apkovl my modloop works. When I do it likely fails because of multiple reasons, one of the reasons being the script wants to loop mount the file. /dev/loop0 is not available because loop.ko is in the modloop itself... 2022-02-06 21:46:23 how to list explicitly/manually installed packages ? (in contrast to `apk list -I` which lists all installed) 2022-02-06 21:51:35 cat /etc/apk/world 2022-02-06 21:52:23 great! 2022-02-06 22:19:26 list explicitly installed packages with apk? 2022-02-06 22:27:06 looks like apk list -I and apk info -vv show pulled dependencies as well 2022-02-07 01:48:05 STATUS 2022-02-07 01:48:13 SENDPASS 2022-02-07 01:48:19 LIST 2022-02-07 04:13:31 Just trying to install docker on alpine 3.15 from repo. apk add docker - but it's not working. 2022-02-07 04:14:11 who #alpine-linux 2022-02-07 04:15:32 what is it doing instead of working? 2022-02-07 04:15:41 do you have an error message to share here? 2022-02-07 04:16:17 I will get the exact prompt, just a sec. 2022-02-07 04:18:17 lol, I messed up the repo's trying to fix that. - oh deer. 2022-02-07 04:18:41 Hey - anyone know an easy way to change mirrors in alpine? 2022-02-07 04:18:50 dudezilla: setup-apkrepos? 2022-02-07 04:19:03 GOLD! 2022-02-07 04:19:06 Then delete the 5 or so lines of the previous mirror in /etc/apk/repositories 2022-02-07 04:44:52 I will try to turn on the docker services according to the info in the wiki 2022-02-07 04:45:22 dudezilla: And probably also add your user to the docker group 2022-02-07 04:51:30 I believe it senses the awesome prowess of the community members present. It works now. 2022-02-07 05:28:52 dudezilla: fyi, adding your user to the docker group is giving them root access, just fyi 2022-02-07 05:31:59 How about openvswitch. - I drop the IP address on eth0, then I install a bridge, "newt" 2022-02-07 05:33:15 add eth0 to a port on the bridge, but when I use "udhcpc newt" I have some issues. 2022-02-07 05:34:09 I get no response. From the DHCP server. 2022-02-07 05:41:01 No experience with openvswitch 2022-02-07 05:42:06 ikke: is there any way to use docker without being in the docker group? 2022-02-07 05:42:29 being root 2022-02-07 05:42:30 use sudo 2022-02-07 05:42:32 or that 2022-02-07 05:42:37 s/sudo/doas 2022-02-07 05:42:37 ikke meant to say: use doas 2022-02-07 05:43:08 so "doas docker ..." for every command? 2022-02-07 05:43:16 yes 2022-02-07 05:45:12 ikke: I see, thanks 2022-02-07 05:50:21 The newer distros make it difficult to use the docker group, and it's not showing up anymore in guides and what not. 2022-02-07 05:52:39 Openvswitch - I love the concept - but... it's complicated 2022-02-07 13:58:48 can also chown the docker socket to give the user you desire to execute docker things permissions to read/write to it 2022-02-07 13:59:41 Death_Syn: makes sense, thanks 2022-02-07 23:19:15 IDENTIFY tungsten 2022-02-08 00:23:08 is the intel/amd firmware needed when you run alpine in a vm or do they even have some kind of effect? 2022-02-08 00:31:13 Apachez: I already replied to you in the other channel... 2022-02-08 00:41:56 minimal: sure but there are two alpine channels with different participants and your answer was basically "I dont know" 2022-02-08 00:42:47 Apachez: nope, my answer was basically that unless QEMU is emulating bugs in specific versions of specific CPUs then not using it won't make a difference 2022-02-08 00:43:03 there are other vm engines than QEMU 2022-02-08 00:43:10 and QEMU also have the cpu type "host" 2022-02-08 00:43:29 which I already told you in the other channel... 2022-02-08 00:43:36 Microcode isn't needed in VM. as for other types of firmware, I don't know. 2022-02-08 00:43:41 Apachez: same principle - microcode is a way for Intel/AMD to patch bugs in specific versions of specific CPUs 2022-02-08 00:49:03 zcrayfish: yeah that is what I was thinking that its a host only feature however with virtualization evolving I wouldnt be surprised if a microcode update might do something for the vm itself 2022-02-08 00:49:58 Apachez: it would also depend on your VM's kernel setting of CONFIG_MICROCODE - it that is not set then there's no way for a VM to even try to update microcode 2022-02-08 00:50:07 minimal: seems like those microcode updates not only contain bugfixes but also thermal profiles and other 2022-02-08 00:50:19 If that's the case, I would expect the microcode updates to be included in the VM's BIOS/UEFI. Getting the OS to do it on a VM is just nuts. 2022-02-08 00:50:39 a regular linux does it upon boot 2022-02-08 00:50:55 except for whatever the bios itself does during powerup 2022-02-08 00:51:35 [ 0.000000] microcode: microcode updated early to revision 0x26, date = 2019-11-12 2022-02-08 00:52:12 Apachez: as zcrayfish pointed out a BIOS also typically includes microcode which it applied first - obviously for a physically machine there comes a point when the vendor stops supplying BIOS updated and that's then when the kernel itself is more likely to then apply rtecent microcode updated during boot 2022-02-08 00:55:38 that comes pretty quickly nowadays that bios updates stops arriving 2022-02-08 00:55:57 also not everybody applies all bios updates so having that delivered through the os as most oses does tend to be more efficient 2022-02-08 00:56:04 but that wasnt the question here 2022-02-08 00:59:12 Was the original question specifically about microcode or just firmware in general? 2022-02-08 00:59:23 that was background. My point was that even if your hypervisor is exposing the same CPU as your host some hypervisors may let you tweak some aspects of what is presented (enable/disable some CPU instructions etc) so even if your VM kernel has microcode support enabled and your OS is configured to check/update microcode whether or not it even tried to do so depends on the revision code (as you example above showed) for that specific CPU and 2022-02-08 00:59:23 whether there is one or more microcode updates to that revision 2022-02-08 01:00:45 so if your host is on revision 0x26 your hypervisor may or may not also announce revision 0x26 also 2022-02-08 01:09:55 Apachez: on both your host and VM try "cat /proc/cpuinfo | grep -i microcode" 2022-02-08 01:12:22 microcode : 0x26 2022-02-08 01:12:22 microcode : 0x26 2022-02-08 01:12:52 zcrayfish: the original question was the firmware packages named intel and amd containing microcode updates and whatelse 2022-02-08 01:13:07 Apachez: on the VM I have here there is NO microcode reported, I assume as I'm using linux-virt which has CONFIG_MICROCODE unset 2022-02-08 01:14:04 so with linux-virt there is no way for microcode to be updated in a VM is the microcode functionality in the kernel is disabled 2022-02-08 01:14:29 brb new kernel :) 2022-02-08 01:14:33 you must be using linux-lts kernel which is why you do see a microcode revision reported 2022-02-08 01:14:42 its year 2022 and one still need to reboot when new kernel arrives for the host :P 2022-02-08 01:14:59 na Im using the non lts version 2022-02-08 01:15:54 is he goalposting or what? 2022-02-08 01:19:08 Apachez: non lts? I'm referring to the Alpine kernel package of which there are 2, either linux-virt or linux-lts (well there's actually 2 more if you're using Alpine Edge but they're derivatives of lts/virt) 2022-02-08 01:40:27 oh 2022-02-08 02:35:50 has anyone used alpine linux on the framework laptop? for some reason libinput won't recognize the framework laptop touchpad. I'm on linux 5.15.21-0-lts and libinput 1.19.3 2022-02-08 02:44:58 adnano: Do you know what the touchpad is driven by on another distro? 2022-02-08 02:47:22 Saijin_Naib[m]: the touchpad works on Fedora, which also uses libinput 2022-02-08 02:49:57 this thread describes my issue: https://community.frame.work/t/fixed-trackpad-in-linux/3747/11 2022-02-08 02:58:39 adnano: I don't see this module in our kernel config 2022-02-08 02:58:39 https://wiki.gentoo.org/wiki/Framework_Laptop 2022-02-08 02:58:45 intel_ishtp_hid 2022-02-08 03:06:47 (the rest that Gentoo says are required are already configured as loadable modules) 2022-02-08 03:11:01 Saijin_Naib[m]: thanks. does that mean that it won't work without rebuilding the kernel? 2022-02-08 03:11:40 Most likely. I'm not sure if it's present in the upstream vanilla kernel config or not. You might just get by by adding that line in and rebuilding to test 2022-02-08 03:11:52 I had to do similar to get my GeoBook 120 supported in Alpine this past summer 2022-02-08 03:12:30 Do you have abuild setup to make a build for you? 2022-02-08 03:13:14 yes 2022-02-08 03:14:44 Apachez: for modern cpus, it is not possible and in fact does not really make sense to update microcode from vm 2022-02-08 03:16:04 the cpu does not maintain the state necessary to have per-vm microcode. suppose there is a problem where DIV returns wrong results in some cases. if you updated it from vm, it would change everybody's DIV results, including host and other vms. 2022-02-08 03:16:12 so the vmm will not permit this 2022-02-08 03:19:20 adnano: What I did, which is likely not a great approach, was to try to compile vanilla/upstream kernel and get the vanilla kernel config, then used GNOME Meld to get a diff of the two, and then modified the config-lts.x86_64 to have the module lines I was missing 2022-02-08 03:19:47 Then used abuild to make the kernel with my modified config-lts.x86_64 and test locally before opening an MR 2022-02-08 03:27:42 isn't that the same as diff .config arch/x86/configs/x86_64_defconfig? or make defconfig && diff .config .config.old 2022-02-08 03:31:18 Saijin_Naib[m]: could CONFIG_INTEL_ISH_HID=m be related? it seems like that should enable it 2022-02-08 04:39:05 Hello71: Probably! Like I said, I'm sure my approach wasn't the best 2022-02-08 04:39:58 adnano: My understanding is that if the kernel module is set to compile as a loadable module, it will be probed/loaded at boot as necessary. According to the Gentoo Wiki, we're missing intel_ishtp_hid module 2022-02-08 04:41:53 Saijin_Naib[m]: it seems like the intel_ishtp_hid module is present, though. at least I see it with lsmod 2022-02-08 04:42:28 I modprobed it 2022-02-08 04:43:49 I also enabled the hid_multitouch module, but it still didn't resolve the issue 2022-02-08 04:50:12 okay, I rebuilt the kernel with CONFIG_INTEL_ISH_FIRMWARE_DOWNLOADER and CONFIG_INTEL_ISHTP_ECLITE, copied from fedora's config. and that still doesn't seem to fix the issue 2022-02-08 05:11:19 another hint: according to https://wiki.archlinux.org/title/Framework_Laptop, the USB id of the touchpad is 093a:0274. however, it doesn't show up with lsusb 2022-02-08 07:06:36 ugh! 2022-02-08 08:29:48 Does alpine linux have a way to describe the complete configuration of a system, i.e what packages to install, what services to enable etc? 2022-02-08 08:30:13 I'm used to guix and I'm moving to alpine. I'm looking for something like https://git.sr.ht/~sircmpwn/builds.sr.ht/tree/master/item/images/guix/system.scm but for alpine. 2022-02-08 08:31:56 The closest I got drawing this parallel is to use mkimage from aports. Is that meant for my use-case? 2022-02-08 08:39:13 There's Alpine Configuration Framework 2022-02-08 08:41:52 Ermine: I looked into it, but it seems that it's for users to configure the host system via a webpage that it hosts as a service. Is this right? 2022-02-08 08:42:25 dhruvin: seems like you're right 2022-02-08 08:43:40 I'd like to have my system configuration in a git repo, and then have some tool create/reconfigure an alpine image (or existing installation) accordingly. Can it be done with Alpine Configuration Framework? 2022-02-08 10:18:47 We don't have a declarative desired state config system 2022-02-08 11:17:43 Okay. Thanks ikke and Ermine. 2022-02-08 11:18:49 What about things like Ansible or Chef? 2022-02-08 11:22:52 I'll most likely go with Ansible for it 2022-02-08 13:38:13 adnano: Is it not showing up since it hasn't initialized properly? Are we still missing a module somehow? 2022-02-08 13:53:40 Saijin_Naib[m]: I'm not sure why. I don't know what could still be missing. 2022-02-08 14:00:31 adnano: Can you try to capture the output of lsmod for a distro that's known working on the same kernel (ish, branch at least) and diff the output against lsmod for our kernel? 2022-02-08 14:00:47 It took me like, 4 months to get my touchpad sorted -_- 2022-02-08 14:39:32 Are there any docs on changes expected in apkv3? 2022-02-08 15:27:37 Saijin_Naib[m]: sure, I'll try that later today 2022-02-08 17:25:19 Hi all :-) 2022-02-08 17:26:24 What do I need to boot AlpineLinux from mmcblk1 ? 2022-02-08 17:26:58 I guess there is something to add to mkinitfs.conf to load required modules ? 2022-02-08 17:30:01 I tried to add mmc_block to /etc/modules but I'm still missing something 2022-02-08 17:46:13 kmmndr: Did you run setup-alpine on the machine? I'm running from a laptop thats eMMC and all I had to do was run setup-alpine and follow the prompts 2022-02-08 17:47:13 Saijin_Naib[m]: yes I did, but it does not manage to mount root device during early boot process 2022-02-08 17:49:41 kmmndr: do you have "mmc" listed in the features list in /etc/mkinitfs/mkinitfs.conf? 2022-02-08 17:50:13 kmmndr: was the iniramfs built when "mmc" was present in the mkinitfs.conf? 2022-02-08 17:59:13 minimal: mmc is set by default 2022-02-08 18:10:17 kmmndr: what are you booting? Raspberry Pi? PC? 2022-02-08 18:10:41 minimal: that's a small x86_64 laptop 2022-02-08 18:11:39 kmmndr: so you booted off USB stick and are installing to the onboard emmc? 2022-02-08 18:12:19 Nope :-) I want to install sys into a micro-sd card 2022-02-08 18:12:47 Now I've mounted and chrooted into root device from live boot 2022-02-08 18:13:33 And I'm wondering what to add into mkinitfs.conf to regenerate init ram disk properly 2022-02-08 18:13:58 I did try to add mmc_block module, but it is not enough 2022-02-08 18:19:08 kmmndr: as you have booted (into ram) off something then won't "lsmod" give you an indication of features you'd need to enable in mkinitfs.conf (as well as "mmc") to build a suitable initramfs? 2022-02-08 18:20:01 you may need also to tweak the bootloader settings (i.e. "modules=") 2022-02-08 22:29:52 Can I use the gcc-aarch64-none-elf package for cross-compiling Linux kernel or do I need a -gnu variant? 2022-02-09 04:35:23 Does anyone care about the zswap kernel module? Should I make an issue or MR to see about having it enabled in our default kernel config as a module for folks who want to use it? 2022-02-09 04:36:11 I'd use it. 2022-02-09 04:36:29 Fantastic! What kernel config/architecture are you? 2022-02-09 04:36:44 I'm x86_64, but on a low-end laptop 2022-02-09 04:37:20 my vps is virt on x86_64, my laptop is lts on x86_64 2022-02-09 04:37:31 Alright, awesome 2022-02-09 04:38:23 Does anyone know if this is something that can't/should not be enabled for all the kernel configs? It seems like upstream has had it in the default config since something like Kernel 3.11 or so 2022-02-09 04:38:42 So to me, that means it's probably safe for everything to enable as an optional module 2022-02-09 04:41:02 i don't see why not 2022-02-09 04:41:33 Do you have any insight into why it hasn't been enabled, or is it just something nobody asked for yet? 2022-02-09 04:42:04 i think zram fills the same niche 2022-02-09 04:42:25 Sort of, it's not great for RAM-constrained devices as it isn't backed by on-disk swap 2022-02-09 04:42:48 it can be 2022-02-09 04:42:50 zswap tries to compress the old pages in memory so more fit, then dumps them to disk when they don't or they're not compressible to some threshold 2022-02-09 04:42:56 i have done disk zram 2022-02-09 04:43:00 it supports both 2022-02-09 04:43:04 Really? That wasn't understanding. huh 2022-02-09 04:43:13 So is that better then? 2022-02-09 04:45:17 i'm not sure 2022-02-09 04:46:59 https://en.wikipedia.org/wiki/Zram 2022-02-09 04:46:59 [WIKIPEDIA] Zram | "zram, formerly called compcache, is a Linux kernel module for creating a compressed block device in RAM, i.e. a RAM disk with on-the-fly disk compression. The block device created with zram can then be used for swap or as general-purpose RAM disk. The two most common uses for zram are for the storage..." 2022-02-09 04:47:15 This is sort of reading like zram is a superset of the functionality of zswap 2022-02-09 04:48:23 yes 2022-02-09 04:48:27 that was my thinking too 2022-02-09 04:48:36 but i am not knowledgeable enough in the space to know of the technical difference 2022-02-09 04:50:10 https://en.wikipedia.org/wiki/Zswap#Alternatives 2022-02-09 04:50:10 [WIKIPEDIA] Zswap#Alternatives | "zswap is a Linux kernel feature that provides a compressed write-back cache for swapped pages, as a form of virtual memory compression. Instead of moving memory pages to a swap device when they are to be swapped out, zswap performs their compression and then stores them into a memory pool dynamically..." 2022-02-09 04:50:27 This seems like zswap would be best for a single-swap device system on disk 2022-02-09 04:51:38 Enable both modules in kernel config and let folks choose? 🤷 2022-02-09 09:07:52 can someone P)ULEASE help me with Wireguard? :) https://www.reddit.com/r/AlpineLinux/comments/snkiwb/still_struggling_with_wireguard_on_alpine/hw4u6zv/?context=3 2022-02-09 09:07:52 [REDDIT] Comment by spicyboi404 | 2 points | Posted at 2022-02-08 - 21:05:13UTC | Just a heads up that you should go ahead and ask in IRC, they are very helpful and a lot more active than here on Reddit. Good luck! 2022-02-09 15:24:28 Does anyone know how to IDENTIFY via the oftc matrix bridge? 2022-02-09 15:54:26 Saijin_Naib: ask in #oftc 2022-02-09 17:09:20 @xion: thanks 2022-02-09 17:25:40 Saijin_Naib talk to @_oftc_NickServ:matrix.org with `IDENTIFY ` 2022-02-09 17:25:52 so basically the same as on IRC 2022-02-09 17:54:25 PureTryOut (matrix.org): Thank you so much. I was having trouble finding the actual username for NickServ via the bridge. 2022-02-09 17:54:52 And yea, just like you said, the exact same syntax/process as in IRC just without needing the /msg NickServ part of the commands 2022-02-09 18:06:33 anyone have USB/IP working from an Alpine host? the init scripts seem a bit obtuse or outdated 2022-02-09 18:39:06 Saijin_Naib: note that any OFTC user is reachable via the @_oftc_:matrix.org format. And since NickServ is just a fancy bot, they too follow that scheme 2022-02-09 18:39:16 s/@_oftc_/@\_oftc\_/ 2022-02-09 18:39:16 PureTryOut meant to say: Saijin_Naib: note that any OFTC user is reachable via the @\_oftc\_:matrix.org format. And since NickServ is just a fancy bot, they too follow that scheme 2022-02-09 18:39:57 Beautiful, thank you! I wasn't using the leading/trailing underscores, so I was not getting anywhere haha 2022-02-09 18:40:43 Yeah that part is a bit weird 😅 I prefer the libera nicks which are just @:libera.org 2022-02-09 18:49:21 Is the registration meant to curb spam or other channel abuse? Part of a COC/TSC decision? 2022-02-09 18:51:09 Curb spam 2022-02-09 18:54:34 Yes, just an anti-spam measure 2022-02-09 19:05:25 Given the change in channel security, is there any way I can go back and check the logs from after I got de-authed? I was discussing zram/zswap and I'm not sure if I missed anything after psykose's comments 2022-02-09 19:06:28 https://irclogs.alpinelinux.org/ 2022-02-09 19:06:51 Oh, that's not impacted? Cool, thank you 2022-02-09 19:33:25 hi 2022-02-10 04:01:22 So zswap seems to be working stable without any disastrous side-effects that I can tell, so that's exciting 2022-02-10 10:46:57 I think i found it. apkovls are used for that purpose, right? 2022-02-10 11:09:00 sry, connectino was interrupted 2022-02-10 11:14:04 So i think i have everything in place, but the overlay doesn't get used. Is there a special variable that has to be set in order to actually load it? 2022-02-10 12:16:16 Where did you place the overlay? 2022-02-10 12:35:49 actually i think, i don't understand how the boot process really works. I see that some files i changed in genapkovl-custom.sh (copied from dhcp) actually are changed, so i guess i'm missing something else in there 2022-02-10 12:36:11 because the packages listed in /etc/apk/world are not installed. 2022-02-10 12:36:23 but i can't find the correct documentation 2022-02-10 12:41:54 different question: makedepends in APKBUILDS only list build time dependencies right? 2022-02-10 12:42:47 mfg: Yes, but if a program links againsts shared libraries from makedepends, the packages providing the libraries will be automatically added to depends by the dependency tracer (runs as part of abuild) 2022-02-10 12:43:29 ktprograms: thx, now it does mkae sense where all those deps come from ^^ 2022-02-10 12:50:05 You can see in the build log, it will list the shared lib deps it found 2022-02-10 12:51:29 mfg: For example https://gitlab.alpinelinux.org/ktprograms/aports/-/jobs/544610#L2256 2022-02-10 12:51:37 @ikke: i think i solved it now :D but i'm not entirely sure how... 2022-02-10 12:52:12 mfg: Fyi, the @ in the irc nick just means that they're a mod, you don't need to use the @ when mentioning them 2022-02-10 12:52:37 ah, i see :D thx 2022-02-10 13:05:47 is it possible to create arbitrary files insinde genapkovl scripts? or is it only possible to write to etc? 2022-02-10 13:07:28 or is there some limitation which doesn't let me add dotfiles? 2022-02-10 13:22:00 arg, what the heck. I should read the last line too ... 2022-02-10 13:25:10 but honestly: the way aports and every alpine related thing i have seen is written/designed is awesome 2022-02-10 15:36:51 Can someone help me troubleshoot my APKBUILD, specifically the subpackage-cli step? 2022-02-10 15:42:09 sure 2022-02-10 15:43:40 Sweet, thanks. I think I just understand how this section works 2022-02-10 15:43:49 ACTION sent a code block: https://matrix.org/_matrix/media/r0/download/matrix.org/FCRaWxnLXcDkvqHJWqfGfBYJ 2022-02-10 15:46:43 `make install gpscorrelate` doesn't install gpscorrelate, but runs those two named steps 2022-02-10 15:46:49 'install' and 'gpscorrelate' 2022-02-10 15:47:01 also PREFIX= doesn't do anything as it's named prefix= in the makefile 2022-02-10 15:47:27 there is no separate install step for the gui/cli, so what you should do is make install in package(), then amove in the subpackage step 2022-02-10 15:51:02 like this https://img.ayaya.dev/p98IyP9PbwFH 2022-02-10 15:54:34 or this https://img.ayaya.dev/pypoAC4IAvq3 if you also want the translations 2022-02-10 15:54:52 also .tar.gz is preferred over zip :) 2022-02-10 16:06:29 Thank you! I'm going to try to fix this all up in an hour or so after I'm done with student support 2022-02-10 16:11:22 So github provides the tar.gz for every commit alongside the zip? I was getting the zip link from the Download repository button, had no idea it offered other formats (thought that was just for the release/tags function) 2022-02-10 16:13:11 And you dropped make/pkgconf because it'll trace those dependencies automatically? 2022-02-10 16:14:23 What is amove? Is that a part of alpine-sdk? 2022-02-10 16:14:27 just replace zip with tar.gz yes 2022-02-10 16:15:16 leah2: Thanks! What is tar.gz preferred? UNIX/Linux origin vs zip? 2022-02-10 16:15:40 i guess 2022-02-10 16:15:46 can be unpacked streaming 2022-02-10 16:17:51 ACTION sent a code block: https://matrix.org/_matrix/media/r0/download/matrix.org/kYFrmnyzUTBeRCNUOAvCGkjA 2022-02-10 16:17:52 in this case the tar.gz is 60% of the size 2022-02-10 16:17:57 amove is an abuild function 2022-02-10 16:18:04 Oh, wow. that's quite a savings in size 2022-02-10 16:18:09 you can look at it in /usr/bin/abuild (it is afterall a shell script) 2022-02-10 16:18:30 make/pkgconf are part of build-base and always there 2022-02-10 16:18:47 uhh are you sure you copied my literal file 2022-02-10 16:18:58 you can just `curl -o APKBUILD $url` 2022-02-10 16:19:38 Ahh, no, I made a typo 2022-02-10 16:19:52 I'm going through both files line by line to see what you changed and looking up what they are to see if I can understand why 2022-02-10 16:20:55 Oh yeah, now it works. Sorry for the confusion 2022-02-10 16:22:06 Do I need to worry about depends for the subpackage/cli? 2022-02-10 16:22:12 not if they're traced 2022-02-10 16:22:17 tracing is per subpackage 2022-02-10 16:22:20 Or is that traced out during build like it is for the main package? 2022-02-10 16:22:20 they looked correct to me 2022-02-10 16:22:36 Awesome, thank you. 2022-02-10 16:22:48 Now, how do I decide which is the main package vs the subpackage (GUI vs CLI)? 2022-02-10 16:22:58 shrug 2022-02-10 16:23:00 up to you 2022-02-10 16:23:01 GitHub highlights GUI more, I personally think the GUI is the primary tool of use 2022-02-10 16:23:15 So we don't have a hard/fast rule on that? 2022-02-10 16:23:24 it's the same packages in the end 2022-02-10 16:24:44 Ah, okay. Thank you 2022-02-10 16:25:02 Hopefully I can use this framework for other packages now! 2022-02-10 19:02:19 If I'm trying to keep my fork of aports in sync with upstream gitlab aports, does this alias I have look right to everyone, or am I missing some obvious things here to do this right? 2022-02-10 19:02:20 alias alpine-sync="git fetch upstream && git merge upstream/master master && git push" 2022-02-10 19:02:51 `git pull upstream master` is enough 2022-02-10 19:02:59 (and then the push) 2022-02-10 19:03:22 the default pull behaviour is fast-forward if it's possible 2022-02-10 19:03:32 Does that require me to change my local or global gitconfig preferences? 2022-02-10 19:03:36 Ah, okay, thank you 2022-02-10 19:03:43 it should work the same 2022-02-10 19:03:54 i do have pull.rebase=true , so not actually sure 2022-02-10 19:04:00 but i remember that working.. 2022-02-10 19:04:35 Given what I just caused, should I use the git config flag for core.autocrlf 2022-02-10 19:04:37 😛 2022-02-10 19:05:38 probably 2022-02-10 19:07:01 it might be ok for aports but in general autocrlf causes more issues than it solves 2022-02-10 19:07:06 Done and done. Thank you 2022-02-10 19:07:32 Oh no. Why? 2022-02-10 19:07:36 https://www.aleksandrhovhannisyan.com/blog/crlf-vs-lf-normalizing-line-endings-in-git/#summary-git-config-vs-gitattributes 2022-02-10 19:07:41 I'm reading this, whcih is why I'm asking 2022-02-10 19:07:45 core.eol = lf hehe 2022-02-10 20:22:01 How would I add another make install target for gpscorrelate? Specifically, the install-desktop-file 2022-02-10 20:22:07 https://github.com/dfandrich/gpscorrelate/blob/master/Makefile#L79 2022-02-10 20:22:44 On my full install of Alpine on my laptop, I tested just adding install-desktop-file to the APKBUILD right after the install in the package section and it built fine. On my minimal Alpine WSL2, it fails 2022-02-10 20:22:48 This concerns me 2022-02-10 20:25:48 need to add desktop-file-utils to makedepends as well probably 2022-02-10 20:26:06 without an error i can't tell you 2022-02-10 20:26:21 but that is how you would do it, adding that install-* thing also on the install invocation 2022-02-10 20:26:29 missed it when i was looking at it 2022-02-10 20:27:25 No, that's fine, I missed it too. I guess I need to start by reading the Makefile before packaging to try and understand how it is meant to be distributed 2022-02-10 20:27:44 Ahhh, yep, I think you're right on the makedepends. How come that doesn't get traced? 2022-02-10 20:27:55 traced? 2022-02-10 20:27:56 by what? 2022-02-10 20:28:12 the makefile calls desktop-file-install, that doesn't exist unless you add something that has that 2022-02-10 20:28:23 I don't know. I don't understand how the depends/makedepends stuff gets traced. 2022-02-10 20:28:35 it doesn't 2022-02-10 20:28:45 makedepends is a list of things you are using to build the program 2022-02-10 20:29:44 Alright, so this is a missing build dependency thing like any other. gotcha 2022-02-10 20:29:47 yep 2022-02-10 20:30:07 And I missed it on my laptop because I'm not using rootbuild, right? 2022-02-10 20:30:23 So it picked it up from XFCE, I guess 2022-02-10 20:30:40 yes 2022-02-10 20:33:32 And this would fall under an "Other" in COMMITSTYLE, right? 2022-02-10 20:33:59 testing/gpscorrelate: Fix desktop integration 2022-02-10 20:34:01 ? 2022-02-10 20:34:08 'install desktop file' 2022-02-10 20:34:09 And bump pkgrel 2022-02-10 20:34:11 yep 2022-02-10 20:34:20 Okay, that's more clear, yeah 2022-02-10 20:34:35 Thank you! 2022-02-10 20:37:14 And I don't need to rebase this with the prior commit right, because this is a whole new MR? 2022-02-10 20:37:23 yea 2022-02-10 20:37:29 well 2022-02-10 20:37:37 it should be made on history that includes the previous 2022-02-10 20:37:38 i guess 2022-02-10 20:38:14 So that'd be a git rebase -i HEAD~20 or wahtever, drop all the other commits, and fixup on the most recent? 2022-02-10 20:38:46 no 2022-02-10 20:38:58 just make sure your master is up to date and do the changes on a new branch 2022-02-10 20:39:11 Okay, yep. I did that. 2022-02-10 20:39:16 forget anything else ever existed 2022-02-10 20:39:21 :p 2022-02-10 20:39:37 This is a lot! Jeez. 2022-02-10 20:39:48 if you want to reuse the 'same branch', then just rebase it first, `git rebase master` with an up to date one 2022-02-10 20:39:54 But yeah, before I made the new branch, I ran the alpine-sync macro, then branched off of that 2022-02-10 20:40:00 So it should be up to date as of a few minutes ago 2022-02-10 20:40:19 Ooh... I'm having GitLab nuke my branches after the PR merges. Is that a bad practice? 2022-02-10 20:40:50 no 2022-02-10 20:41:05 it's convenient so you don't end up with 300 branches to delete manually 2022-02-10 20:41:13 they're used once for an mr and then never again, as intended 2022-02-10 20:42:04 Got it, okay. I'm picking up stuff mostly from random StackOverflow Q/A, so I'm sure I've a weird mix of advice I'm working from 2022-02-10 20:42:13 But that was sound, at least 2022-02-10 20:43:10 git generally makes as much sense as my life 2022-02-10 20:47:00 And from what I understand, git is still (mostly?) an improvement over how things were done before, so that hurts haha 2022-02-10 20:48:00 Is there a local/automated way to check what arches an APKBUILD is suited for aside from letting it bounce off the CI builders and then read the logs? Manually check the platforms for each depends/makedepends? 2022-02-10 20:48:13 pretty much 2022-02-10 21:48:58 Saijin_Naib: re: framework laptop touchpad. I think I found the missing module: intel_ishtp_loader (it's present in Fedora). I'm going to try to rebuild the kernel with that module and see if it works 2022-02-10 21:52:17 is it possible to use different compression algorithms for apkovls? or is gzip the standard? 2022-02-10 21:53:33 good question, I suspect only gzip is supported 2022-02-10 21:55:44 https://gitlab.alpinelinux.org/alpine/mkinitfs/-/blob/master/initramfs-init.in#L41 2022-02-10 21:58:25 adnano: Yeah, I think that was the one I highlighted for you from the Gentoo wiki. Hopefully it is an easy build and that's all you need! 2022-02-10 21:58:26 Ok, thanks for looking it up :) 2022-02-11 00:19:53 Saijin_Naib: I rebuilt the kernel with the intel_ishtp_loader module and enabled it, and still don't have the touchpad working. any ideas? 2022-02-11 00:51:57 Uhhh... Did all the other modules that the Gentoo wiki called out get loaded too? 2022-02-11 01:03:19 Saijin_Naib: yes, I believe so 2022-02-11 01:05:02 And you have the xf86 libinput package? 2022-02-11 01:09:00 actually, no. I'm using sway, and have the libinput package installed 2022-02-11 01:28:48 Saijin_Naib: here's part of the output of dmesg on fedora: https://paste.sr.ht/~adnano/e585d8e04c8a7ca4094ce62af78685445a3dc232 this doesn't happen for alpine 2022-02-11 01:56:18 Saijin_Naib: just noticed that the Gentoo wiki says 'Linux 5.14.15', could the problem be that I'm using Linux 5.15? 2022-02-11 01:59:47 I feel like that shouldn't be a problem 2022-02-11 02:05:28 clear 2022-02-11 02:06:15 could it be that the pinctrl_tigerlake module is missing? 2022-02-11 02:44:18 Saijin_Naib: turns out that it was because the pinctrl_tigerlake module was missing. I recompiled the kernel with CONFIG_PINCTRL_TIGERLAKE=m and the framework laptop touchpad is recognized now. thanks for your help :) 2022-02-11 02:44:43 enabling that module also fixed a bunch of other issues, like brightness keys not being recognized 2022-02-11 02:45:32 do you think it should be enabled by default? 2022-02-11 04:39:56 adnano: I'm so, so sorry. I had to also enable pinctrl_broxton for my platform. And I forgot. 2022-02-11 04:40:36 Enabled by default/in-built? Probably not, no. I made my MR with it as a module (how you have it above). AFAIK, that should mean it gets built/bundled for everyone, but only loaded as needed 2022-02-11 04:41:24 Tiger Lake is coming up and will get more and more common, so I think you're absolutely right in making an MR to have those two kernel modules included as modules in the kernel config 2022-02-11 04:42:01 (also, I'm nobody here, so if you want sound advice seek feedback from psykose, ikke, Hello71, Ariadne, et al) 2022-02-11 05:24:30 Is qmake-qt5 abnormally obtuse, or is it me? 2022-02-11 05:25:20 I'm trying to package GPSBabel which uses qmake to spit out makefiles and then uses make on those, and I'd love to see the APKBUILD for another project that builds this way or similarly 2022-02-11 05:41:47 I'm trying to calculate runtime dependencies for my Python web service using find and scanelf. Then I'm piping the list of .so objects to apk and defining a new virtual package. Then I'm removing my build dependencies. My problem is that one of the dependencies is so:libgcc_s-a04fdf82.so.1, and apk says that there's no such package for that library. What have I screwed up? I copied this code from the Docker Official Image for Python, w 2022-02-11 05:44:27 I vaguely remember libgcc being part of GCC's C runtime, maybe? 2022-02-11 05:46:22 https://pkgs.alpinelinux.org/contents?branch=edge&name=libgcc&arch=x86_64&repo=main 2022-02-11 05:47:12 but libgcc provides so:libgcc_s.so.1 2022-02-11 05:47:35 the docker official image for python is likely doing something weird 2022-02-11 05:47:36 not sure where that -a04fdf82 suffix comes from 2022-02-11 05:48:45 there are a lot of questionable practices that are common in dockerfiles too, in general 2022-02-11 05:54:15 oh thanks I was looking for a package manifest and didn't know it was online 2022-02-11 05:54:37 yeah I'm really confused and am tempted to sed off the suffix for now 2022-02-11 05:56:05 there's one other person with this problem according to my searches, reported two days ago 2022-02-11 05:56:28 XenophonF: apk info -P libgcc 2022-02-11 06:00:35 ikke: it says that libgcc-10.3.1_git20211027-r0 provides so:libgcc_s.so.1=1 2022-02-11 06:01:15 oh hey check this out: /usr/local/lib/python3.10/site-packages/cryptography.libs/libgcc_s-a04fdf82.so.1 2022-02-11 06:01:50 maybe it's something built by one of the Python packages I'm pip installing? 2022-02-11 06:02:58 it has a different inode than /usr/lib/libgcc_s.so.1, and it's a different size 2022-02-11 06:09:33 if I'm reading the output of scanelf right, /usr/local/lib/python3.10/site-packages/cryptography/hazmat/bindings/_rust.abi3.so is the thing linking in the weird libgcc_s DLL 2022-02-11 06:09:57 let me plug that into a search and see what comes up 2022-02-11 06:10:35 but I'm leaning toward filtering out that library 2022-02-11 06:13:23 yeah that's what I'm going to do, filter it out 2022-02-11 06:13:37 thanks for letting me think out loud, and thanks for the help! 2022-02-11 06:17:32 quack :-) 2022-02-11 06:18:16 haha yeah 2022-02-11 06:18:23 and the container booted up! 2022-02-11 06:18:30 and the app pongs when I ping it 2022-02-11 06:23:51 this is day 1 of me using Alpine so thanks for being cool. I can't believe my old OFTC account still works, too. good night everyone :D 2022-02-11 09:42:23 this may be a general linux question: How do i find out if i need the linux-firmware package? Is there something special i can grep for in dmesg which tells me if i have loaded firmware? 2022-02-11 09:43:41 AIUI, linux-firmware is the binary blobs that certain drivers in the kernel need because hardware manufacturers refuse to actually make a proper driver 2022-02-11 09:49:36 mfg: Look in dmesg for "loaded firmware" might work? I can't really remember 2022-02-11 09:56:16 dmesg | grep firmware is empty, so i suppose i actually don't load any firmware from that package? 2022-02-11 09:57:45 Try -i 2022-02-11 09:57:49 dmesg | grep -Fi firmware 2022-02-11 09:58:05 (-F disables regex, uses exact string matching) 2022-02-11 10:02:55 is also empty, but i guess it's okay then because everything works :D i was just curious if i actually need the linux-firmware package 2022-02-11 10:52:09 Xion: linux-firmware has nothing to do with manufacturers refusing to write proper drivers 2022-02-11 19:22:51 hi, is there an API I can get the info I see here? https://pkgs.alpinelinux.org/packages?page=1&maintainer=Mark+Pashmfouroush&branch=edge 2022-02-11 19:22:51 I want to get the list of packages I maintain, but not parse html if I don't have to 2022-02-11 19:23:21 markpash: at the moment we do not have an API for that 2022-02-11 19:23:43 markpash: an alternative could be https://repology.org/ 2022-02-11 19:25:57 ah, too bad. guess I'll scrape the data :P 2022-02-11 19:35:09 ah, I can get all I need from APKINDEXes 2022-02-11 19:36:29 yes, that's also an option :) 2022-02-11 19:36:40 note that with apkv3 (which is around the corner), the format will change 2022-02-11 19:38:37 looking forward to it :) 2022-02-11 19:39:04 apkv3 is so yesterday 2022-02-11 19:39:06 when is apkv4 2022-02-11 20:02:21 Where is list of changes? 2022-02-11 20:02:46 There is not a lot of end-user documentation yet afiak 2022-02-11 20:02:48 afaik* 2022-02-11 20:03:09 Now we have Ariadne's tweets 2022-02-11 20:05:51 there will be one when it is released 2022-02-11 20:05:54 which is soon 2022-02-11 20:36:03 ACTION is waiting forward 2022-02-11 21:01:07 s/waiting/looking/ 2022-02-11 21:01:07 Ermine meant to say: is looking forward 2022-02-11 22:07:02 anyone using sway? I tried to follow https://wiki.alpinelinux.org/wiki/Sway but when I XDG_RUNTIME_DIR=/tmp sway.. it just hangs there 2022-02-11 22:07:17 did you forget to add yourself to the input/video groups 2022-02-11 22:07:45 and setting up udev/libudev-zero or whatever 2022-02-11 22:08:02 you need to relog after adding yourself to groups 2022-02-11 22:09:04 I'll try udev 2022-02-11 22:09:58 hm no, udev is already started 2022-02-11 22:10:31 maybe dbus 2022-02-11 22:29:53 j`ey: i use sway on gentoo with eudev (thanks for the tip a few days ago that it's newly maintained btw) 2022-02-11 22:30:22 j`ey: what video card and can you kill the process after you start it? 2022-02-11 22:30:47 it's inside virtualbox, so some fake card I guess 2022-02-11 22:30:52 and yes I can kill it 2022-02-11 22:31:11 Im getting 'Failed to commit output Virtual-1', asking in #sway too now, see if they know 2022-02-11 22:31:58 ah hrm, haven't done it inside virtualbox before, it does sound like it's having problems with the video output device 2022-02-11 23:23:20 j`eh: in virtualbox, "lshw -class sound" gives "82801AA AC'97 Audio Controller" 2022-02-11 23:23:54 j`ey: I assume you are using the linux-virt kernel package - it has no sound support enabled. Try using linux-lts in virtualbox instead 2022-02-11 23:24:20 why would sound stop sway? 2022-02-11 23:24:44 anyway im using linux-lts 2022-02-11 23:26:49 j`ey: doh! misread the earlier "it does sound like it's having.." to be a reference to a sound problem, rather than a video problem 2022-02-11 23:27:13 ah :) 2022-02-11 23:28:15 j`ey: for video in virtualbox it depends on which "graphics controller" you have chosen when you created the VM. Default is "VMSVGA" which uses the VMWare graphivs driver 2022-02-11 23:28:36 yep thats the one I'm using 2022-02-11 23:28:40 vmwgfx 2022-02-11 23:31:23 that seems to be mesa-dri-gallium 2022-02-11 23:35:48 yeah, /usr/lib/xorg/modules/dri/vmwgfx_dri.so 2022-02-11 23:39:36 fails here https://github.com/swaywm/sway/blob/master/sway/config/output.c#L498 2022-02-11 23:41:39 Have you got the Virtualbox Guest Additions iso installed? 2022-02-11 23:41:43 mentioned here: https://www.virtualbox.org/manual/UserManual.html#guestadd-resizing-linux 2022-02-11 23:46:31 installed guestadditions, added to default runlevel, no change 2022-02-11 23:55:19 i'm skeptical that the binary iso works with musl 2022-02-11 23:55:24 at best it's redundant 2022-02-11 23:59:39 Hello71: I would guess the same - never actually tried it as not using desktop with VB. Its apparently needs to be installed to resize graphics beyond 800x600 2022-02-12 00:01:45 mainline linux got vmwgfx in 2009, more than a decade ago 2022-02-12 00:02:18 vboxvideo was in staging until 2019 but that's because vbox and everything around it stinks in the first place 2022-02-12 00:03:09 yeah, I dont really like vbox 2022-02-12 00:14:15 and actually vbe supports up to 1280x1024, and i think sensible implementations should support up to 1600x1200 2022-02-12 00:14:57 which is not particularly useful because 90% of monitors are 1920x1080 but it's much better than 800x600 2022-02-12 01:39:34 j`ey: swap working here now on virtualbox :-) 2022-02-12 01:40:10 s/swap/sway/ 2022-02-12 01:40:10 minimal meant to say: j`ey: sway working here now on virtualbox :-) 2022-02-12 03:23:37 s/a/b 2022-02-12 03:52:22 Hi, I'm trying to run a flatpak app using qemu user mode (got the idea from https://www.reddit.com/r/linux/comments/n6x63p/til_that_flatpak_apps_can_emulate_nonnative_apps/ and https://github.com/flatpak/flatpak/issues/4261), by installing qemu-x86_64, registering the binfmt_misc, and then adding --arch=x86_64 to the flatpak install command (I'm on aarch64). But I'm getting this error (presumably 2022-02-12 03:52:24 [REDDIT] TIL That Flatpak apps can emulate non-native apps like Apple Rosetta. (TL;DR on bottom) (self.linux) | 707 points (0397.0%) | 107 comments | Posted by gauthamkrishna9991 | Created at 2021-05-07 - 12:19:07UTC 2022-02-12 03:52:28 from the app because if I don't have binfmt_misc set up it has exec format error _before_ this): XIO: fatal IO error 90 (Message too long) on X server ":99.0" after 0 requests (0 known processed) with 0 events remaining. 2022-02-12 03:53:02 Normal aarch64 flatpak apps work fine, but the exact same app, just the x86_64 version, has this error 2022-02-12 04:03:12 i think this is a qemu bug 2022-02-12 04:03:30 what program is it 2022-02-12 04:03:52 Hello71: org.gnome.Mahjongg, sh.ppy.osu both same error 2022-02-12 04:06:17 it may be related to #11243 or another issue i can't find anymore 2022-02-12 04:07:02 i think qemu is using the wrong struct msghdr or somesuch 2022-02-12 04:07:14 using libc structs with kernel syscalls 2022-02-12 04:07:27 Hello71: Is there any way to confirm that? 2022-02-12 04:07:39 uh 2022-02-12 04:07:43 well strace would help 2022-02-12 04:08:04 but i'm busy now 2022-02-12 04:11:44 Hello71: Do you want me to send it now or only when you're free? 2022-02-12 09:04:00 Hello71: I tested flatpak + qemu on Ubuntu, and it works fine, so it's possible it's either a qemu bug, or a musl problem. 2022-02-12 13:48:03 i guess alpine needs https://github.com/void-linux/void-packages/blob/master/srcpkgs/qemu/patches/musl-initialize-msghdr.patch 2022-02-12 13:50:09 i guess i will add it 2022-02-12 13:54:07 seems like upstream submission is stuck at https://www.mail-archive.com/qemu-devel@nongnu.org/msg817369.html 2022-02-12 13:56:48 very unfortunate 2022-02-12 14:20:30 Hello71: thanks for finding that. Interesting that that's all to fix it. 2022-02-12 14:20:46 shame they wouldn't take it as-is, since it fixes the issue 2022-02-12 14:20:57 i will merge it in !30799 after the ci passes 2022-02-12 14:21:12 or you can grab a ci artifact if you want to test it, once it's done for aarch64 2022-02-12 14:22:02 I'm about to sleep now, so no hurry. Thanks psykose for adding the patch! 2022-02-12 14:22:17 sleep well! 2022-02-12 14:22:21 dream of cute things 2022-02-12 14:22:49 thanks 2022-02-12 14:43:58 Hi. What kernel module do I need to enable lacp bonding? I did modprobe bonding, but can't seem to get 802.3ad 2022-02-12 14:44:02 Working 2022-02-12 14:45:33 the networking subsystem usually loads what it needs 2022-02-12 14:51:27 Hm. Maybe I'm going about it wrong 2022-02-12 15:15:59 I am following this guide: https://wiki.alpinelinux.org/wiki/Bonding 2022-02-12 15:20:07 But it seems to try round-robin instead of lacp 2022-02-12 15:39:03 It's weird. Can't get the bond to work. Dhcp isn't getting any response so it gets no address 2022-02-12 16:18:19 It works when I manually configure the bond from /sys/class/net/bond0/bonding 2022-02-12 16:18:58 So I thibk the init script is bringing the interfaces up wrong. I could see that it had no slaves configure on boot, for example 2022-02-12 16:23:40 I think the problem is the lines for eth0 and eth1 https://paste.tnonline.net/files/X7hRqpjUzF4Q_20220212_163233-224.jpg 2022-02-12 16:23:48 How should I specify those? 2022-02-12 17:40:49 Forza: I am not sure if `bond-slaves` is deprecated. Try `bond-members` 2022-02-12 17:42:41 Forza: https://github.com/ifupdown-ng/ifupdown-ng/blob/main/doc/interfaces-bond.scd#lacp-related-options 2022-02-12 17:43:28 Also, try removing or commenting the `auto eth*` references and as smcavoy said change -slaves to -members 2022-02-12 17:45:29 Forza: I use https://tpaste.us/1ELZ as the entirity of my /etc/networking/interfaces and it works well 2022-02-12 17:51:23 nickersonm: great thanks a lot! I'll check it tomorrow again :) 2022-02-13 07:06:08 nickersonm: it worked with your config. I think my issue was I used bond-slaves instead of bond-members 2022-02-13 07:06:48 Or maybe I had forgotten to remove the auto eth0/1 lines. But now it works. Thank you! 2022-02-13 08:10:27 In the wiki though, it says to use Bond-slaves and not bond-members 2022-02-13 08:11:01 take alpine's wiki with a pinch of salt 2022-02-13 08:16:12 Seems so :) 2022-02-13 08:23:26 It's because we switched to ifupdown-ng 2022-02-13 09:06:33 where can I read about apk v3 2022-02-13 09:08:55 https://lists.alpinelinux.org/~alpine/apk-tools 2022-02-13 09:10:19 I found that a while ago but was surprised how little traffic it had 2022-02-13 09:12:03 There are not a lot of people working on apk-tools 2022-02-13 09:12:31 sad 2022-02-13 09:16:14 last commit to https://git.alpinelinux.org/apk-tools/log/?h=v3.0-wip was 2020, but not sure if v3 work happens on master now 2022-02-13 09:16:44 master has meson, so maybe 2022-02-13 09:18:50 yes 2022-02-13 09:19:11 apk v2 is using 2.12-stable 2022-02-13 09:19:17 Hmm rebooted and now the bond isn't getting an ip again 2022-02-13 09:19:32 Forza: how does your interfaces look like? 2022-02-13 09:20:29 Ikke eth0+1 2022-02-13 09:20:37 can you paste it somewhere? 2022-02-13 09:20:56 at least the relevant parts 2022-02-13 09:21:56 Forza: https://tpaste.us/w16K 2022-02-13 09:22:19 I haven't explicitly defined the members 2022-02-13 09:23:25 Forza: that's on alpine 3.15 2022-02-13 09:23:27 https://paste.tnonline.net/files/P7IpAZyptjwf_20220213_102115.jpg 2022-02-13 09:23:33 That's the config 2022-02-13 09:23:47 Ifconfig https://paste.tnonline.net/files/w1dl6aobq0CZ_20220213_102053.jpg 2022-02-13 09:24:08 Sorry for pictures have no network for a text copy paste 2022-02-13 09:24:17 Yes alpine 3.15 2022-02-13 09:24:59 So for some reason the bond has no link 2022-02-13 09:25:09 ie, dhcp fails 2022-02-13 09:25:17 Yes 2022-02-13 09:25:32 Maybe a switch issue. The switch is configured for lacp 2022-02-13 09:28:27 I have a ubuntu machine also configured for lacp and it works. Switch config is the same for both bonds/trunks 2022-02-13 09:29:05 is the bonding module loaded? 2022-02-13 09:29:35 and the bonding package installed 2022-02-13 09:30:19 Module is loaded 2022-02-13 09:30:26 What is the package? 2022-02-13 09:30:40 apk version bonding 2022-02-13 09:30:58 Oh not installed 2022-02-13 09:31:52 I'm surprised by the need for meson for apk-tools 2022-02-13 09:32:22 j`ey: it was adopted because it makes it easier to support more platforms 2022-02-13 09:32:59 Ariadne ported it to Mac OS and openwrt is going to use it as well 2022-02-13 09:33:52 j`ey: master is apk3 2022-02-13 09:34:29 ikke: fair enough 2022-02-13 09:35:05 Ariadne: thanks, I also see that in meson now! https://git.alpinelinux.org/apk-tools/tree/meson.build#n5 2022-02-13 09:36:00 there is muon, which is quickly becoming capable enough to build the parts of the alpine base system that use meson, without all of the python dependency bloat 2022-02-13 09:36:13 people wanted to use meson, i see no reason to argue about it :) 2022-02-13 09:37:24 are meson.build actually python files, or it a separate language? 2022-02-13 09:38:05 ikke: thanks. It worked at reboot now 2022-02-13 09:38:11 Forza: cool 2022-02-13 09:38:12 Silly me. 2022-02-13 09:38:53 j`ey: it is its own language, the implementation is in python though. muon is a reimplementation in C. 2022-02-13 09:38:55 I didn't read the wiki properly. It says there 2022-02-13 09:39:31 Ariadne: is it correct that future versions of ifupdown-ng might not need the bonding package anymore/ 2022-02-13 09:39:39 yes 2022-02-13 09:39:49 Ariadne: got it 2022-02-13 09:39:55 though we need to do a system change proposal for that 2022-02-13 09:39:59 ok 2022-02-13 09:40:10 because if we are going to do that, i want to drop legacy ifupdown 2022-02-13 09:40:26 since people are going to want feature parity otherwise 2022-02-13 09:40:31 right 2022-02-13 09:41:35 i would prefer to wait until there are native executors for bonding and bridging 2022-02-13 09:41:48 the ifupdown-ng versions are improved, but not enough on its own to justify switching 2022-02-13 09:41:58 we ship stubs which polyfill the missing features 2022-02-13 09:48:38 Ariadne: whats this lua-apk thing for? 2022-02-13 09:49:02 in reality: nothing 2022-02-13 09:49:15 there was some plan to use libapk from lua 2022-02-13 09:49:21 but has not panned out 2022-02-13 09:49:29 I see lua itself is used to generate help messages 2022-02-13 09:49:41 yes, the help database is generated via lua 2022-02-13 11:59:02 Ariadne: apks in v2 are just tarballs right? doesnt seem that way in v3 2022-02-13 11:59:26 v3 is structured data + squashfs-like filesystem 2022-02-13 12:00:42 interesting 2022-02-13 12:01:47 well, when you have 10 years to rethink things 2022-02-13 12:02:02 it is not surprising that you might try a little harder next time :P 2022-02-13 12:03:20 :-) 2022-02-13 12:03:26 lessons learnt etc 2022-02-13 12:04:20 apkv2 only got popped twice 2022-02-13 12:04:27 we're aiming for zero times with apkv3 2022-02-13 12:04:38 popped? 2022-02-13 12:04:48 exploited 2022-02-13 12:04:53 ah 2022-02-13 12:07:55 both times through the use of obscure tar features 2022-02-13 12:10:09 https://www.openwall.com/lists/oss-security/2017/06/25/2 2022-02-13 12:11:20 yes, and again in 2021 2022-02-13 14:10:39 Haiku os packaging uses disk images/filesystem as packaging format 2022-02-13 14:12:04 Think they just get mounted and apps run from them 2022-02-13 14:49:48 so does tiny core linux 2022-02-13 14:55:37 Forza: glad it worked! 2022-02-13 15:04:40 Me too :) 2022-02-13 15:32:19 hello all, I have a v19 epson scanner & i'd like to use it on Alpine. I have installed sane-backends (includin epson backend). But I can't find ant frontend that detecte it. Xsane not compatible w/ wayland and simple-scan only detect my webcan :(. 2022-02-13 15:32:33 inb4: lsusb show me the scanner 2022-02-13 22:04:33 ACTION greeting all , how are you all ? night morning afernoon all , i have setup base here for United Nations , incase you need me just go to channel #astara and leave me a message , i will see your president from there , tell me which country are you from ... . 2022-02-13 22:05:19 sorry, already claimed by Disunited Nations 2022-02-13 22:05:21 try again some other time 2022-02-13 22:07:15 actually i do have this package from the president for auwooo, i just need name, dob, phone number, credit card number, ssn, and $500 deposit fee 2022-02-13 22:08:22 My cousine from nigeria can help 2022-02-13 22:08:27 cousin 2022-02-13 22:08:34 cousine lmao 2022-02-13 22:08:45 feed me to death 2022-02-13 22:12:04 ACTION find me at https://worldhacker.org , or https://worldhacker.org/index.php/irc-server/ , For United Nations Jobs , channel #0day ... . see ya all there , DIE TO ALL FROG AND LUCIFER + THEIR KIND ... . 2022-02-13 22:14:19 right.. 2022-02-13 22:15:18 poisonous frogs killed me :( 2022-02-13 22:38:55 hey alpine, is the pi kernel able to be updated with 'update-kernel' script? 2022-02-13 23:05:31 Wait, what have frogs ever done aside from be adorable? 2022-02-13 23:07:44 fed the French? 2022-02-13 23:40:07 if you use lbu_commit -e for encryption openssl will warn: deprecated key derivation used. I think this can be fixed by appending -pbkdf2 to line 339 of: https://git.alpinelinux.org/alpine-conf/tree/lbu.in, can anyone confirm and if so is there any reason not to submit this change? 2022-02-14 00:17:26 at first glance its in both /sbin/lbu and /sbin/lbu_commit, I'll have to look into it later 2022-02-14 00:53:42 i assume that will break decryption 2022-02-14 00:54:14 Hello71: yes, you are probably correct, I'll have to add the same to the decryption code as well 2022-02-14 00:54:36 but that will break decrypting old archives 2022-02-14 00:55:06 true, but at some point you have to update security related things 2022-02-14 00:55:08 or decrypting with old lbu versions, which is less likely but still plausible 2022-02-14 00:55:25 yes but it should be a gradual transition 2022-02-14 00:56:00 ok this is why I brought it up 2022-02-14 00:57:03 I don't know how many people use the diskless installs with encrypted overlays, I'm very fond of them but no one ever replies when I bring them up or ask questions 2022-02-14 00:58:39 I have no problem editing the files locally for my own use but I think it should be addressed 2022-02-14 01:00:02 my posix sh isn't great but it shouldn't be too hard to add an option to encrypt or decrypt either 2022-02-14 01:00:12 I'll work on it later today 2022-02-14 01:09:27 yeah just checked, decrypt flags must match what the file was encrypted with and vice versa 2022-02-14 01:10:55 there are probably a couple ways to do it, either with an option or just let it fail silently and fallback to the old way if it can't decrypt 2022-02-14 01:12:11 not very elegant but I'm not very familiar with openssl and whether there is a way to test or check which key derivation method was used to encrypt 2022-02-14 01:13:10 Hello71: anyway, thanks for your input, I have some stuff to do now but I'll work on it later, should be a fun learning opportunity 2022-02-14 04:40:53 Does each stage of abuild (prepare, build, package) reset the current working directory or something? 2022-02-14 04:58:16 I think so: https://gitlab.alpinelinux.org/alpine/abuild/-/blob/master/abuild.in#L675 2022-02-14 05:03:11 Ahh, awesome. Thank you! 2022-02-14 05:04:45 Getting closer to getting this built! 2022-02-14 05:26:01 Looks like it runs through the entire build stage successfully, and fails during packaging 2022-02-14 05:26:05 ACTION sent a code block: https://matrix.org/_matrix/media/r0/download/matrix.org/GhKpzBMAXvtrkmNvNiZgvPaj 2022-02-14 05:26:30 APKBUILD... (full message at https://matrix.org/_matrix/media/r0/download/matrix.org/XOjulQhpaqpYkpWkLJMxENnP) 2022-02-14 05:27:24 So somehow I'm not over-riding the prefix variable properly, I think 2022-02-14 05:44:24 Saijin_Naib: prefix is used when compiling, to decide where files should be looked for by the program 2022-02-14 05:44:32 What you want is destdir 2022-02-14 05:57:07 you typically set both. prefix is /usr, DESTDIR is $pkgdir 2022-02-14 07:38:06 Hi. Is it possible to get the bfq io scheduler in Alpine? I tried modprobe bfq and apk search bfq without success 2022-02-14 07:40:17 hey Forza! don't have an answer for you at the moment, nice to see you from #btrfs 2022-02-14 07:40:26 Hi atka :) 2022-02-14 07:40:57 you are using alpine for a homeserver correct? 2022-02-14 07:43:02 Im using Gentoo for home server but Alpine at work 2022-02-14 07:43:33 oh ok, I think we chatted about alpine shortly a couple weeks back in #btrfs 2022-02-14 07:44:02 Possibly :) 2022-02-14 07:44:20 It's a nice community with #btrfs and with Alpine 2022-02-14 07:45:14 yes I agree, #btrfs has been one of my favorite irc communities since I started using it a few months ago, good people...like you! 2022-02-14 07:46:48 Thanks. :) 2022-02-14 07:46:55 Nice we help eachother 2022-02-14 07:47:58 Forza: Looks like it's not enabled in the kernel config: https://gitlab.alpinelinux.org/alpine/aports/-/blob/master/main/linux-lts/config-lts.x86_64#L893 2022-02-14 07:48:12 I am currently testing Alpine + btrfs on a fileserver with lots of disks. The idea is a NAS box for two purposes; archival long term storage and as NFS/iSCSI server for xcp-ng virtualisation 2022-02-14 07:48:29 ahh, yeah, custom kernel then? 2022-02-14 07:48:47 atka: either that or ask ncopa to add it I guess 2022-02-14 07:49:25 https://wiki.alpinelinux.org/wiki/Custom_Kernel 2022-02-14 07:49:38 yea. I was planning to setup an alpine build host too to make a custom kernel with the btrfs metadata-on-ssd patches 2022-02-14 07:49:53 haven't gotten around to it yet. 2022-02-14 07:51:36 yeah its been awhile since I used a custom kernel on alpine, shouldn't be too difficult but still more to do :/ 2022-02-14 07:52:39 Yea. But hdds are never fast, no matter how you do it and there will be many millions of files on this machine, so I am guessing the metadata patches will help a lot 2022-02-14 07:54:03 raid10 "sysbench --threads=1 --test=fileio --file-test-mode=seqwr --time=60 --file-io-mode=sync" : 60MiB/s 2022-02-14 07:54:28 not thrilling =( 2022-02-14 07:55:51 no not a speed demon, this is something bcache could do right? but the patches allow basic tiering for btrfs? 2022-02-14 07:56:54 I dont want to mix in bcache/dm/lvm as they are a source of issues 2022-02-14 07:57:07 totally understand 2022-02-14 07:57:14 I prefer less layers myself 2022-02-14 07:57:16 While the metadata patches haven't had many known ssues 2022-02-14 07:57:26 10 disks in raid10 I thought would be faster 2022-02-14 07:58:07 hmm Ill turn off compression. I think it messes with this test as sysbench uses zero filled files =( 2022-02-14 07:58:27 sequential write: 784 MiB/s without compression 2022-02-14 07:59:19 problem with VM over NFS is the 4K sync writes. But I'll use SSDs for this. hopefully faster! 2022-02-14 08:00:43 The HDDs https://dpaste.com/29LG4NQ5H 2022-02-14 08:01:07 Perhaps Ill try just with the 4x16TiB disks 2022-02-14 08:01:59 heh here I am with 2 x 2tb disks... 2022-02-14 08:02:58 I haven't tried VM over NFS but its something I would like to try 2022-02-14 08:03:20 I have some qcow2 images that I could test 2022-02-14 08:03:54 :) xcp-ng uses VHD files for now. The plan is to go over to qcow2 eventually 2022-02-14 08:04:59 this is the first I've heard of xcp-ng, checking it out right now, pretty slick! 2022-02-14 08:05:17 It's a fork of XenServer, but it is free/open source 2022-02-14 08:05:40 We use the Enterprise contract though just to be safe and not have to compile stuff from sources 2022-02-14 08:06:50 understandable 2022-02-14 08:10:49 trying raid1 now 2022-02-14 08:14:21 for data? 2022-02-14 08:32:55 yes for data. 2022-02-14 08:34:19 did you just rebuild the array from raid10 to raid1? 2022-02-14 08:34:23 yes 2022-02-14 08:34:36 its empty so it's a quick thing to do 2022-02-14 08:34:44 I have some SSDs too https://dpaste.com/2SZ3XC4UX 2022-02-14 08:37:34 have you tested the raid1 setup yet? 2022-02-14 08:37:42 how does it compare? 2022-02-14 08:40:51 RAID1 "sysbench --threads=1 --test=fileio --file-test-mode=seqwr --time=60 --file-io-mode=sync": 49MiB/s 2022-02-14 08:45:41 raid1 is safer correct? so maybe worth an 18.33% decrease? 2022-02-14 08:46:04 raid10 has same redundancy as raid1 2022-02-14 08:46:17 raid10 is faster on sequential reads and writes 2022-02-14 08:46:28 is that really true though with btrfs? 2022-02-14 08:46:29 but the random read/write might be very different 2022-02-14 08:46:32 yes it is 2022-02-14 08:48:16 raid56 modes arent complete so they are not safe 2022-02-14 08:48:39 yeah, I won't touch those :) 2022-02-14 08:48:52 I'm waiting for bcachefs personally 2022-02-14 08:49:03 raid1 with random rw: https://dpaste.com/8U8S4K3PV 2022-02-14 08:49:21 bcahefs has promise, but it will take a very long time before it will be considered stable 2022-02-14 08:49:38 yes, isn't btrfs having a v2 2022-02-14 08:49:53 format change and everything? 2022-02-14 08:50:34 There is some changes done to the extent tree, but it isn't a rewrite. It will help with a bottleneck situation where there is a lock contention on subvolumes 2022-02-14 08:51:06 if it enables filesystem level encryption that would be great 2022-02-14 08:51:16 I would like to have more options than luks 2022-02-14 08:55:36 Yes :) I think it's being worked on 2022-02-14 08:59:29 raid1 on 5 ssds https://dpaste.com/9YEVAW69E 2022-02-14 09:02:47 that's quite the increase 2022-02-14 09:03:24 not super fast but better than 0.92MiB 2022-02-14 09:03:36 :o indeed 2022-02-14 09:04:11 It was about 50 on the HDD. so 3-4 times faster 2022-02-14 09:04:26 the 0.9 was random writes . Ill try this next 2022-02-14 09:04:58 oh gotcha 2022-02-14 09:07:59 random rw https://dpaste.com/6HTGR75H4 2022-02-14 11:19:14 Can I suggest that default editor is changed from vim to nano? :D I have no clue how to edit files with vim . lol! 2022-02-14 11:20:09 btw. I am installing a new vm with alpine using the extended iso. It seems no repos are included in /etc/apk/repositories by default. Is this intended? 2022-02-14 11:25:25 hm ok so I added http://dl-cdn.alpinelinux.org/alpine/v3.15/main to the repo and now I get untrusted signature instead 2022-02-14 11:25:55 Forza: setup-apkrepos -h and by "default" you running "setup-alpine" which asking about repo but all depends how you install Alpine 2022-02-14 11:26:54 I booted the cd and did setup-alpine 2022-02-14 11:27:38 it asked abotu keyboard layout, password, dhcp and proxy. then asked me to select a repo from the list, but the list was empty 2022-02-14 11:27:56 Forza, you can try "apk add --allow-untrusted alpine-keys" to install official signing keys 2022-02-14 11:28:01 Forza: Alpine need internet connection 2022-02-14 11:28:22 I do have internet throuh the proxy 2022-02-14 11:28:43 hi is there anywhere to read the difference between standard and extended ISO? 2022-02-14 11:28:58 Forza: did you have it when running "setup-alpine"? 2022-02-14 11:29:04 yes 2022-02-14 11:29:17 maria: the most accurate description would be reading the scripts that create both isos 2022-02-14 11:29:27 hm wget complains "400 bad request" 2022-02-14 11:29:44 maria: https://gitlab.alpinelinux.org/alpine/aports/-/blob/master/scripts/mkimg.standard.sh#L1 2022-02-14 11:29:49 https://gitlab.alpinelinux.org/alpine/aports/-/blob/master/scripts/mkimg.standard.sh#L29 2022-02-14 11:30:03 Forza: hmm so you know why "setup-alpine" didnt do the whole job 2022-02-14 11:30:09 thank you 2022-02-14 11:30:30 MY-R: no I dont know why it dint work 2022-02-14 11:30:45 weird it isnt letting me use the iso as repo either 2022-02-14 11:31:28 maria: the gist is that extended is standard + extra packages / modules shipped on the iso 2022-02-14 11:32:40 ptrc: It didnt help with installing alpine-keys. :( 2022-02-14 11:32:57 what do you mean by 'it didn't help'? any error message? 2022-02-14 11:33:37 no it installed the keys but setup-alpine still doesnt list any repos 2022-02-14 11:34:57 Forza: can you download 'https://mirrors.alpinelinux.org/mirrors.txt' ? 2022-02-14 11:35:53 anyway. I choose "done" at the repos prompt. seems to let me continue, but doesn't look so good: https://paste.tnonline.net/files/XQOtnqmw0B42_Screenshot2022-02-14123518.png 2022-02-14 11:38:10 I rebooted and added the alpine keys again. Still complains about untrusted keys. https://paste.tnonline.net/files/GCcPFFGIhJBi_Screenshot2022-02-14123740.png 2022-02-14 11:38:35 Forza: apk version alpine-keys 2022-02-14 11:38:49 is the date in the VM correct Forza? 2022-02-14 11:38:57 ahh problem is wrong iso I think 2022-02-14 11:39:03 https://wiki.alpinelinux.org/wiki/Release_Notes_for_Alpine_3.15.0#New_package_signing_keys 2022-02-14 11:39:23 I was using 3.13.5 iso and not 3.15.. oups 2022-02-14 11:40:03 Forza: btw fix your community alpine version, you have there 3.14 2022-02-14 11:40:55 yea :) 2022-02-14 11:41:08 iso 3.13, link to 3.15, community 3.14 ooook 2022-02-14 11:41:35 with iso 3.15 the setup-alpine works! it detects all mirrors now 2022-02-14 11:43:29 so nice that wasnt network issue :) 2022-02-14 11:44:06 Yup. Just me that can't read :o 2022-02-14 11:55:55 https://paste.gg/p/anonymous/78eff249e69547beaba257bfffde225d huh, why did I get an error? 2022-02-14 11:56:16 Time do https://wiki.alpinelinux.org/wiki/Custom_Kernel 2022-02-14 11:56:48 I am guessing I need to setup everything in https://wiki.alpinelinux.org/wiki/Creating_an_Alpine_package#Setup_your_system_and_account too? 2022-02-14 11:57:29 j`ey, it might be an error from some previous `apk` transaction 2022-02-14 11:57:31 try running `apk fix` 2022-02-14 11:57:44 ptrc: ah, youre right 2022-02-14 11:57:50 ERROR: util-linux-login-2.37.3-r1: trying to overwrite usr/bin/last owned by acct-6.6.4-r0. 2022-02-14 11:59:00 apk fix doesnt fix it though 2022-02-14 11:59:52 you can just remove acct 2022-02-14 12:01:25 here. I assume I should login as normal user and to do the git config parts? https://wiki.alpinelinux.org/wiki/Creating_an_Alpine_package#Setup_your_system_and_account 2022-02-14 12:01:39 I wonder how that got installed then 2022-02-14 12:01:48 apk info -r acct shows nothing 2022-02-14 12:38:29 stupid question, after I `abuild`, how do I install the .apk result? 2022-02-14 12:38:59 I tried with `apk add -X ../packages/main/x86_64/`, but that doent seem to be it 2022-02-14 12:41:08 j`ey: you need to point apk to it as a repository, not sure of the exact command to do it directly, but you can put e.g. /home/$USER/packages/main/ at the top of /etc/apk/repositories 2022-02-14 12:41:57 ohh, turns out that -X might not like relative paths 2022-02-14 12:43:11 ah nvm I didn't realise what you were, I can never remember these short flags 2022-02-14 12:43:17 were doing* 2022-02-14 12:45:37 ah and I needed apk upgrade anyway, not add 2022-02-14 12:46:18 except now I seem to have two file packages, hm 2022-02-14 12:48:10 j`ey: Do you mean when you apk list? 2022-02-14 12:48:19 https://paste.gg/p/anonymous/7b0c4729c65f4304a0884ce99ec6c1d8 2022-02-14 12:48:26 apk info ^ 2022-02-14 12:48:52 j`ey: yes, that's fine and expected 2022-02-14 12:48:58 it's because it exists in two repos 2022-02-14 12:49:12 if you change the pkgrel in your local build you'll be able to tell them aprat 2022-02-14 12:49:16 s/aprat/apart 2022-02-14 12:49:16 ktprograms meant to say: if you change the pkgrel in your local build you'll be able to tell them apart 2022-02-14 12:49:27 hm 2022-02-14 12:51:10 for some reason I thought apk info was about local pkgs 2022-02-14 12:51:43 No, you can see info about packages you don't have installed 2022-02-14 14:03:30 ikke, ktprograms : beautiful! Thank you so much! That got it packaging perfectly. Now to figure out how to make sure there's a .desktop file for regular usage in a DE, and it's ready to go! 2022-02-14 14:05:52 Saijin_Naib: Either if there's one included upstream you amove it, otherwise look in other distros to see if they have one 2022-02-14 14:08:15 If it isn't upstream, how do I include the one from another distro? As a patch for APKBUILD to add to an empty file I touch during prepare? Is that an okay way to approach it? 2022-02-14 14:10:31 Just put the file in the same directory as the APKBUILD, add it to $source, and amove or "install" it 2022-02-14 14:10:51 Ahhh, okay. Thank you 2022-02-14 14:11:15 You probably should reference where you got the desktop file from with proper attribution too btw 2022-02-14 14:11:43 Yeah, absolutely. Fortunately there is one upstream, I must just have not found the right cmake... uh, target? for it 2022-02-14 14:11:58 Build process does not seem to be documented for this project so I'm just poking around 2022-02-14 14:12:00 What package is it? 2022-02-14 14:12:17 https://github.com/jmlich/geotagging 2022-02-14 14:12:36 It's been pretty friendly thus far, especially with your and ikke's help on the packaging, so that's nice 2022-02-14 14:12:45 Just this last wrinkle so it is nice for DE usage 2022-02-14 14:15:51 Saijin_Naib: Try using -DCMAKE_INSTALL_DATAROOTDIR=/usr/share in the build step 2022-02-14 14:16:04 Because https://github.com/jmlich/geotagging/blob/master/CMakeLists.txt#L94-L95 2022-02-14 14:17:26 Oooh, okay. I didn't set part of the prefix... 2022-02-14 14:18:07 Saijin_Naib: I'm not sure if that's actually correct, but I saw it being used in the freecad APKBUILD 2022-02-14 14:18:56 Ah, didn't know that was cmake build. Thank you! Trying now 2022-02-14 14:23:24 is there a way to manually remap scancode to keycode on USB keyboard without eudev? 2022-02-14 14:24:12 'setkeycodes 90004 30' doesn't work 2022-02-14 14:28:01 If I would like to switch (read: make available) the generic kernel for x86_64 within a virt-kernel system. What stey would I need to take for that? 2022-02-14 14:29:44 skinkie: what is "generic" kernel? 2022-02-14 14:30:06 minimal: any 'unusual' steps you took to get sway in virtualbox? 2022-02-14 14:30:15 minimal: how are you actually launching sway? 2022-02-14 14:30:18 there are 2 kernel packages - linux-lts and linux-virt (plus 2 addition ones if you are running Edge) 2022-02-14 14:30:35 minimal: the kernel shipped with 'standard' 2022-02-14 14:31:15 j`ey: shutdown the VM, go to the Virtualbox settings for the VM, go to "Display", tick the "Enable 3D Acceleration" box, start VM :-) 2022-02-14 14:32:04 minimal: hm, let me try 2022-02-14 14:33:22 j`ey: the only thing within virtualbox i had to do is to have the module in the initramfs 2022-02-14 14:33:41 skinkie: what does "apk info|grep linux-" show? I expect it to show linux-lts which is the usual kernel package 2022-02-14 14:33:49 minimal: ok thanks, that worked! 2022-02-14 14:34:02 never clicked that before in vbox 2022-02-14 14:34:11 minimal: "linux-virt" 2022-02-14 14:34:29 skinie: ok, so do you want linux-lts instead? 2022-02-14 14:34:40 minimal: I guess the other is linux-edge? 2022-02-14 14:34:58 minimal: not really 'instead' I would like to have my clients boot the linux-lts or linux-edge. 2022-02-14 14:35:08 linux-edge and linux-edge4virt are "bleeding edge" alternatives for linux-lts/linux-virt 2022-02-14 14:35:36 skinkie: they boot whichever kernel is installed 2022-02-14 14:36:21 minimal: when trying to apk add linux-lts (or linux-edge) I get an ocean of errors regarding "Failed to create lib/firmware/..." 2022-02-14 14:36:46 skinkie: you are using Alpine run-from-ram? 2022-02-14 14:36:56 minimal: no this is a sys 2022-02-14 14:37:56 skinkie: linux-lts does depend on firmware but by default it tends to install ALL the firmware packages, I typically do "apk add linux-lts linux-firmware-none " 2022-02-14 14:38:23 linux-firmware-none stops all the packages being installed and so you can pick the individual packages instead 2022-02-14 14:42:56 minimal: nifty trick, thanks 2022-02-14 14:43:15 minimal: Just wondering, does the order you put the packages in the apk add matter? 2022-02-14 14:43:22 no 2022-02-14 14:43:30 Ok thanks 2022-02-14 14:43:48 minimal: ERROR: Failed to create lib/firmware/intel/IntcSST2.bin: No such file or directory 2022-02-14 14:44:18 minimal: if I add linux-firmware-intel, does it fail to create /lib/firmware/intel? 2022-02-14 14:45:00 minimal: I now understand why you asked if it was running from ram, I get read-only filesystem. 2022-02-14 14:45:35 skinkie: strange, you said you were running in "Sys" mode which should be rw 2022-02-14 14:45:54 /dev/sda3 on / type ext4 (rw,relatime) 2022-02-14 14:46:00 according to that it is... 2022-02-14 14:46:58 I think it is related to 'modloop' 2022-02-14 14:47:15 skinkie: could the error be when the mkinitfs trigger kicks in? it will rebuild the /boot/initramfs-lts file - is /boot a separate mount and RO? 2022-02-14 14:47:30 dev/loop0 on /.modloop type squashfs (ro,relatime,errors=continue) 2022-02-14 14:47:46 so that is 'resolved' 2022-02-14 14:48:10 modloop? that doesn't sound like a Sys-mode install 2022-02-14 14:48:46 minimal: let put it this way; I did not know what I was doing. But by the way I am approaching Alpine Linux, I still don't ;) 2022-02-14 14:49:18 Sys-mode is a traditional install-and-run-from-disk like other distros, no squashfs/loop 2022-02-14 14:49:42 the default Alpine is run-from-ram using Squashfs 2022-02-14 14:49:52 minimal: yes, I understand that part. What I was trying to do is to reinvent lbu. 2022-02-14 14:50:09 there's a 3rd mode: "Data" mode where its run-from-ram but with data files stored separately on a disk 2022-02-14 14:50:26 LBU == run-from-ram mode, there's no LBU in Sys-mode 2022-02-14 14:51:23 minimal: but even now, I have a VirtualBox installation that works fine, now I want to make that part available on PXE. But since I used the linux-virt image, I obviously have missing modules for my clients. So I think the only option I have is to make a new Virtual Machine, that actually is 'lts' or 'edge' based, and use that as the default. 2022-02-14 14:52:22 minimal: I got lbu to work there ;) But obviously for the wrong reasons. Hence I was using the 'server' also to build the 'clients' I guess that part is something I should not do. 2022-02-14 14:52:30 skinkie: perhaps my script might interest you, though it doesn't handle netbooting (yet): https://github.com/dermotbradley/create-alpine-disk-image 2022-02-14 14:53:21 bookmarked. 2022-02-14 14:53:22 with "normal" run-from-ram I believe you use the update-kernel script to change a kernel and then write it back to the boot media, but I don't currently use run-from-ram so not sure 2022-02-14 14:53:31 minimal: seems like X11 breaks with 3d on heh 2022-02-14 14:53:53 j`ey: I have it running with Qt/EGL 2022-02-14 14:56:55 skinkie: so you are trying to build a PXE boot Alpine installation for physical machines? there is a netboot version of Alpine 2022-02-14 14:57:15 minimal: and have it offline available :) 2022-02-14 14:58:28 skinkie: for a specific use case? or for lots of "random" use cases? 2022-02-14 14:59:29 and somehow the mouse no longer works in wayland :< 2022-02-14 15:01:04 minimal: lets say, the use case is pretty generic. it is narrow casting, but all casts are in separate domains within a ordinary broadcaster. So I have a server, and several clients. And we are currently moving away from microSD cards for obvious reasons. We want to support some legacy devices (UEFI 64bit) and RaspberryPi4. So I am currently prototyping everything within VirtualBox, and just testing 2022-02-14 15:01:10 the first physical client. 2022-02-14 15:01:52 ok got the mouse back with WLR_NO_HARDWARE_CURSORS=1 2022-02-14 15:02:42 minimal: I don't mind to test out boot.alpinelinux.org, but the first change already would be that I would have to recompile since i can't use https://boot.alpinelinux.org/boot.ipxe 2022-02-14 15:02:58 skinkie: the approach I take it to build OS images that are basically read-to-go and use cloud-init on 1st boot to do any specific configuration 2022-02-14 15:03:19 minimal: and I wonder if ipxe.efi doesn't have an issue with this broken uefi network implementation on the board. But I can test that out quick. 2022-02-14 15:04:12 minimal: investigated cloud init as well, but then I understood how ovl could work in my advantage, which it currently does. So from configuration standpoint: i don't have issues. 2022-02-14 15:04:56 skinkie: generally you can chain ipxe from network-card PXE clients (seen this done with server that had its own iPXE limited build on the network card that we used to load/run full iPXE) 2022-02-14 15:07:30 minimal: my biggest frustration with the pxe implementation of that particular client is that the TFTP is very slow. 2022-02-14 15:08:08 skinkie: TFTP is slow by its very nature, that's why if you just load iPXE via tftp and then use iPXE with http/https to do the main work its faster 2022-02-14 15:09:11 Lets say I am going to check first if iPXE has a network connection, if that works, I'll ask you where to get the alpine-ipxe source code from, and how to change the bootscript ;) 2022-02-14 15:09:38 but now I first need to fix my installation, because the linux-lts broke it. 2022-02-14 15:10:42 its been a few years since I last looked at iPXE - we used it for provisioning servers at a very large bare metal private cloud provider 2022-02-14 15:15:19 minimal: I don't know if it is iPXE or a brokeng UEFI implementation. 2022-02-14 15:15:46 But I get "Invalid argument (http://ipxe.org/1c25e082) 2022-02-14 15:17:56 that's referring to drbg, which is RNG related 2022-02-14 15:18:24 what type of machine are you booting? Virtualbox? or a physical machine? 2022-02-14 15:19:43 basically iPXE wants entropy and doesn't have a source for it (e.g. HWRNG, SEED from UEFI, CPU RDRAND instruction, etc) 2022-02-14 15:22:30 minimal: physical machine, best to compare to an Atom based EEE 2022-02-14 15:22:49 minimal: could that be resolved, if no https was required? 2022-02-14 15:24:12 skinkie: if you try http then you *might* be ok, obviously before using encryption (i.e. HTTPS) the drbg should be seeded from entropy 2022-02-14 15:24:54 in the UEFI specs there is a way to provide entropy to a bootloader but not sure if its mandatory or optional functionality 2022-02-14 15:26:57 minimal: i'll it a shot when trying a manual ipxe command, currently evaluating if the lts has the required kernel modules for the network card. 2022-02-14 15:28:15 skinkie: relevant to your machines as Atom is mentioned: https://forum.ipxe.org/archive/index.php?thread-31540.html 2022-02-14 15:28:40 skinkie: Realtek cards? 2022-02-14 15:30:36 minimal: might be, it is a single board computer. But since it does not boot now, I cannot even answer that basic question. 2022-02-14 15:31:35 skinkie: my question about Realtek as that I found in the past for r8169 cards that I also had to add the relevant PHY module to the initramfs as it was not autodetected as a dependancy 2022-02-14 15:32:30 which is why I have this in my script: https://github.com/dermotbradley/create-alpine-disk-image/blob/main/lib/mkinitfs-functions#L1222 2022-02-14 15:35:41 minimal: now you are basically arriving at the destination why I thought I needed to have a different target to create - for example - an initramfs 2022-02-14 15:38:30 skinkie: my currently approach is to boot machines of a USB stick and then write ("dd") the prepared disk image to their HDD/SSD. Eventually I intend to instead/also PXE boot to a small Alpine run-from-ram image that uses HTTPS to pull the disk image down and does the "dd" that way 2022-02-14 15:39:41 minimal: shouldn't we report the missing realtek folder as a bug? 2022-02-14 15:41:22 skinkie: from memory its a upstream linux kernel issue that no direct module dep is defined. I think the Alpine mkinitfs relies on modinfo's "Depend" field to pull in related modules..so if there is no relevant "Depend" field for r8169 then of course it cannot work it out. 2022-02-14 15:42:51 I think the issue is that the realtek folder is not added at all. But I'll validate that later. 2022-02-14 15:44:24 I also find it very annoying that the initramfs always get chmod 400 2022-02-14 15:44:49 skinkie: a workaround might be to modify /etc/mkinitfs/features/network.modules to add the realtek folders but the r8169 module really should declare a dep on the PHY module, that's an upstream issue 2022-02-14 15:46:06 skinkie: one reason for "400" may be that if/when you are using cryptokey you do not want the initramfs group/world readable as it will contain a keyfile to unlock LUKS filesystems 2022-02-14 15:46:59 minimal: if you put that in your initramfs, on the same disk, I really fail to see to use disk encryption at all 2022-02-14 15:49:51 minimal: now my initramfs is appeartly too big, and causes the system to reboot. 2022-02-14 15:49:56 skinkie: I'm referring to FDE with Grub - Grub will prompt you for passphrase and then, unless you use keyfile, the initramfs mounting rootfs will also ask you for passphrase a 2nd time 2022-02-14 15:54:02 skinkie: re initramfs size, not sure - you're currently setup "confused" me as it seems to be partly Sys-mode and partly Run-from-Ram-mode. If your machines do not have a lot of RAM then this could perhaps be a side-effect 2022-02-14 15:54:38 minimal: I am now doing it from a ram based system 2022-02-14 15:54:57 minimal: but it seems that in the standard lts there is a lot less (I expect firmware) 2022-02-14 16:00:15 minimal: ok, so the following is in my perspective an issue, or bug if you want 2022-02-14 16:00:32 minimal: in the standard lts image there is *no* realtek kernel driver 2022-02-14 16:01:56 minimal: and the part that I don't understand is why the initramfs-lts from the standard image ends up in 15MB. While executing the same using mkinitfs becomes 50MB. 2022-02-14 16:02:14 skinkie: looking at package linux-lts I see: 2022-02-14 16:02:21 /lib/modules/5.15.22-0-lts/kernel/drivers/net/phy/realtek.ko.gz 2022-02-14 16:02:34 /lib/modules/5.15.22-0-lts/kernel/drivers/net/ethernet/realtek/r8169.ko.gz 2022-02-14 16:02:45 minimal: I am looking at 5.15.4-0-lts 2022-02-14 16:03:28 https://dl-cdn.alpinelinux.org/alpine/v3.15/releases/x86_64/alpine-standard-3.15.0-x86_64.iso 2022-02-14 16:15:29 minimal: I think I started to read a bit too late. The network card is an usb device. SMSC LAN9514 2022-02-14 16:27:52 ktprograms: I think you got me in the right direction with the FreeCAD APKBUILD, so thanks a ton for that! 2022-02-14 16:50:44 minimal: do you have a clue how to build ipxe.efi? 2022-02-14 16:52:23 skinkie: its build as part of the alpine-ipxe-ipxe_efi package 2022-02-14 16:52:41 Hi. I am trying to figure out the steps to do my custom kernel from readin https://wiki.alpinelinux.org/wiki/Custom_Kernel#Building 2022-02-14 16:53:05 I got as far as doing " git checkout -b my-custom-kernel origin/3.15-stable" 2022-02-14 16:53:12 minimal: I have checked out upstream git, but nothing builds that actual efi thing. 2022-02-14 16:53:37 skinkie: https://git.alpinelinux.org/aports/tree/main/alpine-ipxe?h=master 2022-02-14 16:54:22 but then what do I do? I can see the kernel configs in ~/aports/main/linux-lts 2022-02-14 16:55:21 I guess I need to get some kernel sources too to be able to do "make menuconfig" and then copy the resulting config from 2022-02-14 16:58:16 abuild -rK gets that, next section explains a bit 2022-02-14 16:58:53 "Attempt to build the kernel first. To do that, you do abuild -rK to install most of the dependencies. " 2022-02-14 16:59:06 indeed. the abuilt -rK could be highlited :D 2022-02-14 16:59:59 This seems to install the build stuff for the local user, right? 2022-02-14 17:02:06 ok so it created a "src" dir in "~/aports/main/linux-lts" 2022-02-14 17:02:47 yeah and you can do menuconfig in there 2022-02-14 17:03:26 minimal: ok, abuild checkum and then abuild -r 2022-02-14 17:03:35 just seems to be a bunch of symlinks to the config files and to the linux-5.15.tar.xz 2022-02-14 17:03:37 for that i get /usr/bin/abuild: cd: line 58: can't cd to /home/skinkie/aports/main/alpine-ipxe/src/ipxe-1.20.1/src: No such file or directory 2022-02-14 17:04:33 Forza: is there a build folder? 2022-02-14 17:04:39 no 2022-02-14 17:07:50 minimal: never mind... getting a direction... 2022-02-14 17:10:27 Alright starting over :) 2022-02-14 17:10:49 so 1) https://wiki.alpinelinux.org/wiki/Creating_an_Alpine_package#Setup_your_system_and_account doing the git clone ..aports now 2022-02-14 17:11:28 Forza: aports/main/linux-lts/src/build-lts.x86_64 2022-02-14 17:11:32 for example 2022-02-14 17:12:25 you mean I should do "git clone https://gitlab.alpinelinux.org/alpine/aports/main-linux-lts/src/build-lts.x86_64 2022-02-14 17:13:30 no 2022-02-14 17:13:41 I mean abuild -rK generates that 2022-02-14 17:15:11 Alright. So after I do "git clone https://gitlab.alpinelinux.org/alpine/aports" I do "cd aports". But then? 2022-02-14 17:15:30 cd main/linux-lts ; abuild -rK 2022-02-14 17:17:09 ah ok. 2022-02-14 17:17:47 that starts building, does the config etc. once it's started building properly, you can cancel t and run make menuconfig 2022-02-14 17:18:17 its installing 120 packages now 2022-02-14 17:18:44 that seems like a lot 2022-02-14 17:18:47 oh it seems to build the kernel too 2022-02-14 17:21:57 How do I determine what arch it is building for, or is that autodetected? 2022-02-14 17:22:14 your host 2022-02-14 17:22:45 Alright. I am building in a VM and the idea is to take the kernel package to a bare-metal machine 2022-02-14 17:23:27 the vm is on an EPYC server and the bare metal is a Xeon machine, but I guess the build system is not doing march=native ? 2022-02-14 17:23:33 just plain x86_64? 2022-02-14 17:24:16 yeah 2022-02-14 17:26:44 Taking a while :o 2022-02-14 17:27:23 Should've added some more virtual cpus 2022-02-14 17:27:43 the rest of them are just sitting there unused eh :p 2022-02-14 17:28:03 hi i'm on v3.15 main. edge repos are pinned. if i try to install the new ffirefox version via apk add firefox@ec and start it, i get the error message "Error relocating /usr/bin/firefox: _ZSt28__throw_bad_array_new_lengthv: symbol not found" - (icu-libs are updated @edge-main) 2022-02-14 17:28:54 you can't do that, so just upgrade your whole system to edge 2022-02-14 17:29:24 oh, why doesnt this approach work anymore? 2022-02-14 17:29:34 it was never supported to begin with 2022-02-14 17:29:47 psykose: yes. :) well. its a 24 core/48 thread epyc 2022-02-14 17:30:28 48 thread vm :p 2022-02-14 17:30:38 but it worked. so pinning seems to be useless in this case - now 2022-02-14 17:31:08 psykose: its set on 24 threads in the vm now. probably not going to be _that_ much faster to do the full 48 threads 2022-02-14 17:31:24 not by too much most likely 2022-02-14 17:32:08 Forza: you can kill the build, if you want to make your own config 2022-02-14 17:32:08 is there any meta package for ash/bash completion? 2022-02-14 17:32:17 ash doesn't have them 2022-02-14 17:32:22 bash is bash-completions or something 2022-02-14 17:32:42 ah, no -s 2022-02-14 17:32:43 j`ey: yea. I wanted to test "native" first before changing things 2022-02-14 17:33:06 i think you also needed to source the completion file, i don't remember if that was added to default 2022-02-14 17:35:00 /etc/profile.d/bash_completion.sh 2022-02-14 17:35:03 It's there 2022-02-14 17:35:55 neat 2022-02-14 17:36:15 d_oleovit it only worked since icu on edge was the same as on stable untill a few days ago, firefox has since been rebuild and updated with newer ones not found in 3.15 2022-02-14 17:36:30 that specific symbol is not from icu 2022-02-14 17:36:46 oh, i just saw him mentioning icu 2022-02-14 17:36:50 mb 2022-02-14 17:37:05 i know where it's from, but at that point you are backporting half of edge to your system 2022-02-14 17:37:08 which has no point at all 2022-02-14 17:37:13 just run edge instead 2022-02-14 17:37:25 true 2022-02-14 17:37:27 @Misthios nope, same problem with the previous version 2022-02-14 17:37:58 but why are u mixing repo's? 2022-02-14 17:38:10 if its only 1 application and u rlly need stable u could use flatpak 2022-02-14 17:38:24 i have too... just because of this browser..its frustrating 2022-02-14 17:38:27 you could also use firefox-esr 2022-02-14 17:38:31 which is actually updated in 3.15 2022-02-14 17:39:06 i tried the esr, but its outdadted by dsidn (no fission, et cetera) 2022-02-14 17:39:16 what? 2022-02-14 17:39:27 newer ff has site isolation aka fission 2022-02-14 17:40:03 isn't fission enables by default now in 67 2022-02-14 17:40:49 hmm it seems to be building several kernels 2022-02-14 17:41:10 whatever, i have to decide. everything on edge or not. thx anyway for everything 2022-02-14 17:41:17 ah, well sure, if you want that specific feature, then you would need to upgrade to edge 2022-02-14 17:41:18 it builds lts and virt Forza , I believe 2022-02-14 17:41:20 minimal: alpine-ipxe fails on testing with '-Werror=maybe-uninitialized' 2022-02-14 17:43:46 I rebooted with more cores.. but this is odd. When I run "abuild -rK" again it seems to re-download/reinstall lots of dependencies again. There is also a warning about /home/forza/packages but that dir does exist: https://dpaste.com/DEBXLQMBQ 2022-02-14 17:44:46 should I be running it as root? 2022-02-14 17:45:08 no 2022-02-14 17:45:11 according to abuild -h "-r Install missing dependencies from system repository (using sudo)". I don't have sudo installed 2022-02-14 17:46:45 the warning doesn't mean anything 2022-02-14 17:46:53 and yes, it builds from scratch again 2022-02-14 17:47:10 when you stopped the build it uninstalled the dependencies 2022-02-14 17:47:17 ahh 2022-02-14 17:47:43 ill let it run. going to have a sauna meanwhile :) 2022-02-14 17:47:55 :) 2022-02-14 17:48:00 it will take about an hour 2022-02-14 17:48:08 >>> ERROR: linux-lts: prepare failed 2022-02-14 17:48:30 you probably to rm src 2022-02-14 17:48:58 :) 2022-02-14 17:49:39 Forza: maybe it will help you with the process 2022-02-14 17:50:12 Forza: few commands which using when wanna make menuconfig and build kernel without -virt so it will take doble less time... 2022-02-14 17:50:15 Forza: https://tpaste.us/mqWV 2022-02-14 17:50:17 MY-R: the sauna? yes. I hurt my back this weekend and I cant barely move 2022-02-14 17:50:34 sauna too :) 2022-02-14 17:51:23 MY-R: thanks a lot! 2022-02-14 17:51:30 Ill see you later :) 2022-02-14 17:51:32 can also edit APKBUILD since it is from 3.15 you can change to last kernel for example 5.15.23 2022-02-14 17:51:54 ok! 2022-02-14 17:52:39 skinkie: yes seeing that error when I build it myself locally. There's a newer version of the upstream, let me try that instead 2022-02-14 17:53:02 minimal: multiple errors :) 2022-02-14 17:53:21 minimal: is it possible to continue building, instead of cleaning and starting over from scratch using abuild? 2022-02-14 17:55:42 skinkie: dunno, I'm using abuild inside a docker container so I just kick it off again 2022-02-14 18:00:34 i am increasingly coming to the conclusion that alpine or build-base should come with linux-firmware-none 2022-02-14 18:01:29 build-base for sure 2022-02-14 18:03:07 i think there's an argument for putting it in alpine default docker image 2022-02-14 18:03:55 if you really want to convert docker.io/alpine to bare metal/vm instead of using minirootfs, then you can figure out firmware yourself 2022-02-14 20:35:02 minimal: ipxe is still *slow* (3x) at least the configuring net0 part, download is very fast. It boots moderately fast. 2022-02-14 20:41:41 skinkie: been a while since I looked at ipxe so not sure if there's much in the way of optimisations you can make. 2022-02-14 20:43:35 I might find a free minute to do a MR for alpine-ipxe to update to the current version and add the CFLAG change to get it to build 2022-02-14 20:50:45 >>> linux-lts: Build complete at Mon, 14 Feb 2022 21:49:52 +0100 elapsed time 2h 55m 16s almost 3 hours. wow. Thats a long time. 2022-02-14 20:56:27 for a 24 core machine thats surprising 2022-02-14 20:56:46 It did seem to compile 2 different kernels 2022-02-14 20:57:02 linux-lts-dev and linux-lts ? 2022-02-14 20:59:24 even on my 8-thread laptop it took less than that 2022-02-14 20:59:36 power of xen .. :p 2022-02-14 20:59:39 well 2022-02-14 20:59:45 i have never had a fast vm experience 2022-02-14 20:59:47 on anything, ever 2022-02-14 20:59:51 i avoid them for mostly that reason :p 2022-02-14 20:59:59 i am sure somebody has a fast way of doing it 2022-02-14 21:00:13 even using kvm? 2022-02-14 21:00:58 yea 2022-02-14 21:01:01 at least for desktop usage 2022-02-14 21:01:08 i haven't benchmarked any timed stuff 2022-02-14 21:03:31 could likely be disk io limiting 2022-02-14 21:06:43 is eatmydata availabe for alpine? 2022-02-14 21:07:46 It helps disk io by blocking lots of fsyncs 2022-02-14 21:08:20 obviously it isn't crash safe,... so don't use on important data 2022-02-14 21:08:27 don't see eatmydata in alpine:edge 2022-02-14 21:08:35 it is 2022-02-14 21:08:42 https://pkgs.alpinelinux.org/packages?name=*eatmydata*&branch=edge 2022-02-14 21:08:43 libeatmydata 2022-02-14 21:08:50 you can LD_PRELOAD it 2022-02-14 21:08:55 though careful 2022-02-14 21:09:00 i have unironically lost some files with it 2022-02-14 21:09:06 :( 2022-02-14 21:09:07 like my entire /etc/apk/world 2022-02-14 21:09:50 because i was using it to build things while trying some cpu overclocking, and the board power cycling in the middle of non-fsynced data does funny shit 2022-02-14 21:09:50 :p 2022-02-14 21:11:46 yup thats the risk 2022-02-14 21:13:44 lets see if it is any faster with LD_PRELOAD 2022-02-14 21:14:52 doesnt seem so 2022-02-14 21:15:23 CPU usage is about 2000% so about 20 cores utilized 2022-02-14 21:17:13 I suppose I can try directly on the target machine. but it only has a 6 core xeon 2022-02-14 21:18:02 Intel(R) Xeon(R) E-2126G CPU @ 3.30GHz 2022-02-14 21:18:36 but there is an nvme disk in it that could be used 2022-02-14 21:54:49 install gawk, that'll make it at least half an hour faster 2022-02-14 22:01:04 Hello71: Was it you that recommended that and mawk over built-in busybox awk? Does that help for any/all compilations that look for awk support? 2022-02-14 22:01:28 theoretically yes but the amount of improvement depends on the script 2022-02-14 22:02:02 objdump_reformat.awk processes about a gigabyte of text in a way that is particularly inefficient on busybox awk 2022-02-14 22:02:44 but if you're processing a dozen lines with {print $2} then it doesn't matter 2022-02-14 22:05:04 Ah, got it, thanks 2022-02-14 22:12:20 like, if the total amount of time spent in busybox awk is 1 second, and mawk is 10x faster, then you save 0.9 seconds on the build 2022-02-14 22:12:45 it would be more interesting if mawk was the default awk in, say, build-base 2022-02-14 22:13:37 maybe we save 0.9 seconds per package over 1000 packages, that's pretty good 2022-02-14 22:14:03 i guess it would make more sense to put it in alpine-sdk 2022-02-14 22:14:14 doesn't it not have a awk-mawk symlink 2022-02-14 22:14:17 you would need that too 2022-02-14 22:14:44 like gawk does 2022-02-14 22:14:56 i wonder why one doesn't exist for it 2022-02-14 22:14:59 is it incompatible in some way 2022-02-14 22:15:36 or perhaps just overlooked :) 2022-02-14 22:19:09 Yeah, I don't build anything big enough that it matters for me, but that's a really great point about its impact on the builders 2022-02-14 22:20:50 -G Ninja is mostly the same 2022-02-14 22:21:00 although an average of 20% is quite noticeable for anything non-trivial 2022-02-14 22:42:41 What types of builds can I wrap with the -G Ninja? Anything cmake? 2022-02-14 22:42:45 What about qmake? 2022-02-14 22:49:13 -G Ninja is a cmake argument 2022-02-14 22:50:03 qmake has no ability to specify it to my knowledge, but a lot of qmake projects also usually allow cmake to my knowledge 2022-02-14 22:54:55 Ahh, okay. I'm just asking in preparation of the next APKBUILD I'm looking to tackle. This one seems to be setup weird, but it also has documentation, so I'm hopeful. It looks like it uses qmake to make the Makefile, and then switches to regular make to build 2022-02-14 22:55:00 https://github.com/pawelsalawa/sqlitestudio/wiki/Instructions_for_compilation_under_Linux 2022-02-14 22:56:08 yeah this is qmake only 2022-02-14 22:56:51 Ahh, okay. I'll try to keep that and the gawk/mawk thing in mind for future builds though 2022-02-14 22:59:19 you should never need to add it yourself 2022-02-14 23:03:58 No? What's the expected behavior? 2022-02-14 23:05:13 it just uses busybox awk 2022-02-15 00:03:41 gawk and mawk can't both install /usr/bin/awk, it would conflict 2022-02-15 00:03:52 it works ("works") with busybox using trigger hackery 2022-02-15 00:04:19 but i think everyone agrees that the busybox trigger kinda sucks and doesn't want more of it 2022-02-15 00:30:57 8-) 2022-02-15 00:44:36 anyone want to talk shell scripts? 2022-02-15 00:45:11 Hey, sorry I don't idle here so I don't know if there's been discussion, but is there a public place we can see status updates on the alpinelinux.org outage? 2022-02-15 00:46:03 it is indeed down 2022-02-15 00:46:28 lmao 2022-02-15 00:46:35 even the zabbix instance is down 2022-02-15 00:46:38 or so it seems 2022-02-15 00:46:40 unlucky :p 2022-02-15 00:47:13 A nice reminder to make an internal mirror. 2022-02-15 00:47:27 of the isos? i suppose 2022-02-15 00:48:08 of the packages 2022-02-15 00:48:18 oh 2022-02-15 00:48:24 dl-cdn is down too 2022-02-15 00:48:45 Seemed like that was the first to go down, our pipelines started failing. 2022-02-15 00:50:20 if you can use a mirror i think those should be up 2022-02-15 00:50:29 but yeah 2022-02-15 00:50:37 sadly i don't have access to the *.alpinelinux infra 2022-02-15 00:50:41 everything is down currently 2022-02-15 00:50:45 and ikke is asleep 2022-02-15 00:50:55 :p 2022-02-15 00:50:59 check back in 6-7 hours 2022-02-15 00:52:02 cool 2022-02-15 00:52:19 only a single person with infra access? 2022-02-15 00:52:35 i know of two, maybe another 1-2 that i don't 2022-02-15 00:52:43 but we're all in europe anyway 2022-02-15 00:52:53 ikke manages 90% of it 2022-02-15 01:04:39 Just wanted to mention that the site appears to be down (https://wiki.alpinelinux.org/) 2022-02-15 01:05:20 it is 2022-02-15 01:28:58 it went down just as i was considering installing Alpine for the first time 2022-02-15 01:46:47 Looks like the website and CDN are down? alpinelinux.org and dl-cdn.alpinelinux.org 2022-02-15 01:47:01 yep 2022-02-15 01:54:00 That's odd, *every* DNS record is pointing at 147.75.101.119 2022-02-15 01:54:32 at least, both uk.alpinelinux.org and nld3-dev1.alpinelinux.org as far as I can see so far 2022-02-15 01:55:03 the ones that are CNAMEs to out-of-zone services (e.g. Fastly for dl-cdn) are intact 2022-02-15 01:55:47 (I would have thought uk.alpinelinux points to something in the UK, rather than what appears to be in the Netherlands) 2022-02-15 01:57:15 I'm probably jumping to conclusions here 2022-02-15 02:09:14 Is there a status page with information on the Alpine Linux outage? 2022-02-15 02:11:59 seems like nld3-dev1.alpinelinux.org is down. 2022-02-15 02:12:03 Or is there a different channel discussing it? 2022-02-15 02:12:19 the infra team are in NLD and sleeping right now (presumably) 2022-02-15 02:12:26 All the *.alpinelinux.org sites seem down for the most part. 2022-02-15 02:12:30 channel #alpine-infra 2022-02-15 02:12:49 ncopa, Thanks 2022-02-15 02:14:43 I am trying to figure out what I can from my end 2022-02-15 02:15:49 just got here, but noticed dl-cdn.alpinelinux.org is down as of an hour ago or so 2022-02-15 02:17:23 atka, Yes and most of *.alpinelinux.org 2022-02-15 02:17:55 ncopa, My plan is to ignore and hope when I wake up tomorrow it is fixed ;) 2022-02-15 02:18:24 yeah, I just changed my /etc/apk/repositories for the time being 2022-02-15 03:05:50 I think it should be all back now 2022-02-15 03:10:04 ncopa: Seems to be. Thanks! 2022-02-15 05:39:10 hi friends, is there a similar set of instructions to [1] for installing alpine using a chroot? [1] https://docs.alpinelinux.org/user-handbook/0.1a/Installing/manual.html 2022-02-15 05:41:22 maria: something like this? https://wiki.alpinelinux.org/wiki/Alpine_Linux_in_a_chroot 2022-02-15 05:43:52 thank you! i'm assuming i can run commands like setup-keymap after booting? 2022-02-15 05:44:33 maria: make sure alpine-conf is installed 2022-02-15 05:47:59 note that if you want to boot, you also need to make sure you have a bootloader and kernel installed 2022-02-15 05:49:37 that would be grub, grub-efi, efibootmgr, and linux-lts right? 2022-02-15 05:50:04 yeah, something like that 2022-02-15 05:50:42 maria: Why not just do a normal install then? 2022-02-15 05:51:13 i like to watch youtube while i do mundane things 2022-02-15 05:52:29 maria: A standard (ext4, uefi) install can be done in less than 5 minutes 2022-02-15 06:06:13 why discourage a chroot install with customizations though? I never use the installer for alpine or most other distros. you can set it up how you want and maybe learn something in the process 2022-02-15 06:07:32 atka: I'm not discouraging the customizations, but why use a chroot? 2022-02-15 06:09:32 well I guess it doesn't need to be a chroot, but until recently there was no luks option and I don't use ext4 so using the installer doesn't work for that 2022-02-15 06:09:50 and why trust someone else's scripts 2022-02-15 06:09:55 sometimes necessary when installing alpine on a VPS that has no built-in option 2022-02-15 06:10:12 for instance alpine's setup-bootable is broken somewhat 2022-02-15 06:10:12 atka: ikke: I see, thanks for clarifying 2022-02-15 06:10:30 it populates fstab with /dev/sdaX instead of uuid 2022-02-15 06:11:03 so if you want to use a rootless usb on another system with different disk layout it won't mount properly at boot 2022-02-15 06:12:22 sorry not rootless, diskless 2022-02-15 06:14:06 I plan on submitting a patch to fix that behavior in the next few days though 2022-02-15 06:54:45 Hello71: (re qutebrowser segfaults) I tried booting with the linux-virt kernel (instead of my custom kernel), and qutebrowser works perfectly fine. 2022-02-15 06:55:08 I then tried compiling my custom kernel with 4k pages instead of 16k, and qutebrowser works again. 2022-02-15 06:55:21 So seems like qutebrowser doesn't like running on a 16k pages kernel 2022-02-15 06:56:52 ktprograms: what system are you on? 2022-02-15 06:57:16 j`ey: Alpine aarch64 VM, macos host 2022-02-15 06:59:59 ktprograms: cool 2022-02-15 07:00:39 I was originally testing out 16K pages because the M1 hardware supports it and it's supposed to be faster, but it seems both ripgrep and qutebrowser don't work with it 2022-02-15 07:01:26 Also now that I think about it, I don't even know if the qemu VM uses hardware backed 16K pages 2022-02-15 07:01:34 I think aarch64 only supports 4k and 64k 2022-02-15 07:01:46 also you'll eat more memory I think 2022-02-15 07:01:53 https://github.com/evansm7/vftool/discussions/9 2022-02-15 07:02:00 atka: it also supports 16k 2022-02-15 07:02:31 does the kernel support it though? 2022-02-15 07:02:42 atka: In what way? 2022-02-15 07:03:04 https://www.kernel.org/doc/html/latest/arm64/memory.html 2022-02-15 07:03:09 I see no mention of 16k 2022-02-15 07:03:30 it does 2022-02-15 07:05:08 is qute webkit? 2022-02-15 07:05:20 https://github.com/torvalds/linux/blob/master/arch/arm64/Kconfig#L1046 2022-02-15 07:05:54 atka: qtwebengine 2022-02-15 07:07:16 ok 2022-02-15 07:08:16 Do you think I should ask about 16K page kernels support on the qutebrowser IRC channel, or don't bother since other things aren't working with 16K kernels for me so I'm probably not gonna continue using it? 2022-02-15 07:09:07 if you want to scatch that itch, why not? 2022-02-15 07:09:10 stay curious! 2022-02-15 07:11:10 atka: feeling a bit lazy to do it now. I'll do it when I'm next bored 2022-02-15 07:11:29 :) 2022-02-15 07:13:40 Ill look into ripgrep at some point, I need that 2022-02-15 07:20:29 j`ey: What do you mean by "look into"? 2022-02-15 07:26:18 ktprograms: figure out why it doesnt work 2022-02-15 07:26:33 with 16K pages 2022-02-15 07:26:54 j`ey: iirc it was a problem with jemalloc not supporting 16K pages 2022-02-15 07:28:22 that was my guess 2022-02-15 07:29:05 jemalloc does support 16k, but has to be built specifically for it 2022-02-15 07:31:55 j`ey: I see, interesting. 2022-02-15 09:29:42 hi 2022-02-15 11:00:46 what UEFI bootloaders are available on alpine? 2022-02-15 11:01:55 grub, refind, uboot 2022-02-15 11:11:13 hmm I thought uboot provided the EFI environment itself, not the loader 2022-02-15 11:12:13 going to install either grub or systemd-boot, I guess 2022-02-15 11:12:45 well it can also launch EFI binaries 2022-02-15 11:14:13 oh, no CONFIG_EFISTUB 2022-02-15 11:14:17 I guess grub it is 2022-02-15 11:15:03 wait, d'oh, is efistub even a thing for 32-bit kernels 2022-02-15 11:15:57 this is getting out of hand quite fast, maybe I should just have two separate USB sticks for ancient and modern systems 2022-02-15 11:21:40 I assume EFI stub works on 32bit kernels 2022-02-15 11:23:14 I guess it does but it'll be a 32-bit stub as well, so it won't boot on 64-bit EFI anyway 2022-02-15 11:23:18 Hi, anyone running sway has noticed a font size change (titles, bemenu/ qutebrowser..) ? 2022-02-15 11:56:14 e.g. https://i.postimg.cc/ydsvNrXx/screenshot-2022-02-15-125117.png 2022-02-15 11:57:30 similar behaviour on firefox too 2022-02-15 11:58:35 that looks hilarious 2022-02-15 11:58:44 are you on edge? 2022-02-15 11:58:46 yes 2022-02-15 11:59:08 that hasn't happened to me with either firefox or sway, so probably some weird font stuff 2022-02-15 11:59:09 I think that it started yesterday after doing upgrades 2022-02-15 11:59:21 what's the output of "fc-match"? 2022-02-15 11:59:37 Font Awesome 6 Brands-Regular-400.otf: "Font Awesome 6 Brands" "Regular" 2022-02-15 12:02:37 hmm 2022-02-15 12:02:47 ptrc: that command should output many fonts? 2022-02-15 12:05:28 uhm 2022-02-15 12:05:30 interesting 2022-02-15 12:05:58 without any parameters it should output your "default" font 2022-02-15 12:06:03 which is dejavu sans for me 2022-02-15 12:06:09 i don't think it's supposed to be fontawesome 2022-02-15 12:06:14 all my fonts dir semes modified yesterday 2022-02-15 12:08:52 https://i.postimg.cc/C1CKyVCf/screenshot-2022-02-15-130618.png 2022-02-15 12:09:51 I feel that it's the moment when it started to fail 2022-02-15 12:11:06 hm, fa6 doesn't seem to be in the repos, isn't that somewhere in your local directory maybe? 2022-02-15 12:11:38 are this files directed extracted from packages or generated by the system? 2022-02-15 12:12:03 you mean .ttf files, or..? 2022-02-15 12:12:10 yes 2022-02-15 12:12:16 from packages 2022-02-15 12:12:29 so maybe I just get some package removed while ugprading? 2022-02-15 12:12:43 I don't think that I have any non-packaged font 2022-02-15 12:13:09 can you check "fc-match -b" for the path? 2022-02-15 12:13:40 file: "/usr/share/fonts/X11/OTF/Font Awesome 6 Brands-Regular-400.otf"(w) 2022-02-15 12:14:14 are this fonts X11 only? 2022-02-15 12:15:15 there was a commit 2022-02-15 12:15:23 b348ca9cdf59637fc712072bb1496c014b438b27 2022-02-15 12:15:45 maybe this broke my this 2022-02-15 12:15:49 oh, my font cache must have been stale 2022-02-15 12:16:16 I'm gonna try on alpine-devel 2022-02-15 12:17:47 hm 2022-02-15 12:17:51 do you maybe have something in "~/.config/fontconfig"? 2022-02-15 12:17:53 I could just try to change that font 2022-02-15 12:18:06 doesn't exist 2022-02-15 12:18:13 I don't know where I defined that awesome font 2022-02-15 12:18:29 on sway config I have font pango:SourceCodePro Medium 11 2022-02-15 12:18:55 maybe /etc/fonts? 2022-02-15 12:19:52 it seems unmodified 2022-02-15 12:22:04 the upgrpade seems a major ugprade, from 5.15.4 to 6.0.0 2022-02-15 12:22:49 pkgdesc="Iconic font designed for Bootstrap TTF file only" 2022-02-15 12:22:59 what it means with 'Bootstrap TTF file only'? 2022-02-15 12:24:08 "Bootstrap" is a popular website CSS framework, and "TTF file only" probably means "no OTF, no WOFF, no WOFF2" 2022-02-15 12:25:27 ah, I thought it could mean that it's used for generate something and it could be failed on the upgrade process 2022-02-15 12:25:50 donoban: do you remember what font were you using previously? maybe it's unrelated and your main font just got uninstalled somehow? 2022-02-15 12:26:22 i looked at the ttf-font-awesome packages and they don't seem that different, only .ttf files and no configs 2022-02-15 12:26:30 I'm not sure, the first day I started sway I added that line on config 2022-02-15 12:27:16 font pango:SourceCodePro Medium 11 2022-02-15 12:28:43 does 'fc-match SourceCodePro' output that font's filename? 2022-02-15 12:29:00 lol yes 2022-02-15 12:29:03 Font Awesome 6 Brands-Regular-400.otf: "Font Awesome 6 Brands" "Regular" 2022-02-15 12:29:13 wait how 2022-02-15 12:29:20 I commeted the line and reload sway 2022-02-15 12:29:40 but I feel that a full reboot (or logout/relogin) could be need 2022-02-15 12:29:58 uhM 2022-02-15 12:30:08 if I do: fc-match fooobar 2022-02-15 12:30:19 it also shows 'Font Awesome...' 2022-02-15 12:30:59 so it seems my default fallback font 2022-02-15 12:31:38 on alpine the default font specified in /etc/fonts would be dejavu sans, do you have that installed? 2022-02-15 12:31:56 if not, it probably fallbacks to... first font alphabetically? no idea 2022-02-15 12:32:19 uhM, do you mean on /etc/fonts/fonts.conf ? 2022-02-15 12:32:29 oh, nevermind, dejavu sets _itself_ as the default font 2022-02-15 12:32:36 I have 2022-02-15 12:32:42 ./ttf-dejavu 2022-02-15 12:32:51 on /usr/share/fonts 2022-02-15 12:32:53 i meant /etc/fonts/conf.d/45-dejavu.conf, but now i see that's from ttf-dejavu 2022-02-15 12:33:22 I don't have it intsalled 2022-02-15 12:33:43 [donoban@localhost][/usr/share/fonts]% fc-match Liberation 2022-02-15 12:33:45 DejaVuSans.ttf: "DejaVu Sans" "Book" 2022-02-15 12:33:49 now yes 2022-02-15 12:34:02 ohh 2022-02-15 12:34:06 bemenu returned to normality 2022-02-15 12:34:21 ohh 2022-02-15 12:34:27 that worked! 2022-02-15 12:34:29 thanks ptrc 2022-02-15 12:34:53 it seems that someway that font was removed 2022-02-15 12:35:21 you're welcome :) 2022-02-15 12:36:00 it doensn't seem required by any package that I have installed 2022-02-15 12:43:32 I feel that something works wrong with fc-match 2022-02-15 12:43:45 theorically I have ttf-liberation installed 2022-02-15 12:44:00 but fc-match always returns Dejavu now 2022-02-15 12:50:48 grawity: the esp path is different 2022-02-15 12:51:58 ? 2022-02-15 12:53:22 efi/boot/bootx64.efi, efi/boot/bootia32.efi 2022-02-15 12:53:40 yeah but most systems that I'd be booting UEFI on would be 64-bit UEFI 2022-02-15 12:53:47 and apparently that can't boot 32-bit .efi images 2022-02-15 12:53:55 (at least qemu's tianocore cannot, it seems) 2022-02-15 12:54:01 unless I'm holding something wrong 2022-02-15 12:55:08 this USB stick is being stretched between supporting mid-2000 PCs with 128 MB of memory and modern 2022 PCs with x64 2022-02-15 12:55:36 but got it working with grub in the end, since it *is* still a full linux loader by itself 2022-02-15 12:55:48 mmhmm 2022-02-15 12:56:26 oh I wonder if I could put systemd's 64-bit boot stub on a 32-bit kernel image 2022-02-15 13:07:28 i don't know what you mean by "systemd's 64-bit boot stub". as the arch wiki says, linux 5.7+ supports booting 64-bit kernel with 32-bit efi: https://wiki.archlinux.org/title/Unified_Extensible_Firmware_Interface#Booting_64-bit_kernel_on_32-bit_UEFI 2022-02-15 13:07:54 but as mjg59 points out, this will inevitably lead to every firmware bug, and you're better off using bios boot: https://mjg59.dreamwidth.org/26734.html 2022-02-15 13:07:58 yeah, but I have a 32-bit kernel and 64-bit UEFI 2022-02-15 13:08:09 (on the machines that have UEFI, that is) 2022-02-16 01:11:49 apks are just gzipped archives yes? if I was to extract one and edit it what would I need to do to recompress it? 2022-02-16 01:13:11 atka: It's a tar.gz I think, so "tar -zcf" probably. 2022-02-16 01:13:22 Although the signature will probably be wrong etc 2022-02-16 01:13:41 yeah, that was my next question 2022-02-16 01:13:59 I'm editing the lbu scripts in /sbin/lbu 2022-02-16 01:14:27 so the alpine-conf-3.13.0-r0.apk 2022-02-16 01:16:23 atka: The alternative way would be to build your own alpine-conf package with your changes, but then you'd either have to add patches or completely change the source in the APKBUILD 2022-02-16 01:18:48 ok, I'll ask in devel what they recommend 2022-02-16 01:19:31 there will most likely be some trial and error testing my changes so a quick way to test would be ideal 2022-02-16 01:20:48 atka: For the testing why not just modify /sbin/lbu (keep a backup of the original)? 2022-02-16 01:20:58 s/keep/while keeping/ 2022-02-16 01:21:12 s/keep/while keeping/ 2022-02-16 01:21:20 Where's the bot? 2022-02-16 01:22:41 because I'm making changes on a "diskless" install and those changes are lost when booting and when testing changes to the openssl encryption that takes place too early in the boot from what I see 2022-02-16 01:23:12 atka: Ah I see, not sure what you can do then 2022-02-16 01:24:43 I think I know what needs to be done, I need my patched /sbin/lbu to replace the one in alpine-conf.apk 2022-02-16 01:25:16 shouldn't be too much of an issue unless there are signature problems 2022-02-16 01:28:50 atka: Maybe you could look into making an apkovl with your modified lbu script? (Not sure if that's possible, I don't know much about the diskless area of alpine) 2022-02-16 02:19:56 Hello there. What's the point of the -l switch to `apk update`? The help says "Always choose the latest package by version". So I guess my question is, when would apk _not_ choose the latest version? I thought that's what `apk update` did... 2022-02-16 08:01:07 Hi. Does Alpine support nic Teaming instead of the normal "bonding" mode? 2022-02-16 08:56:31 What would be the difference? You have several bonding modes 2022-02-16 08:56:58 Hi all, is it expected that expat 2.4.4 ( with the CVECVE-2022-23852 fix) will still be deployed in Alpine 3.12 even though eol is in two months? 2022-02-16 08:57:23 Afaik teaming and bonding are different names for the same concept 2022-02-16 09:27:07 yeah but for some reason they're two separate implementations in the kernel 2022-02-16 09:27:54 the former uses a userspace daemon for some things (teamd, teamctl) and there's JSON everywhere 2022-02-16 10:39:31 atka: thats why you sync to the "diskless" with lbu 2022-02-16 10:39:40 so whatever changes you made will remain after reboot 2022-02-16 10:46:44 <__ryan__> Hello folks! 2022-02-16 10:47:29 def __ryan__(self): self.hello() 2022-02-16 10:48:03 <__ryan__> Haha, that's a sweet one-liner. =) 2022-02-16 10:50:47 <__ryan__> I'm trying to get gitlab-runner to run on edge, but, I get this error in the logs: respawned "/usr/bin/gitlab-runner" too many times, exiting. I get other error. When I copy the gitlab-runner command that supervise-daemon runs and try it manually in a terminal everything works. Ideas? :) 2022-02-16 10:50:55 <__ryan__> I get no other error* 2022-02-16 10:55:09 <__ryan__> Ah, I see now, /etc/gitlab-runner is owned by root and is set to 700, while the service runs as gitlab-runner. 2022-02-16 11:18:16 ikke: yea the teaming module is different from the bonding module and apparently is better with ipv6 and other things like less overhead. But I don't know. Just heard it was preferable to bonding these days 2022-02-16 11:18:41 Though I would suspect lacp is still the same in both? 2022-02-16 11:24:02 libteam (which contains teamd) is packaged 2022-02-16 11:26:33 Interesting, never heard about it before on linux 2022-02-16 11:32:46 Yea I was looking at it when building this new server :) 2022-02-16 11:33:22 I expect that there won't be much difference in performance on a couple of 1gbit links 2022-02-16 11:37:26 We have bonds with 2x10Gbit, not sure what the overhead is 2022-02-16 11:38:54 Might be interesting to test. My worry is the config and setup. Normal bonds are straight forward and easy 2022-02-16 11:40:27 Btw the build host in my vm. I got the compilation of the kernel (abuild -r) down from 2.5 to 0.5 hours using ld_preload libeatmydata.so 2022-02-16 11:41:32 interesting 2022-02-16 11:41:55 Forza: But doesn't that not write the finished kernel to disk? 2022-02-16 11:42:19 ktprograms: it does, eventually :) 2022-02-16 11:42:20 It just creates and kernel.apk file 2022-02-16 11:42:45 you just loose data if your machine crashes before it is actually written 2022-02-16 11:42:54 Oh I see, it just removes safety from writing, but it _does_ write 2022-02-16 11:43:20 Io latency in VMs is usually much worse than bare metal, so this is likely why it helped so much 2022-02-16 11:43:37 Yes it makes fsyncs into null operations 2022-02-16 11:44:42 I'm thinking it's worth the tradeoff during compiling 2022-02-16 11:45:43 Forza: did you think about tempfs? 2022-02-16 11:45:53 No :o 2022-02-16 11:46:02 How much space would I need for that? 2022-02-16 11:46:17 It uses memory 2022-02-16 11:46:36 So it would need to fit the entire tree + build artifacts 2022-02-16 11:47:12 Yes I know. I meant the build tree. I could assign up to 64GB ram 2022-02-16 11:47:30 note that tmpfs is even _less_ save :) 2022-02-16 11:47:36 Forza: What kind of CPU do you have? Why does it take 2.5 (or 0.5) hours? 2022-02-16 11:47:59 Forza: did you also install gawk like Hello71 suggested? 2022-02-16 11:48:17 No I didn't do gawk. Must have missed that one 2022-02-16 11:48:36 It's an epyc server running xcp-ng.org hypervisor 2022-02-16 11:48:40 for x86*, it can apparently make a lot of difference 2022-02-16 11:50:12 Forza: Hmm, interesting that it takes that long (although the half hour seems about as much as was previously mentioned when using busybox awk) 2022-02-16 11:55:19 how do I force abuild to build again? it says>>> linux-lts: Package is up to date 2022-02-16 11:55:57 -f 2022-02-16 11:56:15 OK trying again with gawk 2022-02-16 11:57:28 runnign with 24 threads. It did not make it faster with 48 threads. Probably xen scheduling isn't so good to take advantage of the extra smt threads 2022-02-16 11:57:51 How many NUMA domains? 2022-02-16 11:58:21 8 I believe 2022-02-16 11:58:33 That's quite a lot 2022-02-16 11:58:34 but they are not exposed to guests, 2022-02-16 11:58:52 Each CCX is a numa domain on the EPYC 2022-02-16 11:58:58 I can try to turn that off 2022-02-16 11:59:18 or is it called CCD. You know each chiplet 2022-02-16 12:00:19 https://paste.tnonline.net/files/OwyN00p8zHhD_Screenshot2022-02-16125937.png I/O rate is about 25% 2022-02-16 12:00:26 Maybe tmpfs would be helpful 2022-02-16 12:00:51 Do I simply make a tmpfs in ~/aports and copy everything there and try again 2022-02-16 12:01:57 Good thing I have two identical machines to try on 2022-02-16 12:05:27 This stuff "Last-Level Cache as NUMA Node" was enabled. I'll disable and test again 2022-02-16 12:07:42 ikke: why did you think it was a lot? 2022-02-16 12:08:36 I have no experience with NUMA so I don't quite know the implications. I thought the idea that cache centric domains would be good since there is 128MiB L3 cache on this CPU. 2022-02-16 12:09:08 btw why does abuild make two kernels? 2022-02-16 12:09:21 linux-lts and linux-lts-dev? I only need the one :) 2022-02-16 12:09:30 lts and virt 2022-02-16 12:09:47 I removed the virt config 2022-02-16 12:11:04 ok it took 15 minutes with gawk 2022-02-16 12:11:39 https://paste.tnonline.net/files/bpvq4sSlzATC_Screenshot2022-02-16131126.png 2022-02-16 12:11:52 See linux-lts and linux-lts-dev? 2022-02-16 12:12:15 or are they the same thing 2022-02-16 12:12:19 dev is the headers and 3rd party module stuff 2022-02-16 12:12:26 Aha :) 2022-02-16 12:12:47 But impressive gawk made it go to half the time 2022-02-16 12:15:53 heh that warning about noarch 2022-02-16 12:16:24 I don't know what that implies 2022-02-16 12:17:35 Im sure it's unrelated 2022-02-16 12:18:29 Greetings. 2022-02-16 12:18:43 Hi cousin_luigi 2022-02-16 12:19:03 His a "sillysausage" user here? He wrote a page in the wiki I would like to ask him about. 2022-02-16 12:19:07 because https://git.alpinelinux.org/abuild/tree/abuild.in#n1282 probably doesnt know how to handle kernel images 2022-02-16 12:20:37 scanelf /boot/ doesnt show vmlinuz 2022-02-16 12:21:39 (because it's a PE binary, not ELF) 2022-02-16 12:22:06 Is this something to fix, or can it be safely ignored? 2022-02-16 12:27:25 it can be ignored 2022-02-16 12:33:07 going to test on the host without NUMA domains 2022-02-16 12:34:28 running 7zip benchmark (7z b) 2022-02-16 12:35:27 https://dpaste.com/35768GVUR 2022-02-16 12:38:12 a bit better than the xeon bare metal machine https://dpaste.com/GCKLXH4FA 2022-02-16 12:53:12 No real difference on the non-numa config. 16 minutes 42 seconds 2022-02-16 12:55:30 trying on tmpfs instead 2022-02-16 12:57:11 Looking at iotop it doesn't look like that many parallel jobs are actually running. Only 5-8 or so 2022-02-16 20:15:55 So I have been running my own local mirror repository, but had a snap foo and lost that VM. So I setup a new VM and created the script to create the new mirror. One thing I noticed is the download speed is fairly slow and at the rate I am going I would be finished in a month and a half. I know the first time I set up my local mirror it took less then 24 hours to have it all in place. So I guess my question is has something changed from the wiki that 2022-02-16 20:15:55 I am missing or a way to speed it up? Thanks for any input. 2022-02-16 20:39:04 minimal: does your build system also support removing files that the particular installation does not need 2022-02-16 20:40:56 skinkie: depends on what exactly you mean :-) My script has an "--optimise" option that for example tweaks the initramfs to have less things present and so a couple of other things. However it does not provide a mechanism to delete arbitrary files 2022-02-16 20:41:26 so what sort of files had you in mind? 2022-02-16 22:56:48 minimal: name checks out 🤔 2022-02-16 22:56:52 Hehe 2022-02-16 23:40:14 minimal: so it isn't possible to tweak mesa-gallium to only have vc4 files on a raspberry pi and never nvidia or amd stuff? 2022-02-16 23:51:16 pretty sure mesa-dri-gallium contains only one file 2022-02-16 23:52:18 for aarch64 that package has 20+ files: https://pkgs.alpinelinux.org/contents?branch=edge&name=mesa-dri-gallium&arch=aarch64&repo=main 2022-02-16 23:53:30 skinkie: I guess you could do something like create a package that had a trigger on the /usr/lib/xorg/modules/dri/ directory so that it could delete most files from there whenever some installs stuff there 2022-02-17 00:01:40 mesa-dri-gallium contains one file and 36 hard links. you cannot save any meaningful space by deleting files fromit 2022-02-17 00:02:45 Hello71: ok, the pkgs.alpinelinux.org website doesn't indicate those are softlinks 2022-02-17 00:03:06 Hello71: I thought it was much bigger. 2022-02-17 00:03:08 it might be hard to list hard links 2022-02-17 00:03:16 but i don't think it lists symlinks either 2022-02-17 00:03:37 doh! misread as softlinks, yeah harder to list hardlinks 2022-02-17 00:03:37 alpine linux normally tries quite hard to use subpackages appropriately. rarely can you save substantial space by deleting specific files 2022-02-17 00:03:47 But I guess you understand the question. If I want to have a specific platform lets say raspberrypi. Is there any way that would optimise distribution? 2022-02-17 00:04:15 because this also applies to kernel modules etc. 2022-02-17 00:04:22 the only exception i can think of is big language libraries, e.g. python 2022-02-17 00:04:47 now I can fit everything into memory at this moment, but still modloop is > 100MB which is a waste 2022-02-17 00:04:50 if you have some specific program then maybe you could delete the unused modules. but i think there are better ways to do it than manually 2022-02-17 00:04:51 skinkie: apart from the RPI kernel (and a few others like rpi bootloader, etc), all the other armv7/aarch64 packages are not RPI-specific 2022-02-17 00:05:07 ah, yes, kernel modules is the other exception 2022-02-17 00:05:52 skinkie: as I mentioned earlier you might be able to optimise what is in the initramfs 2022-02-17 00:05:58 minimal: I come from a LinuxFromScratch and Gentoo mindset, so... I like what alpine does. But wouldn't mind to have some tuning abilities. 2022-02-17 00:06:12 gentoo is actually super fat 2022-02-17 00:06:50 Hello71: depends what you do with it. 2022-02-17 00:08:24 not really. a standard gentoo installation always stores unpacked headers, docs, full ports trees, inefficient local package database format... yes, you can theoretically make a small distro if you cherry-pick files, but you can do that with any distro 2022-02-17 00:08:30 and it's not a standard installation anymore 2022-02-17 00:08:32 skinkie: Alpine is a general purpose distro 2022-02-17 00:09:51 oh, and portage *still* doesn't support any kind of subpackages, for headers but also debug symbols 2022-02-17 00:10:01 there are "tweaks" you can make depending on your situation - e.g. I use Alpine on RPIs as servers running in Sys-mode rather than run-from-ram so for example I use the cutdown RPI bootloader rather than the "normal" RPI bootloader 2022-02-17 00:10:37 doesn't make a big difference, its just one of various little tweaks possible 2022-02-17 00:26:39 is it possible to make alpine boot without waiting for udhcpc to finish? as in, have it run in the background? 2022-02-17 00:29:53 adnano: this has been asked several times before in the past 2022-02-17 00:30:33 shouldn't it be possible if you have no services that depend on net or whatever? 2022-02-17 00:31:48 grep 'need net' /etc/init.d/* and comment it out :) 2022-02-17 00:33:43 adnano: there are options to wait a defined time is no lease obtained (20 secs) or to background if no lease obtained 2022-02-17 00:34:15 but don't believe there is an option to not try and get a lease in the first place - why else would you run udhpc if you did not want a lease? 2022-02-17 00:35:49 imo waiting for net makes no sense. services that need network connectivity already need to deal with 'network down' condition and sleep and retry whatever they're doing 2022-02-17 00:40:32 dalias: if don't think their issue/complaint is that it is waiting for net in general, rather than udhcpc makes several attempts, with a delay inbetween, trying to get a lease 2022-02-17 00:41:18 assuming ifupdown-ng is installed then this is controlled by /usr/libexec/ifupdown-ng/dhcp 2022-02-17 00:42:27 adnano specifically said "boot without waiting for udhcpc to finish" 2022-02-17 00:42:31 where its running udhcpc with "-b" and "-R" options 2022-02-17 00:42:49 and asked for it to run in background 2022-02-17 00:43:53 dalias: yes but as I said earlier whilst the time between lease retries and number of reties can be configured I don't think udhcpc can be configured to immediately go into background before the 1st lease request 2022-02-17 00:44:33 it doesn't need to be 2022-02-17 00:44:44 the problem is just that other stuff is waiting on it to finish for no reason 2022-02-17 00:45:07 if no other services have a dep on it finishing first, boot should complete without it finishing 2022-02-17 00:49:09 dalias: true, and the "net" dep doesn't guaranteed that network is up - indeed a reliance on "net" could be just to ensure that an interface has been renames or creates (i.e. bridge/vlan interfaces) etc 2022-02-17 01:00:29 yeah. that's basically how i think the dependence as it's implemented now is a logic error 2022-02-17 01:02:05 looks like it's possible by replacing udhcpc with dhcpcd and using the -b flag, which makes it background immediately. should I modify /usr/libexec/ifupdown-ng/dhcp, or is there a better way? 2022-02-17 01:09:42 dalias: there is also a net-online init.d service 2022-02-17 01:12:27 adnano: you don't need to modify /usr/libexec/ifupdown-ng/dhcp, if you look at the top of that file it checks for the presence of dhcpcd, then dhclient, and then udhcpc 2022-02-17 01:13:04 minimal: right, but I want to use it with the -b command line flag 2022-02-17 01:14:55 adnano: then install dhcpcd and edit its settings in /etc/init.d/dhcpcd 2022-02-17 01:33:56 minimal: if I add dhcpcd to the boot runlevel, openrc will try to start it multiple times. 2022-02-17 01:34:24 adnano: ifupdown-ng starts it 2022-02-17 01:34:48 when it reads /etc/network/interfaces and sees "dhcp" listed for an interface 2022-02-17 01:35:38 sorry, forget about /etc/init.d/dhcpcd :-) 2022-02-17 06:15:20 Hi all, is it expected that expat 2.4.4 ( with the CVECVE-2022-23852 fix) will still be deployed in Alpine 3.12 even though eol is in two months? 2022-02-17 07:53:31 Hi again. I tested abuild -r on the linux-lts kernel on tmpfs. It saved about 1-2 minutes. So I am down to 13m 49s :) 2022-02-17 08:03:36 Hi alpine people! Is there a reason why the zfs packages in 3.15.0 are all on 2.1.2 but the zfs-lts package is built with 2.1.1? 2022-02-17 08:04:05 I was hit with #13198 when upgrading from 3.14 to 3.15 without reading the release notes about ZFS (my bad) 2022-02-17 08:05:11 Rebuilding zfs-lts with 2.1.2 worked perfectly, I was wondering if there's a reason 3.15 shipped with 3.15 or if I should send a MR? 2022-02-17 08:26:16 Dustin_: unless someone makes a merge request updating / fixing expat 2022-02-17 08:27:22 m4t: I guess someone upgraded one without the other 2022-02-17 08:27:37 No idea if there was a specific reason 2022-02-17 08:35:21 ikke: thanks, I sent an MR ( 2022-02-17 08:35:21 !31020 2022-02-17 09:44:58 hey I can't seem to get a raspberry pi to boot with alpine, not sure what I'm doing wrong as I've done this before. 2022-02-17 09:45:50 I've been following the headless install guide on the wiki which lists it as tested against alpine 3.14 (which I've tried_ 2022-02-17 10:04:03 ok the trick is use raw disk, not a partition 2022-02-17 10:44:09 What is the alpine way of triggering the creation of devices/partitions in /dev after fdisk? 2022-02-17 10:47:33 doesnt that just happen automatically? 2022-02-17 10:48:05 j`ey: lets say ... reproducably not 2022-02-17 10:48:12 hmm I don't think the *kernel* automatically re-reads partition tables all on its own, but fdisk itself should prod it to do that (at least util-linux fdisk does) 2022-02-17 10:48:41 if not, partx -u /dev/sdX (also util-linux) can do it from userspace... 2022-02-17 10:48:41 do you have devtmpfs on /dev type devtmpfs ? 2022-02-17 10:50:37 j`ey: devtmpfs on /dev type devtmpfs 2022-02-17 10:51:20 oh, is udev running 2022-02-17 10:51:37 grawity: it is not 2022-02-17 10:51:49 skinkie: and: apk info -W $(which fdisk) 2022-02-17 10:52:04 I don't remember the exact mechanism, but IIRC it also triggers the partition rescan whenever a block-device is closed 2022-02-17 10:52:07 /sbin/fdisk symlink target is owned by busybox-1.35.0-r2 2022-02-17 10:55:49 looks like busybox's fdisk should also re-read the pratition table after write 2022-02-17 10:56:42 you can perfectly reproduce this using the standard iso 2022-02-17 10:57:37 can you show the steps? 2022-02-17 10:58:28 boot in your favorite virtual machine, with a disk attached; fdisk a new partition; validate cat /proc/partitions; ls -l /dev/*da* 2022-02-17 10:58:31 https://paste.gg/p/anonymous/44c2498980554529b6cb1b3f91467a73 2022-02-17 11:01:33 skinkie: I can't reproduce it like ^ 2022-02-17 11:11:44 j`ey: one sec had to do call 2022-02-17 11:12:32 j`ey: let say that on the second boot when partitions were deleted I also have a valid vda1 2022-02-17 11:29:37 j`ey: mmm... can't reproduce it anymore. 2022-02-17 11:30:01 :/ 2022-02-17 11:31:18 j`ey: if it happens again, i'll make a pastebin. Crazy. 2022-02-17 11:32:59 j`ey: got it! 2022-02-17 11:33:31 https://paste.gg/p/anonymous/b956145d28ea4542b904f6cebde59a30 2022-02-17 11:33:57 oh, thats different 2022-02-17 11:34:11 j`ey: yes it is different, but that previous problem existed too 2022-02-17 11:34:57 cat /proc/filesystems 2022-02-17 11:36:38 https://paste.gg/p/anonymous/afc79f4eaeea441499f19c1edf1e709e 2022-02-17 11:36:59 fileystems, not partitions! 2022-02-17 11:39:58 https://paste.gg/p/anonymous/74c5855b35344c518ed9dd035dbfaf2b 2022-02-17 11:40:35 you dont have vfat there 2022-02-17 11:40:50 Wouldn't vfat get auto loaded by -t vfat? 2022-02-17 11:40:59 can you try: modprobe vfat, anyway 2022-02-17 11:41:13 modprobe: can't change directory to '/lib/modules': No such file or directory 2022-02-17 11:41:19 that is a missing modloop 2022-02-17 11:41:36 so i tthink what happens is after the reboot, the vfat gets probed from initramfs 2022-02-17 11:41:48 then 'magically' exists 2022-02-17 13:11:32 mh, nginx uses pcre instead of pcre2 2022-02-17 13:11:39 I should send a patch for the APKFILE 2022-02-17 13:12:07 *APKBUILD 2022-02-17 15:32:53 why lua packages have those weird paths? I tried to build a project from source, but its configure could not found lua (lua5.3-dev and lua5.3-libs are installed) 2022-02-17 15:34:33 eg. #include fails 2022-02-17 15:45:17 skinkie: "Wouldn't vfat get auto loaded by -t vfat?" - Busybox's mount (certainly in the past) did not autoload required fs modules whereas util-linux's mount did. I believe Hello71 submitted an upstream patch to address this 2022-02-17 15:47:15 Iirc, it requires the fs type to be defined in /etc/filesystems or something like that 2022-02-17 15:54:13 ikke: yes, that's what I've done in the past - it is noted in my README.alpine for cloud-init 2022-02-17 15:56:05 ikke: though that related to the "-t auto" option behaviour 2022-02-17 16:25:17 to be precise, busybox mount -t auto will try all filesystems in /proc/filesystems then /etc/filesystems. util-linux mount tries whatever blkid returns, then (iirc) /proc/filesystems 2022-02-17 16:26:13 by my reckoning, if busybox is already shipping the blkid detection code, it should also use it for mount -t auto 2022-02-17 16:26:31 but apparently busybox devs don't (care to) agree 2022-02-18 01:14:22 I'm trying to install GRUB to an Alpine chroot in a disk image, however when I run `grub-install --force --boot-directory=/boot --efi-directory=/boot/efi --removable --target=x86_64-efi` I get this error https://pastebin.com/kaGMFKWE 2022-02-18 01:14:22 The same error occurs when I try i386-pc as well. 2022-02-18 01:14:23 I did try manually making a device.map using a guide on the internet, but it always said the device map was invalid. Furthermore, installing GRUB via the host's grub-install (Arch) works fine. I've checked and it looks like I have all the necessary packages installed and the necessary mounts. I compared the build options between Arch and Alpine and I couldn't see any build options that would explain this. 2022-02-18 01:14:23 /etc/mtab: https://pastebin.com/af3qiuV7 2022-02-18 01:14:23 Packages: https://pastebin.com/uD3sxSAS 2022-02-18 01:15:05 Any advice would be much appreciated 2022-02-18 01:19:32 logan2611: yes grub has some issues with dealing with loopback devices 2022-02-18 01:20:18 very rarely is it correct to install to a partition, especially a loop partition 2022-02-18 01:20:49 device.maps basically aren't used by current Grub - in my own script for create disk image via loopback/chroot I ended up temporarily installing eudev as Grub used udev for some logic if it is installed and fallback to its own logic if not 2022-02-18 01:22:17 Hello71: Grub doesn't like partitioned loopback devices 2022-02-18 01:22:30 sure, but, like, why would you want that 2022-02-18 01:22:50 Hello71: when you creating a partition disk image 2022-02-18 01:23:03 s/partiion/partitioned/ 2022-02-18 01:24:23 look, if i say "why are you using partitioned loop devices", if you answer "because i'm using partitioned loop devices" that's not helpful 2022-02-18 01:25:11 I said I am creating disk images which are partitioned, therefore as I am using a loop device then I partition the loop device 2022-02-18 01:26:14 so for a UEFI disk image the "disk" is /dev/loop0 with e.g. 2 partitions, /dev/loop0p1 for the ESP partition and /dev/loop0p2 for the rootfs partition 2022-02-18 01:28:54 Hello71: was that more helpful? 2022-02-18 01:29:18 in this case you should invoke grub-install /dev/loop0 2022-02-18 01:29:29 in the same manner that you would invoke grub-install /dev/sda, not /dev/sda1 2022-02-18 01:30:14 you would only invoke grub-install /dev/loop0p1 if you were subpartitioning /dev/loop0p1, which while possible seems like an extremely unusual use case 2022-02-18 01:30:21 Hello71: I haven't indicated what options I'm passing grub-install and yes I do use /dev/loop0. 2022-02-18 01:30:59 why is there so much snark 2022-02-18 01:31:32 The original person also didn't indicate as far as I saw what grub-install option he is using. He did post a grub-install error but that not details of the options he used 2022-02-18 01:32:09 they did but i misread. sorry 2022-02-18 02:24:56 Hi, is "rc-service start" asynchronous? (Does it return before the service has fully started?) 2022-02-18 02:25:15 it returns as soon as the start() function returns 2022-02-18 02:25:37 i would guess 2022-02-18 02:25:38 the service might not actually be ready yet to perform its functions 2022-02-18 02:25:41 Sometimes when running setup-alpine on certain systems (VMs), setup-apkrepos has a certificate error, I'm wondering if it's because the date/time isn't fully initialized yet 2022-02-18 02:25:42 yep 2022-02-18 02:25:43 no readyness 2022-02-18 02:25:50 it can be 2022-02-18 02:25:51 psykose: raven523: I see, thanks 2022-02-18 02:26:02 but if you use chrony, it should block until time is set 2022-02-18 02:26:11 and without rc_parallel, anything after it should be fine 2022-02-18 02:26:15 since only one service starts 2022-02-18 02:26:21 Should I test by adding a "sleep 10" to the end of setup-ntp? 2022-02-18 02:26:28 well 2022-02-18 02:26:34 you could just use chrony 2022-02-18 02:26:40 or a vm time syncing solution 2022-02-18 02:26:47 or figure out how to make default ntpd block 2022-02-18 02:26:48 or something 2022-02-18 02:29:11 Oh I just remembered what's the problem. setup-ntp isn't called if running in qemu, and the QEMU configuration being used doesn't set rtc time properly, so ntp is needed. 2022-02-18 02:29:31 i would set up the configuration 2022-02-18 02:31:25 psykose: agreed, but it's a VM app (https://github.com/utmapp/UTM), and so far customizing the rtc option hasn't been implemented. 2022-02-18 02:31:34 psykose: What do you recommend it should be set to? 2022-02-18 02:31:48 ah, then no idea 2022-02-18 02:31:56 i would figure out how to run whatever ntp thing correctly 2022-02-18 02:32:53 The problem is here[0], I remember asking about this a few months back when I first tried out Alpine, and iirc the response was because this isn't neccessary on properly set up qemu VMs. [0]: https://gitlab.alpinelinux.org/alpine/alpine-conf/-/blob/master/setup-alpine.in#L189 2022-02-18 02:33:43 psykose: The option used is "-rtc base=localtime". Does that seem correct? 2022-02-18 02:34:12 this is setup-alpine, you can still invoke setup-ntp yourself 2022-02-18 02:34:14 and uhh, no idea 2022-02-18 02:35:35 psykose: The trouble is that you need networking set up before running setup-ntp, and then if you re-run setup-alpine, when it gets to the setup-networking step it suggests a static IP instead of dhcp 2022-02-18 02:37:35 want a really pro solution 2022-02-18 02:37:49 sed -i "someshitinsetupalpine" setupalpine; setup-alpine 2022-02-18 02:38:01 and remove the qemu checks 2022-02-18 02:38:02 :p 2022-02-18 02:38:22 could be as simple as removing the '!' in the if 2022-02-18 02:38:30 it's annoying there's no toggle for it, but you could also do setup-alpine; setup-ntp after 2022-02-18 02:38:31 oh 2022-02-18 02:38:32 right 2022-02-18 02:38:34 apkrepos is in it 2022-02-18 02:38:34 nvbm 2022-02-18 02:38:53 But the problem is it's not very easy to type, and UTM is used on things like iPads so copy paste isn't an option 2022-02-18 02:39:01 yep 2022-02-18 02:42:23 psykose: Do you think adding ",clock=host" to the rtc configuration would help? 2022-02-18 02:42:59 Or should the "is qemu" check be removed since people can just answer "none" if they don't want ntp? 2022-02-18 02:43:17 ktprograms: do you have either of "hwclock" or "osclock" init.d services enabled? 2022-02-18 02:43:31 clock=host should work by the sound of it 2022-02-18 02:44:26 minimal: I'll check, it's just a boot with the default standard ISO 2022-02-18 02:47:15 minimal: The started services (listed with rc-status) doesn't have hwclock or osclock. The list is sysfs, fsck, root, localmount, firstboot 2022-02-18 02:47:53 Is it expected that it starts when booting from the ISO? 2022-02-18 02:48:10 ktprograms: "-rtc base=localtime" as a QEMU option would make sense if you're running Qemu on a Windows machine (as typically they set the RTC to localtime not UTC) 2022-02-18 02:48:31 minimal: It's a Mac host 2022-02-18 02:48:49 ok, not sure which MacOS sets RTC to 2022-02-18 02:48:54 But I'm not in UTC, should it still be base=UTC? 2022-02-18 02:48:59 minimal: Any way to check? 2022-02-18 02:50:07 it depends on the rtc, not whatever system time 2022-02-18 02:50:10 just boot with both 2022-02-18 02:50:16 if it's the wrong one the time will be fucked up 2022-02-18 02:50:22 offset by some hours 2022-02-18 02:52:16 I assume the Mac has something similar to "hwclock" to check the time 2022-02-18 02:53:48 psykose: On a fresh boot from the ISO, it thinks it's in Nov 5 2021. If I run "service hwclock start", then it seemingly gets the correct time. 2022-02-18 02:54:05 minimal: Not sure. Do you mean time from the rtc or what? 2022-02-18 02:54:46 then the hwclock works 2022-02-18 02:54:57 you could just start it then run setup-alpine 2022-02-18 02:54:58 :) 2022-02-18 02:55:18 psykose: Yep, thanks 2022-02-18 02:55:28 Is it supposed to start by itself? 2022-02-18 02:55:32 (the service) 2022-02-18 02:55:39 by default neither that nor swclock is enabled 2022-02-18 02:56:00 not everything has a hwclock, swclock has.. issues, or something 2022-02-18 02:56:07 psykose: I see. ig it's hard to detect which (if any) should be used 2022-02-18 02:56:08 someone that knows more can walk you through the finer points 2022-02-18 02:56:23 but since it's qemu, i would assume for hwclock to always work 2022-02-18 02:57:01 psykose: Yes, I would hope so 2022-02-18 02:57:14 Mac seems to use UTC like Linux 2022-02-18 02:57:42 minimal: Does that mean it should be "-rtc base=UTC" ? 2022-02-18 02:58:13 ktprograms: re hwclock/swclock/osclock,well it depends on whether you expect to have NTP syncing or not 2022-02-18 02:58:46 I see 2022-02-18 02:59:06 ktprograms: according to Qemu manpage, "The default base is UTC" 2022-02-18 02:59:37 swclock is only intended for machines with no RTC (like Raspberry Pis) 2022-02-18 02:59:52 minimal: Does the part about "base=localtime" should be used for correct date on Windows refer to a Windows host or guest? 2022-02-18 03:00:34 osclock (which doesn't actually do anything) is intended for machines expecting to have continuous access to NTP 2022-02-18 03:01:17 ktprograms: that's referring to the host machine 2022-02-18 03:01:51 minimal: So if this VM software is only meant to run on i*OS and macOS, it should just always be "-rtc base=UTC"? 2022-02-18 03:02:44 Will it break any guests? I noticed that with "base=localtime", when I start hwclock it has the local time but with UTC timezone, which is wrong. 2022-02-18 03:03:12 well the manpage says "-rtc base=utc" is the default (i.e. if you do not specify any "-rtc" value) 2022-02-18 03:03:41 minimal: And default _should_ be correct most of the time? 2022-02-18 03:04:24 ktprogram: the Alpine timezone is set by setup-timezone (strangely enough) ;-) 2022-02-18 03:04:32 well should be called by setup-alpine 2022-02-18 03:05:42 ktprogram: the default is correct for Linux and Mac hosts (unless you've purposely changed their RTC to something other than UTC) but not for Windows hosts 2022-02-18 03:05:47 Yes, but if the hwclock is set to UTC, and you set the timezone, then the time will show up as 2022-02-18 03:07:37 that is how it normally works? 2022-02-18 03:07:40 minimal: Look at https://0x0.st/o8yK.png 2022-02-18 03:07:49 psykose: Yes, but then it has the wrong date/time 2022-02-18 03:07:59 is your host UTC 2022-02-18 03:08:08 +8 2022-02-18 03:08:10 ah 2022-02-18 03:08:11 then yes 2022-02-18 03:08:12 in this case 2022-02-18 03:08:14 whatever you have set 2022-02-18 03:08:15 is wrong 2022-02-18 03:08:19 and you want the other one 2022-02-18 03:08:31 That's why the QEMU rtc base should be UTC 2022-02-18 03:08:32 UTC should be 3 am in that picture, in the vm 2022-02-18 03:08:36 yeah 2022-02-18 03:08:39 and that should be the default 2022-02-18 03:08:46 the adding 8 hours should be done by the VM and not QEMU 2022-02-18 03:08:48 i assume that picture is base=localtime? 2022-02-18 03:08:55 psykose: yes 2022-02-18 03:08:56 yes 2022-02-18 03:08:58 so, don't do that 2022-02-18 03:09:04 keep it at the default (=UTC_ 2022-02-18 03:09:08 and that picture should become correct 2022-02-18 03:09:09 ok time to make a PR to UTM 2022-02-18 03:09:09 :) 2022-02-18 03:09:18 ah, right 2022-02-18 03:09:18 If I keep it at the default, is the -rtc argument needed? 2022-02-18 03:09:32 nope, unless you want to make it explicitly 'documented' 2022-02-18 03:09:38 sometimes i leave the defaults out, sometimes i leave them in 2022-02-18 03:09:39 shrug 2022-02-18 03:09:48 confused - you're Mac is UTC+8 and you want your Alpine VM to be UTC rather than UTC+8? 2022-02-18 03:10:11 yeah 2022-02-18 03:10:20 so it's only the 'same' as the host if you also set +8 in the vm 2022-02-18 03:10:25 i believe that is correct generally 2022-02-18 03:10:27 minimal: No, I want the VM to be also UTC+8, but when I start the hwclock before setting the timezone, then the date is but at UTC timezone 2022-02-18 03:10:29 and not setting utc=host 2022-02-18 03:10:30 sorry, clarification: your Mac is set to timezone that is UTC+8? 2022-02-18 03:10:33 yes 2022-02-18 03:10:55 minimal: So when you set the timezone the offset is added to the date, which makes it wrong (the second date in the picture) 2022-02-18 03:11:41 why are you setting timezone AFTER hwclock? 2022-02-18 03:13:21 "date" shows your time taking your timezone into consideration, "date -u" shows you time in UTC timezone regardless of timezone setting 2022-02-18 03:14:23 minimal: Because the simplest for a user to do is to start hwclock then run "setup-alpine", which calls "setup-timezone" 2022-02-18 03:14:55 minimal: Nevermind, I think the correct thing is to use base=utc 2022-02-18 03:17:51 ktprograms: base=utc is the default so if you did not specify any "-rtc" option then you are already using it 2022-02-18 03:19:18 minimal: Yes, but like I said earlier, UTM sets base=localtime. Given what has been found here, I think just removing the rtc option and letting it use the default should be correct. Thanks 2022-02-18 03:19:38 can't comment on UTM as don't have a Mac 2022-02-18 03:19:54 minimal: that's fine, it's basically just a UI wrapper to qemu 2022-02-18 03:21:01 krprograms: Linux typically stores the time internally as UTC, the timezone setting is only used by the system (i.e. things like "date") to display it in the local time 2022-02-18 03:21:32 minimal: Right, thanks 2022-02-18 03:22:16 so you need to ensure that "date -u" show the correct UTC time for now, then sort out the timezone which will then ensure that "date" shows the correct local time 2022-02-18 03:22:31 minimal: Yup, that's what I gathered. Thanks! 2022-02-18 03:23:24 re: hwclock / osclock - as I said the decision relates as to whether NTP sync will be active all/most of the time or not 2022-02-18 03:23:44 minimal: I doubt it will be because setup-alpine skips it 2022-02-18 03:25:13 typically the linux-virt kernel (for a x86/x86_64 VM) will be configured so that the kernel during initialisation will set the system clock from the RTC and then once NTP sync is achieved for the system clock the kernel will approx every 11mins then sync the RTC with the system clock 2022-02-18 03:25:51 minimal: I see. That explains why it works with the Virt ISO but not the Standard one 2022-02-18 03:26:34 krprograms: the above still applies if neither hwclock nor osclock is enabled (osclock doesn't actually anything except satisfy other init.d service dependancy on "clock") 2022-02-18 03:27:09 ktprograms: linux-lts kernel should act in the same way, I was focusing on linux-virt as you're running a VM 2022-02-18 03:29:40 minimal: Hmm interesting. Not sure why the time isn't set then 2022-02-18 03:29:50 It's aarch64 btw 2022-02-18 03:30:24 linux-lts or linux-virt aarch kernel? 2022-02-18 03:30:33 linux-lts (standard iso) 2022-02-18 03:32:12 EFI/UEFI? 2022-02-18 03:32:23 yes 2022-02-18 03:36:32 well CONFIG_RTC_HCTOSYS=y and CONFIG_RTC_SYSTOHC=y appear to be enabled in kernel config, they enable to RTC->System clock at boot and the System clock->RTC at 11 min intervals (*once* NTP sync has occurred) 2022-02-18 03:37:32 ah! CONFIG_RTC_DRV_EFI=m 2022-02-18 03:38:01 that looks like the problem, that's the EFI/UEFI RTC driver as a module, its not built into the kernel 2022-02-18 03:38:19 I see 2022-02-18 03:38:57 so the kernel sets System clock from RTC when it initialises - but if the RTC driver is a loadable module then it is not loaded until after kernel initialises 2022-02-18 03:39:10 and so kernel does not set System clock from RTC 2022-02-18 03:39:21 And that's why manually using hwclock is needed? 2022-02-18 03:39:44 well that's one thing that hwclock init.d will do 2022-02-18 03:41:02 try doing a "dmesg | grep rtc" so see what happened during boot 2022-02-18 03:42:15 rtc-efi seems to have set system clock to the correct time ( UTC, that's just a problem with the "base=localtime" 2022-02-18 03:42:27 minimal: I'm just summarizing, do you want a screenshot? 2022-02-18 03:42:57 But date shows some time in Nov 5 2021 2022-02-18 03:44:08 ktprograms: the first part of dmesg output is a time offset from when the kernel started, so how many seconds does it show for the "setting system time" entry 2022-02-18 03:45:46 minimal: 5.3 2022-02-18 03:47:01 yeah, so 5.3 seconds from kernel init whereas on a machine I'm looking at here its set 0.5 seconds from kernel init 2022-02-18 03:47:37 minimal: Is it also from "rtc-efi rtc-efi.0" ? 2022-02-18 03:51:57 its 5.3 seconds as that's how long it took kernel to initialise, to load and run initramfs, mount rootfs, start openrc, the modules init.d service to load rtc-efi etc (assuming rtc-efi is not included in the initramfs) 2022-02-18 04:06:43 minimal: I see, thanks 2022-02-18 04:08:48 it probably isn't a concern to you but if you want the time to be set as early as possible so that logs etc are accurate then the driver needs to be built into the kernel 2022-02-18 04:09:28 minimal: I don't think that's a big concern. What I'm more wondering is why the date output is so wrong before starting hwclock, even though dmesg shows the correct time 2022-02-18 04:12:17 your earlier screenshot made no sense as the difference is 3+ months, its not a timezone difference as the minute values don't align 2022-02-18 04:13:17 wow it really just has my ip in there lol 2022-02-18 04:13:25 it wasn't, it was feb18 on both (after hwclock started) 2022-02-18 04:13:36 before hwclock it was fucked up for sure 2022-02-18 04:13:58 as for dmesg, that hasn't been posted, so no idea 2022-02-18 04:14:00 could i see it 2022-02-18 04:14:19 minimal: New screenshot, including dmesg: https://0x0.st/o8yj.png 2022-02-18 04:14:43 The time shown in dmesg is correct (or as correct as it can be with base=localtime, ignore the timezone) 2022-02-18 04:15:54 that's showing almost exactly the same date/time as the earlier screenshot 2022-02-18 04:16:12 minimal: The output of the date command is crazy 2022-02-18 04:16:25 like its some default value set during boot/VM initialisation 2022-02-18 04:16:27 i think rtc-efi failed to set the time 2022-02-18 04:16:38 or there is some other thing after that that sets it again 2022-02-18 04:17:32 psykose: I don't see anything implying it failed to set the time 2022-02-18 04:17:48 you could post the whole dmesg, but i doubt there's anything in there 2022-02-18 04:19:06 psykose: Do you want me to post it? I looked through it and I don't see anything important 2022-02-18 04:19:30 sure 2022-02-18 04:19:39 as a general rule nobody complains about more logs :) 2022-02-18 04:19:53 even if they are mostly empty and don't show the issue 2022-02-18 04:20:13 ktprograms: is this run-from-ram Alpine install or Sys-mode? 2022-02-18 04:20:57 var/log/messages too i guess, just in case 2022-02-18 04:23:03 minimal: booting the ISO 2022-02-18 04:23:16 right, so run-from-ram then 2022-02-18 04:24:02 ok, right after a boot can you run "ls -l /lib/rc/cache/" ? 2022-02-18 04:25:07 dmesg: https://tpaste.us/YEK6 2022-02-18 04:25:35 /var/log/messages: https://tpaste.us/5nBx 2022-02-18 04:26:16 minimal: no such file or directory 2022-02-18 04:27:16 ok, then "ls -l /run/openrc" 2022-02-18 04:33:47 krprograms: I think the output of the above will confirm what I suspect is the problem 2022-02-18 04:34:58 don't you want the tree of run/openrc, or something other than top-level 2022-02-18 04:35:09 "ls -l /run/openrc" 2022-02-18 04:35:22 just that 2022-02-18 04:37:08 the /var/log/messages is confirming what I already suspected 2022-02-18 04:46:09 ktprograms: still around? 2022-02-18 04:47:28 psykose: think he's gone 2022-02-18 04:48:34 so what I am suspecting is that after he boots the file /run/openrc/deptree will be timestamped as Nov 5 11:17 2022-02-18 04:50:32 has anyone used diskless mode? I am running alpine on a pi and I don't think I set it up quite right. Can I just use /etc/fstab to mount a persistent /var and /etc? Or is there a specific way that data disk needs to be setup. 2022-02-18 04:51:08 as he's running from RAM then I suspect the APKOVERLAY file (or whatever its called that's managed by lbu) contains a deptree file with this timestamp and its copied to /run/openrc. On a disk-install (Sys-mode) this file is copied from /lib/rc/cache/deptree 2022-02-18 05:30:00 Sorry, had to make lunch 2022-02-18 05:30:59 /run/openrc/deptree is timestamped Nov 21 2021, whereas date says Nov 5 2021 2022-02-18 05:31:58 It seems the 3.15 ISOs were created on Nov 24 2021, is that related? 2022-02-18 05:39:58 Hello, I have a wifi dongle rtl8188ftv/rtl8188fu(both seem similar not exactly sure which one i have). 2022-02-18 05:41:06 The driver for it is available here https://github.com/kelebek333/rtl8188fu 2022-02-18 05:42:03 How can I install it on alpine linux? as far as I can tell, alpine does not have dkms support 2022-02-18 05:42:27 Test555g: We have akms, which has a similar purpose 2022-02-18 05:42:49 but someone would need to write an akmbuild for that 2022-02-18 05:44:16 I have one for another rt driver 2022-02-18 05:44:25 psykose: akmbuild? 2022-02-18 05:44:36 just the akms equivalent 2022-02-18 05:44:56 of whatever dkms is doing 2022-02-18 05:44:59 they look mostly the same 2022-02-18 05:45:16 psykose: I see. Haven't looked into this stuff before 2022-02-18 05:45:37 i don't know why that module is not mainlined 2022-02-18 05:45:42 https://github.com/jirutka/akms 2022-02-18 05:45:57 i quite hate dynamic modules generally 2022-02-18 05:46:05 what a pain in the ass 2022-02-18 05:46:13 yes, but you have little choice for out-of-tree modules 2022-02-18 05:46:25 that is what i mean 2022-02-18 05:46:42 i didn't check, but what is stopping this module being added to mainline 2022-02-18 05:46:51 it's existed since like 4.15 2022-02-18 05:47:07 zfs i understand, people love to argue about licences or something 2022-02-18 05:47:09 Is it in staging perhaps? 2022-02-18 05:47:12 shrug 2022-02-18 05:47:28 all i see is the github pages 2022-02-18 05:49:15 I see 8188eu in staging, but nothing about fu 2022-02-18 05:52:28 hmm, there is also a firmware blob 2022-02-18 05:52:45 does it not work via regular linux-firmware runtime loading 2022-02-18 05:52:52 like every other wifi/eth card 2022-02-18 05:53:51 Probably: https://github.com/kelebek333/rtl8188fu/blob/master/Makefile#L477 2022-02-18 05:53:55 but you need to ship the blob 2022-02-18 05:55:03 this doesn't seem different to every other fell-off-a-truck firmware we literally have 2022-02-18 05:55:14 :p 2022-02-18 05:59:58 Test555g: not tested at all yet: https://tpaste.us/vglk 2022-02-18 06:35:18 can I mount /var and /etc using /etc/fstab in alpine diskless mode? 2022-02-18 08:34:20 Hello, I'm searching for a little help about alpine-linux 2022-02-18 08:35:04 I want to add a menu item to shutdown my machine on openbox, but when invoking doas /sbin/shutdown 2022-02-18 08:37:59 it raise this error "poweroff: (null): Operation not permitted" 2022-02-18 08:38:43 As I'm new of using doas and searching the web lead to very few informations, I ask here for help. 2022-02-18 08:40:37 carlo-alpine: alpine uses poweroff instead of shutdown 2022-02-18 08:41:14 oh sorry in reality is poweroff as the error message is howing 2022-02-18 08:41:26 ok 2022-02-18 08:41:33 what happens when you issue the command manually? 2022-02-18 08:41:38 howing = showing 2022-02-18 08:42:09 i guess you need to provide your password? 2022-02-18 08:42:24 I will try, but if it poweroff YI have to reconnect, let's see, if there are some delay it worked from command line 2022-02-18 08:45:26 Hi, I'm back , it work from the command line 2022-02-18 08:46:24 the issue is the script cannot prompt for your password 2022-02-18 08:47:02 i am not sure if there is a graphical-doas-prompt thing, the 'easiest' way is adding that command as a doas exclusion 2022-02-18 08:47:12 of course, then anything can power off your system as your user without being prompted 2022-02-18 08:47:39 I have put this line on doas.conf `permit nopass :wheel as root cmd /sbin/poweroff ` 2022-02-18 08:48:07 are you in the wheel group 2022-02-18 08:48:13 I think you have to configure elogind or something 2022-02-18 08:48:15 So I've tought that it will suffice as my user is member of Wheel group 2022-02-18 08:48:19 Yes 2022-02-18 08:48:23 that should work 2022-02-18 08:48:49 what's the exact text of the command? that you ran and that's in the file 2022-02-18 08:48:58 but If i didnìt go wrong, probably openbox is not using my user to issue commands 2022-02-18 08:49:34 do you type `doas poweroff` or `doas /sbin/poweroff` 2022-02-18 08:49:39 pretty sure only the latter works 2022-02-18 08:49:45 with that entry 2022-02-18 08:49:56 `/sbin/poweroff ` this is the line in menu.xml 2022-02-18 08:50:13 carlo-alpine: do you see "wheel" it in: groups | grep wheel 2022-02-18 08:50:15 that doesn't look like it has doas in it 2022-02-18 08:51:04 carlo-alpine: FYI if you don't want to make poweroff nopass, you can spawn a terminal that runs doas poweroff (for example "xterm -e doas poweroff") 2022-02-18 08:51:16 ```carlo-alpine disk lp wheel audio cdrom dialout video netdev games users sudo``` 2022-02-18 08:51:59 carlo-alpine: That looks correct, but like others said, you don't have doas in that config line 2022-02-18 08:52:12 Ok I catch, the thing, what stupid error 2022-02-18 08:52:14 check the allowing shutdown and reboot section in https://wiki.alpinelinux.org/wiki/Xfce it should work for any de 2022-02-18 08:52:54 carlo-alpine: Also try what I said for not needing to make poweroff a nopass option, since that's quite dangerous 2022-02-18 08:53:32 let's me try the command, but they are already installed 2022-02-18 08:54:04 One minor thing, i see in the output of apk 1 error listed, but how to see what error is, I've searched around but no luck 2022-02-18 08:54:44 I have migrated from 3.14 to 3.15 and some thing has to be modified like lightdm and xorg not properly set after migration 2022-02-18 08:55:26 apk fix will show you 2022-02-18 08:57:01 thanks let's try and see the output ``` Reinstalling polkit-elogind-libs (0.119-r8)```, maybe it was this that prevent the usual way of using lightdm options to exit 2022-02-18 08:59:08 sorry let's try is not pertinent, It's some time that I don't use IRC so forgive some errors, how to put properly the code block, usual decorations like `` or ``` ``` seems not working 2022-02-18 08:59:45 you don't :) 2022-02-18 08:59:59 IRC is plain-text 2022-02-18 09:01:07 OK maybe I'm remeber wrong being used to other chat, to send some code I rember well on using some external service and put a link here, but it's at least three years that i don't do serious use of IRC 2022-02-18 09:02:10 (that's why I like IRC, because it's plain-text, hehe) 2022-02-18 09:02:27 Let's try if it work, if there are some delay it works, so bye for now, I return here to let you know and thanks you all 2022-02-18 09:03:28 can I ask some packaging questions? 2022-02-18 09:03:34 sure 2022-02-18 09:03:50 currently I have 2 2022-02-18 09:04:09 first is a question from my friend 2022-02-18 09:04:28 he says debian has long package descriptions, while alpine has only the one-liner 2022-02-18 09:04:40 why is that? 2022-02-18 09:04:50 no idea, we simply never added long ones 2022-02-18 09:04:58 i also see no use for them 2022-02-18 09:05:10 me neither :P 2022-02-18 09:05:15 one can always... go to the website of the thing 2022-02-18 09:05:16 but he insists to them 2022-02-18 09:05:28 surely we don't need to encode large amounts of random fluff text in aports 2022-02-18 09:05:37 agree 2022-02-18 09:05:48 I feel like package descriptions are so you know you've got the right thing 2022-02-18 09:06:03 that is all i've ever used them for 2022-02-18 09:06:11 and like 5 words works for that pretty much all the time 2022-02-18 09:06:23 like the alpine fmt package could mean fmt from coreutils, but it doesn't 2022-02-18 09:06:39 yes, this is my usecase too 2022-02-18 09:06:47 then all is good in the world 2022-02-18 09:06:57 he says, "how can you find a package, if you don't know its name?" 2022-02-18 09:07:10 Hello again, it was the "polkit-elogind-libs" that fix the thing so no need the doas nopass anymore, some work on rewriting some code. 2022-02-18 09:07:20 how are you looking for a package if you don't know its name? 2022-02-18 09:08:06 I think he want to find a package for some functionality, grepping the long desc 2022-02-18 09:08:12 oslt 2022-02-18 09:09:29 Ok i think I've solved many of my problems as also the invocation of GUI program from doas seems to be working. 2022-02-18 09:09:33 (he has some weird beliefs, i.e. "if no long pkgdesc in pkg man, then it is no good" :P) 2022-02-18 09:09:40 carlo-alpine: nice :) 2022-02-18 09:09:43 So many thanks to all and Best Regards 2022-02-18 09:10:31 Bye! 2022-02-18 09:10:37 bb 2022-02-18 09:10:37 \disconnect 2022-02-18 09:10:57 then my second question 2022-02-18 09:11:06 the lua packages :) 2022-02-18 09:11:36 they packaged with quite non-standard conventions 2022-02-18 09:11:44 I tried to build a project that uses lua 2022-02-18 09:11:55 and its configure failed on #include 2022-02-18 09:12:34 I checked the file list of various lua versions 2022-02-18 09:12:43 (luaX-dev, luaX-libs, etc) 2022-02-18 09:13:00 all have the same convention 2022-02-18 09:13:09 that sounds normal to me 2022-02-18 09:13:31 lua.h works with lua5.1-dev, as is the standard 2022-02-18 09:13:42 i'm pretty sure you aren't supposed to if you want newer lua, but i could be wrong 2022-02-18 09:13:44 where these paths came from? 2022-02-18 09:13:48 what paths? 2022-02-18 09:14:00 sec 2022-02-18 09:14:02 they could also be expected to be passed with -I 2022-02-18 09:14:03 shrug 2022-02-18 09:14:21 in which case -I /usr/include/lua5.4 would work 2022-02-18 09:14:25 like this: /usr/include/lua5.3/lua.h 2022-02-18 09:14:42 and this: /usr/lib/lua5.3/liblua.so 2022-02-18 09:14:45 i don't see the issue 2022-02-18 09:15:18 if you do pkg-config --libs lua5.3 2022-02-18 09:15:22 you will get the correct paths 2022-02-18 09:16:24 okay, then it seems I have to file a bugreport in that project 2022-02-18 09:16:31 which one 2022-02-18 09:17:01 you might just be doing it wrong 2022-02-18 09:17:04 i can have a look 2022-02-18 09:17:15 send me the git url 2022-02-18 09:17:23 it's a 3rd party 2022-02-18 09:17:24 http://www.repo.hu/projects/fungw/ 2022-02-18 09:17:50 I try to make an APKBUILD for that 2022-02-18 09:19:20 is it supposed to work with lua5.3 2022-02-18 09:19:31 yes 2022-02-18 09:19:41 but I guess no pkg-conf is called 2022-02-18 09:20:10 i mean 2022-02-18 09:20:13 it is checking the dirs 2022-02-18 09:20:20 it executes test_c with each -I 2022-02-18 09:20:27 and they are even correct 2022-02-18 09:20:30 the error is something else 2022-02-18 09:20:35 ah 2022-02-18 09:20:38 missing the -L too 2022-02-18 09:20:40 during the linking 2022-02-18 09:20:44 yes 2022-02-18 09:20:51 uhh, i think they should just stop doing whatever this is 2022-02-18 09:20:54 and use pkg-config 2022-02-18 09:20:57 or any actual build system 2022-02-18 09:21:02 but as for fixing it, idk 2022-02-18 09:21:26 the reason there is no /usr/lib/liblua.so is because then they would conflict 2022-02-18 09:21:46 the lua packages generally pull all the luas at once and build against each, or whatever 2022-02-18 09:22:15 and there is no way to pass args to this 2022-02-18 09:22:44 ah 2022-02-18 09:22:49 you can pass --ldflags to the configure 2022-02-18 09:22:55 and just.. force add -L the path 2022-02-18 09:22:56 lol 2022-02-18 09:23:09 but yeah, i wish you luck 2022-02-18 09:23:19 thanks :) 2022-02-18 09:23:30 there is a --LDFLAGS for global, and a specific per-script ldflags too 2022-02-18 09:23:35 you can figure out how to invoke it for lua only 2022-02-18 09:23:50 dev initially said "no other distro do this" and "the distro is broken" 2022-02-18 09:24:18 passing --LDFLAGS=-L/usr/lib/lua5.3 works 2022-02-18 09:24:23 detected 2022-02-18 09:24:24 :) 2022-02-18 09:24:30 nice! 2022-02-18 09:24:37 I'll do this then 2022-02-18 09:24:47 top level configure passes $@ too 2022-02-18 09:24:51 so you don't need to go into any subdirs 2022-02-18 09:26:45 thanks 2022-02-18 09:27:16 btw, is mailing list considered as dead? 2022-02-18 09:28:54 I can't do the pkg build right now, but I'll let you know about the outcome if you interested 2022-02-18 09:31:32 not dead, but it's preferred to just make normal gitlab mr's 2022-02-18 09:31:48 since you can actually edit them 2022-02-18 09:36:10 hmm 2022-02-18 09:36:18 one more thing 2022-02-18 09:36:47 why is "/usr/lib/lua5.3/liblua.so" better than "/usr/lib/liblua5.3.so"? 2022-02-18 09:37:08 Suggestion for a simple terminal emulator? 2022-02-18 09:37:10 they would be the same 2022-02-18 09:37:30 Test555g: st :D 2022-02-18 09:37:31 jk 2022-02-18 09:37:35 Personally I use xterm 2022-02-18 09:38:26 on wayland, foot/alacritty 2022-02-18 10:41:37 Hi. i just noticed that busybox 'cp' doesnt support files with holes and instead copies the whole thing with zeroes. 2022-02-18 10:44:27 does that mean if i copy my head with busybox cp there will be a lot of zeroes in the world 2022-02-18 10:44:44 i don't seem to remember how cp is supposed to handle sparse files, does coreutils cp do it correctly 2022-02-18 10:52:38 yes coreutils does it right, at least on my other machines. haven't tested with alpine coreutils yet 2022-02-18 10:54:17 Forza: alpine coreutils does it correctly 2022-02-18 11:05:09 aron: I suppose I could install those 2022-02-18 11:05:35 Was just a comment really. Perhaps busybox will support it in the future, like reflinks as well 2022-02-18 11:06:37 idk, maybe :) 2022-02-18 13:03:02 psykose: my fungw package successfully built :) 2022-02-18 13:03:06 thanks again 2022-02-18 13:03:13 :) 2022-02-18 14:04:59 ktprograms: the "clock skew" messages in your logs are an indication that openrc is changing the system clock date/time during boot to Nov 2021, and then later in boot the efi rtc drive loads and date/time is changed again using RTC and so is then accurate - you can see in the time changing of the timestamps of the log entries you provided 2022-02-18 14:09:21 minimal: I see, but why does date show the wrong time? 2022-02-18 14:14:33 because OpenRC i setting the system clock using the date of a file when it first starts. I have seen this before on a RPI which has no RTC - the machine's system clock defaults to a date/time in 1970, then when OpenRC starts it sets system clock to the date/time of the /lib/rc/cache/deptree file and displays a "clock skew" warning. The /run/openrc/deptree file is created as a duplicate of the other deptree file 2022-02-18 14:15:47 then when you shutdown the machine the date/time of the /lib/rc/cache/deptree file is updated to the current date/time of the system clock (on the assumption that NTP has since run to make the system clock accurate) so that the revised date/time of this file will be used on next boot 2022-02-18 14:16:40 as you are booting off an ISO (using run-from-ram) then the date/time of that file (or its alternative) cannot be updated and so every time you boot sets the date/time to the same Nov value 2022-02-18 14:16:57 its meant to be a OpenRC "feature" for systems without a RTC 2022-02-18 14:18:02 whilst your system does have an RTC its handled by a loadable kernel module which is loaded AFTER OpenRC starts. If it was built into the kernel then the system clock would be set from RTC BEFORE OpenRC starts and you would not be seeing this date/time strangeness 2022-02-18 15:39:35 minimal: Thanks, makes a lot of sense. 2022-02-18 15:40:08 So in this case starting the hwclock service before running setup-alpine is correct? 2022-02-18 15:40:38 ktprograms: so I assume the Nov 5 2021 date related to a file on the ISO timestamped shortly before the Alpine 3.15.0 release 2022-02-18 15:42:02 ktprograms: you'll still have the system clock changing at different stages during boot and seeing the "clock skew" messages which are openrc complaining about the time changing 2022-02-18 15:45:04 ktprograms: are you always booting the VM using the ISO or are you booting the VM from the ISO once to then use setup-alpine to install alpine onto a (VM) disk? 2022-02-18 16:00:41 I am having a strange issue with network booting on a RaspberryPi4. Using the standard .tar.gz (which I don't load on an usb stick, but tftp) I get an error that init cannot be found. 2022-02-18 16:01:10 When taking an apkovl I end up with errors for busybox, fontconfig and fontscale. "Script exited with error 127" 2022-02-18 16:02:01 After a search https://github.com/tonistiigi/xx/issues/36 the suggestion there is that the script is of the wrong architecture. But why wouldn't busybox for aarch64 work for aarch64? 2022-02-18 16:57:02 skinkie: isn't error 127 "command not found"? so what command is it trying to run? 2022-02-18 16:58:02 minimal: how can that be observed? 2022-02-18 16:59:57 skinkie: I don't know as it is not clear to me exactly how/when the error(s) are happening 2022-02-18 17:00:52 a shell script can't be "of the wrong architecture" as a shell script is arch-independant, however the shell script may be expecting to run commands (i.e. binaries) that are missing, hence a "command not found" error 2022-02-18 17:03:33 minimal: let me reboot back to tftp, i'll make a screenshot. 2022-02-18 17:03:44 minimal: the actual tar.gz exacted on an usb device does work. 2022-02-18 17:10:36 I had 127 error some days ago because I monted some filesystem as 'noexec' 2022-02-18 17:10:54 s/monted/mounted 2022-02-18 17:11:20 donoban: ouch, how long took it to figure that out? 2022-02-18 17:11:53 few seconds, I asked on #alpine-devel and psykose pointed something similar 2022-02-18 17:12:08 "psykose are you building on some weird -ro filesystem or through qemu or some shit 2022-02-18 17:12:12 :) 2022-02-18 17:13:08 minimal: http://stefan.konink.de/boot-justkernel.png 2022-02-18 17:13:47 so as you can see, this system must have been able to access initramfs 2022-02-18 17:14:36 my guess is that there is a part of the init script preconfigured for a specific operation I cannot see. 2022-02-18 17:19:01 those 2 errors appear to be coming from these parts of initramfs' init: 2022-02-18 17:19:23 https://gitlab.alpinelinux.org/alpine/mkinitfs/-/blob/master/initramfs-init.in#L115 2022-02-18 17:19:48 and https://gitlab.alpinelinux.org/alpine/mkinitfs/-/blob/master/initramfs-init.in#L829 2022-02-18 17:20:05 minimal: My guess it happens already before that, something prevents the alpine-base 2022-02-18 17:22:25 the mount -o binds have not been executed 2022-02-18 17:25:57 or perhaps nothing has actually been installed to the root filesystem? have you tried modifying the cmdline options to remove "quiet" if it is present? 2022-02-18 17:26:23 quiet already has been removed 2022-02-18 17:26:36 want I noticed is that /sysroot/etc/apk/repositories is empty 2022-02-18 17:26:49 so it makes sense it cannot do anything beyond that i suppose 2022-02-18 17:27:19 minimal: I see what happens :) 2022-02-18 17:27:27 https://gitlab.alpinelinux.org/alpine/mkinitfs/-/blob/master/initramfs-init.in#L248 2022-02-18 17:30:41 what file(s) are you using to netboot? 2022-02-18 17:31:30 the files from the ordinary pi image 2022-02-18 17:32:00 does the ordinary RPI image support netboot? 2022-02-18 17:32:26 After providing the alpine_repo in cmdline.txt, it does ;) 2022-02-18 17:33:25 minimal: now we can obviously debate if it should have a fallback, if it doesn't, but it was already helpful to come to this conclusion. 2022-02-18 17:37:27 Anyone have any tips for creating a netboot img file? (not an ISO) netboot is distributed as a tar.gz with a boot dir inside it. 2022-02-18 17:38:08 earboxer: you mean what you need to pxeboot? 2022-02-18 17:38:50 I'm trying to use ventoy (to have multiple images on a USB stick) 2022-02-18 17:39:23 (and they have a bug where on my machine ISO 9660 ISOs don't load) 2022-02-18 17:39:35 earboxer: i can help you with netbooting, but no experience with usb images :) 2022-02-18 17:42:32 wew, I broke my alpine vm, apparently kmod needs zstd, and I somehow uninstalled zstd 2022-02-18 17:42:53 j`ey: welcome in my world :) 2022-02-18 17:43:23 j`ey: i typically fix that part by booting up with a working vm and restore by transfering it from one to the other disk 2022-02-18 17:43:47 if kmod required zstd.. I shouldnt have been able to uninstall it? 2022-02-18 17:45:23 kmod from package kmod or kmod as provided by Busybox? 2022-02-18 17:45:43 not sure 2022-02-18 17:46:27 the kmod package has a dep on zstd, Busybox does not (as its "kmod" doesn't currently support zstd) 2022-02-18 17:46:40 last thing I ran was 'apk del konsole' 2022-02-18 17:47:32 as far as I know, re: apk 2022-02-18 17:48:10 skinkie: so is it booting now that you provide alpine_repo? 2022-02-18 17:48:34 j`ey: apk would let you remove it from world, but it should still keep it installed as a dependency 2022-02-18 17:48:38 minimal: exactly. 2022-02-18 17:48:59 ikke: hm, well I'm not exactly sure what broke, I reinstalled konsole just in case, but it still fails to boot 2022-02-18 17:49:20 minimal: http://stefan.konink.de/booting.png 2022-02-18 17:50:00 in case anyone has seen this before https://i.imgur.com/2z8ygMm.png 2022-02-18 17:50:14 minimal: there is a raspberrypi specific catch21. The cmdline.txt has the path to modloop. How can I assure that modloop-rpi4 is taken, where as modloop-rpi is taken for arm. 2022-02-18 17:53:12 minimal: i think i have solution found for that part too :) cmdline can likely be a different file. Testing that :) 2022-02-18 17:54:38 skinkie: well for tftp, depending on your server setup, you can have different sets of files for each client based on their MAC address 2022-02-18 17:56:12 is there a way to list recently upgraded packages? 2022-02-18 17:56:19 j`ey: not yet 2022-02-18 17:56:28 apk3 has support for log files, apk2 not 2022-02-18 17:56:44 minimal: yes, don't you hate it tftp-root in dnsmasq cannot yet be used with a label? 2022-02-18 17:56:48 other than apk del konsole, I did do an update/upgrade 2022-02-18 17:56:55 so I'm guessing that could have broken things 2022-02-18 17:57:18 minimal: after this part works I am likely upgrade tftp-unique with masking options. 2022-02-18 17:57:28 like partial mac, or better: label 2022-02-18 17:58:12 i think it is called 'tag' in dnsmasq. so i want to have tftp-unique-root=tag 2022-02-18 18:01:04 skinkie: I have not used the tftp parts of dnsmasq for a while (some of dnsmasq's tftp functionality was implemented at my request back in the mid 2000s) 2022-02-18 18:02:15 ikke: any ideas about my issue here https://i.imgur.com/2z8ygMm.png 2022-02-18 18:02:28 maybe I need to regen my initramfs 2022-02-18 18:02:44 j`ey: can it be a partial upgrade? 2022-02-18 18:02:53 not intentionally 2022-02-18 18:03:06 what does apk version zstd-libs return? 2022-02-18 18:03:23 1.5.2r-0 2022-02-18 18:03:30 *-r0 2022-02-18 18:03:43 ok, so it's installed 2022-02-18 18:03:50 and apk info -P zstd-libs 2022-02-18 18:03:51 yeah 2022-02-18 18:04:15 so:libzstd.so.1=1.5.2 2022-02-18 18:04:31 minimal: i think tftp-unique was done by another dutch guy... so i'll pay it forward ;) 2022-02-18 18:04:53 j`ey: apk info -L zstd-libs 2022-02-18 18:05:29 usr/lib/libzstd.so.1 and 1.5.2 2022-02-18 18:05:46 huh 2022-02-18 18:06:09 Why would it say it cannot find libzstd.so.1 2022-02-18 18:06:25 but that's on my rootfs right 2022-02-18 18:06:29 it hasn't mounted that yet 2022-02-18 18:06:32 oh 2022-02-18 18:06:35 then you need to do that 2022-02-18 18:06:45 'do that' == regen initramfs? 2022-02-18 18:06:58 Can you boot a live usb? 2022-02-18 18:07:08 yeah im chrooted right now 2022-02-18 18:07:53 skinkie: dnsmasq 2.40 "Add --tftp-unique-root option. Suggestion from Dermot Bradley.". Thanks me :-) 2022-02-18 18:08:00 ok, so you might've fixed your rootfs already? 2022-02-18 18:08:17 s/Thanks/That's/ 2022-02-18 18:08:31 minimal: I guessed that from your previously github hints ;) 2022-02-18 18:09:16 ikke: i think my rootfs is fine, but it cant seem to mount it? 2022-02-18 18:09:26 'Mounting root failed' 2022-02-18 18:10:30 Im not sure how to regen the initramfs for a kernel than the one thats running 2022-02-18 18:11:09 j`ey: that is mkinitfs with the proper kernel version 2022-02-18 18:11:10 j`ey: mkinitramfs takes an option for the kernel to regen 2022-02-18 18:11:29 oh i missed it in the help, it's just a positional argument 2022-02-18 18:12:38 minimal: http://stefan.konink.de/pickyourbrain.png <- got a hint how to resolve it? This is my thing with an apkovl from a different architecture. 2022-02-18 18:12:39 nope, that didn't help either 2022-02-18 18:14:53 skinkie: I'm guessing the "missing" /sbin/init relates to the earlier busybox trigger errors as /sbin/init on Alpine is provided by Busybox 2022-02-18 18:15:33 minimal: i'm going to create an apkovl for this specific machine, and just compare the two 2022-02-18 18:15:56 looking at https://git.alpinelinux.org/aports/log/main nothing jumps out at me 2022-02-18 18:16:29 skinkie: apkovl from a different architecture? is the machine arch different that than of the apkovl? 2022-02-18 18:17:09 yes :) 2022-02-18 18:17:26 minimal: I wanted to see if alpine would just support that 2022-02-18 18:17:30 j`ey: in the chroot you do have bindmounts setup for /proc, /dev/, and /sys? 2022-02-18 18:17:51 minimal: I can go back into the chroot and do that 2022-02-18 18:17:52 skinkie: how can it? unless you add binfmt/qemu-user into the mix 2022-02-18 18:18:24 if apkovl only contains configurations, there shouldn't be any executable code in that apkov 2022-02-18 18:19:19 skinkie: what exactly does the apkvol contain? 2022-02-18 18:21:30 only my /etc changes 2022-02-18 18:22:01 ikke: in my initramfs, I dont have libzstd 2022-02-18 18:22:33 minimal: so I think the actual problem is etc/apk/arch 2022-02-18 18:22:38 I havent looked into mkinitfs yet 2022-02-18 18:24:03 skinkie: and does /etc/ includes things like /etc/apk/arch? 2022-02-18 18:24:35 skinkie: ah, was writing that whilst you were also ;-) 2022-02-18 18:24:49 minimal: in the apkovl there is indeed an etc/apk/arch 2022-02-18 18:25:08 which will contain the arch of the apkovl file 2022-02-18 18:25:23 minimal: so removing it, might just do the trick ;) 2022-02-18 18:25:50 I wonder if the initramfs' modprobe is supposed to be busybox 2022-02-18 18:27:22 j`ey: it used to be busybox's, then around Alpine 3.15 zstd was used for kernel modules and so kmod package became requirement, then soon after kernel modules were changed to use gzip compression instead 2022-02-18 18:28:08 the dep on kmod package remained, though I expected it to be removed as Busybox's kmod can handle gzip 2022-02-18 18:28:12 I guess mkinitfs then checks `ldd` or something to copy the libraries needed.. 2022-02-18 18:28:40 https://github.com/alpinelinux/mkinitfs/blob/a5f05c98f690d95374b69ae5405052b250305fdf/mkinitfs.in#L48 2022-02-18 18:29:04 j`ey: /etc/mkinitfs/features.d/base.files will list the kmod binary 2022-02-18 18:29:40 right, so it has to copy libzstd.so etc too 2022-02-18 18:29:58 then the script then works out the dependancies of anything going into the initramfs and adds them too 2022-02-18 18:30:13 or in my case, it doesnt :P 2022-02-18 18:32:14 so does your /etc/mkinitfs/features.d/base.files file have /sbin/modprobe listed in it? 2022-02-18 18:32:17 minimal: i still got the script exited errors... is there any way to debug it? 2022-02-18 18:34:00 minimal: yes 2022-02-18 18:34:17 skinkie: add "debug_init" to your cmdline, though it will scroll off the top of the screen 2022-02-18 18:34:57 j`ey: and is /sbin/modprobe (inside the chroot) a softlink to ../bin/kmod? 2022-02-18 18:35:00 Im confused now, when I drop into the emergency shell. thats my initramfs right? 2022-02-18 18:35:08 minimal: this happens inside a booted alpine, I it is certaily an aarch64 vs something else I even get Exec format error now. 2022-02-18 18:35:19 j`ey: yes, should be 2022-02-18 18:35:27 j`ey: yes 2022-02-18 18:35:29 it did not have chance to switch the root to the rootfs 2022-02-18 18:35:46 I ran cpio -t on the initrd-lts file, I could see libzstd in there 2022-02-18 18:35:55 but not when Im in the emergency shell 2022-02-18 18:36:14 j`ey: it needs to load the relevant kernel module for the rootfs (ext4, btrfs, etc) but cannot due to the kmod issue 2022-02-18 18:36:48 right 2022-02-18 18:37:24 inside the emergency shell, there is initrd.image, whats that? 2022-02-18 18:39:41 skinkie: are you *sure* there are no arch-specific files inside the apkvol file? 2022-02-18 18:40:06 minimal: I can reproduce it without that apkvol 2022-02-18 18:40:19 skinkie: reproduce it how? 2022-02-18 18:41:01 minimal: I created a custom repository 2022-02-18 18:41:16 it is signed etc. so i have it for x86_64 and aarch64 2022-02-18 18:42:06 It seems when I change the dl-cdn to mine, it seems to happen. But I'll try to reproduce it first. 2022-02-18 18:43:04 skinkie: not sure I follow, you are booting a x86_64 machine and pointing it to a "custom" aarch64 repo? 2022-02-18 18:44:05 minimal: I am netbooting a raspberry pi 2022-02-18 18:44:24 minimal: Ideally I would like to use the same apkovl 2022-02-18 18:44:42 ok, so you're pointing the RPI to a x86_64 arch repo? 2022-02-18 18:45:03 minimal: actually no... 2022-02-18 18:45:24 so I'm confused how you are reproducing the problem 2022-02-18 18:45:36 http://server/reponame is what i use 2022-02-18 18:45:48 but let me check first 2022-02-18 18:49:11 anyone any idea how come my initramfs has an 'initrd.image' file inside of it? 2022-02-18 18:50:44 j`ey: no idea. Before generating an initramfs you can run mkinitfs with the "-l" option to see what *would* be placed in the initramfs, its a dry-run option 2022-02-18 18:51:48 yeah, that prints libzstd, and cpio -t too 2022-02-18 18:58:39 oh jeez, I didnt mount boot in my chroot.... 2022-02-18 18:58:49 ouch 2022-02-18 18:59:06 well anyyywyay, I knew it was my initramfs! 2022-02-18 18:59:13 no idea how it got broken though 2022-02-18 19:02:57 minimal: the actual issue is also present when I take an apkovl created from that system. But I think that the actual issue is that my local repository serves/has cached the wrong architecture 2022-02-18 19:03:37 minimal: absolutely confirmed, same file sizes 2022-02-18 19:05:43 what would be the correct way to apk fetch a package for a different architecture? 2022-02-18 19:07:32 ok --arch 2022-02-18 19:07:46 --arch only works when creating a new root 2022-02-18 19:07:56 otherwise apk is hardcoded to fetch the current arch 2022-02-18 19:10:51 Hey! Can someone have a look at https://gitlab.alpinelinux.org/alpine/aports/-/merge_requests/30411 ? 2022-02-18 19:12:12 ty! 2022-02-18 19:24:30 skinkie: yeah the only time I've used "--arch" is for creating a base install that I'm going to chroot into 2022-02-19 03:12:58 minimal: The plan is to just boot from the ISO once in order to install to disk, but the setup-apkrepos step can't be done if the clock is wrong 2022-02-19 03:15:11 ktprograms: yeah its a problem. Either the efi-rtc should be compiled into the kernel or else it should be added to the initramfs and then the cmdline could use "modules=efi-rtc" for the initramfs' init to load it 2022-02-19 03:15:41 minimal: Agreed, but I think for now just starting hwclock is _good enough_ 2022-02-19 03:16:20 basically with arm there's no standardised RTC driver - unlike PCs where rtc_cmos is the only driver used for all (BIOS-based) PCs 2022-02-19 03:17:16 so I guess Arm servers can be assumed to have an EFI RTC but other non-server Arm devices either have no RTC or have various chipsets used for RTC 2022-02-19 03:17:35 minimal: Yes, I see how that could be a problem. 2022-02-19 03:17:52 The other solution would just to make setup-alpine run setup-ntp even if running in qemu 2022-02-19 03:18:30 with that you'll still see the "clock skew" warnings 2022-02-19 03:20:26 minimal: That's just a warning, right? If the time is correct it should be fine? 2022-02-19 03:20:47 its a warning every time the time jumps 2022-02-19 03:21:27 so openrc changes the time when it starts, then when hwclock changes (jumps) the time openrc will warning again 2022-02-19 03:22:08 if you look at the log you posted yesterday you can see then warning there plus notice several timestamp jumps 2022-02-19 03:22:43 minimal: Yes, so the fix would be to have rtc-efi in the ISO initramfs? 2022-02-19 03:24:35 if it was in the initramfs and if the cmdline used by the ISO also had rtc-efi in the modules= list that would trigger initramfs' init to load it before openrc is started and so openrc would not see the time jump (assuming your RTC time is generally accurate as then NTP would only have to make very small adjustments that openrc wouldn't complain about) 2022-02-19 03:25:51 the problem if rtc-efi was listed in the ISO's cmdline nmodule list is that it would then be loaded on all aarch64 machines using that ISO, some of which may not have a EFI-RTC 2022-02-19 03:26:05 minimal: Would that break anything, or would it just do nothing? 2022-02-19 03:26:18 perhaps a line could be added to the initramfs's init to only load if is UEFI is active 2022-02-19 03:27:28 so something like this could be added to initramfs-init: 2022-02-19 03:28:12 if [ "$(uname -m)" = "aarch" ] && [ -d /sys/firmware/efi ]; then modprobe efi-rtc; fi 2022-02-19 03:28:45 s/aarch/aarch64/ 2022-02-19 03:29:14 minimal: Makes sense, could you please help to make an MR for it? 2022-02-19 03:30:33 I can add it to my MR list :-) there would have to be another MR to add efi-rtc to a mkinitfs feature file also 2022-02-19 03:31:28 minimal: Shouldn't that be part of the same MR? It's fixing the same problem 2022-02-19 03:32:07 doh, same package mkinitfs, right - for some reason I was thinking there were changes to 2 different packages 2022-02-19 03:33:22 the other approach is to build a aarch64 kernel with the rtc-efi driver compiled in and test it on a non-UEFI aarch64 system and see if that causes or does not cause any problems 2022-02-19 03:34:16 its a different kernel than the one used for the RPIs (linux-lts vs linux-rpi) so would need to test on some other aarch64 machine 2022-02-19 03:34:48 minimal: I think I can configure QEMU to run as non-uefi 2022-02-19 03:35:07 At least there's a checkbox for it in UTM, haven't tested how/if it works 2022-02-19 03:36:23 yes QEMU in general can run either as BIOS or UEFI system, not sure about UTM 2022-02-19 03:37:11 minimal: Ok but since my host is a Mac it'll be a bit complex to make my own ISO. 2022-02-19 03:37:41 And I'll be busy soon so I may have to continue tomorrow 2022-02-19 03:39:15 actually I'm just looking at dmesg output from an Alpine UEFI aarch64 libvirt/Qemu VM I helped someone getting running yesterday and the rtc-efi module appears to activsted during kernel initialisation (i.e. before initramfs is loaded) 2022-02-19 03:39:45 this is kernel version 5.15.16-0 of Alpine linux-virt package 2022-02-19 03:39:53 which kernel version were you testing? 2022-02-19 03:42:49 minimal: Whatever's in the Alpine Standard 3.15 ISO (so 5.15.x, lts) 2022-02-19 03:43:16 ah, they were using Alpine Edge afaik 2022-02-19 03:44:41 (I believe they used their own Alpine mirror which is lagging behind Edge as currently linux-virt for aarch64 in Edge is 5.15.24-r0 2022-02-19 03:45:14 wonder if the kernel config is different for the RTC between Edge and 3.15, let me have a look 2022-02-19 03:48:45 hmm, its currently CONFIG_RTC_DRV_EFI=y in both. linux-virt in 3.15 repo is presently 5.15.16-r0 though it was 5.15.5 when Alpine 3.15 came out (and the ISO was made) 2022-02-19 03:49:19 If it's =y, why isn't it setting the system time at kernel init? 2022-02-19 03:52:16 I think that setting changed to =y AFTER the kernel used for Alpine 3.15.0 release 2022-02-19 03:52:41 so that if/when a v3.15.1 release comes out and ISOs are made then it would be present for you 2022-02-19 03:53:18 I don't see any change in git to enable that module so assume that it was enabling in the upstream linux config rather than by Alpine 2022-02-19 03:53:27 i'm pretty sure it's been =y for a really long time 2022-02-19 03:53:28 let me check 2022-02-19 03:54:32 marked as 2019 in blame 2022-02-19 03:54:39 for virt lts aarch64 2022-02-19 03:54:53 psykose: yeah, I'm not seeing a module for it in 3.13 or 3.12 for aarch64 2022-02-19 03:55:02 mhm 2022-02-19 03:55:14 2019 is like.. alpine 3.2 or something 2022-02-19 03:55:20 haha 2022-02-19 03:55:27 so then why is it appearing in ktprograms dmesg output later in the boot sequence 2022-02-19 03:55:39 that i can't tell you 2022-02-19 03:55:56 if its built in then it should appear in dmesg before the initramfs' init starts 2022-02-19 03:56:11 which is what I seeming in this out aarch64 UEFI server's dmesg 2022-02-19 03:56:36 ktprograms: can you post your dmesg output again? 2022-02-19 03:58:49 psykose: no matter, found your dmesg from yesterday - yes the rtc-efi is appearing very late in the dmesg, long after initramfs' init starts 2022-02-19 04:00:37 oops, found ktprograms dmesg 2022-02-19 04:01:35 this really does make sense as if the rtc is being read to set system clock then how does system clock end up back in Nov 2021 by the time openrc runs? 2022-02-19 04:02:33 s/does/doesn't/ 2022-02-19 04:06:32 s/does/doesn't/ is the story of my life 2022-02-19 04:13:05 ktprograms: is the init.d swclock service enabled or disabled when you boot? 2022-02-19 04:13:39 minimal: Disabled I think 2022-02-19 04:13:45 can you check? 2022-02-19 04:13:52 Also the bot seems to be down 2022-02-19 04:14:54 minimal: Doesn't show up in rc-status 2022-02-19 04:16:20 as swclock sets the system clock to the timestamp of the /sbin/openrc-run file 2022-02-19 04:22:40 minimal: Did you see my message (swclock doesn't show up in rc-status( 2022-02-19 04:23:11 I'm grasping at things that could be changing the clock.... 2022-02-19 04:28:34 ktprograms: looking at your dmesg and the logs again all I can assume it that the EFI RTC is so slow to set the system clock that this occurs AFTER openrc has started and seen the system is wrong (i.e 1970) and changed the time to something wrong (Nov 2021) but more "reasonable" and then the EFI change happen (which openrc warns about) and then you run hwclock which gives another time change 2022-02-19 04:29:17 minimal: I see. But why does "date" still show the Nov date even after efi-rtc has set the sytem time? 2022-02-19 04:30:42 from your dmesg yesterday: 2022-02-19 04:30:43 [ 0.189797] Alpine Init 3.6.0-r0 2022-02-19 04:30:58 that's the initramfs' init starting at 0.18secs 2022-02-19 04:31:31 [ 4.440568] Mounting boot media: ok. 2022-02-19 04:31:42 [ 4.556675] Installing packages to root filesystem: ok. 2022-02-19 04:32:01 [ 5.221239] rtc-efi rtc-efi.0: registered as rtc0 2022-02-19 04:32:01 [ 5.221336] rtc-efi rtc-efi.0: setting system clock to 2022-02-18T12:12:39 UTC (1645186359) 2022-02-19 04:34:14 there's no clear indication of when openrc starts in that sequence but I think it happens before the rtc-efi lines 2022-02-19 04:35:26 minimal: Right. But why doesn't rtc-efi change the output of "date"? 2022-02-19 04:36:05 the date command just reads from the system clock 2022-02-19 04:36:33 minimal: So is rtc-efi failing to set the system clock properly? 2022-02-19 04:36:35 rtc-efi does appear to have set the system clock but I assume something else is chnaging it afterwards 2022-02-19 04:36:48 minimal: Oh ok 2022-02-19 04:36:54 Any way to debug further? 2022-02-19 04:36:55 basically there is insufficient information to know what's going on 2022-02-19 04:38:25 when I investigating this sort of problem in the past (on a RPI with no RTC) I added debug lines (like "date") into various parts of openrc's shell scripts in /lib/rc/ to display the date/time at different stages. Unfortunately as you're booting off an ISO that can't be done 2022-02-19 04:38:48 unless you build/rebuild your own ISO 2022-02-19 04:41:10 minimal: I might do that, but I'm busy this weekend 2022-02-19 04:42:18 busy with something fun? :) 2022-02-19 04:42:33 psykose: kinda 2022-02-19 07:36:53 security wise, does transfering files with syncthing vs usb from android/mobile make any difference? 2022-02-19 07:37:51 i guess its hard to compare, but if there is some kinda of malware i guess it might be tranfered either way 2022-02-19 09:01:32 Being on an uptodate edge install, 'gcc main.c -lbsd' fails with 'cannot find -lmd: No such file or directory'. That command used to work under 3.15. Is this a bug? 2022-02-19 09:03:28 are you sure you have the latest libbsd-dev installed 2022-02-19 09:03:32 it should pull in libmd-dev too 2022-02-19 09:07:21 the reason -lbsd works at all is because for some reason there is an unversioned .so in `libbsd` proper, and something probably pulls libbsd in (quite common) 2022-02-19 09:07:22 psykose: that was the problem. Apparently something was changed in the build container. Thanks. 2022-02-19 09:07:32 if the .so didn't exist you would get 'cant find -lbsd' instead 2022-02-19 09:07:33 yea 2022-02-19 09:07:48 but yeah, just rely on the -dev, and all is fine 2022-02-19 09:09:06 I was just suprised because I did not change anything in my build job. I update my prerequisites. 2022-02-19 09:09:39 i would assume you were just missing the -dev the whole time, and it was a fluke that it worked 2022-02-19 09:09:56 yep 2022-02-19 09:09:59 and something changed that it didn't 2022-02-19 11:43:18 <\pub\bash0r> Hey! If I have suggestions for a community maintained package, how should I proceed? The package repo only lists the maintainers name, but no contact info. 2022-02-19 11:44:37 \pub\bash0r: The APKBUILD file has contact info. 2022-02-19 11:44:49 You could also create an issue or merge request 2022-02-19 11:45:30 <\pub\bash0r> I wasn't sure the tracker would also be about community packages. Just creating a PR seemed to be against community guidelines, since the wiki specifically says "Committers and others should confirm with the maintainer before making any changes of the package.". 2022-02-19 11:45:40 <\pub\bash0r> Then again, a PR is more or less a confirmation .... 2022-02-19 11:46:39 <\pub\bash0r> Ah, indeed, the APKBUILD contains the mail address. I could have thought of that :-/ Thanks! 2022-02-19 11:52:40 \pub\bash0r: You are free to make an MR as a suggestion, then the maintainer can confirm whether it's wanted or not 2022-02-19 11:52:54 though, ofcourse if it's something substantial, it might make sense to discuss it before 2022-02-19 11:53:29 <\pub\bash0r> Sounds reasonable. Thanks again :) 2022-02-19 16:26:54 hi, I became rusty with alpine in recent years. was there any easy way in apk to list packages that do not come from repositories currently set in /etc/apk/repositories? 2022-02-19 16:29:34 przemoc: apk info --orphaned I believe 2022-02-19 16:29:48 a/info/liar 2022-02-19 16:29:53 s/info/list 2022-02-19 16:32:47 thanks, ikke 2022-02-19 19:38:07 What's special about the xen-optimized version of alpine? Is it just the included packages, or are there kernel alterations? 2022-02-19 19:39:39 I want to try and do a network boot of alpine with the xen optimizations and I noticed that it uses the 'vanilla' versions of modloop, vmlinuz and initramfs 2022-02-19 19:40:43 I'm wondering if I'd do just as well to use the network boot version of alpine and just install the xen packages 2022-02-19 19:42:59 QDX45: here's one of the scripts that is used to create the Xen image 2022-02-19 19:43:39 seems to be only just ensuring xen-specific packages are installed 2022-02-19 19:44:56 there's also this: https://gitlab.alpinelinux.org/alpine/aports/-/blob/master/scripts/genapkovl-xen.sh but the reference to it in the 1st script is commented out 2022-02-19 19:48:21 Hi - i would like to install a community package. However, I would like to build it and verify it rather than trust the binary. 2022-02-19 19:48:48 Is there a wiki page for this specific case. The wiki shows how to build new packages rather than existing. 2022-02-19 19:49:29 https://wiki.alpinelinux.org/wiki/Abuild_and_Helpers drudru 2022-02-19 19:50:21 Thank you 2022-02-19 20:10:29 thanks minimal 2022-02-19 21:40:58 I have another question. I'm trying to create a package repository and I need to get the dependencies figured out so I can download all the packages I need. 2022-02-19 21:41:37 I see that I can use the .PKGINFO file inside apks to get something similar, but it doesn't give package names. 2022-02-19 21:41:56 Just individual .so files 2022-02-19 21:44:05 I see that I can use the APKBUILD file here to get a list of actual packages 2022-02-19 21:44:06 https://gitlab.alpinelinux.org/alpine/aports/-/blob/master/main/amavis/APKBUILD 2022-02-19 21:44:35 But I can't see how to get such a list from an .apk file or the APKINDEX.tar.gz file. 2022-02-19 21:45:34 You can ask apk which packages provide those so: packages 2022-02-19 21:48:13 I was hoping to be able to do it on a not-alpine os 2022-02-19 21:48:54 Then you would need to first make a list of provides per package 2022-02-19 21:49:06 A list of provides? 2022-02-19 21:49:39 I don't know what that means 2022-02-19 21:50:35 Does apk keep an internal list of each .so file and what package provides it? 2022-02-19 21:51:01 the apkindex tracks what package provides what so: and what package depends on what so: 2022-02-19 21:51:09 The provides field in the .PKGINFO file 2022-02-19 21:51:10 then it resolves stuff and gets you the 'correct' packages for that 2022-02-19 21:51:42 it doesn't give package names because there can be more than one of something, the dependency is not on a specific package 2022-02-19 21:52:04 unless you depends= on a specific package name, but even then, more than one thing can provides= a name 2022-02-19 21:52:15 provides = so:libz.so.1=1.2.11 2022-02-19 21:52:39 if you want a 'complete' local-built repository, you probably want buildrepo from lua-aports 2022-02-19 21:52:50 instead of dealing with manually checking dependencies 2022-02-19 21:53:47 oooooh. gotcha. so that's the p: and D: lines in the pkginfo. 2022-02-19 21:54:16 You can BTW use apk.static on none alpine systems 2022-02-19 21:54:46 I don't know what that is 2022-02-19 21:55:04 a statically linked apk 2022-02-19 21:55:07 apk-static 2022-02-19 21:55:33 sorry, what's a 'none alpine systems'. 2022-02-19 21:55:47 non-apline systems 2022-02-19 21:55:50 aka other distros 2022-02-19 21:55:57 gotcha 2022-02-19 21:56:29 There was one in my package manager, actually, but it could barely do anything. it mostly threw errors. 2022-02-19 21:57:13 I think I have what I need with the provides thing you mentioned. Looks like that + some text processing should do the trick. 2022-02-19 21:57:56 You mentioned that an .so file can be provided by multiple packages...is there any reason to not grab an arbitrary package containing the .so file needed? 2022-02-19 21:58:29 Or should I worry about different versions or something? 2022-02-19 22:00:16 I'll also look at buildrepo. 2022-02-19 22:00:21 Thanks for your help! 2022-02-19 22:38:59 ikke: BTW using apk.static recently on a non-Alpine system I had to create /etc/apk/ca.pem (by copying equivalent file from host OS) in order to use https repos. Wonder if apk should take an option to point to a CA file rather than hardcode an Alpine-only path 2022-02-19 22:44:12 Perhaps open an issue on. 2022-02-19 22:44:24 The apk-tools project 2022-02-19 22:48:52 yeah, I only had to work around this yesterday, am intending to raise an issue 2022-02-19 22:58:12 pretty sure alpine doesn't use /etc/apk/ca.pem 2022-02-19 22:59:16 https://gitlab.alpinelinux.org/alpine/apk-tools/-/blob/master/libfetch/common.c#L467 it uses SSL_CTX_set_default_verify_paths, which may not work on other distros with incompatible OPENSSLDIR 2022-02-19 22:59:53 afaik all distros have /etc/ssl/certs pointing to hashed ca directory though? 2022-02-19 23:48:14 Hello71: a "strings" on the apk.static binary shows it referencing /etc/apk/ca.pem 2022-02-19 23:50:02 and this https://gitlab.alpinelinux.org/alpine/apk-tools/-/blob/master/libfetch/Makefile#L4 2022-02-19 23:50:20 shows a refernce to $(CONFDIR)/ca.pem 2022-02-19 23:50:47 and this: https://gitlab.alpinelinux.org/alpine/apk-tools/-/blob/master/Makefile#L17 2022-02-19 23:50:56 shows CONFDIR = /etc/apk 2022-02-19 23:51:22 Hello71: this is not the first time your claled me out on IRC ridiculing something I'd said 2022-02-19 23:51:28 have you a problem with me? 2022-02-19 23:51:52 my impression is that the reverse is true 2022-02-19 23:52:48 the only problem i have is that you appear to knock thing I say 2022-02-20 00:01:39 typically only when you are aggressively wrong, which seems to happen disappointingly frequently 2022-02-20 00:03:23 I'm sure I've been wrong in the past, like most people, and I'll happly admit when I'm wrong, but I don't believe I've ever been aggressively wrong 2022-02-20 00:06:03 i mean it literally uses that path https://gitlab.alpinelinux.org/alpine/apk-tools/-/blob/master/libfetch/common.c#L449 , but if it is wrong and finds nothing it would fall back to the ssl_ctx_.. thing 2022-02-20 00:06:07 so you are both correct 2022-02-20 00:06:19 not sure what causes the actual issue on notalpine, but there is nothing to fight about here 2022-02-20 00:06:37 and an explicit arg for it is always nice 2022-02-20 00:08:44 an argument could be made that apk uses that path, but alpine doesn't. if merely referencing a path constitutes using it then alpine uses /usr/local/bin. but this is not a conventional or useful definition 2022-02-20 00:09:55 as i said, different distros use different locations for certificates. the most plausible explanation is that this issue is "caused" by statically linking openssl, not anything related to apk specifically. 2022-02-20 00:11:58 time to dump the huge strace log 2022-02-20 05:54:21 hi 2022-02-20 05:55:42 hello71: the /etc/apk/ca.pem file is used by apk-tools, as a certificate authority bundle of last resort. there is nothing “aggressively wrong” about that assertion. 2022-02-20 05:56:01 the reason it doesn't work on debian, is because debian only has /etc/ssl 2022-02-20 05:56:13 we set the default dir to /etc/ssl1.1 and have a symlink on openssl 2022-02-20 05:56:22 so the static version, linked into apk-tools-static, only checks there 2022-02-20 05:56:28 yes because reasons 2022-02-20 05:56:30 this sym does not exist on some other distros, so apk-tools fails 2022-02-20 05:56:30 yes 2022-02-20 05:56:34 it was done during the openssl3 switch 2022-02-20 05:56:50 https://gitlab.alpinelinux.org/alpine/aports/-/commit/832bbeb063dbb7ef6b6dd6487a9ed26f32666977 2022-02-20 05:57:11 so now the openssl .a/.so only checks there.. which doesn't work 2022-02-20 05:57:12 yes by hello71:) 2022-02-20 05:57:16 not sure if it's possible to amend 2022-02-20 05:57:18 of course :) 2022-02-20 05:57:29 we should probably clean up this situation 2022-02-20 05:57:46 probably, the issue from the linked issue was openssl1.1 was reading 3.0 config files 2022-02-20 05:57:47 but i’m already spent for this weekend 2022-02-20 05:58:12 i went to visit my grandmother for her birthday, and that was A Thing 2022-02-20 05:58:12 i guess this just happens when you run two openssls like a boss 2022-02-20 05:58:48 it was 2022-02-20 14:54:05 I'm making a rough copy of my alpine setup so far, since I'm going to be wiping this machine soon. I copied etc/apk/world and my runlevels + rc-status, is there anything else I might have missed? 2022-02-20 14:56:35 why don't you copy all of /etc 2022-02-20 14:56:38 and maybe /var too 2022-02-20 17:36:48 aw, I wish QDX45 were still around. I got over the hurdles of what they are attempting a week ago 2022-02-20 22:42:49 Hi, I'm struggling to install Alpine with an encrypted /boot using the LVM_on_LUKS guide on the wiki. When booting, even if I give grub a correct passphrase it tells me it was an invalid passphrase. Could someone point me in the right direction? 2022-02-20 22:47:59 sorry if someone replied, I had to restart my client. 2022-02-20 22:58:46 not sure if grub in alpine 3.15 supports luks2 2022-02-20 22:59:52 The guide *did* warn it was an experimental method... 2022-02-20 23:00:09 yes, the 3.15 one did not have the fixes backported 2022-02-20 23:00:14 i forget if even edge grub was fixed 2022-02-20 23:01:06 The ArchLinux AUR seems to have a fix for it. Would it be ill advised to figure out away to install their version of grub to see if it works? 2022-02-20 23:01:56 it's compiled against glibc and probably won't work at all 2022-02-20 23:04:40 AUR is PKGBUILDs only 2022-02-20 23:04:47 ah 2022-02-20 23:05:22 hm 2022-02-20 23:05:26 which fix are you referring to 2022-02-20 23:05:35 it is normally easier to convert a package recipe than a compiled package 2022-02-20 23:06:24 although alpine has pacman so theoretically you would makepkg -d 2022-02-20 23:09:05 psykose: This fix -- https://aur.archlinux.org/packages/grub-improved-luks2-git 2022-02-20 23:09:33 ah 2022-02-20 23:09:36 that is a lot of patches 2022-02-20 23:14:37 Hello71: I'm kinda curious to see if that can work 2022-02-20 23:21:35 you can try 2022-02-20 23:21:48 Before I do all that, would downloading the grub.apk from edge and installing it locally do anything weird? Just to see if those fixes did happen to be implemented there. 2022-02-20 23:24:04 you can upgrade the whole system to edge, should be fine 2022-02-20 23:24:14 from the manually-chroot-into-disk environment 2022-02-20 23:24:37 but yes, you can also install only the edge grub 2022-02-20 23:25:01 should be ok, no guarantees from me though 2022-02-20 23:25:21 copy that. 2022-02-20 23:29:24 Didn't work :p 2022-02-21 00:59:24 Quick update, I followed the part of the wiki that recommends LUKS1 and got everything to boot. Thank y'all for your help! 2022-02-21 01:01:04 :) 2022-02-21 01:01:15 i never got luks2 to work either, but luks1 has always worked for me for 5 years 2022-02-21 01:01:49 next time i try setting something up i will try it sans grub at all, probably easier that way 2022-02-21 01:38:39 imo there's fairly little reason to use luks for /boot 2022-02-21 01:40:19 why's that 2022-02-21 03:19:09 at most it hides the kernel version and initramfs and command line arguments, assuming luks works perfectly 2022-02-21 03:20:25 there's some vague theory about making it harder to install a rootkit, but i reckon any attacker putting in the effort to make a linux bootkit can also make it work on grub+luks+linux 2022-02-21 03:21:16 secure boot might actually be some hindrance; even if it's not terribly secure, it's at least some additional per-target work 2022-02-21 03:25:16 psykose: luks2 boot works but only with pbkdf2 which has to be explicitly stated when creatink the luks2 container as it defaults to argon2 2022-02-21 03:25:33 s/creatink/creating 2022-02-21 03:25:33 atka meant to say: psykose: luks2 boot works but only with pbkdf2 which has to be explicitly stated when creating the luks2 container as it defaults to argon2 2022-02-21 03:26:12 yes, i think that is what the argon patches in that aur package are for 2022-02-21 03:27:02 I haven't looked into it for a couple months but argon2 support should be in grub anytime as the patches were accepted for the next release 2022-02-21 03:27:49 so in a year 2022-02-21 03:27:53 :p 2022-02-21 03:28:32 yeah, I'm not sure on their release cadence I think they were planning on q1 2022 2022-02-21 03:28:47 but like I said, I haven't looked into it for a bit 2022-02-22 04:32:18 psykose: After the linux-lts bump (and GCC), I'm now no longer able to access my microSDXC from my regular user, only root, while this was resolved under linux-lts 5.15.22 - 5.15.24-r0 2022-02-22 04:32:45 Something else that got pulled in during the GCC bump broke this, or kernel you think? 2022-02-22 04:34:43 tumbler doesn't work right, takes multiple minutes to open thunar, etc. This happened a month or so ago to me and I couldn't figure out any resolution 2022-02-22 05:31:33 that doesn't sound related 2022-02-22 05:31:36 nothing got changed in it 2022-02-22 05:31:49 it's just a rebuild against gcc, which now has like... 40 more commits in it 2022-02-22 05:31:50 or something 2022-02-22 05:32:06 there's really not much difference in anything 2022-02-22 10:44:48 some missing optimizer flag? 2022-02-22 12:08:26 Saijin_Naib: right, i dont think it should be related to a kernel change, kernel does not set the permissions afaik. 2022-02-22 17:13:19 psykose: clandmeter yep, back-level kernel doesn't help. It's something else that got updated or rebuilt during the same time... No idea how to track this down again 2022-02-22 22:04:31 hmm.. maybve GVFS? It was updated/rebuilt around the same time as 5.15.24-r2 got pushed and I upgraded everything 2022-02-22 22:04:36 I get this on the cmd when I launch thunar:... (full message at https://matrix.org/_matrix/media/r0/download/matrix.org/ppnzmtdzJLhxSsPJmYItrZLd) 2022-02-23 00:20:31 ACTION sent a code block: https://matrix.org/_matrix/media/r0/download/matrix.org/VkKokSJrNtdObmTPkkOKoyMg 2022-02-23 00:21:07 I'm assuming that since the newapkbuild -r template uses the --locked flag, it is advised that I do not follow the console message and change that to --offline, right? 2022-02-23 00:22:36 ACTION sent a code block: https://matrix.org/_matrix/media/r0/download/matrix.org/jvtkTyefvnVopytEyWBDSDqj 2022-02-23 02:09:57 just installed alpine, having some trouble with apk. Every time I try to install a program with apk add, it gives me an error message: 2022-02-23 02:10:40 ERROR: unable to select packages: 2022-02-23 02:13:33 util-linux-misc-2.37.4-r0: conflicts: agetty-2.37.4-r0[cmd:agetty=2.37.4-r0] fstrim-agetty-2..37.4-r0[cmd:fstrim=2..37.4-r0] losetup-2.37.4-r0[cmd losetup=2.37.4-r0] 2022-02-23 02:15:58 losetup-2.37.4-r0 conflicts util-linux-misc-2.37.4-r0[cmd;losetup=2.37.4-r0] fstrim2.37.4-r0: conflicts: util-linux-misc2.37.4-r0[cmd:fstrim=2.37.4-r0] agetty-2.37.4-r0: conflicts: util-linux-misc-2.37.4-r0[cmd:agetty=2.37.4-r0] 2022-02-23 02:16:49 if it helps, this seems to have broken after I ran setup-xorg-base 2022-02-23 02:17:27 what do you have in /etc/apk/world 2022-02-23 02:17:38 and also /etc/apk/repositories 2022-02-23 02:18:18 cat /etc/apk/world | curl -F 'f:1=<-' ix.io 2022-02-23 02:49:00 i don't have curl 2022-02-23 02:51:54 in /etc/apk/world I have acct alpine-base blkid btrfs ca-certificates cfdisk chrony cryptsetup device-mapper dialog doas e2fsprogs eudev gptfdisk haveged linux-lts lvm2 mesa nano openssh openssl sgdisk st syslinux util-linux vim wget wireless-tools wpa_supplicant xf86-input-libinput xorg-server 2022-02-23 02:55:16 i deleted util-linux from that file and now it seems like it might be working 2022-02-23 02:56:03 i think that might have been a bad idea imma be honest i dont really know what i'm doing 2022-02-23 03:00:28 ok i think that fixed it not sure what i did 2022-02-23 03:41:06 now i can't start x as anything but root 2022-02-23 03:41:23 here's the log file http://ix.io/3QsC 2022-02-23 03:55:22 you should probably just start over 2022-02-23 03:56:25 noooooo 2022-02-23 03:56:28 took me so long 2022-02-23 03:59:18 well there's no telling what you did and if you broke apk at the beginning and are having issues with xorg right off the bat you're probably just asking for more problems 2022-02-23 03:59:30 did you follow a guide or something? first time with alpine? 2022-02-23 04:02:46 i mostly followed the wiki 2022-02-23 04:02:55 yes it's my first time with alpine 2022-02-23 04:04:12 I can probably give you some pointers if you can describe what you're wanting to acheive 2022-02-23 04:05:38 i just want to be able to launch bspwm really, to be honest this is probably not alpine specific so this is probably the wrong place to ask 2022-02-23 04:35:31 sp0rk: it shouldn't be an issue really, but the apk issues from before leads me to believe something else is going on 2022-02-23 04:37:19 guess I'll start from scratch then 2022-02-23 04:40:46 after you run through the installer try setup-xorg-base and then add bspwm 2022-02-23 05:06:12 apk broke in the same way again 2022-02-23 05:06:25 this time i'm sure i didn't do anyting weird, no idea why that happened 2022-02-23 05:10:31 I have to run right now, maybe post your .ash_history and link what guide/wiki you're using. I honestly can't say what's going on without more info, been using alpine since 2015 and never had an issue with apk 2022-02-23 05:22:29 yep it's still not working, i guess i'll give it another go tomorrow thanks for the help 2022-02-23 07:49:40 Hi. Does hddtemp exist in Alpine? I can only find collectd-hddtemp-5.12.0-r5 2022-02-23 07:53:16 I've just updated my alpine and got issue like this: https://gitlab.alpinelinux.org/alpine/aports/-/issues/9124 2022-02-23 07:53:41 And, like in the issue, af_packet module is missing 2022-02-23 07:57:31 so you dont have /lib/modules//net/packet/af_packet.ko.gz ? 2022-02-23 07:59:48 No, and I don't have the whole net/ directory tree 2022-02-23 08:00:09 what kernel are you using? 2022-02-23 08:00:15 But I have /lib/modules//kernel/net/packet/af_packet.ko.gz 2022-02-23 08:00:25 5.15.16-0-virt 2022-02-23 08:01:06 ok so you do have the module 2022-02-23 08:01:24 (I missed the /kernel/ part) 2022-02-23 08:01:46 it is present in /kernel/ 2022-02-23 08:01:56 thats the right place 2022-02-23 08:02:02 modprobe doesn't see it 2022-02-23 08:02:59 can you paste the command/output? 2022-02-23 08:03:06 Ah, somewhy I got 5.15.15 kernel and 5.15.16 module tree 2022-02-23 08:13:00 I've booted from chroot and reinstalled linux-virt 2022-02-23 08:13:36 package manager reports that it is 5.15.16, but the kernel reports itself as 5.15.15 2022-02-23 08:42:16 not having experience with this in Alpine, don'y you just have to reboot? 2022-02-23 08:43:29 that was poorly said. I meant "I don't have experience with this in Alpine. Don't you just have to reboot?" 2022-02-23 08:43:50 Actual reason was as follows: I forgot to mount /boot, so new kernel image got into wrong place, while system continued booting with old kernel 2022-02-23 09:10:54 Hello, quick question about cups. I have installed it with all the packages descibed into the wiki page but I have no answer from http://localhost:631/ 2022-02-23 09:14:07 You need to start the service, apk does not do that 2022-02-23 09:44:40 I did 2022-02-23 09:45:09 but no change 2022-02-23 09:53:23 try http://127.0.0.1:631/ 2022-02-23 09:57:08 unfortunately no change: The connection has timed out 2022-02-23 09:57:24 should I install a web server or it is embedde into cups ? 2022-02-23 09:57:44 no, it should work 2022-02-23 09:57:47 ps -aux | grep cups 2022-02-23 09:58:09 26460 root 0:00 /usr/sbin/cupsd -f -c /etc/cups/cupsd.conf 2022-02-23 09:58:31 it may be a pb of rights ? 2022-02-23 09:58:42 grep Listen /etc/cups/cupsd.conf 2022-02-23 09:58:57 Listen localhost:631 2022-02-23 09:58:59 Listen /run/cups/cups.sock 2022-02-23 09:59:07 (thanks btw ;) ) 2022-02-23 09:59:38 np 2022-02-23 09:59:47 look into /var/log/cups/error_log 2022-02-23 09:59:54 maybe something there 2022-02-23 10:00:20 Unable to open listen socket for address [v1.::1]:631 - Address not available 2022-02-23 10:01:06 Unable to open listen socket for address 127.0.0.1:631 - Address not available 2022-02-23 10:02:36 that means something else is already listening there 2022-02-23 10:02:43 probably a duplicate cups that doesn't work 2022-02-23 10:07:38 stop the cups daemon 2022-02-23 10:07:54 and check ps for remaining cups 2022-02-23 10:08:15 if daemon is stopped, and cupsd is still in ps, kill it 2022-02-23 10:13:43 daemon stopped and cups is killed as well 2022-02-23 10:14:03 is it possibale that somthg else is using that port...? 2022-02-23 10:14:06 stopping the daemon killed the cups process, or it was you? 2022-02-23 10:14:39 check whether it is bound: netstat -an | grep -e tcp -e udp 2022-02-23 10:15:43 stopping the daemon kill the process 2022-02-23 10:17:06 oh 2022-02-23 10:18:39 if the process is launched or not: there is no port 631 used with `netstat -an` 2022-02-23 10:25:39 you may check this: https://www.toptip.ca/2010/02/linux-eaddrnotavail-address-not.html 2022-02-23 11:02:44 Hello. In what package can I find kexec? (mentioned in https://wiki.alpinelinux.org/wiki/Custom_Kernel ). apk search kexec gives no results 2022-02-23 11:03:32 oh kexec-tools is only in edge 2022-02-23 11:03:34 https://pkgs.alpinelinux.org/packages?name=*kexec*&branch=edge 2022-02-23 11:14:29 I am guessing runnint kexec -e isn't a good idea because it won't cleanly shutdown services and mount points. Is there a openrc script that can handle this nicely? 2022-02-23 11:17:53 you could add the kexec command to /etc/inittab as ctrlaltdel 2022-02-23 11:18:15 or just 'openrc shutdown && kexec ...' 2022-02-23 11:18:26 but i'm not sure how well it would work 2022-02-23 11:19:53 didn't work well https://dpaste.com/7MMDTJ874 2022-02-23 11:20:23 network died, but I could use the system normally through ikvm console. 2022-02-23 11:20:36 no dmesg output at all 2022-02-23 11:20:54 # CONFIG_KEXEC is not set 2022-02-23 11:22:17 ahh yes I see"syscall kexec_file_load not available." 2022-02-23 11:23:52 would have been cool. saves the long boot times on the server 2022-02-23 11:36:27 aron: indeed cat /proc/sys/net/ipv4/ip_local_port_range return `32768 60999` 2022-02-23 11:36:43 does it mean that by default I can bind port 631 ? 2022-02-23 11:38:08 can´t* 2022-02-23 11:39:37 local port range is what the kernel allocates for new tcp connections 2022-02-23 11:39:59 I'm trying to reconstruct your scenario 2022-02-23 11:40:22 but something is definitely not right 2022-02-23 11:41:20 it seems cupsd can't bind those socket, but they are not bound by other process 2022-02-23 11:41:52 I thought of permission problem, but no, it's running under root 2022-02-23 11:42:41 how many interfaces you have in 'ifconfig'? 2022-02-23 11:43:00 only 1 2022-02-23 11:43:05 my wan interface 2022-02-23 11:43:24 ahhh 2022-02-23 11:43:28 that could be a problem 2022-02-23 11:43:34 there is no "lo"? 2022-02-23 11:43:39 (local loopback) 2022-02-23 11:44:11 Ohhhh... you are right no lo 2022-02-23 11:45:04 add these 2 lines to /etc/network/interfaces 2022-02-23 11:45:07 auto lo 2022-02-23 11:45:07 iface lo inet loopback 2022-02-23 11:45:37 how did they gone? 2022-02-23 11:45:46 should be there by default 2022-02-23 11:47:21 aron: nice find! 2022-02-23 11:47:43 yes definitly !!! 2022-02-23 11:47:50 I can now ping myself ! 2022-02-23 11:48:43 and of course.... 2022-02-23 11:48:46 CUPS works 2022-02-23 11:49:06 :) 2022-02-23 11:50:31 awesome 2022-02-23 11:50:45 ACTION give an ovation to aron ! 2022-02-23 11:50:54 impressive investigation ! 2022-02-23 11:50:59 Big thanks 2022-02-23 11:51:04 yw 2022-02-23 11:51:06 :) 2022-02-23 12:00:42 I have installed alpine from usb and since the beggining I have installed iwd for my wan connection so I never thought about the interfaces file. I would have imagine that it was created by default but it was not even created 2022-02-23 12:20:51 weird 2022-02-23 12:25:42 Hi again. I just noticed that my alpine does not remount rootfs with the full mount options from fstab. What should I do to make this happen? 2022-02-23 12:26:44 Forza: I personally use UUID in fstab 2022-02-23 12:28:26 aron: I do 2022-02-23 12:28:48 fstab: UUID=558ec0b4-869e-4f8e-a143-258d4d380847 / btrfs noatime,compress=zstd:1,ssd,subvol=/volume/root 0 0 2022-02-23 12:29:06 but mount output is "/dev/sdb3 on / type btrfs (rw,relatime,ssd,space_cache=v2,subvolid=256,subvol=/volume/root)" 2022-02-23 12:29:14 so compress and noatime is missing 2022-02-23 12:30:37 cat /proc/cmdline "BOOT_IMAGE=/volume/root/boot/vmlinuz-lts root=UUID=558ec0b4-869e-4f8e-a143-258d4d380847 ro rootflags=subvol=volume/root modules=sd-mod,usb-storage,btrfs quiet rootfstype=btrfs" 2022-02-23 12:32:53 I suppose I could edit /etc/default/grub to add the options, but this seems wrong, doesn't it? 2022-02-23 12:33:48 Hi, is there a standard tftp server i should use with alpine? on packages.alpinelinux.org/packages?name=ftp&branch=edge actually doesn't return anything ^^ 2022-02-23 12:34:30 mfg: that just means there's no package called ftp 2022-02-23 12:35:42 i see, does the search support regex then? 2022-02-23 12:35:45 yeah 2022-02-23 12:36:02 there's tftp-hpa, I don't know if that's what you're looking for 2022-02-23 12:37:13 alright, thank you :-) 2022-02-23 12:41:30 Forza: this is during a very early stage, right? 2022-02-23 12:48:35 aron: no? 2022-02-23 12:49:08 I expected init.d/localmount or something similar to remount / with the mount options from fstab. 2022-02-23 12:49:45 on my system rootfs is mounted twice during boot, first in ro, then remount rw ;) 2022-02-23 12:49:53 yes 2022-02-23 12:50:15 kernel mounts it as ro, then init remounts it to rw. but for some reason the compress and noatime flags aren't applied 2022-02-23 12:51:00 at the ro mount? 2022-02-23 12:51:51 localmount just calls mount -a, i noticed mine has the options applied but for me /bin/mount is from util-linux-misc and not the busybox one 2022-02-23 12:51:54 no after when init is mounting it to rw 2022-02-23 12:52:00 if you install util-linux-misc it might just magically fix it 2022-02-23 12:52:04 ah 2022-02-23 12:52:08 and i have no idea why the busybox one does not respect options properly 2022-02-23 12:52:11 good point psykose 2022-02-23 12:52:18 i think this might be the cause in general, as i have seen this 'error' like 10 times now 2022-02-23 12:52:19 in this channel 2022-02-23 12:52:42 so give the util-linux thing a try first, and then we can go see why busybox is dumb :) 2022-02-23 12:53:03 (and make sure apk info -W /bin/mount is not util-linux already) 2022-02-23 12:53:06 nice one 2022-02-23 12:53:17 I always install util-linux 2022-02-23 12:53:20 bad habit? :P 2022-02-23 12:53:23 maybe 2022-02-23 12:53:27 i have -misc and -login 2022-02-23 12:53:43 the main package is empty now as we split it up 2022-02-23 12:54:49 /bin/mount symlink target is owned by busybox-suid-1.34.1-r4 2022-02-23 12:56:46 Trying with utili-linux-misc now 2022-02-23 12:57:53 is there a linux-lts-dev package with debug info (non-stripped kernel)? 2022-02-23 12:58:38 util-linux-misc did not help 2022-02-23 13:01:17 I'll edit /etc/grub.d/10_linux instead 2022-02-23 13:01:23 rip 2022-02-23 13:02:38 Problem is that might get overridden by and apk upgrade? 2022-02-23 13:02:50 probably 2022-02-23 13:02:56 but you can make a custom new file 2022-02-23 13:03:06 well, it wouldn't be overriden 2022-02-23 13:03:12 just make a 10_linux.apk-new 2022-02-23 13:03:20 but either way you should make your own 2022-02-23 13:05:16 probably best 2022-02-23 13:05:18 thanks 2022-02-23 13:05:33 I do think this is a slight bug in localmount though? 2022-02-23 13:06:01 you can see what it runs 2022-02-23 13:06:08 idk, i would assume `mount -a` would follow options 2022-02-23 13:06:21 it doesn't seem to 2022-02-23 13:07:06 mount -avf : mount: according to /proc/mounts, /dev/sdb3 is already mounted on / 2022-02-23 13:11:32 sad 2022-02-23 13:11:34 :( 2022-02-23 13:12:02 then i assume `mount -o remount -a` would work 2022-02-23 13:12:10 not sure if that works for things not mounted at all yet 2022-02-23 13:12:17 if it does, i guess localmount is 'easily' amended 2022-02-23 13:12:38 has anyone successfully replaced openrc/bb init with s6? 2022-02-23 13:12:50 by hand, multiple people 2022-02-23 13:12:58 i forget who though 2022-02-23 13:13:07 and would be willing to provide instructions :P 2022-02-23 13:13:17 you can ask in the usual s6 channel 2022-02-23 13:13:34 #s6 2022-02-23 13:13:52 it's easier to get the hang of it by starting with containers or something 2022-02-23 13:14:00 than trying to hotpatch an alpine install 2022-02-23 13:14:19 I'm trying to replace init first 2022-02-23 13:14:55 mount -a -o remount worked 2022-02-23 13:15:03 at least with util-linux-misc installed 2022-02-23 13:15:53 but not with busybox mount 2022-02-23 13:16:09 i don't think just replacing init is very easy 2022-02-23 13:17:02 which init do you want to install? 2022-02-23 13:17:10 s6-l-i probably 2022-02-23 13:17:18 skarnet.org lists just 3 steps for that :P 2022-02-23 13:18:52 panekj: well write it up if you get it working :P 2022-02-23 13:18:53 These steps will allow you to run s6-svscan as PID 1, but you need also to prepare service directories for it 2022-02-23 13:19:08 j`ey: i intend to do so 2022-02-23 13:20:17 Ermine: from the content of rc.init I assumed I can just start openrc first 2022-02-23 13:20:46 Should I create an issue on the bugtracker for localmount and btrfs mount options? 2022-02-23 13:21:35 yes 2022-02-23 13:21:45 i have no idea why my xfs ones are respected 2022-02-23 13:21:46 panekj: probably you can do like this, but in this case openrc will launch services unsupervised by s6 and you will have no benefit from using s6-l-i 2022-02-23 13:22:08 yes, but you at least replaced one little part 2022-02-23 13:22:12 and things continued to work :) 2022-02-23 13:22:14 then you continue 2022-02-23 13:22:21 step by step 2022-02-23 13:23:03 yes 2022-02-23 13:23:04 exactly what psykose said 2022-02-23 13:23:44 i do recommend chatting about it in #s6 if you do anything though, then you can get some technical answers to whatever questions you have along the way 2022-02-23 13:30:43 hmm all mounts are missing the noatime and compress mount options: https://dpaste.com/GHDP2WE5B 2022-02-23 13:33:24 3 of them are noatime 2022-02-23 13:43:25 yes youre right. Odd this 2022-02-23 13:47:44 I did a report https://gitlab.alpinelinux.org/alpine/aports/-/issues/13544 2022-02-23 14:36:10 How can I compile a kernel with debuginfo ? 2022-02-23 14:36:17 I'm trying to setup syslog to use remote logging, so on the destination machine I've installed syslog-ng (because busybox syslogd does not seem to support that) 2022-02-23 14:36:29 and on the client I've changed /etc/conf.d/syslog to add -R 192.168.1.221 2022-02-23 14:36:43 but nothing happens on the remote 2022-02-23 14:39:05 I think I need. CONFIG_DEBUG_INFO. Trying 2022-02-23 14:41:15 markand: I have a syslong-ng server accepting network trafic. I have the following lines: source network { syslog(transport(tcp) port(514)); syslog(transport(udp) port(514)); }; destination network { file("/var/log/remote/syslog.log"); }; log { source(network); destination(network); flags(final);}; 2022-02-23 14:43:19 markand: and on your local machine do syslogd -R $REMOTE_LOGGING_SERIVCE:$PORT -L -O /var/log/messages 2022-02-23 14:46:57 hmmm I've switched to sysklogd on both machines, much simpler 2022-02-23 14:47:13 client : echo "*.* @192.168.1.221" > /etc/syslog.conf 2022-02-23 14:47:21 remote : add -r in /etc/conf.d/sysklogd 2022-02-23 14:47:22 :-) 2022-02-23 14:51:34 unfortunately it just seems that you can't filter a hostname on the remote (to write messages into a different file for example) 2022-02-23 14:51:42 maybe a little bit too minimalist :P 2022-02-23 14:51:58 Ah yea if you have several senders. 2022-02-23 14:52:10 Maybe several daemons on different IP ports? 2022-02-23 14:59:08 Hello :) I am having trouble while doing a sys installation of Alpine 3.15, after confirming that the disk should be erased, I am told that "partition X contains a Y signature" for several partitions 2022-02-23 14:59:36 Then it says "The file /dev/sdaX does not exist and no size was specified" 2022-02-23 14:59:39 and mounting fails 2022-02-23 15:00:51 I tried to understand what is going on in the install script, is it possible that sfdisk is failing because it detects existing signatures from a previous alpine installation? 2022-02-23 15:01:04 Should this not be expected and handled properly? 2022-02-23 15:01:30 Any help is appreciated since I am doing all this for writing a guide :) 2022-02-23 16:34:03 Hi. I installed a package manually from an .apk file. How can I revert this by installing the one from the official repo? 2022-02-23 16:34:26 I tried apk fix -r but it gives me [APK unavailable, skipped] 2022-02-23 16:35:19 apk add -u 2022-02-23 16:35:27 I use this 2022-02-23 16:35:37 I think it works 2022-02-23 16:36:04 nope. =( 2022-02-23 16:36:19 OK: 2299 MiB in 324 packages 2022-02-23 16:36:30 is there a version difference? 2022-02-23 16:36:59 I just had a package like yours 2022-02-23 16:37:04 right now 2022-02-23 16:37:12 No. same version.I had compiled it manually 2022-02-23 16:37:19 linux-lts package 2022-02-23 16:37:25 then nothing needs to be done 2022-02-23 16:37:32 it should be okay 2022-02-23 16:38:07 (for mine apk add -u, I got this: "(1/1) Upgrading rsvg-zoompan (1.1.0-r0 -> 1.1.0-r1)") 2022-02-23 16:38:13 Well what is odd is the huge difference in size of the initramfs-lts 2022-02-23 16:38:39 115.9M vs 14.2M 2022-02-23 16:38:48 But I did compile with debug_info enabled 2022-02-23 16:39:43 then apk del :) 2022-02-23 16:40:00 hmm scary with the kernel :D 2022-02-23 16:40:08 but yea. that's an option 2022-02-23 16:40:35 I noticed my update-grub script is bonkers now. But it could be due to other reasons and I just noticed 2022-02-23 16:40:46 apk upgrade -a 2022-02-23 16:40:55 It adds both my RAID1 members and forgets to add the initrd lines to grub.cfg 2022-02-23 16:41:10 I think the problem is the same version number 2022-02-23 16:42:16 aron: you might be right 2022-02-23 16:46:57 directly: apk fix linux-lts ? 2022-02-23 16:47:00 maybe 2022-02-23 16:47:47 apk fix --no-cache linux-lts 2022-02-23 16:47:53 seems to work with a random package 2022-02-23 16:54:14 Thanks. Ill try again tomorrow when I am back at work =) 2022-02-23 17:22:57 bananicorn: I saw something similar some days ago, it seems a bug in the script 2022-02-23 17:39:21 https://www.mail-archive.com/bug-grub@gnu.org/msg15384.html 2022-02-23 17:39:48 I got this bug so I don't think it's my custom kernel build 2022-02-23 17:44:34 I have a raid1 machine and grub-mkconfkg adds both sda3 and sdb3 with a line breake 2022-02-23 17:44:47 And it omits initrd completely 2022-02-23 18:10:53 donoban: I see, I will try to figure out a solution by playing with it a bit 2022-02-23 18:45:52 The behaviour is very strange and inconsistent 2022-02-23 18:46:15 Sometimes one of the devices does not give an error, sometimes it works, sometimes both give an error 2022-02-23 18:47:01 Repeatedly running setup-disk ends up working at some point, but also inconsistently 2022-02-23 18:52:18 I have never seen something like this, sometimes it just works, other times it does not 2022-02-23 19:00:01 Hi, I'm trying to boot alpine on a bit old computer and it is stuck at "Mounting security filesystem ..." any ideas what I can do? 2022-02-23 19:07:23 donoban: I believe I have found the issue, it is something I have observed before with sfdisk. It works consistently if you first do `echo "label: $DISKLABEL" | sfdisk $diskdev` and then create the partitions afterwards 2022-02-23 19:08:16 It does not seem to work consistently if you do both in a single step 2022-02-23 19:08:33 I will submit a patch to alpine-conf 2022-02-23 19:26:12 nice! 2022-02-23 19:33:27 dcreator: if you can edit the fileystem try to remove/comment it from '/etc/init.d/sysfs' 2022-02-23 19:37:32 donoban: Thanks for reply. Unfortunately it is running from a CD. 2022-02-23 19:38:35 I see.. 2022-02-23 19:40:18 donoban: I changed the date in the bios from 2001 to 2022, rebooted and then it mounted it instantly.. now it is at "starting busybox mdev". 2022-02-23 19:42:27 ahh great 2022-02-23 19:46:49 I will try to remember it, so if I have to travel back in time, and wants to bring an alpine linux installation to at least have some modern software, when modern hardware aren't available ... 2022-02-23 19:54:20 I disabled a lot of unneeded stuff in the bios to see if it is able to get past "starting busybox mdev". I guess that is related to /dev and enumeration of hardware? 2022-02-23 19:55:10 probably 2022-02-23 20:06:27 `grub-mkconfig` adds a lot of stuff to the kernel cmdline including stuff that's just plain wrong on my system so I have to overwrite it in `GRUB_CMDLINE_LINUX_DEFAULT` which then means I have the same option multiple times but with different values. How can I stop this tool from adding that stuff? 2022-02-23 20:06:43 e.g. my rootfs is btrfs but it addes `rootfstype=ext4` 2022-02-23 20:09:26 PureTryOut: isn't that based on the contents of /etc/grub.d/*? 2022-02-23 20:11:41 PureTryOut: btrfs works for me. Sort of. I use raid1 and mkgrubconfig adds both devices with newlines on kernel command line. 2022-02-23 20:11:58 ikke: I do not know, maybe? I'll check 2022-02-23 20:11:59 It worked before I converted to raid1 though 2022-02-23 20:12:21 PureTryOut: there is GRUB_CMDLINE_LINUX and GRUB_CMDLINE_LINUX_DEFAULT 2022-02-23 20:12:25 Not sure how they are defined 2022-02-23 20:12:40 GRUB_CMDLINE_LINUX_DEFAULT="modules=${modules} ${default_kernel_opts} ${GRUB_CMDLINE_LINUX_DEFAULT}" 2022-02-23 20:12:45 in 10_linux 2022-02-23 20:12:46 In what file? 2022-02-23 20:12:59 ah 2022-02-23 20:12:59 PureTryOut: i use Alpine 3.15 and it detects btrfs. Look in /etc/grub.d/10_linux 2022-02-23 20:13:51 There's /etc/default/grub as well 2022-02-23 20:14:07 right, 2022-02-23 20:14:13 that's where rootfstype is set for me 2022-02-23 20:14:25 ikke: I posted a link to redhat bug tracker earlier.. It's the same bug 2022-02-23 20:14:42 oh strange it gets it's stuff from `/etc/update-extlinux.conf` where indeed rootfstype is set for some reason 2022-02-23 20:15:03 Oh no it was https://www.mail-archive.com/bug-grub@gnu.org/msg15384.html 2022-02-23 20:15:13 Same error and corrupt config 2022-02-23 20:15:15 PureTryOut: https://wiki.archlinux.org/title/User:Eschwartz/Grub#Write_the_configuration_from_scratch 2022-02-23 20:15:49 mkconfig is a pile of garbage hacks, grub.cfg can be 3 lines 2022-02-23 20:15:49 elibrokeit: I didn't ask how to write it from scratch, I asked how to fix the tool doing it wrong 2022-02-23 20:16:09 you can edit the inaccurate shell scripts that generate the wrong info :) 2022-02-23 20:16:40 you're asking how to stop a bad tool from being bad, it's a fundamentally illogical question to have 2022-02-23 20:17:25 i mean 2022-02-23 20:17:31 one can certainly make bad things not bad 2022-02-23 20:17:34 That's not so nice. It usually works well and is the default tool in a lot of distros 2022-02-23 20:18:16 psykose: I agree and I experimented with that thought process. The result was https://paste.xinu.at/WxTG8Oeb/ 2022-02-23 20:20:06 it was bad for me because of bad configuration, thanks to the pointers here I fixed that. I have no need to write the file myself and neither do I want too 2022-02-23 20:21:47 this doesn't warn you of anything if you somehow have more than one ucode by accident 2022-02-23 20:22:13 and the error echo also outputs to stdout, even though it looks this writes the config to stdout (i.e. grub-gencfg > grub.cfg) 2022-02-23 20:22:19 aside from that, it's a nice start :) 2022-02-23 20:24:41 It was a quick hack which I never bothered to polish, I'm not sure why I forgot to redirect errors to stderr :/ 2022-02-23 20:26:07 also i think there is a confusion, in bash arr=(1 2 3); echo $arr prints 1, in zsh it prints 1 2 3 (by default?) 2022-02-23 20:33:37 well I have code there to explicitly check if [0] has something sensible and is a file, and if so prefix it with $BOOTPREFIX/ 2022-02-23 20:34:03 but not touch any other indexes at all, and only make use of $arr not ${arr[1]} and on 2022-02-23 20:34:21 basically the point there is assuming there is exactly one, but I wanted to do glob expansion :) 2022-02-23 20:47:34 actually, no, i'm confused again. i don't think there's any reason for $ucode on 36 and ${ucode[0]} on 37? 2022-02-23 20:51:08 hmm, idk either but in this case they are being used interchangeably. I think I probably updated line 37 to be explicit but forgot to update line 36 2022-02-23 21:30:30 mmhmm 2022-02-24 08:08:17 The weird grub-mkconfig issues seems to be related to the custom kernel I built 2022-02-24 08:09:10 I removed it with `apk del linux-lts`and then added it back from official repo. Now grub-mkconfig outputs a correct config with no line-break in the root= line. How odd, isn't it? 2022-02-24 13:01:30 I have a very strange issue with a python program using ZeroMQ and Python Watchdog, a file monitoring solution. When a file changes, my code gets terminated. Is this something for musl upstream? 2022-02-24 13:02:18 Is there any way to get musl and/or python symbols in gdb? 2022-02-24 13:02:38 skinkie: yes 2022-02-24 13:02:45 musl-dbg python3-dbg 2022-02-24 13:03:04 these pkgs contain debug symbols for gdb 2022-02-24 13:04:48 Ok, I narrowed it down to a completely different issue. A modification in vi, is causing the file to be deleted, and created. Instead of modified :S 2022-02-24 13:07:23 I can't help you with that, it is vi internals 2022-02-24 13:08:10 that sounds like you should write your file monitoring to also handle deletes and creates 2022-02-24 13:12:49 psykose: it was... thats where the terminated came from 2022-02-24 13:13:07 psykose: and why in that way, because watchdog is unable to handle unmounts :( 2022-02-24 13:13:32 psykose: but the problem now is that it does not get the 'create' 2022-02-24 13:13:41 sounds like an issue in watchdog then 2022-02-24 13:22:20 psykose: it sure is 2022-02-24 16:10:30 Hi! Is there any documentation where I can read up on the Alpine Linux build infrastructure? Like, is this physical machines? Do they have access to the internet? How is it prevented that building a package compromises the build infra and hence all further packages? 2022-02-24 16:11:06 there's an #alpine-infra 2022-02-24 16:11:58 Thx, I'll ask there 2022-02-24 16:32:03 any idea why udhcpc keeps running when interface gets changed in /e/n/i from dhcp to static? 2022-02-24 16:32:13 after restarting networking, that is 2022-02-24 16:32:40 ptrc: I suspect udhcpc is started in the background and nothing is keeping track of it 2022-02-24 16:34:50 there's a pidfile, according to /usr/libexec/ifupdown-ng/dhcp 2022-02-24 16:35:25 But what is tracking that it should be stopped 2022-02-24 16:37:10 oh, i guess ifupdown doesn't know that it should run it, because it's already changed in /e/n/i to static 2022-02-25 04:06:17 May I get a sanity check on this APKBUILD? 2022-02-25 04:06:22 ACTION sent a code block: https://matrix.org/_matrix/media/r0/download/matrix.org/XZKjaNfoCgpOzrtbOSlDklFX 2022-02-25 04:06:58 It's failing packaging, I believe. I must have something wrong with the make install line, I think... or maybe I forgot to pass something to the qmake to make chain 2022-02-25 04:07:07 Saijin_Naib: Does the pkgver actually work? The variable is $_gitdate but you use $gitdate 2022-02-25 04:08:05 ktprograms: Huh... I didn't see that, it hasn't got far enough for that to matter, I don't think 2022-02-25 04:08:08 Let me fix that up now 2022-02-25 04:10:02 Saijin_Naib: I think the problem is that INSTALL_ROOT should be "$pkgdir" and DESTDIR should be removed 2022-02-25 04:10:29 it seems that the Makefiles generated by qmake use INSTALL_ROOT to specify the base directory that files should be installed into 2022-02-25 04:10:30 That very well may be... I'm just following their build directions verbatim currently 2022-02-25 04:10:34 ACTION sent a code block: https://matrix.org/_matrix/media/r0/download/matrix.org/KxswPxgQSZFvDsMMCuqizHUr 2022-02-25 04:10:44 That does seem to track with these permission errors in fakeroot, right? 2022-02-25 04:11:04 Saijin_Naib: Yes, because you're supposed to write to pkgdir, not /usr 2022-02-25 04:11:55 Thank you so much. Re-running now. Do you think that would cause the testsuite to fail? I wasn't getting that to run either, it was also throwing Error 3 2022-02-25 04:12:35 The README says to set INSTALL_ROOT=/usr/local, which is correct if you're installing it for yourself, but when packaging Alpine software, the files need to be installed to $pkgdir 2022-02-25 04:12:52 Saijin_Naib: You can try re-enabling the tests and sending the error log 2022-02-25 04:13:19 Saijin_Naib: Did you read https://github.com/nickbnf/glogg#tests 2022-02-25 04:14:12 ktprograms: Yes, I basically used that whole block for my test commands in the APKBUILD, minus the env-vars 2022-02-25 04:14:26 BTW This project was last updated 4 years ago 2022-02-25 04:14:40 As it says, you need Google Mocks, but that doesn 2022-02-25 04:14:52 doesn't even exist any more: https://github.com/google/googlemock 2022-02-25 04:14:58 I am aware. There is a fork (klogg) that uses a different backend that is faster, but doesn't perform the same in all cases in temrs of matching 2022-02-25 04:15:15 the google-test seems to have googlemock in it 2022-02-25 04:15:47 Saijin_Naib: Probably, but I 2022-02-25 04:16:03 I'm not sure how to use it, because you need to set the GMOCK_HOME env var 2022-02-25 04:16:46 Hmm... maybe that's why my tests were failing. I wasn't setting the env vars since I dind't know how I could 2022-02-25 04:16:59 I think you fixed up my APKBUILD though, so mega-thanks for that 2022-02-25 04:17:07 Saijin_Naib: With the export line stated in the readme 2022-02-25 04:17:14 klogg looks to build the same, so I can likely do that one up easily too, now! 2022-02-25 04:17:43 ktprograms: and that will function the same in the abuild process? fakeroot or however it works? 2022-02-25 04:18:00 Saijin_Naib: It _should_, after all it's just running a shell 2022-02-25 04:19:15 So would installing gtest and doing a which gmock and getting that path work for the env var path to gmock? 2022-02-25 04:19:16 Or am I approaching this wrong? 2022-02-25 04:20:21 Saijin_Naib: You should find out the path manually and hard code it in the APKBUILD 2022-02-25 04:21:56 Our gmock package is .so... does that matter? 2022-02-25 04:23:54 I'm sorry, I have no experience with gmock 2022-02-25 04:24:35 Fair enough! Me either, haha 2022-02-25 04:32:06 Is this something that would raise eyebrows if I skipped tests on if I can't get gmock to cooperate? 2022-02-25 04:32:55 Saijin_Naib: If you want just leave your best attempt in with it disabled, and mention it clearly in the MR, someone might be able to help 2022-02-25 04:33:53 If you just add gmock to checkdepends, does it work? 2022-02-25 04:34:24 That's queued up next after this rebuild (hopefully) succeeds... N3450 not fast 2022-02-25 04:36:57 ktprograms: Beautiful! Builds fine, minus I guess .desktop file going in the wrong place 2022-02-25 04:45:29 ktprograms: it is installing to /bin for the program, and /share for docs/icons/applications... Did I mess up the prefix? It needs to go under /usr for Alpine, right? 2022-02-25 04:50:59 ktprograms: adding gmock alongside gtest-dev to checkdepends does not fix it. I think you're right about needing to figure out/set env vars 2022-02-25 04:55:16 Saijin_Naib: Try setting prefix in the qmake step with 'PREFIX=/usr qmake' 2022-02-25 05:43:27 ktprograms: I did this, and it worked. I assume this was not the right approach? 2022-02-25 05:43:35 package() { 2022-02-25 05:43:35 } 2022-02-25 05:43:35 make install INSTALL_ROOT="$pkgdir/usr" 2022-02-25 05:54:13 Saijin_Naib: It's a bit of a hack, because programs sometimes use PREFIX for certain functionality 2022-02-25 05:54:24 You can do that, and point it out in the MR for review 2022-02-25 16:45:18 Ariadne: afaict the 'BOOTSTRAP=nobase' thing is never actually checked? 2022-02-25 16:45:50 although it might just be implicitly be checked by seeing if BOOTSTRAP is not empty.. 2022-02-26 00:16:49 Is there a way to have the data disk also house the apkovl using UUIDs? 2022-02-26 00:17:12 the problem I am currently having is that when the boot disk scans for the apkovl it mounts the disk as /media/sdx 2022-02-26 00:17:25 which makes the uuid based fstab entry fail 2022-02-26 10:21:21 ACTION hey yah all skraito here what's up , i am here 2022-02-26 10:22:25 ACTION brb 2022-02-26 10:23:33 ACTION after sabbath day guys i join you all at 1:48 more hour rest well ... . is your sabbath day soon , at friday night to saturday night ... . 2022-02-26 10:24:34 ndut2: Are you in the correct channel? 2022-02-26 10:24:47 no theyre spamming a bunch 2022-02-26 10:25:09 ACTION get the fuck off libera none of your concern here ... . 2022-02-26 11:39:20 hi 2022-02-26 11:39:28 has anyone used alpine on a rockpro64 before? 2022-02-26 11:59:32 Trying to set up gitolite, I've followed the steps on alpine wiki but get denied permission when trying to clone gitolite-admin. Does it have something to do with not having installed PAM (as the archwiki suggested that) or the strict config that gitolite puts on .ssh/authorized_keys? 2022-02-26 13:32:20 Fixed it, turns out I need to unlock the git account or enable PAM (which I don't have installed) 2022-02-26 19:34:12 I'm doing a PXE boot and getting an error when the kernel tries to load modules: 2022-02-26 19:34:38 modprobe: can't change directory to /lib/modules: no such file or directory 2022-02-26 19:35:48 Any idea what I might be doing wrong? I have an NFS mount, and I'm using the alpine_dev option to point to it. 2022-02-26 19:36:20 I can't find any other indications on https://wiki.alpinelinux.org/wiki/PXE_boot as to what I'm missing 2022-02-26 19:36:46 Thanks 2022-02-26 19:36:49 afaik there is no modloop with the netboot image 2022-02-26 19:40:55 Apparently the nfs share is intended to host the modules. I dunno where to get those from though. 2022-02-26 19:41:52 But also, there's an option to specify the modloop when PXE booting. So I should have a modloop. 2022-02-26 19:42:14 This is my ipxe cmd: kernel http://${net0/next-server}/vmlinuz ip=dhcp modloop=http://${net0/next-server}/modloop modules=loop,squashfs,sd-mod,usb-storage alpine_dev=nfs:${net0/next-server}:/pxesrv/nfs alpine_repo=http://${net0/next-server}/ apkovl=http://${net0/next-server}/apkovl-01.tar.gz 2022-02-26 19:42:24 (sry its so long) 2022-02-26 19:49:53 QDX45: is the modloop service started? 2022-02-26 19:52:58 I don't know. I assumed the boot cmd above would start it? 2022-02-26 19:56:31 When I try and run modprobe from within the pxe boot it says theres no /lib/modules directory 2022-02-26 19:56:39 So I guess something didn't start 2022-02-26 19:56:51 rc-status 2022-02-26 19:56:57 and rc-service modloop start 2022-02-26 19:57:09 the modloop services is responsible for mounting the modloop to /lib/modules 2022-02-26 19:57:41 does the aarch64 kernel config of alpine contain any extra options compared to the defconfig? 2022-02-26 19:58:20 or does it only disable things 2022-02-26 19:59:05 ikke thank you! that totally worked. 2022-02-26 19:59:40 How should I start the service in the future? put that cmd in a script or is there something more correct? 2022-02-26 20:07:51 QDX45: not sure how that's suppose to work with PXE boot 2022-02-26 20:08:01 fair enough 2022-02-26 20:13:15 if you are loading an apkovl, isn't it possible to build a custom one then that writes those commands to /etc/rc.local? 2022-02-26 20:13:23 or some similar location? 2022-02-26 20:13:41 mfg: if you have an apkovl, services are part if it 2022-02-26 20:13:42 ()i don't know exactly how openrc works) 2022-02-26 20:13:45 I am loading an apkovl 2022-02-26 20:13:57 then rc-update add modloop 2022-02-26 20:14:17 Where do I put services manually inside the apkovl? 2022-02-26 20:14:46 I'm making it manually on a not-alpine os 2022-02-26 20:14:50 in the aports repo under scripts are genapkovl scripts, you can copy one customize it and set the according variable in mkimg.netboot.sh toi use it 2022-02-26 20:14:54 /etc/runlevels/default/* 2022-02-26 20:14:57 as symlinks 2022-02-26 20:15:03 awesome thanks 2022-02-26 20:15:13 there is also boot 2022-02-26 20:15:19 okay it's just symlinks, nice :D 2022-02-26 22:34:41 Sorry for all the questions. I noticed that when I run alpine in a VM, the /sys/class/dmi/id/board* files are missing. Is that also true of bare-metal alpine installations? 2022-02-26 22:36:00 ACTION hi all how are ya ? skraito here ... . what you call me for ? ... . 2022-02-26 22:36:38 i am not sure gqx45 2022-02-26 22:36:40 there is couple of OS we released 2022-02-26 22:36:44 better choose Our OS 2022-02-26 22:36:54 and port the package to alpine linux 2022-02-26 22:44:23 ok. fair enough 2022-02-26 23:38:57 ACTION brb 2022-02-26 23:39:47 ACTION bah installing AstaraOS now 2022-02-26 23:44:11 please stop spamming 2022-02-26 23:51:08 it's a bot 2022-02-26 23:53:32 yeah yeah is a bot 2022-02-27 01:07:55 ACTION back all 2022-02-27 01:08:57 ACTION skraito here just ask all busy building blog now 2022-02-27 01:11:09 ACTION there is secret for wireless super fast ... . you can make it something like 9 bit or alot of bit ... . which mean the packet from different2 antena . .. . that's all i can reveal to you ... . think for yourself ... . 2022-02-27 01:29:48 ACTION skraito here 2022-02-27 01:35:21 ACTION i am actually looking for some team mate ... . :) anyone interested ? ... . 2022-02-27 01:54:17 QDX45: I don't know if you got any actual answer, I do have some board_* on this laptop. 2022-02-27 04:35:15 ACTION hi anyone wanna code something together ? may be we can meet at channnel #rabbithacker 2022-02-27 05:21:26 thanks for checking lucidiot. good to know. I guess qemu doesn't like showing those. 2022-02-27 05:51:23 In light of mp3gain landing, could I get some sanity checks on my APKBUILD for wxmp3gain? 2022-02-27 05:51:25 ACTION sent a code block: https://matrix.org/_matrix/media/r0/download/matrix.org/ggekCRSXzejmRMvZikivtjLb 2022-02-27 05:51:53 Compiles fine, but all of the icons are missing in the GUI when run... Am I missing something with a prefix or other cmake variable? 2022-02-27 05:53:33 ACTION uploaded an image: (32KiB) < https://matrix.org/_matrix/media/r0/download/matrix.org/ymUNWLZohGcbanHaTkPBuZXR/image.png > 2022-02-27 05:54:16 ACTION uploaded an image: (54KiB) < https://matrix.org/_matrix/media/r0/download/matrix.org/sWhNSBoepqkCUFVxVTkSAXjE/image.png > 2022-02-27 05:54:33 Those can't go there, so I have something wrong in the build 2022-02-27 15:19:36 dammit, missed QDX45 again 2022-02-27 20:08:30 APKBUILD overall seems fine but please version it as `pkgver=4.0_git20220221` and remove the `_gitdate` variable 2022-02-27 21:10:20 hello good people 2022-02-27 21:10:40 how tf do I install this package https://gitlab.alpinelinux.org/alpine/aports/-/tree/master/community/anbox 2022-02-27 21:11:31 I just want this package to run, only alpine has it. I don't know anything about alpine, I intalled it on a physical machine. now I don't know how to install this package 2022-02-27 21:12:33 make sure you have the community repository in /etc/apk/repositories , and then you can just add it if you are on one of the supported architectures for it 2022-02-27 21:13:45 hi 2022-02-27 21:14:14 having issues with ffmpeg https, not sure what the issue is 2022-02-27 21:14:53 trying to use ffmpeg to play https streams says "https protocol not found, reocmpile ffmpeg with openssl, gnutls, or securetransport enabled." 2022-02-27 21:15:03 i see the aports community ffmpeg is built with gnutls enabled 2022-02-27 21:15:16 and https is listed in the ffmpeg binary protocols 2022-02-27 21:15:41 so why can't i play https streams with ffmpeg? 2022-02-27 21:17:37 i tried using abuild to build the aports ffmpeg myself, but the result is the same, https not working 2022-02-27 21:18:02 tsuki: I don't see ffmpeg depending on gnutls or openssl at runtime 2022-02-27 21:18:21 alpine's community repo ffmpeg is not built with openssl support 2022-02-27 21:18:27 it's built with gnutls support 2022-02-27 21:18:34 yes, but I see neither 2022-02-27 21:18:34 but it doesn't work for some reason 2022-02-27 21:19:01 apkh, ffmpeg-libs 2022-02-27 21:19:01 if you run ffmpeg -h, you should see --enable-gnutls in the configuration specified 2022-02-27 21:19:03 oh*' 2022-02-27 21:19:10 then ffmpeg -protocols should show https 2022-02-27 21:19:23 give me a sample https url 2022-02-27 21:19:32 Then I don't know why it does not work 2022-02-27 21:19:34 psykose that was easy, thank you very much. now I need a window manager :D 2022-02-27 21:19:37 :) 2022-02-27 21:25:31 https://wdcb-ice.streamguys1.com/wdcb128 2022-02-27 21:25:50 or you can use a search engine to find icecast https servers 2022-02-27 21:26:19 i typed `ffplay thaturl` and it worked 2022-02-27 21:26:54 some calm music 2022-02-27 21:27:28 are you on edge or some other release 2022-02-27 21:27:30 yea, sunday's and monday's on that station are usually pretty chill 2022-02-27 21:27:34 and what error do you get 2022-02-27 21:27:38 i see that it works fine with ffplay too tho, that's weird 2022-02-27 21:28:00 `ffmpeg` by itself can't load a url like that i don't think, you need more arguments 2022-02-27 21:28:43 i know that, but i am using mplayer which uses ffmpeg 2022-02-27 21:28:53 mplayer -nocache -loop 0 ${url} 2022-02-27 21:29:01 then i get the error about lacking https support in ffmpeg 2022-02-27 21:29:02 no idea why 2022-02-27 21:30:02 i am on edge 2022-02-27 21:31:02 you see 2022-02-27 21:31:06 it has a vendored ffmpeg 2022-02-27 21:31:13 and there is no gnutls/openssl in its makedeps 2022-02-27 21:31:23 so the vendored ffmpeg it builds does not actually have the support 2022-02-27 21:32:43 that makes sense, why doesn't it use the system ffmpeg? i don't remember mplayer working like that 2022-02-27 21:33:04 shrug, maybe someone forgot to unvendor it or they didn't realise or it was too hard 2022-02-27 21:33:06 i can look into it 2022-02-27 21:33:20 i am looking at the apkbuild file rn 2022-02-27 21:33:41 rebuilding with gnutls-dev then doesn't print that error anymore 2022-02-27 21:33:57 nice 2022-02-27 21:34:08 gonna give it a try in a few minutes 2022-02-27 21:34:11 although, it still barely works 2022-02-27 21:34:15 takes like.. 10 seconds to start 2022-02-27 21:34:15 haha 2022-02-27 21:34:18 but then it runs 2022-02-27 21:34:18 :) 2022-02-27 21:34:25 https://img.ayaya.dev/GuKMKkJaOXat.png 2022-02-27 21:35:04 yea, that's typical icecast connections lol 2022-02-27 21:35:18 cool, good to see 2022-02-27 21:37:46 adding https://img.ayaya.dev/iaqTmRB6Xsys seems to make it use system ffmpeg 2022-02-27 21:42:37 alright, trying to build now 2022-02-27 21:43:21 Size difference for mplayer: 27 MiB -> 5328 KiB 2022-02-27 21:43:21 haha 2022-02-27 21:43:35 lol 2022-02-27 21:43:53 this old laptop does not enjoy building stuff 2022-02-27 21:44:37 i merged it, so it will be on edge in a bit 2022-02-27 21:46:00 and by in a bit i mean like 5-10 minutes on dl-cdn 2022-02-27 21:46:23 that's convenient 2022-02-27 21:46:34 thanks 2022-02-27 21:46:39 anytime :) 2022-02-27 21:49:38 the build worked fine for me which is nice to see too 2022-02-27 21:49:43 that vendored ffmpeg through me for a loop 2022-02-27 21:49:57 mhm 2022-02-27 21:49:59 unexpected 2022-02-27 22:23:09 anyone here using sway? 2022-02-27 22:23:36 me 2022-02-27 22:24:11 when I press the predd the respective keys to open a terminal or dmenu, nothing happens. what's the issue lol 2022-02-27 22:24:57 am I missing a package or some conf? 2022-02-27 22:25:12 by default the terminal is foot, and not installed 2022-02-27 22:25:19 I installed that 2022-02-27 22:25:19 neither is dmenu 2022-02-27 22:25:22 that too 2022-02-27 22:25:31 hm 2022-02-27 22:26:35 do you see your mouse moving in it 2022-02-27 22:26:45 yeah, that's working fine 2022-02-27 22:26:55 no clue then 2022-02-27 22:26:58 I can click stuff and it works 2022-02-27 22:27:02 damn 2022-02-27 22:27:05 you could edit the config to specify some other terminal 2022-02-27 22:27:07 and install that one 2022-02-27 22:27:13 I did 2022-02-27 22:27:19 still nothing 2022-02-27 22:27:26 what config file are you editing 2022-02-27 22:27:34 but I wonder about why dmenu doesn"t work 2022-02-27 22:27:46 the one I put in .config/sway 2022-02-27 22:27:59 .config/sway/config ? 2022-02-27 22:28:01 ye 2022-02-27 22:28:02 could you post it 2022-02-27 22:28:24 even the standard conf doesn't work mate 2022-02-27 22:29:32 it's ridiculious 2022-02-27 22:30:05 well, if you don't want to post it, could you start sway with `sway -dV &> sway.log` and then do some stuff and post the log 2022-02-27 22:31:38 also are you sure you are pressing the correct mod key for the shortcuts, since the default is super and not alt 2022-02-27 22:33:03 yea, I changed it to alt, but neither does work. nor win key as default nor mine 2022-02-27 22:33:16 I'm sure it's something stupid 2022-02-27 22:34:22 I'll use dwm, that usuakky works. 2022-02-27 22:35:15 you could post the log i asked for 2022-02-27 22:35:48 I won't use sway anymore, I did not know it before I consulted the wiki, then it sucked for me 2022-02-27 22:36:01 I might try again if dwm doesn't work either 2022-02-27 22:48:53 well, dwm just works as expected. 2022-02-27 22:49:11 sorry for swiping you off, but using wayland made me ill 2022-02-27 22:53:48 i mean you could literally just post the log and then it would help me figure it out 2022-02-27 22:53:48 and help the next person with the same issue 2022-02-27 22:57:55 I decided to not use alpine at all, I just installed it because you had a package nobody else had 2022-02-27 22:58:36 but the package also didn't work (don't worry, the devs are not sanest people) so I just use a completely different solution. 2022-02-27 22:58:57 I'm very thankful for you being so helpful, don't get me wrong 2022-02-27 23:01:05 well, you can obviously use whatever you want, but if you never actually fix any issues how do you expect anything to improve 2022-02-27 23:01:17 if i used something else every time something didn't work i would retire from computers 2022-02-27 23:01:18 :p 2022-02-27 23:01:48 I do fix issues, but wayland is an issue somebody else has to worry about 2022-02-27 23:02:05 and sway is now on my list of bad software 2022-02-27 23:02:10 my friend 2022-02-27 23:02:13 posting the log takes like 2022-02-27 23:02:14 3 minutes 2022-02-27 23:02:27 right, but that's not what I installed the OS for 2022-02-27 23:02:37 psykose: 30s not 3 mins! 2022-02-27 23:02:46 yeah, I agree 2022-02-27 23:02:56 you have spent more time typing about stuff not working 2022-02-27 23:02:58 than just posting the log 2022-02-27 23:02:59 fyi 2022-02-27 23:03:13 :p 2022-02-27 23:03:23 but yes, good luck 2022-02-27 23:03:23 I got some time while stuff is compiling, I don't see the problem 2022-02-27 23:04:10 all I can say is 2022-02-27 23:04:16 >using wayland 2022-02-27 23:06:26 bb 2022-02-28 03:17:35 PureTryOut (matrix.org): Is that dangerous to do, or just not done by convention? 2022-02-28 03:18:39 convention 2022-02-28 03:18:49 you probably need to install gtk icons theme 2022-02-28 03:18:57 i also looked at it, and couldn't figure out why it tried to load the icons from bindir 2022-02-28 03:19:08 oh is that the issue 2022-02-28 03:19:22 but it also uses wxgtk (2) so perhaps better left unpackaged 2022-02-28 03:19:59 apparently it's hardcoded: https://github.com/cfgnunes/wxmp3gain/blob/1bfe4c619ab546ce7ed99c6847d21dff363c8814/src/Constants.h#L90 2022-02-28 03:19:59 yeah, it tried to load usr/bin/resource/toolbar/something.png or whatever 2022-02-28 03:20:22 ah 2022-02-28 03:20:40 if that doesn't exist then you just get exe path 2022-02-28 03:20:42 neat 2022-02-28 03:21:30 but __LINUX__ isn't a real macro 2022-02-28 03:22:31 easily fixed with a -D :) 2022-02-28 03:22:42 or patch it to be __linux__ like it should be 2022-02-28 03:22:46 or just don't package it 2022-02-28 03:30:06 i kinda don't see the point of this kind of program. it's not really simpler than opening terminal and running mp3gain *.mp3, even for inexperienced users. yes, find . -name '*.mp3' -exec mp3gain {} + is complicated, but the gui doesn't actually let you do that anyways 2022-02-28 03:32:42 also mp3gain is nigh-obsolete anyways, since it only supports mp3 which isn't lossless even at very high bitrates. modern software like r128gain and loudgain support all common media formats (even video), the only limitation is that extremely old players don't support the metadata 2022-02-28 03:39:44 Missing icons got packaged to /usr/share/wxmp3gain/resource, but those errors indicate it's looking in /usr/bin/resource... I need to patch the source, right? Or is this a build/configure flag somehow? 2022-02-28 04:23:50 Hi, I need some help with configuring the trackpad on my Thinkpad X1 Carbon (Gen 1), using the libinput driver. I can't seem to figure out how to enable pressing to click. I added 'Option "Tapping" "on"' to the libinput file in xorg.conf.d, and I indeed can tap to click, but pressing doesn't click. 2022-02-28 05:53:24 Saijin_Naib: is what dangerous to do? What I mentioned is the standard to do it in Alpine, I'm not sure how it would be dangerous? 2022-02-28 09:02:36 Sorry, turns out I just didn't understand how clicking on the Thinkpad trackpad works (clicking only works in the bottom left) 2022-02-28 13:33:41 "Saijin_Naib: is what dangerous..." <- I meant how I had it with the _gitdate variable. 2022-02-28 13:47:06 I'm still unsure what would be dangerous about it, but it's a variable you don't use 2022-02-28 13:47:21 Also you can remove any empty variables like `depends=""` 2022-02-28 15:41:45 ACTION sent a code block: https://matrix.org/_matrix/media/r0/download/matrix.org/TUPZApGSHfKZRlxvhzwwhTbI 2022-02-28 15:41:53 That's the CMakeLists.txt 2022-02-28 15:42:16 It seems like I need to adjust line 61, correct? 2022-02-28 16:39:16 well you obviously shouldn't install icons to /usr/bin 2022-02-28 16:39:29 why don't you patch the wrong __LINUX__ to __linux__ 2022-02-28 18:35:27 Hi everyone , I am planning on doing a root on zfs (with native encryption) .If someone could explain the processes specific to alpine . how does the kernel modules are built in alpine ? and etc .. ?. Thanks . 2022-02-28 18:37:35 Luckily you don't need to build kernel modules. There's a pre-built package in the repos. https://pkgs.alpinelinux.org/packages?name=zfs-lts&branch=v3.15 2022-02-28 18:42:28 There's a page on the wiki for setting up ZFS on root. I'm not sure how up to date it is: https://wiki.alpinelinux.org/wiki/Root_on_ZFS_with_native_encryption 2022-02-28 18:44:11 very nice . Another question . Can I install firmware packages only specific to my hardware ? . I've searched the package mirror and the firmwares are well split . is it possible ? 2022-02-28 18:46:58 xordspar0: I am reading the wiki page . It's upto date ( 12 february 2022) . Thanks for mentioning it .