2022-10-01 07:27:10 it was a reply to Ermine 2022-10-01 09:17:55 lopid: And I replied to paul, but they turned out to have quit. 2022-10-01 11:41:55 Hello 2022-10-01 11:41:55 It has suddenly stopped working 2022-10-01 11:41:55 what is wrong with the alpine wiki? 2022-10-01 11:44:13 database error 2022-10-01 11:44:31 working on it 2022-10-01 11:47:54 Restored 2022-10-01 15:50:25 hi, I installed xfce and cannot control brightness via power-manager while logged as a normal user, from a root it works, so it seems like xfce4-power-manager is launched from root account and as user cannot modify values in /sys/class/backlight/intel_backlight/brightness i went through google searching already, but couldnt find any info, so anyone knows how can i fix this? 2022-10-01 15:51:19 must you be in the correct group? 2022-10-01 15:52:37 i checked that, there is no group for power management and my user is already in video group 2022-10-01 17:27:37 Hello guys. Sorry for offtop but iirc someone recommended met pv to see progress bar with cp copy. 2022-10-01 17:28:02 Could you guys write an example how to use it with cp? 2022-10-01 17:28:07 We do have an off-topic channel 2022-10-01 17:28:21 #alpine-offtopic 2022-10-01 17:30:29 pv works with pipes 2022-10-01 17:30:51 rsync -ah --progress source-file destination-file 2022-10-01 19:39:43 Hello, I just installed Alpine Linux and am struggling to install Flathub apps. I followed the wiki instructions and was able to install flatpak, and I added flathub as a repository, but when I run a command to install a flathub app I get a message that the app was not found. I copied the app names directly from flathub's site and tried multiple apps (weather, clocks, libreoffice). I ran remote-list -d and see flathub listed. Not sure what to do to 2022-10-01 19:39:44 install flatpaks. Thanks for your help 2022-10-01 19:40:53 if you `flatpak search app` what do you see 2022-10-01 19:41:36 No matches found 2022-10-01 19:42:48 what did you search 2022-10-01 19:43:33 'weather' , 'firefox', and 'libreoffice' 2022-10-01 19:44:17 hmm 2022-10-01 19:44:39 could you paste bin that with `-vv` added 2022-10-01 19:46:01 F: No installations directory in /etc/flatpak/installations.d. Skipping 2022-10-01 19:46:11 (there's more) 2022-10-01 19:46:26 F: Opening system flatpak installation at path /var/lib/flatpak 2022-10-01 19:46:49 F: Opening user flatpak installation at path /home/jared/.local/share/flatpak 2022-10-01 19:47:16 F: flathub:i386 appstream age 2875 is less than ttl 86400 2022-10-01 19:47:30 F: Opening system flatpak installation at path /var/lib/flatpak 2022-10-01 19:47:50 F: Opening user flatpak installation at path /home/jared/.local/share/flatpak 2022-10-01 19:47:53 No matches found 2022-10-01 19:48:10 what hardware do you have 2022-10-01 19:48:12 (This was from the command 'flatpak search libreoffice -vv) 2022-10-01 19:48:51 maybe due to i386? Are there 32bit packages on Flathub? 2022-10-01 19:50:38 Dell Inspiron, processor Intel Core I7-7500U with integrated graphics 2022-10-01 19:50:59 you installed the wrong image 2022-10-01 19:51:07 x86 is 32-bit 2022-10-01 19:51:14 x86_64 is what you want 2022-10-01 19:51:26 of all of alpine? 2022-10-01 19:51:32 dang, no wonder flatpak wasn't working 2022-10-01 19:51:47 Thanks for your help, I'll reinstall. 2022-10-01 19:51:50 well by itself it doesn't imply anything, but i guess they don't ship 32-bit anything on flatpak anymore 2022-10-01 19:52:05 regardless you definitely don't want that one with that hardware haha 2022-10-01 19:52:05 Oh so reinstalling woudn't fix this issue 2022-10-01 19:52:08 it would 2022-10-01 19:52:57 Oh I get it, right. I'll reinstall, appreciate the help. Have a good one 2022-10-01 21:14:15 hi, is nvme-cli still in repos? or what do i need to install on nvme disk? 2022-10-01 21:17:45 ghormoon_: it's in community 2022-10-01 21:18:53 https://pkgs.alpinelinux.org/packages?name=nvme-cli&branch=edge&repo=&arch=&maintainer= 2022-10-01 21:37:56 hm, it doesn't list anything even after modprobe nvme, though 2022-10-01 21:38:26 what did you type 2022-10-01 21:39:34 nvme list shown only headers. /dev/mmc* doesn't exist, dmesg doesn't look like it would detect the hw 2022-10-01 21:39:52 did you run it as root for the first part 2022-10-01 21:40:18 yup 2022-10-01 21:40:27 what storage do you have 2022-10-01 21:46:01 hm, maybe i'ts just sata, but i don;t see it in fdisk anyway 2022-10-01 21:46:21 could you post nvme list -v and dmesg in a pastebin 2022-10-01 21:46:49 sfsa2048u1br2-c-ms-236-fts-to 2022-10-01 21:47:13 ugh, does it have some commandline pastebin tool or i need to install one? :D 2022-10-01 21:47:49 it's sata anyway 2022-10-01 21:48:20 lspci sees the sata controller at least 2022-10-01 21:48:33 the drive is sata2 yeah 2022-10-01 21:48:41 but it's not in /dev/sd*, there's inly /dev/sda which is the boot flash disk 2022-10-01 21:49:46 either something is broken or missing firmware or a kernel module for something 2022-10-01 21:49:51 dmesg may or may not have more 2022-10-01 21:49:55 hard to know without the whole device 2022-10-01 21:50:32 Fujitsu Futro S720 2022-10-01 22:04:00 ok, pastebin-cli and tpaste failed me, this works though: http://sprunge.us/IKSLTy dmesg, but i don't see any obvious issue there 2022-10-01 22:06:13 lspci -v http://sprunge.us/0JAgoJ 2022-10-01 22:12:55 the sata interface comes up fine 2022-10-01 22:13:15 if fdisk -l doesn't see it i would guess the actual drive is not detected/broken/whatever 2022-10-01 22:15:15 well, the drive boots, but there's some ikk linus that is limited as much as possible xD 2022-10-01 23:11:43 ha, i somewhere read an article about this hwsaying that alpine has no issues ... so based on the post date, i guessed they used 3.8.2 and that one sees the disk 2022-10-01 23:22:03 could you post an lspci -k 2022-10-01 23:23:33 shortly :) i just finished installing 3.8, i'll try to upgrade version by by version to see when i die 2022-10-01 23:26:24 http://sprunge.us/ukBK68 2022-10-01 23:29:30 i assume that's working 2022-10-01 23:30:31 yeah that's on 3.8 2022-10-01 23:30:47 none of those come across as storage, weird 2022-10-01 23:31:00 is the storage on usb or something haha 2022-10-01 23:31:09 no, it's the sata one 2022-10-01 23:31:48 ah, right 2022-10-01 23:46:21 is the linux-vanilla correct package to have? i'm already at 3.14 and i'm still on 4.19.x 2022-10-01 23:47:35 linux-edge, linux-vert, linux-lts are the ones I know of. hm 2022-10-01 23:48:07 ok at some point that changed and the upgrade doesn't handle that 2022-10-01 23:48:40 i'll try to replace that but i'm pretty sure that is the thing that will break it now as i'm already upgradingt to 3.15 2022-10-01 23:48:45 those would bring you to 5.19.12, 5.15.71, 5.15.71 respectively 2022-10-01 23:49:45 if your bootloader is setup to allow booting from old kernels, there shouldn't be a lot of harm in trying to upgrade. 2022-10-01 23:50:41 ok last is 3.10, 3.11 doesn't have linux-vanilla anymore 2022-10-01 23:53:12 yup generating initramfs with linux=lts thorws bunch of errors no such file or directory on modules 2022-10-01 23:54:31 vanilla -> lts 2022-10-01 23:54:40 what errors 2022-10-01 23:54:49 can't copy that xD i'm still on basic console 2022-10-01 23:56:05 yup lts won't boot, vanilla will 2022-10-02 00:00:06 http://sprunge.us/5ufgmO none of these look like disk, though 2022-10-02 00:00:12 anyway, i'll have to continue in the morning, ppl wanna sleep xD 2022-10-02 00:06:46 tomorrow i'll get back to the lts on 3.11 to see if that works or not 2022-10-02 08:46:52 hi, can i ask if Alpine will follow the "Fedora drop of the VAAPI h264,h265 and av1 hardware decoders"? Look like it is spreading across distros and i would like to know if Alpine will also follow 2022-10-02 08:55:17 i planned to use Alpine for a future computer and it would be sad to not be able to fully use the hardware i will buy ๐Ÿ˜ข 2022-10-02 08:55:50 also i didn't understand if this is AMD specific or will apply also to Intel hardware 2022-10-02 08:57:24 it applies to all vaapi via mesa, so every open source driver in essence 2022-10-02 09:02:07 I made a blunder. can I go back one step in the installer? 2022-10-02 09:02:46 ctrl + c, and re-run step you want 2022-10-02 09:02:55 setup-alpine just executes multiple setup-* scripts 2022-10-02 09:03:20 great, thanks 2022-10-02 09:03:35 but you will have to complete by running rest of scripts 2022-10-02 09:20:31 "it applies to all vaapi via mesa..." <- i see, thanks ๐Ÿ˜ข 2022-10-02 09:20:39 (the stack) 2022-10-02 09:20:45 the disablement, no idea 2022-10-02 09:21:44 i haven't seen anyone do it except redhat 2022-10-02 09:21:56 i also don't care what redhat is doing, and i have 0 lawyers outside my door 2022-10-02 09:21:58 wink emoji 2022-10-02 09:22:03 and that's the end of my thoughts on that one 2022-10-02 09:23:03 psykose: OpenSUSE is doing it too 2022-10-02 09:23:19 ah, redhats sister, right 2022-10-02 09:23:51 german red hat 2022-10-02 09:24:02 indeed 2022-10-02 09:27:17 also i'm not sure how they would drop the av1 decoder 2022-10-02 09:37:06 Hi, running into a strange behavior where my gitlab ci (using alpine:3.16 ) is pulling in nodejs-current when I run `apk add nodejs`, however when I run `docker -ti --rm alpine:3.16`, and run `apk add nodejs` it pulls in nodejs-16.x. Anybody seen something like this with gitlab saas before? 2022-10-02 09:39:18 can you post /etc/apk/world of the ci environment where -current is pulled 2022-10-02 09:46:24 https://gitlab.com/winny.tech/blog.winny.tech/-/jobs/3115327846 alpine-baselayout alpine-keys apk-tools busybox libc-utils 2022-10-02 09:47:12 i think i will have some more details shortly, i didn't actually try that apk add line 2022-10-02 09:47:27 was just confused that apk add nodejs in my docker container pulled in a different version 2022-10-02 09:49:30 ah think i figured it out, i was running apk add npm, which seems to prefer the latest nodejs, so if i instead do `apk add nodejs npm` it'll pull in the 16.x version 2022-10-02 09:50:49 "i also don't care what redhat is..." <- yeah, but i was asking because other distros, like Opensuse, were doing the same; of couse i know policies one distro have do not apply to all other ones ๐Ÿ˜ 2022-10-02 09:51:03 thanks for the info! 2022-10-02 09:52:03 winny: ah, it's because npm depends on cmd:node, but that is versioned based on pkgver, and so -current is a higher number.. but if you add nodejs, that's more of a virtual and prefers the lts one, and in the end that is picked 2022-10-02 09:52:28 Uniformbuffer[m]1: that and this doesn't apply to the other codecs (av1/vp9) even if they're disabled 2022-10-02 09:52:28 ah makes sense thanks :) 2022-10-02 09:52:42 youtube doesn't even serve h264 unless you ask specifically pretty sure, just those 2022-10-02 09:52:44 is there a way to blacklist packages i especially have no intention of accidentally installing? 2022-10-02 09:52:49 (as an example) 2022-10-02 09:52:54 winny: `apk add !nodejs-current` 2022-10-02 09:53:01 ha, couldn't be easier 2022-10-02 10:30:12 what provides gid_t / uid_t type 2022-10-02 10:35:01 seems like it should be in unistd.h 2022-10-02 10:39:52 panekj: bits/alltypes.h on musl 2022-10-02 10:40:15 that's annoying 2022-10-02 10:41:16 but if you include unistd.h uid_t and gid_t will be defined 2022-10-02 10:41:49 I'm getting error: unknown type name 'gid_t' so that's not true 2022-10-02 10:43:18 with #include 2022-10-02 10:44:47 works for me 2022-10-02 10:46:03 o, ok, it fails in utils.h not utils.c 2022-10-02 10:48:15 btw system_data_types(7) is very good resource to check for types stuff 2022-10-02 10:48:47 documentations are for nerds 2022-10-02 10:49:08 you people are amazing, alpine is by far my most favorate distro 2022-10-02 10:49:09 dude on irc says 2022-10-02 10:49:17 :P 2022-10-02 10:49:43 excuse me, but I'm using a web gui for irc 2022-10-02 10:49:46 document is not enough, or say, sucks, i read source code. 2022-10-02 10:49:50 that makes me less nerdy 2022-10-02 10:50:22 i don't read source code, I grep through it 2022-10-02 10:51:25 the alpine documentation is pretty casual and rough. 2022-10-02 10:51:53 but the code is pretty amazing. 2022-10-02 10:51:58 *Did you mean "nonexistent"?* 2022-10-02 10:52:21 i mean, half made. 2022-10-02 10:52:41 uncompleted to be honest 2022-10-02 10:53:02 I wouldn't call the code amazing 2022-10-02 10:54:56 mostly C and shell scripts. i don't understand the C part, but reading the shell scripts helped a lot for configuration alpine 2022-10-02 10:55:57 now i made a setup-alpine.conf and a firstlogin.sh to install alpine on the server, it takes less than 5 mins to get a full featured server. 2022-10-02 20:35:51 so regarding the issue with sata disk from yesterday ... it breaks between 3.14 and 3.15 (kernel 5.10.146 -> 5.15.71) having all packages from 3.15 except linux-lts works, so obviously some driver issue 2022-10-02 20:40:37 will anything bad happen on alpine if i add both repos (3.14 and 3.15, or maybe even 3.16) and try to pin linux-lts<5.15? so i assume it would be able to take updates from 3.14 branch so i get at least security ones for a while 2022-10-02 20:42:56 for just the kernel it's probably ok 2022-10-02 20:43:13 you could also try linux-edge, but the config is completely different 2022-10-02 20:43:19 if you don't mind losing time to it 2022-10-02 20:43:35 the 5.15 update didn't "remove" random modules or anything, so it sounds like just a regular kernel issue 2022-10-02 20:43:36 hm 2022-10-02 20:50:55 i can try to boot the installer again to see if it binds any driver to it. i'll check linux-edge if it fits on my disk still xD i'm quite tight, the upgrades made the installation go from like 200mb to like 1.8gb 2022-10-02 20:52:13 :) 2022-10-02 20:52:14 good luck 2022-10-02 20:52:28 i've been awake for like 36 hours so it's bedtime for me 2022-10-02 20:54:25 i need edge repo for that i guess? 2022-10-02 20:54:33 3.16 community has it too 2022-10-02 20:55:02 ah community 2022-10-02 20:55:41 ok 3.15 has there 5.17.9 2022-10-02 20:56:35 nope, same issue 2022-10-02 20:58:31 ah no, it ttied to boot lts again. if i select edge, it takes ages to get past grub, but it tries to boot 2022-10-02 20:59:03 1ae0591341617ef23007c24c0570d5c9937d80b9 was the 5.15 upgrade, you can look at the change to x86_64 kconfig if you know what to look for 2022-10-02 20:59:06 just in case 2022-10-02 21:05:59 ok the edge one boots exactly 30 seconds longer, so there's something timeouting on grub level 2022-10-02 21:07:28 it stays 30s more on the page with boot selection with blibking cursor before printing first lines 2022-10-02 21:10:57 dmesg, http://sprunge.us/0qCtU9 it seems the drive is detected at 32s after some tioemout, so yeah, ithat makes sense 2022-10-02 21:13:56 but hey it works don't it 2022-10-02 21:13:57 :) 2022-10-02 21:15:13 ghormoon_: what is the "disk" though? its 2GB is size, is it a SATA DOM? 2022-10-02 21:19:49 sfsa2048u1br2-c-ms-236-fts-to 2022-10-02 21:21:20 2gb msata disk in thin client (that's why i'm trying alpine for the first time, i need something that can fit there including some minimal VM and firefox :D) 2022-10-02 21:22:37 you'd probably need 16 at least 2022-10-02 21:23:34 i assume if i get something lightweight like xfce i can fit there 2022-10-02 21:29:58 ghormoon_: so it could be some "quirk" setting is missing in kernel config that affects that device 2022-10-02 21:33:14 i found some people iwt hsimmilar error using libata.force=noncq, but it has no effect on me. linux-vanilla boots like 10s, linux-edge like 41s 2022-10-02 21:36:07 are there somewhere install images with linux-edge kernel? so i can install directly 3.16 and not waste so much space? 2022-10-02 21:40:53 ghormoon_: when you add "libata.force=noncq" to the cmdline with linux-edge kernel is there any noticable difference in the dmesg output? 2022-10-02 21:41:35 no 2022-10-02 22:07:01 Hello, I'm installing Alpine Linux on my laptop and so far it's going well, but I've come to the part where I put plasma KDE on my laptop and I've hit a snag. I went with the lightweight plasma installation and began rc-update add ***, and I got an error with adding polkit. It says service polkit does not exist 2022-10-02 22:08:09 Same with polkit-kde-agent 2022-10-02 22:08:30 Is polkit-openrc installed? 2022-10-02 22:08:53 Lightweight had me install polkit-elogind 2022-10-02 22:09:19 plasma-desktop systemsettings sddm breeze elogind polkit-elogind dbus 2022-10-02 22:10:02 is there some neat trick to make it install only linux-firmware-* that are actually needed for the hw i have instead of all of them? 2022-10-02 22:10:48 polkit-openrc doesn't conflict with those. I suspect installing it will fix your rc-update issue. 2022-10-02 22:12:06 It did! Installing polkit-openrc helped me do the next step in the wiki of rc-update 2022-10-02 22:12:13 Appreciate the help ,have a good one 2022-10-02 22:13:16 No problem. =) 2022-10-02 22:14:05 ghormoon_: the latest version of the installer should already do that 2022-10-02 22:15:54 i can;t use anything over 3.14. i can upgrade to 3.16 after installing 2022-10-02 22:16:41 but if i remove linux-lts i'm at 57mb, if i add linux-edge i'm at 1gb. i assume i don't need all those firmware packages 2022-10-02 22:19:29 correct 2022-10-02 22:19:46 it used to install just all firmware packages 2022-10-02 22:20:13 but if you would add one or more specific packages, the others are automatically removed 2022-10-02 22:21:02 can i somehow run the part that detects it? or i need to do manually? 2022-10-02 22:21:35 https://gitlab.alpinelinux.org/alpine/alpine-conf/-/blob/master/setup-disk.in#L474 2022-10-02 22:23:28 can i just rerun setup-disk from the installed system or it would break things? 2022-10-02 22:34:53 ok i found a workaround, download 3.14 iso, setup-alpine, change /etc/apk/repos to 3.16, apk update, apk ugrade alpine-conf, change repo back to 3.14, setup-alpine (and setup-disk -s 0, not sure if i can somehow pass that directly to setup-alpine), then i get bootable 3.14, reboot, upgrade to 3.16, apk del linux-lts, apk add linux-edge, reboot and there i get bootable 3.16 2022-10-02 22:35:08 *repositories 2022-10-02 22:35:47 except that it boots over 40s, 3.14 boots 10s 2022-10-02 22:51:29 ok i managed to squeeze it all in including firefox with 327mb left. not sure if i won't need some more codecs for netflix though :D 2022-10-02 23:11:13 Hi! Is there a preferred VM software for running Alpine on macOS? I'm trying in VMware 22h2 preview and it's hanging on boot 2022-10-03 00:00:21 utm is probably the last worst 2022-10-03 00:00:46 afaik there is a bug, probably in vmware, where it will hang if you use the default lsilogic scsi controller 2022-10-03 00:10:47 parallels is probably more mac-centric than vmware, if you're going with something non-free. otherwise virtualbox or qemu 2022-10-03 00:27:37 s/last/least/ 2022-10-03 02:28:37 I'm on an M1 Mac, so the only real options are VMware, Parallels, and some free stuff 2022-10-03 02:28:43 sorry, M2 Mac 2022-10-03 02:31:27 I'll try UTM next 2022-10-03 04:46:56 I keep getting no such package linux-firmware-b43 2022-10-03 04:47:04 when trying to install the latest alpine linux 2022-10-03 04:47:12 this is a fresh install on this device 2022-10-03 04:47:39 Im using the setup-alpine script. Not sure why it's pulling that dependancy 2022-10-03 04:47:50 and if it is, why is it failing to find it? Can anyone help? 2022-10-03 04:48:24 linux-firmware-b43 does not exist, because we're not allowed to distribute it 2022-10-03 04:48:42 ikke: ok, I dont mind that it doesnt install it but the installer doesnt finish the script 2022-10-03 04:49:00 https://gitlab.alpinelinux.org/alpine/alpine-conf/-/blob/master/setup-disk.in#L474 2022-10-03 04:49:22 https://gitlab.alpinelinux.org/alpine/alpine-conf/-/merge_requests/104/diffs 2022-10-03 04:50:48 cheers, thank you, I will try that ikke 2022-10-03 04:51:06 not sure what that firmware is needed for.... 2022-10-03 04:52:23 looking at that script, the package name is based on module name, so `modinfo b43 | less` says "Broadcom B43 wireless driver" i.e. your wifi something or other 2022-10-03 04:55:20 oh no 2022-10-03 04:55:35 ok....I might not be able to use alpine for this wifi router =( 2022-10-03 04:56:55 https://wireless.wiki.kernel.org/en/users/Drivers/b43 2022-10-03 04:57:15 actual wifi router, or a PC with a broadcom usb wi-fi stick? 2022-10-03 04:57:47 grawity: a laptop I have lying around to use as a wifi-stick 2022-10-03 04:57:47 sorry, that's not the page I was looking for 2022-10-03 04:58:15 ah, laptop wifi hardware might not make a good access point even if you do get it working, though 2022-10-03 04:59:04 grawity: eh, it' alright. itll work better than my hodge podge of openwrt+isp provided wifi routers 2022-10-03 04:59:28 the isp provided wifi routers dont resolve dns properly 20% of the time.... 2022-10-03 04:59:38 then use the laptop for the "router" part, not for the "wifi" part 2022-10-03 05:00:36 (though, for that matter, it can be a DNS server without being a router either) 2022-10-03 05:02:51 https://wireless.wiki.kernel.org/en/users/drivers/b43/firmware 2022-10-03 05:06:07 grawity: I have a box running openwrt as my router 2022-10-03 05:06:18 the isp routers are purely for broadcasting the wifi signals 2022-10-03 05:06:30 (please let me know if this discussion is going offtopic) 2022-10-03 05:06:46 btw, after install, the wifi is working 2022-10-03 05:06:49 ACTION shrugs 2022-10-03 05:09:21 ok, then the isp routers aren't the ones resolving DNS in the first place, are they 2022-10-03 05:10:11 anjan: I don't see any issue :) 2022-10-03 08:56:21 how can I audit which packages on my system come from which repositories? say I want to know all packages installed from repo X 2022-10-03 09:06:56 ddevault: something like apk policy alpine-base ? 2022-10-03 09:07:21 that's just for one package, though 2022-10-03 09:07:34 I want the reverse: find all packages from a repo 2022-10-03 09:14:06 i think you will have to parse policy * or something 2022-10-03 09:40:04 Hi all - Looking for some advice for using knockd on a new Alpine build as part of a CTF challenge. Any pointers appreciated. 2022-10-03 09:54:28 apk does not keep track of what repository a package came from. So it can only see what repositories provide it at the moment 2022-10-03 09:54:33 Never mind - answed my own question :) 2022-10-03 09:58:10 # f () { la ;} ; alias la="ls -la" ; f 2022-10-03 09:59:06 if the alias is defined after the function which uses the alias. then function failed to run? 2022-10-03 10:00:00 # alias la="ls -la" ; f () { la ;} ; f 2022-10-03 10:00:39 sorry, wrong channel 2022-10-03 10:00:45 should be #bash 2022-10-03 10:13:32 ddevault: apk policy $(apk info) | grep -B3 $repo_url | grep -v -e -- -e '^ ' 2022-10-03 10:13:54 that'll do, thanks 2022-10-03 10:14:42 however, keep in mind ikke's disclaimer probably rings true 2022-10-03 20:08:16 Is it possible to run alpine linux live(from RAM) on a raspberry pi? 2022-10-03 20:11:33 Yes, that's the default way of running it 2022-10-03 20:21:10 Hi, I installed alpine with gnome and am having an issue removing things. I removed gnome-apps-core but I can't remove the application Tour. When I try my terminal tells me that gnome-tour can't be removed due to "gnome-tour: gnome". Any advice on ripping it out? 2022-10-03 20:22:06 tbh I'd rather be using KDE but it had some bugs so I figured it'd be easier just to switch to gnome, but the number of default apps they push on me is surprising 2022-10-03 20:23:22 Tour just seems to be a 7-picture slideshow intro to gnome, and while it was nice it doesn't seem necessary for the system to run 2022-10-03 20:23:41 nicholas: it means that the package 'gnome' pulls in tour 2022-10-03 20:23:54 and apk will not let you actually delete it 2022-10-03 20:24:47 is there a config file or something I can edit to stop it pulling tour in? Or can I install individual pieces of gnome without the package 'gnome'? 2022-10-03 20:24:54 the latter 2022-10-03 20:25:03 If you look under depends here: https://pkgs.alpinelinux.org/package/edge/community/x86_64/gnome 2022-10-03 20:25:07 you see what it installs 2022-10-03 20:25:19 you can manually install the things you want / need and remove gnome 2022-10-03 20:26:21 Great, thanks for your help! 2022-10-03 20:27:21 No problem 2022-10-03 21:56:46 When I boot I have sound as long as I keep using it. When I stop checking the sound and leave it be for a few seconds, I no longer have sound until I reboot. I installed pipewire, alsa, pipewire-alsa, and added my user and root to the audio group. Not sure what the problem is 2022-10-03 22:13:19 When I typed 'doas rc-service alsa start' I got a warning that a file was missing and I needed to unmute my card. I'm not sure what that means, Alsamixer can see my chip but I don't get audio when I change the volume controls there, and hitting M doesn't make a difference 2022-10-03 22:41:50 nicholas: Is pulseaudio in the mix at all? 2022-10-03 22:46:21 No, it wasn't. I added pulseaudio, pulseaudio-alsa, and alsa-plugins-pulse, then I rebooted. No change :( 2022-10-03 22:47:17 Ok, well I would avoid it like the plague whenever possible, because by default it dies after so many seconds. 2022-10-03 22:47:46 the other thing I can think of checking is the audio codec power management and the runtime power management for the audio card; it's rare, but some cards don't handle it properly 2022-10-03 22:47:51 having pipewire-alsa installed and using not-pipewire would give you no audio because only pipewire can be used then 2022-10-03 22:47:58 and then you have to start it 2022-10-03 22:48:19 /usr/libexec/pipewire-launcher can do that if you want, it's 3 daemons 2022-10-03 22:48:37 aside from that "working but dies after a few seconds" implies another issue[27;2;13~ 2022-10-03 22:48:57 insert keyboard escape code here 2022-10-03 22:49:08 there might be something in dmesg, random stuff like that sounds like it might print it 2022-10-03 22:49:17 aside from that something might have crashed, harder to see 2022-10-03 22:49:35 but i'd start by not installing every audio daemon at once (though you forget sndio if you want to collect them all) 2022-10-03 22:51:01 And definitely avoid daemons that exit automatically after 20 seconds (not a joke) 2022-10-03 22:51:25 lol I started with just pipewire then added more to try and get sound, I can uninstall the rest and reboot to see if helps 2022-10-03 22:54:05 gnome is unhappy and apk won't let me remove alsa-lib, alsa-plugins-pulse, pulseaudio, and pulseaudio-alsa 2022-10-03 22:56:30 Well, heaven forbid you are forced to live with pulseaudio, make sure you edit pulse-daemon.conf (or whatever they're calling it now) and give exit-idle-time a negative value so that you're not running into the same problem from that particular source. haha. 2022-10-03 22:56:35 (the default is 20 seconds) 2022-10-03 22:58:09 can you pastebin the error 2022-10-03 23:03:26 I have sound now. It was fixed after I deleted Wireplumber, which I had added during install because the PipeWire page said to install pipewire and wireplumber together 2022-10-03 23:03:45 ah, happens 2022-10-03 23:03:50 I guess if wireplumber was the problem that's why i never had sound 2022-10-03 23:03:53 means it got replaced by pipewire-media-session 2022-10-03 23:04:03 though there's probably more to it 2022-10-03 23:04:11 like wireplumber didn't start for some reason, etc 2022-10-03 23:04:19 i wonder what started it to begin with 2022-10-03 23:05:46 So if I have pipewire I don't need alsa or pulseaudio? Every gnome apk I have says they need those and won't let me uninstall, so I'll probably just uninstall the rest of gnome and see if it's still a problem on reinstall. Or try KDE again 2022-10-03 23:06:05 nope 2022-10-03 23:06:15 can you just post the error 2022-10-03 23:06:32 "World updated, but the following packages are not removed due to: 2022-10-03 23:07:02 alsa-lib: gnome-shell gdm gnome-control-center network-manager-applet xdg-desktop-portal-gnome gnome-tweaks konsole pipewire firefox 2022-10-03 23:07:30 alsa-plugins-pulse: gdm gnome-shell gnome-control-center network-manager-applet xdg-desktop-portal-gnome gnome-tweaks konsole pipewire firefox 2022-10-03 23:07:46 pulseaudio: gnome-tweaks gnome-shell gdm 2022-10-03 23:08:11 alsa-lib is not alsa and always needed regardless for anything that uses the alsa api 2022-10-03 23:08:20 pulseaudio-alsa: gdm gnome-tweaks gnome-shell gnome-control-center network-manager-applet xdg-desktop-portal-gnome konsole pipewire firefox" 2022-10-03 23:09:32 as for alsa-plugins-pulse/pulseaudio-alsa `apk add pipewire pipewire-pulse` and the rest should vanish 2022-10-03 23:11:50 Nice, thanks 2022-10-04 01:43:20 I'm on KDE and have switched from wpa_supplicant to iwd. My internet is connected, it automatically connects after reboot, no issues there. But my plasma gui always is red and says I'm disconnected and isn't able to see any networks in my area. I've changed network manager's config file to have wifi.backend=iwd so I'm not sure why they aren't talking to each other 2022-10-04 01:44:20 I'm unable to do any network managing through networkmanager GUI. I checked my rc-status and it says networkmanager is running and does not mention any iwd or wpa_supplicant, which I think is how it should be 2022-10-04 01:44:41 I'm not seeing any error messages either 2022-10-04 01:45:53 for context I don't think my network manager has worked at all since original install, even when I was on gnome and using wpa_supplicant 2022-10-04 01:46:59 you have to make sure /etc/network/interfaces doesn't have any interfaces in it for networkmanager to manager them 2022-10-04 01:48:10 I can't cd into /etc/network/interfaces it says not a directory 2022-10-04 01:49:32 which is strange because when I ls inside /etc/network I can see interfaces there 2022-10-04 01:49:46 it's file. 2022-10-04 01:49:49 It's a text file* 2022-10-04 01:50:31 use `ls -l' to see attributes... directories will start with d 2022-10-04 01:52:12 gotcha, I opened the file with nano. There are 4 lines, one says 'auto wlan0' and the only line under it says 'iface wlan0 inet dhcp' is that the problem? 2022-10-04 01:53:14 sure 2022-10-04 02:01:13 I deleted everything in the file, rc-service networkmanager restart, and rebooted for good measure. Wifi connection is gone but network manager still seems to think I live on an uninhabited asteroid 2022-10-04 02:01:29 did you add networkmanager-wifi 2022-10-04 02:01:40 how's the rent there? 2022-10-04 02:02:53 How high do superjumbo jets fly? 2022-10-04 02:03:11 considering its great location, wonderful unobstructed view of the stars, and lack of an HOA it's surprisingly low 2022-10-04 02:04:10 I did not add networkmanager-wifi, didn't know it was a thing. I will figure out how to restore my wifi and add it 2022-10-04 02:05:21 you would have to unedit the files 2022-10-04 02:05:33 ooor just use the iwctl cli i guess 2022-10-04 02:05:38 ah 2022-10-04 02:05:45 but you'd have to start a dhcp by hand 2022-10-04 02:05:48 eh, edit the files 2022-10-04 02:06:10 good thing I wrote it in the chat before I deleted it 2022-10-04 02:19:24 hooooly canoli that's all I needed 2022-10-04 02:21:13 I restored the file, connected to my network with iwctl, restarted services, added the networkmanager-wifi, and bam. Wifi. Rebooted a few times just to be safe, and it's working great. Thanks for your help 2022-10-04 02:23:02 :) glad it works 2022-10-04 02:23:32 should I add the networkmanager-wifi to the wiki or does it usually come packaged with other things 2022-10-04 02:23:46 (so this isn't usually an issue) 2022-10-04 02:26:33 i think it is isn't it 2022-10-04 02:26:38 or did it only get a migration notice 2022-10-04 02:26:38 hm 2022-10-04 02:26:47 you can add whatever if you want 2022-10-04 02:36:56 done, thanks for all your help. Have a good one 2022-10-04 09:40:29 say I want to apply a trivial patch to the gtk3 package for my personal usage but still keep the package up to date with new versions from the official repositories. 2022-10-04 09:40:56 what would be the best way for me to manage version numbers for my locally patched APKBUILD? 2022-10-04 10:22:34 Probably a high pkgrel number and pin the version? 2022-10-04 10:26:33 ah, yeah that makes sense, thanks 2022-10-04 13:58:47 i pinned linux-virt=5.19.71-r0 2022-10-04 14:23:53 *5.15.71-r0 2022-10-04 16:31:17 does musl libc use /etc/protocols or it's own database? I'm trying to compile netrek-client-cow (C program) and getprotobyname("TCP") fails where getprotobyname("tcp") succeeds 2022-10-04 16:34:29 I'm writing a patch file to fix it in the program for my APKBUILD, but just curious as to why and what other issues I might run into 2022-10-04 16:39:20 I see its own DB here, is why I'm asking: http://git.musl-libc.org/cgit/musl/tree/src/network/proto.c 2022-10-04 16:39:50 it's only lowercase in proto.c 2022-10-04 16:40:05 There is also #musl for musl related questions :) 2022-10-04 16:41:11 was curious if its an Alpine issue or a Musl issue 2022-10-04 16:44:07 or a PEBCAK issue :) 2022-10-04 16:59:06 kwadroke[m]: nice PEBCAK 2022-10-05 10:51:46 anyone here using gopass? I noticed more than 320k "--- SIGURG {si_signo=SIGURG, si_code=SI_TKILL, si_pid=23205, si_uid=1000}" on strace 2022-10-05 10:52:08 I'm not sure if it's related with https://github.com/golang/go/issues/37942 2022-10-05 10:52:42 which should't affect go 1.19 2022-10-05 10:54:59 i don't see why it wouldn't affect go 1.19 2022-10-05 10:55:52 hmm, I missunderstood 'This program, when running under go 1.14, ' 2022-10-05 11:09:04 well, with http://ix.io/4cki it went to ~22k 2022-10-05 11:09:37 and I see the gopass prompt, the other run was interrupted before showing it so probably there are more than 320k 2022-10-05 17:42:33 In your name, O God There is no God except you, the Greatest, the one, no partner exists to you 2022-10-05 17:42:33 O people, Avoid blasphemy and polytheism in God, which is infact taking equals with God, and nullifying all deeds and all that what has been given to man, so that you may escape from God’s eternal punishment 2022-10-05 17:42:48 The soldiers and participants of the ruling oppressors regimes all over the world now are required: 2022-10-05 17:42:48 To leave the government and leave army and leave their positions and to create strife, confusion and revolutions against the ruling regime in your country now using all means possible, otherwise they will be tormented with a greater punishment from God the redeemer and through the Messenger of God, Moses the Greatest . 2022-10-05 17:56:09 is alpine switching to yash as default? 2022-10-05 18:06:37 no 2022-10-05 18:06:54 dinit looks interesting 2022-10-05 18:07:11 https://github.com/davmac314/dinit 2022-10-05 19:00:21 So, I have heard some rumors that Alpine is going to switch away from sysvinit (or was it OpenRC?) in the near future. Is there a page or something where I can read more about it. 2022-10-05 19:25:32 townsfolkPravda: There is no concrete proposal yet. Before we can decide to switch, we first need something to switch to. Work is being done on an alternative (s6), but that's still in progress 2022-10-05 19:26:40 ikke, So the idea is to possibly switch OpenRC to something else? 2022-10-05 19:26:49 yes 2022-10-05 19:26:52 ikke, Ok, thanks. 2022-10-05 19:29:21 psykose: okok, heard a rumor 2022-10-05 19:29:32 there are proposals 2022-10-05 19:42:10 Is there some kind of problem with OpenRC? Or is s6 just better? 2022-10-05 19:50:03 development is mostly geared towards gentoo 2022-10-05 19:50:16 is one issue 2022-10-05 19:59:47 townsfolkPravda: trying to find skarnets overview of init systems 2022-10-05 20:00:03 it used to be on skarnet.com, but it seems it's no longer there 2022-10-05 20:00:16 ooh 2022-10-05 20:00:19 https 2022-10-05 20:00:42 https://skarnet.com/projects/service-manager.html#comparison 2022-10-05 20:02:33 ikke, thanks! Reading that page now... 2022-10-05 20:54:29 is there a tool like pmbootstrap for regular alpine? would be cool to make customized live iso 2022-10-05 21:02:44 ikke: there is one piece of comparison too https://raw.githubusercontent.com/davmac314/dinit/master/doc/COMPARISON 2022-10-05 21:08:04 artix linux uses like 3 service management, s6, openrc and dinit 2022-10-05 21:10:07 i took at look at the dhcpcd service daemon from their mirror, looks like dinit is very user friendly very similar to systemd's configuration. A=B style 2022-10-05 21:19:55 is there a way to see last and lastb? i installed util-linux-login, but looks like it needs extra configuration to record the login history 2022-10-05 21:22:43 setup-utmp and reboot 2022-10-05 21:22:47 some things still don't work iirc 2022-10-05 21:23:02 but last should from util-linux 2022-10-05 21:24:35 last/lastb is from util-linux-login. 2022-10-05 21:26:07 setup-utmp is from utmps which is from skarnet.org 2022-10-05 21:32:01 sudo lastb | wc -l = 82357 2022-10-05 21:32:21 on one of my server. i really shoudn't use default 22 2022-10-05 21:34:20 why not 2022-10-05 21:36:38 i remember one time sshd uses 30% of my cpu, i got attacked 2022-10-05 21:37:09 changing the port will not meaningfully improve that 2022-10-05 21:37:42 if you want to eliminate issues with attacks, using something like fwknopd is best 2022-10-05 21:37:50 reducing the DDOS attack maybe 2022-10-05 21:37:53 nope 2022-10-05 21:38:11 with services like shodan, it's easy to find ssh servers regardless of port 2022-10-06 01:41:03 I'm looking to add neofetch to start on new window of my terminal. Where previously I would add to /bashrc, I'm seeing a note in that file to put my configs in /etc/profile.d/*.sh However, when I open that and add neofetch to the end of the file, neofetch does not open when I start a new terminal. What am I missing? Thanks 2022-10-06 01:42:18 Notably when I type "source /etc/profile.d/*.sh" neofetch immediately runs in that window of the terminal, but when I close that terminal and open a new one it does not appear 2022-10-06 02:47:59 all the guides I can find talk about adding "neofetch" to ~/.bashrc but I don't seem to have that, making it doesn't do anything, and adding "neofetch" to the end of /etc/profile.d/*.sh doesn't do anything either 2022-10-06 04:36:11 nicholas: The profile is only sources for login shells 2022-10-06 04:36:45 What you can do is define ENV to be a file that ash sources for non-login shells as well 2022-10-06 04:38:50 *nod* 2022-10-06 04:39:29 my .profile has export ENV="$HOME/.env" ; . "$ENV" 2022-10-06 08:08:45 identify qwetyDirtyIRC123890 AppleME 2022-10-06 08:08:54 oh s**t 2022-10-06 12:19:37 besides changing the password, I suggest adding an alias to your irc client, so that - worst case - you post only your alias instead 2022-10-06 12:45:51 ACTION saw identify ******************* AppleME there 2022-10-06 12:47:33 this worked once :P 2022-10-06 13:41:33 Hi everyone, the libtiff library has some vulnerability in the version available (4.3.0) for alpine 3.16. Would it be possible to update this library to version 4.4.0 which is available for edge? 2022-10-06 13:53:20 what vulnerability is it 2022-10-06 14:00:58 https://nvd.nist.gov/vuln/detail/CVE-2022-2868 2022-10-06 14:01:02 https://nvd.nist.gov/vuln/detail/CVE-2022-2869 2022-10-06 14:01:05 https://nvd.nist.gov/vuln/detail/CVE-2022-2867 2022-10-06 14:04:13 ooh, fuzzer spam cves 2022-10-06 14:04:16 sure, gimme a second 2022-10-06 14:48:08 great, thank you 2022-10-06 15:06:35 trying out yash, seems default path is only '/bin:/usr/bin', and my xorg crashes ... anyone know why xorg would crash? 2022-10-06 15:09:50 Check Xorg logs 2022-10-06 15:10:04 Or run it with full PATH 2022-10-06 15:12:44 with full path it works ja 2022-10-06 15:19:27 frag: why not dash 2022-10-06 15:20:22 i don't know who brought up the yash idea, but why not dash. 2022-10-06 15:20:49 uuidNuniq: why dash? 2022-10-06 15:21:11 small, light weight, posix... 2022-10-06 15:21:12 yash seems to work well out of the box, and kinda fell in love 2022-10-06 15:21:53 AND seems to be quite good for scripting too 2022-10-06 15:22:26 for GOOD, you mean, it has extra features other than posix? 2022-10-06 15:23:02 ah i dunno how shells are posix or not tbh, but i guess its extra 2022-10-06 15:23:15 It depends on the usecase 2022-10-06 15:23:40 But many things in Alpine expect an extended posix shell, not strict posix 2022-10-06 15:23:52 https://0x0.st/oyPI.txt 2022-10-06 15:24:35 so tiny! 2022-10-06 15:24:45 uuidNuniq: you use dash as loginshell? 2022-10-06 15:25:12 why no busybox ash?? 2022-10-06 15:25:16 ikke: most scripts in apline are based on busybox ash right? 2022-10-06 15:25:35 frag: ash is not a separate package 2022-10-06 15:25:54 uuidNuniq: implicitly, yes 2022-10-06 15:25:58 so its kinda 0 :p 2022-10-06 15:26:12 then dash should work out of box. 2022-10-06 15:26:26 The issue with ash is that we ran into bugs that were very difficult to debug due to the way ash is written 2022-10-06 15:26:37 uuidNuniq: no, dash is strictly POSIX afaik 2022-10-06 15:27:45 im more talking about convience like history/completion etc working out of the box 2022-10-06 15:27:58 dash is not really meant as a login shell 2022-10-06 15:28:48 actually history/completion works in ash.. 2022-10-06 15:28:58 frag: well, for login shell, i use zsh. 2022-10-06 15:29:24 with zimfw framework :) 2022-10-06 15:30:00 uuidNuniq: thats what ive been using too, when i tried yash i noticed big difference in startup time (slow computer, could be my bloathed zsh config tho...) 2022-10-06 15:30:10 there is nothing that is more powerful then zsh on login shell or scripting. 2022-10-06 15:30:33 yeah i cant soon delete zsh :p 2022-10-06 15:30:52 frag: zsh seems fine on my Amlogic S905d soc 2022-10-06 15:31:35 uuidNuniq: never heard of, im on rpi4 2022-10-06 15:32:02 it's A53 4core. much weaker than rpi4's A72 2022-10-06 15:32:20 :o 2022-10-06 15:32:32 prob my bloathed zsh config then :| 2022-10-06 15:32:48 bunch of aliases/functions and stuff, dunno what makes it slow... 2022-10-06 15:33:10 maybe, so i just installed a light weight framework called ZIMFW 2022-10-06 15:33:12 now even my PATH seems extremely bloated :p 2022-10-06 15:33:54 duplicated items in $PATH? you mean 2022-10-06 15:34:32 yes actually, but for yash its just '/bin:/usr/bin' i think it was ... 2022-10-06 15:34:48 Code of conduct has been updated: https://alpinelinux.org/community/code-of-conduct.html 2022-10-06 15:35:17 frag: /bin:/usr/bin is fine to be honest. i think every bin should be in just /usr/bin 2022-10-06 15:35:27 or just in /bin 2022-10-06 15:36:15 uuidNuniq: that would be nice! 2022-10-06 15:36:25 i have a function to remove duplicated items in path on zsh. 2022-10-06 15:36:40 ncopa: is there a diff available from the previous version? 2022-10-06 15:37:11 frag: https://0x0.st/oyZf.txt 2022-10-06 15:37:15 i just write it out like: export PATH=/bin:/usr/bin:/usr/local/bin:/usr/sbin:/sbin:/1/z in my new .yashrc, i think that should solve it 2022-10-06 15:40:55 is ncopa the creater Natanael Copa ? 2022-10-06 15:41:10 Yes 2022-10-06 15:41:15 hi 2022-10-06 15:41:32 but please be aware that there are two Natanael Copa on telegram and one is not me :) 2022-10-06 15:41:40 you'd doing fantastic jobs. 2022-10-06 15:42:06 ncopa: i see a mention of political attacks in this coc. do political attacks include an expression of support for one or other side of the infowars? 2022-10-06 15:42:07 thanks, but the fantastic job is done by many others nowdays 2022-10-06 15:42:11 you should probably have a @ prefix 2022-10-06 15:43:19 i thought i had... 2022-10-06 15:45:31 autoop is on, but for some reason not applied 2022-10-06 15:45:58 ikke: thanks! 2022-10-06 15:46:21 You can always ask chanserv for op 2022-10-06 15:46:27 better now. 2022-10-06 15:47:42 grisha: I guess it depends. In my mind "attacks" and "expression of support" is not the same thing 2022-10-06 15:48:31 ncopa: is there a diff available from the previous version of the COC? 2022-10-06 15:48:43 yeah i guess they are a little different. i just wanted to clarify) 2022-10-06 15:49:07 minimal: a diff would not make a lot of sense 2022-10-06 15:49:15 it's a completely different text 2022-10-06 15:49:48 minimal: yes: https://gitlab.alpinelinux.org/alpine/infra/alpine-mksite/-/commits/master/community 2022-10-06 15:49:58 ikke: completely? ok, wasn't aware 2022-10-06 15:50:24 minimal: We first had something that was written by ourselves, now we adopted the contributor covenant 2022-10-06 15:50:46 Hello everyone ! Is it possible to install packages through APK without the system actually running ? 2022-10-06 15:51:16 zaghp: sure 2022-10-06 15:51:22 My point is, starting from the Alpine minirootfs, can I install a few packages from another system ? (ie. like debootstrap on debian) 2022-10-06 15:51:36 yes, via chroot. 2022-10-06 15:51:40 zaghp: apk --root path/to/root 2022-10-06 15:51:44 or that. 2022-10-06 15:52:01 If you need to run it on a non-alpine system, you can use the static apk binary 2022-10-06 15:52:03 ikke: Hmm, so I need the "apk" executable anyway 2022-10-06 15:52:08 yes 2022-10-06 15:52:18 That's great 2022-10-06 15:52:25 Thank you, I'll give it a try 2022-10-06 15:52:43 zaghp: https://gitlab.alpinelinux.org/alpine/apk-tools/-/releases 2022-10-06 15:52:58 static binaries for all supported arches 2022-10-06 15:54:10 Oh, very nice, I'll try those, then 2022-10-06 15:54:14 Thanks a lot 2022-10-06 15:54:28 Win up to $1000 in crypto trading when you invest with just the minimum of $50... (full message at ) 2022-10-06 15:55:56 it is amazing that i have all the base tools installed but the size is only OK: 220 MiB in 276 packages 2022-10-06 15:57:32 https://0x0.st/oyN1.txt 2022-10-06 16:01:34 is ash and sh in busybox the same? 2022-10-06 16:01:42 yes 2022-10-06 16:03:20 the command prompt string is different?? 2022-10-06 16:03:49 (but not is i run ash first and the sh from ash ..) 2022-10-06 16:04:30 *then 2022-10-06 16:04:34 i think if you use yash then, the /bin/sh symlink needs an update 2022-10-06 16:04:52 uuidNuniq: im testing busybox ash/sh now :] 2022-10-06 16:05:15 kinda suprised at how well it works out of the box tbh 2022-10-06 16:05:38 ash reads ~/.profile 2022-10-06 16:07:04 command prompt you mean the $PS1 ? 2022-10-06 16:07:27 but ash does not seem to have documents 2022-10-06 16:08:49 uuidNuniq: yes, and hard to find documentation indeed 2022-10-06 16:10:13 when i need docs for ash, i read dash-doc 2022-10-06 16:10:57 i don't know how much differences there. but i guess 90% in similarity 2022-10-06 16:12:00 uuidNuniq: good idea :] 2022-10-06 16:17:37 frag: ash ; export PS1="\[\e[1;32m\]\u@\h\[\e[0m\]:\[\e[1;34m\]\$(pwd) \[\e[0m\]-> " 2022-10-06 16:17:45 now it's colorful 2022-10-06 16:18:23 uuidNuniq: just like yash hehe 2022-10-06 16:37:04 is there no place to store aliases for sh/ash then? 2022-10-06 16:38:38 frag: if you export ENV in your profile, that gets sourced for every shell 2022-10-06 16:38:43 (every non-login shell) 2022-10-06 16:39:09 $HOME/.profile gets sourced for login shells, as is /etc/profile 2022-10-06 16:42:20 ikke: hmm, im not familiar with ENV, what does it have to do with aliases? 2022-10-06 16:43:49 frag: You define the file that corresponds with .bashrc in bash 2022-10-06 16:44:39 ikke oh! 2022-10-06 16:45:08 But in many cases, adding it to $HOME/.profile works as well 2022-10-06 17:36:01 frag: echo '[ -f ~/.alias ] && . ~/.alias' >> ~/.profile 2022-10-06 17:37:34 uuidNuniq: thanks :) 2022-10-06 17:37:52 ash by default reads ~/.profile and /etc/profile only. nothing else. 2022-10-06 17:38:15 so you got to source your custom settings from there, as many as you want 2022-10-06 17:41:38 yeah, pretty cool, learning today! 2022-10-06 17:44:57 [ -f ~/.alias ] && export ENV="~/.alias" also works i think 2022-10-06 18:08:37 http://landley.net/toybox/bin/ 2022-10-06 20:18:34 In your name, O God There is no God except you, the Greatest, the one, no partner exists to you 2022-10-06 20:18:34 Avoid: 2022-10-06 20:18:34 O people, Avoid the following matters that are blasphemy and polytheism in God, which is infact taking equals with God, and nullifying all deeds and all that what has been given to man, so that you may escape from God’s eternal punishment 2022-10-06 20:18:35 1- Glorifying for those other than God as glorifying God 2022-10-06 20:18:35 2- Judging by injustice, empowering the oppressors, appointing them to rule, or preparing them for people to rule, or ruling the opposite of what God the great has revealed, or doing injustice and tyranny; 2022-10-06 20:18:35 For Injustice is darkness, and as you do, it is received by you, asfor tyranny is neglection; that is ending yourself by the redeemer asfor the redeem does not die. 2022-10-06 20:18:48 3- Taking refuge to other than God as you take refuge (resorting) to God, and supplicating to other than God, i.e. requesting and seeking help from other than God, in matters that God alone has power to do. All of that is blasphemy 2022-10-06 20:18:49 and polytheism with God the Greatest, {and nullifies the work of man and what he has been given}, and it is the habit of monsters, that is, devils, not humans; and hence, your great contribution to [the ruling of the evil that is going on now] because of you - will return on you if you don't quit now, and your rivers of blood will return red .... on you 2022-10-06 20:19:01 The soldiers and participants of the ruling oppressors regimes all over the world now are required: 2022-10-06 20:19:02 To leave the government and leave army and leave their positions and to create strife, confusion and revolutions against the ruling regime in your country now using all means possible, otherwise they will be tormented with a greater punishment from God the redeemer and through the Messenger of God, Moses the Greatest ( ZRZ ). 2022-10-06 20:19:02 Moses of Elohim 2022-10-06 20:26:43 interesting 2022-10-06 23:47:07 hey, the docs link on alpinelinux.org homepage is 404'd. just thought i'd report that incase nobody has already 2022-10-06 23:47:25 nvm it apparently just got fixed? 2022-10-07 04:56:43 Hi guys, how are you? 2022-10-07 04:56:43 rc-update add local default, rc-service local start, put the file in /etc/local.d/test.start, made it executable, restarted and it didn't work. Any tips? 2022-10-07 04:56:43 Well, I'm trying to run a script in /etc/local.d but it's not working. I followed all the steps in the documentation, such as: 2022-10-07 04:57:12 what does the script start with? 2022-10-07 04:58:44 start a nodejs script with pm2, like pm2 start main.js 2022-10-07 04:58:58 I mean, what is the first line of the file? 2022-10-07 04:59:03 no docker 2022-10-07 05:00:21 #!/bin/sh 2022-10-07 05:00:36 usr/local/bin/pm2 start /root/server/node/abnr/main.js 2022-10-07 05:00:43 that should be right 2022-10-07 05:01:11 does it work when you do rc-service local start? 2022-10-07 05:01:12 exactly as it is in the script 2022-10-07 05:01:22 yes 2022-10-07 05:01:36 ops, no 2022-10-07 05:02:37 eu tentei isso: rc-service local restart -v 2022-10-07 05:02:55 no error in the output 2022-10-07 05:03:22 but still it didn't work 2022-10-07 05:03:28 does it say "starting local"? 2022-10-07 05:04:54 yes 2022-10-07 05:05:36 but not "Execuring test.start"? 2022-10-07 05:06:52 starting and executing 2022-10-07 05:07:27 but when I reboot, nothing happens 2022-10-07 05:07:46 I wonder if it's missing some environment that this script needs 2022-10-07 05:09:17 when I do it via command line, the script runs normally 2022-10-07 05:09:31 as root? 2022-10-07 05:10:21 yes 2022-10-07 05:10:59 and as normal user too 2022-10-07 05:11:09 I would guess that it's missing some environment variable 2022-10-07 05:11:28 the line with usr/local/bin/pm2 start has / before usr, right? 2022-10-07 05:12:51 yes 2022-10-07 05:13:34 '/usr/local/bin/pm2 start /root/server/node/abnr/main.js' 2022-10-07 05:13:43 weird. I think the most oikely explanation is still that it's missing some environment variable. printenv should tell you what your environment is 2022-10-07 05:14:56 ok, I will check and test 2022-10-07 05:15:23 thanks for the help 2022-10-07 08:09:40 is it hard to install alpine on a computer thats not on the internet? (and the packages you want) 2022-10-07 08:11:27 the extended iso should be able to install offline 2022-10-07 08:12:00 as for the packages you want you can download them to an external drive and install them locally 2022-10-07 08:13:48 Misthios: possibly the new packages i donwload would be newer version thatn the install img tho? 2022-10-07 08:14:38 one option would be to just download all the packages onto an external drive and edit the mirror file to the usb during install 2022-10-07 08:15:23 just using the installs packages should be fine if you use stable 2022-10-07 08:15:35 perhaps few minor updates 2022-10-07 08:16:21 okok :] 2022-10-07 09:39:21 could swear a small function in .profile i made for sh has been working, suddenly stop working untill i do '. .profile' or 'sh -l', what could cause this? 2022-10-07 09:39:46 i use sh as login-shell also now, and have logged out/in 2022-10-07 09:41:02 ei i get: "/bin/sh: z: not found" 2022-10-07 09:41:11 what package name do we use for fonts? `ttf-` or `font-`? I see it's being used both 2022-10-07 10:34:57 I'm not able to get shared folders or copy and paste between host and guest working in Alpine with VMware Workstation. I installed open-vm-tools following the guide on the wiki, but when I add a shared folder with the VMware GUI it doesn't appear in Alpine. I added vmhgfs-fuse /mnt/hgfs fuse defaults,allow_other 0 0 to /etc/fstab and rebooted, /mnt/hgfs is still an empty directory. Windows 11 host, VMwareยฎ Workstation 16 Pro 16.2.4 build-20089737, 2022-10-07 10:34:57 Alpine Linux 3.16.2 Virtual. 2022-10-07 10:46:07 zaphida: have you tried manually mounting it? 2022-10-07 10:48:12 does anyone know how I could route some app with different gateway without using containersORvm's ? maybe some fancy iptables 2022-10-07 10:51:07 vrfs probably 2022-10-07 10:51:12 never used them myself though 2022-10-07 10:52:07 (on linux) 2022-10-07 10:52:31 hmm, it looks promising, thanks :) 2022-10-07 10:56:57 donban just tried with mount -t vmhgfs .host:/ /mnt/hgfs and got mount: mounting .host:/ on /mnt/hgfs failed: No such device 2022-10-07 10:57:04 donoban 2022-10-07 10:57:42 does .host resolves to something? 2022-10-07 10:58:05 I don't know how vmhgfs works, can it use an ip address? 2022-10-07 10:59:22 vmhgfs is only the driver that supports shared folders directly between the host and a virtual machine. no it doesn't use IP address 2022-10-07 11:10:18 there is something on dmesg? 2022-10-07 11:48:09 ikke: this is too much complex for me 2022-10-07 11:48:24 I'm thinking to just use docker or try lxc/lxd 2022-10-07 12:02:05 https://elou.world/en/kubernetes/crio-alpine-error-127 2022-10-07 12:05:01 uhM, I wonder why it needs chroot() for just install a package 2022-10-07 12:05:32 he is installing v3.12 2022-10-07 12:10:54 it's on apk_db_run_script() 2022-10-07 12:13:42 nothing on dmesg that looks like another drive connected 2022-10-07 12:36:42 donoban: i think it says alpine needs chroot 2022-10-07 12:36:51 apk needs chroot to install pkg 2022-10-07 12:57:06 when i open a terminal with my hotkeys, 'printf "%s" $SHLVL' gives me "5", does it matter or not? (i guess i use a shellscript to open hotkeys...........) 2022-10-07 13:04:16 frag: you are in subshell 4 2022-10-07 13:05:22 is it like the nine circles of hell? j/k :p 2022-10-07 13:05:29 frag: pstree -a 2022-10-07 13:07:27 mhmhm, maybe i could cut some shells out :/ 2022-10-07 13:11:10 i don't think it matters. 2022-10-07 13:11:50 i always always start a tmux when open any terminal or any ssh connection. 2022-10-07 13:23:50 uuidNuniq: it's only needed if --root option is passed, it seems that the author is based on some bugged version 2022-10-07 13:24:24 I tried to reproduce http://ix.io/4cw8 2022-10-07 13:25:19 anyway changuing default capabilites doesn't look a good solution 2022-10-07 13:33:32 oh, i see donoban thanks 2022-10-07 13:40:50 are you using CRI-O? 2022-10-07 13:42:45 Exagone313: are you the author? could you check it again? 2022-10-07 13:43:57 donoban: no, i am not, just found that interesting. 2022-10-07 13:44:14 the strace part 2022-10-07 13:45:37 it's curious that it let's to run strace but not chroot. When I tried with docker the strace failed due permissions 2022-10-07 15:43:55 free 2022-10-07 15:45:30 Hello together. When is the next Release 3.16.3 planned? 2022-10-07 15:47:22 Guest2574: it is not planned yet 2022-10-07 15:47:35 Usually it happens when there is a specific need for it 2022-10-07 15:49:17 I am looking for a security patch for the CVE 2022-10-07 15:49:20 โ”Œโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”ฌโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”ฌโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”ฌโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”ฌโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”ฌโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ” โ”‚ Library โ”‚ Vulnerability 2022-10-07 15:50:06 seems to be already patched in edge-release but not yet as stable release 2022-10-07 15:50:27 What CVE? You didn't paste any details 2022-10-07 15:50:29 For which CVE? 2022-10-07 15:50:44 CVE-2022-40674 2022-10-07 15:50:56 libexpat 2022-10-07 15:51:01 https://security.alpinelinux.org/vuln/CVE-2022-40674 2022-10-07 15:51:15 Guest2574: is already fixed in 3.16, no released necessary for that 2022-10-07 15:51:20 release* 2022-10-07 15:52:14 d36d9eadf3b6c62f3d74c38b595f4322cd88851c 2022-10-07 15:52:27 https://gitlab.alpinelinux.org/alpine/aports/-/commit/d36d9eadf3b6c62f3d74c38b595f4322cd88851c 2022-10-07 15:53:02 This one for 3.16: https://gitlab.alpinelinux.org/alpine/aports/-/commit/85cc4d8bfc2d300c3628c60f584ec4d734d893f4 2022-10-07 15:57:22 ok thanks. did not see that. need to check the built-up container 2022-10-07 16:55:53 Hello, I am new to linux (altough I do understand the basics about package managers, distros, terminal and desktops etcetera), and alpine Linux has an interesting name in my opinion. 2022-10-07 16:55:53 So might I ask what makes it different from other distro's? 2022-10-07 16:56:55 It's relatively small and light-weight 2022-10-07 16:57:08 ah 2022-10-07 16:58:06 Some people like that alpine does not use systemd 2022-10-07 16:58:36 I did hear that name before, what does it mean? 2022-10-07 16:59:14 systemd is an init system, the components that are involved with starting everything after the kernel runs 2022-10-07 16:59:38 The most visible part is taking care of services 2022-10-07 17:02:29 oh okay 2022-10-07 17:03:07 is that better with something else than systemd? 2022-10-07 17:06:21 That's subjective 2022-10-07 17:10:49 ah 2022-10-07 17:11:21 is systemd closed source, that people dont like it? 2022-10-07 17:12:37 no, it's open source 2022-10-07 17:13:01 It's a large project that tries to do a lot on it's own 2022-10-07 17:14:46 nope 2022-10-07 17:19:54 thanks for the explanation ikke 2022-10-07 18:06:32 xxxsmmsnns: Alpine also strives to be "minimal" - with most distributions, if you install a new service (ssh, samba, postgres, etc) they generally autostart the service as part of the install. 2022-10-07 18:07:01 Alpine does not do that. You only run what you, as the admin, decide to run; even if the "package" is installed 2022-10-07 18:07:50 I'm minimal and so is my OS! ;-) 2022-10-07 18:08:27 LOL 2022-10-07 18:09:11 nangel: they left. 2022-10-07 18:25:04 very similar to arch (without systemd) 2022-10-07 18:28:16 but alpine has stable tree which i value a lot. a decent Debian replacement 2022-10-07 18:32:21 since this conversation started, may i ask if alpine was inspired by arch or vice versa? i as a previous arch user noticed a lot of similarities during the transition (similar similar install scripts, package recipes and logo, hah) 2022-10-07 18:33:22 I've sure noticed that the Arch wiki is so incredibly helpful; and a lot of the stuff there applies to other distros. 2022-10-07 18:40:30 For a diskless install of alpine, where/when does the EFI partition get mounted? 2022-10-07 19:00:31 grisha: Alpine was started on its own. Arch and Alpine seem to be "twins separated at birth", but they had different starting points. 2022-10-07 20:34:36 is .ash_history ever pruned or will it grow forever? 2022-10-07 20:42:37 frag: try export HISTSIZE=5 for a test ? if works increase to 50000 or whatever you like 2022-10-07 20:43:03 backup before testing! 2022-10-07 20:46:11 uuidNuniq: doesnt seem to do anything 2022-10-07 20:59:02 frag: then you need a cron job. sort uniq tail... :) 2022-10-07 21:05:00 uuidNuniq: not a bad idea haha 2022-10-07 21:07:21 Hi :) 2022-10-07 21:08:05 Is there a fix for the gnome-calendar segfault issue? 2022-10-07 21:09:55 Does busybox ln support -t option? 2022-10-07 21:14:40 Ermine: -T Treat LINK as a file, not DIR 2022-10-07 21:14:57 Forza: I mean t small 2022-10-07 21:15:07 according to a source, no. 2022-10-07 21:15:09 It's not listed 2022-10-07 21:15:35 It is listed in the usage string. 2022-10-07 21:18:46 Yea. Maybe a bug 2022-10-07 21:19:08 I'd say a typo. 2022-10-07 21:21:25 this is debian, but I'd expect there not to be a significant difference: 2022-10-07 21:21:25 $ busybox ln -t tmp . test 2022-10-07 21:21:26 ln: invalid option -- 't' 2022-10-07 21:37:42 Ermine: busybox ln --help 2022-10-07 22:05:40 I'm installing Alpine KDE on my device, and it went well up until I tried logging in at the end via SDDM into XOrg. I went from the SDDM login page to black screen. I'm wondering whether this is because I have a NVIDIA gpu. I found https://wiki.alpinelinux.org/wiki/Radeon_Video and https://wiki.alpinelinux.org/wiki/Nvidia_Video, but I'm using XOrg because it's KDE 2022-10-07 22:06:29 I installed xf86-video-ati like the first link said as well as the mesa driver just in case, but nothing 2022-10-07 22:08:08 I tried to do "apk add modprobe fbcon" but it failed. I also found a github post (https://github.com/NVIDIA/nvidia-docker/issues/1606) that I think means NVIDIA doesn't work on Alpine Linux. Is this distro doable for me or should I wait to switch to Alpine until after I get an AMD GPU 2022-10-07 22:35:53 nvidia isn't very usable without the proprietary drivers which you can't use on musl 2022-10-07 22:36:59 depends on how well-supported your hardware is with nouveau, though 2022-10-07 22:39:57 oof, and nouveau is installed with the mesa package right? So if I've already installed it and it isn't working, i'm SoL on this PC 2022-10-07 22:41:26 darn you nvidia! 2022-10-07 22:41:38 nouveau the kernel driver should be present in lsmod and functional with linux-firmware-nvidia 2022-10-07 22:41:58 as for userspace there is xf86-video-nouveau for xorg (?) and mesa-dri-gallium has the nouveau component in there 2022-10-07 22:42:23 at best you're going to get a desktop though 2022-10-07 22:42:32 ah so no games or anything 2022-10-07 22:42:38 my intel laptop uhd igpu is better than like a gtx970 with nouveau which can't even open a game 2022-10-07 22:42:57 nvidia makes sure nouveau works well enough to open a web browser to download their drivers :p 2022-10-07 22:43:04 maybe the older gen cards work better 2022-10-07 22:43:05 lol well that makes sense 2022-10-07 22:43:30 so yeah sadly you won't do too well and i'd recommend either buying a gpu (expensive) or using a distro with the proprietary drivers, unless you don't play games 2022-10-07 22:43:39 if you don't play games, it's "ok", though it might crash here and there 2022-10-07 22:44:18 nah this pc is going to be my gaming one, I've already setup Alpine on me and my partner's laptops. I'll save up for a new GPU and install alpine on it then. Appreciate the help! 2022-10-07 22:44:30 np :) 2022-10-07 23:06:19 ACTION 2022-10-07 23:28:21 how is progress on reproducible builds going? 2022-10-08 07:08:18 Hi, is the busybox adduser hardcoded to create home directories that are world readable? https://git.busybox.net/busybox/tree/loginutils/adduser.c#n302 2022-10-08 07:09:35 ktprograms what? 2022-10-08 08:01:54 ktprograms: true. just added a user, and can easily cd to it's home, not a root btw 2022-10-08 08:12:48 https://github.com/brgl/busybox/blob/abbf17abccbf832365d9acf1c280369ba7d5f8b2/loginutils/adduser.c#L314 2022-10-08 08:12:52 0 2022-10-08 08:13:43 the user's home dir is 755 2022-10-08 08:16:47 this is terrible by default 2022-10-08 08:17:31 700 is something desired? 2022-10-08 08:37:10 looks like some distro uses 700 and some uses 755 for /home/$user 2022-10-08 10:54:35 actually, busybox uses 2755 for /home/$user permission 2022-10-08 10:55:15 2 is the setgid bit. 2022-10-08 10:55:40 755 makes everyone including nobody view the files under your home 2022-10-08 14:51:28 it seems that comes from 1fca558799350fb6caff97f55aa6d1237b08fccb , year 2002.. altough probably 750 and remove setgid should better, I'm not sure if it's a big problem 2022-10-08 14:55:03 It looks that tried to emulate std adduser at that time, https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=64806 2022-10-08 15:27:44 is the man-db trigger noise normal on edge (when using apk)? haven't seen it before 2022-10-08 15:30:06 can busybox ash use floating point in aritmetics? (it errors for me) 2022-10-08 15:32:49 ill ask in another chan too.. 2022-10-08 15:49:09 frag: no, sh/bash/ash/dash can not do floating point, but zsh can 2022-10-08 15:50:24 for that, use bc 2022-10-08 15:53:05 printf "%s" "scale=15;10/3" | bc 2022-10-08 15:54:33 donoban: ubuntu uses 755 by default, archlinux uses 700 2022-10-08 15:54:51 uuidNuniq: why not just printf "scale=15;10/3" | bc? 2022-10-08 15:55:21 ikke: well, that is fine too :) 2022-10-08 15:58:56 uuidNuniq: thanks! 2022-10-08 21:16:39 I'm not able to get shared folders or copy and paste between host and guest working in Alpine with VMware Workstation. I installed open-vm-tools following the guide on the wiki, but when I add a shared folder with the VMware GUI it doesn't appear in Alpine. I added vmhgfs-fuse /mnt/hgfs fuse defaults,allow_other 0 0 to /etc/fstab and rebooted, /mnt/hgfs is still an empty directory. Windows 11 host, VMwareยฎ Workstation 16 Pro 16.2.4 build-20089737, 2022-10-08 21:16:39 Alpine Linux 3.16.2 Virtual. 2022-10-08 21:43:14 inconsistant output of -v ? apk search -v -x linux-virt : linux-virt-5.15.71-r0 - Linux lts kernel ; apk info -v -e linux-virt : linux-virt-5.15.71-r0 ; apk info -e linux-virt : linux-virt 2022-10-08 21:44:32 Zaphida: there is a service you may need to start : open-vm-tools-openrc 2022-10-08 21:44:56 i am just guessing, i don't use this tool 2022-10-08 21:44:58 That is in the guide on the wiki. 2022-10-08 21:45:12 Wouldn't hurt to run it again just to make sure, though. 2022-10-08 21:47:48 Zaphida: which wiki? 2022-10-08 21:48:16 The Alpine Linux wiki. 2022-10-08 21:48:27 yeah, which URL? 2022-10-08 21:49:45 https://wiki.alpinelinux.org/wiki/Open-vm-tools 2022-10-08 21:49:56 rc-service open-vm-tools start 2022-10-08 21:52:08 Zaphida: fstab is not like that 2022-10-08 21:52:21 what? 2022-10-08 21:52:45 .host:/ fuse.vmhgfs-fuse nofail,allow_other 0 0 2022-10-08 21:53:02 oh ok 2022-10-08 21:53:04 thanks 2022-10-08 21:53:07 read arch wiki : https://wiki.archlinux.org/title/VMware/Install_Arch_Linux_as_a_guest#Shared_Folders_with_vmhgfs-fuse_utility 2022-10-08 21:54:45 well even though open-vm-tools is installed which is verified by the fact that the service is running 2022-10-08 21:54:55 'rc-update add open-vm-tools' to enable on boot 2022-10-08 21:55:11 vmware-hgfsclient and vmhgfs-fuse are not found 2022-10-08 21:55:58 Zaphida: apk add open-vm-tools-hgfs 2022-10-08 21:56:51 it's always something simple 2022-10-08 21:57:06 most of the time alpine splits packages 2022-10-08 21:57:31 that really should be in the wiki. I'm going to edit it. 2022-10-08 21:57:47 so you must : `apk search open-vm-tools` and take a good look at all the package names 2022-10-08 21:58:04 yes and the wiki page for open-vm-tools on alpine linux lists the packages it's split into but open-vm-tools-hgfs is not listed in that list 2022-10-08 21:58:35 ok I must note that search argument for future reference 2022-10-08 21:59:42 there are a lot of packages that aren't included on the wiki 2022-10-08 21:59:43 Zaphida: the wiki did not mention shared folders, it is meant to say 'how to see guest info from vmware host' 2022-10-08 22:01:23 most people want vmware tools resizeable screen and shared folders 2022-10-08 22:01:32 Zaphida: you'll want to include reading archlinux wiki no matter what distro you use. it is the best linux wiki to be honest. 2022-10-08 22:01:48 sure will do 2022-10-08 22:02:20 e.g. you first search alpine wiki, if there is no info you want, then go to arch wiki 2022-10-08 22:03:59 Zaphida: also if a command not found , you can search like this : apk search cmd:vmhgfs-fuse 2022-10-08 22:04:11 beautiful 2022-10-08 22:05:14 and if a lib not found : e.g. 'linux.h not does not exist' , you can 'apk search so:linux.h' 2022-10-08 22:05:48 shared object? 2022-10-08 22:05:53 cmd:XYZ so:XYZ 2022-10-08 22:06:20 .h is a source code header file which is shared with many programs source code 2022-10-08 22:06:36 linux.h being the core kernel header 2022-10-08 22:06:52 yeah. .h or .so or whatever lib 2022-10-08 22:07:50 this also helps : https://pkgs.alpinelinux.org/contents 2022-10-08 22:08:13 if you want to find what package provide a file you are looking for 2022-10-08 22:12:46 nice hexchat crashed trying to use find 2022-10-08 22:17:33 zaphida: oh, based on Xchat? 2022-10-08 22:17:43 there's no such thing as "linux.h the core kernel header": https://pkgs.alpinelinux.org/contents?file=linux.h&path=&name=&branch=edge&repo=main&arch=x86_64 2022-10-08 22:17:57 and you can't search for header files with so: because it's so: not file: 2022-10-08 22:18:02 Hello71: it was just an example 2022-10-08 22:21:31 now I'm having more trouble 2022-10-08 22:21:43 I'll just pastebin the terminal output 2022-10-08 22:23:55 one thing at a time zaphida 2022-10-08 22:24:19 this is a really weird one though 2022-10-08 22:24:59 can't mkdir /mnt/hgfs because file exists and can't rm /mnt/hgfs because no such file or directory 2022-10-08 22:25:03 very contradictory 2022-10-08 22:25:17 lol. that is cool 2022-10-08 22:25:39 ls -ld hgfs 2022-10-08 22:25:42 pastebin.com/H3gidjFA 2022-10-08 22:26:03 ls: hgfs: No such file or directory 2022-10-08 22:26:25 i think you mounted it to an unknow place 2022-10-08 22:26:32 and doing an ls in /mnt returns ls: ./hgfs: No such file or directory 2022-10-08 22:26:34 mount | grep hgfs 2022-10-08 22:26:47 I did indeed 2022-10-08 22:26:51 now im trying to reverse that 2022-10-08 22:28:44 vmhgfs-fuse on /mnt/hgfs type fuse.vmhgfs-fuse (rw,nosuid,nodev,relatime,user_id=0,group_id=0,allow_other) 2022-10-08 22:29:21 umount /mnt/hgfs works just fine 2022-10-08 22:29:40 yay 2022-10-08 22:30:04 /mnt/hgfs now contains the shared folder 2022-10-08 22:30:14 now for copy and paste between host and guest 2022-10-08 22:30:46 forgot the .host:/ ? 2022-10-08 22:31:16 no I put .host:/steam when I should have put .host:/steam_linux 2022-10-08 22:31:39 because I had changed the name 2022-10-08 22:31:49 and which folder in windows it is 2022-10-08 22:32:10 should have double checked with vmware-hgfsclient 2022-10-08 22:33:19 vmware-user - Tool to enable clipboard sharing (copy/paste) between host and guest. 2022-10-08 22:33:31 need a GUI interface i guess? 2022-10-08 22:34:01 ^^ for copy and paste 2022-10-08 22:34:21 yep and the package is open-vm-tools-gtk 2022-10-08 22:34:50 and vmware-user should be run after X11 2022-10-08 22:35:37 ok. did you solve the shared folder issue? 2022-10-08 22:36:59 yes :) 2022-10-08 22:37:01 see above 2022-10-08 22:37:19 grea 2022-10-08 22:37:22 t 2022-10-08 22:37:42 now to add it into steam 2022-10-08 22:38:22 small and beautiful alpine. a nice journey 2022-10-08 22:38:46 and hope CSGO runs useably. it ran at like 2-3 frames per minute under antix, not sure what the problem was 2022-10-08 22:39:09 just in the menu 2022-10-08 22:41:09 yes very much so :)) 2022-10-08 23:41:37 2-3 frames per minute ! 2022-10-08 23:48:11 yes Lord 2022-10-08 23:48:31 unfortunately it looks to be the same in my alpine vm after finally getting everything set up 2022-10-08 23:49:14 black screen for a loooong time and if it's the same as antix it will finally load and play music but run at 2-3 frames per minute 2022-10-08 23:49:19 oh it just loaded 2022-10-08 23:49:28 just saw the CS GO splash 2022-10-08 23:50:07 menu is up, well this is better than 2-3 frames per minute 2022-10-09 01:08:34 ikke: apk add dhcpcd ; pkill -x /sbin/udhcpc ; rc-service networking restart ; output: no such user dhcpcd , however dhcpcd is working but as root 2022-10-09 01:25:40 it is a dhcpcd bug 2022-10-09 01:29:18 https://github.com/NetworkConfiguration/dhcpcd/blob/4627c5d80fd677531878a819d30ad9f6d385b50b/src/privsep.c#L104 2022-10-09 09:48:34 does anyone tried to share host /etc/apk/cache while doing 'docker build'? 2022-10-09 09:56:29 uhM, looks something like RUN --mount=type=cache,target=/var/cache/apk 2022-10-09 10:02:56 NE#NEthe --mount option requires BuildKit. Refer to https://docs.docker.com/go/buildkit/ to learn how to build images with BuildKit enabled 2022-10-09 10:03:00 meh 2022-10-09 11:11:42 I have buildkit enabled by default nowadays 2022-10-09 14:45:08 yeah it looks great 2022-10-09 14:45:31 ikke: do you know if I could share the host cache with write access? 2022-10-09 14:46:44 I currently have RUN --mount=type=cache,target=/var/cache/apk , it runs fine but I feel that it creates cache folder somewhere internally, and I don't sure if it will be shared with other images 2022-10-09 15:45:56 donoban: yeah, I don't see it documented, but it's indeed likely to be cached per image / layer 2022-10-09 15:46:52 uuidNuniq: yeah I spotting that privsep wasn't working with dhcpcd some time ago, it's been on my list of things to fix for some time (along with the multi interface pid file issue), I haven't gotten a Round Tuit yet ;-) 2022-10-09 15:50:26 ACTION hands minimal a Square Tuit 2022-10-09 16:01:01 psykose: you can't fit a Square Tuit in a Round hole lol 2022-10-09 16:01:14 :) 2022-10-09 16:04:06 You can if the Square Tuit is small enough :P 2022-10-09 16:09:12 minimal: the reason i use dhcpcd is that udhcpc does not detect cable unplug. 2022-10-09 16:10:25 tested in qemu : set_link virtio-net-pci.0 off , the ip addr is still there with udhcpc 2022-10-09 16:10:54 uuidNuniq: nothing wrong with dhcpcd, I like it, was just pointing out 2 issues I'd spotted in the past with its Alpine package that I've been meaning to fix 2022-10-09 16:11:00 while dhcpcd will detect and flush the eth0 2022-10-09 16:12:45 minimal: got it. i tried to create a dhcpcd user. the log says something like permission error. i forgot to record that down. 2022-10-09 16:12:57 unplugging the networking makes the bits sad :( 2022-10-09 16:13:39 psykose: e.g. router rebooted ... such case 2022-10-09 16:15:22 does mdevd need hwdrivers? 2022-10-09 16:15:49 from setup-devd i see that mdev needs it, but the other 2 does not need it 2022-10-09 16:16:37 uuidNuniq: from memory there's an outstanding MR that skarnet submitted before he went on holiday that hasn't yet been merged, think it adds hwdrivers for mdevd 2022-10-09 16:17:43 i was thinking about del it from sysinit, looks like networking needs it too 2022-10-09 16:17:49 ok, then let it be there 2022-10-09 16:20:11 uuidNuniq: https://gitlab.alpinelinux.org/alpine/alpine-conf/-/merge_requests/100 2022-10-09 16:20:46 I pinged skarnet about this a week ago but he wasn't "officially" finished his holiday then lol 2022-10-09 16:21:11 aha, it is nice to have holidays 2022-10-09 16:21:41 uuidNuniq: the other thing about dhcpcd is that the last upstream release was 1 year ago and there's been quite a few changes to it (including related to privsep) 2022-10-09 16:23:30 while dhclient is about to EOL 2022-10-09 16:24:12 uuidNuniq: dhclient is already EOL, it is dhcpd (the server) that is about to EOL 2022-10-09 16:24:51 there was talk about the server EOL in Debian channel yesterday 2022-10-09 16:24:58 i see : ISC has ended development on the ISC DHCP client as of early 2022 2022-10-09 16:25:25 ISC has announced the end of life for ISC DHCP as of the end of 2022. 2022-10-09 16:26:17 which dhcp client and servere are most people moving to? 2022-10-09 16:26:31 dnsmasq has a server 2022-10-09 16:27:05 raven523: ISC expect people to move from dhcpd to kea 2022-10-09 16:28:38 however kea doesn't provide a client AFAIK 2022-10-09 16:29:07 looks like networkmanager has everything 2022-10-09 16:33:38 is this maintained ? https://boot.alpinelinux.org/boot.ipxe 2022-10-09 16:37:42 anyway i parse this script to 3 lines of code, to run in a netbootxyz IPXE shell, works on both aarch64 and x86_64 with the latest_stable tag 2022-10-09 16:39:44 kernel http://dl-cdn.alpinelinux.org/alpine/latest-stable/releases/aarch64/netboot/vmlinuz-lts modules=loop,squashfs nomodeset alpine_repo=http://dl-cdn.alpinelinux.org/alpine/latest-stable/main modloop=http://dl-cdn.alpinelinux.org/alpine/latest-stable/releases/aarch64/netboot/modloop-lts console=ttyS0 console=tty0 console=ttyAMA0 initrd=initramfs-lts 2022-10-09 16:39:58 initrd http://dl-cdn.alpinelinux.org/alpine/latest-stable/releases/aarch64/netboot/initramfs-lts 2022-10-09 16:40:02 boot 2022-10-09 16:55:46 the alpine bootable ISO, regardless which flavor, lts virt edge, when it boots the date is always not 'now', this cause SSL issue during installation. my solution is to run 'hwclock -s' before doing anything. to make sure the date is correctly synced 2022-10-09 16:56:42 i think this needs to get some attention :P 2022-10-09 16:57:47 ncopa ^^^ 2022-10-09 16:58:01 uuidNuniq: better to create an issue on gitlab 2022-10-09 16:58:28 i don't have an account there :) 2022-10-09 16:58:34 uuidNuniq: suggestions how this would work for SBCs with no RTC (e.g. RPIs) ? ;-) 2022-10-09 16:59:09 minimal: start a network and start a ntpd 2022-10-09 16:59:16 uuidNuniq: generally the ISO is used as an installer rather than a runtime 2022-10-09 17:00:00 yeah, but it needs to download some packages from repos i think? 2022-10-09 17:00:34 uuidNuniq: yes if HTTPS is being used that would be an issue (unless the Extended ISO is used) 2022-10-09 17:01:17 yeah, https, i use that in setup-alpine -f setup.conf 2022-10-09 17:01:30 if http, no issue. 2022-10-09 17:07:16 uuidNuniq: "hwclock -s" should not be needed for x86_64 (and likely x86) as CONFIG_RTC_HCTOSYS=y is set in the kernel config, this means the kernel sets system clock to RTC clock time when kernel initialises 2022-10-09 17:07:25 should appear in "dmesg" output 2022-10-09 17:16:03 I am trying to install on a raspberry pi cm4. I have formatted the emmc into 256m vfat and the remainder ext4, set boot and lba flags, then decompressed the aarch64 tarball into the vfat partition and made an overlay with a wifi SSID and password, but I get no connection to the wifi. I also don't see any blinking activity lights on the carrier board when I power on, so I am thinking that it fails pretty early in the boot process. 2022-10-09 17:16:03 What should I check? 2022-10-09 17:19:51 uuidNuniq: note you may need to add something like "log_buf_len=32768" to the bootloader's cmdline to ensure that the full set of kernel messages is shown by "dmesg" 2022-10-09 17:23:05 cache_return[m]: does the emmc present itself as /dev/mmcblk0? How are you booting in the first place? via your module's interface board's USB? SDCard? 2022-10-09 17:51:50 It mounts as a block device. It is sdb1 and 2. I am using the emmc onboard. 2022-10-09 18:06:54 cache_return[m]: I'm confused - how did you boot the CM4 in the first place? 2022-10-09 18:17:05 Let's start over. I formatted the emmc of a compute module 4 as described above--I did that by putting the module in a special mode (disable emmc boot with a jumper and run a script on USB host). Then I turned it on and ... nothing. So I do not know if it's booting at all. I suspect it's not since there is no activity light. 2022-10-09 18:17:46 There is no monitor attached, so I was using the instructions for headless setup. 2022-10-09 18:18:14 https://github.com/macmpi/alpine-linux-headless-bootstrap 2022-10-09 18:19:31 cache_return[m]: so do the RIP bootloader files reflect that eMMC is /dev/sda rather than /dev/mmcblk0? 2022-10-09 18:19:38 s/RIP/RPI/ 2022-10-09 18:19:38 minimal meant to say: cache_return[m]: so do the RPI bootloader files reflect that eMMC is /dev/sda rather than /dev/mmcblk0? 2022-10-09 18:20:39 oops, sdb 2022-10-09 18:24:22 Well, to be clear, I am saying that the emmc reports as sdb when mounted as a block device for setup. That's not relevant to booting is it? 2022-10-09 18:24:53 Also, what rpi bootloader files are you talking about? 2022-10-09 18:26:38 cache_return[m]: the files the RPI's bootlaoder requires to boot, e.g. config.txt, cmdline.txt, bootcode.bin, etc 2022-10-09 18:27:50 typically in cmdline.txt there will be a "root=" entry that typically says something like "root=/dev/mmcblk0p3", in your case that would need to refer to the eMMC device 2022-10-09 18:31:54 This seems like a likely cause of my problem. But I wonder how to be sure what root will be. 2022-10-09 18:33:01 How it appears to a random USB host does not predict how it will enumerate when it boots. 2022-10-09 18:34:10 cache_return[m]: what do you mean? rootfs is where you installed the rootfs. As the only storage you have is the eMMC (unless you have a USB or something else to install one) then /dev/sdbX will be the rootfs, where "X" depends on how many partitions are created. What is device /dev/sda BTW? 2022-10-09 18:36:35 cache_return[m]: you could specify UUID or LABEL for the rootfs instead of a device name 2022-10-09 18:39:43 What I mean is that sdb is arbitrarily assigned by the host. When the device boots by itself it will not be called that. 2022-10-09 18:41:35 cache_return[m]: that is why I mentioned you can use UUID or LABEL to refer to it instead 2022-10-09 18:41:53 It's called sdc if I plug something else in first. Yeah, UUID might be the way. 2022-10-09 19:18:28 Well, there was no root= line in cmdline.txt 2022-10-09 19:19:09 Adding root=PARTUUID=blahblah did not fix it. 2022-10-09 19:20:10 hey motherfuckers 2022-10-09 19:20:14 cache_return[m]: so what was in cmdline.txt? 2022-10-09 19:20:21 enlight: mind your language 2022-10-09 19:23:27 minimal: I'll need a minute to go back and look. 2022-10-09 19:28:04 cache_return[m]: are you trying to install Alpine in Sys-mode ("run-from-disk") rather than "run-from-ram", right? 2022-10-09 19:35:59 Yes 2022-10-09 19:38:42 cache_return[m]: so the RPI image you have is designed with run-from-ram in mind, if you have booted that and then run setup-alpine and specified "sys" mode then it would create a cmdline.txt file that does specify "root=" 2022-10-09 19:40:03 cache_return[m]: as can be seen here: https://gitlab.alpinelinux.org/alpine/alpine-conf/-/blob/master/setup-disk.in#L470 2022-10-09 19:41:13 All the instructions from the wiki say to download the aarch64 image, then setup sysmode on first boot. 2022-10-09 19:44:17 cache_return[m]: yes, but my understanding if you haven't actually booted Alpine yet in order to run setup, you have instead copied files across onto the eMMC from the host computer 2022-10-09 19:45:05 Yes, copied the files and turned it on. 2022-10-09 19:46:15 right, but is what you have copied across actually bootable for a CM4 using eMMC? 2022-10-09 19:50:03 Probably not in my opinion, so I came in here to ask what specific configuration I might need to consider for this hardware. 2022-10-09 19:50:39 Weirdly, there are howtos for CM4 that do not mention any additional configuration. 2022-10-09 19:51:27 cache_return[m]: but are they written with using a SDcard via a interface board to boot rather than the mechanism you are using? 2022-10-09 19:52:51 No. 2022-10-09 19:55:34 getting this both when i set it as dash/ash script and with or without '-m', whats the solution?: /1/z/x: line 0: can't access tty; job control turned off 2022-10-09 19:58:18 cache_return[m]: ok, I haven't seen such guides so I can't comment on them 2022-10-10 06:47:10 Hello I'm new here 2022-10-10 06:47:12 How alpine ISO file can be so small compare other raw distro like Arch Linux or Debian? 2022-10-10 06:55:34 faustzero1: much less bloat ;) 2022-10-10 07:01:35 aron, yeah, but what is it? 2022-10-10 07:02:07 what stuff that has been trimmed on Alpine Linux that other distro hasn't aron 2022-10-10 07:02:17 no systemd for start, and the ISO does not include any GUI 2022-10-10 07:03:30 I see Artix linux has no systemd and the ISO does not include any GUI too, but the size difference is much larger than Alpine aron 2022-10-10 07:04:30 *also Devuan has no systemd and not include any GUI, but the ISO are larger 2022-10-10 07:05:06 list the preinstalled packages on them if you can 2022-10-10 07:05:15 is by using musl + busybox instead of glibc and gnu coreutils will save much file's size? 2022-10-10 07:06:50 gnu stuff are large, stuffed with enormous amount of features 2022-10-10 07:07:25 the ISO simply contains much less preinstalled (or installable) packages 2022-10-10 07:18:00 which doc do I have to install to get "man ls" working? 2022-10-10 07:19:11 man-pages-posix 2022-10-10 07:21:25 thanks. is there a way to figure this out on my own? i.e. which package to retrieve to get "man X"? 2022-10-10 07:23:00 usually just the -doc of the thing it comes from, and man-pages/man-pages-posix is an exception 2022-10-10 07:23:26 oh, I see. thanks. 2022-10-10 07:24:25 you can also try something like https://pkgs.alpinelinux.org/contents?file=man.1&path=&name=&branch=v3.16&arch=x86_64 2022-10-10 07:25:06 https://pkgs.alpinelinux.org/contents?file=ls.1.gz&path=&name=&branch=v3.16&arch=x86_64 rather 2022-10-10 11:32:57 How do I disable apk invoking grub-mkconfig ? 2022-10-10 11:34:29 uuidNuniq: i would appreciate an issue about hwclock in initramfs, otherwise it will likely be forgotten 2022-10-10 11:57:00 Is there some breakage with ncurses on edge? It does not seem to want to load terminfo entries for me anymore. 2022-10-10 11:57:20 ncopa: as I pointed out to uuidNuniq, the x86/x86_64 kernels config is such that the kernel itself should get the system time from the RTC when it initialised, other archs are likely to be a different matter as they may not have RTC drivers compiled into the kernel 2022-10-10 11:58:52 minimal: sure. but if there are no issue where it can be tracked, it will likely not be followed up 2022-10-10 11:59:52 ncopa: agreed 2022-10-10 12:01:21 Ah, the issue seems to be that apk fix -r doesn't actually reinstall files? The terminfo/l/linux file was 0 bytes. Had to extract the apk with tar manually. 2022-10-10 12:06:13 I also would like apk to stop invoking grub-mkconfig 2022-10-10 12:06:18 this is more like a debian thing 2022-10-10 12:08:18 markand: it is being triggered whenever changes occur in the /boot directory, this is to ensure that the grub.cfg reflects the actual contents of /boot 2022-10-10 12:08:59 sure, but that is can be configurable would be great, I have grub installed because I need it to create bootable m.2 disks through an adapter 2022-10-10 12:09:16 but I don't use it myself so each time there is an upgrade I get my host /boot polluted with unneeded grub files 2022-10-10 12:10:58 markand: all I can think would be for the trigger to check for the existance of some file to prevent the autorun of grub-mkconfig. Your setup is sort of non-standard 2022-10-10 12:11:53 I don't know what kind of standard is, everyone is free to have its bootloader of choice :) 2022-10-10 12:12:47 markand: right, but you've installed a bootloader (package) but don't actually want it to act as bootloader 2022-10-10 12:26:04 minimal: ncopa : there are aarch64 VMs, a lot, for instance, i run Alpine Linux on Apple M1 Qemu as well as on VPS. 2022-10-10 12:27:08 uuidNuniq: ah, that makes sense, the RTC driver is likely a loadable kernel module then rather than being compiled into the kernel 2022-10-10 12:27:47 it is not a hwclock issue. it is that from what i experienced so far, aarch64 ISO does not sync time from host when it boots. 2022-10-10 12:28:59 the kernel code to auto set system time from RTC won't run if the RTC is not available (as RTC module not loaded at that point) 2022-10-10 12:32:36 wait a moment. on Virt : * Setting system clock using the hardware clock [UTC] ... [ ok ] 2022-10-10 12:32:54 let me try another ISO 2022-10-10 12:33:02 virt syncs clock fine. 2022-10-10 12:34:24 alpine-virt-3.16.2-aarch64.iso this one syncs clock fine on boot. 2022-10-10 12:36:01 uuidNuniq: that output means the hwclock init.d service ran 2022-10-10 12:38:00 as it is a virt machine then it is likely using the UEFI RTC driver which is compiled into the linux-virt aarch kernel 2022-10-10 12:38:33 whereas on a M1 you would have to load whichever kernel module is needed for the M1's RTC 2022-10-10 12:39:22 alpine-standard-3.16.2-aarch64.iso : https://0x0.st/otto.txt 2022-10-10 12:40:13 uuidNuniq: that boot output shows a machine using swclock init.d rather than hwclock 2022-10-10 12:40:42 as it is using linux-lts which type of physical machine are you booting this on? 2022-10-10 12:40:54 so , it is different standard vs virt ISO 2022-10-10 12:41:23 the above are all boot from Apple M1 Qemu 7.0 2022-10-10 12:41:48 linux-virt is a kernel intended for a virtual machine, linux-lts is a kernel intended for a physical machine 2022-10-10 12:42:40 so if you are using linux-lts you need to load the RTC driver for the "physical" machine, even if you are booting it on a Qemu VM emulating a physical machine 2022-10-10 12:42:57 aha, i see, that is a design consideration. actually i did not know virt iso before. but since i used it, it is so efficient for my needs 2022-10-10 12:44:31 minimal: either a check in trigger (like update-extlinux has) or separate package to do the config autogeneration 2022-10-10 12:45:10 grub-mkconfig just doesn't work for all setups. Just for the common simple cases. 2022-10-10 12:45:42 ccx^xmpp: in which setups have you seen it not work? 2022-10-10 12:46:41 uuidNuniq: linux-virt kernel is not intended to be used to emulate an Apple M1 2022-10-10 12:47:01 minimal: really? even under Qemu? 2022-10-10 12:48:34 i have no issues so far running a linux-virt kernel under qemu on macOS 2022-10-10 12:49:43 uuidNuniq: Yes, the linux-virt kernel has very few drivers for physical hardware included as it is intended to be used on hypervisors 2022-10-10 12:50:12 https://0x0.st/ottZ.txt 2022-10-10 12:50:30 uuidNuniq: that's a different situation - it is not about what you run Qemu on, it is about what type of "machine" you want a VM to be 2022-10-10 12:50:48 all virtio, i don't have to attach physical hardware to it 2022-10-10 12:51:10 do you want a VM to appear to be a physical machine, then use the linux-lts kernel, if you want the VM to be a virtual machine then use linux-virt 2022-10-10 12:51:50 uuidNuniq: I thought you wanted to emulate a M1 Mac in Qemu? 2022-10-10 12:51:58 see this output : https://0x0.st/ottZ.txt 2022-10-10 12:52:17 minimal: no, i don't emulate M1 mac, i run alpine on Qemu on m1 mac 2022-10-10 12:52:30 uuidNuniq: ok, then use linux-virt package 2022-10-10 12:52:38 then do some software test, that is all. 2022-10-10 12:53:03 minimal: mine is LUKS setup with detached header and LVM 2022-10-10 12:53:42 uuidNuniq: I assume you are always booting the VM via UEFI in which the UEFI RTC driver will be available and hwclock init.d will set system clock from RTC 2022-10-10 12:54:28 yeah, so far, all my cases are UEFI boot 2022-10-10 12:54:40 ccx^xmpp: LUKS and LVM should not be a problem, not sure about detached header as have not used it before, but I assume it is just a matter of you setting up /etc/default/grub appropriately 2022-10-10 12:54:58 i have 2 VPS running alpine linux and one local qemu instance 2022-10-10 12:55:14 uuidNuniq: so then you should not have a problem with system time 2022-10-10 12:56:03 minimal: yes, virt iso has no problem. i will use virt for all future instance. 2022-10-10 12:57:22 i wanted to run alpine on my Amlogic S905d tv box but all the alpine kernel config do not support Amlogic as i checked 2022-10-10 12:57:51 so currently, the s905d runs archlinux arm. 2022-10-10 13:00:24 uuidNuniq: likely because no-one has submitted patches/changes to enable Alpine to run on that machine 2022-10-10 13:00:43 aha, i have all the files and patches :) 2022-10-10 13:01:13 minimal: It's matter of modifying the scripts that attach the LUKS device. For which I use scripts in /boot rather than initrd. So there's lot of difference from the default setup and grub-mkconfig doesn't detect the systems to boot correctly. 2022-10-10 13:01:29 uuidNuniq: so there's an opportunity for you to submit changes to get it with Alpine ;-) 2022-10-10 13:01:37 a dtb patch and a config to enable selection of amlogic in `make nconfig`, that's all 2022-10-10 13:01:54 Though the obvious issue to me is that every config file is protected by apk from unintended modifications โ€ฆ except of this one. 2022-10-10 13:03:28 ccx^xmpp: grub.cfg is not provided by apk, it is autogenerated (by grub-mkconfig) and so apk cannot "protect" it 2022-10-10 13:04:45 what do you mean protected ? like apk audit? 2022-10-10 13:04:56 Indeed. But it also doesn't have to force overwrite on every install. Extlinux doesn't 2022-10-10 13:06:47 ccx^xmpp: that's why I suggested the trigger be modified to check for a "don't run" flag/file. For typical Alpine installs people would want it to autogenerate/update to ensure it correctly reflects the files they have in /boot 2022-10-10 13:08:16 At very least it could check if the grub.conf file has been autogenerated or hand-written. grub-mkconfig creates a recognizable header there 2022-10-10 13:10:53 ccx^xmpp: so submit a MR for the trigger to check if grub.cfg is handwritten? 2022-10-10 13:12:48 How can I make lbu save installed packages? Every time I reboot I have to run apk update and apk fix in order to run programs installed from repos? 2022-10-10 13:17:17 Are triggers ran with sh -e or does it follow the shebang? 2022-10-10 13:17:59 Because currently it has absolutely no error handling. 2022-10-10 13:18:34 -e isn't error handling 2022-10-10 13:20:00 impressive : https://uploads-ssl.webflow.com/6228fdbc6c97145dad2a9c2b/624e2337f70386ed568d7e7e_chainguard-all-about-that-base-image.pdf 2022-10-10 13:23:08 It's exit on error for simple statements. And the trigger script doesn't check for success of those nor does it `set -e`, making it ignore those errors. 2022-10-10 13:24:06 it's exit on non-zero exit status in some cases but not all 2022-10-10 13:25:35 if a script should exit in some situations then it needs to check for that situation 2022-10-10 13:31:41 Yes, the "some cases" are simple statements (i.e. not compound commands). So yeah, that's another thing to fix in the script. Just wanted to check apk doesn't do any magic there. 2022-10-10 13:37:15 the grub uses device name instead of UUID by default. which is emm... 2022-10-10 13:38:09 doesn't know your uuid beforehand? 2022-10-10 13:39:21 GRUB_DISABLE_LINUX_UUID=false 2022-10-10 13:39:43 this is by default true, need to change it to false 2022-10-10 13:40:33 BTW here's the trigger patch, until I figure out where and how to submit it: https://bpa.st/GOBQ 2022-10-10 13:44:07 ccx^xmpp: submit it as a git Merge Request for aports/main/grub ? 2022-10-10 13:44:36 On the gitlab instance? I'll go ahead and try to make an account. 2022-10-10 13:45:33 ccx^xmpp: well you fork the aport repo, do the change in your fork, then submit a MR to merge it to the main aports 2022-10-10 13:46:14 yeah, i don't have an account either. i'll create one, alpine is more and more interesting to me after 20days try on 2022-10-10 16:33:38 what happened https://git.alpinelinux.org/aports/log/?h=3.16-stable 2022-10-10 16:33:53 imagemagick 2022-10-10 16:34:08 why it has to upgrade one version at a time 2022-10-10 16:35:58 psykose: ^ 2022-10-10 16:50:51 i think that was an article title 2022-10-10 16:51:07 please give us the url, uuidNuniq 2022-10-10 16:52:10 lopid: what url 2022-10-10 17:06:23 https://thenewstack.io/chainguard-its-all-about-that-base-image/ 2022-10-10 22:19:54 j0wi person just picks the upgrades like that 2022-10-10 22:19:58 shrug 2022-10-11 04:31:01 if apk search could support some basic regex that would be great 2022-10-11 04:31:04 apk search zsh | grep '^zsh' 2022-10-11 11:15:13 apk list zsh\* 2022-10-11 11:32:51 Hello71: yes, i am aware that list can use * 2022-10-11 11:34:20 but anyway, pipe to grep and use grep's regex also works, so it is not a must have feature 2022-10-11 11:44:24 something something unix principle 2022-10-11 11:47:28 what, you don't want a regex implementation compiled into 95 programs on your system? 2022-10-11 12:00:38 well, there is no some 'good' regex as/in some lib? 2022-10-11 12:01:41 don't worry about programs linking -lregex, worry about them implementing a custom LISP 2022-10-11 12:02:01 psykose: 95 lol, how did you come up with that number 2022-10-11 12:02:32 random number 2022-10-11 12:02:53 anyway for apk scope pipeline is good enough 2022-10-11 12:04:24 let's pipe 2022-10-11 12:04:55 ACTION . o O (If you start me up, If you start me up...) 2022-10-11 12:43:56 you make a grown man cry 2022-10-11 16:55:42 hi here. cryptkey does not support arguments in alpine ? full device encryption work's with cryptkey (no argument) in initramfs. But no in external media. Eq: cryptkey=UUID=C237-B385:vfat:/crypto_keyfile.bin 2022-10-11 17:06:13 tetsumaki: that would require the initramfs' init to mount the vfat filesystem somewhere and then read the file from it - it does not appear to support this 2022-10-11 17:35:33 ok minimal :( 2022-10-11 17:43:58 :D https://0x0.st/otCV.txt 2022-10-11 18:07:21 tetsumaki: there is a pending MR: https://gitlab.alpinelinux.org/alpine/mkinitfs/-/merge_requests/108 that partly does what you want 2022-10-11 23:04:46 is this page from 2016 still the recommended way to make packages from git refs? https://wiki.alpinelinux.org/wiki/APKBUILD_examples:Git_checkout 2022-10-11 23:05:39 there seems to be a "snapshot" method in abuild now, which refers to $giturl, etc 2022-10-12 00:23:31 Hey yโ€™all, is there a package for `arcan` in any of Alpineโ€™s repos? Even one in testing or unmaintained, I just want to take a look, see if I can maybe skip packaging it myself. 2022-10-12 00:24:02 you can search here: https://pkgs.alpinelinux.org/ 2022-10-12 00:24:56 there may also be APKBUILDs in aports/testing that don't have built packages 2022-10-12 00:49:14 I donโ€™t see any unfortunately. Going to try searching the mirror on GitHub because I canโ€™t find the search button on gitlab 2022-10-12 00:51:49 Nothing 2022-10-12 00:52:20 It seems I will have to speak to letoram about how best to package it. 2022-10-12 01:40:27 Good night, vlc does not think chromecast, do I have to configure something? 2022-10-12 01:41:41 it's missing the deps for it 2022-10-12 01:41:45 i guess i can have a look 2022-10-12 01:59:10 added, when vlc upgrades to -r10 try again 2022-10-12 02:03:17 Obrigado 2022-10-12 02:03:26 Thank's 2022-10-12 02:03:30 rsrs 2022-10-12 05:48:53 hi, i don't have any other installation for comparison, but is there any issue with 4k on amdgpu? i had crashes when playing vlc in radeon driver (on some videos, i guess it depends on codec), so i switched to amdgpu, which works fine, but only allows me to select 1080p, not 4k resolution 2022-10-12 05:49:09 edge kernel 2022-10-12 05:53:53 you mean the radeon/amdgpu kernel modules? 2022-10-12 05:53:56 what gpu do you have 2022-10-12 06:14:46 psykose: yup. AMD Radeon HD8280E 2022-10-12 06:15:28 could you apk add drm_info and pastebin drm_info 2022-10-12 06:15:33 where exactly do you "select 4k" 2022-10-12 06:16:06 display settings inxfce, but i assume xrandr won't see it either, i'll confirm 2022-10-12 06:19:34 http://sprunge.us/nOvZQq 2022-10-12 06:19:44 and xrandr doesn't see it (on radeon it does) 2022-10-12 06:20:05 indeed this doesn't see 4k at all from the driver directly 2022-10-12 06:21:18 hm 2022-10-12 06:21:23 curious if you have other ports on the gpu 2022-10-12 06:21:52 dvi, but i can't connect that to the tv 2022-10-12 06:24:10 there's probably a few things to try but i don't remember what they are 2022-10-12 06:26:22 that's what i get for trying to get low power machine working as tv player xD it's still the same that has sata drive issues on lts kernel, that's why i'm on edge :D 2022-10-12 06:36:55 upgrade from 3.16.2 to edge, stuck at Scanning hardware for mdevd ... 2022-10-12 06:46:46 aha, installed linux-edge fixed the issue 2022-10-12 06:47:40 the classic 2022-10-12 06:59:43 * Setting system clock using the hardware clock [UTC] ...grep: warning: stray \ before - 2022-10-12 07:00:13 but there is linux-virt and linux-lts in edge right? i think it was because features changed? 2022-10-12 07:01:07 psykose: what do you mean? use mdev instead of mdevd? 2022-10-12 07:02:18 i don't mean anything 2022-10-12 09:05:33 could someone guide me howto install jupyter-notebook pls ? 2022-10-12 09:05:48 seems I cannot figure out the pkgs needed 2022-10-12 09:09:29 `apk add jupyter-notebook` should work 2022-10-12 09:09:50 currently testing on v3.16(aarch64), pkgs installed https://tpaste.us/5kKj 2022-10-12 09:10:43 I tried, keep getting -> ModuleNotFoundError: No module named 'jupyter_server' 2022-10-12 09:35:57 there is another one, 'Error loading server extension ipyparallel' 2022-10-12 09:55:04 vkrishn: using pip? 2022-10-12 09:55:21 why don't you follow their official website for installation 2022-10-12 09:55:27 uuidNuniq: no, just apk add 2022-10-12 09:55:47 try using : pip install notebook 2022-10-12 09:55:52 ok 2022-10-12 09:59:19 hmmm... install went ok 2022-10-12 10:00:57 vkrishn: but? 2022-10-12 10:02:45 same errors with previous install, now trying to del pkgs and pip install all 2022-10-12 10:05:40 do I need to do, pip install jupyterlab ? 2022-10-12 10:05:57 coz, its says jupyter: not found 2022-10-12 10:06:24 ok did it 2022-10-12 10:06:35 not sure. you can open their website and follow the docs 2022-10-12 10:06:41 looks like all you need is pip3 2022-10-12 10:22:38 progress, re-installed previous pkgs over it, now works with error 2022-10-12 10:22:47 except it tryin 2022-10-12 10:22:54 except it try to find kernel 2022-10-12 10:22:58 but still works 2022-10-12 10:23:09 now cleaning, todo 2022-10-12 11:02:49 Hello. I want to use in Alipine Edge but cannot open .svg files with it because librsvg dropped from this package. 2022-10-12 11:02:54 Is ot possible to add it? 2022-10-12 11:03:08 here is the issue on the github 2022-10-12 11:03:09 https://github.com/artemsen/swayimg/issues/54#issuecomment-1275977693 2022-10-12 11:03:54 swayimg-full has it 2022-10-12 11:04:23 and please don't report these things upstream as it's not their problem and a waste of their time 2022-10-12 11:07:19 got it. Sorry. What is the proper place to report it? 2022-10-12 11:10:34 aports 2022-10-12 11:10:58 ok. any irc channel for aprots? Or report only to repo? 2022-10-12 11:14:15 https://gitlab.alpinelinux.org/alpine/aports/-/issues 2022-10-12 11:18:38 you're already here 2022-10-12 13:05:24 Does anyone here use Giara? I packaged it a good while ago because I thought it could be nice to have but I don't really have any use for it personally and last time I tried using it to make sure it still works it was unbearably slow. 2022-10-12 13:05:43 I'm thinking about moving it to unmaintained 2022-10-12 14:52:40 hi! i'm trying to use the raspi aarch64 image for a pi zero 2w, but somehow it doesn't contain the wifi firmwarew 2022-10-12 14:52:44 is there a way I could add them? 2022-10-12 14:52:46 thanks! 2022-10-12 14:53:29 pipppero: Does the pi zero 2w even have a 64 bit CPU? 2022-10-12 14:53:42 yup, it's a pi3 in a different form factor 2022-10-12 14:53:45 it boots fine! 2022-10-12 14:55:10 Nice, didn't know that. Is the pi zero 2 new? 2022-10-12 14:55:36 hmm i think it's a couple years old 2022-10-12 14:57:31 you have to somehow copy the firmware over because of a bug in the image generation 2022-10-12 14:57:42 the firmware does exist in the usual firmware packages, just not on the install iso 2022-10-12 14:57:51 right, linux-firmware-cypress should do the trick right? 2022-10-12 14:57:52 sounds like time for an iso respin 2022-10-12 14:58:01 the bug is in the thing that generates the iso 2022-10-12 14:59:22 https://gitlab.alpinelinux.org/alpine/alpine-conf/-/merge_requests/108 and the following links 2022-10-12 14:59:26 pipppero: Seems like I haven't been up to date on raspberry pi products for a few years now ... xD 2022-10-12 15:00:06 pipppero: the linux-rpi/4 kernels already depend on it 2022-10-12 15:00:09 brcm/cypress yeah 2022-10-12 15:00:20 but you have to somehow finish the install and have it install the packages 2022-10-12 15:00:25 or copy stuff by hand i guess 2022-10-12 15:00:32 huh 2022-10-12 15:02:07 uname says i'm on 5.15.55-0-rpi 2022-10-12 15:03:06 so can i just download https://pkgs.alpinelinux.org/package/v3.16/main/aarch64/linux-firmware-cypress, and apk-add it? 2022-10-12 15:03:40 if you have networking on the device you can merely finish the install and reboot 2022-10-12 15:03:51 otherwise, yeah, you have to install that by hand 2022-10-12 15:03:56 sadly I don't, the zero doesn't have ethernet 2022-10-12 15:03:57 and remove/reprobe the modules i think 2022-10-12 15:04:07 then the interface should exist 2022-10-12 15:04:21 and i think even then you need to somehow connect to wifi and i forget if wpa_supplicant is even there 2022-10-12 15:04:41 ye wpa_supplicant is there, in the local cache i think 2022-10-12 15:05:03 neat 2022-10-12 15:05:22 uh, errors 2022-10-12 15:05:46 this https://dl-cdn.alpinelinux.org/alpine/v3.16/main/aarch64/linux-firmware-cypress-20220509-r1.apk and the brcm one together 2022-10-12 15:05:53 https://www.irccloud.com/pastebin/uzoD0IY3/ 2022-10-12 15:06:09 ah right 2022-10-12 15:06:25 the memory thing is readonly 2022-10-12 15:06:29 ahh filesystem is RO 2022-10-12 15:06:30 right 2022-10-12 15:07:59 ` mount / -o rw,remount` doesn't seem to do the trick 2022-10-12 15:09:34 i forgot the magic invocations 2022-10-12 15:09:53 iirc there is no disk, so your 512m of memory is the limit 2022-10-12 15:10:03 i tried invoking various gods but apparently they don't like raspberry pi's 2022-10-12 15:10:07 so you probably have no more space unless you mount into a real disk 2022-10-12 15:10:30 so i have to proceed with a real setup before having net working i guess 2022-10-12 15:10:43 or you can copy the uh 2022-10-12 15:10:58 brcmfmac43436* 2022-10-12 15:11:00 set of files 2022-10-12 15:11:01 more or less 2022-10-12 15:11:06 into the correct folder 2022-10-12 15:11:08 and nothing else 2022-10-12 15:11:20 maybe it works, maybe it doesn't 2022-10-12 15:12:17 i have the cypress firmware thing 2022-10-12 15:12:27 does apk "extracts" as well? 2022-10-12 15:12:35 or are apks tarballs or something 2022-10-12 15:13:44 just a targz yeah 2022-10-12 15:13:53 alright let's see 2022-10-12 15:14:01 in the meantime I created a ext4 partition just in case 2022-10-12 15:28:32 no way 2022-10-12 15:28:45 you can drop firmware files in `firmware` o nthe boot partition 2022-10-12 15:31:44 yay got internet 2022-10-12 15:31:47 let's get this bread now 2022-10-12 15:41:02 grats 2022-10-12 18:06:03 is there a way to prevent elogind-daemon (launched by an openrc service) from starting? 2022-10-12 18:06:19 it comes with kwin and i don't want it to be running 2022-10-12 18:13:25 vyivel: disable the elogind service? 2022-10-12 18:17:01 minimal: how exactly? `rc-update del elogind -a` says it isn't in any runlevel; or do you mean something else? 2022-10-12 18:19:51 vyivel: perhaps some other service has declared a "need" for it 2022-10-12 18:20:48 kscreenlocker depends on it apparently, but i don't need its functionality either 2022-10-12 18:22:54 vyivel: it depends on the elogind package, that is not the same as another init.d server needing the elogind init.d service 2022-10-12 18:23:25 try grepping for "elogin" in /etc/init.d/ to see if any other service references it? 2022-10-12 18:23:48 nope, no references 2022-10-12 18:26:39 try "rc-status -a" to see if elogind shows up in the "needed/wanted" section? 2022-10-12 18:28:04 it doesn't appear there at all 2022-10-12 18:28:12 i can only see it in `rc-service -l` 2022-10-12 18:28:50 i also can't really tell when does it launch 2022-10-12 18:29:13 that only shows what services are available, not which ones are enabled 2022-10-12 18:29:27 so how do you know it is being run? 2022-10-12 18:30:33 it's in htop (elogind-daemon), and i know when it's running because it puts my pc back to sleep when it wakes up (as it has its own lid close handler) 2022-10-12 18:30:48 it's not running right now btw 2022-10-12 18:31:00 vyivel: I'm asking how do you know that the elogind init.d service is being run, not the elogind binary 2022-10-12 18:31:24 hm 2022-10-12 18:31:31 maybe i don't 2022-10-12 18:31:36 could something be running that binary directly? 2022-10-12 18:31:46 ...don't think so? 2022-10-12 18:32:58 if the elogind service is running then "rc-status -a" should give some info about how it is run. Try that next time you see 2022-10-12 18:33:00 the elogind initscript in aports uses supervise-daemon, so if it was started by the initscript, supervise-daemon would be its parent process (htop tree mode is great for this) 2022-10-12 18:33:32 iirc it was a child of `init` last time i saw it running 2022-10-12 18:33:47 so not launched by the initscript then 2022-10-12 18:34:01 hang on, "elogind-daemon"? I don't see that file in any Alpine package in Edge. the elogind package has a elogind file, not elogind-daemon 2022-10-12 18:34:24 (also /etc/init.d/elogind status would only show 'running' if it was started by the initscript) 2022-10-12 18:35:02 i'm fairly sure it was elogind-daemon 2022-10-12 18:35:50 so what does "which elogind-daemon" give? 2022-10-12 18:35:54 minimal: elogind sets its process name to elogind-daemon 2022-10-12 18:36:03 ^ 2022-10-12 18:36:09 dwfreed: ah, ok 2022-10-12 18:36:26 minimal: nothing, the executable is /usr/libexec/elogind/elogind 2022-10-12 18:37:19 vyivel: the point is next time you boot alpine if the elogind processing is running and "rc-status -a" doesn't show elogind then it wasn't started via the init.d script 2022-10-12 18:37:45 okay 2022-10-12 18:37:56 something else is starting it then 2022-10-12 18:38:23 it would also only say elogind-daemon as process name if it was daemonized, and the initscript does not do that (uses supervise-daemon instead) 2022-10-12 18:39:26 seems like /usr/share/dbus-1/system-services/org.freedesktop.login1.service is responsible for that 2022-10-12 18:41:28 dbus auto-activation 2022-10-12 18:46:19 added a dbus config to blacklist the service, hopefully it'll work 2022-10-13 03:11:44 can vi wrap lines? also in vi command line can it navigate backwards in the line? eg go back to foo to change it in :foo bar 2022-10-13 03:19:11 I don't know about vi, but vim has the textwidth setting 2022-10-13 03:20:00 you could always pass the buffer through the fold command (ie :%!fold -sw 80 ) 2022-10-13 03:23:23 dwfreed: hmm doesnt seem to work for busybox vi 2022-10-13 03:24:08 if you're planning to use these fancy tricks for actual editing i recommend getting a real editor and not a 4kb minimalist accomplishment for "can input some characters" 2022-10-13 03:24:22 vim itself isn't very big 2022-10-13 03:26:18 yeah... 2022-10-13 03:27:41 > can input some characters 2022-10-13 03:27:48 indeed, tested with some symbols :p 2022-10-13 03:28:28 (i do use busybox vi all the time but wanting it to autowrap is pushing it a little) 2022-10-13 03:28:36 granted, you can always use fold on the file after saving 2022-10-13 03:33:32 its suprisingly functional tbh 2022-10-13 05:18:31 Putting .xpi file in /usr/lib/firefox/browser/extensions/ worked today! \o/ 2022-10-13 05:18:56 With no changes to firefox buildconfig 2022-10-13 05:19:41 I think it did not work before because I might have had a firefox window open in some other workspace. Killing all windows and starting firefox shows the extension. 2022-10-13 05:21:29 what firefox do you have installed 2022-10-13 05:21:33 pkgver-pkgrel 2022-10-13 05:21:45 firefox-105.0.3-r1 2022-10-13 05:21:53 i added the sideload in -r1 2022-10-13 05:22:13 Oh! Thank you so much! 2022-10-13 05:22:50 during the conversation we had before 2022-10-13 05:22:53 somehow forgot to mention it 2022-10-13 05:22:59 maybe i was going to bed or something 2022-10-13 05:23:53 I tried building it on my laptop as you suggested, but I gave up because it was already 3 hours and the compilation seemed frozen. 2022-10-13 05:24:47 Anyhow. Thanks :) 2022-10-13 05:26:43 the final part takes a while, if you build it yourself for testing you can drop the --enable-lto=cross 2022-10-13 05:27:55 ack 2022-10-13 10:52:43 anyone managed to install windows 11 in virt-manager? it says "This PC doesn't meed the minimum system requirements" 2022-10-13 10:52:54 i have added TPM 2.0 2022-10-13 11:01:38 There is a /var/log/swtpm/libvirt/qemu/win11-swtpm.log and there seems to be no errors with the TPM 2022-10-13 12:00:21 are you sure it's complaining about tpm? 2022-10-13 12:00:27 ncopa: ^ 2022-10-13 12:00:53 who can be sure about anythign regarding windows :p 2022-10-13 12:01:07 me 2022-10-13 12:01:11 I followed the sysmode on raspberry pi guide on the wiki, and I am getting an error on boot that "filesystems couldn't be fixed". I ran e2fsck on the ext4 partition and fsck.vfat on the vfat partition. It didn't help. What else should I do? 2022-10-13 12:12:32 panekj: no. i have no idea what it complains about 2022-10-13 12:13:03 if you didn't gave it enough space it can also fail 2022-10-13 12:15:26 maybe it does not find the disk due to virtio disk 2022-10-13 12:16:05 no it is SATA disk in virt-manager 2022-10-13 12:16:37 allocated 64G. The error happens right before the disk selection 2022-10-13 12:17:47 does it meet the CPU requirement? (does libvirt pass through all the CPUID etc?) 2022-10-13 12:18:13 this is what i am following. the diff in my setup is that I use raw lvm disk 64G https://www.youtube.com/watch?v=i-OHcENVMG0 2022-10-13 12:23:13 I have host CPU in host-passthrough 2022-10-13 12:23:23 4 vCPUs 2022-10-13 12:23:49 Video is Virtio 2022-10-13 12:24:53 im gonna test qco2 disk now 2022-10-13 12:24:58 qcow2 2022-10-13 12:27:02 it does not even show the license agreement 2022-10-13 12:34:29 ncopa: how does win11 see cpus? It may complain at too old processor. It's my guess though 2022-10-13 12:43:13 no idea how it sees it. its in host-passthough mode, so it should see it as i9-10900 2022-10-13 13:01:45 if only it could just mention _why_ it does not meet minimual requirements 2022-10-13 13:02:15 ikke: I like to keep my cards close to my chest ;-) 2022-10-13 13:03:51 :D 2022-10-13 13:03:59 cache_return: that message is coming from /etc/init.d/fsck so it is after the rootfs is already mounted (by initramfs' init). What entries do you have in /etc/fstab? 2022-10-13 13:06:29 minimal: I have restarted the intallation since last night. I am currently at the step where you run setup-disk -m sys /mnt 2022-10-13 13:07:05 it finishes with the message "you might need to fix the MBR to be able to boot" 2022-10-13 13:07:12 what does that mean? 2022-10-13 13:09:17 It gave the same message last time, but as you said, it did boot after this. 2022-10-13 13:13:54 cache_return: that message is coming from setup-disk but not sure the exact situation that is triggering it 2022-10-13 13:15:36 I'm going to ignore it for now 2022-10-13 13:15:56 I'll try to boot sysmode and let you know if I get another fsck error 2022-10-13 13:17:03 tangent - is 'ln -s media/mmcblk0p1/boot boot' correct? should it have a leading '/'? 2022-10-13 13:17:58 In the current environment of the howto, the absolute path will be /mnt/media..., but in the new environment, it will map to /media/ 2022-10-13 13:18:11 but I don't know how ln -s really works 2022-10-13 13:18:42 cache_return: I assume that if run in the /mnt directory and the issue is to have the created link be /mmt/boot which when booted is actually then /boot 2022-10-13 13:19:00 s/if /is / 2022-10-13 13:19:01 minimal meant to say: cache_return: I assume that is run in the /mnt directory and the issue is to have the created link be /mmt/boot which when booted is actually then /boot 2022-10-13 13:19:15 yes 2022-10-13 13:19:19 so run the command as is 2022-10-13 13:20:34 would 'ln -s /media/mmcblk0p1/boot boot' give the correct absolute path with the new root environment? 2022-10-13 13:21:57 cache_return: if you are in /mnt when you rrun that command then it file create a boot softlink in /mnt, when you boot that softlink will then be /boot 2022-10-13 13:22:41 yes, I understand that part of it, I am asking about the relative (media...) vs absolute (/media...) paths 2022-10-13 13:31:34 cache_return: once it has booted the storage's partitions are mounted as /media/mmcblk0p1 & /media/mmcblk0p2 typically AFAIK. In your case using eMMC then yes I guess you replace mmcblk0 with sda or sdb or wherever your eMMC is mounted. I guess you use a relative path for that as this is happening inside /mnt (does the "installer" system even have /media/mmcblk0p1, or sda/sdb in your case, mounted?) 2022-10-13 13:32:18 yes, it's mounted earlier in the how-to 2022-10-13 13:32:26 https://wiki.alpinelinux.org/wiki/Classic_install_or_sys_mode_on_Raspberry_Pi 2022-10-13 13:41:34 okay, new installation, same problem. File system couldn't be fixed 2022-10-13 13:43:14 I have four entries in fstab - / which is by UUID, /media/mmcblk0p1, /dev/usbdisk. and tmpfs 2022-10-13 13:44:16 I originally added mmcblk0p1 by UUID too, but that failed with the same error, so I tried with the same line as the wiki, but it fails too. 2022-10-13 13:44:34 mounting manually after boot works as expected 2022-10-13 13:50:25 http://sprunge.us/t3e2M1 2022-10-13 13:55:26 cache_return: minor point, the usbdisk entry in /etc/fstab should have the last value be "1" if you want it to be fsck'ed at boot 2022-10-13 13:56:06 shouldn't you also change the mmcblk0p1 reference in fstab to be sda1 or sdb1 depending on which device the eMMC appears as? 2022-10-13 13:57:50 usbdisk was auto-generated, but point taken. 2022-10-13 13:58:34 mmcblk0p1 is correct. That's how it appears in /dev when in the ext4 root. 2022-10-13 13:58:51 And I can mount it manually that way 2022-10-13 14:00:49 cache_return: ok, am just confused as you originally indicated that eMMC was appearing as sdb when you originally copied the files across 2022-10-13 14:02:54 forget about that. we were having a misunderstanding 2022-10-13 14:37:44 I ran the fsck init script with -vd and got this https://ibb.co/rfw2210 2022-10-13 14:45:10 cache_return: do you have the dosfstools package installed? 2022-10-13 14:48:11 afk. I'll check when I get back. 2022-10-13 14:48:30 I would think not, though. 2022-10-13 14:49:06 that package contains the fsck.fat and fsck.vfat programs, without those then obviously the fsck won't succeed 2022-10-13 14:54:04 in addition to this error, sshd is not coming up. Is there a way to bypass fsck in order to get a more functional sysyem while I troubleshoot? 2022-10-13 14:55:50 cache_return: sshd is not coming up? or it just isn't configured to let you in remotely (via password or key)? 2022-10-13 14:57:37 I configured it with an overlay to come up automatically for the install 2022-10-13 14:58:23 it works until I try to reboot into sysmode 2022-10-13 14:59:24 I am suspecting that the startup was exiting before sshd started because of fsck error 2022-10-13 15:19:39 cache_return: I just realised I was incorrect earlier, the initramfs' init mounts the rootfs readonly, then after /etc/init.d/fsck succeeds then /etc/init.d/root remounts it read-write - if it never changes to rw then the sshd init.d will be unable to create SSH hosts keys 2022-10-13 15:21:02 so I'd guess it is rootfs, not the bootfs that is cannot be fsck'ed - though why is not clear - when the setup-alpine finished and you rebooted the fs should have been cleanly unmounted 2022-10-13 16:27:43 so I installed fish on a fresh 3.16.2 install a couple days ago and I'm seeing these extra characters very frequently, like when I first log in (or `cd ~`) I see 004h and then the expected prompt, which by default is bright green user, uncoloured @, brown hostname, green path, uncoloured > 2022-10-13 16:29:19 is there anything I can do to make that stop happening? it didn't happen with busybox ash and it doesn't happen on other distros I use fish on (ubuntu, macOS, Adรฉlie). 2022-10-13 16:31:30 hmm. it also only happens in alacritty on Windows, so that's interesting. 2022-10-13 18:02:45 correct TERM? 2022-10-13 18:05:29 can the busybox fsck even check ext4 filesystems? 2022-10-13 18:17:02 what is going on here?: I am using the busybox fsck to try to check partitions, but all it does is return the busybox version 2022-10-13 18:17:24 I mean the filesystems in the partitions 2022-10-13 18:20:11 even with -V it doesn't really tell me if it is succeeding 2022-10-13 18:43:54 hello, how do you upgrade a diskless sys with lbu when there is a new major release of alpine? 2022-10-13 18:44:52 you would need to update the boot medium 2022-10-13 18:45:21 ikke: ok, that is enough? 2022-10-13 18:45:30 yes 2022-10-13 18:46:14 thanks, if instead I want to just upgrade the kernel before a release? I guess it's not that easy? 2022-10-13 18:46:47 There is update-kernel, but it can take quite some memory 2022-10-13 18:47:47 ballpark? 2022-10-13 18:48:05 I've heard figures of 8G, but not sure 2022-10-13 18:48:56 gosh, can there be security fixes to the kernel without there being a release? 2022-10-13 18:49:27 potentially, we do not explcitly track kernel issues 2022-10-13 18:51:45 ok thanks, I'm not sure this is great for an internet facing server 2022-10-13 18:59:31 I think it's the case, but just to make sure: Am I be able to upgrade the boot media from within the system while running it? 2022-10-13 19:00:18 We have done that plenty of times 2022-10-13 19:00:57 You need to stop the modloop service to be able to unmount the modloop 2022-10-13 19:04:57 ok, I thought the whole disk was loaded into ram so there was no need to do anything 2022-10-13 19:05:49 The rootfs is running in ram, but it uses a modloop for the kernel modules to reduce the amount of ram it's using 2022-10-13 19:07:14 is it possible to go to a previous line in a heredoc? 2022-10-13 19:07:59 afaik no 2022-10-13 19:08:50 not while writing it the first time, but you can while editing it the second time 2022-10-13 19:09:40 dwfreed: editing it the second time? 2022-10-13 19:09:48 frag: cancel, arrow up 2022-10-13 19:09:50 when you go back to it in history 2022-10-13 19:10:07 can't speak to busybox ash, but in bash, if you end a command with \ and ctrl+c in the continuation, it saves the command in history even though it wasn't run 2022-10-13 19:11:18 oh :] 2022-10-13 19:13:08 im surprised ^v and another key(comb) doesnt work in ash tho 2022-10-13 19:13:41 I mean, it's busybox, so you're getting the MVP 2022-10-13 19:14:21 but does busybox try to be posix? 2022-10-13 19:14:35 ish? 2022-10-13 19:14:59 :]] 2022-10-13 19:15:14 but see what we said yesterday about vi :P 2022-10-13 19:15:28 :} 2022-10-13 19:16:02 ash seems to work pretty good too imo 2022-10-13 19:17:38 is it normal that `lbu st` does not print anything while `lb diff` does? Shouldn't st list everything that changed? 2022-10-13 19:22:59 it happened before, now I cannot reproduce, maybe I was just not being attentive 2022-10-13 19:24:53 okay, if I put all partitions in fstab and run the fsck init script before trying to switch to sysmode, it passes 2022-10-13 19:25:27 but if I follow the sysmode tutorial and reboot, I get an error at fsck 2022-10-13 20:45:14 what is this error? https://bpa.st/WTEA 2022-10-13 20:47:13 either bad internet or out of date certificates 2022-10-13 20:49:09 what if it's that the system time is wrong? 2022-10-13 20:49:24 or that 2022-10-13 20:49:46 yep, time is 10 hours slow 2022-10-13 20:49:55 how do I fix that? 2022-10-13 20:51:17 ntpd -q just gives me the busybox version 2022-10-13 20:52:25 same with -ddnq 2022-10-13 20:56:49 busybox ntpd -q should work 2022-10-13 20:57:45 wait. _that's_ how you're supposed to do it? 2022-10-13 20:58:09 no, ntpd -q does the same if ntpd is a symlink to busybox 2022-10-13 20:58:46 well neither one works 2022-10-13 20:59:07 cache_return: fsck is a frontend to various fs-specific fsck binaries 2022-10-13 20:59:26 does it actually print just the version ..? that's weird 2022-10-13 20:59:48 so for ext4 you need the e2fsprogs packagw installed to get fsck.ext4 2022-10-13 21:00:44 minimal thanks, I was wondering about that 2022-10-13 21:01:04 ntpd -q and busybox ntpd -q print nothing 2022-10-13 21:01:06 cache_return: your RPI doesn't have a RTC does it? 2022-10-13 21:01:30 they're not supposed to print anything so that sounds fine to me 2022-10-13 21:01:56 but it doesn't set the time 2022-10-13 21:02:32 you don't need an RTC to keep the time 2022-10-13 21:02:47 Oh, lol, it's off by ten hours but it also thinks it's may 2nd 2022-10-13 21:02:50 cache_return: a standard RPI has no RTC to remember the time when the RPI is turned off 2022-10-13 21:03:01 heat: first of all, I do have an rtc 2022-10-13 21:03:31 the problem here is setting the time, though 2022-10-13 21:03:41 I will worry about retaining it later 2022-10-13 21:03:45 cache_return: aftermarket RTC? 2022-10-13 21:04:53 cache_return: if it think today is the 2nd May then it's not off by 10 hours it is off by 5+ months... 2022-10-13 21:04:57 cm4 carrier board strikes again 2022-10-13 21:05:36 regular ntpd had a flag for large time change, is that the issue here? 2022-10-13 21:06:08 are you running ntpd as root? 2022-10-13 21:06:21 yes 2022-10-13 21:06:35 does passing -d say anything funny? 2022-10-13 21:06:46 cache_return: so what RTC chip does the IO board have? a driver for it perhaps is not compiled, or if it is the module may not be loaded... 2022-10-13 21:07:28 I can't add any packages becasue I can't set the time 2022-10-13 21:08:13 minimal: there definitely is no /dev/misc/rtc 2022-10-13 21:08:29 cache_return: can't you set system clock via "date"? 2022-10-13 21:08:54 maybe, let's see 2022-10-13 21:09:34 cache_return: "date" is the standard Lniux day to set system clock (not the RTC) 2022-10-13 21:09:42 s/day/way/ 2022-10-13 21:09:42 minimal meant to say: cache_return: "date" is the standard Lniux way to set system clock (not the RTC) 2022-10-13 21:11:43 okay, date worked, ntpd did not 2022-10-13 21:11:49 we are back on the internet 2022-10-13 21:12:01 apk is working 2022-10-13 21:12:34 I thought that chrony was supposed to solve this 2022-10-13 21:13:12 cache_return: CM4 IO board has a PCF85063AT RTC chip, which uses the pcf85063 driver, try doing a "modprobe pcf85063" 2022-10-13 21:13:37 cache_return: chrony was supposed to solve what exactly? 2022-10-13 21:15:10 set the system time at boot if no rtc detected 2022-10-13 21:17:29 so right now I am connected over ssh, the system is diskless, I have not attempted to transfer to sysmode again. 2022-10-13 21:18:32 I did setup-alpine and tried to fsck the partitions manually to see if there was an error, but it seemed to work 2022-10-13 21:18:48 the results were not super clear to me, however 2022-10-13 21:23:57 cache_return: chrony or nttpd *can* set the system time once the machine boots AND gets an Internet/network connection, and ntpd or chronyd start, AND if they are configured to permit a large jump in time (if required) 2022-10-13 21:26:09 have you tried loading the RTC module to set if the RTC is detected? 2022-10-13 21:31:09 I tried it, but modprobe failed because there is no /lib/modules 2022-10-13 21:35:25 cache_return: this is run-from-ram? 2022-10-13 21:47:46 yes, like I said above, it's the initial install before the sysmode conversion. 2022-10-13 22:00:45 the pcf85063 kernel module appears to be present in the modloop-rpi4 file so I guess you could edit cmdline.txt to add "pcf85063" to the modules list 2022-10-13 22:06:03 I'll try that. Any thoughts on the fsck thing? Should I just install dosfstools and e2fsprogs then try sysmode? 2022-10-13 22:06:42 ^ does the setup-disk process commit the installed packages so they survive reboot? 2022-10-13 22:07:41 cache_return: I was offline earlier for a while so don't know and you found anything further, I'm suspect the rootfs was having the fsck issue for some (unknown) reason 2022-10-13 22:09:35 I was suspectng that the predicate to fsck for ext4 was not installed 2022-10-13 22:10:20 so I was trying to run the init script before transitioning to sysmode to see if it would work, but results inconclusive. 2022-10-13 22:12:28 cache_return: if you run the init.d/fsck script in run-from-ram mode then you're not checking the sys-mode rootfs (as it is /mnt, not the "/" listed in the run-from-ram's /etc/fstab) 2022-10-13 22:14:42 cache_return: to check if the fsck programs have been installed for sys-mode, look for /mnt/sbin/fsck.* after running setup-alpine 2022-10-13 23:16:52 minimal: I was thinking that it would check the ext4 part if I added it to fstab. 2022-10-13 23:23:10 cache_return: you could just run fsck.ext4 manually to check 2022-10-13 23:24:34 I will try installing everything before converting to sysmode and we will see. 2022-10-13 23:24:54 this time, I'm putting it in a script to save time next time. 2022-10-14 05:35:59 hey all, trying to use Window Maker but its not coming up as a choice in the session list 2022-10-14 05:50:17 brb 2022-10-14 07:04:01 morning. I figured why I cannot install win11 in qemu. It really is TPM 2.0 that is not found 2022-10-14 07:05:38 it is possible to disable the check 2022-10-14 07:05:41 TPM: TPM not detected 2022-10-14 07:06:06 i wonder how to fix TPM. 2022-10-14 07:07:07 qemu can emulate TPM? 2022-10-14 07:07:13 yes 2022-10-14 07:07:17 swtpm 2022-10-14 07:07:24 and UEFI with secure boot 2022-10-14 07:07:28 at least that is what i understand 2022-10-14 07:07:31 did you start swtpm first and then pass it to qemu with -tpmdev emulator,id=tpm0,chardev=chrtpm or whatever 2022-10-14 07:07:53 i believe libvert does that for me 2022-10-14 07:08:29 it would be under in the config 2022-10-14 07:08:34 if not, probably not 2022-10-14 07:08:44 i explicitly added it 2022-10-14 07:08:52 and selected 2.0 2022-10-14 07:09:06 2022-10-14 07:09:06 2022-10-14 07:09:06 2022-10-14 07:09:06 2022-10-14 07:09:31 21239 ? S 0:00 /usr/bin/swtpm socket --ctrl type=unixio,path=/var/run/libvirt/qemu/swtpm/14-win11-swtpm.sock,mode=0600 --tpmstate dir=/var/lib/libvirt/swtpm/9a62c23e-351a-4f2e-8b83-31101f375b59/tpm2,mode=0600 --log file=/var/log/swtpm/libvirt/qemu/win11-swtpm.log --terminate --tpm2 2022-10-14 07:10:35 could try tpm-crb instead of -tis i guess, not many other ideas 2022-10-14 07:10:58 ... -chardev socket,id=chrtpm,path=/var/run/libvirt/qemu/swtpm/14-win11-swtpm.sock -tpmdev emulator,id=tpm-tpm0,chardev=chrtpm -device {"driver":"tpm-tis","tpmdev":"tpm-tpm0","id":"tpm0"} ... 2022-10-14 07:11:14 i tried both 2022-10-14 07:11:40 /usr/share/qemu/edk2-x86_64-secure-code.fd 2022-10-14 07:11:57 so i wonder if maybe edk2 or swtpm is broken 2022-10-14 07:12:09 could be 2022-10-14 07:14:32 i wonder if i need to enable it in boot menu in tianocore 2022-10-14 08:20:13 ok. i have booted alpine-standard in the same VM, and it does have a /dev/tpm0 device. The qemu tpm setup works, but windows 11 does not find it 2022-10-14 08:20:40 could it be that windows does not like or accept the certs? 2022-10-14 08:27:27 cache_return: the date issue, i mentioned here few days ago. the virt kernel setup date using hwclock -s, but the lts/edge kernel does not do that... 2022-10-14 08:27:59 not sure 2022-10-14 08:29:55 even if one is booting lts/edge ISO from a vm, the date is not synced during boot. only ISO with virt kernel sync date during boot. 2022-10-14 08:34:43 there are 2 solution, as a user : 1. run hwclock -s or date xxxxxxxxx the first time ISO is boot , OR 2. use http repo . As alpine developer : make add a hwclock daemon to lts/edge ISO 2022-10-14 08:40:25 but oftern, fresh users do not notice the incorrect date, and then SSL issue, and then spend hours figuring out what was the causes. that is my first experience with alpine a year ago, and i gave up... easily. until recently i decide to try alpine again. 2022-10-14 09:53:19 tpm 1.2 works but something with tpm 2.0 seems to be broken 2022-10-14 09:53:29 https://github.com/stefanberger/swtpm/issues/669 2022-10-14 12:02:41 I'm on the way out of the door, but just catched this link right now. It's not my style to just post things in random channels, but this might be real serious and it'd be nice if someone picked it up so it doesn't get lost: https://www.openwall.com/lists/oss-security/2022/10/13/2 Sorry. 2022-10-14 12:06:57 thats for #alpine-security. I did push linux-lts 5.15.73. I havent checked if it fixes those 2022-10-14 12:07:12 re my TPM 2.0 problem in qemu, I figured out what the problem is 2022-10-14 12:07:34 its a regression in qemu's edk2 due to config knob name change in edk2 2022-10-14 12:13:30 it doesn't, .74 which is not out yet 2022-10-14 12:13:54 ok 2022-10-14 12:40:51 uuidNuniq: yes, that tracks. 2022-10-14 12:42:32 I now have a working sysmode installation, but the time is wrong on every boot. I tried to configure chronyd to fix this by adding 'maxchange 1000 1 2' to the conf, but that doesn't work on boot, it only works when I manually restart chronyd. Whats up with that? 2022-10-14 13:05:15 Hey yโ€™all. Everyone here has good taste in linux distributions. 2022-10-14 13:06:05 For some reason the tftp prompt provided by the tftp-hpa package is kind of broken 2022-10-14 13:06:15 Does anyone here know why thatโ€™s the case? 2022-10-14 13:10:18 i have a hunch 2022-10-14 13:14:30 try -r5 of it from edge 2022-10-14 13:14:38 when it builds 2022-10-14 13:20:38 uniqNuuid: when you raised this issue several days agO I suggested you check "dmesg" output to see if the linux-virt kernel sets the system clock from the RTC, I don't believe you gave any indication whether you checked this 2022-10-14 13:22:57 cache_return: I already suggested you modify the cmdline.txt modules entry to load the RTC driver module... 2022-10-14 13:32:55 cache_return: re chrony, you'd want to use "initstepslew" in your chronyd.conf file 2022-10-14 14:23:16 minimal: here is the rtc update. I tried adding to cmdline in the diskless system, it did not create a rtc device. I also tried to modprobe pcf85063, but that failed (not in /lib/modules I think) 2022-10-14 14:24:07 the sysmode update is that sysmode conversion works if you apk add e2fsprogs before setup-disks 2022-10-14 14:24:46 So now I have a sysmode installation and I'm happy with that. Thanks for all help. 2022-10-14 14:25:20 I can try adding pcf85063 to sysmode cmdline. Stand by. 2022-10-14 14:28:06 cache_return: setup-disks specifically does a "apk add e2fsprogs" for a ext2/3/4 rootfs so I don't see why that appears to be going wrong in your scenario 2022-10-14 14:28:38 https://gitlab.alpinelinux.org/alpine/alpine-conf/-/blob/master/setup-disk.in#L815 2022-10-14 14:30:27 as for the RTC, adding it to the modules list should have made the RTC device available. Did "dmesg" show anything relating to the RTC? One potential solution would be for the linux-rpi4 armv7/aarch64 kernels to build the PCF module inside the kernel 2022-10-14 14:32:04 did you just add that line? 2022-10-14 14:32:55 cache_return: i don't have a CM4 to test 2022-10-14 14:34:07 cache_return: ah, from looking at RPI Forums it seem you may also need to add 2 lines to config.txt: 2022-10-14 14:34:17 "dtparam=i2c_vc=on" 2022-10-14 14:34:17 I mean, are you referring to a line that is in setup-disk already or did you commit it? 2022-10-14 14:34:25 it's been there for 6+ years 2022-10-14 14:34:51 "dtoverlay=i2c-rtc,pcf85063a,i2c_csi_dsi" 2022-10-14 14:35:26 cache_return: for setup-disk I'm referring to a line that is already present 2022-10-14 14:36:17 I dont know what to make of that. 2022-10-14 14:37:34 cache_return: when you do a Sys install it uses ext4 by default and so should auto-add e2fsprogs 2022-10-14 14:39:36 I have no idea why I had to do it manually. 2022-10-14 14:40:42 cache_return: you said you were scripting things, have you a script to show? 2022-10-14 15:32:48 is there an universal way to copy the line you have in commandline to primary (without mouse)? 2022-10-14 15:41:03 minimal: not a real script, but just a list of the commands. Will share later. 2022-10-14 15:43:28 those two lines in usercfg.txt and now the hwclock is online. 2022-10-14 15:43:47 but the system time is not being set from hwclock at boot 2022-10-14 15:45:09 cache_return: is the hwclock init.d script enabled? normally for RPIs the swclock script is enabled instead - disable that and enable hwclock 2022-10-14 15:55:45 that was it. First time running rc-update 2022-10-14 15:55:59 wow, that was better than systemd 2022-10-14 16:56:23 cache_return: good news 2022-10-14 16:57:43 cache_return: wondering whether a linux-rpicm4 kernel package should be created rather than adding some cm4-specific drivers to the linux-rpi4 kernel package 2022-10-14 17:09:39 it seems like just configuration though, right? 2022-10-14 17:11:25 unrelated - what should be in /boot/boot fpr a sysmode installation? After following the guide, I have a lot of duplicates between /boot and /boot/boot 2022-10-14 17:11:55 minimal: someone would need to maintain it 2022-10-14 17:18:58 ikke: yeah, the other option is adding (some) more modules compiled into the linux-rpi kernel that only apply for the CM4 2022-10-14 17:57:27 do I need to figure out the wifi hardware and packages for it manually? 2022-10-14 17:57:59 it's probably a difference of at best 100kb of modules 2022-10-14 17:58:43 I have the CYW43455, I think the driver is called brcmfmac. I see a linux-firmware-brcm and linux-firmware-cypress, but I'm not sure what package provides what 2022-10-14 17:59:25 apk info not very helpful 2022-10-14 17:59:52 https://pkgs.alpinelinux.org/contents?file=&path=&name=linux-firmware-brcm&branch=edge&arch=x86_64 2022-10-14 18:02:23 Thanks, for some reason I was getting zero search results on that page 2022-10-14 18:06:08 brcm seems to have brcmfmac43455 2022-10-14 18:07:49 Yes, I have it installed now. It even has a nice note that says Compute Module 4 right on it 2022-10-14 18:08:08 I just have to figure out the rest. 2022-10-14 18:12:18 I thought just adding the packace would give me a /dev device, but this is not the case 2022-10-14 18:14:40 you'd have to reload the module so it even tries to pick up the firmware 2022-10-14 18:15:22 the brcm one with modprobe/modprobe -r 2022-10-14 18:15:29 and perhaps something else in that chain 2022-10-14 18:16:06 I rebooted 2022-10-14 18:16:10 let me reload 2022-10-14 18:17:36 no dice 2022-10-14 18:18:52 cache_return: check "dmesg" output to see what it says? 2022-10-14 18:20:27 direct firmware load failed with error -2 2022-10-14 18:21:52 cache_return: does it mention a filename/path? does that actually exist? 2022-10-14 18:30:04 the path in the error is brcm/brcmfmac43455-sdio.raspberrypi,4-compute-module.bin 2022-10-14 18:31:45 post the full dmesg 2022-10-14 18:33:13 cache_return: don't see that in the linux-firmware-brcm package contents 2022-10-14 18:33:22 http://sprunge.us/tLvLQA 2022-10-14 18:34:18 minimal, nope: http://sprunge.us/bakyWv 2022-10-14 18:36:21 I only see people posting logs with the same error, but nothing that suggests how to fix it 2022-10-14 18:36:33 ^ that's my /lib/firmware after installing the brcm package 2022-10-14 18:37:13 The question is where to get that blob from 2022-10-14 18:38:22 ikke: there have been issues with the cypress firmware - basically the RIP Foundation did put it in their repo but it didn't go into the general linux-firmware repo, not sure if that has been resolved yet 2022-10-14 18:38:38 questions around the licensing of the Cypress firmware... 2022-10-14 18:38:48 I can grab it from an rpi-os image, but that's... suboptimal? 2022-10-14 18:41:24 https://github.com/raspberrypi/linux/issues/4248 2022-10-14 18:41:41 "When I say that Direct firmware load for brcm/brcmfmac43455-sdio.raspberrypi,4-model-b.txt failed with error -2 is expected I mean that you should see it." 2022-10-14 18:42:39 it should be here: https://github.com/RPI-Distro/firmware-nonfree/tree/bullseye/debian/config/brcm80211/cypress/ 2022-10-14 18:42:54 They are talking about symlinks 2022-10-14 18:45:09 ikke: that is crazy 2022-10-14 18:45:39 I guess it just tries to see if that firmware exists and load it when it does? 2022-10-14 18:45:50 from memory the RPI Foundation were in discussions with someone (EFF? GNU? Linux Foundation?) to sort out the licensing confusion in order for these to appear in the upstream linux-firmware repo 2022-10-14 18:45:57 They are saying that it throws this error in dmesg if the firmware is actually correctly loaded?? 2022-10-14 18:46:19 cache_return: I think they say the module is correctly loaded, which then tries to load firmware 2022-10-14 18:46:44 but I still have no wifi device, so either I am getting the error in error and there is actually an error, or there is another issue 2022-10-14 18:50:57 i guess we need one more symlink 2022-10-14 18:50:59 i'll add that shortly 2022-10-14 18:51:06 actually, it's this. They are saying the error is not causing the problem described, but if you look at my dmesg vs the posted one, their brcm driver does report a firmware version, where mine does not 2022-10-14 18:53:44 also that error is for the .txt, not for the .bin 2022-10-14 18:53:49 not sure if matters 2022-10-14 18:55:20 minimal: https://img.ayaya.dev/mhmYBuJ4Ocac this look ok or nah 2022-10-14 19:03:55 should I just try that? 2022-10-14 19:05:34 by hand? sure 2022-10-14 19:05:39 psykose: I guess so, I've never used Wifi on any of my RPIs so far ;-) 2022-10-14 19:05:43 create the link with the same command 2022-10-14 19:05:52 just delete the $pkgdir part and it should work 2022-10-14 19:05:52 oh fun, firefox showing empty again for that mimetype 2022-10-14 19:05:57 always does 2022-10-14 19:05:58 :) 2022-10-14 19:06:24 the only way to bypass that is that mime ext extension to define actions per type but it gets annoying because you will get prompted for every single new mimetype ever 2022-10-14 19:12:38 I'll see if making that link works 2022-10-14 19:18:37 with the link - see the module reload at bottom http://sprunge.us/urxNFV 2022-10-14 19:18:49 still nothing in /dev though 2022-10-14 19:28:54 hello, dumb question, but anyway: is there any drivers (and mesa/vulkan packages) for nvidia gpus on alpine? 2022-10-14 19:29:31 I feel like everything is running on the intel integrated graphics card, so basically the gpu is just making space hehe 2022-10-14 19:32:31 i dont think the proprietary blob runs, you can try with nouveau maybe 2022-10-14 19:32:47 and the open source nvidia vk driver (nvk) has not been released yet 2022-10-14 19:34:18 this would be the mesa-dri-nouveau package, right? 2022-10-14 19:34:41 no, just gallium 2022-10-14 19:34:44 there is no nouveau 2022-10-14 19:35:17 there is mesa-dri-nouveau-21.3.8-r2 & mesa-dri-gallium-21.3.8-r2 2022-10-14 19:35:24 what happens if I install both haha 2022-10-14 19:35:40 nothing, the former is just a shitty provides= thing kept around for like 10 alpine releases 2022-10-14 19:35:47 ah 2022-10-14 19:36:27 and removed for 3.17 so then you would actually have to delete it, but that's for the release notes 2022-10-14 19:37:46 all right, thanks 2022-10-14 19:38:07 np 2022-10-14 19:43:38 the firmware reports loaded in dmesg, but rfkill is empty and setup-interfaces still only sees eth0 2022-10-14 19:45:35 cache_return: is this different that the last dmesg output you posted? 2022-10-14 19:45:50 also are you running Alpine 3.16.x or Edge? 2022-10-14 19:58:06 3.16 2022-10-14 19:58:31 the most recent one was with the module reload at the bottom 2022-10-14 19:58:45 I have rebooted, I can post that dmesg if you want 2022-10-14 20:09:01 wondering if anyone has a CM4 working with Alpine 3.16.x (e.g. potentially there could be a kernel-related issue affecting Wifi, just don't know) 2022-10-14 20:27:40 should I try edge? will it make a difference? 2022-10-14 20:40:15 minimal: rtc-efi rtc-efi.0: setting system clock to 2022-10-14T12:27:13 // this is linux-virt 2022-10-14 20:41:42 dmesg 2022-10-14 20:42:36 with the rtc, once the modules were loaded, i had /dev/rtc0 and could interect with it. How is it supposed to work with wifi? I was thinking there was a /dev/net/wlan0 or something, but that's actually not what happens, is it? 2022-10-14 20:47:38 uniqNuuid: yes, that is because linux-virt for aarch64 has the UEFI RTC driver into the kernel, whereas linux-lits for asrch64 has it as a loadable module. Nothing to do with "hwclock -s", if it is built into the kernel then the kernel will set system time from the RTC directly 2022-10-14 20:48:00 deja-vu 2022-10-14 20:48:01 s/driver into/driver built into/ 2022-10-14 20:48:01 minimal meant to say: uniqNuuid: yes, that is because linux-virt for aarch64 has the UEFI RTC driver built into the kernel, whereas linux-lits for asrch64 has it as a loadable module. Nothing to do with "hwclock -s", if it is built into the kernel then the kernel will set system time from the RTC directly 2022-10-14 20:52:12 cache_return: I don't know if the kernel/firmware in Edge would make a difference or not 2022-10-14 20:55:44 minimal: understand now. i guess it is possible to load the RTC module into initramfs 2022-10-14 20:56:30 /etc/mkinitfs/mkinitfs.conf ? 2022-10-14 20:58:43 uuidNuniq: try running "mkinitfs -l" which does NOT regenerate but shows you what *would* be placed in the initramfs if it were generated, look for "rtc-efi.ko.gz" 2022-10-14 20:59:17 however you would ALSO need to add "rtc-efi" to the grub or other bootloader's "modules=" list in the cmdline as well 2022-10-14 20:59:44 looks like it is possible, 1) create a file named rpirtc.modules in /etc/mkinitfs/features.d/, which contains the rpi rtc module names and 2) add it to /etc/mkinitfs/mkinitfs.conf 2022-10-14 21:00:31 however you would STILL then need to run hwclock (i.e. enable the hwclock init.d service) for the time to be set as the kernel itself will ONLY auto-set the system time if rtc-efi is compiled into the kernel rather than being a loadable module 2022-10-14 21:01:23 uuidNuniq: what has a RPI got to do with aarch64 UEFI? 2022-10-14 21:01:39 it is not always rtc-efi: on one of my x86 vm : rtc_cmos 00:00: registered as rtc0 2022-10-14 21:02:07 uuidNuniq: you originally were talking about linux-virt and linux-lts for aarch64 2022-10-14 21:02:44 yeah, i am not talking general aarch64 cases :) 2022-10-14 21:02:56 e.g. aarch64 on real hardware 2022-10-14 21:03:06 i am *now 2022-10-14 21:04:31 uuidNuniq: which real hardware? a RPI doesn't use linux-lts or linux-virt and it doesn't boot via UEFI 2022-10-14 21:04:56 ok. so it has another customized kernel. 2022-10-14 21:05:00 so if you have a RPI then which RTC are you using with it? 2022-10-14 21:05:21 anyway, i am good on linux-virt. all my use cases are on virtual machine. 2022-10-14 21:07:10 uuidNuniq: unlike x86/x86_64 machines on Arm machine there is no completely standardised way to present a RTC, there are instead a large number of various RTC chip drivers 2022-10-14 21:07:30 i have a tv box, no rtc either. it uses ntpd to update the clock 2022-10-14 21:08:05 for x86/x86_64 BIOS there is only cmos_rtc (name from memory) and on x86/x86_64 UEFI there is "rtc-efi" 2022-10-14 21:10:48 not sure if that's correct. the x86_64 i am running alpine on, has /sys/firmware/efi/efivars/* and it uses rtc_cmos 2022-10-14 21:11:33 you're not forced to use efi if efi 2022-10-14 21:11:45 there's the bunch of csm garbage 2022-10-14 21:12:02 psykose: CSM is BIOS emulation though 2022-10-14 21:12:09 yea 2022-10-14 21:12:36 right, some efi firmware has a compatible layer CSM 2022-10-14 21:14:51 uuidNuniq: the point being that for PCs there's 2 drivers whereas for Arm there's probably 20-30 or more drivers for individual RTC chips 2022-10-14 21:15:27 yeah, that's true 2022-10-14 23:54:22 what happened? I did service networking restart and it said "warning, you are stopping a boot service" now ssh is down for good? 2022-10-14 23:55:07 that is not what I expected. 2022-10-15 00:08:25 minimal: raspi os all links back to cyfmac43455-sdio.bin and brcmfmac43455-sdio.txt 2022-10-15 00:08:37 as far as I can tell. 2022-10-15 00:10:47 pmaybe I need to install cypress firmware and link back to that. 2022-10-15 00:10:57 when I get networking back on 2022-10-15 00:11:29 cache_return: the warning can be ignored. 2022-10-15 00:12:33 but sshd never came back up 2022-10-15 00:12:39 even after reboot 2022-10-15 00:13:00 rc-update see if it is on? 2022-10-15 00:13:16 it's headless 2022-10-15 00:13:33 is there a file in disk I can check? 2022-10-15 00:13:47 I can mount the fs and fix if so 2022-10-15 00:15:04 ls /etc/runlevels 2022-10-15 00:15:56 ln -s /etc/init.d/sshd /etc/runlevels/default/sshd 2022-10-15 00:16:06 it is a symlink 2022-10-15 00:16:29 if it is not there, just 'ln -s' 2022-10-15 00:45:09 cache_return: init.d/sshd does not depend on init.d networking (unless /etc/ssh/sshd_config has a specific listen address defined which it does not by default) 2022-10-15 00:49:03 minimal: what did I maybe break, then? 2022-10-15 00:50:08 there is no /etc/runlevels 2022-10-15 00:50:45 sorry, my typo. It's there. 2022-10-15 00:52:14 I have .../boot/networking and .../default/sshd 2022-10-15 00:53:50 cache_return: dunno, what's in your /etc/network/interfaces file? 2022-10-15 00:54:17 nothing! 2022-10-15 00:54:26 how did that happen? 2022-10-15 00:58:10 this mirrors some weird problems I was having earlier with e.g. fstab becasue the sysmode conversion left behind two copies of everything. 2022-10-15 00:58:34 that's why I was asking about /boot and /boot/boot earlier 2022-10-15 00:59:00 anyway, that fixed it. I am not going to restart networking again to try to replicate. 2022-10-15 01:06:48 no change with the cypress firmware 2022-10-15 01:10:06 you would only use the cypress firmware for a Cypress-based wifi chip - the RPI spec sheet didn't clearly indicate if CM4 has Broadcom (Infinion? or whatever it's now called) or Cypress chip 2022-10-15 01:12:14 I thought it was clear that it is a cyfmac43455 2022-10-15 01:17:46 cache_return: you're right, I checkied it again, section 2.1 of datasheet, CYW43455 2022-10-15 02:21:13 I think the firmware setup is the same as rpi-os. Is there anything else I should check? 2022-10-15 14:24:11 do somebody can have a look at https://gitlab.alpinelinux.org/alpine/aports/-/merge_requests/40237 2022-10-15 14:24:43 I guess no CVE's available yet? 2022-10-15 14:25:06 it's important ๐Ÿ˜… 2022-10-15 14:25:10 yes 2022-10-15 14:25:12 understood 2022-10-15 14:25:47 and there will likely be 3 2022-10-15 14:25:55 ikke: is in request 2022-10-15 14:29:15 CVE-2022-32149 we got one 2022-10-15 14:29:33 * got one - well it's unrelated to mentioned aboth 2022-10-15 14:31:12 I'll add all of them in a new MR 2022-10-15 14:32:11 That one seems to just affect go itself? 2022-10-15 14:49:07 well the dependency we use, the other issues are gitea related and I just have to request the CVEs 2022-10-15 14:49:12 doing so atm 2022-10-15 15:13:39 3.16 builds are failing due to patches not applyign 2022-10-15 17:45:38 hello folks, i'm trying to do a fresh install of Alpine with an encrypted root / EFISTUB / LVM on LUKS 2022-10-15 17:46:02 this is my usual setup - but normally Arch / GRUB / FDE / LVM on LUKS 2022-10-15 17:47:00 it looks like the initramfs is failing to decrypt the Luks container......and then fails to mount the root device stating 'No such file or directory' 2022-10-15 17:47:14 i can successfully mount everything in the recovery shell 2022-10-15 17:47:43 has anyone any experience with EFISTUB here? 2022-10-15 17:48:48 i'm passing "kms ext4 lvm cryptsetup keymap" to mkinitfs.conf -> but never get a password prompt on the boot sequence (only a 5 second timeout between 3 & 8 secs of boot time before crashing to the shell) 2022-10-15 17:49:28 i've passed both UUID and /dev/ paths to the kernal via EFIBOOTMGR entries, and these appear ok in the recovery shell at /proc/cmdline 2022-10-15 17:49:52 which crypt-related options are you passing on cmdline? 2022-10-15 17:50:26 only cryptdevice=/dev/sda2:lvmcrypt root=/dev/mapper/vg0-root rw 2022-10-15 17:50:39 besides the two initrd statements..... 2022-10-15 17:51:19 that's the same as my GRUB params except I don't specific the root device (believe its in the Grub entiries automatically) 2022-10-15 17:56:47 ah, and i noticed in the recovery shell that mount /dev/mapper/vg0-root /root -> 'no such file or directory' 2022-10-15 17:57:02 but if I mount -t ext4 /dev/mapper/ vg0-root /root it works 2022-10-15 17:57:36 probably not the issue as I don't get a passphrase prompt and I don't see the lvmcrypt in the shell until I manually luksOpen it 2022-10-15 17:58:42 noyau: I use "root=/dev/mapper/cryptroot cryptroot=UUID=....... cryptdm=cryptroot" 2022-10-15 17:59:55 i'll change it real quick.....are you using EFISTUB successfully or a boot loader? 2022-10-15 18:00:11 i also note the hooks in Arch are 'encrypt' where as the Alpine wiki says to use 'cryptsetup' 2022-10-15 18:00:22 what do you use in your mkinitfs.conf? 2022-10-15 18:00:22 this is with a bootloader but I don't think that should matter 2022-10-15 18:02:00 noyau: is "base" missing from your mkinitfs.conf FEATURES list? 2022-10-15 18:02:42 also are you passing "rootfstype=ext4" on cmdline? 2022-10-15 18:04:17 base is there, rebooting.....will specify the rootfs next 2022-10-15 18:08:03 noyau: rootfstype you mean? that is used to tell the initramfs' init which kernel module to load for the rootfs' type 2022-10-15 18:08:49 yes, sorry meant rootfstype 2022-10-15 18:09:04 no change either on the cryptdm=lvmcrypt nor rootfstype=ext4 2022-10-15 18:09:25 i'm really inclined to say EFISTUB doesn't support the cryptdevice unlock 2022-10-15 18:09:55 noyau: ?? The unlock is done by the initramfs' init 2022-10-15 18:10:34 you use the cryptsetup hook right? not 'encrypt'? 2022-10-15 18:10:38 so you have cryptdm=lvmcrypt and also root=dev/mapper/lvmcrypt? 2022-10-15 18:10:53 noyau: I'm not sure what you are referring to 2022-10-15 18:11:11 root has to be the lvmcrypt or the vg0-root? 2022-10-15 18:11:47 root= has to refer to the same name as specified for cryptdm= 2022-10-15 18:12:50 noyau: you can't refer to any LVM naming as that can't be known/seen until AFTER the luks device is unlocked 2022-10-15 18:12:59 appreciate the help.....changing again..... 2022-10-15 18:13:07 would have thought i would get the unlock prompt first..... 2022-10-15 18:13:59 noyau: you get the unlock prompt once it finds the LUKS part you want to unlock, if you don't specified the right cmdline options then it won't know what you want to unlock 2022-10-15 18:14:38 and do you have the correct UUID specified in the "cryptroot=" option? 2022-10-15 18:15:47 noyau: actually what is the name of your PV? 2022-10-15 18:17:34 lvmcrypt same as the guide 2022-10-15 18:17:47 and no joy changing root=/dev/mapper/lvmcreate 2022-10-15 18:18:12 still no prompt for the unlock.....i started with UUID refs but then changes to simple paths 2022-10-15 18:18:26 as above ^ 2022-10-15 18:19:25 noyau: lvmcrypt? lvmcreate? 2022-10-15 18:20:54 really sorry, been working this for 10 hours..... /dev/mapper/lvmcrypt 2022-10-15 18:22:35 noyau: was thinking too much about the way *I* set things up, so try "rootfstype=ext4 root=/dev/mapper/lvmcrypt cryptroot=UUID= cryptdm=lvmcrypt" 2022-10-15 18:23:47 argument order doesn't matter correct? 2022-10-15 18:23:54 nope 2022-10-15 18:24:19 the trouble with having all this scripted is later forgetting the precise details of everything lol 2022-10-15 18:24:53 yeah, also when it just works and you don't have a habit of needlessly rebuilding your systems 2022-10-15 18:25:05 it was the same shit show with Arch for me, but a bit more documentation 2022-10-15 18:25:13 EFISTUB has been on my list of things to try for some time, I just haven't had an opportunity to try it out 2022-10-15 18:25:26 UUID is the /dev/sda2 that is Luks formatted correct? the parent container 2022-10-15 18:25:41 i got to it first :D 2022-10-15 18:25:42 noyau: well the documentation is basically the initramfs' init shell code :-) 2022-10-15 18:25:52 i'm about 2 mins from just dumping syslinux and seeing if that is the issue 2022-10-15 18:26:56 use "lsblk2 to see the UUID for the "crypto_LUKS" entry 2022-10-15 18:27:12 noyau: Syslinux? I'm confused you said you were using EFISTUB 2022-10-15 18:28:06 i am..... because i wanted to experiment away from Grub or alternatives......but if no one can tell me its a working combination on Alpine....process of elimination 2022-10-15 18:28:14 UUID made no difference either.... 2022-10-15 18:28:48 the problme you are having is with the initramfs' init, I don't see why any particular bootloader or EFISTUB would be a fact in this 2022-10-15 18:29:36 well the only other part that is different between the two (arch and alpine) is the mkinitfs script 2022-10-15 18:29:36 by the time the kernel has loaded, initialised and the initramfs is available and its' init run the bootloader/EFISTUB has done its job 2022-10-15 18:29:50 and grub unlocks the partition before passing to the initramfs is my understanding 2022-10-15 18:30:06 which is why you can hav ea fully encrypted (including the boot partition) 2022-10-15 18:30:17 nayau: no, grub does not unlock LUKS for the initramfs 2022-10-15 18:30:54 i'm not on a FDE setup this time around, but when it is a FDE, the only thing unencrypted is the grub.efi... 2022-10-15 18:31:31 Grub unlocks (for itself) whichever filesystem contains /boot so that Grub can get to the kernel and initramfs files, that's all 2022-10-15 18:32:32 i've asked twice, but you are sure your hook is 'cryptsetup' and not 'encrypt' 2022-10-15 18:32:33 Grub's decryption of LUKS to get those files is separate from the initramfs' init's decryption of rootfs 2022-10-15 18:32:45 noyau: what hook? 2022-10-15 18:32:45 i'm surprised the kernal hooks aren't the same regardless of the mkinitfs.conf 2022-10-15 18:33:08 dit the /mnt/etc/mkinitfs/mkinitfs.conf file and append the cryptsetup module to the features parameter: features="... cryptsetup" 2022-10-15 18:33:18 you mean cmdline options? 2022-10-15 18:33:52 after you added cryptsetup to features list did you rebuild the initramfs? 2022-10-15 18:34:22 happy to have my terminology corrected, but there are two areas that have to be edited the kernal paramaters passed to EFISTUB, and the Hooks used by mkinit when building the initramfs 2022-10-15 18:34:40 yes....but i'll do it again just to make sure i didn't miss it 2022-10-15 18:34:58 cryptsetup didn't have to be added manually, it was already there, and i remember running it 2022-10-15 18:35:05 noyau: you mean initramfs features, not hooks 2022-10-15 18:35:22 as there is a special command to pickup the install directory and not the bootable USB 2022-10-15 18:35:38 Grub calls it 'hooks' , yes mkinitfs calls it 'features' 2022-10-15 18:35:39 ? 2022-10-15 18:36:03 noyau: i think you mean dracut calls them hooks (rather than Grub) ? 2022-10-15 18:36:43 I don't understand your sentence about "special command" 2022-10-15 18:37:28 # mkinitfs -c /mnt/etc/mkinitfs/mkinitfs.conf -b /mnt/ $(ls /mnt/lib/modules/) The command uses the settings from the mkinitfs.conf file set in the -c parameter to generate the RAM disk. The command is executed in the /mnt/ directory and the RAM disk is generated using the modules for the installed kernel. Without setting the kernel version using the $(ls /mnt/lib/modules/) option, mkinitfs tries to generate the RAM disk using the kernel version in 2022-10-15 18:37:48 noyau: question, in which filesystem/partition are you kernel and initramfs files? 2022-10-15 18:38:18 in both /boot and also manually copied to /boot/efi as there was a comment about relative to the EFI directory 2022-10-15 18:38:26 noyau: generally you just need to run "mkinitfs" with no options on a running system to rebuild the initramfs 2022-10-15 18:40:01 so you've put them in the ESP partition (/boot/efi) as the they need to be available in an unencrypted location... 2022-10-15 18:40:48 yes they are also in /mnt/boot/efi 2022-10-15 18:40:59 there is also /mnt/boot/efi/EFI as a subfolder in that same directory 2022-10-15 18:41:12 well if that's the only encrypted location then that's the only place they can be loaded from 2022-10-15 18:41:29 its on a separate vfat partition 2022-10-15 18:41:36 s/encrypted/unencrypted/ 2022-10-15 18:41:36 minimal meant to say: well if that's the only unencrypted location then that's the only place they can be loaded from 2022-10-15 18:41:50 yes, that's your ESP partition 2022-10-15 18:42:01 and i wouldn't get a shell if it wasn't loading the initramfs correct? 2022-10-15 18:42:09 yes 2022-10-15 18:42:30 --loader is supposed to be /vmlinuz-lts? 2022-10-15 18:42:43 now when you updated the initramfs (via mkinitfs) after adding the cryptsetup feature, did you copy the updated initramfs file from /boot to /boot/efi/ ? 2022-10-15 18:42:46 apologies if its a newb question, i'm not entirely clear the difference between vmlinuz and the initram 2022-10-15 18:42:53 yes 2022-10-15 18:42:56 vmlinuz is your kernel 2022-10-15 18:43:57 anyway to verify the initramfs is loading the cryptsetup modules? 2022-10-15 18:44:02 in the recovery shell? 2022-10-15 18:44:08 lsmod? 2022-10-15 18:44:48 lsmod would only show what's loaded (which isn't done until nl-plugfs is run by init I think) 2022-10-15 18:45:15 well its not there, and on the other hand it works just fine in the shell.... 2022-10-15 18:45:45 noyau: the shell is a stripped down rootfs running from memory 2022-10-15 18:47:36 noyau: have you had alpine running on this machine without encryption? 2022-10-15 18:47:54 also what is the root device? SATA disk? SATA ssd? NVME ssd? 2022-10-15 18:48:09 no, but the install usb works 2022-10-15 18:48:25 its an older dell latitude (yes that has caused issues in the past :D ) 2022-10-15 18:48:28 sata disk 2022-10-15 18:49:20 so the disk is GPT partitioned/labelled? with 2 partitions - ESP for UEFI, and 2nd for LUKS? 2022-10-15 18:50:47 efi was definitely labelled as esp yes - 2 partitions 2022-10-15 18:51:27 GPT or MBR partitioned? 2022-10-15 18:52:15 gpt 2022-10-15 18:52:21 esp has boot and esp flags 2022-10-15 18:54:24 i liked your question on the type of drive 2022-10-15 18:54:33 i have a raid bus controller / sata controller 2022-10-15 18:54:46 wondering if that requires different 'features' 2022-10-15 18:55:00 drive type was more to do with historic issues related to NVMe (where the module wasn't being loaded) 2022-10-15 18:55:21 what make/model is the RAID controller? 2022-10-15 18:55:37 also, what files are in the ESP partition? 2022-10-15 18:56:11 intel 82801 2022-10-15 18:56:39 to be clear, esp is /boot or /boot/efi 2022-10-15 18:57:22 ESP is the ESP (i.e. 1st) partition regardless of where it may be mounted 2022-10-15 18:57:52 is the controller set to RAID or AHCI in the UEFI settings? 2022-10-15 18:58:58 -> /boot/efi/{initramfs-lts, intel-ucode.img, vmlinuz-lts} 2022-10-15 18:59:06 -> /boot/{system.map-lts, config-lts, efi/, grub/, initramfs-lts, intel-ucode.img, vmlinuz-lts 2022-10-15 19:00:02 raid on.... 2022-10-15 19:00:12 shall i change it to ahci? i don't use raid 2022-10-15 19:00:24 or ATA? 2022-10-15 19:01:00 you're currently in the rescue shell or in the Alpine run-from-ram that you booted off USB? 2022-10-15 19:01:16 rescue shell 2022-10-15 19:02:20 ah ok, if you'd been in the Alpine booted off USB I would have asked you to do a "lsmod" and look for either "vmd" of another module that might relate to the RAID controller 2022-10-15 19:02:53 if you switch it to AHCI then it should look like just another SATA controller 2022-10-15 19:03:02 already tried with no change 2022-10-15 19:03:46 no vmd in lsmod 2022-10-15 19:03:52 but a scsi_mod 2022-10-15 19:04:06 which of /boot/efi and /boot is (the root of) your ESP partition? 2022-10-15 19:04:29 -> /boot 2022-10-15 19:04:33 noyau: you won't see "vmd" is a rescue shell as it is not included in the initramfs by default... 2022-10-15 19:04:40 already rebooted 2022-10-15 19:04:41 s/is a/in a/ 2022-10-15 19:04:41 minimal meant to say: noyau: you won't see "vmd" in a rescue shell as it is not included in the initramfs by default... 2022-10-15 19:05:27 i'll probably throw in the towel and start again tomorrow with a fresh vanilla install 2022-10-15 19:05:34 just to make sure its not a hardward issue 2022-10-15 19:05:52 really wish i knew more to debug the init process 2022-10-15 19:06:20 noyau: so /boot/efi/ has no "BOOT" subdirectly with a BOOTX64.EFI file? 2022-10-15 19:06:27 subdirectory 2022-10-15 19:06:45 no it does not 2022-10-15 19:06:59 i understood that EFISTUB relied directly on the initramfs & kernel 2022-10-15 19:07:11 there is no .EFI file 2022-10-15 19:07:45 noyau: there has to be *some* EFI file for a UEFI to load, and the point of EFISTUB is that it is used to create a EFI file to directly boot 2022-10-15 19:08:53 normally UEFI relies on EFI variables being defined/registered for each bootable OS to tell them where its EFI file is 2022-10-15 19:09:30 ***##@## i might have found an issue.... 2022-10-15 19:09:35 so Debian will put its bootloader related stuff in EFI/debian/ of the ESP partition and then create a EFI boot variable to point to that 2022-10-15 19:09:36 someone said arch uses cryptdevice 2022-10-15 19:09:43 and i didn't notice that alpine uses cryptroot 2022-10-15 19:09:52 give me a min 2022-10-15 19:10:15 if no boot variables exist then the FALLBACK it for UEFI to run EFI/BOOT/BOOTX64.EFI automatically 2022-10-15 19:10:48 noyau: I only typed "cryptroot=" a few times earlier ;-) 2022-10-15 19:12:26 awwwwwww jesus 2022-10-15 19:12:30 that was it 2022-10-15 19:12:45 i only typed cryptdevice 20 times........... 60 times ..... 100 times 2022-10-15 19:12:54 what a painful lesson that one is going to be 2022-10-15 19:13:41 still fails mounting with an 'invalid argument' 2022-10-15 19:14:37 the lvmcrypt PV is created 2022-10-15 19:14:43 but not /dev/mapper/vg0-root 2022-10-15 19:15:45 noyau: what I said above, I don't understand how the EFISTUB is even booting in the 1st place if you didn't set up EFI variables or named the created stub file as EFI/BOOT/BOOTX64/.EFI 2022-10-15 19:17:02 the 'feature' order counts right? does lvm need to be before cryptsetup? 2022-10-15 19:17:19 noyau: no it does not matter 2022-10-15 19:17:43 noyau: what's your thoughts as to how the stub is even being loaded? 2022-10-15 19:18:51 so a) i made a really dumb mistake, b) I wouldn't profess to have any sort of expertise when it comes to boot chains 2022-10-15 19:19:04 but i understood the UEFI firmware can directly load the initramfs 2022-10-15 19:19:22 as long as you have the correct boot entry with efibootmgr 2022-10-15 19:19:47 noyau: the UEFI can load a EFI, which may be a bootloader or a file created by combining EFISTUB with a kernel & initramfs 2022-10-15 19:20:34 noyau: efibootmgr *READS* EFI boot variables, something has to create them. That's my point, do you have EFI boot variables as you didn't mention them so far... 2022-10-15 19:20:51 oops, efibootmgr read and can create EFI boot variables 2022-10-15 19:21:42 i understood the EFI boot variables are stored in the boards NVRAM 2022-10-15 19:21:58 noyau, fyi there's a guide i use to install alpine with encryption: https://codeberg.org/grisha/gsetup/src/branch/master/wiki/operating-systems/alpine-linux/installation-guide about a year has passed since the last installation, so something could change, but i'm sure it's not critical 2022-10-15 19:22:00 i.e. its what you are modifying when you edit the boot entries? 2022-10-15 19:22:01 noyau: yes but they don't magically appear, something or someone has to create them 2022-10-15 19:22:14 Grub creates them when grub-install is run for example 2022-10-15 19:22:25 but you're not apparently using Grub 2022-10-15 19:22:36 so I'm wondering what created them 2022-10-15 19:22:41 i'm using efibootmgr to modify them 2022-10-15 19:23:03 deleted the old entries from Arch and creating the new one passing the kernel command line params 2022-10-15 19:23:13 noyau: so that's a missing piece of information, so you created it/them and therefore that is how it boots the stub 2022-10-15 19:23:33 and therefore the fallback file does not need to be used 2022-10-15 19:24:53 noyau: re the 'invalid argument', what's the full cmdline options list? 2022-10-15 19:27:38 cryptroot=UUID=XXXXX cryptdm=lvmcrypt root=/dev/mapper/lvmcrypt rootfstype=ext4 initrd=\intel-ucode.img initrd=\initramfs-lts 2022-10-15 19:28:35 @grisha ty, but its grub based, looking through none the less 2022-10-15 19:32:03 noyau: no other cmdline options at all? also aren't those slashs the wrong way around for the initrd= options? 2022-10-15 19:32:46 that's it and apparently \ is part of the UEFI standards 2022-10-15 19:32:52 for the initrd calls 2022-10-15 19:36:03 I would have expected more options like "ro console=tty0 quiet" 2022-10-15 19:39:03 if i change root back to point to /dev/mapper/vg0-root i get a very helpful error asking me if the dm_mod kernel is loaded 2022-10-15 19:39:16 is that a straightforward feature including of 'dm_mod'? 2022-10-15 19:42:03 hmmm its supposed to be loaded as part of lvm 2022-10-15 19:42:13 so seems to imply that it isn't being loaded or not in the right order 2022-10-15 19:50:15 noyau: dm_mod would be included by the lvm feature 2022-10-15 19:52:01 i had to fix this to get FDE to work on a Pinephone Pro, could it be related? 2022-10-15 19:52:08 https://wiki.postmarketos.org/wiki/Full_disk_encryption 2022-10-15 19:52:36 Install and enable the following: # apk add cryptsetup cryptsetup-openrc # rc-update add dmcrypt boot Add the following lines to /etc/conf.d/dmcrypt: target= source='/dev/' 2022-10-15 19:53:26 is it possible that you need dmcrypt vs archlinu using crypttab to mount the VG0 devices once opened? 2022-10-15 19:56:10 noyau: ? you've got the initramfs to decrypt the LUKS partition so that (LUKS-related) bit is done 2022-10-15 19:56:32 VGs are nothing to do with LUKS 2022-10-15 19:56:47 so if root=/dev/mapper/lvmcrypt -> it decryptes the container but never creates the vg0 entries 2022-10-15 19:56:58 so it stops before the equivalent of vgchange -ay 2022-10-15 19:57:20 if it's changed to root=/dev/mapper/vg0-root it gives the device mapper failure and points to dm_mod not being loaded 2022-10-15 19:58:23 jumping back a bit, why are you providing the "initrd=" options on the cmdline? normally those would be used to tell the bootloader to load those files but you aren't using a bootloader... 2022-10-15 19:58:55 https://wiki.alpinelinux.org/wiki/Bootloaders#efibootmgr 2022-10-15 19:59:13 To boot directly from your motherboard's UEFI boot menu, a boot entry needs to be created with either a UEFI shell or efibootmgr. efibootmgr Install efibootmgr: apk add efibootmgr Create a boot entry. It's recommended to do this in a script, as efibootmgr does not allow editing entries. #!/bin/sh params="root=/dev/sdXZ rw \ initrd=\intel-ucode.img \ initrd=\initramfs-lts" efibootmgr --create --label "Alpine Linux" \ --disk /dev/sdX --par 2022-10-15 20:01:32 noyau: actually for Grub those files are defined in the grub.cfg, not specified on the commandline 2022-10-15 20:02:29 ? i'm not using grub? 2022-10-15 20:02:55 I know, but I'm trying to work out what exactly is intended to make use of the initrd= params 2022-10-15 20:07:45 my point is that generally speaking it is the bootloader's job to load the initramfs, but when you don't have a bootloader as you're using EFISTUB I thought the initramfs was bundled together with the kernel and the stub 2022-10-15 20:10:03 i got it 2022-10-15 20:10:29 i made two changes at once which threw me off, root has to be /dev/mapper/vg0-root not lvmcrypt 2022-10-15 20:10:50 (and don't change the features at the same time ..... particularly if you accidently delete keymap) 2022-10-15 20:11:00 booted to a root shell 2022-10-15 20:12:38 minimal much obliged for your help, i'll try to draft some changes to the wiki in the coming days, just to add the efibootmgr section and maybe explicitly call out the delta between the arch and alpine crypt params 2022-10-15 20:12:51 i'm surprised the root= isn't explicitly described in the wiki as well 2022-10-15 20:13:32 to honest I rarely look at the Wiki 2022-10-15 20:13:55 also as I said earlier I've not yet tried stub booting 2022-10-15 20:14:02 i hate asking questions, and every time i do it never fails its something like this 2022-10-15 20:14:20 well at least now you know it works :D 2022-10-15 20:14:32 I just get it working and script it to use from then onwards 2022-10-15 20:15:05 thanks again and have a good night 2022-10-16 07:40:01 good morning. 2022-10-16 07:40:57 i'm trying to setup alpine with a ppp dialup connection, but despite following the wiki, /etc/init.d/networking insists on "device ppp0 not found" and fails to start the pppd to actually dial up. how do i fix thisรŸ 2022-10-16 07:40:59 ? 2022-10-16 10:28:32 do you use yt-dlp or youutube-dl? 2022-10-16 10:29:36 yt-dlp 2022-10-16 10:31:03 seems to be the new prefered one .. recently learned about it.. 2022-10-16 10:31:37 it downloads really fast compared to youtube-dl 2022-10-16 10:32:58 when i use mpv it starts with lowest quality tho (not sure if my settings or default) 2022-10-16 10:33:38 uhM 2022-10-16 10:33:50 could you share some URL? 2022-10-16 10:34:13 I just have this http://ix.io/4djq 2022-10-16 10:35:15 donoban: happens with all urls, eg https://www.youtube.com/watch?v=UrGw_cOgwa8 2022-10-16 10:35:35 VO: [gpu] 640x480 yuv420p 2022-10-16 10:35:36 ah nvm, its my settings 2022-10-16 10:35:54 uhM, looks low qualitty 2022-10-16 10:36:14 tbh i dont mind low q toooo much, but sometimes nice with better q 2022-10-16 10:36:41 maybe would be nice with something between worst and best... oh well 2022-10-16 10:37:40 hmm i guess that meanst yt-dlp can give worse q than youutube-dl :D 2022-10-16 10:38:36 well, youtube-dl was slow for me with any qualitty 2022-10-16 10:39:07 I think that you can switch the qualitty with some option, I didn't care since I rarely see videos on fullscreen 2022-10-16 10:40:46 Hello everyone, do you have any idea why podman breaks my host IPv6 (my route) on Alpine please ? I have been searching since yesterday without success. If I recreate the route manually it is functional but it is not great. 2022-10-16 10:40:52 Paste here: https://privatebin.tetsumaki.net/?ed28eb74b6f646f2#9HSD8QLZndJzzApJzzpLa4AXHmr5SWjfU1RLDyjGwKvN 2022-10-16 10:44:59 frag: I opened that video in youtube 2022-10-16 10:45:11 max quality avaialbe is 480p 2022-10-16 10:45:33 additionally to getting iface ppp0 started during system boot - it seems net-snmp's snmpd segfaults more or less reliably in libnetsnmpmibs.so.40 - any hints on how to debug this? 2022-10-16 10:45:50 VO: [gpu] 3840x2160 yuv420p10 2022-10-16 10:45:59 try this https://www.youtube.com/channel/UCYq-iAOSZBvoUxvfzwKIZWA 2022-10-16 10:46:46 Marquel: do you know how to get a backtrace with gdb 2022-10-16 10:47:44 but all videos by default opens in a tiny 4cm wide box on my 40" tv, so i think there is some difference :] 2022-10-16 10:48:14 anyway, its not really a problem :] 2022-10-16 10:48:28 what quallity do you get with the last one? 2022-10-16 10:49:12 it bottlenecks my cpu 2022-10-16 10:50:05 sorry the URL was https://www.youtube.com/watch?v=LXb3EKWsInQ 2022-10-16 10:50:13 oh thought you meant youtube-dl .. which is abou 20cm wide 2022-10-16 10:50:42 thats a channel, but ill try 2022-10-16 10:50:52 oh new url.. 2022-10-16 10:51:11 yeah I don't how why I copied it wrong 2022-10-16 10:51:31 basically the same with this last url 4-6 cm 2022-10-16 10:51:48 psykose: not precisely but with a bit of assistance, i can try. 2022-10-16 10:52:03 apk add gdb net-snmp-dbg openssl-dbg musl-dbg 2022-10-16 10:52:08 i guess yt-dlp can find lower quality when using the "worst" option 2022-10-16 10:52:10 gdb --args 2022-10-16 10:52:10 r 2022-10-16 10:52:12 bt full 2022-10-16 10:52:27 uhm, I'm not sure what is the problem 2022-10-16 10:52:31 do you want that it oppens it fullscreen? 2022-10-16 10:52:41 psykose: it's a daemon running and it crashes at some point after it's been queried. 2022-10-16 10:53:05 you'd have to run it by hand in the same fashion and make it stay in foreground 2022-10-16 10:53:09 and hope it still reproduces 2022-10-16 10:53:16 or you can raise ulimit and have it dump a core somewhere 2022-10-16 10:53:24 that one is a bit more work so you can try the other first 2022-10-16 10:53:37 donoban: nah, still testing but maybe yt-dlp finds too low quality (or too high), maybe can be fixed in setting i duno.. 2022-10-16 10:54:33 you can config ytl-format=xxxx 2022-10-16 10:54:53 check https://github.com/ytdl-org/youtube-dl#format-selection 2022-10-16 10:57:47 donoban: yeah basically worst or best, or manually set every time 2022-10-16 10:57:56 psykose: i gather i need to tell the snmpd to not background, correct? 2022-10-16 10:58:19 you can select best and limit resolution and others things 2022-10-16 10:58:57 `snmpd -Le -f` should suffice 2022-10-16 10:59:08 e.g. ylt-format="bestvideo[height<=480]+bestaudio/best[height<=480]" 2022-10-16 10:59:14 psykose: got to that point, yeah ;) 2022-10-16 10:59:54 psykose: just as a side-node while waiting for it to crash - you don't happen to know how to get init.d/networking to auto-start a PPP iface? 2022-10-16 10:59:57 sorry it's ytdl-format 2022-10-16 10:59:59 donoban: oh thanks! 2022-10-16 11:00:05 sadly i don't 2022-10-16 11:00:15 ah 2022-10-16 11:00:28 perhaps try adding ifupdown-ng ifupdown-ng-ppp 2022-10-16 11:00:32 then it should be recognised 2022-10-16 11:00:37 it's not. 2022-10-16 11:00:40 https://github.com/ifupdown-ng/ifupdown-ng/blob/34753136b86d5b1d5df2b60b22f1bfac16f28416/doc/interfaces-ppp.scd 2022-10-16 11:00:48 you have to add those 2022-10-16 11:00:54 networking just says "ppp0 does not exist" 2022-10-16 11:01:06 and yes, i do have both ;) 2022-10-16 11:01:37 can you post the /etc/network/interfaces 2022-10-16 11:07:53 psykose: iface ppp0\nppp-physdev eth0.7\nppp-provider ISP 2022-10-16 11:08:56 do you also have pppd started 2022-10-16 11:10:44 no. 2022-10-16 11:11:18 there's no ppp* service in /etc/init.d, i also only need to dial out, not dial in. 2022-10-16 11:11:53 well it's what makes the ppp0 interface at all, no idea how it's meant to work 2022-10-16 11:12:05 *sigh* 2022-10-16 11:12:28 the wiki is not exactly helpful. "just add that stanza, done". yeah... no. 2022-10-16 11:13:04 so, the crash happens in var_extensible_disk() of libnetsnmpmibs.so. 2022-10-16 11:13:22 pastebin the backtrace 2022-10-16 11:14:07 can i make gdb put it into a file? 2022-10-16 11:15:17 set logging on 2022-10-16 11:15:25 set logging file path/to/file 2022-10-16 11:15:30 https://stackoverflow.com/questions/5941158/gdb-print-to-file-instead-of-stdout 2022-10-16 11:20:23 ikke: did not exactly get the bt full into the logfile. 2022-10-16 11:31:42 psykose: http://dpaste.com/7B9NQCQ74 - here you go 2022-10-16 11:32:06 did you install the relevant -dbg packages/ 2022-10-16 11:32:07 did you apk add the -dbg packages first 2022-10-16 11:32:10 :D 2022-10-16 11:32:37 obviously not ;รผ 2022-10-16 11:34:00 musl-dbg as well 2022-10-16 11:34:42 i hope gdb will find them w/o being restarted. 2022-10-16 11:35:37 that's the cool part 2022-10-16 11:35:39 it won't 2022-10-16 11:35:41 :D 2022-10-16 11:35:56 but there's probably a command for it 2022-10-16 11:36:59 if you use `file` to load the binary again and pass the args again after (r list of args) it would work i g uess 2022-10-16 11:37:10 https://unix.stackexchange.com/questions/572639/reload-symbol-table-after-recompilation-in-gdb has now answer 2022-10-16 11:37:35 i don't care. 2022-10-16 11:37:41 restarted the service anyway. 2022-10-16 11:37:55 and _maybe_ already fixed it with a different system configuration. 2022-10-16 11:49:55 I have found solution for me: https://gitlab.alpinelinux.org/alpine/aports/-/issues/12553 2022-10-16 11:52:50 psykose: it's just 6MB of backtrace now. how do you like it? 2022-10-16 11:53:01 spicy 2022-10-16 11:54:37 with a flush of death? 2022-10-16 11:54:55 however, wgetpaste fails to paste it, so any other option is appreciated. 2022-10-16 11:55:11 note sure what paste service accepts 6M of input 2022-10-16 11:55:22 s/note/not/ 2022-10-16 11:55:22 ikke meant to say: not sure what paste service accepts 6M of input 2022-10-16 11:55:45 deer meatbot, i can read regex. 2022-10-16 11:56:15 some people like / expect that it does that 2022-10-16 11:56:27 i guess. 2022-10-16 11:56:49 0x0.st should be fine 2022-10-16 11:58:41 psykose: http://0x0.st/owiS.log enjoy. 2022-10-16 11:59:03 that's 14kb 2022-10-16 11:59:21 thankies 2022-10-16 12:01:19 tetsumaki: the combination of accept_ra + forwarding is always tricky 2022-10-16 12:01:37 (in the sense that it surprises people) 2022-10-16 12:01:49 strange segfault 2022-10-16 12:02:35 seems to be in the entry->flags deref but it's not null and initialised right there 2022-10-16 12:03:09 it works on gentoo *hide* 2022-10-16 12:04:56 funnily, this seems related https://github.com/net-snmp/net-snmp/issues/453 2022-10-16 12:05:08 even if it crashes somewhere else 2022-10-16 12:05:57 ah 2022-10-16 12:05:59 known for years 2022-10-16 12:06:00 https://github.com/net-snmp/net-snmp/issues/293 2022-10-16 12:06:41 and already in https://gitlab.alpinelinux.org/alpine/aports/-/issues/12309 2022-10-16 12:07:07 i guess if you comment out the includeAllDisks 90% it will.. work 2022-10-16 12:07:08 but 2022-10-16 12:10:44 psykose: i don't mind. 2022-10-16 12:11:24 but why is includeAllDisks causing the segfault 2022-10-16 12:11:28 psykose: the system in question has one disk, which is, with about 2GB of size _waaaaayyyyyy_ too big for what it's holding and as such i couldn't care less about "allDisks". 2022-10-16 12:15:14 rerunning w/o the incriminating configuration option and we'll see what happens next. 2022-10-16 12:15:23 ikke: no idea, i don't code good 2022-10-16 12:15:38 the entry-> in question is a top level static in another TU 2022-10-16 12:15:42 after that it's down to "how do i get ppp0 to auto-dial-in when networking[.ppp0] gets started? 2022-10-16 12:15:48 so i guess either it's a threading issue or an initialisation issue 2022-10-16 12:20:49 i'm like 10 files deep just to find where this is initialised 2022-10-16 12:21:00 i think i've passed like 3 kinds of object systems on the way 2022-10-16 12:21:09 my conclusion is to uninstall this from your system 2022-10-16 12:21:17 anyway 2022-10-16 12:21:18 more to go.. 2022-10-16 12:23:19 psykose: actually, no. that's not an option. 2022-10-16 12:23:56 i am querying the system via snmp and a not-so-unimportant part is netfilter statistics. 2022-10-16 12:24:14 i rather reconfigure to "working" instead of uninstalling ;) 2022-10-16 12:24:56 ok i give up tbh 2022-10-16 12:25:03 psykose: no worries. 2022-10-16 12:25:33 as i said, if it is caused by a configuration option - and it very much seems that way - i can configure it to stable operation and be done with it. 2022-10-16 12:25:39 thanks for your tries. 2022-10-16 12:25:56 and, tbh, _I_ can try to fix it myself anyway. 2022-10-16 12:26:01 it's not, the actual issue is what code gets ran with the configuration option (well, in the sense of that's not a real fix) 2022-10-16 12:26:14 it's not like that'd be the first C program I've took a too-deep look into. 2022-10-16 12:26:26 i'd guess it's either the things above or one of these weirdass containers point to a stack local 2022-10-16 12:26:35 probably the stack local given the things i saw along the way 2022-10-16 12:26:51 which is invalid and hence gets optimised out 2022-10-16 12:27:01 if you follow the upstream issue you see the issue gets removed with -O0 2022-10-16 12:27:06 or clang, different opts, maybe not true anymore 2022-10-16 12:27:31 sounds like recompiling with a different APKBUILD would help. 2022-10-16 12:27:56 i would sooner remove the software from the repositories than add -O0 to work around weird bugs 2022-10-16 12:28:01 but maybe i'm a bit of a zealot 2022-10-16 12:28:05 i shall simply not look at it again 2022-10-16 12:28:18 as i said, maybe i can get it fixed. 2022-10-16 12:28:39 apparently people want me to add my APKBUILDs to the official tree, might as well get that done too. 2022-10-16 12:28:40 asan probably makes this more easier 2022-10-16 12:28:55 but i would also bet it crashes on 25 other issues 2022-10-16 16:54:16 so, after looking into it, it seems ifupdown-ng tries the link module for my ppp-connection first, which - obviously - fails. 2022-10-16 16:56:17 "pon isp" works, "ifup -f -i /etc/network/interfaces ppp0" kind-of works, "/etc/init.d/networking[.ppp0] start" does not, nor does it accurately report it does not work, instead, it reports coming up w/o the pppd actually being invoked. 2022-10-16 17:12:08 okay, seems i got it to work as far as starting the interface, still it says "interface not found", apparently trying the link executor despite that executor being unable to perform any task here. 2022-10-16 17:39:18 I'm trying to access localhost but am having trouble on firefox and a fresh install of chromium. I was able to access it before and am wondering if my rc-services are to blame. If I'm using IWD via NetworkManager, is it right to have dbus enabled as boot and default, networking as default, and networkmanager as default? 2022-10-16 17:41:45 These may actually be separate issues; I noticed that on both firefox and chromium there's a small "i" that says "you are not securely connected to this site," warns about people maybe trying to steal my data, yadda yadda. Not sure how to force it to just connect anyway 2022-10-16 18:10:09 Is there any script or process that edits /etc/network/interfaces? I have now had eth0 entry deleted twice. 2022-10-16 18:13:42 nick_: depending on what you are trying to access, it may have been disabled by recent changes to the browser. 2022-10-16 18:14:17 e.g. file:///tmp/test.html doesn't work anymore in some builds 2022-10-16 18:19:21 what builds would those be 2022-10-16 18:20:04 cache_return: only setup-interfaces writes to it 2022-10-16 18:21:22 psykose: that file url doesnt work in chrome at all without a flag, and has been removed from firefox focus and firefox android. 2022-10-16 18:21:42 ACTION checks 2022-10-16 18:22:00 still works for me in ff package for arch on x86_64 2022-10-16 18:22:21 works in chromium 106 and firefox and firefox-esr and firefox-nightly 2022-10-16 18:22:23 it would be pretty weird if file urls didn't work in chrome, since chrome is the primary pdf viewer for a lot of users 2022-10-16 18:22:34 the flag in chrome is something like --enable-file-access-from-files 2022-10-16 18:22:40 not needed 2022-10-16 18:22:44 it depends on what yoy are doing 2022-10-16 18:23:18 if it's just standalone html, it eorks, but for something more complicated it needs the flag 2022-10-16 18:23:27 Running the application opened the localhost for me. I don't know how it was able to when I can't, but hey, it works. 2022-10-16 18:23:54 Still curious about the rc-services, I don't want redundant networking services 2022-10-16 18:24:12 something more complicated like what 2022-10-16 18:25:48 anything that loads from the local filesystem 2022-10-16 18:25:53 https://stackoverflow.com/questions/39007243/cannot-open-local-file-chrome-not-allowed-to-load-local-resource#46480984 2022-10-16 18:26:07 the solutions are to run a local http server 2022-10-16 18:27:54 okay, so I am deleting my own network interfaces file by running setup-interfaces to try to configure wifi 2022-10-16 18:28:02 annoying 2022-10-16 18:28:29 well, I have loaded the firmware from rpi os and still no wifi 2022-10-16 18:30:44 ah that thing 2022-10-16 18:30:46 js api is separate 2022-10-16 18:30:57 that flag is also not in chrome://flags and doesn't do anything and you get the same permission denied 2022-10-16 18:33:47 I probably just remembered the flag wrong 2022-10-16 18:38:43 --allow-file-access-from-files 2022-10-16 18:40:01 nope 2022-10-16 18:42:33 who knows, maybe I imagined it all. I'll get back to deleting my networking configs over and over again. 2022-10-16 18:43:19 i don't think you have to run setup-interfaces more than once 2022-10-16 18:43:34 all it does is write that file pretty sure, so if it's already written it's not really doing anything except deleting changes 2022-10-16 18:48:11 cache_return: when are these changes to interfaces file occurring? during boot? or after the RPI finishes booting? 2022-10-16 18:48:21 I feel as though I've learned that the hard way. 2022-10-16 18:49:11 minimal: psykose got it, I think I was causing it by running setup-interfaces to see if it would detect wlan0 2022-10-16 18:49:33 nah, that's based on the module/firmware loading correctly then it would just be there 2022-10-16 18:52:50 dmesg reports that firmware is loaded 2022-10-16 18:53:57 other thing to try is linking the .txt the same way as the .bin 2022-10-16 19:03:16 I can look into that. 2022-10-16 19:03:49 also though, what do you mean by "it would just be there"? 2022-10-16 19:04:22 be where? Network devices don't show up in /dev, right? 2022-10-16 19:05:03 ip a 2022-10-16 19:05:22 do we know what ip is detecting? 2022-10-16 19:05:32 sys/class/net/ would have it 2022-10-16 19:05:40 ah, ok 2022-10-16 19:05:48 it wouldn't have an ip until you added wpa_supplicant/iwd to connect to something 2022-10-16 19:05:55 and then added a dhcp client, etc 2022-10-16 19:18:45 yes, /sys/class/net is what I was looking for 2022-10-16 19:19:27 just didnt want to be relying on something at a higher level of abstraction 2022-10-16 19:26:07 literally just reads that folder 2022-10-17 08:42:58 exit 2022-10-17 08:59:32 finally got tmux started on boot 2022-10-17 09:05:46 the key arg is 'new -d -s 1 weechat' 2022-10-17 09:06:14 without -d, openrc will fail to start tmux 2022-10-17 09:18:28 Because tmux probably expects a tty otherwise which is not available 2022-10-17 09:20:05 yeah, i was trying to find out a background args for tmux, but no, -d is kind of that one 2022-10-17 09:21:35 but one issue, when i attach to this session with tmux attach -d -s 1, rc-status shows tmux crashed 2022-10-17 09:22:55 but anyways, it does not matter, i only needed tmux to start on boot and start weechat backgroud, i don't care it crashed 2022-10-17 09:26:30 here is the init.d i created https://0x0.st/owqI.txt 2022-10-17 09:38:34 here is another example i haven't tried, it looks very complicated : https://gist.github.com/kirelagin/d4e13b027a384cb6ab65 2022-10-17 09:38:42 and scary 2022-10-17 09:55:23 0 2022-10-17 09:55:29 1 2022-10-17 10:08:52 10 2022-10-17 10:10:21 11 2022-10-17 10:26:45 uuidNuniq: cannot view https://0x0.st/owqI.txt ^ 2022-10-17 10:27:52 you wrote script on gist in 2014 and released in 2022 :-) 2022-10-17 10:36:19 nm, got it via free webproxy 2022-10-17 10:42:14 vkrishn: the gist was not written by me 2022-10-17 10:42:43 i am considering launch tmux via /etc/init.d/local 2022-10-17 10:49:31 pls share via tpaste.us 2022-10-17 10:50:41 astlep4: please fix your connection 2022-10-17 10:50:49 currently I find this useful, tmux new -d -s $(uname -n); tmux new-window weechat 2022-10-17 10:51:52 tmux new -d -s $(uname -n) \; new-window weechat 2022-10-17 10:52:42 if you start interactively , just tmux -s $(uname -n) weechat 2022-10-17 10:52:58 tmux new -s $(uname -n) weechat 2022-10-17 10:56:03 starting together means quiting weechat also quits tmux, guess thats intended 2022-10-17 11:05:11 vkrishn: nope, after attach you can create new windows. 2022-10-17 11:05:22 ok 2022-10-17 11:12:38 gonna try this : https://0x0.st/owb0.sh 2022-10-17 11:12:41 brb 2022-10-17 11:14:15 aha, it works, i am back 2022-10-17 11:33:08 another method from #gentoo suggestion : https://0x0.st/owbd.txt 2022-10-17 11:34:50 so we have 3 method to start a program on boot : via standalone openrc daemon, via local daemon, via crontab @reboot. the last one is easist to start a one-shot program 2022-10-17 11:58:54 Hi! I understand that this is a very odd question (especially since both code-oss and electron are only present in edge/testing). Does anybody on here use code-oss on a daily basis? 2022-10-17 11:59:05 Trying to figure out if it only broke for me after updating to latest electron from `testing` (from v19 to v21, which was published five days ago), or if it's a global problem (for me, even when launching code-oss with a fresh profile, its "extensionHost" process consumes an entire CPU core; and some extensions are unresponsive). 2022-10-17 12:09:44 Never mind, it was my fault. I have looked through the entire code-oss bug tracker, but didn't think to look at alpine code-oss package history. Turns out there was a fix for electron 21 compatibility in 1.72.1-r1, I just didn't have it updated because I had pinned code-oss version. (I'm using it with SSH remote plugin, which installs VSCodium remote execution host, so I have to patch code-oss product.json every time to mak 2022-10-17 12:17:53 Inga: you can pin a package if you don't want 'every time do something' any more 2022-10-17 12:29:02 uuidNuniq: that was the issue 2022-10-17 12:29:24 The package was pinned, so it was not updated to a version with a fix 2022-10-17 12:32:19 i mean, pin the related packages too 2022-10-17 12:34:30 it would be nice, if apk upgrade output messages like this : package xyz not upgrading due to world settings 2022-10-17 12:35:00 currently, it outputs nothing? 2022-10-17 16:22:26 ikke: one would need to know what packages do they need to pin. "All dependencies and subdependencies" is too broad; and as I understand, after the electron package was updated, there was no way of going back besides building an earlier version from sources. That's OK, it was my mistake, and if I understand alpine philosophy correctly, nothing would have prevented me from making it, unfortunately. 2022-10-17 16:23:16 curious what adjustments to the product.json there are 2022-10-17 16:23:24 uuidNuniq: sorry, should have mentioned you instead 2022-10-17 16:25:57 @psykose: just changing a bunch of fields in product.json and package.json to correspond to the appropriate release of VSCodium. changing nameShort, nameLong, etc to vscodium, setting release field in package.json to vscodium build number, commit field in product.json to vscodium commit hash, etc 2022-10-17 16:26:18 that sounds useless 2022-10-17 16:28:10 the thing is, code-oss does not have official builds for remote execution host (which is when you run the IDE on one host but the actual code is on another host). vscodium does have these, but vscodium is not packaged for alpine (and thank you so much for packaging code-oss and electron!) 2022-10-17 16:28:47 yeah, i'm just confused why the extension would care about some version field to be installable 2022-10-17 16:28:49 however, remote execution host requires that the IDE is the same product and the same version 2022-10-17 16:29:08 could you link the extension 2022-10-17 16:29:24 and it checks this by all these fields (I guess the main reason was to prevent people from mixing code-oss/vscodium with VS Code) 2022-10-17 16:30:02 @psykose: the extension is https://open-vsx.org/extension/jeanp413/open-remote-ssh 2022-10-17 16:31:54 there is also a non-free official version for VS Code: https://marketplace.visualstudio.com/items?itemName=ms-vscode-remote.remote-ssh (but as I understand, trying to use it from code-oss would violate its license) 2022-10-17 16:35:31 it appears to work without changing anything in the product.json 2022-10-17 16:35:56 unless the ssh authentication failure is what happens when you don't do that 2022-10-17 16:38:46 @psykose: IIRC ssh authentication failure happened for me on the first step, long before the extension started to use any of the fields from product.json 2022-10-17 16:40:25 @psykose: here the extension relies on some of these fields to compute the URL of the REH package: https://github.com/jeanp413/open-remote-ssh/blob/v0.0.16/src/serverSetup.ts#L218 2022-10-17 16:40:52 yeah noticed that 2022-10-17 16:40:52 dunno 2022-10-17 16:40:55 they should just not do that 2022-10-17 16:41:12 not that i would ever use something where the first action on connection is to run a bash script to download and start something lol 2022-10-17 16:41:37 and then from what I understood, REH itselfs checks whether versions installed on the client and on the server match each other. so even after I fixed enough fields for the extension to work, I just got REH complaining about version mismatch, and had to fix some more fields 2022-10-17 16:42:39 yep 2022-10-17 16:43:01 @psykose: from what I understand, the original MS idea was something like headless VS Code, connecting to the non-headless VS Code on your desktop 2022-10-17 16:43:45 it is, yeah 2022-10-17 16:46:07 and since I'm doing webdev, I'm just running a linux container for development (to isolate node_modules etc from my actual workstation), and use code-oss with this extension on my desktop. so that I can do webdev yet none of the node_modules (or extensions working with node_modules) can escape the container and affect anything on my desktop 2022-10-17 16:54:28 perhaps the code-oss package should have a REH subpackage created? 2022-10-17 16:56:28 minimal: I think it's the fault of code-oss maintainers, not of alpine, that they don't provide REH binaries 2022-10-17 16:56:57 regardless of distro, nobody can use remote ssh from code-oss without tweaking these jsons 2022-10-17 16:57:29 it's just that most other distros package vscodium, but there is only code-oss for alpine. 2022-10-17 16:57:35 Inga: well I'm thinking of where both "ends" are running Alpine 2022-10-17 16:57:51 (which is a huge feat, and thank you all for packaging code-oss!) 2022-10-17 16:58:55 if the remote host was as simple as building it and then running `code-oss-remote-host --listen` on the remote or something, and the extension can connect to that, i could fix this in 5 minutes 2022-10-17 16:59:02 the issue is for some reason they don't work that way 2022-10-17 16:59:07 minimal: even then you'd also have to tweak that extension, so that it would install REH subpackage to the remote, instead of downloading that binary. and that's assuming it's even possible to create REH subpackage from code-oss 2022-10-17 17:00:15 minimal: from my unexperienced point of view, it seems that it would be easier to package vscodium for alpine (by following the established path with code-oss), than to make code-oss to work with REH properly. but of course I don't know anything 2022-10-17 17:04:13 but yeah i guess it would make more sense to replace it with vscodium as a whole 2022-10-17 18:12:56 ๐Ÿ’จDon't miss your chance to be rich... (full message at ) 2022-10-17 18:19:37 hello I am wondering why the alpine virt kernel does not offer THP (transparent huge pages) support in x86_64, but does for aarch64 (https://git.alpinelinux.org/aports/tree/main/linux-lts/virt.aarch64.config#n116) 2022-10-17 18:20:00 should I ask this in #alpine-devel ? 2022-10-17 18:21:01 schwicht: probably best to open an issue on our gitlab instance 2022-10-17 18:25:47 @ikke any suggestion about the project I should file this under? I looked for kernel, but did not locate that 2022-10-17 18:25:56 https://gitlab.alpinelinux.org/alpine/aports 2022-10-17 18:31:53 done https://gitlab.alpinelinux.org/alpine/aports/-/issues/14272 2022-10-17 18:33:25 Inga, psykose: there's also https://github.com/xaberus/vscode-remote-oss which seems to be a open-remote-ssh alternative (though a REH server is still required) 2022-10-17 18:35:28 schwicht: I've added the category:kernel tag. ncopa will look at those when upgrading kernels 2022-10-17 20:05:09 minimal: as I understand, vscode-remote-oss is a really minimal alternative to open-remote-ssh where one has to manually do everything that open-remote-ssh would do for them (to install REH manually, to start it up, to set up all tunnels etc) 2022-10-17 20:09:57 vscode-remote-oss is also much older (with its latest release just 6 days after the first prototype of open-remote-ssh). so I'd say it's more like the other way around, open-remote-ssh is a much more user-friendly alternative to vscode-remote-oss. and the problem that the product+version on the client and on the server should match each other applies to both, as I understand 2022-10-17 20:18:16 Hey, is there a minimum architecture for the x86 version of alpine? I think I saw somewhere it's going to be pentium4 in the future, but i don't remember where I read that 2022-10-17 20:18:37 i586 now, pentium4 when it's changed (not yet) 2022-10-17 20:18:46 whether packages respect that or not depends 2022-10-17 20:19:48 https://gitlab.alpinelinux.org/alpine/tsc/-/issues/35 2022-10-17 20:20:32 Which package in alpine carries the finger command please? 2022-10-17 20:21:02 Guest1537: i don't think there is one 2022-10-17 20:21:04 ucspi-tcp (or tcp6) 2022-10-17 20:21:16 if you want finger@ anyway 2022-10-17 20:21:20 maybe it's a very spicy one 2022-10-17 20:21:45 What's with the @ suffix? 2022-10-17 20:21:51 some fancy way of namespacing? 2022-10-17 20:22:06 oh, i pinged the wrong guest lol 2022-10-17 20:22:17 they're gone now 2022-10-17 20:24:51 some djb convention probably 2022-10-17 20:24:56 aha 2022-10-17 20:26:21 Nyet on ucspi-tcp6 carrying finger. 2022-10-17 20:27:06 Shufei_: https://pkgs.alpinelinux.org/contents?file=finger%40&path=&name=ucspi-tcp&branch=edge&arch=x86_64 2022-10-17 20:27:23 it's 'finger@' 2022-10-17 20:27:59 ah 2022-10-17 20:28:58 i don't think that finger exists 2022-10-17 20:29:43 Yea, just plain oldschool finger 2022-10-17 20:30:14 We don't ship it 2022-10-17 20:30:55 I reckon not. Looking but not finding. 2022-10-17 20:31:11 coreutils pinky should do the same 2022-10-17 20:31:29 lol 2022-10-17 20:31:43 whether it does what you want to at all without utmps.. no idea 2022-10-17 20:36:44 psykose, ikke thanks for the x86 infos... i'm probably going to follow up on the mailing list pointing out it's probably not worth packaging that ancient intel firmware since most machines that use it are nearing alpine EoL 2022-10-17 20:37:06 what machine do you have 2022-10-17 20:37:47 Thanks, kids, looks like no finger to be had and pinky wonโ€™t do. 2022-10-17 22:14:55 psykose: i think my machine will remain supported thankfully. (Thinkpad x31) just barely lol 2022-10-17 22:16:08 that is technically after the pentium4 x) 2022-10-17 22:16:26 or is it 2022-10-17 22:16:26 hm 2022-10-17 22:16:44 yea 2022-10-17 23:11:06 I understand sodo has EOL-ed in Alpine 3.16, I spotted today that https://wiki.alpinelinux.org/wiki/Creating_an_Alpine_package -> "abuild-keygen -a -i" relies on sudo to be functional. also doas-sudo-shim did not work for me 2022-10-17 23:11:25 I meant sudo of course 2022-10-17 23:14:43 SUDO=doas abuild-keygen also works 2022-10-17 23:14:45 i think 2022-10-17 23:16:35 I helped myself in building it quickly ... 2022-10-17 23:17:51 oh wow 2022-10-17 23:17:55 why was sudo EOL'd? 2022-10-17 23:19:24 Probably because doas has a better security track record so far... 2022-10-17 23:20:27 it's merely in community instead and you can install it all the same 2022-10-17 23:20:35 ah cool 2022-10-17 23:20:42 i thought it was removed or something 2022-10-17 23:20:53 it helps to read the actual changelog and not an irc channel 2022-10-17 23:21:15 yea 2022-10-17 23:21:22 overheat :-) 2022-10-18 00:50:01 swinnenb: echo '/usr/bin/doas' | doas tee /usr/bin/sudo ; doas chmod +x /usr/bin/sudo 2022-10-18 00:52:15 note that the doas on alpine is opendoas, not the openbsd doas 2022-10-18 00:54:35 Wouldn't it be best to just apk add doas-sudo-shim instead of messing with the system? 2022-10-18 00:57:18 Saijin_Naib: no, that is a shell script. 2022-10-18 00:58:48 most of the time, we just run sudo/doas without any options 2022-10-18 01:00:09 or simply alias sudo='doas' for interactive use 2022-10-18 01:44:21 i've been just installing doas and symlinking it to sudo 2022-10-18 01:44:27 which works till a package installs sudo over top 2022-10-18 01:44:48 i haven't converted all my stuff to alpine...plus years of habit is hard to break 2022-10-18 02:21:50 dewdude: I think you can blacklist sudo from being installed, i can't remember, maybe you simply put a ! before the package in /etc/apk/world 2022-10-18 02:22:32 maybe. not a big deal 2022-10-18 02:22:42 it errors, i flip to doas 2022-10-18 02:22:52 I'm down to two ubuntu VMs anyway 2022-10-18 02:24:26 https://img.pickmy.org/tbw9nuzk.png 2022-10-18 02:27:30 I found this distro sometime last week when I was looking for a stupid small system do literally just bridge two interfaces together. Holy cow. I had a few hours of yelling at my monitor but I feel like I'm really getting the hang of it 2022-10-18 02:40:09 dewdude: what distro 2022-10-18 02:40:17 alpine 2022-10-18 02:40:25 it's great distro 2022-10-18 02:40:29 the one for the channel we're in 2022-10-18 02:41:20 i f love it. 2022-10-18 02:43:10 my old xenorchestra vm wanted tons of ram. now i give it 512MB and it's only using half of that. I'm happy. 2022-10-18 02:57:10 Alpine is pretty freaking incredible, for sure. Stopped a nearly two decade distro-hop habit for me. I daily it on my desktop and laptops 2022-10-18 02:59:24 i found great success building appliances around alpine, due to its easy of modification 2022-10-18 02:59:25 Saijin_Naib: i have alpine on my server 2022-10-18 03:00:11 the setup-* scripts are amazing 2022-10-18 03:01:05 Ahh... I do not have a server. I, unfortunately, have an ARM NAS with a horrid operating system based on an ancient fork of OpenWRT that I'd love to replace with Alpine, but I don't think all of the hardware is in mainline kernel and I don't know how to make the BSP so... All end-user for me currently 2022-10-18 03:01:15 i'm talking to you through a alpine-powered irc-bouncer with a GUI provisioned mostly by ansible 2022-10-18 03:03:23 ARM is chaos sure 2022-10-18 03:03:50 but ARM vm is fine, almost the same as x86 2022-10-18 03:03:51 that's the first time i've heard the setup scripts are amazing 2022-10-18 03:04:29 well with setup-alpine -f custom.conf it installs within 60s 2022-10-18 03:04:33 I was able to walk someone through a working Alpine setup via text in about 15min when they've never installed an OS before 2022-10-18 03:04:45 That's a testament to the general setup-alpine flow 2022-10-18 03:04:47 what distro installs within 60s? 2022-10-18 03:06:26 MenuetOS doesn't count, right? 2022-10-18 03:07:50 i once did a presentation where i installed alpine then rm -rf'ed it in 15 minutes 2022-10-18 03:07:58 that's with talking it through 2022-10-18 03:07:59 and then i reboot. and i have another sh script to configure the system i wanted, that took another 60s? 2022-10-18 03:08:28 so, 2 mins, i have a full working system with all the tools and a webserver and a weechat client and logined here 2022-10-18 03:09:19 let's try gentoo :P 2022-10-18 03:10:41 apples to oranges, most serious gentoo users will roll their own stage4's 2022-10-18 03:11:01 (tarball with a kernel that's just a untar + make bootable) 2022-10-18 03:12:01 i mean, apk is super fast. 2022-10-18 03:17:29 plenty of "serious" gentoo users do not make stage4s 2022-10-18 03:17:47 source: I'm a "serious" gentoo user 2022-10-18 03:19:05 double source: was a "serious" gentoo user 2022-10-18 03:19:14 maybe not that serious since i didn't participate in development though 2022-10-18 03:19:37 if you've done the gentoo install enough, you could certainly complete a stage3 installation in 15 minutes 2022-10-18 03:19:47 hah, yeah 2022-10-18 03:19:49 it's a fun speedrun 2022-10-18 03:20:01 especially if you've predownloaded the stage3 and ebuild tree 2022-10-18 03:20:03 hardware permitting, but assuming you don't rebuild gcc it's not super long 2022-10-18 03:21:38 I mean, partition, format fs, mount, unpack stage3 and ebuild tree, chroot, install a prebuilt kernel and grub, configure grub, reboot 2022-10-18 03:22:04 unpacking the stage3 and ebuild tree will take the longest out of all that 2022-10-18 03:31:51 dwfreed: you're missing out friend :) 2022-10-18 03:35:44 the only thing I'd add to a stage3 is kernel, grub, and config management agent 2022-10-18 03:49:44 good mix 2022-10-18 12:20:20 who could this not work when started in the background as ash script (but works as zsh script): setsid nohup $@ &> /dev/null & 2022-10-18 12:20:28 s/who/why 2022-10-18 12:20:28 frag meant to say: why could this not work when started in the background as ash script (but works as zsh script): setsid nohup $@ &> /dev/null & 2022-10-18 14:22:36 hello, there is no qemu sound card supported with alpine virt? 2022-10-18 15:28:30 could someone remind me some a simple progam for file direct sharing? I remeber that I used it once 2022-10-18 15:31:12 donoban, Do you mean something like rsync/unison or NFS or OpenSSH + SFTP? 2022-10-18 15:31:42 it was something like tpaste , you pass some file and it generates a URL for downliading it 2022-10-18 15:37:24 uhm, it seems something like this https://www.tecmint.com/share-files-over-a-local-network-in-linux/ 2022-10-18 15:37:40 but I feel that it was something already packaged on alpine 2022-10-18 15:39:59 maybe it was onionsahre 2022-10-18 15:51:01 dpaste? 2022-10-18 16:43:37 it was onionshare, it provides a onion url for share a file without need to open firewalls, redirect, etc.. 2022-10-18 16:50:54 well, and anonymously obv 2022-10-18 18:28:57 cozy delusion this anonymity in web. there's always party involved who aware of all your file handholding ;) 2022-10-19 01:29:57 and here i am the kind of guy that just pokes holes in his firewall and drops shit in to httpd 2022-10-19 02:08:35 I am missing something or I don't understand something about networking. 2022-10-19 02:09:24 I have eth0 and wlan0. I installed iwd. iwd says that I am connected to the wireless network, but I seem to have no network connection while eth0 is connected 2022-10-19 02:09:57 s/eth0 is connected/eth0 is disconnected 2022-10-19 02:09:57 cache_return meant to say: I have eth0 and wlan0. I installed iwd. iwd says that I am connected to the wireless network, but I seem to have no network connection while eth0 is disconnected 2022-10-19 02:11:20 I do "ifconfig eth0 down" and suddenly I can ping the router and ssh into this computer, but cannot connect to anything on the internet. 2022-10-19 02:17:46 this is the interfaces: https://bpa.st/MXIQ 2022-10-19 02:41:09 cache_return: you need bonded interface 2022-10-19 02:41:26 man interfaces-bond 2022-10-19 02:46:59 systemd networkd has an option Metric or RouteMetric, i don't find similar ones in ifupdown-ng 2022-10-19 02:49:35 Bonding allows connection sharing through multiple interfaces, so if e.g. the wired interface is unplugged, the wireless is still connected and the network connectivity remains up seamlessly. // from arch wiki 2022-10-19 06:49:46 Hej 2022-10-19 06:50:11 Is there support in Alpine for network interface renaming? 2022-10-19 06:50:35 I face a problem with ethX switches after reboots 2022-10-19 06:52:57 Oh I think I found in the wiki. Using mdev and nameif 2022-10-19 07:05:18 Forza: there is also macifrename 2022-10-19 07:32:59 ikke: didn't see that option. Seems to work with nameif and /etc/mactab 2022-10-19 07:34:49 I am a little unsure about mdev.conf though. In the wiki it says to have "-SUBSYSTEM=net....." but should it not be "SUBSYSTEM=net......" without the leading dash? 2022-10-19 07:36:27 if in doubt, you could use eudev. a good chunk of alpine software requires it anyways 2022-10-19 07:38:01 I can't reboot to test at the moment, but in a few hours I have a window to do so. If the mdev.conf isn't working I might change. I prever to stay as stock config as possible as each change has to be documented 2022-10-19 10:19:35 Hmm. Now my system doesn't boot. I have "loading initial ramdisk... " and then nothing.. Sigh.. 2022-10-19 10:23:56 Any idea how to debug that? I tried restoring a backup from earlier 2022-10-19 10:25:20 Can you change the kernel cmd line? Maybe remove quiet? 2022-10-19 10:25:48 Yes. I can access grub so I'll try that 2022-10-19 10:28:43 It booted early this morning so I thought I restored a backup since midnight. But I didn't restore all kernel mods, just the /boot files 2022-10-19 10:33:18 Seems to have made no change with removing quiet 2022-10-19 11:04:49 not sure, but can this be a lack of simpledrm in your initram (would only make sense if haven't updated for a while) 2022-10-19 11:21:53 does that mean I'm supposed to install ifupdown-ng if I want both network interfaces to be functional at the same time? 2022-10-19 11:22:04 uuidNuniq: ^ 2022-10-19 11:22:23 ifupdown-ng should be the default 2022-10-19 11:22:49 uuidNuniq: was saying to use interfaces-bonded which is not on the system 2022-10-19 11:23:20 You need to separately install the bonding package if you want bonding 2022-10-19 11:24:03 do I need bonding? the issue is that wlan0 doesn't fully work unless eth0 is plugged. 2022-10-19 11:24:25 although I have not.tried removing all eth0 configs 2022-10-19 11:24:30 no, you don't need bonding 2022-10-19 11:24:47 what do I need? 2022-10-19 11:25:00 You should not need anything specific 2022-10-19 11:25:30 Not sure why uuidNuniq mentioned bonded interfaces 2022-10-19 11:26:19 cache_return: what does ip route return? 2022-10-19 11:32:44 10.0.8.0/24 dev wlan0 scope link src 10.0.8.20 2022-10-19 11:32:59 no default route? 2022-10-19 11:33:21 that's what ip route returns 2022-10-19 11:33:31 That explains why you don't have ihnternet 2022-10-19 11:33:37 I posted my interfaces file earlier 2022-10-19 11:33:42 yes 2022-10-19 11:33:49 but they both have the same gateway 2022-10-19 11:34:30 what do you mean? why is that bad? 2022-10-19 11:34:36 thats the router 2022-10-19 11:34:44 yes, not necessarily bad 2022-10-19 11:35:09 if you do: ifdown wlan0 && ifup wlan0 2022-10-19 11:36:10 what is the issue that I have created? I don't understand it. 2022-10-19 11:36:52 At least from what I see now, you don't have a default gateway 2022-10-19 11:36:59 that would explain why you do not have internet 2022-10-19 11:37:12 you could only reach the local lan 2022-10-19 11:40:11 how am I supposed to set it up? 2022-10-19 11:40:38 are you still using iwd? 2022-10-19 11:40:49 yes 2022-10-19 11:41:18 I have no experience with iwd, or if it could interfere with what you setup in /etc/network/interfaces 2022-10-19 11:41:43 Ikke seems it freezes. Can't do ctrl alt del either 2022-10-19 11:41:48 I think if I have the flag off, it uses interfaces instead of its own sauce 2022-10-19 11:43:38 if I understand how I am misconfiguring interfaces, I can figure out the iwd thing, I think 2022-10-19 11:44:39 You need to 1) make sure the interface is up. 2) the interface has a valid address. 3) make sure you have a default route 2022-10-19 11:46:41 I can boot efi and bios USB sticks OK. So hmm 2022-10-19 11:46:51 I'll go back to an older backup maybe 2022-10-19 11:56:26 ikke: default route or default gateway? How do I set it? 2022-10-19 11:57:26 `ip r add default via $IP` 2022-10-19 11:57:56 you should not need to set this when using dhcp 2022-10-19 11:58:34 is that persistent? 2022-10-19 11:58:39 no 2022-10-19 12:03:03 I went back two weeks. Now it boots 2022-10-19 12:03:06 Odd 2022-10-19 12:10:07 cache_return: that's the gateway statement in /etc/network/interfaces 2022-10-19 12:12:46 what should I change? 2022-10-19 12:13:08 both interfaces have the gateway defined 2022-10-19 12:13:28 there is some basic thing I'm missing 2022-10-19 12:17:58 I reinstalled the new Linux-lts using 'apk fix' and rebooted. It still works. How odd 2022-10-19 12:19:05 So hum. The issue is maybe related to the mdev fix for nic renaming 2022-10-19 12:19:59 The wiki said I should regenerate initramfs. But perhaps I had an updated kernel installed but not rebooted, so it made a mismatch between the kernel and initramfs 2022-10-19 12:25:07 ikke: is the answer to delete one of the gateway lines, so it's only under one interface? If that's true, then what about when that one is not connected? 2022-10-19 12:25:26 per this: https://serverfault.com/questions/589673/changing-the-default-gateway-multiple-nics 2022-10-19 13:08:56 cache_return: the problem is that you have configured both the ethernet and wifi interface to be on the "same" IP subnet 2022-10-19 13:10:01 that should work because of metrics magic 2022-10-19 13:10:05 it might be slow tho 2022-10-19 13:11:59 satmd: well if his RPI wants to get to any machine on that IP subnet (such as the router for the default route) then which interface is it going to use? Also how will another host on the same subnet get traffic back to his RPI (which which interface)? 2022-10-19 13:12:23 it will use the interface with the lower metric value if the route is identical 2022-10-19 13:12:41 wired devices usually have a lower metric 2022-10-19 13:13:18 ... at least for ipv4, this seems to not work with v6 2022-10-19 13:14:13 the wifi is going to an AP though and general an AP will use a different subnet for its wired connection than for its wireless connections 2022-10-19 13:15:26 so it is not clear how the wired network and the Wifi network/AP interact 2022-10-19 13:30:09 I didn't think of any of this. Wired is a rare use case for me. 2022-10-19 13:31:13 I didn't explicitly configure a wired subnet. Maybe it exists anyway. 2022-10-19 13:31:32 Will investigate. 2022-10-19 13:34:27 "the wifi is going to an AP though and general an AP will use a different subnet for its wired connection than for its wireless connections" <- this is not generally true 2022-10-19 13:40:17 satmd: it depends, basically we're talking about an unknown as no background information has been provided 2022-10-19 13:41:17 the router/ap is openwrt 2022-10-19 13:42:12 I recall changing the device IP and setting a specific range for dhcp but not much other configuration 2022-10-19 13:42:48 cache_return: is the wired connection also via a port on the OpenWRT router/ap? 2022-10-19 13:43:06 yes 2022-10-19 13:43:46 I think it is set up to assign dhcp on the same subnet for wired and wireless, but I do not recall. 2022-10-19 13:44:13 "it depends" and "this is not generally true" are congruent ;) It's not safe to make either assumption 2022-10-19 13:44:48 satmd: indeed, no assumptions would be needed if adequate information was available 2022-10-19 13:48:40 cache_return: is this OpenWRT router/ap also your internet router? 2022-10-19 13:49:03 yes 2022-10-19 13:49:16 one stop shop 2022-10-19 14:09:01 I think I'll try macifrename next 2022-10-19 14:09:08 ร…รฅ 2022-10-19 14:12:35 Oups :) pocket typing 2022-10-19 14:13:29 https://pastebin.com/raw/9G05D5KJ 2022-10-19 14:14:54 Good morning. I am using edge and today this error appeared in the update. 2022-10-19 16:02:29 Does alpine have an equivalent to systemd-detect-virt? 2022-10-19 16:03:36 something like this https://github.com/kaniini/slabbed-or-not? 2022-10-19 16:04:04 I'm trying to reimplement systemd-hostnamed from scratch, and part of that is detecting when I am running in QEMU (or VMWare or VirtualBox or an apple vm or a windows VM etc) 2022-10-19 16:04:40 That looks promising 2022-10-19 16:04:43 That looks promising 2022-10-19 16:05:48 Looks like cpuid is what I want 2022-10-19 16:17:06 CameronNemo[m]: virt-what perhaps? 2022-10-19 16:18:54 I was trying to avoid reading other people's code, for licensing reasons. 2022-10-19 16:22:33 CameronNemo[m]: you asked for an alpine equivalent to systemd-detect-virt, virt-what appears to be that (as slabbed-or-not is not packaged for Alpine), so what's the issue? 2022-10-19 16:24:27 Good point 2022-10-19 16:25:14 I guess I was just hoping for something more like a command snippet 2022-10-19 16:36:02 anyone know a way to do 3 way merges for .apk-new files? I'm using vimdiff but it would be nice to distinguish my changes from the package update 2022-10-19 16:50:48 Good afternoon. I installed guake on edge, but when it runs this error appears. 2022-10-19 16:51:04 https://pastebin.com/RaPw1XtS 2022-10-19 17:14:40 edsonwolf, Try installing py3-gettext and see if it works. 2022-10-19 17:55:14 ah that thing again 2022-10-19 18:05:12 hm, I'm dealing with some weirdness, running apk update or add just... hangs 2022-10-19 18:05:32 I can, however using curl grab the APKINDEX file, so I can connect 2022-10-19 18:05:39 try strace? 2022-10-19 18:05:41 -v doesn't tell me anything else 2022-10-19 18:07:18 yeah, it hangs on connect(7, {sa_family=AF_INET6, sin6_port=htons(80), sin6_flowinfo=htonl(0), inet_pton(AF_INET6, "2a04:4e42:33::645", &sin6_addr), sin6_scope_id=0}, 28 2022-10-19 18:07:42 why did out of blue, started to use ipv6 2022-10-19 18:13:27 does apk have any pre/post upgrade hooks? 2022-10-19 18:15:48 I think packages do 2022-10-19 18:31:26 so if I set eth0 for dhcp everything works and this is ip route: http://sprunge.us/ah7jKp 2022-10-19 18:31:59 so the probem is caused by two static ips at once, but I don't think related to subnets. 2022-10-19 20:44:48 ACTION got struck with paranoia 2022-10-19 20:45:16 Is it possible to encrypt rootfs of the running system or do I have to reinstall it? 2022-10-19 22:58:39 Ermine: allegedly "cryptsetup-reencrypt --new" (think that is now "cryptsetup reencrypt --new" in recent cryptsetup) can be used to encrypt whilst keeping existing data, however it can't be done on a mounted fs AFAIK, you'd have to boot off another device 2022-10-19 23:06:17 minimal: ok, thank you. I guess I need to setup LVM also? 2022-10-19 23:07:12 Ermine: why? 2022-10-19 23:07:47 LUKS and LVM are 2 different unrelated things, you can use one without the other, but can also use both together as well 2022-10-19 23:08:47 Ah, ok, thank you again. 2022-10-20 10:22:59 can please help me for creating custom kernel for alpine linux 2022-10-20 10:23:23 rlofc: what is it you want to achieve? 2022-10-20 10:26:14 In aports/main/, I am trying to build linux-lts. I want to create my own package and then build custom-kernel for that. 2022-10-20 10:27:04 Ok, and where are you stuck> 2022-10-20 10:27:07 ? 2022-10-20 10:27:53 in the link Alpine Linux provided for custom build, we need to navigate to linux-vanilla in aports/main. That I didnt get it. 2022-10-20 10:28:07 linux-vanilla is now linux-lts 2022-10-20 10:30:00 When I am performing abuild -rK in linux-lts folder, I am getting DSO support routines:dlfcn_load:couldnot load shared library:crypto/dso/dso_error loading shared library libproviders.so 2022-10-20 10:33:43 unknown module name:crypto/conf/conf_mod.c 2022-10-20 10:34:43 What are these errors related to? 2022-10-20 10:34:46 rlofc: on what kind of system are you running this? 2022-10-20 10:36:12 did you make changes to the APKBUILD? 2022-10-20 10:38:14 Yes, But, I don't have a clarity on what all changes we need to perform in APKBUILD 2022-10-20 10:38:41 What is it you want to customize? 2022-10-20 10:40:07 I took gnu plot APKBUILD sample and tried it. 2022-10-20 10:42:05 I am not worried about particular customization, I am just trying to build linux-lts by taking any sample 2022-10-20 10:47:02 It helps to be clear about what you are doing, what you want to achieve (details matter) 2022-10-20 10:51:35 Thanks for that. We don't have any particular thing to customize for now. But we need to check if we can customize Alpine Linux or not and the difficulty in it.So that we can start our project based on client requirements. 2022-10-20 10:52:48 So on what kind of system are you trying to build these packages? 2022-10-20 10:53:31 Having a proper build system saves a lot of head-aches 2022-10-20 10:54:46 x86 2022-10-20 10:54:54 32-bits 2022-10-20 10:54:57 ? 2022-10-20 10:55:01 64 bit 2022-10-20 10:55:27 Alpine Linux installation? 2022-10-20 10:55:40 (I assume so) 2022-10-20 10:56:32 We already installed Alpine Linux in VMware. 2022-10-20 10:56:44 ok 2022-10-20 10:56:54 Did you install the alpine-sdk package? 2022-10-20 10:57:00 yes 2022-10-20 10:57:12 Do you have a dedicated build user? 2022-10-20 10:57:12 added 2022-10-20 10:57:15 (non root) 2022-10-20 10:57:24 yes created. 2022-10-20 10:57:37 What is in your /etc/apk/repositories? 2022-10-20 10:58:28 3.16/main, community 2022-10-20 10:58:44 edge/main, community, testing 2022-10-20 10:58:59 everything is uncommented 2022-10-20 10:59:22 Ok, you should not combine repositories from different versions, that can cause issues 2022-10-20 10:59:41 I would recommend removing the 3.16 repos and then running apk upgrade -Ua 2022-10-20 11:00:32 Or, if you want to build for 3.16 specifically, use that, but then you should also use the 3.16-stable branch of aports 2022-10-20 11:00:53 Thankyou. 2022-10-20 11:01:29 I am switching to 3.16-stable version. 2022-10-20 11:01:48 ok, then remove the edge repositories, and then the command I mentioned 2022-10-20 11:02:34 Ok. Thanks for this. 2022-10-20 12:14:43 is there some simple "log integrity check" for syslog or similar? I wonder if it could be very simple to implement with sha256um the current the message + the last log checksum message 2022-10-20 12:45:12 what is "integrity" here? 2022-10-20 12:47:36 that once it logged the first line, there are no external modifications between it and the last one 2022-10-20 12:48:37 making sure only the syslog daemon can write to it? 2022-10-20 12:49:56 root can modify it, but 'he' can't fake it if has some auto cryptographic signature 2022-10-20 12:50:33 he could just modify subsequent checksums 2022-10-20 12:51:07 but he would probably be more interested on modifying the past 2022-10-20 13:43:42 meh is not so easy, he could rewrote the past until present, you will need checkpoints for determine that it hasn't changed, so new dilema about how/when/where save this checkpoints.. 2022-10-20 13:45:54 just log to a blockchain 2022-10-20 13:45:57 donoban: did you do any internet search for this at all? I did a quick search and the 3rd result was info on how to do this with syslog-ng 2022-10-20 13:47:15 yeah I even read https://www.rfc-editor.org/rfc/rfc5424 2022-10-20 13:47:49 what did you search exactly? 2022-10-20 13:48:16 "syslog integrity" 2022-10-20 13:48:41 exactly same as I did 2022-10-20 13:49:17 nothing useful in first results, what URL did you see? 2022-10-20 13:49:30 fun, i always wanted to implement that 2022-10-20 13:50:42 Oxffrin: it could be a solution if I was really worried about this, maybe it could log only the checkpoints, but if I do something it should be extremally simple 2022-10-20 13:50:45 FOSDEM 2020 talk, "Secure logging with syslog-ng, Forward integrity and confidentialty of system logs" 2022-10-20 13:51:07 not in: https://duckduckgo.com/?q=syslog+integrity&ia=web 2022-10-20 14:01:06 how often does ppc64le package repo update? i noticed that there is a rather large delay for packge between getting into the repository and getting into the apkindex 2022-10-20 14:01:37 same as everything else 2022-10-20 14:01:43 the builder has just been offline for over a week 2022-10-20 14:01:58 probably another day to catch up 2022-10-20 16:00:31 Hi. Just upgraded 7.1 to 7.2 and after logging in some commands stopped working like id or top. hey just create new core files. ktrace showing the last statement to access /var/run/ypbind.lock, and pledge access rpath denied it. and after sigabrt and core created 2022-10-20 16:02:40 running gdb id id.core for example says about "Dwarf error: wrong version in compilation unit header (is 4, should be 2) [ in module /usr/libexec/ld.so] and after access () at /tmp/-:3 and no such file ot directory /tmp- 2022-10-20 16:02:46 what it could be ? 2022-10-20 16:03:09 system was not configured to use ypserver or ypbind 2022-10-20 16:05:34 amnesiac: 7.1 to 7.2 of what? 2022-10-20 16:06:13 ikke, sorry wrong chat 2022-10-20 16:06:38 aha 2022-10-20 16:07:19 ikke: openbsd 2022-10-20 16:07:44 they forgot which channel they were in? ;-) 2022-10-20 16:08:21 psykose, right 2022-10-20 16:08:40 just pick wrong window, guys. Glad you joke about this. 2022-10-20 16:10:43 Happens to everyone who's on IRC for a long time. no big deal. Better to joke about it rather than fuss about it 2022-10-20 20:14:49 Are PKGBUILDs from official repos GPL? 2022-10-20 20:14:56 *APKGBUILDs 2022-10-20 20:37:38 don't think there is a licence at all 2022-10-20 20:55:21 can apk list/fix tracked files if the checksum differs from what the package contained upon install? 2022-10-20 20:55:44 is that apk-fix ? 2022-10-20 20:56:26 how can I list the files instead of doing applying the fix? 2022-10-20 21:01:04 apk audit? 2022-10-20 21:04:09 is there a guide somewhere about how to actually make an APK? I searched the wiki a bit but couldn't find anything 2022-10-20 21:04:27 abuild right? 2022-10-20 21:05:45 something like https://wiki.alpinelinux.org/wiki/Abuild_and_Helpers#Setting_up_the_build_environment is probably a place to start elly 2022-10-20 21:06:33 or rather the whole doc: https://wiki.alpinelinux.org/wiki/Abuild_and_Helpers 2022-10-20 21:06:44 aha, thank you! 2022-10-20 21:06:52 newapkbuild looks like the right starting point 2022-10-20 21:07:38 if you follow those instructions you can build apk files for packages in aports (their repo) or your own directory with an APKBUILD in it. 2022-10-20 21:08:02 yup! I am going to try to create an APKBUILD from scratch, we'll see how it goes 2022-10-20 21:08:05 thanks :) 2022-10-20 21:08:08 cool. 2022-10-20 21:08:27 I find them quite easy. Especially with copy/paste/modify as an option ;) 2022-10-21 00:36:57 WhyNotHugo: https://gitlab.alpinelinux.org/alpine/aports/-/issues/9074 2022-10-21 02:02:15 unrznbl[m]: what about kernel, it has many flavor in there 2022-10-21 03:40:47 is it possible to setup zfs via akms? 2022-10-21 07:44:48 Hello guys ! I'm trying to port Alpine on STM32 board 2022-10-21 07:44:48 Now, I've done some shenanigans with the armhf ISO for Raspberry, so I've added the missing openrc to my distribution and it is OK (except that I cannot login since root is not recognised, but I'll manage, I think) 2022-10-21 07:44:48 I cannot help but wonder why the OpenRC is missing in the minirootfs ? I mean it is aimed mostly for virtual machines and stuff, but wouldn't it be more convenient to add the binary directly in it ? Even more that the inittab reaches for the missing binary 2022-10-21 07:44:48 So in my process I took a OpenSTLinux and wiped out the rootfs to put on the armhf mini rootfs available on https://dl-cdn.alpinelinux.org/alpine/v3.16/releases/armhf/alpine-minirootfs-3.16.2-armhf.tar.gz 2022-10-21 07:44:48 Since STM32MP157 needs a specific partition table in order to boot I cannot take the generic ARM delivery that is available on downloads 2022-10-21 07:44:48 It successfully boots, but search for openrc which is not present in the rootfs linked above (and for any other platforms) 2022-10-21 09:41:57 rgallaispou: minirootfs is mostly targeted towards docker 2022-10-21 09:42:17 So there is no need for openrc there 2022-10-21 10:31:16 ikke: Okay I understand, what bothers me is that in the rootfs inittab reaches for /sbin/openrc, while it is not here 2022-10-21 12:22:55 Hello, does anyone know a program or similar that shows a few infos like weather forecasts or timetables on an infoscreen graphically in "beautiful"? 2022-10-21 13:14:42 I have alpine installed on a laptop, encrypted lvm. The password prompt doesn't display, but if I wait a few seconds and enter the passphrase then it starts booting. Is there config in the initramfs I have to do for it to output to the display? I'm using a thinkpad x220, coreboot, and this shows up with both syslinux and grub. 2022-10-21 13:16:42 I've set GRUB_TERMINAL_OUTPUT="console". THe display doesn't show up until "Loading hardware drivers" is already on the screen in the boot sequence. 2022-10-21 13:22:26 elagost: on which Alpine version? Does the Grub menu not appear? is simpledrm module in your initramfs? 2022-10-21 13:24:46 the initramfs (actually: the kernel) expects console= 2022-10-21 13:25:13 console= can be specified multiple times, but only ONE of them actually accepts input 2022-10-21 13:35:30 The grub module menu appears, shows "loading linux lts..." and "loading initial ramdisk..." 2022-10-21 13:37:17 I did not have simpledrm in /etc/mkinitfs/mkinitfs.conf. Adding that and testing now. 2022-10-21 13:38:45 No, still no prompt. After the hardware drivers load, I can see the "staring openrc..." screen and just above it there's "enter password for /dev/sda2" or something like that. 2022-10-21 13:39:44 Oh, and alpine 3.16. 2022-10-21 13:45:40 elagost: simpledrm would not be in the features list of mkinitfs.conf, the file /etc/mkinitfs/features.d/base.modules should have an entry for "kernel/drivers/gpu/drm/tiny/simpledrm.ko*" 2022-10-21 13:47:42 minimal: thanks. It does have that. 2022-10-21 13:48:54 Anywhere else I can look? I realize it's a weird setup - I don't think this thing actually has a vgabios in it, just a text console, since coreboot and all. 2022-10-21 13:50:35 I also don't have any firmware packages installed. Is there one I need? 2022-10-21 13:51:12 oh, I do have -other, -intel, and -i915, so that might cover it? 2022-10-21 13:52:21 hello, short question. i am trying to build a package which requires /usr/bin/gencat from glibc. is there a way to go around this? i believe other software also uses gencat in the build process... 2022-10-21 13:57:07 elagost: try passing "nomodeset" on the cmdline via Grub 2022-10-21 14:01:19 no dice, same thing. 2022-10-21 14:05:42 elagost: "I don't think this thing actually has a vgabios in it, just a text console" - confused about this, on BIOS systems the text console is provided by VGA 2022-10-21 14:07:14 or do you mean it only has a serial console? 2022-10-21 14:08:01 my apologies, I am probably not saying it correctly, I only have a passing undersanding of it. 2022-10-21 14:08:49 It has vga out, but not a full graphical bios, just a text console. 2022-10-21 14:14:58 Recorded a video. the little pause/cut is where I put my passphrase in http://elagost.com/alpinex220.webm 2022-10-21 14:19:55 elagost: when Alpine finishes booting check whicg graphics driver is loaded (via something like "dmesg" or "lsmod" or "lshw") 2022-10-21 14:21:33 is there a way to list all packages from testing? 2022-10-21 14:21:39 installed packages* 2022-10-21 14:24:06 i915 is in there; https://paste.centos.org/view/fc699f2c 2022-10-21 14:34:56 elagost: I think this is going to be hard to diagnose without seeing info such as your /etc/default/grub, /boot/grub/grub.cfg, dmesg output, /etc/mkinitfs/mkinitfs.conf, lsmod output, etc 2022-10-21 14:35:22 elagost: to clarify, you are booting using BIOS and not UEFI, right? 2022-10-21 14:37:49 everything is basically default. I'm booting BIOS. I can get you that info in a sec; just figure it's a weird problem to have. 2022-10-21 14:42:14 http://elagost.com/logs.tar.gz there's the info. 2022-10-21 15:16:11 elagost: try adding "i915" to cmdline "modules=" list? 2022-10-21 15:22:16 minimal: did not work either. 2022-10-21 15:23:56 elagost: just realised, "kms" is missing from your mkinitfs.conf features list, add it to that, rebuild initramfs, try adding "i915" to cmdline 2022-10-21 15:24:19 s/to cmdline/to cmdline modules list/ 2022-10-21 15:24:19 minimal meant to say: elagost: just realised, "kms" is missing from your mkinitfs.conf features list, add it to that, rebuild initramfs, try adding "i915" to cmdline modules list 2022-10-21 15:25:52 That worked! My hero! :D 2022-10-21 15:26:13 Now how do I make that i915 permanent? 2022-10-21 15:27:14 with those changes the i915 driver is in the initramfs (plus its firmware) and it is loaded by the initramfs (before it tried to unlock the LUKS device). Without those setting only simpledrm is loaded before you unlock and then i915 is loaded afterwards when Alpine boots 2022-10-21 15:28:08 Just tried with_out_ i915 in the cmdline. Still works. 2022-10-21 15:28:16 elagost: where did you put thr crypt settings? did you edit grub.cfg directly? 2022-10-21 15:28:28 So all I needed was kms in /etc/mkinitfs/mkinitfs.conf. 2022-10-21 15:29:31 And I don't know how but the default installation of grub has a pretty barebones /etc/default/grub, and it still manages to pick up my encrypted install. I did it as part of the installer so maybe that's it? 2022-10-21 15:31:58 setup-disk (used by setup-alpine) puts stuff like GRUB_CMDLINE_LINUX_DEFAULT into /etc/default/grub but I don't see that in the file you provided 2022-10-21 15:32:29 I installed it with syslinux (default) and removed/installed grub after the fact. 2022-10-21 15:33:18 /etc/grub.d is populated with a bunch of scripts though. 2022-10-21 15:34:05 I installed alpine this morning so I've not done any customizations beyond the install, swapping out syslinux for grub, and installing a desktop. 2022-10-21 15:34:36 elagost: ah, I keep forgetting Alpine's Grub will read /etc/update-extlinux.conf if present 2022-10-21 15:35:04 so it is likely getting some settings from there 2022-10-21 15:40:44 cool. Well all I really needed was 'kms'. Thank you so much! I appreciate the help. 2022-10-21 15:40:47 elagost: so if you look at a dmesg now that it is working is there any noticable difference in "dmesg" output shortly after the "Alpine Init 3.6.2-r0" line? 2022-10-21 15:41:34 between that line and "Mounting root: ok." 2022-10-21 15:44:10 There are some lines about i915drmfb (fb0) as primary device. 2022-10-21 15:46:36 elagost: right, so the i915 module is being loaded earlier which happens now because the "kms" feature setting makes that module be included in the initramfs and so it can be loaded before the rootfs is mounted. I though you'd also have needed to add "i915" to the cmdline modules list to get it loaded but if it works without that then fine. 2022-10-21 15:50:29 Yeah I figured it was something along those lines, just unfamiliar with how alpine did it. I can't thank you enough. This has been bugging me for better part of a year now. 2022-10-21 15:52:57 whilst searching online I noticed someone here had what sounds like the same problem almost 1 year ago with a coreboot based Thinkpad X230 2022-10-21 15:53:19 haha, also may have been me, I have one of those too! 2022-10-21 15:53:57 it was a different IRC nickname 2022-10-22 01:34:56 Yโ€™all, akms is giving me an error about not having a linux-6.0.2-dev package? I only see 6.0.3 on the repos so this is kind of annoying, I just want to build a ding dang kernel module 2022-10-22 01:40:49 chexo4[m]: which arch? 2022-10-22 01:43:23 because it looks like only riscv64 is on that version. All the other arches moved to 6.0.3 already? 2022-10-22 01:56:16 zcrayfish: 6.0.2 aarch64. Not sure how to update my kernel. Using postmarketOS btw 2022-10-22 02:00:10 I usually use `apk upgrade -a --prune -s' to see if any updates are available... likewise `apk upgrade -a --prune' to do the actual update. I don't know what repos postmarketOS uses so dunno if that will help, but definitely worth a shot. 2022-10-22 02:01:15 I guess I can try that later 2022-10-22 02:02:03 One more thing: 2022-10-22 02:02:03 zcrayfish: how can I turn off this aarch64 laptopโ€™s LCD overnight? I want to be able to get it back by pressing a key but I want to avoid burn in from the overnight build Iโ€™m running 2022-10-22 02:02:17 Wait nevermind itโ€™s already done?? 2022-10-22 02:07:31 are you running in console, Xorg, or Wayland? 2022-10-22 07:26:55 TIL printf "foo %s bar\n" * 2022-10-22 07:58:41 Dear Community, I want to install ossec on postmarketos but it does not work. I tried once via apk to install the package ossec-hids. But here ossec-control is not installed, so ossec-hids can not be started. Secondly, I tried to compile ossec, but in the compilation process I get an error that it can not find the file systemd. This is clear Alpine Linux does not work with systemd. Can you help me to find a so 2022-10-22 08:39:45 hi folks, i've had issues in the past with dhcp leases on alpine linux 2022-10-22 08:39:53 dhcpcd works fine 2022-10-22 08:40:18 but i'm trying to figure out why the default udhcpc doesn't 2022-10-22 08:40:44 udhcpc gets an ip address lease, but never sets the routing table 2022-10-22 08:40:54 and RTNETLINK says 'Network Unreachable' 2022-10-22 09:11:41 did the network device get the ip assigned? 2022-10-22 09:14:43 Yes 2022-10-22 09:21:20 Static address works fine and the same DHCP setup works fine as soon as I add dhcpcd and ifupdown uses that dhcp client as default 2022-10-22 09:21:39 Static address successfully populated the route table 2022-10-22 09:22:00 Dhcp with udhcpc leases the address but with no routing 2022-10-22 09:34:41 after doing setup-interfaces? 2022-10-22 09:40:35 Yes 2022-10-22 09:47:29 I get the same error when trying to add a default route manually 2022-10-22 09:47:56 ip: RTNETLINK answers: Network unreachable 2022-10-22 09:48:26 This when I just try to run: 2022-10-22 09:48:26 Ip route add default via 192.168.0.1 dev wlan0 2022-10-22 09:49:10 that means the kernel has no idea how to get to 192.168.0.1 via wlan0 2022-10-22 09:49:40 are you missing the subnet route? 2022-10-22 09:50:16 There is no output at all from ip route list 2022-10-22 09:50:37 that would explain it, then 2022-10-22 09:50:57 you don't have the subnet route, therefore you can't add routes that would depend on it 2022-10-22 09:54:51 Would you mind spelling that out for me 2022-10-22 09:55:03 I've never messed with routing manually 2022-10-22 10:05:45 normally you would see `192.168.0.0/24 dev wlan0 proto dhcp scope link src ` in that list, and only then you can add a default via .1 2022-10-22 10:05:51 dunno why it wouldn't be there though 2022-10-22 11:16:20 So I can add routes manually without issue 2022-10-22 11:16:39 Don't really see any good debug options in udhcpc to help identify the root cause 2022-10-22 11:29:44 hey everyone 2022-10-22 11:31:25 I figured out how to add debugging to /usr/share/udhcpc/default.sh 2022-10-22 11:31:51 And can see that $ip = 0.0.0.0/0 2022-10-22 11:32:08 So the dhcp lease isn't picking up the assigned address 2022-10-22 11:32:19 I have a strange issue. I am trying to setup diskless mode 2022-10-22 11:32:36 But my USB stick is randomly mounted to either /media/usb or /media/sda1 on subsequent reboots 2022-10-22 11:32:46 I created the bootable USB as follows: 2022-10-22 11:32:59 1.) Format USB stick with MBR scheme and single FAT32 partition 2022-10-22 11:35:23 2.) bsdtar xvf alpine-standard-3.16.2-x86_64.iso -C /path/to/mounted/fat32partition 2022-10-22 11:35:34 3.) boot from said USB disk. 2022-10-22 11:36:01 4.) sometimes I can see the drive mounted as read-only at /media/usb sometimes at /media/sda1 2022-10-22 11:36:12 How could I make that consistent? 2022-10-22 11:36:23 Thank you all <3 2022-10-22 12:03:17 noyau[m]: try wireshark 2022-10-22 12:05:57 hi 2022-10-22 12:06:16 what could i do to fix this https://bpa.st/NLWQ ? thanks 2022-10-22 12:10:24 mmmm: it's an innocent warning 2022-10-22 12:10:38 so the error wont affect? 2022-10-22 12:15:26 No 2022-10-22 12:16:28 thanks 2022-10-22 12:16:38 it keeps saying whtever i do so 2022-10-22 12:27:45 Also an important information I left out is that I am booting the stick via EFI not CSM 2022-10-22 12:30:22 iluvalpine: try writing .iso to stick with dd 2022-10-22 12:33:51 Ermine: Thank you, but then I would not be able to remount it as read-write :( 2022-10-22 12:34:32 Why do you need it read-write? 2022-10-22 12:35:06 Installation mediums are meant to be immutable I suppose 2022-10-22 12:35:13 to store custom configs on it 2022-10-22 12:35:20 and use LBU 2022-10-22 12:35:48 the problem is that if I use LBU it will make an apkovl, which contains references to the mountpoint (APK cache location for example) 2022-10-22 12:36:08 so if /etc/apk/cache gets linked to /media/usb/cache and I do an LBU commit 2022-10-22 12:36:26 then I only have the cache available if the system mounts my USB disk at /media/usb on next boot 2022-10-22 12:36:42 if it mounts it as /media/sda1 on next boot then no cache and packages can not be installed :( 2022-10-22 12:36:52 Anyway, does this issue reproduce if you try to insert your stick with another distributions running? 2022-10-22 12:37:09 sounds like a race in the default mdev naming mount thing 2022-10-22 12:37:12 wouldn't be the first time 2022-10-22 12:38:09 Ermine: this is specific to how alpine works in Diskless mode, checking other distros is not relevant here 2022-10-22 12:38:42 @psykose thanks, this is really strange. I've tried it on multiple computers, all behave the same (all boot via EFI) 2022-10-22 12:39:41 What I can do is manually hack the initramfs and insert some autorun after switch_root to make sure that /media/usb is mounted, and if not then symlink it to /media/sda1, which is super ugly hack to do 2022-10-22 12:39:49 open an issue https://gitlab.alpinelinux.org/alpine/mdev-conf here i guess 2022-10-22 12:40:17 Thank you, will do that! 2022-10-22 13:39:04 "noyau: try wireshark" <- Yeah so comparing two different tcpdumps, my mobile hotspot provides a subnet mask in the dhcp lease 2022-10-22 13:39:17 But the dhcp lease from my pfsense firewall does not 2022-10-22 13:39:26 Just the address and gateway information 2022-10-22 13:47:02 what's your /etc/network/interfaces 2022-10-22 13:49:27 i guess udhcpc doesn't handle the case of no mask 2022-10-22 13:49:33 not sure if it's required or not by spec 2022-10-22 13:51:54 I got it, pfsense breaks RHEL in the same way 2022-10-22 13:52:07 You have to add a parameter to the dhcp config to add the subnet into the dhcp lease packets..... 2022-10-22 13:53:21 heh 2022-10-22 13:53:45 i guess dhcpcd defaults to a /24 mask or something to be more resilient 2022-10-22 14:00:56 This has been a real PITA for me across multiple devices as dhclient I believe breaks as well.....glad I found it 2022-10-22 14:01:16 I'll look through dhcpcd to see if it does indeed apply a default 2022-10-22 14:06:55 "/* If we haven't been given a netmask for our requested address, * set it now. */ if (ifp->options->req_addr.s_addr != INADDR_ANY && ifp->options->req_mask.s_addr == INADDR_ANY) ifp->options->req_mask.s_addr = ipv4_getnetmask(ifp->options->req_addr.s_addr);" 2022-10-22 14:27:08 finally used up almost all RAM, by running multiple ffmpeg :p 2022-10-22 14:51:26 mandatory reminder that pfsense registered opnsense.com and filled it with defamatory information before they were ordered to cut it out 2022-10-22 14:51:41 https://opnsense.org/opnsense-com/ 2022-10-22 15:06:29 "mandatory reminder that pfsense..." <- Thanks for sharing 2022-10-22 15:06:54 It looks like OPNsense has better monitoring now too...gonna look into switching 2022-10-22 15:32:52 hi, i just installed mpv on alpine 3.16 but i can't launch it. I get libavutil: 57.28.100 -> 57.17.100 2022-10-22 15:33:49 is there anything special to do to make it works ? 2022-10-22 15:35:05 could you pastebin the full output 2022-10-22 15:35:13 and /etc/apk/repositories 2022-10-22 15:35:28 scruffy:~$ mpv 2022-10-22 15:35:30 libavutil: 57.28.100 -> 57.17.100 2022-10-22 15:35:32 Aborted 2022-10-22 15:35:38 this is the full output ^___^ 2022-10-22 15:36:05 hooo 2022-10-22 15:36:17 in repositories i may have mixed versions 2022-10-22 15:36:25 community on edge and main on 3.16 2022-10-22 15:36:46 :) 2022-10-22 15:36:53 ho yes it works \o/ 2022-10-22 15:36:54 thanks a lot 2022-10-22 15:38:13 on rpi, will you get all upgrades when upgrading the system with apk compared to new install with new tarball? 2022-10-22 20:46:50 frag: if you have a run-from-ram system, you won't get a newer kernel 2022-10-22 20:51:27 ikke: i installed on hd/usb :] 2022-10-22 20:52:25 Then yes, you should get everything 2022-10-22 20:53:35 sweet! 2022-10-22 22:07:32 what is the package syncthing-utils for? 2022-10-22 22:08:33 I see what's in it, but I don't know what that stuff does. What functionality does it enable? 2022-10-22 22:12:24 hi everyone, when i install putty with apk, i only have puttygen in commands. How can I get the basic part of putty 2022-10-22 22:12:27 ? 2022-10-22 22:16:13 qhat does putty do on linux? 2022-10-22 22:17:23 It's for ssh, serial and telnet connection with a graphic interface 2022-10-22 22:18:41 I thought it was for that on windows because windows doesn't have openssh. 2022-10-22 22:20:27 admin-sys1: are you looking for putty on purpose for some specific functionality, or do you just need to ssh? 2022-10-22 22:23:32 I agree it's not usefull 2022-10-22 22:23:32 to fully explain, it s a project for school and they ask us to install putty... 2022-10-22 22:25:04 maybe they don't expect you to be on linux with a ssh client? 2022-10-22 22:26:49 i don't know, i only need to get the graphical part of putty to be installed as it is on windows 2022-10-22 22:27:05 I can see the packet on alpine package website 2022-10-22 22:27:23 and when i install it with apk, it doesnt work 2022-10-22 22:30:13 it looks like it has different executables for different uses. 2022-10-22 22:30:19 https://pkgs.alpinelinux.org/contents?branch=edge&name=putty&arch=riscv64&repo=main 2022-10-22 22:31:12 it looks like even the x86-64 version only includes some utilities from putty 2022-10-22 22:31:32 yeah, sorry random arch. 2022-10-22 22:33:39 arch's putty package has everything you'd want as far as I can see 2022-10-22 22:34:04 so maybe you should ask the alpine pkg's maintainer 2022-10-22 22:34:05 omg i didnt see those commands, i feel dumb right now 2022-10-22 22:35:02 never mind for the graphical part 2022-10-22 22:35:14 thank you guys !! 2022-10-22 22:37:55 was mistakenly missing, should exist in -r1 2022-10-22 23:19:53 the putty executable? 2022-10-22 23:32:14 cache_return, I think the Putty package was build without GUI. It looks like the Putty package got just updated (on Edge). 2022-10-22 23:33:29 interesting 2022-10-23 00:56:49 root should always own /etc right? 2022-10-23 01:14:11 cache_return: yes 2022-10-23 02:05:00 it's weird. I created a nonroot user and it owns /etc and like three files in there. 2022-10-23 02:14:26 I think it's because I used an overlay on initial boot. 2022-10-23 02:14:38 It's all startup stuff. 2022-10-23 03:34:45 My alpine experiment continues.....how do you folks handle glibc compatibility? Gcompat still isn't linking all the necessary stuff (getdirentries64: symbol not found) 2022-10-23 03:35:50 I have 1-2 funky applications (one being citrix receiver) that will for sure be the most difficult part of moving to musl 2022-10-23 04:08:10 just run debian in a chroot 2022-10-23 07:18:48 Done and works, but with arch as the base 2022-10-23 07:19:21 And apparently sudo has a dependency on systemd-homed now......I won't miss systemd that is for sure 2022-10-23 08:14:42 noyau[m]: I tried flatpak for my stuff, it apparently works. 2022-10-23 08:26:55 on that note recently i had an idea to recomplie all pip packages against musl and setup a mirror 2022-10-23 08:26:55 i wonder if anyone else has already done this 2022-10-23 09:03:32 I started there, but flatpak contains all the dependencies for each app....a chroot is cleaner if you have more than a handful no? 2022-10-23 09:03:53 Can anyone point me to an article on recompiling glibc projects with musl? 2022-10-23 09:04:26 I can compile from source but because its a single make command.....the subtleties of one c library to another escape me 2022-10-23 09:05:00 Do I change make flags to point to the musl tool chain or? 2022-10-23 09:09:50 noyau[m]: glibc may have features that musl doesn't have, so if program doesn't build with musl, it might need patching. 2022-10-23 09:14:44 And if program is non-free, you cannot recompile it against musl. You'll need to use chroot or flatpak, or a container. 2022-10-23 09:22:26 noyau[m], generally, compiling a project on alpine will make it use musl without extra effort 2022-10-23 09:34:16 "noyau, generally, compiling a..." <- So just the same make install and make will automatically pick up the musl tool chain? 2022-10-23 09:34:34 I've only tried one which builds with gradle.....but it wasn't that easy 2022-10-23 09:34:48 Granted its a Java app so yet another dependency.... 2022-10-23 09:35:44 Purely an intellectual exercise for me, but postmarketos got me interested and I'm curious to see if I can get alpine as my daily driver 2022-10-23 10:47:00 hello, I want to create a new aport. Should I put myself as contributor or mantainer in the file? 2022-10-23 10:52:12 also, it's actually two packages, one depends on the other. Should I make a single MR? 2022-10-23 10:54:03 elris: You can put yourself as maintainer indeed, and it can be a single MR 2022-10-23 10:57:19 ikke: thank you. For the MR how do I do it? Do I need to make a new branch on my fork? 2022-10-23 10:57:59 Yes, create a new branch based on master and push it to your fork. 2022-10-23 12:24:00 openrc question. What does this mean? command_user="${SYNCTHING_USER:-syncthing}" 2022-10-23 12:24:41 actually a shell scripting question 2022-10-23 12:25:02 ${parameter:-word} 2022-10-23 12:25:04 Use Default Values. If parameter is unset or null, the expansion of word is substituted. Otherwise, the value 2022-10-23 12:25:06 of parameter is substituted. 2022-10-23 12:25:09 from man bash, but i trust it also applies to alpine's default shell :) 2022-10-23 12:25:37 that tracks. 2022-10-23 12:25:43 :) 2022-10-23 12:26:19 so what is the right way to set the user for this rc script? how do I pass that option? 2022-10-23 12:26:43 (now it's an openrc question, right?) 2022-10-23 12:26:56 it is looking for an environment variable called SYNCTHING_USER; where to set that, i don't know (because yes :) ) 2022-10-23 13:49:32 where are apks actually stored on disk? 2022-10-23 13:49:50 they are unpacked 2022-10-23 13:50:44 ah, but the original apks are not cached anywhere? 2022-10-23 13:51:17 Only if you enable caching 2022-10-23 13:51:35 alright 2022-10-23 13:51:54 By creating /etc/apk/cache (or more commonly symlinking it) 2022-10-23 13:51:56 I would like to learn how to package things for alpine so I went looking for a sample one 2022-10-23 13:52:18 we have setup-cache 2022-10-23 14:07:08 elly: they have that on the wiki: https://wiki.alpinelinux.org/wiki/Creating_an_Alpine_package 2022-10-23 14:08:55 elly, better look at APKBUILD files then 2022-10-23 14:09:27 dewdude: ah, thank you! I was reading https://wiki.alpinelinux.org/wiki/Abuild_and_Helpers and trying to figure it out backwards 2022-10-23 14:10:53 the APKBUILD files are also a good resource. don't forget this is musl based; so MOST of your stuff will need to be patched to compile 2022-10-23 14:11:27 i built asterisk from source using the APKBUILD and Aports files as a guide. holy cow. i learned a lot though 2022-10-23 14:11:58 yup :) I believe the program I'm trying to package (angband) will build fine with musl already but if not I'll fix it upstream 2022-10-23 14:12:56 i'm trying to figure out now if there's a limit to the total number of system pipes I can use. I got something that doesn't run despite being essentially similar to 5 other things running 2022-10-23 14:13:14 but alpine is not the target system for this so i might just move to a different lab machine 2022-10-23 17:52:02 is there a package which contains man pages for libc functions like malloc(3)? 2022-10-23 17:54:55 man-pages 2022-10-23 17:55:23 and then man-pages-posix for a few extra things (but not malloc) 2022-10-23 17:56:15 cheers, thank you 2022-10-23 17:56:19 I was looking for musl-doc 2022-10-24 10:21:18 hello. In alpine 3.16-stable trying to start nut by "upsdrvctl start" gives me "Can't chdir to /var/run/nut: No such file or directory". nut installed with apk. any ideas? 2022-10-24 10:23:56 why not "rc-service nut start"? 2022-10-24 10:29:44 that's it 2022-10-24 10:29:46 thanx 2022-10-24 10:48:39 hrm 2022-10-24 10:48:44 I seem to have put myself in a bit of a bind 2022-10-24 10:48:48 $ doas apk fix apk-tools 2022-10-24 10:48:53 (1/1) [APK unavailable, skipped] Reinstalling apk-tools (2.12.9-r7) 2022-10-24 10:51:19 apk upgrade -a fixed it, don't ask me why 2022-10-24 10:59:51 can I get logs from prior boots ร  la 'journalctl -b-1'? 2022-10-24 11:01:58 if they were saved in /var/log 2022-10-24 11:04:09 under what? doesn't seem to be anything in messages 2022-10-24 11:04:34 depends what logs you're looking forโ€ฆ 2022-10-24 11:05:06 the dmesg 2022-10-24 11:05:44 it's enough if I can just set it to record the next one. 2022-10-24 11:06:32 maybe add it to logrotate 2022-10-24 15:16:55 cache_return: in /etc/conf.d/bootmisc set "previous_dmesg=yes" 2022-10-24 15:34:55 what's the best way to check whether the community repository is enabled on an alpine host? 2022-10-24 15:35:12 for a script, I mean, not a human 2022-10-24 15:37:56 grep community /etc/apk/repositories 2022-10-24 15:39:17 that would fail with the default configuration, which has community commented out but still present 2022-10-24 15:39:35 yep 2022-10-24 15:39:50 i figured there should be some apk command to probe but i don't see it 2022-10-24 15:40:33 something like grep "[^#].*/community$" /etc/apk/repositories 2022-10-24 15:40:45 right now I'm using: grep -E '^[^#]+/community$' /etc/apk/repositories 2022-10-24 15:40:54 but I thought maybe there was a nicer way to do it 2022-10-24 17:46:20 hello there, is openjdk available for alpine? 2022-10-24 17:46:35 it used to but can't find it anymore on oracle website 2022-10-24 17:47:17 https://pkgs.alpinelinux.org/packages?name=openjdk*&branch=v3.16&repo=&arch=x86_64&maintainer= 2022-10-24 17:47:19 why would you be looking on the oracle website? 2022-10-24 17:47:28 if it's available in alpine it's in the alpine package repository 2022-10-24 17:48:01 the whole point of having a distro is that you don't get things from sketchy upstreams like oracle :) 2022-10-24 17:48:52 thanks lopid 2022-10-24 17:48:57 & @dal 2022-10-24 17:49:02 dalias* 2022-10-24 17:56:31 hey 2022-10-24 17:56:51 on 32bit alpine the firefox is the version that requires SSE2 capable CPU ? 2022-10-24 17:57:17 probably 2022-10-24 18:19:44 is there any usable web browser that does not require SSE2 in the repo? 2022-10-24 18:20:46 no idea but the alpine baseline for all of x86 is technically sse2 (pentium4) now 2022-10-24 18:21:06 maybe the webkit ones work, i don't have an sse-less machine to check 2022-10-24 18:22:30 then there is no sense to use alpine 32bit on sse2-less machine? 2022-10-24 18:22:39 even like Athlon XP 2022-10-24 18:25:30 technically for now it mostly works, but soon enough it wont 2022-10-24 18:25:32 so i guess not 2022-10-24 18:27:13 from what you said it will boot up but almost nothing more when on almost every step I will get into lack of SSE2 2022-10-24 18:27:56 I would guess you will need to compile some things yourself with -mno-sse2 2022-10-24 18:28:08 no, for now it's still technically the i586 no-sse build. but not far in the future it won't be, and then practically every single program will fail all of a sudden 2022-10-24 21:25:02 <@psykose> no, for now it's still technically the i586 no-sse build. but not far in the future it won't be, and then practically every single program will fail all of a sudden <-- may i ask why that is? I'd say without support for non-sse2 cpus there really is not much point in supporting 32 bit x86 at all since there aren't many non x86_64 cpus that do support sse2.. 2022-10-24 21:27:43 https://gitlab.alpinelinux.org/alpine/tsc/-/issues/20 2022-10-24 21:28:34 oh, that is really unfortunate :/ 2022-10-24 21:28:50 alpine was really nice for running on i586 cpus 2022-10-24 21:31:06 but if that's not longer wanted, why not drop x86 support altogether? pentium 4's are probably the least useful old cpus and pentium m's are really uncommon by now 2022-10-24 21:31:32 on the other hand, there are a couple of embedded cpus, that are indeed i586 compatible, but do not support sse2 2022-10-24 21:33:18 no idea, i don't really think the decision makes sense 2022-10-24 21:33:42 mh, yeah :( 2022-10-24 21:33:51 well, that issue says it's kept 2022-10-24 21:33:54 then in https://gitlab.alpinelinux.org/alpine/tsc/-/issues/35 it's changed 2022-10-24 21:35:06 i would very much like alpine to keep x86 support with 586 baseline. it is actually pretty usable on such old systems since alpine is so lightweight 2022-10-24 21:35:56 i don't really get the point that's being made in the issue, that many packages already require i686. most of those packages will never be run on 586 class computers anyway 2022-10-24 21:40:21 what even is the difference between 586 and 686 2022-10-24 21:40:24 sse? mmx? 2022-10-24 21:40:53 difference between i586 vs. i686 is mostly CMOV 2022-10-24 21:41:09 i686 does not require mmx nor sse 2022-10-24 21:41:27 ah 2022-10-24 21:41:27 mmx was available on later i586 cpu, e.g. pentium mmx 2022-10-24 21:41:56 the original SSE was first available on pentium III 2022-10-24 21:42:14 SSE2 on pentium 4 and athlon 64 2022-10-24 21:42:50 so there really aren't many cpus at all that support sse2 but don't support x86_64 2022-10-24 21:43:00 even some later pentium 4's are 64 bit 2022-10-24 21:43:12 and all amd cpus are 2022-10-24 21:53:58 Hello71 did raise a good point in that issue 35 when it comes to floating point precision in certain programming languages, but at the same time also mentioned that requiring sse2 for php or go was not such a big deal and i would agree with that. i still find it kind of sad that that meant all of alpine dropping non-sse2 cpu support. 2022-10-24 22:06:15 technically the change hasn't been made yet 2022-10-24 22:07:18 wej, the k6 is not 64bit :) 2022-10-24 22:07:45 dalias: no, but it does not support sse2 either 2022-10-24 22:08:15 dalias: AMD's earliest cpu to support sse2 is the athlon64 2022-10-24 22:08:41 even the athlon xp doesn't support sse2 2022-10-24 22:10:39 ah you're counting machines with sse2 but not 64-bit 2022-10-24 22:11:34 dalias: yeah, because that's what alpine 3.17 will support if the change by the TSC is done 2022-10-24 22:14:42 that seems rather useless 2022-10-24 22:15:03 my preference is to keep the i586 target profile minimal. i actually wish it had been i486 2022-10-24 22:15:18 (i586 has almost nothing useful at the isa level over i486, btw) 2022-10-24 22:15:40 yeah, exactly, that's what i thought 2022-10-24 22:15:45 (and the only useful things i486 has over i386 are bswap and cmpxchg) 2022-10-24 22:15:52 yep 2022-10-24 22:16:15 pentium has cpuid, cmpxchg8b, rdtsc 2022-10-24 22:16:16 you need cmpxchg to implement a working multitasking system, so there's no point in making the isa profile target i386 2022-10-24 22:16:23 i would say those are pretty useful 2022-10-24 22:16:26 i disagree 2022-10-24 22:16:43 rdtsc is very useful as a fast clock source for vdso 2022-10-24 22:16:51 but that's not exposed to direct use by userspace 2022-10-24 22:17:06 the kernel can give it out just if it's available and it falls back gracefully if not 2022-10-24 22:17:24 cpuid likewise is only useful to kernel. userspace has the hwcap vector 2022-10-24 22:18:09 and cmpxchg8b is not something you can portably rely on having 2022-10-24 22:18:17 most archs don't have double-pointer-sized cas 2022-10-24 22:18:52 so in order for it to be useful you have to be writing target-specific synchronization primitives that use very different design depending on target 2022-10-24 22:18:56 which is a big no-no 2022-10-24 22:19:41 it's worth pointing out that the kernel is looking into dropping 486 support as you may know 2022-10-24 22:20:01 :/ 2022-10-24 22:20:40 aren't there nice clone (possibly even open hardware) x86's that do 486 but not 586+ ? 2022-10-24 22:20:53 at much higher clock than original 486 of course 2022-10-24 22:21:04 iirc there were like 1-2 ghz ones 2022-10-24 22:21:23 *shrug* 2022-10-24 22:21:33 i very much understand their reasoning 2022-10-24 22:21:45 does it cost a lot to keep it in the kernel? 2022-10-24 22:21:48 there are open source 486 cpu designs that can be used on fpgas 2022-10-24 22:21:56 wej, that's what i thought 2022-10-24 22:22:05 meaning you can fab them in silicon too if you want 2022-10-24 22:22:14 yeah 2022-10-24 22:22:26 i don't know if anyone has done that, but it should be possible 2022-10-24 22:22:49 https://t.me/+PEc8FSRwFiFiYzc0 2022-10-24 22:22:55 a lot of stuff pentium introduced is common to modern architectures and maintaining hairy codepaths (like not being to assume cpuid, or u64 atomics) is kinda meh 2022-10-24 22:23:22 they can't assume u64 atomics in general tho 2022-10-24 22:23:23 there are also some embedded x86 cpus that are technically i486 like the earlier dm&p vortex86 cpus 2022-10-24 22:23:29 because most 32bit archs don't have them 2022-10-24 22:23:36 (for context https://lore.kernel.org/lkml/CAHk-=wikUaRM5H_y1Bc+QyvGi40dKDL8fnCTyz7ECbwK7aHNPQ@mail.gmail.com/) 2022-10-24 22:23:46 so only x86-specific code could be assuming that with 486 removed 2022-10-24 22:23:52 not the kernel at large 2022-10-24 22:24:27 anyway i don't think 486 is on the table for alpine 2022-10-24 22:24:39 so for alpine purposes, my vote is keeping i586 as real i586 baseline 2022-10-24 22:24:57 yeah, that would be my preference too 2022-10-24 22:25:15 i'm curious what kind of graphics you can run on such a machine in practice 2022-10-24 22:25:22 why would you run any graphics? 2022-10-24 22:25:24 i really don't see the point in keeping 32 bit x86, if the baseline is raised to sse2 2022-10-24 22:25:41 mesa has sse2 in it on x86 (easy to change fwiw), mesa-dri-classic doesn't exist so no old gpus work that won't be supported by i915 gallium 2022-10-24 22:25:42 i mean you can. X runs fine 2022-10-24 22:25:58 the xorg drivers work i guess and no acceleration does but i mean 2022-10-24 22:26:07 if i had to use something like that i would rather just live in the woods 2022-10-24 22:26:13 you're not gonna be gaming on it 2022-10-24 22:26:17 so who cares 2022-10-24 22:26:26 i'm not even referring to gaming 2022-10-24 22:26:27 (or if you are it's retro gaming and runs fine :) 2022-10-24 22:26:40 on the i586 computers i have run alpine on, i usually didn't use any graphics 2022-10-24 22:27:46 bizarre considering you can most likely buy an equivalently powerful arm board and make back the cost in electricity in under 2 years or something 2022-10-24 22:27:50 where i'd run i586 alpine is on a box with 2-4 GB of ram where the memory workload is pointer-bound and running 32-bit basically doubles your ram 2022-10-24 22:27:52 unless you have 20 solar arrays in your house 2022-10-24 22:28:06 x32 would be better for that if we had an x32 distro but *shrug* 2022-10-24 22:28:32 is there anyone shipping x32 i wonder 2022-10-24 22:28:46 sabotage isn't a binary distro but supports x32 2022-10-24 22:28:51 eh 2022-10-24 22:28:53 sort of gentoo and sort of debian 2022-10-24 22:28:58 i don't think x32 every really took off. i don't know of any distro using it 2022-10-24 22:29:15 linux constantly threatening to remove it has basically ensured it wouldn't take off :( 2022-10-24 22:29:16 oh, interesting that there are some 2022-10-24 22:29:36 if linux had a viable mechanism for the x32 support to be in userspace so that could be ruled out 2022-10-24 22:29:43 then it might take off 2022-10-25 05:26:29 quick question if anyone is here 2022-10-25 05:26:46 how does lbu commit keep track of changes to the filesystem 2022-10-25 05:27:15 how does it know what to files to save, for the next boot overlay 2022-10-25 05:27:49 ikke: hola 2022-10-25 05:29:46 oneinsect: https://gitlab.alpinelinux.org/alpine/alpine-conf/-/blob/master/lbu.in#L16 2022-10-25 05:30:01 thanks (as always) 2022-10-25 05:30:05 you are a savior 2022-10-25 05:30:30 aha so these are predefined paths 2022-10-25 05:30:37 interesting 2022-10-25 05:32:11 so no INCRON or INOTIFY required to keep track of file changes 2022-10-25 05:32:20 very interesting 2022-10-25 05:33:16 yeah, it does not actively track it 2022-10-25 05:33:44 it creates a tar file when you run lbu commit from the list of files that are tracked 2022-10-25 05:36:31 basically from ${ROOT}etc/apk/protected_paths.d/lbu.list 2022-10-25 05:36:41 which itself is modifiable 2022-10-25 05:36:55 yes 2022-10-25 05:37:41 where is the source of this lbu.list file ikke 2022-10-25 05:39:42 I think by default it's empty, which would default to /etc 2022-10-25 05:43:02 hmmmm 2022-10-25 05:43:31 oneinsect: I think it's closely related to the files that apk itself would not overwrite if there are changes 2022-10-25 05:44:16 the other aspect is that packages themselves and installed packages list file are also preserved and reinstalled on every subsequent boot 2022-10-25 05:44:32 that's in /etc/apk/world 2022-10-25 05:44:38 which is included in the overlay 2022-10-25 05:45:17 indeed 2022-10-25 06:24:33 morning. i seem to have issues configuring foreground colors in vim. i'm running a terminal with a dark background and vim's default dark-blue color for comments isn't exactly helpful here. putting 'highlight Comment ctermfg=cyan' into /etc/vim/vimrc.local doesn't seem to help either, so how would i do that on alpine? 2022-10-25 06:25:44 before anything else, "set bg=dark" should hint it to choose default colors more suitable for a dark bg 2022-10-25 06:27:56 vimrc.local should work on 3.16 or edge, but you should verify /etc/vim/vimrc actually loads it and it's not in .apk-new pending overwrite 2022-10-25 07:22:50 psykose: that is already verified. 2022-10-25 07:24:32 grawity: seems to do the trick. 2022-10-25 07:25:21 grawity: thx. 2022-10-25 09:31:43 f 2022-10-25 11:14:43 Greetings. After upgrading from 3.13 to 3.16, I noticed that I can no longer use ip netns. Is there a way to get it back? 2022-10-25 11:15:34 apk add iproute2 2022-10-25 11:19:18 It works, thanks a lot! ๐Ÿฅณ 2022-10-25 19:20:22 Hi @ all 2022-10-25 19:22:23 I try to install "qemu-guest-agent" an start it but ist alway goes to crashed ... I don't know why? 2022-10-25 19:22:53 feuerteufel: do you see the process running? 2022-10-25 19:23:10 Let me see 2022-10-25 19:24:46 ps aux | grep qemu-guest-agent <-- say's me no see yust the grep 2022-10-25 19:25:40 Does /etc/init.d/qemu-guest-agent -d start return something relevant? 2022-10-25 19:28:56 Ok "start" dosn't work it say's it is running. 2022-10-25 19:29:53 I tryed "restart" and there I get a lot of output ... 2022-10-25 19:30:12 Have to find a plase to paste 2022-10-25 19:31:37 https://tpaste.us 2022-10-25 19:34:15 https://tpaste.us/Jrla 2022-10-25 19:34:35 That's just 2 lines 2022-10-25 19:34:40 But that's not what I see here 2022-10-25 19:34:54 are you only capturing stdin, not stderr? 2022-10-25 19:36:34 Looks like 2022-10-25 19:41:00 Hope you able to read that one ... https://pastebin.com/TsGAXPJU 2022-10-25 19:45:38 Let me know, if you need a nother place to paste 2022-10-25 19:46:31 No, that's fine 2022-10-25 19:46:59 Does /var/log/qemu-ga.log contain something? 2022-10-25 19:47:26 I allready used that before, nevver eany probs yust apk add ... - rcupdate add ... 2022-10-25 19:49:03 Now here ;) 2022-10-25 19:49:16 https://tpaste.us/8blR 2022-10-25 19:49:48 Ok, not sure how to fix that though 2022-10-25 19:51:11 Did I something wrong? 2022-10-25 19:51:15 No idea 2022-10-25 19:58:48 I already have a VM where it works, it's the same Alpine version, may that help to look up? 2022-10-25 20:00:29 It does mean it's not some issue with qemu-guest-agent in general 2022-10-25 20:03:37 ikke: I rebooted one mor time (6 or 7 times allready did) and now it looks like ist works 2022-10-25 20:03:46 lol 2022-10-25 20:04:28 But I only started the service by hand 2022-10-25 20:05:57 I'll add it to default and try again 2022-10-25 20:06:39 I'm sorry about that noise! 2022-10-25 20:07:33 But first I Think i'll get a coffee more :-)) 2022-10-25 20:07:35 No worry 2022-10-25 20:07:45 Thank you ikke! 2022-10-26 00:28:13 Does anyone know what needs to be fixed in Rhythmbox on Edge with this error message? 2022-10-26 00:28:16 Trace/breakpoint trap 2022-10-26 00:28:16 (rhythmbox:25450): libsoup-ERROR **: 20:28:02.626: libsoup2 symbols detected. Using libsoup2 and libsoup3 in the same process is not supported. 2022-10-26 00:28:41 Is that an APKBUILD fix for a depends/makedepends, or is that something that needs fixing upstream? 2022-10-26 00:38:38 it is strange, grub is able to boot alpine-virt.ISO but after sys installation, it is not able to boot the same OS on HD. it is a vm started by github.com/cirruslabs/tart which uses Apple's Virtualization.Framework on macOS 13.0 2022-10-26 00:40:29 when boots, the grub menu shows, CPU usage is 1core 100%, but stucks at loading vmlinuz-virt after 2 seconds autoboot, and then the CPU useage is 4core 100%. i think i may need to uninstall grub-efi and grub, and use efibootmgr instead. 2022-10-26 00:43:56 looks like setup-disk does not have an option to NOT install a bootloader. 2022-10-26 00:47:44 time to get rid of vmware and qemu. 2022-10-26 00:48:29 the linux performance is awesome native via apple's Virtualization.Framework 2022-10-26 00:53:47 maybe the /etc/defualt/grub is lack of some options, i found on my archlinux arm installation (via tart) there are addtional GRUB_GFXMODE=auto and GRUB_GFXPAYLOAD_LINUX=keep GRUB_TERMINAL_INPUT=console . i'll try to add these options before trying efibootmgr 2022-10-26 03:03:23 aha, no success, is there a difference between the grub in the ISO and the grub in the alpine repo? 2022-10-26 03:15:51 the grub in the repo installs core.efi to ESP/boot/bootx64.efi. while the grub in the ISO has all these modules builtin : grub_mod="all_video disk part_gpt part_msdos linux normal configfile search search_label efi_gop fat iso9660 cat echo ls test true help gzio" 2022-10-26 03:16:09 i think that is the difference part 2022-10-26 03:26:49 the ESP mount point in setup-disk is hardcoded to /boot/efi, instead of /boot, which means, the kernel in not in a vfat partition, which means efibootmgr method is not going to work. 2022-10-26 03:29:07 i'm gonna change 'boot/efi'it to 'boot' in the setup-disk script 2022-10-26 03:50:41 also changed the 64 / 1000 + 2 to 640 / 1000 + , so the boot size is now 320M, lol 2022-10-26 03:50:50 instead of 32M 2022-10-26 03:51:14 now fire up efibootmgr! 2022-10-26 09:51:15 fail) ; can you point me where I can find more logs about what's happening ? or is it a known issue ? 2022-10-26 09:51:15 Hello everyone ! I have an installation issue, using the `alpine-minirootfs-3.16.2-aarch64.tar.gz` rootfs with the `v2.12.10/x86_64/apk.static` apk binary running on debian, when I run `apk.static --root /path/to/minirootfs update` then `apk.static --root /path/to/minirootfs add nginx`, I get `ERROR: nginx-1.22.1-r0.pre-install: script exited with error 127` (nginx post-install and busybox-1.35.0-r17.trigger also 2022-10-26 09:54:16 my guess is you mix aarch64 with x86_64 2022-10-26 09:54:55 that looks like not running as root 2022-10-26 09:55:04 no permissions for chroot 2022-10-26 09:56:09 If I don't run the procedure as root, I get `ERROR: nginx-1.22.1-r0.pre-install: chroot: Operation not permitted` instead (it was running under sudo) 2022-10-26 09:56:59 hmm, mixing aarch64 and x86_64, the system doing the installation is in fact x86_64, the target system is aarch64 2022-10-26 09:57:11 do you have qemu? 2022-10-26 09:57:30 I have, yes 2022-10-26 09:59:55 The full build log is here : https://pastebin.com/ME1jnH5N 2022-10-26 10:28:08 so apk will execute the trigger and post-install scripts in the chroot 2022-10-26 10:28:29 that wont work if you try execute aarch64 binaries on an x86_64 machine 2022-10-26 10:28:45 you could make it work though, with qemu and binfmt 2022-10-26 10:29:22 yup 2022-10-26 10:29:22 but you might need a static qemu-aarch64 binary temorarily in your chroot 2022-10-26 10:37:00 ncopa: is this behavior new ? I mean this procedure was working perfectly fine last week (I don't think I ran it earlier this week) 2022-10-26 10:37:44 no, nothing new 2022-10-26 10:38:54 so there must be something wrong on my system them, I used it for days without problem and today it is failing, I might have broken something ; I'll have a look into that direction, thank you 2022-10-26 11:46:20 is it allowed to ask for gentoo's etc-update instead of update-conf to be available in one of the next alpine releases? 2022-10-26 11:50:04 it is always alloed to ask :) 2022-10-26 11:50:57 "it is just you might not like the answer"? ;-) 2022-10-26 11:52:25 yeah... 2022-10-26 11:52:40 okay, i'll elaborate on the question: i am used to my configuration update scripts not only showing me differences between current active and distribution configuration files. gentoo's etc-update also allows for integration of changes between both, that is, new parts can be integrated with the current admin-modified version. alpine's update-conf doesn't do that. 2022-10-26 11:53:30 i suppose "patches are welcome" is the answer 2022-10-26 11:53:32 instead in alpine i have to go through each affected file manually copy new snippets into the current configuration or, vice versa, transfer my changes into the new distributed file. 2022-10-26 11:54:17 would be nice to fix this, indeed. unfortunately, my hands are more than full already 2022-10-26 11:54:45 and if we want this to be a part of alpine-conf, then it can not depend on too many external tools 2022-10-26 11:54:54 well... 2022-10-26 11:55:08 i can try to get my hands dirty there. 2022-10-26 11:55:08 but it might be an idea to provide an etc-update package that pulls in the deps 2022-10-26 11:55:32 IIRC, the busybox tools are a bit limited 2022-10-26 11:55:35 (it's not like it'd be my first alpine APKBUILD) 2022-10-26 11:56:04 well, apparently it's basically a bash (and that's a point there) script. 2022-10-26 11:56:14 do you have an url to the etc-update sources? 2022-10-26 11:56:25 i suppose it also depends on a proper diff and other stuff 2022-10-26 11:58:00 ncopa: ikke: it started to work again, it seems I didn't reboot after an update or qemu was not properly updated ; Things are running well now ; Thank you 2022-10-26 11:58:32 https://gitweb.gentoo.org/proj/portage.git/tree/bin/etc-update 2022-10-26 12:08:26 false positive, sorry, it is still failing 2022-10-26 12:09:06 ACTION deconfuses everyone by fixing this to: 'false negative' about the fix 2022-10-26 12:10:03 -de? 2022-10-26 12:10:43 Marquel: if you install vimdiff, it shows you an option 2022-10-26 12:12:20 Marquel: looks like its not too big job to port it to alpine 2022-10-26 12:12:33 would be nice to port it to posix shell 2022-10-26 12:26:01 yeah, i already figured busybox is not exactly easy going. 2022-10-26 12:26:57 ikke: the point is - i want it to go, like update-conf, through all the files there are with updates and then give me the option to merge them properly. 2022-10-26 12:32:36 Marquel: yes, that's what happens, though you would merge it through vimdiff 2022-10-26 12:32:47 But it gives you the option 2022-10-26 12:34:03 But you need to have vimdiff installed, otherwise it would not show the option 2022-10-26 13:03:16 I definitely fixed the issue with `proot -q qemu-aarch64 -S /path/to/minirootfs -w / apk ...` I have no idea why it stopped working as before with the static apk build; Anyway, thank you for your help 2022-10-26 15:38:38 etc-update? but alpine has a update-conf already 2022-10-26 15:43:00 uuidNuniq: re you earlier questions, setup-disk (currently) does not handle bootloader-less situation, that might change in future (i.e. for EFISTUB) 2022-10-26 15:45:45 uuidNuniq: also "the ESP mount point in setup-disk is hardcoded to /boot/efi" is not correct, depending on the type of install either /boot or /boot/efi may be the ESP filesystem, as can be seen in lines 298 and 307 of setup-setup 2022-10-26 19:29:16 is there a way to get `apk upgrade` to give me a view of the changelogs of packages I'm upgrading? I know `apk upgrade -s` shows me that they're going from version x to version y, but is there a way to see diffs or similar? 2022-10-26 19:29:51 in particular I would love to know whether linux-lts 5.15.74-r1 is new enough to have a fix for those wifi CVEs :) 2022-10-26 19:29:58 changelogs? no 2022-10-26 19:30:39 it is new enough yes 2022-10-26 19:31:00 lopid: hmm, okay 2022-10-26 19:31:19 psykose: that is good to know - is there a way I could have found that out myself? 2022-10-26 19:32:33 hmm, that's quite hard to find 2022-10-26 19:32:36 grr 2022-10-26 19:32:37 you could check the git log for aports? 2022-10-26 19:32:45 why is everyone so terrible at this 2022-10-26 19:32:51 the aports log isn't useful for this specifically 2022-10-26 19:33:48 yeah it's somewhat dependent on the maintainer providing the info, Omni does provide the CVE info in his commit for 5.15.74 though 2022-10-26 19:34:24 I could look at changes to the APKBUILD and see if the underlying git revision the package was built from changed maybe? 2022-10-26 19:35:06 elly: You would check what kernel versions fixed that issue, and verify if you have that kernel or newer 2022-10-26 19:36:56 that's the thing :p finding that is quite hard 2022-10-26 19:37:18 we don't really backport kernel patches (just update to the point releases, so there's nothing fancy) but just finding that .74 fixed it is a challenge 2022-10-26 19:37:32 all the google results are random blogspam or distros backporting patches into x.y.z 2022-10-26 19:37:53 i only knew it because of the oss-sec mailing list that it was picked into .74-rc 2022-10-26 19:39:06 maybe there's some one good source that keeps track of all of them 2022-10-26 19:53:31 for linux-lts specifically, there is a comment in the APKBUILD that says: "secfixes:" and then lists package revisions and CVEs 2022-10-26 19:54:37 That list is severely incomplete 2022-10-26 19:54:37 it's like that for every package, linux-lts is just missing like all of them except for this one time 2022-10-26 19:54:43 yeah 2022-10-26 19:54:59 If we would track all CVEs for the kernel, that list would be 1000th of lines long 2022-10-26 19:55:27 right, I do not think that a comment in the APKBUILD is the ideal way to do this 2022-10-26 20:00:05 hm, looking at the git logs on the APKBUILD files is indeed pretty un-illuminating... it looks like pkgrel and such is often bumped in a commit of its own, not as part of another commit 2022-10-26 20:00:10 oh well 2022-10-26 20:00:35 a pkgrel being bumped alone is usually just for rebuilds against some dependency 2022-10-26 20:00:54 gotcha 2022-10-26 20:02:58 unrelated (well, maybe a bit related) question: if I'm packaging something, is there an expectation that I just do the build of it myself and then package the resulting binaries? or is there some build service that is supposed to do them 2022-10-26 20:03:28 The package is being built by the alpine builders 2022-10-26 20:04:11 You typically create a merge request against aports providing a new APKBUILD, which then first builds it in a CI environment 2022-10-26 20:04:17 once merged, the builders will pick it up 2022-10-26 20:04:24 cool, thank you :) 2022-10-26 20:04:36 The APKBUILDs are expected to build everything from source 2022-10-26 20:05:00 so, with some exceptions, an APKBUILD that just packages pre-built binaries is unlikely to be accepted 2022-10-26 20:07:51 that is very good to hear, I was wondering how the actual binaries get built 2022-10-26 20:08:59 ncopa: you own the chromium package, right? are you interested in trying to upstream some of the patches in the tree? I am a chromium developer and can probably help if so 2022-10-26 20:09:45 elly: in general, we prefer patches to be upstreams 2022-10-26 20:09:48 upstreamed* 2022-10-26 20:10:10 yeah, I also would prefer if these were upstreamed :) but I don't own the copyright on them to do so 2022-10-26 20:10:14 So I expect that help to be appreciated 2022-10-26 20:10:28 understood 2022-10-26 20:10:31 looking around, some of these can't go upstream but it looks like a lot can 2022-10-26 20:13:53 elly: I would love to have as much as possible upstreamed 2022-10-26 20:14:26 but honestly, for the last few months other people have done the real work of maintaining chrome in alpine 2022-10-26 20:14:38 gotcha 2022-10-26 20:16:15 realistically it's been me maintaining it 2022-10-26 20:16:25 you can take authorship of every patch there, i wrote like 40% of them 2022-10-26 20:16:26 yeah 2022-10-26 20:16:30 they're not really upstreamable though 2022-10-26 20:16:47 they're the equivalent of an #if 0 a lot of the time, unless chromium likes more __GLIBC__ macro checks than now 2022-10-26 20:16:50 some of them definitely aren't but I bet the ones for building against musl are (missing includes of sys/types.h or whatnot) 2022-10-26 20:17:05 sure, yeah 2022-10-26 20:17:27 also check out the epic hack i have in partition-atfork.patch 2022-10-26 20:18:03 oh gosh 2022-10-26 20:18:06 what's going on in there? 2022-10-26 20:18:48 yeah, what does that solve? 2022-10-26 20:19:07 idk why it works, but the reason it doesn't work on musl is one of those atfork handlers call something that isn't AS-safe or something 2022-10-26 20:19:17 you get an infinite spinlock otherwise 2022-10-26 20:19:18 O_o yikes 2022-10-26 20:19:29 ah, no 2022-10-26 20:19:40 well, I'll check with one of the linux folks and see if they care to support building against musl 2022-10-26 20:19:43 not as-safety but a specific different thing that uses a glibc internal semantic 2022-10-26 20:19:46 if so I bet a lot of this can go upstream 2022-10-26 20:19:47 that is exactly what atfork is supposed to prevent happen 2022-10-26 20:20:08 i should really document the damn patches more often 2022-10-26 20:20:16 ah 2022-10-26 20:20:16 right 2022-10-26 20:20:22 psykose: please do :) 2022-10-26 20:20:25 it would be convenient if it had a bug number, yeah :P 2022-10-26 20:20:36 pthread_atfork in musl calls malloc 2022-10-26 20:20:40 in glibc it does not 2022-10-26 20:20:45 that is undefined, but the issue is 2022-10-26 20:20:51 this specifically is used with the malloc shim 2022-10-26 20:21:05 to the partition alloc shim calls pthread_atfork.. which calls malloc.. which calls back into the shim 2022-10-26 20:21:08 which already holds the spinlock 2022-10-26 20:21:11 so it spins forever 2022-10-26 20:21:12 oh! 2022-10-26 20:21:23 I've fixed other bugs of that exact form before, with malloc re-entering the RNG and similar 2022-10-26 20:21:27 not valid to assume no malloc in pthread_atfork i guess? 2022-10-26 20:21:29 aha, sounds good 2022-10-26 20:21:29 that's probably fixable (and should be fixed) upstream 2022-10-26 20:21:32 yeah 2022-10-26 20:21:40 just personally don't have the energy for it so it's there 2022-10-26 20:21:51 partition alloc often just does things that work under glibc and then moves on 2022-10-26 20:21:56 also fwiw chromium builds without the partition alloc too but i wanted to keep it more official-build style 2022-10-26 20:22:13 do you know which atfork handler it happens in? 2022-10-26 20:22:19 it's not in a handler 2022-10-26 20:22:21 I would leave it in if you can - it has a bunch of security mitigations in it that are fairly valuable 2022-10-26 20:22:22 it's int he call itself 2022-10-26 20:22:36 pthread_atfork in musl immediately calls malloc with redirects to the code (it overrides malloc) and hangs 2022-10-26 20:22:52 the partition allocator is a custom malloc, chromium doesn't use malloc-ng 2022-10-26 20:23:12 unless you disable it, but it has much more tuned memory/speed semantics for chromium itself 2022-10-26 20:23:14 oh, its their own allocator 2022-10-26 20:23:17 quite normal and not unwanted 2022-10-26 20:23:17 ye 2022-10-26 20:24:06 i spent like 6 hours debugging that, it was fun :) 2022-10-26 20:24:24 the "rand calls partition alloc calls rand" bug took me quite a while too 2022-10-26 20:24:30 not very fun to figure out 2022-10-26 20:24:44 indeed. but i guess it's fun in a puzzle sense :) 2022-10-26 20:24:50 but yeah, you can ping me about anything there 2022-10-26 20:24:55 cheers 2022-10-26 20:24:57 i've reworked the whole thing i think so i know most about it 2022-10-26 20:25:08 happy to help if you want to improve anything 2022-10-26 20:25:13 thank you, both of you 2022-10-26 20:25:18 I am guessing that you are not the original author of most of these patches though? 2022-10-26 20:25:29 I'd be happy to provide anything to help if needed too 2022-10-26 20:25:42 the ones that have a patch-source are from where it says 2022-10-26 20:25:56 the rest are mostly me if you want to look at the log, the non-trivial ones aren't 2022-10-26 20:26:17 alright 2022-10-26 20:26:18 the arm64 thing and the sandbox thing etc 2022-10-26 20:26:32 but it's probably not upstreamable as is so any modifications would be fine 2022-10-26 20:26:32 I'll get with the linux team a little bit and see what we can take upstream, we love reducing these kinds of divergences when possible 2022-10-26 20:26:39 and aports isn't exactly very licence-caring 2022-10-26 20:26:49 I noticed that, yeah :P 2022-10-26 20:27:03 (not in terms of upstream code but in terms of itself and patches in it) 2022-10-26 20:27:21 right 2022-10-26 20:27:24 Wonder how much of it is considered copyrightable 2022-10-26 20:27:41 90% is boilerplate 2022-10-26 20:27:58 not a lawyer but by most country laws i'd assume a lot is copyright by default if it's not very trivial 2022-10-26 20:28:03 I don't know, but if I tried to just commit 30 random patches upstream to chromium and my answer to "who wrote these?" was "not sure really" it would probably get rejected :) 2022-10-26 20:28:16 i wouldn't say even a whole apkbuild is trivial? certainly not a ton of patches in there manually written 2022-10-26 20:28:32 as I remember, part of the challenge with chromium code base was that it doesn't to build time detection (meson, configure etc) to detect OS capabilities. it just assumes LINUX == __GLIBC__ 2022-10-26 20:28:35 an #include fix is whatever for sure 2022-10-26 20:28:48 it does have quite a few glibc checks, funnily 2022-10-26 20:28:55 just not universal as they don't test the other platforms 2022-10-26 20:29:09 ncopa: it does a lot of that detection during gn rather than actually at compile time 2022-10-26 20:29:28 yeah the BUILDFLAG stuff is the macros for the most part 2022-10-26 20:29:51 probably how we would do this upstream is add a 'use_musl' gn arg, have the APKBUILD set it, then condition as needed on BUILDFLAG(USE_MUSL) 2022-10-26 20:30:57 I'll investigate and see what happens 2022-10-26 20:31:00 cheers, thanks folks! 2022-10-26 20:31:04 o/ 2022-10-26 20:31:07 thank you! 2022-10-26 20:31:16 :) surprised to see someone that's an upstream dev 2022-10-26 20:31:26 a full time chromium developer, in fact :P 2022-10-26 20:31:30 hope you have fun with the archaelogy here 2022-10-26 20:31:39 thinking the same. I wish Firefox devs would have done this too 2022-10-26 20:31:56 thankfully firefox is a small touch easier 2022-10-26 20:32:04 but really, this is very much appreciated 2022-10-26 20:32:10 yep 2022-10-26 20:32:32 upstreaming will also help thing things like electron I suppose 2022-10-26 20:33:14 ideally! we'll see 2022-10-26 20:34:16 I initially went looking in the APKBUILD to make sure nobody was disabling the sandbox, which I've definitely seen distros do 2022-10-26 20:34:49 electron is the same patchset yeah 2022-10-26 20:34:54 +4 patches for their changes 2022-10-26 20:35:06 sandbox was a bit tricky, because musl uses different syscalls 2022-10-26 20:35:28 that is the change I would most like to upstream and am least confident will get upstreamed 2022-10-26 20:35:34 firefox needed the same finagling for syscall numbers 2022-10-26 20:49:16 I'll get with the linux team a little bit and see what we can take upstream, we love reducing these kinds of divergences when possible 2022-10-26 20:49:31 can you get the sandbox people to stop rejecting ppc64le patches for stupid reasons 2022-10-26 20:50:11 hi! possibly but probably not is the answer 2022-10-26 20:50:23 can you link me an example of one? 2022-10-26 20:50:25 last time these were submitted the answer was basically "we don't care because it adds ifdefs and it's not commercially interesting to us" 2022-10-26 20:51:34 I have no power to change the incentives of other developers I'm afraid 2022-10-26 20:51:36 e.g. https://chromium-review.googlesource.com/c/chromium/src/+/1490513 2022-10-26 20:52:37 it's honestly quite frustrating to maintain a 5KLOC patch downstream (times 5 because also qt5-webengine, electron and other junk) because google doesn't care 2022-10-26 20:52:50 that does sound pretty unpleasant, oof 2022-10-26 20:53:35 reading over this sandbox code it looks like it wants more than a little bit of reworking honestly 2022-10-26 21:30:51 elly: this is the patch i maintain https://github.com/void-linux/void-packages/blob/master/srcpkgs/chromium/patches/xxx-ppc64le-support.patch plus a bunch of minor patches on top of that 2022-10-26 21:31:02 i tried doing big endian too a while back, but in case of chromium that seems completely futile 2022-10-26 21:31:40 the way skia is written is not endian-clean at all, and there are other things where google ripped out correct endianness handling, e.g. in boringssl 2022-10-26 21:32:18 yup, boringssl basically only supports little-endian targets I think 2022-10-26 21:32:39 very intentionally apparently 2022-10-26 21:33:00 where openssl had endian-handling macros boringssl just ripped that out and replaced with raw byte access and type punning 2022-10-26 21:33:43 ya, it's a fork of openssl customized for chromium use and chromium doesn't run on any big-endian targets (iirc) 2022-10-26 21:34:44 chromium's codebase is often very non-portable in places where i don't understand the reasoning why 2022-10-26 21:36:46 psykose: is there a convenient way to run a trial build from an APKBUILD file? I want to see if the build breaks if I remove one of these patches 2022-10-26 21:37:08 if you set up a normal abuild dev environment just running abuild -r should build it 2022-10-26 21:37:14 as for how to do that uhh 2022-10-26 21:37:53 there was a wiki entry somewhere, but it's just clone aports, apk add alpine-sdk, SUDO=doas abuild-keygen 2022-10-26 21:37:59 ah 2022-10-26 21:38:05 abuild-keygen -ai 2022-10-26 21:38:20 adduser abuild 2022-10-26 21:38:23 and that 2022-10-26 21:38:31 and having doas configured (or sudo without that added) 2022-10-26 21:38:39 and make sure you're on edge to work with it 2022-10-26 21:39:44 and distfiles 2022-10-26 21:40:00 i guess https://wiki.alpinelinux.org/wiki/Creating_an_Alpine_package#Setup_your_system_and_account does have it all 2022-10-26 21:40:05 thanks! 2022-10-26 21:45:06 news 2022-10-26 21:46:28 news 2022-10-26 21:46:40 and then you leave 2022-10-26 21:50:13 as if i'd make it so easy for you 2022-10-26 21:50:36 Doesn't prevent me from trying :P 2022-10-26 21:50:43 what news do you have today q67 2022-10-26 21:51:18 how long till q 2022-10-26 21:51:55 i've been q66 for almost 2 decades 2022-10-26 21:51:58 you'll have to wait a while 2022-10-26 21:52:20 only 60+ more years to go 2022-10-26 21:52:24 'sfine, I've got time 2022-10-26 23:12:01 psykose: ncopa: what's the deal with quiche-arena-size.patch for chromium? 2022-10-26 23:13:11 it predates me. reasoning is https://gitlab.alpinelinux.org/alpine/aports/-/merge_requests/23366#note_169719 2022-10-26 23:13:56 it was adjusted in.. 102? to the 1280 amount or something 2022-10-26 23:14:09 maybe it's not needed anymore, but i didn't have the time to test at the time 2022-10-26 23:14:17 hmm, okay 2022-10-26 23:14:38 thanks 2022-10-26 23:15:21 this was back when the apkbuild actually ran tests, hah 2022-10-26 23:15:36 then some time passed and effort ran out i guess 2022-10-26 23:16:12 building the tests needs another bunch of patches, then there's more requirements, but iirc not impossible 2022-10-26 23:16:17 just always have a ton of things to do 2022-10-26 23:16:22 yeah, valid 2022-10-26 23:33:36 tbh those test suites take a good while to run these days 2022-10-26 23:33:40 I would not care to regularly run all of them 2022-10-26 23:38:33 yeah, but some minimal tests are probably quite useful 2022-10-26 23:38:58 you can see based on check() what was ran before 2022-10-26 23:39:26 you would likely get 80% of the value from running base_unittests since that is where a lot of the platform-specific goo is, and the other 20% from unit_tests 2022-10-26 23:42:42 sounds about right tbh 2022-10-26 23:43:09 you can get them to build if you want :) or i'll look at it next go-around 2022-10-26 23:43:29 ah, there was another hacky patch 2022-10-26 23:43:30 my todo list for this evening includes setting up a dev environment for that so I'll try it 2022-10-26 23:43:39 the scoped_file_linux one 2022-10-26 23:43:52 yes, I saw that... that seems like it's papering over a real bug 2022-10-26 23:44:17 but even if we do wish to continue papering, there is a better way (that g_is_ownership_enforced knob) 2022-10-26 23:44:18 it's a different hacky fix for https://gitlab.alpinelinux.org/alpine/aports/-/merge_requests/23366#note_169692 iirc 2022-10-26 23:44:24 yeah well 2022-10-26 23:44:27 that's why it's hacky 2022-10-26 23:44:31 i did toggle that 2022-10-26 23:44:37 it crashed in the same place anyway, idk how 2022-10-26 23:44:40 interesting 2022-10-26 23:44:45 hm! 2022-10-26 23:44:48 at runtime it took that codepath anyway 2022-10-26 23:45:10 well, my plan is basically to turn some of these into upstream chromium bugs, some of them into upstream patches, and ideally just plain delete some of them 2022-10-26 23:45:17 note that this is only with thinlto plus the partition alloc 2022-10-26 23:45:20 just in case if it's of relevance 2022-10-26 23:45:25 i did all of those at once as an improvement 2022-10-26 23:45:31 before that, the patch was in a different form 2022-10-26 23:45:32 sec 2022-10-26 23:48:00 ah, it was that same merge request 2022-10-26 23:48:01 https://gitlab.alpinelinux.org/alpine/aports/-/blob/2418074f45e25d08fc0f0fb367b4ef27b4eb971e/community/chromium/scoped-file.patch 2022-10-26 23:48:16 used to do something like this but i think this was a different workaround 2022-10-26 23:48:25 anyway those validation checks only failed with lto enabled for some reason 2022-10-26 23:48:45 i also tried CFI but that had a violation in sqlite somewhere at runtime so i never ended up keeping it 2022-10-26 23:48:53 gotcha 2022-10-26 23:49:01 I don't think we ship CFI in official builds of chrome either still 2022-10-26 23:49:12 yeah it's a tricky beast 2022-10-26 23:49:40 right now i remember, that patch is just because __close doesn't exist on musl 2022-10-26 23:49:58 i assume on glibc they override a symbol with close() then call real __close after some validation 2022-10-26 23:50:03 so the workaround does the reverse 2022-10-26 23:50:09 eh, seems unrelated to the toggling and fd checks 2022-10-26 23:51:40 well, we'll see :) the linux team are on board with "non-invasive" musl support patches although they do not want to officially support it 2022-10-26 23:52:55 i'm surprised anyone planned it at all :p 2022-10-27 00:13:52 could anyone using dnsmasq confirm if config option address=/#/1.2.3.4 as in https://wiki.archlinux.org/title/Dnsmasq#Override_addresses works in v2.86, pls ? been trying to figure out that for hours now :( 2022-10-27 00:15:05 is 1.2.3.4a legitimate dns 2022-10-27 00:15:28 have tried debug mode too, https://bpa.st/JL5A 2022-10-27 00:15:34 ah 2022-10-27 00:15:34 hm 2022-10-27 00:15:34 sec 2022-10-27 00:15:39 psykose: that is an example 2022-10-27 00:17:34 2.87 changelog says 2022-10-27 00:17:36 Replace --address=/#/..... functionality which got 2022-10-27 00:17:36 missed in the 2.86 domain search rewrite. 2022-10-27 00:17:52 i assume 2.86 has something broken with it? 2022-10-27 00:19:12 hmm, thanks 2022-10-27 00:19:45 but i also can't get it to work so 2022-10-27 00:20:27 note that /#/ in address only works for things not in /etc/hosts or dhcp leases 2022-10-27 00:20:57 so while it would match example.com, it wouldn't match your device's DHCP hostname 2022-10-27 00:21:32 is there a way to test it when using dnsmasq as only a dns cache on localhost 2022-10-27 00:22:20 ah 2022-10-27 00:22:31 yeah it gave me 1.2.3.4 for a random domain 2022-10-27 00:23:28 yes its like catch-all 2022-10-27 00:24:23 I tryied dig @1.2.3.4 ANY 2022-10-27 00:27:10 will try to install and see if djbdns works with similar options 2022-10-27 00:29:49 thanks 2022-10-27 00:32:17 ah, yes, changelog say, 2022-10-27 00:32:27 Fix problem with the configuration 2022-10-27 00:32:29 --server=/some.domain/# --address=/#/ --server= 2022-10-27 00:32:31 Yes 2022-10-27 01:17:36 minimal: thanks for the info. tried many methods, the alpine-virt.ISO started by tart works fine except it keeps trying /dev/ttyAMA0 which is not found, a simle touch /dev/ttyAMA0 makes the message disappear. however after installation, it won't boot the HD, grub menu shows, then stuck there. i tried efibootmgr to add the kernel image directly to the UEFI menu, it does not boot 2022-10-27 01:17:37 the kernel. then i made a unified kernel image to replace grub's /boot/efi/bootaa64.efi, it boots to blank. 2022-10-27 01:22:18 uuidNuniq: /dev/ttyAMA0 should be the serial port for a console on aarch64 2022-10-27 01:22:40 yeah i know, tty0 works. 2022-10-27 01:22:58 it works? I though you said screen was blank 2022-10-27 01:22:59 AMA0 not found on this vm. 2022-10-27 01:24:07 i mean the ISO boots. and it uses tty0. 2022-10-27 01:25:35 right, but the ISO isn't necessarily using the same initramfs contents and cmdline as the Sys-mode install you did 2022-10-27 01:26:10 i mean, the ISO boots fine, it has 2 "console=" args, one is tty0 one is ttyAMA0, on the login: prompt, it keeps saying ttyAMA0 not found every second. so i just login and 'touch /dev/ttyAMA0' :) 2022-10-27 01:26:16 the message then disappear 2022-10-27 01:26:32 does "tart" support serial devices? 2022-10-27 01:27:17 https://0x0.st/o3Oz.txt 2022-10-27 01:29:07 tart is dead simple, not many options. 'tart create alpine --linux --disk-size 20' 'tart run alpine --disk /path/to/iso' 2022-10-27 01:29:44 right, I'm assume it provides no serial device which is why ttyAMA0 is missing 2022-10-27 01:29:45 file @ is considered ok, I guess https://pkgs.alpinelinux.org/contents?branch=edge&name=dnscache&arch=armv7&repo=main ? 2022-10-27 01:30:15 minimal: i added the console=/dev/tty0 to the hd installation of grub.cfg though, same result. 2022-10-27 01:30:45 uuidNuniq: just "console=tty0" 2022-10-27 01:30:55 i am just so curious, why alpine-virt ISO boots, but hd installation stucks 2022-10-27 01:31:08 minimal: yeah, like that. typo 2022-10-27 01:31:23 uuidNuniq: some modules missing from initramfs? 2022-10-27 01:31:31 uuidNuniq: as I said it might be due to difference in initramfs contents 2022-10-27 01:32:04 uuidNuniq: what cmdline value are you using? 2022-10-27 01:32:06 aha. that's a thing i was missing to check. 2022-10-27 01:32:32 so, chroot to edit the FEATURES? and rebuild the initramfs? 2022-10-27 01:32:44 looks like tart might only support virtio graphics 2022-10-27 01:33:03 root=UUID=5b307815-eb17-4d63-a9fa-152c92aba677 ro modules=ext4 quiet console=tty0 rootfstype=ext4 2022-10-27 01:34:14 it is funny when boot the installed HD, when grub menu shows, CPU 100% :P 2022-10-27 01:34:26 from the view of macOS 13 2022-10-27 01:35:37 https://0x0.st/o3OP.txt 2022-10-27 01:35:52 from the archlinux arm in the tart vm. 2022-10-27 01:37:19 uuidNuniq: yupe, virtio GPU, so is virtio_gpu.ko.gz in the initramfs and then try adding virtio_gpu to "modules=" on cmdline? 2022-10-27 01:37:45 s/virtio_gpu.ko.gz/virtio-gpu.ko.gz/ 2022-10-27 01:37:45 minimal meant to say: uuidNuniq: yupe, virtio GPU, so is virtio-gpu.ko.gz in the initramfs and then try adding virtio_gpu to "modules=" on cmdline? 2022-10-27 01:38:21 ok. i am gonna try that today, thanks a lot! 2022-10-27 01:39:00 is it possible to enable that during installation or has to chroot after installation to do the job 2022-10-27 01:40:02 uuidNuniq: if "kms" is in the mkinitfs.conf features list that will add virtio-gpu.ko.gz into the initramfs 2022-10-27 01:40:41 you could chroot (adding setting up /dev, /sys, etc) and run mkinitfs 2022-10-27 01:41:16 yeah, i'll chroot then. 2022-10-27 01:41:53 I assume you already have "virtio" in the features list 2022-10-27 01:42:18 but when the grub menu shows, the apple's vm process uses 100% cpu that is weird, 2022-10-27 01:44:37 I've seen (rather heard from fans) excessive CPU sometimes when sitting in a VM in Grub editing a boot menu 2022-10-27 01:46:48 uuidNuniq: I could probably tweak my Alpine VM creation script to handle tart 2022-10-27 01:47:45 https://0x0.st/o3Om.txt 2022-10-27 01:48:01 minimal: for your info. how tart/apple handles vm 2022-10-27 01:48:32 that's the files and ports open by the vm process 2022-10-27 01:49:53 once am able to boot alpine from HD via tart, i am gonna get rid of vmware and qemu 2022-10-27 01:50:44 the performance and resources usage is far better than qemu on macOS 2022-10-27 01:55:42 this is the default features in a alpine virt installation: features="ata base ide scsi usb virtio ext4" 2022-10-27 01:57:18 uuidNuniq: from the info I've seen so far you only really need "base kms usb virtio ext4", and maybe "scsi" if it can provide something like a CDROM devices 2022-10-27 01:58:18 as the majority of devices it presents are virtio: virtio-blk, virtio-net, virtio-rng, virtio-gpu, virtiofs 2022-10-27 01:59:21 no virtio-crypto device though it seems 2022-10-27 02:00:59 minimal: wait a minute, i also tried install alpine lts, same result. 2022-10-27 02:01:55 i think lts already includes kms by default? 2022-10-27 02:02:44 maybe it is something else? 2022-10-27 02:04:14 uuidNuniq: the ISO or when setup-disk is run you mean? 2022-10-27 02:05:29 both virt lts ISOs boots fine. then i run setup-alpine and do a sys installation, then both stuck after grub 2022-10-27 02:07:28 from a quick look, setup-disk does not put "kms" in features list 2022-10-27 02:09:06 ok, i am tring again now, by adding kms and cmdline 2022-10-27 02:10:04 by the way, the archlinux arm kernel was build against alpine-virt kernel .config :P, i built it manually 2022-10-27 02:16:18 lsmod from alpine-virt ISO : https://0x0.st/o3OY.txt 2022-10-27 02:18:17 cmdline from virt ISO: BOOT_IMAGE=/boot/vmlinuz-virt modules=loop,squashfs,sd-mod,usb-storage quiet console=tty0 console=ttyAMA0 2022-10-27 02:19:16 uuidNuniq: a "dmesg" would be nice ;-) though it might be truncated unless you add "log_buf_len=32768" to cmdline 2022-10-27 02:20:14 dmesg : https://0x0.st/o3O6.txt 2022-10-27 02:20:37 yeah, kms is a must :) 2022-10-27 02:28:02 uuidNuniq: there's no sign of simpledrm being loaded in that dmesg output, just virtio-gpu 2022-10-27 02:31:28 now am tring to mkinitfs. there are some options i am confused like : -f use fstab instead of /usr/share/mkinitfs/fstab 2022-10-27 02:32:12 general inside a chroot I just run "mkinitfs" with no options 2022-10-27 02:32:13 or just run: mkinitfs 5.15.74-0-virt 2022-10-27 02:32:33 done 2022-10-27 02:32:41 if you're not in a chroot then yeah you'd have to provide various options to point to the files inside the dir 2022-10-27 02:33:13 BTW "mkinitfs -l" does not create a initramfs but shows you, based on the current config, what *would* be placed inside the initramfs 2022-10-27 02:33:30 features="base scsi usb virtio ext4 kms" 2022-10-27 02:34:38 minimal: thanks for the '-l' info. 2022-10-27 02:35:14 in chroot has to provide a version, or else it says /lib/modules/5.15.59... not found 2022-10-27 02:37:33 uuidNuniq, yeah you're right, otherwise I think it does a "uname" to find the version (which will be of the host machine which may not be the same as the version installed inside the chroot) 2022-10-27 02:39:22 no luck 2022-10-27 02:39:41 still stuck there and vm process uses 400% CPU because i give it 4cores 2022-10-27 02:41:25 one sec, it might be missing fat module (for ESP partition) 2022-10-27 02:43:25 nope, the "usb" feature should add the fat module 2022-10-27 02:56:58 removed the 'quiet' from grub menu, but still not able to see any detailed info. it just shows 2 lines: loading linux virt, loading ramdisk 2022-10-27 02:57:41 I guess you won't see anything until/unless virtio-gpu is loaded, did you add virtio-gpu to the "modules=" list? 2022-10-27 03:02:27 i have modules=ext4,virtio-gpu,virtio_gpu,simpledrm,drm,fb 2022-10-27 03:03:00 anything that can be added was added :) by editing the menu 2022-10-27 03:05:29 so you're back to using grub rather than booting a unified EFI kernel/initramfs file? 2022-10-27 03:05:57 yeah, i am trying grub first. 2022-10-27 03:06:08 easier to do. 2022-10-27 03:06:27 UKI is a lot of steps :P 2022-10-27 03:08:22 the thing about tart/apple vm framkework is that when the hd install finished, you can't boot to ISO any more, it always boot from hd, even with --disk /path/to/iso and there is no way to enter UEFI SHELL, no such thing. 2022-10-27 03:08:42 so to try another method like UKI i have to delete the vm, and create again 2022-10-27 03:08:44 that sucks 2022-10-27 03:08:46 when you had it booted in past (i.e. via ISO) were there any serial devices present (i.e. /dev/ttyS* or /dev/AMA*) - wondering if you should get serial console working to debug (if Tar supports that) 2022-10-27 03:10:21 as for UEFI shell, yeah some motherboards don't provide those either, which is why EDK2 provides a shell EFI file that can be used 2022-10-27 03:10:22 from alarm i see there is : /dev/ttyS0 /dev/ttyS1 /dev/ttyS2 /dev/ttyS3 2022-10-27 03:10:32 no A* 2022-10-27 03:12:25 tryed adding 'console=ttyS0' to grub menu, same thing 2022-10-27 03:13:00 so try adding a getty on /dev/ttyS0 in inittab (i.e. "ttyS0::respawn:/sbin/getty -L ttyS0 115200 vt100") and when booting use "console=ttyS0,115200" (you mau 2022-10-27 03:13:34 that is weird. i need to dig into how the virt-ISO was created, what modules it adds to "initramfs" and "grub" 2022-10-27 03:14:02 thats in the setup scripts in aports/scripts/ 2022-10-27 03:14:18 https://gitlab.alpinelinux.org/alpine/aports/-/blob/master/scripts/mkimg.standard.sh 2022-10-27 03:14:25 yeah reading this rn 2022-10-27 03:14:26 but I don't believe it tweaks Grub modules (grub-install takes care of that itself) 2022-10-27 03:14:44 it has a grub-mkimage 2022-10-27 03:15:15 https://gitlab.alpinelinux.org/alpine/aports/-/blob/master/scripts/mkimg.base.sh#L321 2022-10-27 03:15:39 maybe efi_gop? 2022-10-27 03:16:23 so, right now 2 things: grub menu uses 100% cpu, kernel won't boot, and uses 400% cpu 2022-10-27 03:17:02 ah ok, didn't realise it used grub-mkimage directly 2022-10-27 03:17:35 efi_gop is one of the 2 Grub modules to do with EFI display, the other is efi_uga 2022-10-27 03:19:00 efi_uga is for "old" EFI graphics apparently 2022-10-27 03:24:48 https://gitlab.alpinelinux.org/alpine/aports/-/blob/master/scripts/mkimg.base.sh#L319 2022-10-27 03:25:06 that's the modules included in ISO ^^^ 2022-10-27 03:26:22 yes as it is covering all bases as to various devices that it might be installed on 2022-10-27 03:26:36 and profile virt is nothing more than a base : https://gitlab.alpinelinux.org/alpine/aports/-/blob/master/scripts/mkimg.standard.sh#L74 2022-10-27 03:27:14 ohh, initfs_features="$initfs_features dasd_mod qeth zfcp" 2022-10-27 03:27:33 oh no, that's for s390x 2022-10-27 03:30:34 uuidNuniq: I have to disappear, I'll catch up with you later to see if you've had a brainwave 2022-10-27 03:36:01 i boot it success!!! 2022-10-27 03:36:50 entered grub cmdline, and insmod all the modules mentioned in above code and ... load linux init, then run boot 2022-10-27 04:20:52 oh my god, spends hours and found that a simple 'insmod linux' on grub command line fixed the booting 2022-10-27 04:20:55 what the hell 2022-10-27 04:23:53 doesn't need kms at all. but grub.cfg needs : 'insmod linux' , that's all, and all. holy crap 2022-10-27 04:25:24 linux - This module is just like the "linux16" module except that this is the standard "linux" command. 2022-10-27 04:33:00 why the hell that mod for grub is not enabled, i wonder 2022-10-27 04:36:03 grub module 'linux' depends on 'boot fdt' , it is actually the 'boot' module was missing on default grub? 2022-10-27 04:43:07 so, final report : add GRUB_PRELOAD_MODULES="linux" to /etc/default/grub and regen the grub.cfg 2022-10-27 04:46:00 minimal : https://0x0.st/o3Vn.txt 2022-10-27 04:59:47 https://0x0.st/o3Ws.txt 2022-10-27 05:43:53 on archlinux arm, i did a lsmod in the grub command line, and 'linux' isn't loaded either by default. but why do i need to 'insmod linux' on alpine's grub to boot in this case? 2022-10-27 05:44:19 interesting 2022-10-27 05:52:15 finally, i can get rid of qemu and vmware! 2022-10-27 05:52:56 using apple's virtualization framework via tart only 2022-10-27 06:03:00 ikke: so, essentially you're saying that with vimdiff installed update-conf does offer a "merge" option? 2022-10-27 06:04:15 MarqueI: yeah, v opens vimdiff, then you can edit or use 'dp' command of vimdiff to sync the config and then :w 2022-10-27 06:04:33 ^ 2022-10-27 06:04:39 uuidNuniq: i guess i'll have to have a look into it. 2022-10-27 06:04:45 update-conf works like pacdiff in archlinux 2022-10-27 06:05:52 uuidNuniq: that's, unfortunately, a comparison i have nothing to say for. arch linux made it to the point where i wanted to install apache on a headless server and had to pull in a lot of X11 stuff for that. at that point the system in question went power-down hard and arch was never seen again. 2022-10-27 06:06:43 i mean, pacdiff uses vimdiff too :) 2022-10-27 06:08:24 or apknew in $(find "$ROOT/etc" -name '*.apk-new') 2022-10-27 06:08:45 yeah, i guess ;) 2022-10-27 06:09:00 so i made an alias to check myself first : 2022-10-27 06:09:07 apknew is aliased to `doas find /etc -type f -name "*.apk-new"' 2022-10-27 06:19:41 still don't understand why it has to 'insmod linux' to boot alpine via tart 2022-10-27 06:20:15 that leaves to devs to figure out :P 2022-10-27 06:21:17 well, i'll admit i'm not strictly a fan of modular kernels everywhereโ€ฆ 2022-10-27 06:21:35 that's a grub module 2022-10-27 06:21:53 i'm not strictly a fan of a modular boodloader either ;) 2022-10-27 06:21:55 I am trying to build a package in aports/testing/py3-wifi. While building, I am facing the error-Could not open file or uri for loading private key from key.TORE routines:ossl_store_get0_loader_int:unregistered scheme:crypto/store/store_register.c:237:scheme=file system library:file_open:No such file or directory:providers/implementations/storemgmt/file_store.c:267:calling stat(key) aports/testing/py3-wifi/pkg/.control.py3-wifi 2022-10-27 06:24:49 alpb: you'd better use a paste service to show us the whole log 2022-10-27 06:25:22 command 2>&1 | curl -F"file=@-" https://0x0.st 2022-10-27 06:26:19 command is your command to the abuild 2022-10-27 06:27:32 I am using abuild -r command to build. I haven't modified APKBUILD. 2022-10-27 06:28:04 so : abuild -r 2>&1 | curl -F"file=@-" https://0x0.st 2022-10-27 06:33:53 >>> zip: Building main/zip 3.0-r10 (using abuild 3.10.0_rc1-r2) started Thu, 27 Oct 2022 06:30:18 +0000 >>> zip: Checking sanity of /home/ramya/aports/main/zip/APKBUILD... >>> zip: Analyzing dependencies... >>> zip: Installing for build: build-base unzip WARNING: Ignoring /home/ramya/packages//main: No such file or directory (1/2) Installing unzip (6.0-r13) (2/2) Installing .makedepends-zip (20221027.063019) Executing busybox-1 2022-10-27 06:35:25 you'll get a URL from 0x0.st 2022-10-27 06:35:29 paste the URL here 2022-10-27 06:39:03 Sorry, I didn't get it. 2022-10-27 06:39:12 Where should I find that URL 2022-10-27 06:39:15 HTTP POST files here: curl -F'file=@yourfile.png' https://0x0.st You can also POST remote URLs: curl -F'url=http://example.com/image.jpg' https://0x0.st 2022-10-27 06:45:30 fvbnm: it returns a url for you 2022-10-27 06:46:14 file=@ was missing on your command 2022-10-27 06:50:22 should we give any file name in the place of @ 2022-10-27 06:51:34 The above command is to create a log file in that server, right! 2022-10-27 06:55:26 https://0x0.st/o3WV.txt 2022-10-27 06:59:46 https://0x0.st/o3WV.txt This is my log file. 2022-10-27 07:02:48 I wish I could run alpine on beaglebone, damn the official image is so utterly bloat 2022-10-27 07:03:15 I'll never understand why people do like these boards 2022-10-27 07:08:15 dwfreed: ^^ it wouldn't match your device's DHCP hostname 2022-10-27 07:08:33 does it mean I need to make an entry in hosts file explicitly ? 2022-10-27 07:09:14 vkrishn: if you want to override the response when querying a hostname that was handed out via DHCP, yes 2022-10-27 07:09:51 can't I just add * in hosts file ? 2022-10-27 07:09:58 note this only applies if you're also using dnsmasq's dhcp *server* 2022-10-27 07:10:05 markand: what is the SOC? 2022-10-27 07:10:10 ok 2022-10-27 07:11:32 uuidNuniq, AM335x 1GHz ARMยฎ Cortex-A8 2022-10-27 07:12:32 markand: you can always build a kernel for it 2022-10-27 07:14:58 markand: https://git.alpinelinux.org/aports/tree/main/linux-lts/lts.armv7.config#n52 2022-10-27 07:15:05 it is there :) supported i think 2022-10-27 07:18:15 yes that's not the issue, but installing a custom Linux on a BBB is harder than on a pi, you have to deal with uboot and many other bootloader files and partitions 2022-10-27 07:20:14 uboot is a separate thing. it is just a bootloader. you can borrow it from beaglebone official 2022-10-27 07:22:21 you just need to figure out how the current linux was booted by uboot... e.g. what are the conf needed 2022-10-27 07:22:42 then, the rest are easy. 2022-10-27 07:37:09 #postmarketos 2022-10-27 13:50:14 hello, how do you remove a user from a group with busybox ? 2022-10-27 13:56:29 does busybox not have usermod? 2022-10-27 13:57:14 apparently not 2022-10-27 13:57:34 delgroup user group 2022-10-27 13:57:42 linex[m]: use delgroup 2022-10-27 13:57:55 https://busybox.net/downloads/BusyBox.html 2022-10-27 14:32:02 busybox delgroup --help 2022-10-27 14:32:08 or -h 2022-10-27 14:32:12 whatever works. 2022-10-27 15:20:37 thanks 2022-10-27 17:14:40 is theres a package that provides execinfo.h? 2022-10-27 17:15:03 https://pkgs.alpinelinux.org/contents?file=execinfo.h&path=&name=&branch=edge 2022-10-27 17:15:22 nothing directgly 2022-10-27 17:15:30 only zig 2022-10-27 17:15:46 hm 2022-10-27 17:15:53 how was this code compiling before 2022-10-27 17:17:26 There used to be libexec apparently 2022-10-27 17:17:32 ah yeah that one 2022-10-27 17:17:40 it was removed recently right 2022-10-27 17:18:07 50795a14dee639ce2dcc836e2b2baca9bad4a1b1 2022-10-27 17:18:30 https://gitlab.alpinelinux.org/alpine/aports/-/commit/50795a14dee639ce2dcc836e2b2baca9bad4a1b1 2022-10-27 17:19:08 ACTION shakes fist at psykose 2022-10-27 17:19:24 time to comment large batches of code 2022-10-27 17:28:21 easier with an early return and a few comments 2022-10-27 17:28:29 ACTION shakes fist at bl4ckb0ne 2022-10-27 17:34:10 dwfreed: psykose: dnsmsq 2.87 works as aspected, thanks 2022-10-27 17:34:28 seems v2.86 was the problem 2022-10-27 17:35:22 if not a load, maybe backport it to AL-3.15 2022-10-27 17:36:08 as expected ^ 2022-10-27 18:21:33 Does the APK cache and the APK overlay file need to be on the boot media device when using Alpine on a Raspberry Pi? 2022-10-27 18:22:07 Assuming I'm not using the apkovl parameter 2022-10-28 05:15:05 does API eg make it easy to show the content of a tweet-url in the terminal? 2022-10-28 05:16:31 ? 2022-10-28 05:17:43 ? 2022-10-28 05:18:04 'tweet-print http:twitter.com/...' > shows content of tweet 2022-10-28 05:19:27 I still have no idea what you're asking about 2022-10-28 05:21:05 id want a easy way to view tweets in the terminal 2022-10-28 05:21:37 and wonder why it doesnt exist, i guess APIs would make it easy (but im no programmer) 2022-10-28 05:22:17 I'm pretty sure there is some kind of twitter cli client made in go/rust/python 2022-10-28 05:31:08 Hi, I am trying to build kernel. When I am trying make command after setting up menuconfig, I am facing this error. 2022-10-28 05:31:11 make[1]: *** No rule to make target 'arch/x86/entry/syscalls/syscall_32.tbl', needed by 'arch/x86/include/generated/uapi/asm/unistd_32.h'. Stop. make: *** [arch/x86/Makefile:217: archheaders] Error 2 2022-10-28 05:42:33 I am using linux-headers-5.15.74.0-lts 2022-10-28 06:59:44 Hlo I was trying to build kernel but facing error while building it 2022-10-28 11:01:28 anyone running edge as a desktop os? how usable is it? I like alpine but a few of the tools I want are in testing. Otherwise I would no consider using edge. Is there a more advisable way to do this? 2022-10-28 11:24:13 elagost: why not use 2 repos at the same time? 2022-10-28 11:25:24 For packages from "testing" 2022-10-28 11:31:37 I suppose I could - wiki mentions stable+testing is a bad idea though. 2022-10-28 11:33:03 elagost: yes mixing repos from a release and Edge is not supported - packages in Edge may be built against newer library versions than those in a release 2022-10-28 11:34:27 True. I have run into that issue before - can't just expect to add a package in testing and have it work. 2022-10-28 11:36:28 probably best to run a container or chroot I guess. 2022-10-28 11:36:45 Or build package by hand? 2022-10-28 11:36:52 :) 2022-10-28 11:37:23 Yep, I have done that before too. alpines build tools are really friendly. 2022-10-28 11:37:30 yeah 2022-10-28 11:37:54 different topic - is there a way to specify a different size for /boot in the installer, other than custom partitioning? setup-disk accepts a parameter for the size of swap it looks like. 2022-10-28 11:40:55 I couldn't find how to do this when I installed the system, so I just partitioned and mounted them in one place. And then alpine-setup... 2022-10-28 11:41:02 elagost: BOOT_SIZE= setup-disk ? 2022-10-28 11:47:26 minimal: perfect, it works. Tested in a vm. I can just run `BOOT_SIZE=512 SWAP_SIZE=512 setup-alpine` and those values get respected. 2022-10-28 11:51:15 the default 100M is a little small for both lts and stable kernels :) 2022-10-28 11:51:54 elagost: I have an Issue open relating to the boot sizes used by setup-disk: https://gitlab.alpinelinux.org/alpine/alpine-conf/-/issues/10536 2022-10-28 11:52:56 elagost: lts and stable kernels? you mean linux-lts and linux-edge? 2022-10-28 11:53:32 minimal: correct, I forgot it was called -edge for a second. 2022-10-28 11:53:50 I'm not using uefi so I didn't notice, but 34MB is very tiny! 2022-10-28 14:36:02 Should an APK overlay file be inside the same partition as the one containing initramfs or can it be in a different partition? 2022-10-28 14:36:43 I had an Alpine installation on a RPI4 before and lbu saved an overlay in a different partition and it was applied at boot. 2022-10-28 14:37:19 I'm trying to install Alpine on another RPI and if I keep the overlay in a different partition (on the same device), the overlay isn't detected. 2022-10-28 14:38:57 i think the idea is that it is supposed to be detected 2022-10-28 14:39:18 but it is apparently buggy 2022-10-28 14:41:08 ncopa: the behavior seems inconsistent to me, the overlay is applied when saved by lbu in a different partition but not applied when saved manually in a different partition when installing Alpine 2022-10-28 14:41:31 it's only applied when the overlay is in the same partition where initramfs is kept 2022-10-28 14:41:46 sounds like a bug 2022-10-28 14:42:16 ncopa: not sure if it's a bug in nlplug-findfs or initramfs-init.in 2022-10-28 14:42:27 i'd say initramfs-init.in 2022-10-28 14:42:39 i saw a MR for a fix for somethign related 2022-10-28 14:43:00 its a fix for respecting the /etc/fstab for the partition where apkovl is 2022-10-28 14:44:31 ncopa: this one? 2022-10-28 14:44:32 https://github.com/alpinelinux/mkinitfs/pull/65 2022-10-28 14:45:41 no 2022-10-28 14:45:43 let me find it 2022-10-28 14:46:22 https://gitlab.alpinelinux.org/alpine/mkinitfs/-/merge_requests/111 2022-10-28 16:46:37 How do I get zswap to use lz4? Already put lz4 in /etc/modules but its still falling back to lzo 2022-10-28 16:51:46 maybe add lz4 and lz4_compress to initramfs? 2022-10-28 16:52:14 by adding them to /etc/mkinitfs/fatures.d/something.modules 2022-10-28 17:08:58 zswap.compressor=lz4 with the modules present i'd guess 2022-10-28 17:53:26 Anyone experiences routing issues in linux where the first times routes are added, they appear are throw routes (throw 172.16.0.254 proto static) preventing traffic reaching that route? I have to remove those routes and then manually add the expected routes again 2022-10-28 18:56:43 ncopa: that issue where apkvol wasn't being detected in a separate partition? it happens when that partition is xfs but not when it's ext4 2022-10-28 18:57:06 https://gitlab.alpinelinux.org/alpine/aports/-/issues/11589 2022-10-28 19:00:22 Why is XFS not included in ARM images? Space constraints? 2022-10-28 21:28:09 hello 2022-10-28 21:28:32 is it possible for busybox vi to display utf8 characters? 2022-10-28 23:40:51 are there any tutorial for alpine-virt-x.y.z-aarch64.iso? 2022-10-28 23:43:37 I currently extract the vmlinuz-virt and initramfs-virt out of the .iso file to launch a VM, also I need specify the .iso as a virtual disk and I can see login prompt. I am not sure if there are simpler ways to use the .iso directly on ARM virtual machine? 2022-10-29 02:16:59 the default /boot created by setup-alpine is 32MB, which is not big enough to upgrade linux-lts-5.15.74-r0 to 5.15.75-r0 2022-10-29 02:17:07 rather awkward 2022-10-29 02:38:10 so remove the kernel, apply update ;) 2022-10-29 02:43:08 it is a bug indeed https://gitlab.alpinelinux.org/alpine/alpine-conf/-/issues/10536 2022-10-29 03:24:00 Apachez: I did move the old /boot contents aside, but it was a bit scary :) 2022-10-29 03:48:08 fear leads to happiness, something something 2022-10-29 08:53:34 hello 2022-10-29 08:54:54 how to I set caps as control in the console? do I need the kbd pkg or is it possible with the loadkmap service? 2022-10-29 09:28:27 *how do I 2022-10-29 10:24:26 I had to disconnect, but asking how to get capslock as control in console, i.e. whether kbd pkg is required or if this can be achieved with just the loadkmap service? 2022-10-29 10:25:26 loadkmap requires a compressed binary keymap but has no documentation 2022-10-29 10:25:46 https://gitlab.com/interception/linux/plugins/dual-function-keys 2022-10-29 10:27:16 lopid: sorry that's not really what I'm after 2022-10-29 10:27:48 sounded like it was, but ok 2022-10-29 10:28:03 i don't know how to do it with what the distro already comes with 2022-10-29 10:29:09 that is an interesting package, but I am not looking for a tap-vs-hold solution, I just wanna 'keycode 58 = control' 2022-10-29 10:29:50 it supports hold 2022-10-29 10:30:09 which is easy with loadkeys(1) but I see loadkmap is part of busybox so perhaps kbd package is unnecessary 2022-10-29 10:30:14 yep that was clear 2022-10-29 10:30:32 then i can't help, sorry 2022-10-29 10:30:39 it's cool, thanks! 2022-10-29 10:31:11 maybe it's not distro specific, try a more general channel? 2022-10-29 10:32:17 it's specific to the loadkmap service, which is part of busybox 2022-10-29 10:32:32 which isn't specific to alpine 2022-10-29 10:32:46 I guess my question could be better phrased as: how does one create a binary keymap suitable for loadkmap 2022-10-29 10:32:50 ..? 2022-10-29 10:33:38 or even, where might one find the documentation for creation of binary keymaps? 2022-10-29 11:25:57 aha! okay so the answer to above is: kinda. in order to create a binary keymap, you need loadkeys(1) (which is park on kbd package) but once a keymap is created, you can `loadkeys -b my/new/key.map.asc > /my/new/key.map && gzip my/new/key.map` then add that to /etc/conf.d/loadkmap and `rc-update add loadkmap; apk del kbd` 2022-10-29 11:27:02 kinda wish linux would use a more BSD approach to man pages, since finding this info was a catch-22 of reading loadkeys.1, which of course is not mentioned in the doc to loadkmap 2022-10-29 13:10:20 elly: I raised an issue on this a while ago: https://gitlab.alpinelinux.org/alpine/alpine-conf/-/issues/10536 2022-10-29 13:58:05 minimal: yup, hopefully that is merged soon :) but for my system I think I will need to resize it manually anyway 2022-10-29 14:52:57 hello 2022-10-29 15:31:39 using ash 'read' i have to escape prepending blanks to make it register(like '\ \ indented line')? read var;echo "var:$var" 2022-10-29 15:33:00 'read' splits input by $IFS 2022-10-29 15:33:10 if you want exact input, `IFS= read -r var` 2022-10-29 15:34:05 grawity: interesting! does that unset IFS and run the command? 2022-10-29 15:34:16 yes 2022-10-29 15:34:31 well, sets it to empty string, but that's good enough 2022-10-29 15:35:03 very good! 2022-10-29 15:35:06 splitting-by-IFS becomes useful when you do e.g. `read col1 col2 col3 therest` 2022-10-29 15:35:19 whereas in your case you probably want to disable it 2022-10-29 15:35:33 don't forget the -r to prevent `read` from interpreting other kinds of backslash escapes 2022-10-29 15:36:16 yes i knew about -r but it didnt ork without IFS p 2022-10-29 17:04:34 Does anyone know how I can mount an nfs export from openbsd to linux? 2022-10-29 17:06:50 the same way you mount nfs shares? 2022-10-29 17:52:15 Just installed Alpine in data disk mode but when restarting all installed packages aren't there (running `apk add` fixes it, so cache is ok). Do I have to make a service on startup to install them every time or is there an expected method? 2022-10-29 18:25:51 truita: Is /etc/apk/cache symlinked to your APK cache location? 2022-10-29 18:44:14 if you've configured one, yes 2022-10-29 19:06:44 you do have to install them every time pretty sure 2022-10-29 19:42:02 I'm trying to make sshd listen the private IP assigned to eth0 but there's a warning inside /etc/init.d/sshd that recommends using rc_need="net.eth0" in /etc/conf.d/sshd. When I add this parameter, the sshd service fails to restart. 2022-10-29 19:42:24 you can ignore the warning 2022-10-29 19:43:16 psykose: oh, ok, so I can simply use the private IP assigned to eth0 in sshd_config? The device in question is a Raspberry Pi 4 2022-10-29 19:43:54 probably yeah 2022-10-29 19:47:28 psykose: rebooting the rpi made sshd start as expected, thanks 2022-10-29 20:03:26 one of the issues when using a private IP for sshd is that if I restart my router and alpine at the same time, Alpine probably doesn't get an IP in time and the sshd service fails. 2022-10-29 20:03:58 I think this happened even when I had 0.0.0.0 as the ListenAddress for sshd in Alpine 2022-10-29 20:05:02 Do I need to modify the init scripts for Alpine to fix this? 2022-10-29 20:05:16 s/for Alpine/for sshd/ 2022-10-29 20:05:17 ayushnix meant to say: Do I need to modify the init scripts for sshd to fix this? 2022-10-29 20:07:36 I was facing the same issue when I using the adguard service in Alpine 2022-10-29 20:07:40 sshd[2179]: error: Bind to port 22 on 192.168.1.3 failed: Address not available 2022-10-29 20:07:59 Hi, by default it seems output of stuff like `git diff` is not colorized, any way to achieve colorization of that output? 2022-10-29 20:10:35 git diff | less -R 2022-10-29 20:12:10 git diff --color=always 2022-10-29 20:12:10 @panekj: Still no color :( 2022-10-29 20:12:34 @psykose: Same. 2022-10-29 20:12:45 what are you running it in 2022-10-29 20:12:55 I do have color in my Bash prompt, but not in output it seems. 2022-10-29 20:12:57 inb4 TERM=linux 2022-10-29 20:13:18 if you have diff long enough, I presume your pager strips colours 2022-10-29 20:13:36 TERM=screen-256color 2022-10-29 20:14:36 diff was like 2 lines and resulted in no colors, then I added some stuff to my git config forcing colors and now it probably goes through a pager, but it's the default pager of Alpine (maybe BusyBox?) 2022-10-29 20:14:42 @psykose: A VM 2022-10-29 20:16:54 bb less is incapable of colours 2022-10-29 20:17:15 command -v less 2022-10-29 20:19:37 Ah. Could be the culprit then. 2022-10-29 20:19:41 `lrwxrwxrwx 1 root root 12 Jun 24 2021 /usr/bin/less -> /bin/busybox*` 2022-10-29 20:21:55 One `apk add less` later and everything is all and well, thanks :) 2022-10-29 21:06:04 ayushnix: it is 2022-10-29 21:06:50 I also set it up with setup-apkcache just in case 2022-10-29 21:08:00 for now I have `apk add` in my crontab @reboot 2022-10-29 21:08:27 but I guess it won't be enough if I later have to start services at boot 2022-10-29 21:09:41 Stupid question, how do I rotate the tty on Alpine? echo 1 > /sys/class/graphics/fbcon/rotate_all doesn't do anything. 2022-10-29 21:15:34 hm... 2022-10-29 21:15:36 # CONFIG_FRAMEBUFFER_CONSOLE_ROTATION is not set 2022-10-29 21:15:42 is this why? but why does the sysfs node exist then? 2022-10-29 21:16:26 can't tell if its default is n or y though 2022-10-29 21:20:58 well, I guess that's my answer - apparently Alpine simply doesn't support FB rotation unless I build my own kernel - fine then 2022-10-29 21:21:28 don't really care about this when done installing, but setup-alpine would work much better if rotated the right way round 2022-10-29 21:21:42 "not set" always means =n 2022-10-29 21:22:12 regardless of the default 2022-10-30 00:19:11 xen-dom0 image hangs on setting system clock to hardware clock 2022-10-30 00:58:45 nope. 2022-10-30 00:58:51 the xen-dom0 image doesn't work. 2022-10-30 00:59:06 you go to apk add xen-hypervisor, GRUB errors 2022-10-30 01:14:18 xen needs qemu to install and it's not even on the dom0 image. wtf? 2022-10-30 01:26:59 anyway, is there a concrete reason why Alpine's default kernel has CONFIG_FRAMEBUFFER_CONSOLE_ROTATION turned off? 2022-10-30 01:27:11 Seems like many small tablets (that are actually Windows tablets) need that 2022-10-30 01:27:37 probably no 2022-10-30 01:27:44 at least on the "extended" install image that'd be quite nice to have 2022-10-30 01:28:00 just wondering if kernel size is the issue, like, on some devices the bootloader can barely load the kernel right now 2022-10-30 01:28:07 I do remember some Chromebooks have a 32MB limit 2022-10-30 01:28:21 open issue on gitlab 2022-10-30 01:28:34 gitlab.a.o 2022-10-30 01:29:31 sure, will do 2022-10-30 01:29:45 was just first wondering if there was a strong reason against, as virtually every other distro has it enabled 2022-10-30 01:33:23 https://gitlab.alpinelinux.org/alpine/aports/-/issues/14311 2022-10-30 08:54:36 tmpfs mounted on /tmp is 4G but it only uses what it needs of ram? also is there a application to see whats using your RAM? 2022-10-30 08:56:07 doing some ffmpeg work on a slow computer, it basically freezes for longer periods :p 2022-10-30 08:58:18 yes. top 2022-10-30 09:03:56 kinda seems cpu/mem is not used that much (less than 30%) still it slows down user experience 2022-10-30 09:04:08 i/o? 2022-10-30 09:04:21 depends what you're doing 2022-10-30 09:05:10 lopid, oh, its the hdd? yeah rpi4 2022-10-30 09:05:37 spinning disk? more likely to be 2022-10-30 09:07:08 its a samsung fit usb stick, i also use a SD card, but i dont think thats involved in this 2022-10-30 09:07:37 the system and the file im working on is on the samsung fit 2022-10-30 09:08:15 usb2? 2022-10-30 09:09:24 lopid: hmm should be usb3 2022-10-30 09:10:53 maybe would be better to work on files on the SD? 2022-10-30 09:12:15 maybe 2022-10-30 09:24:28 seems to be going better from the SD indeed :] 2022-10-30 09:24:35 how do i check if i am using usb3? 2022-10-30 11:07:22 the port will probably be blue 2022-10-30 11:09:27 lsusb also shows it 2022-10-30 11:11:57 does this mean its usb3? /: Bus 02.Port 1: Dev 1, Class=root_hub, Driver=xhci_hcd/4p, 5000M 2022-10-30 11:12:19 |__ Port 2: Dev 2, If 0, Class=Mass Storage, Driver=usb-storage, 5000M 2022-10-30 11:12:26 if 5000M is the speed then i think it is 2022-10-30 11:13:08 yeah: https://en.wikipedia.org/wiki/USB_3.0 2022-10-30 11:13:08 [wikipedia] USB 3.0 | "USB 3.0, released in November 2008, is the third major version of the Universal Serial Bus (USB) standard for interfacing computers and electronic devices. Among other improvements, USB 3.0 adds the new transfer rate referred to as SuperSpeed USB (SS) that can transfer data at up to 5 Gbit/s (500 MB/s after encoding overhead), which is about 10 times faster than Hi-Speed (maximum for USB 2.0 [โ€ฆ]" 2022-10-30 11:13:30 cool :] 2022-10-30 11:13:43 (i guess busybox lsusb doesnt show it tho) 2022-10-30 11:13:50 had to install haha 2022-10-30 11:14:16 probably just reading from /sys anyway 2022-10-30 11:16:08 yeah it opens files under /sys/bus/usb/devices 2022-10-30 11:17:15 hmm there is a usb1 and usb2 dir in /sys/bus/usb/devices/, no usb3... 2022-10-30 11:17:39 i think they're ports 2022-10-30 11:17:46 not versions 2022-10-30 11:18:00 would be nice with a script with overview that showed info from those kinds of files 2022-10-30 11:18:17 lsusb, for example? 2022-10-30 11:19:02 well why not a shell script with overview of more info 2022-10-30 11:19:40 lsusb -v. why a shell script? 2022-10-30 11:19:43 in general i mean(not just usb), im not too familiar with those kinds of files tho 2022-10-30 11:19:59 lsusb, lspci, lscpuโ€ฆ 2022-10-30 11:20:18 there are any number of hardware info tools 2022-10-30 11:20:45 yeah, why not one shell script 2022-10-30 11:21:03 my ls output: ls lsattr lsmod lsof lsusb 2022-10-30 11:21:41 why a shell script? 2022-10-30 11:22:15 hwinfo is just one of those tools 2022-10-30 11:22:36 i guess those binaries are in "bloated" packages 2022-10-30 11:23:00 im a scripter tho, so maybe thats why i want scripts :] 2022-10-30 11:23:19 easy to read/modify 2022-10-30 11:23:49 maybe on file that can give overview of available info too .. 2022-10-30 11:24:44 mv a 1GB file from SD to usb basically halts my system for minutes... 2022-10-30 12:13:50 "the same way you mount nfs..." <- Ok. I went through the steps in the openbsd faq about nfs (setting up to NFS server) and then started the client with "rc-service nfsclient start" and tried to mount it with "mount x.x.x.x:/foo/bar /bar". Then it says: mount.nfs: access denied by server while mounting x.x.x.x:/foo/bar 2022-10-30 12:34:50 the nfs server likely says that you're not authorized in its logs 2022-10-30 13:30:28 Hello there. I have recent Alpine Linux installed on a Raspberry pi 3b (sys mode). /etc/fstab has this 'UUID=... / ext4 rw,realtime 0 1' line. But boot process halts, informing about ro filesystem. When i log in as a root, and issue 'mount | grep mmc' the reply is '/dev/mmcblk0p2 on / type ext4 (ro,relatime)'. 2022-10-30 13:33:00 Then i do 'mount -o remount,rw /' and start all the necessary sshd and whatnot. Could you, please give me some hint? 2022-10-30 13:33:36 brown_fox: what are the contents of cmdline.txt in the boot partition? 2022-10-30 13:33:43 Hi, when I do `startx`, I get an error. I am hopelessly bad at troubleshooting Xorg issues (since on Ubuntu it "just works"), are there otherwise any peculiarities in this log I should explore? https://0x0.st/oY1s.log.old 2022-10-30 13:33:56 which version of Alpine is this? is it armv7 or aarch64? 2022-10-30 13:34:08 brown_fox: ^^^ 2022-10-30 13:34:10 Guest4174: Xorg.log is a good starting point 2022-10-30 13:34:43 minimal: aarch64. 2022-10-30 13:34:55 Guest4174: open /dev/dri/card0: Permission denied 2022-10-30 13:35:06 brown_fox: which release? 2022-10-30 13:35:11 You need to be part of the `video` group 2022-10-30 13:35:42 minimal: alpine-rpi-3.16.2-aarch64.tar.gz 2022-10-30 13:35:57 brown_fox: what are the contents of cmdline.txt in the boot partition? 2022-10-30 13:37:02 minimal: root=UUID=04d6f356-b0a5-4f4f-8f2a-3172147683ac modules=sd-mod,usb-storage,ext4 quiet rootfstype=ext4 2022-10-30 13:37:49 brown_fox: and that UUID matches the one in /etc/fstab? 2022-10-30 13:38:14 @ikke: Yeah I was already searching for that but couldn't find much to go on. It's running in a VM, so perhaps that's an issue. I tried doing `doas startx` but that also errored out. 2022-10-30 13:38:47 minimal: sure 2022-10-30 13:38:59 alpine:/boot$ cat /etc/fstab 2022-10-30 13:38:59 UUID=04d6f356-b0a5-4f4f-8f2a-3172147683ac / ext4 rw,relatime 0 1 2022-10-30 13:40:43 Guest4174: What hypervisor? 2022-10-30 13:41:07 @ikke: KVM / Qemu using virt-manager. 2022-10-30 13:41:27 Running on host Ubuntu 18.04 2022-10-30 13:41:57 brown_fox: no other errors during boot? Do you see "Remounting root filesystem read/write..." during boot? 2022-10-30 13:44:43 minimal: let me check this out. 2022-10-30 13:44:59 has anyone noticed odd behavior with gnome terminal tab shortcuts? ctrl-pageup doesn't work (does nothing) and editing the shortcut to use ctrl-pageup then inserts a tilde 2022-10-30 13:44:59 minimal: may be dmesg has more? 2022-10-30 13:45:05 Guest4174: Did you install any xf86-video-* package? 2022-10-30 13:46:50 brown_fox: edit /etc/rc.conf to set rc_logger="YES" and then after boot look at /var/log/rc.log 2022-10-30 13:48:00 @ikke: `apk -vv info | grep -i xf86` results in `libxxf86vm-1.1.4-r2` and its docs. 2022-10-30 13:48:21 Should I install the xf86-video packages? 2022-10-30 13:49:09 try xf86-video-vesa or xf86-video-qxl 2022-10-30 13:49:56 I'll try installing them both just in case :) 2022-10-30 13:50:36 and probably xf86-input-libinput as well 2022-10-30 13:50:59 (And make sure you are memer of the video and input group 2022-10-30 13:54:41 minimal: will this do? i mean the file system is read only, /var is root descendant. Or is it going to log into ram? 2022-10-30 13:55:43 @ikke: I will meme the hell out of them if it means that this will work, but sadly no :( 2022-10-30 13:55:51 This is the new Xorg log: https://0x0.st/oY1K.0.log 2022-10-30 13:56:33 Ok, an experiment: Could this be because of the i3wm I installed via Nix? Let's find out. 2022-10-30 13:58:09 brown_fox: ok, then do "rc-status -a | less" and look for the "root" entry to see its status (i.e. "started") 2022-10-30 14:00:37 Guest4174: Can you also try xf86-video-modesetting? 2022-10-30 14:00:51 Ok. 2022-10-30 14:01:57 I think in my local vm, it uses modesetting 2022-10-30 14:02:48 @ikke: Nope: https://0x0.st/oY1q.0.log 2022-10-30 14:04:41 Do you know what graphics card is being used? 2022-10-30 14:04:50 I start it manually with qemu, and I just pass -vga 2022-10-30 14:05:57 minimal: alpine:/var/log$ rc-status -a | grep root 2022-10-30 14:05:57 root [ started ] 2022-10-30 14:06:30 minimal: but sorry i've already done mount -o remount,rw / 2022-10-30 14:06:33 so /etc/init.d/root is the script that switches / from "ro" to "rw" 2022-10-30 14:06:54 you're command is unrelated as to whether that init.d ran during boot 2022-10-30 14:07:04 s/you're/your/ 2022-10-30 14:07:04 minimal meant to say: your command is unrelated as to whether that init.d ran during boot 2022-10-30 14:08:11 brown_fox: so did you try adding rc_logger="YES" ? 2022-10-30 14:08:28 minimal: yes i did. 2022-10-30 14:08:35 @ikke: I'm just using my integrated GPU. 2022-10-30 14:08:42 brown_fox: and was /var/log/rc.log created? 2022-10-30 14:09:00 Guest4174: I mean for the VM 2022-10-30 14:09:07 ACTION uploaded an image: (17KiB) < https://matrix.org/_matrix/media/r0/download/matrix.org/PXBgycuCUXMVOtlXbKENjIha/image.png > 2022-10-30 14:09:08 In the virt-manager settings I'm using a Spice server it seems? 2022-10-30 14:11:19 minimal: yes but the last record is 'rc shutdown logging stopped at Sun Oct 30 16:59:00 2022' so it wrote this line while halting. 2022-10-30 14:11:43 I don't think it's i3wm because installing it via apk didn't work as well: https://0x0.st/oY1a.0.log 2022-10-30 14:12:33 brown_fox: and you rebooted after adding this line? and you checked all then entries in the created rc.log file to see what happened during boot? 2022-10-30 14:12:52 Guest4174: Do you happen to have a /etc/X11/xorg.conf file? 2022-10-30 14:13:16 brown_fox: or do you mean nothing was written to the file since the shutdown portion of the "reboot"? 2022-10-30 14:14:50 minimal: the question is, is it possible to write something into /var/log/rc.log while the root file system is read only? 2022-10-30 14:15:12 minimal: indeed. 2022-10-30 14:16:32 brown_fox: no, but it wasn't clear if your problem was happening BEFORE /etc/init.d/root would be run - exactly what errors do you see during boot? is the fsck running ok? 2022-10-30 14:18:33 @ikke: These are the contents of my `/etc/X11/` - https://0x0.st/oY1M.txt 2022-10-30 14:19:02 minimal: nope. * Filesystems couldn't be fixed * rc: Aborting! * fsck: caught SIGTERM, aborting (this messages repeat twice). 2022-10-30 14:19:37 brown_fox: ok, so there's a problem with your roofs, so why aren't you fixing that? 2022-10-30 14:19:41 @ikke: But I have a local `~/.xinitrc` which is usually used since I'm launching `startx` as a normal user. 2022-10-30 14:20:33 yes, that's expected 2022-10-30 14:22:10 minimal: oh i see! I'm going to give it a try. Thank you. 2022-10-30 14:35:11 minimal: is it okay to fsck the sd on my laptop? 2022-10-30 14:37:31 brown_fox: yes, that's how I would no it - just make sure the filesystems are not mounted when you do it (you should fsck both the RPI boot fs and the rootfs) 2022-10-30 14:37:38 s/no /do / 2022-10-30 14:37:38 minimal meant to say: brown_fox: yes, that's how I would do it - just make sure the filesystems are not mounted when you do it (you should fsck both the RPI boot fs and the rootfs) 2022-10-30 14:51:50 minimal: both partitions are clean, but nothing's changed: all the errors persist. 2022-10-30 14:52:57 brown_fox: well the "Filesystems couldn't be fixed" output you mention seems to indicate a problem 2022-10-30 14:55:08 brown_fox: have you any other filesystems apart from boot and / ? 2022-10-30 14:56:10 minimal: alpine:~$ lsblk 2022-10-30 14:56:10 NAME MAJ:MIN RM SIZE RO TYPE MOUNTPOINTS 2022-10-30 14:56:10 โ”œโ”€mmcblk0p1 179:1 0 512M 0 part /media/mmcblk0p1 2022-10-30 14:56:10 โ””โ”€mmcblk0p2 179:2 0 14.1G 0 part / 2022-10-30 14:56:10 mmcblk0 179:0 0 14.6G 0 disk 2022-10-30 14:57:11 brown_fox: no other filesystems in /etc/fstab being mounted from another device? 2022-10-30 14:58:40 minimal: do you mean some other usb drive? The sd is the only block device that is mounted to my rpi. 2022-10-30 15:00:10 brown_fox: e2fsprogs and dosfstools packages are installed? 2022-10-30 15:01:04 minimal: may be i should get rid of this 'quiet' option in cmdline.txt and i'll have more information? 2022-10-30 15:01:53 brown_fox: yupe. can you confirm both those packages are installed? 2022-10-30 15:04:00 brown_fox: what other entries are in your /etc/fstab? 2022-10-30 15:04:46 minimal: alpine:~$ apk list | egrep "e2fsprogs|dosfstools" 2022-10-30 15:04:46 e2fsprogs-static-1.46.5-r0 aarch64 {e2fsprogs} (GPL-2.0-or-later AND LGPL-2.0-or-later AND BSD-3-Clause AND MIT) 2022-10-30 15:04:46 e2fsprogs-1.46.5-r0 aarch64 {e2fsprogs} (GPL-2.0-or-later AND LGPL-2.0-or-later AND BSD-3-Clause AND MIT) 2022-10-30 15:04:46 libcom_err-1.46.5-r0 aarch64 {e2fsprogs} (GPL-2.0-or-later AND LGPL-2.0-or-later AND BSD-3-Clause AND MIT) [installed] 2022-10-30 15:04:48 e2fsprogs-libs-1.46.5-r0 aarch64 {e2fsprogs} (GPL-2.0-or-later AND LGPL-2.0-or-later AND BSD-3-Clause AND MIT) [installed] 2022-10-30 15:04:51 dosfstools-doc-4.2-r1 aarch64 {dosfstools} (GPL-3.0-or-later) 2022-10-30 15:04:53 dosfstools-4.2-r1 aarch64 {dosfstools} (GPL-3.0-or-later) [installed] 2022-10-30 15:04:57 e2fsprogs-extra-1.46.5-r0 aarch64 {e2fsprogs} (GPL-2.0-or-later AND LGPL-2.0-or-later AND BSD-3-Clause AND MIT) 2022-10-30 15:05:00 e2fsprogs-dev-1.46.5-r0 aarch64 {e2fsprogs} (GPL-2.0-or-later AND LGPL-2.0-or-later AND BSD-3-Clause AND MIT) 2022-10-30 15:05:03 e2fsprogs-doc-1.46.5-r0 aarch64 {e2fsprogs} (GPL-2.0-or-later AND LGPL-2.0-or-later AND BSD-3-Clause AND MIT) 2022-10-30 15:05:20 minimal: alpine:~$ cat /etc/fstab 2022-10-30 15:05:20 UUID=04d6f356-b0a5-4f4f-8f2a-3172147683ac / ext4 rw,relatime 0 1 2022-10-30 15:05:20 # /dev/usbdisk /media/usb vfat noauto 0 0 2022-10-30 15:05:20 # /dev/cdrom /media/cdrom iso9660 noauto,ro 0 0 2022-10-30 15:05:22 tmpfs /tmp tmpfs nosuid,nodev 0 0 2022-10-30 15:07:33 minimal: oops, last paste is incomplete. sorry 2022-10-30 15:08:29 brown_fox: just wondering if fstab has entry for the boot partition, so whether the fsck is only run against root or also against boot partition 2022-10-30 15:08:51 minimal: sure it has. 2022-10-30 15:09:00 alpine:~$ grep -v "^#" /etc/fstab 2022-10-30 15:09:00 UUID=04d6f356-b0a5-4f4f-8f2a-3172147683ac / ext4 rw,relatime 0 1 2022-10-30 15:09:00 tmpfs /tmp tmpfs nosuid,nodev 0 0 2022-10-30 15:09:43 It's weird. The boot line vanishes away while pasting. Sorry for the flood guys. 2022-10-30 15:10:17 brown_fox, it's the leading "/" being interpreted as an IRC command 2022-10-30 15:10:32 I see. Slash must be escaped. 2022-10-30 15:10:52 /dev/mmcblk0p1 /media/mmcblk0p1 vfat defaults 0 0 2022-10-30 15:11:07 Long time since using irc. 2022-10-30 15:11:31 Anyone on Edge also experiencing Gnome Keyring daemon pegging one thread constantly after login? I have it across both my Alpine Edge machines 2022-10-30 15:12:17 brown_fox: so it's not to do with boot part/fs as the "0 0" means it is not fsck'ed 2022-10-30 15:12:34 brown_fox: so its am fsck issue with rootfs then 2022-10-30 15:14:21 minimal: should it be 0 1 then? 2022-10-30 15:16:40 minimal: but wiki says and i quote: 2022-10-30 15:16:45 echo "/dev/mmcblk0p1 /media/mmcblk0p1 vfat defaults 0 0" >> etc/fstab 2022-10-30 15:16:45 sed -i '/floppy/d' etc/fstab 2022-10-30 15:16:45 sed -i '/cdrom/d' etc/fstab # Of course, you don't have any cdrom or floppy on the Raspberry Pi 2022-10-30 15:16:45 cd /media/mmcblk0p1 2022-10-30 15:16:51 end quote 2022-10-30 15:19:14 brown_fox: I didn't say it should be anything, I just pointed out that "0 0" means the boot part is not fsck'd and so is not your problem (if you want it fsck'd then you would change it to "0 2") 2022-10-30 15:19:18 minimal: okay. let's postpone it. I'm going to remove this 'quiet' parameter and i guess i'll have more. Thank you for your time anyway. Take care. I think i'll manage to launch my pi on alpine. 2022-10-30 15:20:19 or visa-versa :-) 2022-10-30 15:25:30 minimal: last guess. Seeing this 'after clock' dependency in /etc/init.d/root and i've noticed this "* Clock skew detected with '(null)'" may be this is the line i should focus on? 2022-10-30 15:29:44 nope, that's a typical line as you are using a Raspberry which has no RTC 2022-10-30 15:30:38 minimal: okay. Thank you very much again. 2022-10-30 15:31:05 every time you boot your RPI it has (initially) no idea what the date/time is and so defaults to 00:00:00 1st Jan 1970.........during boot the time is then changed, which causes these "clock skew" messages 2022-10-30 16:19:49 is there a way to get XDG_RUNTIME_DIR create/set to /run/user/$UID without using a display manager? The instructions on the sway wiki page put it in /tmp and that seems not to work well with podman. 2022-10-30 16:22:50 elagost, you can add something like this to your .profile file: https://wiki.gentoo.org/wiki/Weston#Usage 2022-10-30 16:29:47 Thanks, but that's the same way the alpine wiki page on Sway says to do it. distrobox seems to expect a /run/user/$UID directory, along with some other stuff too. 2022-10-30 16:31:38 Normally that's handled by some PAM plugin, but default login tools do not have PAM integration 2022-10-30 16:33:24 just display managers I guess? Hm. Guess I will just have to get used to that. 2022-10-30 16:34:45 also maybe more of a sway specific question, but the cursor contantly disappears and I have to move it over the bar to see it again. Is that a configuration issue? 2022-10-30 17:34:35 elagost: does it happen when you open and close a window? 2022-10-30 18:02:30 It seems like sway just has no global cursor config, and whatever the last application is that sets a cursor, it keeps that until something else sets it. So when I play a video in mpv and the cursor goes away, the cursor won't come back until I move it on to a GTK application or something. 2022-10-30 18:10:12 :/ 2022-10-30 18:10:22 that sounds awful. what is sway? 2022-10-30 18:10:49 https://swaywm.org/ 2022-10-30 18:11:26 sounds like a hardware cursor issue 2022-10-30 18:13:00 Afaik in Wayland clients are responsible for the cursor. 2022-10-30 18:18:36 it does not do that for me 2022-10-30 18:18:40 you can try WLR_NO_HARDWARE_CURSORS=1 before starting it 2022-10-30 18:22:31 bug? 'seq 1 4 > file;busybox ed file' 2i test . w q ... it replaces first line with "test" .. 2022-10-30 18:26:23 if you use 'seq 1 7' it seems to work.../ 2022-10-31 10:23:22 is there a way to get `journald` logging on alpine ? 2022-10-31 10:23:43 * is there an 'easy'/strightforwared way to get `journald` logging on alpine ? 2022-10-31 10:24:05 not sure if you'd really get any advantages of journald without systemd 2022-10-31 10:24:42 and getting `systemd` would be a pain right ? 2022-10-31 10:25:11 try another distro? 2022-10-31 10:30:18 linex[m]: systemd (and journald) does not support POSIX and musl libc. They only support glibc so its not possible to build/maintain systemd in a sustainable way in alpine 2022-10-31 13:19:07 "linex: systemd (and journald..." <- thank you 2022-10-31 14:37:43 are the gitlab issues the best way to ask questions about a package in aports? 2022-10-31 14:42:16 Guest34: do you have questions or issues/problems ? 2022-10-31 14:45:59 a question, wondering if the clojure package could also include the cli tools 2022-10-31 14:46:28 I think it just installs the clojure jar atm, upstream provides a separate script to install the cli tools 2022-10-31 14:47:42 perhaps the package could install the jar and the tools? I'm new to Alpine so not sure how things are done usually 2022-10-31 14:50:14 Guest34: looking on pkgs.alpinelinux.org that package has 2 files, a JAR and a /usr/bin/clojure - is that the cli? 2022-10-31 14:50:55 Guest34: ah, that's just a wrapper to run Java with the JAR 2022-10-31 14:50:58 yeah 2022-10-31 14:51:16 I installed that at first, I think it's just the wrapper 2022-10-31 14:52:26 so you could either raise an Issue or email the maintainer directly 2022-10-31 14:52:36 OK cool 2022-10-31 17:54:13 _maybe_ a bit obscure q: does feh and/or mpv have "command-lines/read" abilities to take user input? 2022-10-31 17:55:55 example? 2022-10-31 18:00:25 oh, not sure i thought this thorugh, but maybe in actions you could get user input to use in a shell script, like feh can pas vars to a shellscript 2022-10-31 18:07:45 i use a pop-up/dmenu style now, but kinda often get problems with it, like not killing the window when its done (but thats another topic)