2024-07-01 00:45:33 "Don't follow those links..." <- For me it is blocked by uBlock Origin AND by opendns. Kinda sus ;) 2024-07-01 00:46:16 So... can't tell if malware or not, but SURE nothing good. 2024-07-01 00:46:55 also, same URLs were posted in an entirely different room I am in 2024-07-01 00:46:58 so it's obviously spam 2024-07-01 01:01:42 if you're posting shortened urls with the message "We're having a party" in a linux distro channel, i think that's already enough to be hella sus 2024-07-01 02:42:26 What RISC V boards are people using with Alpine to daily drive using a tiling windows manager? 2024-07-01 11:55:50 is alpine impacted by this? https://blog.qualys.com/vulnerabilities-threat-research/2024/07/01/regresshion-remote-unauthenticated-code-execution-vulnerability-in-openssh-server 2024-07-01 11:56:25 v3.20 is still on "OpenSSH_9.7p1, OpenSSL 3.3.1 4 Jun 2024" 2024-07-01 11:56:37 yMVV_HsHcX0: unclear, but the report mentions glibc based systems 2024-07-01 11:58:36 thx 2024-07-01 11:58:46 Apparently musl does not use malloc in it's syslog implementation 2024-07-01 12:13:06 Humm, sorry if that's a well-known answer, but is there a way to magane RAR archives on Alpine? 2024-07-01 12:13:09 *manage 2024-07-01 12:13:13 Morning. Short question: Is it possible to bind Alpine Linux user authentication to an AD without an overly complex setup? I've seen SSSD in the edge/testing repo, but is that all there is? 2024-07-01 12:15:50 doesn't 7z handle rar? 2024-07-01 12:16:00 Not with the Alpine build 2024-07-01 12:16:30 try `bsdtar xf file.rar` from libarchive-tools. 2024-07-01 12:16:39 from gitlog comment 2b3abc74c5f33e4ab6b2b23a1716c813af73edc1 2024-07-01 12:16:49 AFAIK, it only handle like older RAR format 2024-07-01 12:16:53 But let-me try again 2024-07-01 12:17:49 Worked for those archives at least! 2024-07-01 12:17:51 Thank you yMVV_HsHcX0 2024-07-01 12:19:49 you know you can always just clone aports, go back to a version where unrar was not yet purged and do an abuild -r there 2024-07-01 12:20:32 just because alpine doesn't distribute it, noone stops you from building it yourself, and git has a handy history that enables you to do this very simply :) 2024-07-01 12:21:02 wait... librewolf is in edge/community now? 2024-07-01 12:21:03 yay 2024-07-01 12:22:40 Sure, but then it's a local package to manage locally 2024-07-01 12:22:49 If there's a working alternative, I'll take it :) 2024-07-01 12:24:26 true, and it leads to all kind of dependencies not being upgraded later... but for short, one-off things this is pretty ok. 2024-07-01 12:34:58 runxiyu_: Does RFP not break too many websites? Do you keep it enabled? 2024-07-01 12:35:06 I do like that they drop all the other crap/ads from mozilla. 2024-07-01 12:36:53 WhyNotHugo: It does break many websites... but I do more things that are guaranteed to break websites (e.g. disabling JavaScript) so I do an allowlist on an individual basis anyway 2024-07-01 13:52:21 upgrading from 3.17 to 3.20 2024-07-01 13:52:39 fingers crossed 2024-07-01 13:52:58 can I go straight to 3.20 or should I upgrade incrementally 3.17 -> 3.18 -> 3.19 -> 3.20? 2024-07-01 13:53:18 You can go straight 2024-07-01 13:53:37 Don't forget to add --available 2024-07-01 13:54:21 ok cool 2024-07-01 13:57:49 apk add --upgrade apk-tools shows conflicts with musl-dev-1.2.5-r0 and libc-dev-0.7.2-r3 2024-07-01 13:58:10 e.g. for musl: 2024-07-01 13:58:14 conflicts: libc-dev-0.7.2-r3[libc-dev=1.2.5-r0] 2024-07-01 13:58:19 satisfies: libc-dev-0.7.2-r3[musl-dev] 2024-07-01 14:07:03 any reason the wiki instructions shouldn't say apk upgrade --available apk-tools instead of apk add --upgrade apk-tools? 2024-07-01 14:07:36 bleb: wiki page link? 2024-07-01 14:16:54 You don't need to manually upgrade only apk-tools first. Apk handles that itself 2024-07-01 15:01:27 WhyNotHugo: https://wiki.alpinelinux.org/wiki/Upgrading_Alpine?useskin=vector#Upgrading_packages_2 2024-07-01 15:02:28 so I did apk upgrade -a apk-tools, expecting it would only upgrade apk-tools and its dependencies, but instead it upgraded everything 2024-07-01 15:03:58 the man page apk-upgrade(8) says "If list of packages is provided, only those packages are upgraded along with needed dependencies." 2024-07-01 15:04:04 so I think this is a bug? 2024-07-01 15:04:44 I know it upgraded texlive which apk-tools should not depend on... 2024-07-01 15:05:54 ANYWAY, after running that command, my system is pretty messed up. 2024-07-01 15:06:05 running any command gives a bunch of "symbol not found" errors 2024-07-01 15:06:16 for example if I type ls, I get "Error relocating /lib/libutmps.so.0.1: timeval_from_tain: symbol not found" 2024-07-01 15:06:25 and 7 other similar errors 2024-07-01 15:07:49 any suggestions for what to do now? 2024-07-01 15:09:45 I guess I'll try rebooting? 2024-07-01 15:11:09 WhyNotHugo: oops the proper link is here but it has the same instructions: https://wiki.alpinelinux.org/wiki/Upgrading_Alpine?useskin=vector#Upgrading_packages 2024-07-01 15:12:15 bleb: those are two separate commands 2024-07-01 15:12:21 that you need to execute in sequence 2024-07-01 15:12:35 `apk upgrade -a apk-tools` is not the same 2024-07-01 15:13:03 yes. the first one, when I executed it, showed conflicts with musl-dev-1.2.5-r0 and libc-dev-0.7.2-r3 2024-07-01 15:13:28 I read the man page for apk-upgrade and it seemed to describe the same process, so I tried that, and it upgraded all of my packages 2024-07-01 15:14:48 apk add -u: "Upgrade packages and it's dependencies." 2024-07-01 15:15:40 apk upgrade: "apk upgrade upgrades installed packages to the latest version available from configured package repositories (see apk-repositories(5))." 2024-07-01 15:16:28 by experiment, apk upgrade -a apk-tools seems to upgrade everything anyway... but at least as documented it seems like it should have been the same operation 2024-07-01 15:17:05 with the one difference that apk upgrade -a resets package versions 2024-07-01 15:17:56 anyway I rebooted and now I can't ssh in... have to drag out the vga monitor 2024-07-01 15:18:15 I have no idea what I will do on the system, if anyone has suggestions 2024-07-01 15:18:45 seems like I am probably screwed and just have to reinstall? 2024-07-01 15:30:45 so the screen after reboot shows Failed to update the dependency tree; failed to load deptree 2024-07-01 15:31:01 and, "Welcome to Alpine Linux 3.20" ! 2024-07-01 15:31:08 so at least that part was successful! 2024-07-01 15:32:11 ah, but my keyboard is not being detected 2024-07-01 15:32:25 yeah... this would have been good to avoid 2024-07-01 15:43:37 lol my vps is also on 3.17 2024-07-01 15:43:42 I guess I will just compile got this time... 2024-07-01 15:44:04 or just use git 2024-07-01 15:44:06 hope its ok to ask a bit offtopic here, anyone know if using 4/5g-router and wifi instead of 4/5g directly to smart-phone will make a big difference? 2024-07-01 19:10:44 Hi there. 2024-07-01 19:11:02 echo $SHELL gives me: /bin/sh 2024-07-01 19:11:18 which aapt gives me: /tmp/foobar/build-tools/30.0.0/aapt 2024-07-01 19:11:29 When I run aapt I get: -sh: aapt: not found 2024-07-01 19:11:55 Is this sh specific? 2024-07-01 19:12:22 head -n1 /tmp/foobar/build-tools/30.0.0/aapt 2024-07-01 19:12:52 https://termbin.com/eets 2024-07-01 19:13:21 ldd /tmp/foobar/build-tools/30.0.0/aapt 2024-07-01 19:14:01 https://termbin.com/i33j 2024-07-01 19:14:13 libc.so.6 => /lib64/ld-linux-x86-64.so.2 (0x7ffabcbee000)\ 2024-07-01 19:14:19 that's glibc 2024-07-01 19:15:16 Hm what di I do already to make it run on a laptop 2024-07-01 19:15:52 gcompat probably 2024-07-01 19:17:18 Yhank you ikke 2024-07-01 19:17:24 Sometimes I forget I'm on musl 2024-07-01 19:17:39 s/Yhank/Thank 2024-07-01 19:24:30 does alpine linux or the installer spoof its mac address by default? 2024-07-01 19:24:52 Not that I'm aware of 2024-07-01 19:24:54 trying to figure out why I have 4 dhcp leases for 4 different mac addresses within the past 2 hours 2024-07-01 19:25:18 for the same machine where I haven't moved the ethernet cable 2024-07-01 19:25:38 Maybe because that machine doesn't have a fixed mac address 2024-07-01 19:26:01 meaning the nic doesn't have a mac address hard coded? 2024-07-01 19:26:19 I don't know what hardware you're running 2024-07-01 19:26:39 hp microserver n40l 2024-07-01 19:26:53 If it's a NIC, most likely it has 2024-07-01 19:27:34 it's not a separate card, just what is built into the motherboard 2024-07-01 20:31:04 quinq: libarchive got RAR5 support a few years ago 2024-07-01 20:37:03 That's nice 2024-07-01 22:19:44 Hi. My speakers and microphone aren't working on my ASUS Flip C434 laptop running Alpine Linux. Trying to play sounds with the speaker or record sounds with the microphone results in no sounds being played or recorded, respectively. Running speaker-test results in the error, "Playback open error: -2,No such file or directory". 2024-07-01 22:19:57 If I attach a USB to 8mm adapter to the laptop and then plug in earbuds into the adapter, I can hear sound, though my microphone still doesn't work. I've tried connecting to bluetooth headphones that have a built-in mic, but even when it says I've successfully connected, no sound plays. 2024-07-01 22:20:06 I found this post (https://www.reddit.com/r/linux4noobs/comments/u4upa2/converting_the_asus_c433t_chromebook_to_linux/), which talks about audio issues and solutions with the seemingly closely-related ASUS c433t Chromebook. However, the proposed solution didn't work on my laptop. Any advice would be appreciated. 2024-07-02 00:07:10 one of my /media dirs has drwxrwxrwx@ how do I get rid of the extended attributes? I tried this but it didn't work the @ is still there. sudo setfacl --remove-all -R /media/dir 2024-07-02 00:08:36 Aerith: do you use pipewire or pulseaudio? 2024-07-02 00:26:39 Pipewire, it seems 2024-07-02 00:58:14 voidastro[m]: @ means extended attribute. extended access control list is marked by +. you can manage extended attributes with the getfattr and setfattr commands 2024-07-02 00:59:23 Hello71: I see. but there is no option to remove-all. only: -x name, --remove=name 2024-07-02 01:00:08 no recursion either. is there a way to remove all the xattrs and get rid of the @? 2024-07-02 08:21:30 voidastro[m]: getfattr -d will list all attributes 2024-07-02 08:27:10 Hi there, I have an issue with sshd, it says that error occurred with networking. I tried to run it manually, here are logs: 2024-07-02 08:27:19 https://www.irccloud.com/pastebin/Qp4FRwaV 2024-07-02 09:19:59 sasha00066: share the output of `mount` 2024-07-02 09:20:09 also `ls -l /proc` 2024-07-02 09:47:59 Will the openssh package be updated in 3.18/3.19/3.20, due to "regresshion"? 2024-07-02 09:54:45 Andreas[m]: they write "The vulnerability tracked as CVE-2024-6387 is related to OpenSSH servers running with the GNU C Library (glibc) in Linux environments. 2024-07-02 09:54:49 " 2024-07-02 09:55:10 alpine is musl-based ;) 2024-07-02 09:57:54 consus: I don’t know what is mount, but here are ls -l / proc: 2024-07-02 09:58:01 https://www.irccloud.com/pastebin/NknwTiF7 2024-07-02 10:07:13 "Andreas: they write "The..." <- I wouldn't count on that. This is our syslog(): http://git.musl-libc.org/cgit/musl/tree/src/misc/syslog.c#n85 2024-07-02 10:08:14 Yes, it doesn't malloc/free. It calls into stdio (snprintf, vsnprintf, dprintf), these can be quite some hell to make fully safe - not sure if musl's are. 2024-07-02 10:12:09 sasha00066: hmmmm... `apk info | grep linux` 2024-07-02 10:13:07 divVerent: doing things in signal handler is known to be problematic, I wonder why ssh devs still do that 2024-07-02 10:13:11 definitely some stuff also not thread safe in there, or rather, only mostly thread safe - probably behaves right on amd64, but on platforms where pointers are too big for a single MOV (x64 is one of them), this looks awful 2024-07-02 10:13:34 that's actually odd as POSIX says syslog functions must be thread safe 2024-07-02 10:13:40 oh wait, it is 2024-07-02 10:13:46 __vsyslog itself takes a lock 2024-07-02 10:14:18 ok, I don't know 100% what the LOCK function does in musl, but I'd bet it deadlocks then in async signal use - still not GREAT, likely a DoS attack possibility - but at least no remote code execution 2024-07-02 10:18:37 yeah, LOCK nowhere uses the current thread ID or something similar, so it must deadlock if acquired twice by the same thread 2024-07-02 10:20:17 so yeah... in Alpine, regreSSHion exists but is just a deadlock 2024-07-02 10:20:29 definitely do want to patch that 2024-07-02 10:21:33 the way to deadlock Alpine is, of course, causing SIGALARM right when the process is calling into some printf-like function 2024-07-02 10:22:02 that'll create one sshd subprocess that runs as root and does nothing forever... enough of those and the machine can no longer create new processes and hangs 2024-07-02 10:23:15 to mitigate against whole-system impact, set a num processes limit on the sshd process chain... might already be the case; if that's done, then the impact of the bug is still that no more SSH logins will be accepted until sshd restart 2024-07-02 10:24:44 s/some/syslog/, s/printf-like function/too/ 2024-07-02 10:26:42 musl's printf family on first glance actually looks signal and thread safe BTW, I see no abuse of static or globals right there 2024-07-02 10:32:24 divVerent: see also https://dustri.org/b/notes-on-regresshion-on-musl.html 2024-07-02 10:32:57 which also confirms that a deadlock is possible 2024-07-02 10:32:58 consus: result is: 2024-07-02 10:33:03 https://www.irccloud.com/pastebin/gzX70qKP 2024-07-02 10:39:34 sasha00066: ah so you have a container? 2024-07-02 11:12:37 lol ASLRn't - didn't see that convention anywhere else other than in Mario Maker troll level community (like "yesn't", "smartn't"). 2024-07-02 11:14:55 haha 2024-07-02 11:18:45 BTW ASLRn't seems to not be fixed on Alpine 2024-07-02 11:20:43 sysctl vm.mmap_rnd_bits=32 vm.mmap_rnd_compat_bits=16 is Ubuntu's fix from https://bugs.launchpad.net/bugs/1983357 2024-07-02 11:23:20 they say there's a chance of things breaking with this setting if one runs 32bit applications that can only handle limited address space or maybe use the vsize resource limit (not sure if that affects it?), so I guess the obvious thing to check if it still works is WINE 2024-07-02 11:25:25 https://www.openwall.com/lists/oss-security/2024/07/01/13 2024-07-02 11:43:06 if my server restart before my internet provider box, networking stay stucked. How can I configure the default networking stack to re-trigger udhcp later by itself? 2024-07-02 12:38:33 staceee: bb-udhcpc you mean? openwrt uses that too 2024-07-02 12:38:52 on its wan interface 2024-07-02 12:39:15 yap 2024-07-02 12:39:43 the vanilla udhcpc cli, run by networking, on the configured interfaces 2024-07-02 12:40:09 try to debug this with -s then? 2024-07-02 12:40:21 -t 0 might be what you want? 2024-07-02 12:41:25 here'w what my openwrt box has: 2024-07-02 12:41:25 udhcpc -p pidfile -s script -f -t 0 -i eth1 -C -R -O 121 2024-07-02 12:42:04 https://www.busybox.net/downloads/BusyBox.html 2024-07-02 12:51:56 thanks for the clues 2024-07-02 13:42:06 hello, I'm trying to figure out how to make virtual machine with direct kernel booting work. Actually it works almost, I just don't understand why some kernel modules cannot be loaded, like af_packets. Kernel with its modules are installed locally in the VM root disk, even though the kernel I use to boot is located on the hypervisor. I made sure it was the exact same version of the kernel. 2024-07-02 13:42:56 when I try to modprobe af_packet, I get: modprobe: ERROR: could not insert 'af_packet': Invalid argument 2024-07-02 14:15:45 raspbeguy, anything in dmesg? 2024-07-02 14:19:38 Habbie, this is what happens when I try to modprobe af_packet https://termbin.com/x1h6 2024-07-02 14:20:06 aha, that's not a lot to go on indeed 2024-07-02 14:23:27 Greetings, a question, I'm using alpine 3.20.1, running libvirt with a session uri (per user), I've configured a specific bridge to be shared between the sessioned vms but when I try to start one that uses the shared bridge, I get this error: error: Unable to create tap device vnet0: Operation not permitted. aaccording to the libvirt devs, it should work, what am I missing? 2024-07-02 14:34:59 Habbie, I now wonder if it could be linked to the modules signature, i don't know 2024-07-02 14:50:52 im merging openssh upgrade to 9.8 now. with this upgrade you will not be able to establish new ssh connections before you restarted the service 2024-07-02 14:51:06 so make sure you restart sshd when you upgrade edge 2024-07-02 15:10:08 Hi i am trying to debug a script and i dont really understand why [[ -n $TAG ]] is True if TAG="" 2024-07-02 15:11:27 is it? 2024-07-02 15:12:00 how do you see that? 2024-07-02 15:12:02 is there some non-printable character in there? 2024-07-02 15:13:10 do a `declare -p TAG` to see the variable's contents 2024-07-02 16:20:26 Just out of curiosity, I'm fairly new to Alpine, but my take would be that no versions of Alpine would be vulnerable to regreSSHion due to it being specific to glibc built openSSH that is vulnerable. Is this correct? 2024-07-02 16:22:55 h0bble5: https://dustri.org/b/notes-on-regresshion-on-musl.html explains it 2024-07-02 16:23:09 summary, no RCE, but potentially still a deadlock 2024-07-02 16:33:17 Ah thanks a ton for the link. I was just wondering which systems I should prioritize patching on. 2024-07-02 17:09:09 @lopid will try that 2024-07-02 17:10:31 thing is that i encountered this issue in a gitlab ci pipeline so debugging is a charm 2024-07-02 17:15:42 @lopid declare -p did return '' 2024-07-02 17:22:20 hutbuerger: what job? 2024-07-02 17:23:54 i am runnin a gilab ci job on a self hosted gitlab-runner and the condition is " if [[ -n ${CI_COMMIT_TAG} ]]; then" 2024-07-02 17:24:18 and it alwaws runs into the TRUE block. 2024-07-02 17:28:32 when you say it returns '', do you mean it said ｢declare -- CI_COMMIT_TAG=""｣? or did it actually say "''"? 2024-07-02 17:31:08 as suggested by lopid i did declare declare -p ${CI_COMMIT_TAG} 2024-07-02 17:31:43 no, declare -p CI_COMMIT_TAG 2024-07-02 17:31:54 oh i see 2024-07-02 17:34:09 it seems like the gitlab-runner:alpine dosnt have declare with "/bin/sh: eval: line 146: declare: not found" 2024-07-02 17:34:20 it's a bash command 2024-07-02 17:36:57 didnt know that. will run with bash 2024-07-02 17:39:33 declare -p CI_COMMIT_TAG returns declare -- CI_COMMIT_TAG 2024-07-02 17:40:52 no =? 2024-07-02 17:41:09 i copied that straight over 2024-07-02 17:42:36 and that declare statment ran only after the [[ -n CI_COMMIT_TAG ]] 2024-07-02 17:51:26 the thing is that the statement is running inside of an gitlab/gitlab-runner:alpine@3c497077c445 inside of a docker:stable@b0757c55a1fd 2024-07-02 17:52:18 do you concur in the assumt 2024-07-02 17:52:45 *ption that the check is at least performing wierd? 2024-07-02 17:55:14 i'm more suspecting there's a preparser involved 2024-07-02 17:57:02 Thank you so much for your help. Do you have an idea for debugging this further? 2024-07-02 17:57:58 not right now, i have to go 2024-07-02 17:58:38 Either way i appreaciate the support. Thanks 2024-07-02 18:32:41 consus: Sorry for delay, I have x86 emulator called "iSH" that comes with pre-installed alpine. It’s the only choice for iPhone. 2024-07-02 19:11:42 ah, yes, that's what I have to use 2024-07-02 19:30:54 Hi, suppose my bootup is being kinda slow, what should I use to see what's causing the holdup? Like a flamegraph or something. 2024-07-02 19:32:52 OpenSSH upgrade should be in https://alpinelinux.org/posts/ ? 2024-07-02 19:33:25 It broke CI on sr.ht, because build VMs are accessed via ssh 2024-07-02 19:34:31 Not sure if other automated scenarios can break too 2024-07-02 19:34:32 Why did it break CI? 2024-07-02 19:34:59 (I know the issue with ssh needing restarts, but ncopa did something to address it) 2024-07-02 19:35:19 CI starts a VM for the build, accesses it via ssh. During a phase in the CI pipeline, `apk` upgrades ssh. It becomes impossible to access the job via ssh. 2024-07-02 19:35:30 WhyNotHugo: what release? 2024-07-02 19:35:47 I assume edge 2024-07-02 19:36:10 Oh, I also thought it was broken on 3.20, but now I see that I mis-tested. 2024-07-02 19:36:53 in 3.20, only the vulnerability was patched 2024-07-02 19:37:36 oh: "I'm gonna merge this as is, and we can see if we can work with upstream to solve the lose connectivity on upgrades." 2024-07-02 20:25:33 hi! what did I do? 2024-07-02 20:26:13 The openssh upgrade on edge 2024-07-02 20:26:39 maybe a post on mastodon and x 2024-07-02 20:26:53 in addition to the post on website 2024-07-02 20:27:03 I've boosted your post on mastodon 2024-07-02 20:27:10 ok. good 2024-07-02 20:27:13 I originally thought this affected 3.20, my bad. 2024-07-02 20:27:17 thanks 2024-07-03 00:40:03 🫃 2024-07-03 09:28:07 I want to fetch packages over http, and not https, when building Alpine 3.20 images. I've replaced https with http in /etc/apk/repositories. Should I now only need to run apk add now, without any trafik over https? 2024-07-03 09:29:26 Yes 2024-07-03 09:32:38 ikke: And Alpine doesn't do DNS over HTTPS? I'm not very familiar with that, so maybe that's a dumb question. 2024-07-03 09:33:12 Not out of the box, no. It just uses the resolvers specified in /etc/resolv.conf 2024-07-03 09:33:38 And just plain dns (port 53) 2024-07-03 09:36:25 Context: I'm trying to do this behind a corporate proxy. I've done that I descriped, but when I try to do apk I get the "temporary error" message. 2024-07-03 09:37:09 But when I manually download the ca-certificates apk package and copy that during building of docker image, and then manually install it, it succeeds 2024-07-03 09:37:35 Not fully, but going from failing to the first apk command to the last one 2024-07-03 09:50:16 ikke: Thanks! 2024-07-03 12:59:02 sasha00066: it has some weird kernel that lacks common files in /proc 2024-07-03 12:59:08 Probably custom 2024-07-03 12:59:30 Kernel config would be nice 2024-07-03 15:10:40 I think this can be merged already: https://gitlab.alpinelinux.org/alpine/aports/-/merge_requests/65880 2024-07-03 15:10:48 !65880 2024-07-03 15:57:46 Could somebody advice me how to run i386 alpine container? When running `podman run -it --arch=i386 i386/alpine /bin/uname -m` I get still x86_64 2024-07-03 15:58:34 mcepl: execute the shell with linux32 2024-07-03 15:58:58 sorry, where? Parameter of podman? 2024-07-03 15:59:23 for example as entrypoint (if that's the same as docker) 2024-07-03 15:59:47 podman run -it --arch=i386 --entrypoint linux32 alpine uname -m 2024-07-03 15:59:55 (I assume podman supports manifests) 2024-07-03 16:00:40 yup, that works! Thank you! 2024-07-03 17:26:40 One more question: could anybody explain why this Python script fails on Alpine/i386 with OSError: Invalid argument (https://gitlab.com/m2crypto/m2crypto/-/jobs/7258943848), while it works perfectly on other Linuxes with glibc? The only difference I can observe is that socket.SO_RCVTIMEO is 20 with glibc, but 66 with musl. 2024-07-03 17:28:00 (this script means https://paste.sr.ht/~mcepl/9e3f6da0dd9794c0b2b4dc0f546c9fd24dd61a08) 2024-07-03 17:44:58 The man page mentions this: "optlen invalid in setsockopt(). In some cases this error can also occur for an invalid value in optval (e.g., for the IP_ADD_MEMBERSHIP option described in ip(7))." 2024-07-03 17:45:32 I guess the first part is only relevant for C 2024-07-03 17:49:54 Hi everyone! I’m interested in the cloudflared package which seems to exist in testing: https://gitlab.alpinelinux.org/alpine/aports/-/tree/master/testing/cloudflared?ref_type=heads but I can find it in neither https://pkgs.alpinelinux.org/packages nor the actual package mirrors. Why could that be? 2024-07-03 17:51:03 https://gitlab.alpinelinux.org/alpine/aports/-/blob/master/testing/cloudflared/APKBUILD?ref_type=heads#L8-11 2024-07-03 17:52:26 oh, that makes sense, thanks! 2024-07-03 17:52:55 It has been 9 months ago, so perhaps the quick module has been updated 2024-07-03 17:55:00 good point, I'll try building it but need to set up the environment first 2024-07-03 17:55:21 quic-go has been updated multiple times 2024-07-03 19:14:49 cloudflared 2024.6.1 compiles with Go 1.22, but then I run into https://github.com/cloudflare/cloudflared/issues/1151 where it turns out they use a fork a Go compiler to build their binaries. Oh well, it was worth a try. 2024-07-03 19:16:44 🤦 2024-07-03 19:17:36 indeed 2024-07-03 19:19:16 haha clownfare 2024-07-03 19:40:03 let me interject for a moment, what you call clownfare, i recently took calling clownfart, or clown+fart. 2024-07-03 20:18:46 yMVV_HsHcX0, I suppose that works too 2024-07-03 21:24:36 I have a new laptop so I will be migrating off my current one. Current laptop runs 3.20.1, is there a good guide for migrating from one laptop to another? I have backups. 2024-07-03 21:25:02 I guess my first question is, can I dump the list of packages (ie `apk list -I`) and use that to install the same packages on the new machine? 2024-07-03 21:31:37 pjs: use /etc/apk/world 2024-07-03 21:31:42 that is the list of explicitly installed packages 2024-07-03 21:38:19 morning. 2024-07-03 21:40:11 if i had a number of Alpine systems and want them to synchronize their repositories against a local mirror, including a download proxy with the goal to reduce the workload on the infrastructure as well as avoiding multiple downloads of the same package - how would i do that? both in terms of what is needed on the repo mirror as well as the download proxy? 2024-07-03 21:40:14 ikke: ok cool. And how do I feed that into apk on the new machine? Or is it something like `doas apk add < world.copy` ? 2024-07-03 21:40:35 add the packages to world and then apk fix 2024-07-03 21:41:11 oh cool! thanks 2024-07-03 21:46:16 Marquel: personally, I stole and adapted this nginx config from sr.ht: https://git.sr.ht/~sircmpwn/builds.sr.ht/blob/master/contrib/repo-cache.nginx.conf 2024-07-03 22:02:31 if I remember right, the setup-alpine script will now let you setup encrypted drive, etc. right? Last time I did a fresh install was I think 3.17 and I remember not having to do anything but answer some questions for the encrypted stuff 2024-07-03 22:22:37 is there any way to make the console font bigger *during install* I literally can't read the questions for setup-alpine 2024-07-03 22:27:44 well using zoom on my phone camera is working out OK I guess lol 2024-07-03 22:30:06 https://unix.stackexchange.com/a/49823 2024-07-03 22:30:11 try the setfont thing 2024-07-03 22:30:31 but I don't know what's present on the alpine ISO by default 2024-07-03 22:32:44 yea, consolefonts aren't there by default 2024-07-03 22:33:37 install some? 2024-07-03 22:34:16 hrmm, can I do that before even running setup-alpine? I mean the disk isn't even selected,e tc. 2024-07-03 22:34:20 mirrors, wifi, etc. 2024-07-03 22:34:24 i mean instal it in the iso 2024-07-03 22:34:30 not for the actual setup 2024-07-03 22:35:36 well yea. Of course I'll do that after the setup but I can't even read the prompts for the setup. It's all good. I'm working through it with camera zoom lol 2024-07-03 22:35:49 sounds crappy 2024-07-03 22:35:52 godspeed 2024-07-03 22:36:17 haha yea 2024-07-04 09:37:17 i was wondering, do you guys rebuild linux entirely on every build on the linux packages? or are the build artifacts cached and then it just builds whatever changes? 2024-07-04 09:38:47 i ask because usually i see the artifacts get deleted after a successful when i use abuild, and i was thinking to make a linux package but also keep the artifacts cached to avoid rebuilding every time entirely 2024-07-04 09:38:59 *after a successful build 2024-07-04 09:40:55 rdbo: the builders only need to build what's missing. They keep a complete repository locally (as a matter of fact, that's the canonical source for the repo) 2024-07-04 09:42:11 Just in case it's not clear, our builders are separate from our CI infrastructure 2024-07-04 09:45:08 @ikke the builders doin't do clean up then? (i assume `abuild build` does clean up by default) 2024-07-04 09:46:30 Cleanup what? 2024-07-04 09:47:18 for example, the package has the sources with checksum. it downloads the sources, runs the build(), installed the content on $pkgdir, and then removes the artifacts 2024-07-04 09:47:41 but by removing the artifacts it will have to recompile everything again (in the case of linux kernel can take a long while) 2024-07-04 09:48:58 Compared to chromium, the kernel builds fast ;-) 2024-07-04 09:49:57 XD true 2024-07-04 09:50:24 rdbo: abuild puts the package in ~/packages. As long as the package is there (with the right version), it knows it does not need to build it any more 2024-07-04 09:50:41 The rest can be removed 2024-07-04 09:51:22 ah ok, i thought maybe the artifacts for big packages were being kept to save build time 2024-07-04 09:54:14 The builder does cache all downloaded source files 2024-07-04 09:54:29 (and abuild in general) 2024-07-04 09:57:16 i will try to avoid running the cleanup part of abuild to keep the build cache of linux, and start using a alpine package instead of copying my custom kernel all the time 2024-07-04 09:58:00 will be more convenient to update the kernel also 2024-07-04 09:59:27 abuild -K will keep the src and pkg directories 2024-07-04 10:00:40 oohhh that's better than what i initially though. i was thinking of running the steps individually `abuild prepare`, `abuild package`... lol 2024-07-04 10:00:44 thanks @ikke 2024-07-04 13:00:19 Anyone using Signal Desktop from Flatpak? I cannot access any files from my harddrive. Any Idea how to fix? 2024-07-04 13:03:08 i believe flatpack has some commandline switches that allow to mount host filesystems in the jail 2024-07-04 13:03:27 sounds more like a portals issue 2024-07-04 18:48:12 wizzard: try using flatseal to grant access to the directory 2024-07-04 21:42:40 ikke: so, when i386/alpine is the only image where M2Crypto test suite fails in setsockopt(SOL_SOCKET, SO_RCVTIME0, ...) in M2Crypto tests? Known bug in musl libc? 2024-07-05 04:15:52 Perhaps ask in #musl on irc.libera.chat 2024-07-05 09:16:35 Hi! I'm trying to set up an nfs server with kerberos authentication. Each works fine on its own but I can't get them to work together. Without kerberos (sec=sys) it works fine, but with kerberos (sys=krb5), I can mount, but when I try to access the mount (ls for example), I get `Operation Not Permitted` on the client. I added `rpc.idmapd` to NFS_NEEDED_SERVICES in /etc/conf.d/nfs. Do I need to do 2024-07-05 09:16:41 anything on the client side as well? Where do I see the client logs for the nfs mount? 2024-07-05 09:35:15 tmrlvi: would be nice with an wiki article if you figure it out ;) 2024-07-05 09:41:50 ncopa: I thought about writing one. Arch has a nice article about it, but I'm not sure how to translate the client part to Alpine 2024-07-05 11:08:18 Ok! Solved it. Was just missing the rpc.gssd that is used for kerberos authentication. So on the client you need to enable both rpc.idmapd and rpc.gssd service 2024-07-05 13:01:27 ah! nice find! 2024-07-05 13:01:54 Maybe add to this doc: https://wiki.alpinelinux.org/wiki/Setting_up_an_NFS_server 2024-07-05 15:36:51 hello 2024-07-05 15:36:52 ping 2024-07-05 15:37:38 is this room bridged from matrix to irc ? 2024-07-05 15:37:47 i am on irc 2024-07-05 15:37:50 if you are on matrix 2024-07-05 15:37:52 we have our answer 2024-07-05 15:38:05 perfect !! (yes i'm on matrix) 2024-07-05 15:38:07 (i see the [m] in yhour nick :) ) 2024-07-05 15:38:10 *your 2024-07-05 15:38:17 and the org.matrix: in your realname 2024-07-05 15:38:56 is it possible to change an alpine sysdisk installation to a diskless one ? 2024-07-05 15:43:08 i'm not looking to re-install in diskless mode .. I'm facing problems with the iso and the only way to boot in a sane way was to untar the minirootfs and chroot to setup kernel and bootloader. 2024-07-05 15:44:17 but that got me a sys disk mode installation 2024-07-05 16:06:50 jbara[m]1: dont unpack the minirootfs onto your target disk 2024-07-05 16:07:09 unpack it somewhere else, chroot into it and use setup-alpine from there to install to the target 2024-07-05 16:08:35 any other users of imv(1): it's crashing on startup for me with "Assertion failed: vi (../src/x11_window.c: imv_window_create: 124)" 2024-07-05 16:08:38 anyone else seen that? 2024-07-05 16:09:56 nero: understood. I just found this too (in `mkinitfs-bootparam(7)` ) 2024-07-05 16:09:57 > If no root= parameter is given, the initramfs will build a live system in memory from scratch. This is also called diskless mode. 2024-07-05 16:11:15 yes, this is how diskless mode / the iso installer boots 2024-07-05 16:11:43 but you want to setup that using the regular setup-* scripts 2024-07-05 16:14:24 elly: just installed it to check, seems to work at my end 2024-07-05 16:14:37 I'm somewhat lost , so the setup-* scripts do install a bootloader and stuff for you ? 2024-07-05 16:14:38 hm! I wonder what's wrong on my system then 2024-07-05 16:14:40 on x11 fwiw 2024-07-05 16:14:52 i vaguely remember seeing that error before 2024-07-05 16:14:56 let me check what's there in the source 2024-07-05 16:15:30 nero: i believe, I read in the installation instructions that you should set the disk to none if you wanna install diskless , but that doesn't seem to do anything to any disk. 2024-07-05 16:16:22 assert(vi); 2024-07-05 16:16:22 XVisualInfo *vi = glXChooseVisual(window->x_display, 0, att); 2024-07-05 16:16:46 right - what I don't understand is why glXChooseVisual() is returning null there 2024-07-05 16:16:54 yea 2024-07-05 16:17:23 it is not exactly asking for anything weird 2024-07-05 16:17:31 it's probably a setup issue 2024-07-05 16:17:39 yep, I just can't figure out what it is 2024-07-05 16:17:45 I will try to debug it more deeply later I guess 2024-07-05 16:18:20 are you using x11? 2024-07-05 16:18:57 > If no conforming visual exists, 2024-07-05 16:18:57 NULL is returned 2024-07-05 16:18:58 hm 2024-07-05 16:19:04 I am using x11, yeah 2024-07-05 16:19:15 i really don't know anything about opengl, but maybe check the attr declaration just above 2024-07-05 16:19:19 att* 2024-07-05 16:19:37 maybe there's something there that doesn't apply for you? 2024-07-05 16:19:57 could be although my system definitely has 24-bit color depth and obviously has double-buffering 2024-07-05 16:20:17 well, I will investigate and maybe open a ticket on the imv repo or somesuch 2024-07-05 16:20:23 yea sure 2024-07-05 16:21:13 i guess i didn't find anything new for you, heh 2024-07-05 16:43:25 jbara[m]1: yes, iirc it was setup-bootable in particular 2024-07-05 17:01:08 btw, nlplug-findfs isn't interpreting usbdelay correctly in my case (causing a mount error on every boot) 2024-07-05 17:01:57 what do you mean, "not correctly" ? 2024-07-05 17:03:14 i believe it should wait for my usb to be detected before trying to mount .. I tried setting the boot parameter as high as 120 seconds .. but I always get a mount fail (and 7 seconds later kprint shows sdb being detected , which is when I do the mount manually and type exit for the boot to continue) 2024-07-05 17:03:55 > what do you mean, "not correctly" ? 2024-07-05 17:03:55 nero: I mean not working at all, lmao. 2024-07-05 17:14:44 jbara[m]1: this might be worth a ticket on mkinitfs 2024-07-05 17:15:10 i suspect the conditions for "boot devices" found trigger ahead and thus cause the exit before the USB hotplug 2024-07-05 17:19:57 is there something i can share in order to help with this ? 2024-07-05 17:21:17 question: does nlplug-findfs not wait for the entirety of the usbdelay parameter ? 2024-07-05 17:22:33 I think it's not consistent (or maybe it depends on the order of the position of the parameter in kernel cmdline, although I think the latter is not likely) 2024-07-05 17:37:34 jbara[m]1: no, if it finds the boot, repo and apkovl before that, the timer gets changed 2024-07-05 17:37:52 https://gitlab.alpinelinux.org/alpine/mkinitfs/-/blob/master/nlplug-findfs/nlplug-findfs.c?ref_type=heads#L941 2024-07-05 17:38:24 im still figuring out the logic behind the several timeouts 2024-07-05 17:42:00 jbara[m]1: do you have an AL gitlab account? 2024-07-05 17:42:20 if you open an issue there we'll have an place to write it down 2024-07-05 17:43:43 jbara[m]1: also, what is on the usb device that is failing to mount? the iso? your apkovl? 2024-07-05 17:57:10 nero: I don't have an AL gitlab acc (I'm kinda new to alpine) 2024-07-05 17:58:25 but I'm sure the iso had always failed to mount (with my 2.0 flash drive) but I'm not sure if it still failed with usbdelay , weirdly it did work for the new sdb1 that was setup with setup-bootable (will retry it a couple of times to make sure) 2024-07-05 17:58:51 it consistently failed on my manual disk mode installation on the flash drive though 2024-07-05 18:00:01 I had two problems with it , the usbdelay and the fact that ext4 doesn't load despite the fact that it is present in /etc/modules (it is only loading when modules parameter is passed) 2024-07-05 18:00:10 and yes, i didn't forget to rebuild the initramfs 2024-07-05 18:37:58 nero: can I get some help ? 2024-07-05 18:39:00 jbara[m]1: im just some random person in here with you 2024-07-05 18:39:22 i think you are off best by creating an ticket in the alpine linux gitlab, on the mkinitfs project 2024-07-05 18:39:24 I actually thought you're a maintainer lol .. sorry 2024-07-05 18:39:52 current setup is just a vfat /dev/sdb1 setup with setup-bootable (and both apkovl and cache are in that same partition) 2024-07-05 18:39:54 i contributed before but i have no more privileges than you 2024-07-05 18:40:58 usbdelay still not working, and so i mount it myself when i'm dropped on a shell during boot , but it doesn't load the apkovl 2024-07-05 18:41:24 maybe I'll try passing it via boot param , but does anybody know a way to force loading it ? 2024-07-05 18:43:22 nero: i understand 😁 still though , I'm grateful for your help so far 2024-07-05 18:55:11 setting boot parameter as a path solves this btw. 2024-07-05 18:55:48 (solves the loading of apkovl once switch_root is executed .. so at least I'm able to persist across reboots now) 2024-07-05 19:17:36 I updated the init script with a loop that waits for blkid | grep -q "$KOPT_root" to happen (right before the nlplug-findfs command) 2024-07-05 19:18:48 ncopa: Updated the wiki entry :) 2024-07-05 19:18:55 tmrlvi: great :) 2024-07-05 20:09:02 would someone post the alpine public key fingerprint so i can verify from another source please? then i will 'trust' it thanks 2024-07-05 20:12:12 0482D84022F52DF1C4E7CD43293ACD0907D9495A 2024-07-05 20:12:28 That's the fingerprint for ncopa.asc 2024-07-05 20:12:33 thx 2024-07-05 20:16:24 which package provides the default /etc/X11/xorg.conf file? Doing a package contents search with that path provides no results 2024-07-05 20:16:55 bascially I copied over my configs from my old laptop and I want to restart with the defaults because X isn't starting lol 2024-07-05 20:19:50 I don't have that file 2024-07-05 20:19:58 So it's not managed by any package 2024-07-05 20:21:54 Gotcha, even better lol 2024-07-05 20:33:33 tmrlvi: idmapd should be optional on new kernels, the nfsidmap upcall will be enough 2024-07-05 20:33:44 tmrlvi: er, on new clients* 2024-07-05 20:45:30 grawity: ah, thanks. I'll fix it (I thought it was needed because of the comment in the conf file, and when enabled in the server, you must also enable in the client) 2024-07-05 22:04:03 Hi, any idea why Alpine 3.19.2 at boot gets stuck for ~5min here: https://0x0.st/XBNj.jpg ... and then ~3m here: https://0x0.st/XBNe.jpg ? 2024-07-05 22:04:12 I'm running this on Thinkpad X200, so not the most obscure platform I figure. 2024-07-05 22:22:24 ForeverNoob[m], after it boots, do you have "spew" in dmesg output? 2024-07-05 22:22:50 ForeverNoob[m], and are you using a stock kernel/config or did you make you own kernel config? 2024-07-05 22:28:17 tdr: `doas dmesg | grep spew` doesn't result in any output, and everything is as stock as it can be. 2024-07-05 22:28:45 not literal match of the word "spew" i meant ugly messages 2024-07-05 22:31:04 Not that I can see. 2024-07-05 22:31:08 Might also mention I run LibreBoot on there, but that should make things more transparent than less so. 2024-07-05 22:34:00 the libreboot shouldnt matter 2024-07-05 22:34:28 Thought so too. 2024-07-05 23:05:06 Ok, the dmesg package requires me to be on v3.20, but when upgrading, I suddenly see: 2024-07-05 23:05:10 (191/273) Downgrading linux-lts (6.6.37-r0 -> 6.6.36-r0) 2024-07-05 23:07:14 Huh, apparently 3.19 has a newer kernel than 3.20: https://pkgs.alpinelinux.org/packages?name=linux-lts&branch=v3.19&repo=&arch=&maintainer= vs. https://pkgs.alpinelinux.org/packages?name=linux-lts&branch=v3.20&repo=&arch=&maintainer= 2024-07-05 23:18:10 `dmesg -H` shows me this: https://paste.debian.net/plain/1322450 - Could this be relevant? "clocksource: Long readout interval, skipping watchdog check" 2024-07-05 23:18:24 (that paste is an excerpt btw) 2024-07-06 06:09:39 has anyone succeeded in building Rstudio on alpine before? 2024-07-06 09:05:45 Hi, when I reconfigure the network with `rc-service networking restart`, I get a warning message 2024-07-06 09:05:48 networking | * WARNING: you are stopping a boot service 2024-07-06 09:06:01 Is there a better way to reconfigure the network than by doing this? 2024-07-06 09:33:20 It's only a "* WARNING" because it is started at boottime, I do it extly the same way 2024-07-06 10:08:27 ok, but what does that warning mean? 2024-07-06 10:08:33 What is it warning against? 2024-07-06 10:09:46 there are some services you put in boot runlevel that you may not want to stop/restart 2024-07-06 10:14:42 Hummm 2024-07-06 10:14:51 But it's too late anyway, isn't it? 2024-07-06 10:17:07 well you've done it now, the bed is made 2024-07-06 10:41:04 Hello, is it possible to install Alpine to a disk with PXE and without internet access? 2024-07-06 10:42:07 yes 2024-07-06 10:43:58 Also will Alpine x86 work with a CPU that doesn't have SSE? 2024-07-06 10:45:27 no 2024-07-06 10:45:42 my non-SSE i586 died like that https://pomf2.lain.la/f/ow945n7.jpg 2024-07-06 10:45:44 RIP 2024-07-06 10:46:45 also, alpine had been one of the last linux holdouts with support for machines without SSE 2024-07-06 10:47:31 that is a fine picture 2024-07-06 10:48:19 nero: :'( 2024-07-06 10:50:50 nero: thanks. So I guess only choice is NetBSD :\ 2024-07-06 10:52:04 dpirate2: same for me, but was too lazy 2024-07-06 10:52:25 I wonder if you could get modernish gentoo (even if you had to bootstrap with old gentoo). but that's a long compile nowadays. and a bunch of stuff might not compile at all (rust, apparently...) 2024-07-06 10:53:44 lopid: here, have more https://pomf2.lain.la/f/3w4sc37.jpeg 2024-07-06 10:54:25 when you hilight me, it makes a physical beep 2024-07-06 10:54:48 nice 2024-07-06 10:55:08 how much is it costing you to run irc like that? :) 2024-07-06 10:55:14 the slackware page still says it will run on a 486, though I dunno if that's still accurate 2024-07-06 10:55:32 fission: iirc i saw similar complaints with SSE about slackware, too 2024-07-06 10:55:33 tumbled on this a couple of days ago. have not yet prepared old hardware to test it, though: http://wiki.kolibrios.org/wiki/Main_Page 2024-07-06 10:55:54 the slackware kernel works without SSE, but the userspace doesn't 2024-07-06 10:56:18 yeah. that doesn't surprise me. too many things expect it nowadays 2024-07-06 10:57:18 lopid: didn't measure, i blank it when unused to avoid burn-in 2024-07-06 10:58:59 does plan9 need sse? 2024-07-06 10:59:00 heh 2024-07-06 11:00:04 im surprised how many people seem to use pre-SSE devices 2024-07-06 11:00:25 the SSE question comes up here like twice a week 2024-07-06 11:01:10 me too. I think I only ever had one, a Pentium 75 (P54C) 2024-07-06 11:04:15 i would migrate to an RPi in an instant 2024-07-06 11:05:22 but current alpine (3.20) on my RPi 3B results in a crap taint 2024-07-06 11:05:27 "whats sse2? cant you just compile it?" haha oof 2024-07-06 11:05:48 due to bcm2835 driver 2024-07-06 11:06:19 I watched some video on kolibrios and the guy was like... "it fits on a floppy, but I've never even seen one in real life!" damn I'm old 2024-07-06 11:06:56 nero: that's suboptimal 2024-07-06 11:40:07 haha, I just updated and got the funniest bug... in some fonts ligatures are not confused, e.g. if I type into Firefox address bar "ffl", it shows up as "IL" 2024-07-06 11:40:18 s/not/now/ 2024-07-06 11:40:37 * haha, I just updated on edge and got the funniest bug... in some fonts ligatures are now confused, e.g. if I type into Firefox address bar "ffl", it shows up as "IL" 2024-07-06 11:45:50 Can't make it happen in inkscape, and it also doesn't happen in Element, it does happen in the address bar of the still running Firefox. 2024-07-06 11:48:27 divVerent: this depends on which fonts you have installed and get selected 2024-07-06 11:48:40 yeah, currently playing around with this in element inspector 2024-07-06 11:48:43 the chosen font is Cantarell 2024-07-06 11:48:50 but Cantarell in Inkscape doesn't do it 2024-07-06 11:50:08 except, I don't even have that font 2024-07-06 11:50:11 WAT is firefox doing there 2024-07-06 11:50:44 i remember typing "123456789" in the firefox address bar and i think the "4" was one or two pixels larger 2024-07-06 11:50:55 so not the first time i see firefox doing funny font things 2024-07-06 11:55:58 fc-match sans-serif spits out Noto Sans, but Noto Sans doesn't do it either in GIMP or Inkscape 2024-07-06 11:56:37 and it appears to happen only in elements or the address bar, not on page elements - so yeah, Firefox fault 2024-07-06 11:58:19 nah, happens on page elements too if they're sans-serif ! important 2024-07-06 11:59:22 yes, is active in Noto Sans too, so - indeed Firefox only problem, fun 2024-07-06 12:00:53 fun problem gone after firefox restart; I guess what happens is that Firefox cached something about the ligatures and apk upgrade updated the font? 2024-07-06 12:01:11 and then it reloaded the font but not the ligatures? Weird... 2024-07-06 12:04:30 yeah, the fonts were indeed updated; weird that Firefox can't handle that properly 2024-07-06 12:52:23 divVerent, funny, I saw that exact same problem a couple days ago (fi/fl) 2024-07-06 12:52:39 It depends on the zoom for me 2024-07-06 12:52:56 Sometimes you also get ff 2024-07-06 12:54:03 (well, the zoom behavior part wasn't observed on the address bar obviously) 2024-07-06 16:37:19 tmrlvi: it *used to be* needed on both sides, yes, but they added a simpler way for clients sometime later (it's still needed on servers) 2024-07-06 17:13:38 grawity: When I put idmapd on the server only, I get weird ids on the client side. When I remove it from the server, I get the correct ids on the client side 2024-07-06 17:15:05 could be a side affect of having the same ids (except client root is mapped to nobody) 2024-07-06 18:17:56 hi, I'm trying to build dino package from git master on Alpine edge ARM64 (postmarketOS), but I'm getting error: file Failed to run ldconfig 2024-07-06 18:18:09 Any idea how I could fix it? 2024-07-06 18:18:43 I'm wondering if it's related to Alpine peculiarities or anything 2024-07-06 18:20:48 ldconfig is a glibc thing. Alpine uses musl instead of glibc. 2024-07-06 18:20:55 martin123456: by "dino" you mean https://github.com/dino/dino, right? 2024-07-06 18:21:05 yes 2024-07-06 18:21:23 raven523: we have ldconfig, it's pretty much a stub that pretends it does something 2024-07-06 18:21:41 it's an alpine package too 2024-07-06 18:22:07 to be fair, community/dino doesn't seem to be doing anything special 2024-07-06 18:22:21 but I'm trying t build frm git instead 2024-07-06 18:22:25 yeah; are you using cmake, or `./configure`? 2024-07-06 18:22:44 ./configure 2024-07-06 18:23:05 try installing musl-utils ( which provides /sbin/ldconfig ), if that doesn't help then check if cmake works 2024-07-06 18:23:29 thanks, I'll report back shortly :) 2024-07-06 18:23:58 alternatively just use cmake because all the configure script does is call cmake for you 2024-07-06 18:24:50 so just cmake command with no args? 2024-07-06 18:25:06 cmake -B build -DCMAKE_BUILD_TYPE=Release -G Ninja 2024-07-06 18:25:08 I already had that pkg installed, btw 2024-07-06 18:25:09 ninja -C build 2024-07-06 18:25:20 thank you 2024-07-06 18:27:57 CMake error: coyld not create named generator ninja 2024-07-06 18:28:18 *could 2024-07-06 18:28:44 ptrc, does this error mean anything to you? 2024-07-06 18:29:12 `apk add ninja` 2024-07-06 18:30:17 is there a reason why the Unix Makefiles backend wouldn't work 2024-07-06 18:30:49 it would work 2024-07-06 18:30:53 it's also just 20% slower and sucks 2024-07-06 18:31:05 literally zero reason to use it 2024-07-06 18:34:00 hmm I already had ninja installed. Someone replied over in the dino room saying add --with-libsoup3 toconfigure 2024-07-06 18:42:10 that got me further, but now error: "install path of target requires changing an RPATH from the build tree, but not supported unless on ELF-based or XCOFF-based platform. The CMAKE_BUILD_WITH_INSTALL_RPATH variable may be set to avoid this relinking step 2024-07-06 18:42:29 not sure what that variable is or how to set it 2024-07-06 19:35:41 I got it to build, but now I have new errors when I try to execute 2024-07-06 19:35:59 command I used to build: cmake -B build -DCMAKE_BUILD_TYPE=Release -DCMAKE_BUILD_WITH_INSTALL_RPATH=ON -G Ninja 2024-07-06 19:36:47 martin123456: https://tpaste.us/RBmO 2024-07-06 19:37:00 These are the errors when I run the executable: 2024-07-06 19:37:00 Error relocating build/dino: dino_avatar_manager_get_avatar_file: symbol not found 2024-07-06 19:37:00 Error relocating build/dino: xmpp_xep_consistent_color_string_to_rgb: symbol not found 2024-07-06 19:37:03 Error relocating build/dino: xmpp_xep_consistent_color_string_to_rgbf: symbol not found 2024-07-06 19:37:06 Error relocating build/dino: dino_model_conversation_display_name_get_type: symbol not found 2024-07-06 19:37:09 Error relocating build/dino: dino_plugins_encryption_preferences_entry_get_widget: symbol not found 2024-07-06 19:37:12 Error relocating build/dino: dino_contact_models_get_display_name_model: symbol not found 2024-07-06 19:37:16 Error relocating build/dino: dino_weak_timeout_add_seconds_once: symbol not found 2024-07-06 19:37:19 Error relocating build/dino: dino_avatar_manager_unset_avatar: symbol not found 2024-07-06 19:37:22 Error relocating build/dino: dino_contact_models_get_type: symbol not found 2024-07-06 19:37:25 Error relocating build/dino: dino_contact_models_IDENTITY: symbol not found 2024-07-06 19:40:47 j_v, thanks. I tried the command in that file, but it gave me errors about changing RPATH 2024-07-06 19:42:29 martin123456: using abuild rootbld, it built fine (though checks fail), so could be something in your build environment, dunno 2024-07-06 19:58:47 j_v what is abuild rootbld? 2024-07-06 20:03:27 heh, wasmtime-dev installs /usr/include/wasmtime/conf.h.in rather than /usr/include/wasmtime/conf.h, oops 2024-07-06 20:08:10 I wonder why it does that... 2024-07-06 20:08:45 maybe the new C standard has a second preprocessor ;) 2024-07-06 20:08:49 j_v I got dino running, had to make install, and those errors went away 2024-07-06 20:09:34 oh, conf.h is generated by the cmake file but we directly run `cargo build` instead 2024-07-06 20:12:53 that's annoying, not sure how to fix that 2024-07-06 20:24:02 i should have suggested him to google LD_LIBRARY_PATH, since that was obviously his issue. oh well, he's gone 2024-07-06 20:56:41 elly: i outsourced a fix 2024-07-06 20:56:51 will be pushed to edge hopefully still today 2024-07-06 20:56:53 :p 2024-07-06 20:57:09 ( and by outsourced i mean nerdsniped someone else to fix it for me ) 2024-07-06 21:00:20 that's the way to do it tbh 2024-07-06 21:04:27 s/the/one/ 2024-07-06 21:49:20 g 2024-07-06 22:08:47 I noticed the ladybird package was deleted from apk. Does anybody have a work in progress? See 2024-07-06 22:17:30 it was over a year ago 2024-07-06 22:18:02 and it's still borderline unusable for practical purposes 2024-07-06 22:45:16 Hi, I noticed that system-wide dbus-daemon can and will autostart services as root on my machine. I have waydroid installed and running it as non-root user will prompt D-Bus to start `id.waydro.Container.service` with no confirmation or password necessary. Is this the expected behaviour? Regardless if it is expected, is there a way to prevent this? 2024-07-07 01:19:39 For alpine 3.20.1 I noticed that /usr/lib/libreadline.so.8 and /bin/bash where corrupted. Any thoughts on what might have gone wrong? The rest of the OS seems fine. apk fix solved the isues. 2024-07-07 01:24:17 Corrupted file system 2024-07-07 07:25:05 ashie: you can delete the priviledged helper that starts services: `apk add !dbus-daemon-launch-helper` 2024-07-07 09:36:24 Can I have permissions to delete wiki pages? I'd like to delete some of the obvious candidates like https://wiki.alpinelinux.org/wiki/MIPS64, https://wiki.alpinelinux.org/wiki/EyeOS or https://wiki.alpinelinux.org/wiki/TODO:py3_packages 2024-07-07 09:36:42 A lot of empty pages show up in https://wiki.alpinelinux.org/wiki/Category:Articles_for_deletion 2024-07-07 14:07:37 Don't suppose there are any Thinkpad P14 or T14 Gen 3+ AMD users here? Switched over and mostly things work fine but sleep/hibernate is turning into a real pain. Any tips? 2024-07-07 14:09:57 got a T14 G3 here, but with Arch Linux atm. but had some sleep issues aswell 2024-07-07 14:20:12 this page was last edited over ten years ago https://wiki.alpinelinux.org/wiki/APKBUILD_examples:Ruby 2024-07-07 14:33:24 WhyNotHugo: thanks, appreciate it 2024-07-07 14:35:29 fossdd[m]: yea, it's an issue unfortunately. 2024-07-07 14:37:11 side note, I find it funny how often I come across you when trying to solve some of my issues, for example when searching on how to solve this I saw your issue about autostarting on the dbus repo 2024-07-07 14:37:20 ^.^ 2024-07-07 18:49:22 good evening 2024-07-07 18:49:33 I was wondering if anyone else has mesa upgrade issues? 2024-07-07 18:49:59 When I run apk upgrade -a, I get: mesa-21.3.8-r1: breaks: mesa-gles-24.1.2-r1[mesa=24.1.2-r1] 2024-07-07 18:52:45 have you got a pinned mesa version in your world 2024-07-07 18:52:50 check /etc/apk/world 2024-07-07 18:53:34 in world is only "mesa-dri-intel" 2024-07-07 18:54:23 uh, I just noticed the version difference is massive 2024-07-07 18:54:30 idk maybe somebody else does 2024-07-07 18:57:01 apk list '*mesa*' doesn't show 'mesa-dri-intel' on edge, so I'm thinking mesa-dri-intel was removed as mesa subpkg a while back. 2024-07-07 18:57:59 i seem to remember that mesa was deprecating a bunch of dri drivers in favor of gallium, not sure how long ago 2024-07-07 19:03:44 ok, deleting that one makes apk upgrade work 2024-07-07 21:11:43 can anyone help us with the LKML? 2024-07-07 21:34:13 Soni, help with what about the LKML? 2024-07-07 21:34:27 oh, https://lkml.org/lkml/last100/ is messed up again 2024-07-07 21:34:33 i've passed the message on 2024-07-07 21:34:37 if that is not what you meant, please say so 2024-07-07 21:35:37 we want to define a .sigalttext for, well, objects, so we can implement a "signal sanitizer" in the compiler 2024-07-07 21:36:48 the idea being that you can't call signal-unsafe code if said signal-unsafe code doesn't exist, which basically means swapping out .text for .sigalttext when running signals 2024-07-07 21:37:02 what does that have to do with LKML? 2024-07-07 21:37:22 allegedly the object format is defined by linux 2024-07-07 21:37:31 go on 2024-07-07 21:37:53 well, this would require changes to the object format, no? 2024-07-07 21:39:44 i'm just trying to figure out what you want from LKML 2024-07-07 21:40:09 did you perhaps never mean lkml.org when you said LKML? 2024-07-07 21:40:42 linux kernel mailing list? 2024-07-07 21:41:08 yes 2024-07-07 21:41:35 allegedly such changes have to go through the mailing list? 2024-07-07 21:41:49 (instead of being pushed into distros?) 2024-07-07 21:42:05 alleged by whom? who are you asking? 2024-07-07 21:42:08 so much context is missing 2024-07-07 21:42:33 well we would like help posting this idea to the mailing list 2024-07-07 21:42:53 (indeed it would be preferable for distros to carry fewer patches, not more) 2024-07-07 23:04:13 Hmm why is yt-dlp installed and symlinked as youtube-dl when those two applications don't have feature or cli switch parity? 2024-07-07 23:05:03 so programs who were still hardcodes to youtube-dl would use yt-dlp 2024-07-07 23:05:12 usecase: watching youtube with mpv 2024-07-07 23:05:38 so just hoping they use almost no command line switches then 2024-07-07 23:08:10 no? they are largely the same 2024-07-07 23:08:30 yt-dlp and youtube-dl are branches of the same codebase 2024-07-07 23:08:54 as of years ago, and don't keep parity with each other. 2024-07-07 23:09:18 and youtube-dl is still actively maintained 2024-07-07 23:10:49 by one guy who is porting stuff from yt-dlp 2024-07-07 23:10:57 yt-dlp is where the devs are 2024-07-07 23:12:27 feel free to continue with youtube-dl, i had to switch to yt-dlp because some sites didn't work otherwise 2024-07-07 23:13:03 youtube-dl if you install it also just gives you yt-dlp which is also pretty fun. 2024-07-07 23:27:18 iirc MPV also checks for yt-dlp 2024-07-07 23:28:18 but yeah yt-dlp is more well-maintained nowadays afaik 2024-07-07 23:32:45 xulfer: the command-line interface is mostly the same 2024-07-07 23:32:48 not exactly, but mostly 2024-07-07 23:33:49 https://github.com/yt-dlp/yt-dlp?tab=readme-ov-file#differences-in-default-behavior documents these 2024-07-08 08:55:18 Hi 2024-07-08 08:56:02 What is the state of OpenRC service supervision? Seems most services still don't use it. Why? 2024-07-08 08:56:35 https://github.com/OpenRC/openrc/blob/master/supervise-daemon-guide.md 2024-07-08 09:09:27 unixery: Because if you need to assume your service dies, it's better to fix your service. 2024-07-08 09:09:56 unixery: Well, at least that would be the reason the OpenBSD guys would bring up at that point... 2024-07-08 09:11:50 Actually yes, but... :-D 2024-07-08 09:13:43 On Illumos SMF can not only restart services but even send an e-mail/snmp if a service fails. Which I think is quite nice. 2024-07-08 09:14:43 unixery: That's conceptually nice, yeah, but technically a 3 line wrapper script in BASH if you really need that. 2024-07-08 09:15:19 unixery: Still, ask yourself what makes a service fail, it's still better to never receive such an email. 2024-07-08 09:18:44 unixery: Of course we could go all in like concurrent init systems and make OpenRC send email, too, but I guess you see where that's going. ;-) 2024-07-08 09:26:17 unixery: The way described on https://wiki.gentoo.org/wiki/OpenRC/supervise-daemon looks like you can utilize supervise-daemon for this to introduce "supervision" to a service, though, I'm not sure if that's "the state of supervision", though. :-D 2024-07-08 09:26:23 "unixery: Because if you need..." <- So how do I fix my service so it never gets OOM-killed just because a logged in user is playing with malloc while the service is receiving a sudden load spike? 2024-07-08 09:26:44 The kernel OOM killer is far from reliable in finding the right "culprit", it ultimately is guesswork anyway. 2024-07-08 09:26:48 divVerent: By not exceeding your RAM limits. 2024-07-08 09:27:10 divVerent: If you reach the state where OOM killer comes and randomly kills your processes, you have a very different problem. 2024-07-08 09:27:18 I mean, a regular used should not be able to kill a system wide service, ever. 2024-07-08 09:27:30 Linux ulimit/setrlimit is too weak to ensure that. 2024-07-08 09:28:32 divVerent: Well one could say that the system administrator didn't set up limitations properly, but again, if you fear that your program eats up all your RAM, maybe fix your program? 2024-07-08 09:28:55 armin: Arguably the problem is the user, or the Java software the user decided to run... e.g. I got an OOM for loading a too big binary into Ghidra. 2024-07-08 09:29:24 But my point is, a non-root user should never be allowed to kill a root-owned service by their actions. 2024-07-08 09:29:48 armin: How do I set up those limits then? 2024-07-08 09:29:57 divVerent: If they are, OOM-killer has a weird scoring system. 2024-07-08 09:30:14 divVerent: bubblewrap, containers, whatever 2024-07-08 09:30:23 It usually gets the right task but not always. 2024-07-08 09:30:41 So how do I put an entire user account into a container on Alpine? 2024-07-08 09:30:54 divVerent: No clue, I would use FreeBSD and jails at that point. 2024-07-08 09:31:13 I had ONE case in my life where the OOM killer killed sshd by accident. 2024-07-08 09:31:31 divVerent: That should never happen and I consider that a mistake. 2024-07-08 09:31:40 How it happened? Accidental memory eater running as user, AND a spike of ssh login attempts by those password guessing bots. 2024-07-08 09:32:15 So ssh was the one allocating last. Usually the OOM killer has scoring to avoid that, and usually it works... this was an exception. 2024-07-08 09:32:36 The scoring is heuristics though and not guaranteed to always do the right thing, it just gets close. 2024-07-08 09:33:18 armin: Yeah, a mistake by the scoring. Happens. Cannot guarantee 100% this never happens. 2024-07-08 09:33:41 Now arguably a supervisor daemon could be hit by the same. 2024-07-08 09:33:51 So the question arises, who supervises the supervisor. 2024-07-08 09:34:02 PID EINS! 2024-07-08 09:34:20 The supervisor can operate without allocating RAM though. 2024-07-08 09:35:05 So I wonder, if a process works like one of these supervisors, basically a fork, exec, wait loop... 2024-07-08 09:35:06 would mlockall() be enough to ensure the OOM killer will never consider it? 2024-07-08 09:35:30 And yes, I know processes running as root can "adjust" their own OOM score. 2024-07-08 09:35:46 But that too just changes probabilities, provides no guarantees. 2024-07-08 09:35:59 divVerent: good question. 2024-07-08 09:37:24 So that basically is my case for supervisor daemons - now ideally Linux should provide better protections than that, like, strict stacked OOM priorities. 2024-07-08 09:38:00 Containers can also set total resource limit which is nice, but Alpine provides no way to run user accounts in one. 2024-07-08 09:38:08 systemd actually does, I think... 2024-07-08 09:39:20 And it MIGHT be possible to do it on Alpine with a special login shell, but that is... voluntary again and thus nothing a sysadmin can rely on. 2024-07-08 09:39:51 push him into a cgroup? 2024-07-08 09:40:01 cgred daemon might help 2024-07-08 09:40:05 yeah, that is the idea 2024-07-08 09:40:43 https://kb.brightcomputing.com/knowledge-base/how-do-i-set-per-user-resource-limits-with-cgroups/ 2024-07-08 09:43:40 Yeah, while a systemd unit file is a nice abstraction of that kernel feature, you of course can totally use that without systemd, too, so that's probably reasonable. 2024-07-08 09:44:07 Not sure if "bubblewrap" would be a good alternative here. 2024-07-08 09:46:08 cgred... is that polling processes and moving them to cgroups every few sec? That might actually work... 2024-07-08 09:46:49 but that too is dirty as it relies on cgred seeing the process soon enough 2024-07-08 09:47:25 wonder if that can somehow integrate with PAM so login does not finish before the process is in its cgroup. 2024-07-08 09:48:22 ah, there is pam_cgroup 2024-07-08 09:49:17 with that it can actually catch 100% of user processes, maybe (not 100% sure about cron or at invoked processes) 2024-07-08 09:51:16 also some fun netlink hackery in cgred so it does not have to scan procfs so much 2024-07-08 09:55:23 it would be nice to have a kernel module for that 2024-07-08 09:55:42 And move things based on rules 2024-07-08 09:55:55 ebpf lol :DDDDD 2024-07-08 10:46:35 yeah, technically it allows that ;) 2024-07-08 10:47:59 Till then we do need supervisors. Or rather we always need them, just they need not be local daemons. They can be remote servers that try to ssh in, and if they fail N times in sequence, they actuate a reboot externally. 2024-07-08 10:47:59 Because those are needed anyway and be it just to handle HW failures. 2024-07-08 10:50:02 wha 2024-07-08 10:50:35 hw supervisors are kinda orthogonal, no? 2024-07-08 10:50:51 hw supervisor basically makes sure your kernel is responsive 2024-07-08 10:51:14 it does not care about cgroups 2024-07-08 11:09:01 IIRC OOM killer would first kill the new memory eating process, not long running processes 2024-07-08 11:09:39 but supervision won't help here anyway 2024-07-08 11:10:12 supervision helps for stuff that is directly related to the service 2024-07-08 11:11:37 a bug in the service that makes it crash once a month. fixing that bug is of course best, but still iprefer automatic restart of the service over doing it by hand 2024-07-08 11:13:07 its a matter of personal taste whether one crash per month is acceptable or not 2024-07-08 11:13:08 unixery: I would much prefer to actually locate the issue and fix it, but that's just me. 2024-07-08 11:13:18 normal amount of crashes is zero 2024-07-08 11:13:57 and if a crash happens, thats better be a hardware fault or an OOM 2024-07-08 11:14:13 unixery: If you just want to re-start a service when it dies and go all yolo, installing runit on top of OpenRC I remember to work nicely. 2024-07-08 11:19:35 armin: openrc already has supervise-daemon 2024-07-08 11:22:59 that's why I've asked what the state of it is ;-) 2024-07-08 11:23:03 unixery: Correct. So what do you suggest? Using it for *all* service files in Alpine? :) 2024-07-08 11:23:11 ok :D 2024-07-08 11:23:36 e.g. if it is considered unstable or production ready 2024-07-08 11:23:45 ah! :) 2024-07-08 11:24:36 and if it is stable, why not use it? the only real drawback I can see are crashing supervise-daemons 2024-07-08 11:24:51 it works fairly well 2024-07-08 11:25:16 who is this? https://codeberg.org/alpinelinux 2024-07-08 11:25:18 would love to see chimera's dinit however, it looks cleaner 2024-07-08 11:26:06 this actually came to my mind again, because I've looked at Chimera and dinit :-D 2024-07-08 11:28:19 + user services 2024-07-08 11:28:23 all in all a win 2024-07-08 11:52:48 unixery: I've tried to install Chimera with full disk encryption but so far I had no success - without disk encryption was easy, though 2024-07-08 14:04:58 Hey! I'm upgrading a Alpine 3.16 with docker and docker-compose (from community repo) to an alpine 3.20. When executing 'apk upgrade --available' my docker-compose is purged but never reinstalled. If I execute a second time 'apk upgrade --available' my docker-compose is correctly reinstalled... Is it a bug? 2024-07-08 14:31:15 Cyanat: it should have installed docker-cli-compose on the first upgrade, which replaces old docker-compose 2024-07-08 14:32:01 if it didn't, then it's probably some kind of bug 2024-07-08 14:32:07 but i can't reproduce it with just docker-compose 2024-07-08 14:39:54 ptrc: Thank you for your response. Well, in that case this bug is probably linked to a prerequisite package. 2024-07-08 14:40:25 For info, here are the different step of my upgrade relative to docker: 2024-07-08 14:40:32 (25/158) Purging docker-compose (1.29.2-r2) 2024-07-08 14:40:37 (33/158) Purging py3-dockerpty (0.4.1-r4) 2024-07-08 14:40:42 (46/158) Purging docker-py (5.0.3-r1) 2024-07-08 14:40:47 (47/158) Purging dockerpy-creds (0.4.0-r3) 2024-07-08 14:40:52 (104/158) Upgrading docker-engine (20.10.20-r0 -> 26.1.3-r1) 2024-07-08 14:40:56 (105/158) Upgrading docker-openrc (20.10.20-r0 -> 26.1.3-r1) 2024-07-08 14:41:01 (106/158) Upgrading docker-cli (20.10.20-r0 -> 26.1.3-r1) 2024-07-08 14:41:06 (107/158) Installing docker-cli-buildx (0.14.0-r2) 2024-07-08 14:41:11 (108/158) Upgrading docker (20.10.20-r0 -> 26.1.3-r1) 2024-07-08 14:42:03 Anyway, it's not a big deal... If I have to apk upgrade two time, then will do that. 2024-07-08 15:25:21 hi all 2024-07-08 15:25:46 i just installed tigervnc here in alpine 3.20 2024-07-08 15:26:24 it was ok, but cannot find the program 2024-07-08 15:26:48 is it working with alpine 3.20? 2024-07-08 15:27:29 csena2k2, depends on the program you're looking for, but most likely that'd be "vncviewer" 2024-07-08 15:29:23 Hummm, that's from tigervnc-client actually, tigervnc is the server, so you get vncserver there (and other programs that you can list with "apk info -L tigervnc") 2024-07-08 15:30:11 if i use lbu commit with "-e" what happens after reboot ? ask for password ? should i add PASSWORD=xx in lbu.conf to boot without asking ? 2024-07-08 15:42:48 thanks for the help 2024-07-08 15:43:10 i will try it here 2024-07-08 15:47:07 installed tigervnc-client and it works 2024-07-08 15:51:40 csena2k2, perfect :) 2024-07-08 23:12:46 hi :) i bought a sbc, radxa zero 3e, and would like to get alpine running on it. where do i start in order to get the proper u-boot and .dtb working. i've tested armbian and it works fine, but i prefer alpine 2024-07-08 23:39:00 Is anyone daily driving a RISC V SBC with Alpine? I'm interested in buying one, so would love to hear your experience. 2024-07-09 01:00:50 jahway603[m]: drew devault has been using one for a hot minute. https://drewdevault.com/2022/01/15/2022-01-15-The-RISC-V-experience.html 2024-07-09 01:01:20 His description makes it sound pretty viable to me 2024-07-09 01:54:29 note that that article was published more than two years ago 2024-07-09 01:54:35 things have probably changed a lot since then 2024-07-09 02:03:08 that's a long minute 2024-07-09 06:24:49 jahway603[m]: #alpine-riscv64 might be a good place to ask as well 2024-07-09 09:26:34 Hi! I use python-iptables to manage some iptables rules. However, it seems that python-iptables is compatible only with iptables legacy while alpine 3.20 use now by default nf_tables. Is it possible to configure my alpine to force use of iptables-legacy instead of nf_tables? 2024-07-09 11:09:29 you get the binaries with the iptables-legacy package 2024-07-09 11:10:06 quick question, what happens when two repositories have the same package? for example, i wanna make my own remote repo that "overwrites" some packages from the main repos, could i do that? 2024-07-09 11:10:49 Cyanat: i suppose you can create the symlinks to point to legacy manually, but it may need to be re-done if iptables packages are updated 2024-07-09 11:12:18 rdbo: apk will prefer the newes package. if they have same version, I *think* it will pick it from the repo listed first in /etc/apk/repositories (but I may be wrong) 2024-07-09 11:12:53 oh, that makes sense. thank you! 2024-07-09 11:13:22 rdbo: https://wiki.alpinelinux.org/wiki/Alpine_Package_Keeper#Repository_pinning 2024-07-09 11:17:07 we want KPTI for signal handlers and nobody's interested in extending ELF for that, what do we do? 2024-07-09 11:17:59 oh that's cool as well clandmeter, didn't know about this 2024-07-09 11:37:25 ("secure distro", doesn't even have signal handler page table isolation...) 2024-07-09 11:46:14 "("secure distro", doesn't even..." <- damn it is security oriented 2024-07-09 12:05:01 we know alpine has signal-safe malloc but why not also have SHPTI on top of it? 2024-07-09 12:13:39 "SHPTI" doesn't seem like an initialism used by anyone, and "signal handler page table isolation" returns 0 results in any search engine - what do you mean? 2024-07-09 12:14:02 i was able to find https://www.kernel.org/doc/html/next/x86/pti.html 2024-07-09 12:14:22 and that just looks like a single option in the kernel config, but it doesn't mention signal handlers anywhere 2024-07-09 12:18:45 what is SHPTI? 2024-07-09 13:04:00 https://discourse.llvm.org/t/signal-sanitizer-fsanitize-signals/79991 2024-07-09 13:04:22 can't call signal-unsafe functions if they don't exist in the address space ^^ 2024-07-09 13:05:33 (this turns RCE issues into DoS issues, which is allegedly better than having RCE issues.) 2024-07-09 13:07:50 (tho we mean... we don't know, what do y'all prefer, RCE or DoS?) 2024-07-09 13:23:45 unfortunately we don't know if anyone here understands the concept of page tables well enough to get how the idea of "mapping things out" can improve security (and turn RCE into DoS), it's a pretty low-level concept after all... 2024-07-09 13:27:09 do you need to be so patronising 2024-07-09 13:27:52 triallax: we're telling on ourselves - we don't know how to explain the idea to others 2024-07-09 13:28:33 we? 2024-07-09 13:28:55 system 2024-07-09 13:30:06 we got exactly one person to understand the idea. haven't been able to replicate it since. 2024-07-09 13:30:49 just dont put code in aignal hanlders 2024-07-09 13:30:56 they are cursed 2024-07-09 13:31:19 can't put code in signal handlers if code is mapped out when running signal handlers :p 2024-07-09 13:31:50 well, it will still crash 2024-07-09 13:32:09 certainly, would you prefer an RCE instead? 2024-07-09 13:32:24 i would prefer no code in signal handlers 2024-07-09 13:32:34 jus dont 2024-07-09 13:32:43 leave it alone 2024-07-09 13:32:54 set a flag and go do somewhere else 2024-07-09 13:33:09 JVMs use signal handlers for hardware-accelerated null pointer exceptions 2024-07-09 13:33:37 my point exactly 2024-07-09 13:33:40 cursed 2024-07-09 13:34:53 however mapping out stdlib sounds like fun :DDD 2024-07-09 13:35:22 consus: for better or worse that isn't enforced 2024-07-09 13:35:44 That C, nothing enforced here 2024-07-09 13:35:52 I'm aware 2024-07-09 13:35:57 cursed language 2024-07-09 13:36:21 metaprogramming through memoey corruption was a mistake 2024-07-09 13:37:22 given how slow ssh is its probably doable to just implement it in go 2024-07-09 13:37:24 but yeah like, just... stick ppl's code in a .sigunsafetext ELF section that gets unmapped in signal handlers 2024-07-09 13:37:54 if that makes any sense 2024-07-09 13:39:24 given the amount it probably better to put signal handlers in a specific section 2024-07-09 13:39:39 and unmap everyting else 2024-07-09 13:39:58 well the original idea is a signal sanitizer, i.e. make this opt-in 2024-07-09 13:40:11 (except you set the whole stdlib to opt-in, system-wide) 2024-07-09 13:40:31 but this leaves other libs 2024-07-09 13:40:51 they'll usually call into stdlib sooner or later 2024-07-09 13:41:06 true, but this may cause further havoc 2024-07-09 13:41:15 and you can roast rust with this, while opt-out makes it hard to roast rust 2024-07-09 13:41:26 also true 2024-07-09 13:42:14 ("alpine linux has safe signals in pure C" would make some tasty roasted crabs) 2024-07-09 13:44:14 its easy enough, just ask rusturbators about any complex data structure 2024-07-09 13:44:34 and see how people struggle with pinned memory and phantom data 2024-07-09 13:45:23 pinned memory and phantom data are easy... 2024-07-09 13:45:39 as a concept? yes 2024-07-09 13:45:48 as in practical code? lol no 2024-07-09 13:45:55 rust can't do loadable libraries 2024-07-09 13:46:00 (at all) 2024-07-09 13:46:00 well it can 2024-07-09 13:46:04 with c api 2024-07-09 13:46:10 it's completely incompatible with the rust memory model 2024-07-09 13:46:21 yep 2024-07-09 13:46:23 also we should take this to an offtopic channel or something 2024-07-09 13:46:29 probably 2024-07-09 13:46:40 ##rust? :DDD 2024-07-09 13:47:16 lmao 2024-07-09 13:47:36 let's just get back to getting alpine linux developers to extend ELF, that seems more exciting 2024-07-09 13:49:04 (and mmap, we guess) 2024-07-09 13:49:31 (or mprotect? something like that) 2024-07-09 14:07:01 ncopa: Thanks for your response regarding iptables legacy. Indeed, I was trying to use alias so "iptables" command will target "iptables-legacy" but it seems quite dirty and I'm afraid to pay the bill later... But, I suppose there is no better solution for now. 2024-07-09 14:08:14 `iptables` is just a symlink 2024-07-09 14:08:37 maybe you could change the python code to use the iptables-legacy binary 2024-07-09 14:08:46 or even better, fit it to work with the nft backend 2024-07-09 14:45:30 the python code does not use the binary, i checked 2024-07-09 14:46:39 ncopa, unfortunately, python-iptables doesn't use binary but C libraries of iptables legacy (see https://github.com/ldx/python-iptables/issues/319#issuecomment-796320521), and I don't find any alternative to python-iptables. So, I guess you're right, either I modify my script to manage rules manually, either I wait for an update of python-iptables to work with nf_table :) 2024-07-09 14:54:22 > getting alpine linux developers to extend ELF 2024-07-09 14:54:24 imsorry what 2024-07-09 15:44:04 Can one create like a script directly on the commandline using ; ? 2024-07-09 15:44:26 this is whe line i tried to run 2024-07-09 15:44:30 while true do ; ping 8.8.8.8 -c 1 -i 0.3 -OU ; sleep 0.7 ; done 2024-07-09 15:45:05 but there are something i don't know here, what should it look like? 2024-07-09 15:45:54 why upstream things when you can shove them in distros instead 2024-07-09 15:46:22 isn't that what DVCS is for 2024-07-09 15:46:29 Babba: while true ; do ping 8.8.8.8 -c 1 -i 0.3 -OU ; sleep 0.7 ; done 2024-07-09 15:47:14 ah, major thanks mid :) 2024-07-09 15:57:45 Nice, Firefox -1 is out! 2024-07-09 16:01:56 yup, now time to fix all the issues that somehow didn't happen with developer edition :) 2024-07-09 16:06:14 u guys now a simple initramfs generator with live support? i was looking at booster and dracut, but i'm not sold on either yet. i had made my own also but handling kernel modules by hand can be a pain sometimes XD 2024-07-09 16:43:01 rdbo: alpine uses mkinitfs 2024-07-09 16:44:24 ptrc: the expectation that developers should work on addons on a version entirely different from their audience is absurd. 2024-07-09 16:45:20 eeh, it's only one version apart, and i'd say it's mostly so that they can test ahead with versions that are not yet released 2024-07-09 18:27:41 close 2024-07-09 18:58:42 ikke: Hi, are you there? 2024-07-09 19:00:43 In limited capacity 2024-07-09 19:00:58 can you force build https://pkgs.alpinelinux.org/package/v3.20/main/x86_64/lxc 2024-07-09 19:01:29 it seams that it has not been build with the patch 2024-05-13 17:36:15 2024-07-09 19:02:19 i think the lxc-top-cgroupv2.patch is missing in that build 2024-07-09 19:03:17 Or, is it only ncopa that can trigger a build? 2024-07-09 19:06:22 hmm, looks that the patch was applied. but lxc-top does not work in 3.20 -- lxc-top-cgroupv2.patch: OK 2024-07-09 19:17:35 does not work in 3.19 2024-07-09 22:05:39 hey, can anyone tell me what license the files in apports are, specifically the patches for the nose python package? 2024-07-09 22:05:58 s/license the/license for the/ 2024-07-09 22:07:36 ncopa: ^? 2024-07-09 22:13:00 Patch 2024-07-09 22:24:46 wdym 2024-07-09 22:36:48 Apachez, i guess you have your resons, IRL or something else for tbg.nu not being up, if you would ever find the desire and time to get it back up again it would be pleasant for atleast me. either way i'm happy that it has been in existence for many years :) 2024-07-09 22:59:51 Babba: it will return... soon... 2024-07-09 23:00:05 vacation starts on monday then I should have some time to fix the last bits :) 2024-07-09 23:00:23 unfortunately work got overhand 2024-07-09 23:01:31 yea, work always come first :) 2024-07-09 23:02:13 a thing i often says while playing games :) 2024-07-09 23:02:19 :) 2024-07-09 23:21:01 hexa-: afaict there's no license on aports, but then, patches to LGPL software are probably LGPL themselves? 2024-07-09 23:21:29 no license? oO 2024-07-09 23:22:31 i mean, there isn't any license in the repo itself, so... idunno 2024-07-09 23:22:42 https://gitlab.alpinelinux.org/alpine/aports/-/issues/9074 2024-07-09 23:22:58 apparently this was brought up but just never really resolved 2024-07-09 23:25:55 yeah uh, there's also the case of licensing APKBUILDs/.initd files vs licensing patches ( which i'm pretty sure have to follow some rules wrt license of upstream project ) 2024-07-09 23:27:30 but then, i've had my own patches to aports just copied into other distributions without any attribution, and i don't really mind; don't think many people would 2024-07-09 23:28:02 s/patches to aports/patches to packages in aports/ 2024-07-09 23:29:31 anyhow, would be nice if someone opened a tsc issue about it, since i don't think it can be resolved otherwise 2024-07-09 23:29:34 good point, the patches would be a derivative work 2024-07-09 23:30:56 but if it came to actually adopting a license for aports, we'd just wait forever for all contributors to sign off their contributions under the chosen license, and quite a lot of them just abandoned their packages 2024-07-09 23:31:40 i don't use alpine obviously, but i do think this is a pretty important matter to resolve 2024-07-09 23:31:55 not having a license is legally tenuous ground 2024-07-09 23:32:27 (not a lawyer but) i think in case of no license it's just an "all rights reserved" moment? 2024-07-09 23:32:47 i'm not a lawyer either, but i think so 2024-07-09 23:34:05 and maybe i'm mildly concerned because i sometimes copy patches from alpine :D 2024-07-09 23:34:30 pretty sure distros copy patches between each other all the time 2024-07-09 23:34:36 yea 2024-07-09 23:34:44 for alpine it is worse, since it is the most prominent source for musl compat 2024-07-09 23:35:39 eeh, i'd say we steal as many patches from gentoo/chimera/adelie as they do from us :p 2024-07-09 23:36:02 ( and void, which i always forget also does some musl stuff ) 2024-07-09 23:36:13 (( and also nixos kinda?? seen people experiment with it) )) 2024-07-09 23:36:14 hm i practically look only at alpine at this point 2024-07-09 23:36:26 and maybe void sometimes 2024-07-09 23:36:52 but also, most of my patches were yoinked by arch and they've been non-musl-specific 2024-07-09 23:36:55 just general fixes 2024-07-09 23:37:05 yea i meant for musl patches 2024-07-09 23:40:14 but then yeah, would be nice to clear at least the patch licensing issue - do they follow the project's license (should when it's a copyleft license), or are they also all-rights-reserved 2024-07-09 23:40:40 they don't have to follow the project's license even if it's copyleft 2024-07-09 23:40:42 armin@subwave ~/aports/community/py3-nose $ grep license APKBUILD 2024-07-09 23:40:42 license="LGPL-2.0-or-later" 2024-07-09 23:40:43 armin@subwave ~/aports/community/py3-nose $ 2024-07-09 23:40:48 hexa-: Does this help by any chance? 2024-07-09 23:40:58 a gpl project can incorporate mit code, but the whole thing would be under the gpl 2024-07-09 23:41:06 armin: that is the license of the package, not the patches nor the apkbuild 2024-07-09 23:41:14 ah 2024-07-09 23:41:21 i think they're asking about nose specifically because of the license fiasco 2024-07-09 23:41:30 gotcha 2024-07-09 23:42:10 armin: as was discovered further up, the patch can probably be considered a derivative work 2024-07-09 23:45:01 when someone sends a patch to a project it is implied that the patch is licensed under the same license as the project, no? I think that could apply here 2024-07-09 23:45:19 implied, yes, but legally i have no idea what the case is 2024-07-09 23:45:21 again, not a lawyer 2024-07-09 23:45:48 triallax: tbf i was taught that if you're using the GPL project downstream with some changes, those changes have to be GPL as well 2024-07-09 23:45:59 pretty sure not 2024-07-09 23:46:11 well 2024-07-09 23:46:18 if you're *distributing* the project 2024-07-09 23:46:23 otherwise you couldn't incorporate mit libraries in gpl projects 2024-07-09 23:46:35 ptrc: if you're distributing it, then the whole distribution is under the gpl, yes 2024-07-09 23:46:50 i think there's an exception for linking in GPL? 2024-07-09 23:47:03 as in, that's a boundary between different projects 2024-07-09 23:47:12 gpl doesn't have a linking exception iirc 2024-07-09 23:47:21 it's viral across project boundaries anyway 2024-07-09 23:47:34 well no 2024-07-09 23:47:40 how come 2024-07-09 23:47:41 it's viral one way 2024-07-09 23:47:49 triallax: > implied yes...I think to be legally covered a project could have a rule for contributors saying that to contribute the patch has to be 2024-07-09 23:47:59 the same license as the project 2024-07-09 23:48:06 (pressed enter too soon by accident) 2024-07-09 23:48:15 ptrc: can you elaborate, i don't quite get that 2024-07-09 23:48:16 you can use MIT code in a GPL project because they're compatible but when you're distributing it as a whole, it becomes GPL 2024-07-09 23:48:28 that's more or less what i'm saying 2024-07-09 23:48:29 but nobody does that and project ever just implies it 2024-07-09 23:48:54 s/project/every project/ 2024-07-09 23:48:55 ah, maybe i misunderstood 2024-07-09 23:49:31 what i'm confused by is you're saying that doesn't apply if you were to incorporate code into the project itself, licensing that code under the mit license? 2024-07-09 23:49:35 i don't think there's any issue with that 2024-07-09 23:49:59 just the whole when distributed would be under the gpl, as you say 2024-07-09 23:51:08 i was thinking about shared libraries and how if you have one library using GPL, that doesn't mean its dependencies have to be GPL as well 2024-07-09 23:51:25 as long as their licenses aren't incompatible yea 2024-07-09 23:52:40 but now my mind is riddled with stuff like: is Alpine breaking the license of a GPL project if we release a MIT-licensed project that uses a GPL-licensed library and we didn't check for that 2024-07-09 23:53:21 unsure 2024-07-09 23:53:23 licensing hurts my head 2024-07-09 23:53:32 yeah same 2024-07-09 23:53:57 wouldn't that be on the maintainer of the project? and if the issue arose it should be brought up with them and only afterwards fixed downstream in alpine 2024-07-09 23:54:16 the issue is with distribution 2024-07-09 23:55:21 if they just posted their code on a private gitlab instance and never provided prebuilts then i don't think they'd necessarily be breaking the license 2024-07-09 23:56:31 but if we distribute e.g. cargo-built executables with GPL code inside of them, but the project itself using MIT/BSD/Apache-2.0/etc... 2024-07-09 23:57:56 so the project using a GPL or maybe even proprietary library in code is fine as long as it directly doesn't contain it and does not provide binaries? 2024-07-09 23:58:21 yea 2024-07-09 23:58:28 but people would have to build it privately and not distribute 2024-07-09 23:58:33 i think? 2024-07-09 23:59:04 yeah uhhh 2024-07-10 00:02:56 debian had similar thing: https://lists.debian.org/debian-legal/2023/09/msg00003.html 2024-07-10 00:03:33 and I don't think they solved it, I think it is just best effort to abide by licenses in cases like these and solve individual issues as they are brought up 2024-07-10 00:05:29 and I do not think most authors of the applications care about the licensing of libraries they use and just use whatever is on crates.io, npm, pypi or whatever language relevant registry with no regard for this 2024-07-10 00:05:39 okay it took me less than 5 minutes to find a project in aports that uses GPLv2 and an Apache-2.0 dependency ( which is not compatible ) 2024-07-10 00:05:47 sigh 2024-07-10 00:05:54 -or-later or -only? 2024-07-10 00:06:16 -only 2024-07-10 00:06:20 triallax: has to be -only 2024-07-10 00:06:20 :/ 2024-07-10 00:06:23 -or-later would mean GPLv3 applies as well 2024-07-10 00:06:24 welp, clearly its time to convert alpine to purely source-based distro /s 2024-07-10 00:06:29 because gpl 3 is compatible with apache 2 2024-07-10 00:06:33 that's why i asked 2024-07-10 00:06:33 yeah 2024-07-10 00:06:38 what is it? 2024-07-10 00:06:43 wanna check if it's in chimera too 2024-07-10 00:06:58 found another 2024-07-10 00:07:10 git-branchless and radio-cli 2024-07-10 00:07:14 you might have the former? dunno 2024-07-10 00:07:24 we have the former 2024-07-10 00:07:33 actually 2024-07-10 00:07:33 not the latter 2024-07-10 00:07:35 a common way to circumvent gpl for the linux kernel by companies was to compile their stuff as linux kernel module instead of having the code compile within the kernel 2024-07-10 00:07:41 but are you sure it's an issue 2024-07-10 00:07:45 i think for git-branchless someone messed up the license in APKBUILD lol 2024-07-10 00:07:49 e.g. calling out to a binary shouldn't be an issue 2024-07-10 00:07:49 you get like 1-2% performance penalty but avoided the licensing nightmare 2024-07-10 00:08:02 ptrc: i noticed plenty of incorrect licenses in aports 2024-07-10 00:08:47 eyup 2024-07-10 00:09:07 Apachez: does anyone still bother with that 2024-07-10 00:09:12 except for idk nvidia probably 2024-07-10 00:09:34 if somebody's ever feeling bored they could diff alpine's package licenses against those of another distros and check the discrepancies heh 2024-07-10 00:10:07 could also use repology for that i guess 2024-07-10 00:10:11 https://repology.org/project/git-branchless/information 2024-07-10 00:10:18 probably a good idea anyway 2024-07-10 00:10:24 shows... that others also have GPL-2.0 in their build scripts 2024-07-10 00:10:42 we've got MIT OR Apache-2.0 2024-07-10 00:10:50 which seems to be right 2024-07-10 00:11:06 yeah 2024-07-10 00:11:11 that's also what i fixed ours to 2024-07-10 00:11:14 nice 2024-07-10 00:11:23 nixos has `licenses.gpl2Only` 2024-07-10 00:11:34 arch has `GPL2` 2024-07-10 00:11:35 lol 2024-07-10 00:11:36 can see some cross-distro copying happening there 2024-07-10 00:11:47 might also be that the project *did* in fact have GPL-2.0-only 2024-07-10 00:11:54 but then relicensed ( somehow ) 2024-07-10 00:11:58 i doubt it tbh 2024-07-10 00:12:27 uh, wait 2024-07-10 00:12:32 https://github.com/arxanas/git-branchless/blob/c54d6c895b29935e2d82e833f7b4558f68bdbeaa/CHANGELOG.md?plain=1#L353 2024-07-10 00:12:33 ??? 2024-07-10 00:13:06 yeah but then line 59 ( newer ) 2024-07-10 00:13:11 > git-branchless is now dual-licensed as MIT/Apache 2.0. 2024-07-10 00:13:22 right 2024-07-10 00:13:41 i guess this really was a relicensing thing 2024-07-10 00:13:58 but it doesn't seem like there was any consensus or signing-off from any of the other contributors 2024-07-10 00:14:06 just the project author relicensing 2024-07-10 00:14:07 lol 2024-07-10 00:14:09 epic 2024-07-10 00:14:19 well 2024-07-10 00:14:22 regarding the original question about aports, they could really use a license, just something like MIT or Unlicense, so it is clear that the APKBUILDS are free for the taking 2024-07-10 00:14:26 https://github.com/arxanas/git-branchless/discussions/813 2024-07-10 00:14:26 ah no 2024-07-10 00:14:27 the author does seem to have written the vast majority of the thing 2024-07-10 00:14:29 there was 2024-07-10 00:14:41 great 2024-07-10 00:15:24 ashie: agreed, i'll try to push the topic forward in the following days 2024-07-10 00:15:47 good luck with that 2024-07-10 00:15:52 hope you figure out something 2024-07-10 00:15:56 ptrc: juniper did for their IDP appliances which used linux 2024-07-10 00:16:02 oh yeah it's not gonna be easy 2024-07-10 00:16:12 i'm just not sure how you'll figure it out anyway... 2024-07-10 00:16:17 "and I do not think most authors..." <- In Go this is indeed rather complicated. For AAAAXY I went with integrating go-licenses as part of the build process to embed all licenses of dependencies into the binary. This collection of license files can then be viewed with a command line flag or an in game menu option. 2024-07-10 00:16:17 But if a package changed license to something incompatible with AAAAXY's license, I may or may not notice. 2024-07-10 00:16:21 other than by trying to relicense 2024-07-10 00:16:29 which seems very difficult 2024-07-10 00:16:31 but we can probably use the same strategy of "if somebody contributed a single file then it can be treated as legally insignificant" 2024-07-10 00:17:00 or if the files are simple enough that they can be waved off as too simple to be copyrightable anyway 2024-07-10 00:17:35 you might be forced to drop a few things ig 2024-07-10 00:17:38 Apachez: oh huh, that's fun; i would have thought they'd do open-source, but.. well 2024-07-10 00:17:42 > just the project author relicensing... this is why I don't really look down on contributor license agreements even tho a bunch of people dislike them, it makes things a lot clearer 2024-07-10 00:17:45 if they're substanial contributions and you can't reach the author... 2024-07-10 00:18:13 triallax: if it's a substantial contribution by an author that's been inactive for a long while, it might be even worth dropping, who knows what can of worms it might be 2024-07-10 00:18:18 heh 2024-07-10 00:18:19 true 2024-07-10 00:18:25 but like, what about e.g. firefox 2024-07-10 00:18:34 that's been added ages ago and cant really be dropped 2024-07-10 00:18:37 and if 90% of it was rewritten by active contributors, then.. idk? 2024-07-10 00:19:16 also a good question, what about the ship-of-theseus'd APKBUILDs 2024-07-10 00:19:36 that look nothing like original ones and the blame for the entire file doesn't even contain the original author anymore 2024-07-10 00:19:53 can of worms 2024-07-10 00:20:07 worm of cans 2024-07-10 00:20:15 wan of corms 2024-07-10 00:20:16 anyway 2024-07-10 00:20:31 well, it's something to worry about for future me 2024-07-10 00:20:38 as long as i don't forget 2024-07-10 07:47:48 Hi~ How should we apply to be included in the official mirror list? 2024-07-10 07:59:03 seadog007__: try https://mirrors.alpinelinux.org/ it explain how part way down the page. 2024-07-10 08:00:51 @cyclisme24[m] I sent the email on Tue, Apr 2, never get reply 2024-07-10 08:03:01 hi seadog007__ 2024-07-10 08:03:03 let me check 2024-07-10 08:03:18 <3 2024-07-10 08:03:36 which mirror was this? 2024-07-10 08:03:49 mirror.twds.com.tw 2024-07-10 08:04:33 We are the currently fastest mirror in Taiwan, rest of the mirror have different kind of performance issue :( 2024-07-10 08:05:10 ok found it 2024-07-10 08:05:39 a heads up but we are having issues with our mirrors page. need to debug what is wrong. 2024-07-10 08:07:57 No problem, just remember to add us to the mirror list when it is fixed 2024-07-10 08:07:58 thanks 2024-07-10 08:08:20 your request is in a ticket on gitlab. ill try to push it. 2024-07-10 08:09:06 Great, thanks for your efforts 2024-07-10 09:58:49 hi, I'm trying to build a package using an APKBUILD file in my home dir, do I need to become root to run abuild? 2024-07-10 10:01:59 getting abuild-fetch: /var/cache/distfiles/pcsx2-v1.7.5973.tar.gz.lock: Permission denied 2024-07-10 10:03:51 ah seems just adding my user to the abuild group works 2024-07-10 10:08:04 Yes 2024-07-10 10:34:03 hexa-: see https://gitlab.alpinelinux.org/alpine/aports/-/issues/9074#note_53927 2024-07-10 10:34:56 we had a very long and extended discussion about it :) 2024-07-10 10:35:09 with the conclusion being "licenses are an absolute pain" 2024-07-10 10:35:17 and "we should fix this" 2024-07-10 10:46:46 ncopa: noted, thanks 2024-07-10 10:48:40 ncopa: for the python 3.12 patch, it looks like that code was lifted from the cpython codebase https://github.com/mdmintz/pynose/issues/33 2024-07-10 10:49:01 and the patches seem to based on pynose work? 2024-07-10 10:51:08 hexa-: you do know that the whole of that pynose repo was lifted from the original with attribution etc. removed? 2024-07-10 10:51:51 and also illegally relicensed 2024-07-10 11:00:09 I am aware 2024-07-10 11:00:18 alright, just wanted to confirm 2024-07-10 11:42:16 hexa-: which python 3.12 patch? 2024-07-10 11:42:18 $ ls */*/python*3.12*patch | tpaste 2024-07-10 11:42:18 https://tpaste.us/XyLe 2024-07-10 11:43:03 hexa-: merge requests are welcome 2024-07-10 11:48:30 The one for py3-nose 2024-07-10 11:49:03 ah, ok 2024-07-10 11:49:05 yeah 2024-07-10 11:49:15 looks like it was reported upstream 2024-07-10 11:51:31 py3-nose does not seem to have a maintainer 2024-07-10 11:51:38 maybe its easier to just purge it? 2024-07-10 11:52:24 depends on whether you run tests using nose 2024-07-10 11:52:38 but yeah, unmaintained since ~2015 upstream 2024-07-10 11:52:53 $ grep -w py3-nose */APKBUILD | wc -l 2024-07-10 11:52:53 56 2024-07-10 11:53:13 maybe we can delete everything using py3-nose as well? 2024-07-10 11:53:18 slightly above 150 in nixpkgs :) 2024-07-10 11:53:26 $ grep -w py3-nose */APKBUILD | tpaste 2024-07-10 11:53:26 https://tpaste.us/LRJW 2024-07-10 11:53:42 loads of stuff will disapear 2024-07-10 11:53:59 the whole celery stack does not depend on nose anymore 2024-07-10 11:54:15 billiard, kombu 2024-07-10 11:54:17 but otoh, i think python ppl nowadays prefer pip install anyway, so maybe its not that bad to delete those 2024-07-10 11:54:41 pyproject/pep517 installs actually 2024-07-10 11:54:53 build/installer usually 2024-07-10 11:56:11 python depency managers are like js frameworks at this point 2024-07-10 11:56:56 there are a bunch of packages in testing as well 2024-07-10 11:57:01 also doesnt pip install invoke pyproject/setup.py if the project contains it? 2024-07-10 11:59:39 I think so 2024-07-10 12:01:18 i wonder if i should continue work on upgrading linux kernel today. build fails with latest fortify-headers. or if I should start work on gcc14, or spend rest of the day on py3-nose and licensing issues 2024-07-10 15:43:39 honestly, the only python dependencies I find useful to have in alpine repos are ones that require compiling... everything else we manage via other tools 2024-07-11 00:36:37 python itself tells people to use the OS package manager if not using a venv though I think 2024-07-11 02:13:12 hi, when i plug in ethernet the interface just works, while when i plug in my phone for usb tethering seems like the networking service needs to be restarted 2024-07-11 02:13:46 i think it's because the usb0 interface is only created when the phone is plugged in, while etho is always there 2024-07-11 02:14:08 but how can i achieve "hotplugging" with usb tethering? 2024-07-11 02:26:27 or it might be because of dhcp, but i don't know how to make it run automatically or even manually other then restarting the networking service... i don't think i have a standalone dhcp service, just `inet dhcp` in the config, not sure what it uses 2024-07-11 04:50:38 currently alpine does not have any good ways to do this. you can run ifup usb0 instead of restarting all networking 2024-07-11 04:51:08 and for a slightly more automated solution you can add ifup usb0 to a mdev or udev rule, or use networkmanager or similar 2024-07-11 07:35:36 Hello 2024-07-11 07:41:01 South Korea must be defeated. 2024-07-11 07:46:45 김여정 2024-07-11 07:51:05 heh wat 2024-07-11 07:52:33 we all just go to a hole don't we? buried or burnt to ash it's just the same! This whole effort of living is not worth it. We will all decompose eventually. Forgotten. Unloved. Unheard. 2024-07-11 07:56:41 ok then 2024-07-11 08:28:36 Hi. New to alpine. How would I go about building chromium with some extra patches? I need some patched for onscreen keyboard. 2024-07-11 08:29:16 More precise these two 0001-ozone-wayland-implement-text_input_manager-fixes.patch 0001-ozone-wayland-implement-text_input_manager_v3.patch as found in https://aur.archlinux.org/packages/ungoogled-chromium 2024-07-11 08:29:30 https://git.alpinelinux.org/aports/tree/community/chromium?h=master 2024-07-11 08:32:45 qaqland: so thats the source for chromium, how do I patch it and build a new package? 2024-07-11 08:35:43 add your patch to the source list and run abuild checksum 2024-07-11 08:35:44 https://wiki.alpinelinux.org/wiki/Creating_an_Alpine_package 2024-07-11 08:36:42 see here to setup abuild 2024-07-11 08:39:51 qaqland: thanks, will give it a try... 2024-07-11 08:42:46 JLindeberg: i have never tried compiling webbrowser, wish you good luck :) 2024-07-11 08:45:10 It takes quite some resources and time 2024-07-11 10:20:03 is there something that modprobes devices automatically when they get recognized (i'm in initramfs)? 2024-07-11 10:20:34 i notice that openrc does something called "Loading hardware drivers" or smth like that, ideally i want something like that 2024-07-11 10:22:47 e.g detect and modprobe automatically the ATA controller, video drivers, etc 2024-07-11 10:55:48 nlplug-findsfs will modprobe modaliases that are detected 2024-07-11 11:01:17 oh, interesting, i'll take a look 2024-07-11 14:15:09 see also the hwdrivers openrc script 2024-07-11 14:15:23 oh.. not for initramfs though 2024-07-11 15:13:20 For some odd reason my net-working onboard LAN started to work after upgrading to v3.20.. and my external usb-to-lan stopped working. :| 2024-07-12 00:45:27 test 2024-07-12 00:46:22 test failed successfully! 2024-07-12 02:00:48 zcrayfish, :D 2024-07-12 08:16:24 What's the usual way if one would like to suggest something about how package should be compiled? For example, would love that crystal package would be compiled with interpreter, To send mail directly to Maintainer, or to open a new thread on mailing list or smth else? 2024-07-12 10:40:32 (1) Are there supported alternative methods of installing an Alpine system on "real hardware", other than using the Alpine ISOs? I am aware that I could just use a chroot installation but I want to know how "different" of a system would I end up with 2024-07-12 10:41:47 network booting 2024-07-12 10:42:01 (2) How does the x86_64 ISO boot? Could I make a version of the ISO that uses GRUB? My SeaBIOS gives it a limited environment that cannot switch graphics modes, and I suppose that that's causing issues with the ISO 2024-07-12 10:42:21 Unfortunately my firmware doesn't have ipxe 2024-07-12 10:42:30 No UEFI at all for now 2024-07-12 10:43:40 SeaBIOS is pretty well tested because a) open source and b) used in QEMU, so i doubt thats the problem 2024-07-12 10:44:06 nero: I'm loading SeaBIOS from the GRUB menu provided by Libreboot 2024-07-12 10:45:22 that does not sound right. 2024-07-12 10:47:19 I guess I'll do the chroot for now 2024-07-12 10:47:45 how will you install the chroot on the disk? 2024-07-12 10:48:25 I'll boot with the Arch ISO and bascially do an Arch install, except I'll use apk-static instead of pacstrap 2024-07-12 10:48:30 Well, at least that's what I'm thinking 2024-07-12 10:48:54 if you setup an aplpine chroot in ram, you can use the setup-alpine from there 2024-07-12 10:49:03 and basically use the official installer 2024-07-12 10:49:22 actually that makes sense 2024-07-12 10:49:23 that way you also get the standard disk and bootloader layout 2024-07-12 10:49:43 I need to read a bit about Alpine's standard bootloader layout though 2024-07-12 10:51:00 seabios and syslinux should be no issue 2024-07-12 10:51:31 I probably need grub, because chainloading grub is the most common/supported way to boot from libreboot 2024-07-12 10:55:13 runxiyu: try exporting BOOTLOADER=grub before setup-alpine 2024-07-12 10:55:24 Thanks! 2024-07-12 10:55:46 /sbin/setup-disk (called by setup-alpine) is where your POI is 2024-07-12 11:00:52 (turns out that the arch iso uses a similar isolinux setup and has graphics issues too, I'm contemplating life choices while grabbing Debian live) 2024-07-12 11:03:14 or... it actually gets stuck while loading the initramfs 2024-07-12 11:03:35 I think the fact that it's on USB 2.0 is making me confused because loading things takes such a long time I think it's stalled 2024-07-12 11:25:40 > setup-xorg-base: "is also required for Wayland sessions." "installs xorg-server"... why 2024-07-12 11:29:58 what are you quoting 2024-07-12 11:31:21 there is setup-wayland-base 2024-07-12 11:32:34 ah, i found it: https://wiki.alpinelinux.org/wiki/Alpine_setup_scripts#setup-xorg-base 2024-07-12 11:32:38 "It configures a graphical environment, installing basic Xorg packages and eudev (replacing mdev), and is also required for Wayland sessions. " 2024-07-12 11:32:41 yes, that is wrong. 2024-07-12 11:36:45 runxiyu: did you succeed in installing and booting the base system ? 2024-07-12 11:37:18 nero: Not yet, even Arch doesn't work... I'm going to try Alpine with another flash drive just in case my flash drive is broken (and in that case, sorry for all the hassle) 2024-07-12 11:39:49 ok. nevermind, sorry for bothering for the past hour 2024-07-12 11:43:45 I'm now trying to figure out using the installation medium with WiFi only 2024-07-12 11:44:44 I don't see wpa_supplicant anywhere so this is going to be hard 2024-07-12 11:47:41 hi 2024-07-12 11:50:29 i tried to install Alpine Linux 3.20 on an old laptop (asus eeepc 1101ha, i386 from ~2009), and it failed to install libssl during the disk install 2024-07-12 12:04:55 dok: memtest? 2024-07-12 12:05:21 nero: BOOTLOADER=grub doesn't seem to work 2024-07-12 12:05:34 I've installed my system and I could boot from it in SeaBIOS 2024-07-12 12:06:07 runxiyu: i haven't though of DDR issue, i will run a memtest, i though that might be caused by a bad HDD 2024-07-12 12:11:29 (for my question: whatever, I just uninstalled syslinux, installed grub, and I'm fine now) 2024-07-12 12:11:35 (I'll work WiFi out in a moment) 2024-07-12 12:12:03 dok: Is it deterministic? e.g. if you install multiple times, does it always fail on libssl? 2024-07-12 12:14:06 i dont know 2024-07-12 12:14:43 i've download libssl on another computer (and moved the file with an usb stick, because no wifi) and installed it manually 2024-07-12 12:15:07 you 2024-07-12 12:15:12 you're installing each package manually? wow 2024-07-12 12:15:18 no, only libssl 2024-07-12 12:20:52 Why? 2024-07-12 12:22:46 it was the only missing package 2024-07-12 12:23:00 that failed during the disk install 2024-07-12 12:24:31 the question is why it fails during the disk install 2024-07-12 12:24:41 no clue 2024-07-12 17:02:30 hi 2024-07-12 20:06:02 Hi, does anyone know where I might be able to find documentation on what APKBUILD variables and stages are provided? For example what is `$srcdir`? 2024-07-12 20:08:20 https://man.uex.se/5/APKBUILD#Automatic%20Variables 2024-07-12 20:10:12 Thanks! 2024-07-12 20:11:36 also https://wiki.alpinelinux.org/wiki/APKBUILD_Reference 2024-07-12 20:12:38 and https://wiki.alpinelinux.org/wiki/Creating_an_Alpine_package 2024-07-12 20:12:50 Oh and while I'm here. I found one missing character in one README file https://lucamatei.com/paste/b136b7f3-cd5f-4873-91a2-7be078ade5a2.txt 2024-07-12 20:13:14 I'm not an alpine linux user, nor do I have an account on alpine's gitlab. Is it enough for me to mention it here? 2024-07-12 20:16:17 ayakael97: ^ 2024-07-13 04:11:36 Is there a package for wayland-only emacs? 2024-07-13 04:11:48 I've tried emacs-gtk3-nativecomp but it still complains about not finding DISPLAY :0 2024-07-13 04:12:56 (if I unset the DISPLAY environment variable it just spawns inside my terminal) 2024-07-13 04:13:02 XDG_SESSION_TYPE is wayland 2024-07-13 04:13:07 (xwayland is absent) 2024-07-13 05:27:14 solution is to use emacs-pgtk-nativecomp (p for pure gtk) 2024-07-13 05:27:53 and now I have a different issue: I've installed libcurl and curl-dev but I get ld errors on "cc -Wall -Wextra -pedantic -O3 -lcurl -o sjauth sjauth.c" 2024-07-13 05:28:11 currently on 3.21.0_alpha20240606 (edge) 2024-07-13 05:30:11 runxiyu: put the '-lcurl' after the '-o sjauth sjauth.c' ? 2024-07-13 05:30:28 thanks 2024-07-13 05:30:32 I didn't know that makes a difference 2024-07-13 05:32:23 I no expert on it, that is for sure, but have had similar issues 2024-07-13 05:33:00 interesting.. 2024-07-13 05:35:18 something about symbol lookup and undefined references, but i know that if i put my $(LIBS) at end of a makefile link command, usually that issue doesn't bite me 2024-07-13 05:36:20 hmm 2024-07-13 07:14:29 Hi, I see that Qt has been at v 6.6.3 for some time, is there anything blocking an update to 6.7? 2024-07-13 07:14:48 https://gitlab.alpinelinux.org/alpine/aports/-/merge_requests/63448 2024-07-13 07:16:07 oh right, thanks! 2024-07-13 11:59:09 What is CBUILDROOT supposed to point to? 2024-07-13 11:59:55 A directory 2024-07-13 12:00:13 It's only relevant when cross-compiling (bootstrapping) 2024-07-13 12:02:12 So like a cross platform /usr/lib ? 2024-07-13 12:03:45 that's the way to do it tbh8ööömUklmm8[#hl] 2024-07-13 12:04:42 aninternettroll: no, more like a chroot 2024-07-13 12:05:47 my bad, didn't mean to write garbage in here 2024-07-13 12:06:08 I thought all aports are built in chroots 2024-07-13 12:06:15 just blame it on a cat 2024-07-13 12:07:11 i blame it on not turning my phone off before putting it in my pocket 2024-07-13 12:08:17 That's inherent problem to pocket computers 2024-07-13 12:08:22 aninternettroll: not necessarily 2024-07-13 12:28:27 Is it possible to make dhcp fork earlier, or run asynchronously, during boot? I don't want my system to hang for 5 seconds just for dhcp 2024-07-13 12:29:02 (Or does this have to be this way due to how openrc schedules services that depend on networking) 2024-07-13 12:39:36 You can install dhcpcd as an alternative, but then you'll have to deal with its integration into ifupdown 2024-07-13 12:40:04 (I mean it already is integrated, but you might encounter some problems) 2024-07-13 12:44:11 There is a -A parameter that specifies how long it should wait before going to background (which defaults to 20 seconds) 2024-07-13 12:58:18 I'm not really sure where could I pass parameters to udhcpd 2024-07-13 12:58:57 I think udhcpd is being started as part of `networking' (/etc/network/interfaces: iface wlan0 inet dhcp) 2024-07-13 12:59:38 though tbh I've only touched networkmanager systems on all other distros 2024-07-13 13:01:17 runxiyu: https://github.com/ifupdown-ng/ifupdown-ng/blob/main/doc/interfaces-dhcp.scd#dhcp-related-options 2024-07-13 13:01:45 'dhcp-opts -A5' for example 2024-07-13 13:02:20 > .scd 2024-07-13 13:02:20 Sorry, what's the corresponding package name to install man pages again 2024-07-13 13:02:49 ifupdown-ng-doc 2024-07-13 13:02:56 Thanks 2024-07-13 13:03:02 or docs if you generally want all man pages 2024-07-13 13:03:20 alright that's what I've been looking for for a long time, thanks! 2024-07-13 13:03:32 (hopefully that means i'll spam this channel less) 2024-07-13 14:07:35 is there a way to install manpages for stdlib? for example on my desktop which is not running alpine I can do `man stdlib.h` and I havent been able yo find that on alpine 2024-07-13 14:07:59 s/yo/to 2024-07-13 14:08:30 also somehow it's udhcpc_opts 2024-07-13 14:23:54 ashie: instal man-pages-posix package 2024-07-13 14:49:46 sdl2 seems to only depend on libc, according to apk info --depends 2024-07-13 14:50:07 how does it link to e.g. libx11/libwayland? 2024-07-13 14:50:55 maybe dlopen? 2024-07-13 14:58:05 -DSDL_STATIC=ON 2024-07-13 14:58:06 :( 2024-07-13 15:12:00 I'm trying to set up dwm on 3.20 (x86) and I'm having what can only be described as a massive struggle, so i apologize in advance if what i'll be writing is half a blog post, I'll try to split it in resonable places 2024-07-13 15:16:32 What i did: setup-alpine, reboot, setup-xorg-base dwm (that pulls in st and dmenu too, no idea why the wiki explicitly adds them, and enables community repo), change to my user, startx, that didn't work, from what i can gather modesetting driver with very early i915 is not great, so i installed xf86-video-intel. 2024-07-13 15:19:54 correction, i first ran "Xorg --configure" to see if that would get me anywhere, then added xf86-video-intel 2024-07-13 15:25:52 At this point i had "working" X11, but no input, so set up seatd as described in the wiki, which didn't change anything as far as i could tell. Since this was my second attemt at doing this (my first was a slightly custom alpine install, and i didn't want that variability) i had a look at what groups my user was a member of, and saw i was a member of the "video" group, since 2024-07-13 15:25:55 that is strongly discouraged by the wiki, and wasn't the case in my custom install, i looked why i was a member of that group, turns out when you make a user as part of "setup-alpine" it gets added to the video group. 2024-07-13 15:27:57 Questions: What is actually the right way to handle "video" and "input" for xorg? Did i miss something when setting up seatd? 2024-07-13 16:09:54 socksinspace: did you add your user to the input group? 2024-07-13 16:18:50 you shouldn't need seatd for xorg. 2024-07-13 16:19:26 just have your user in input and video groups and it should just work. seatd might be mucking things up 2024-07-13 16:21:23 well, if i just add my user to input and video that works, but as i understand it that is very much not the way to do it 2024-07-13 16:22:13 if you only have one user on the system, it doesn't matter 2024-07-13 16:22:31 wayland requires seat management where xorg does not. 2024-07-13 16:22:51 hm, okay then i shall be lazy for now :D 2024-07-13 16:23:36 since wayland on the i915gm is not working out in my favour anyway 2024-07-13 16:25:24 also i might have misspoke. i'm not 100% sure wayland technically requires seat management, but stuff like wlroots does. 2024-07-13 17:07:59 invoked: in the case of a multi user system what would you do instead of adding each user to input/video? 2024-07-13 17:15:03 for xorg? i dunno, seatd or something like it i suppose. these are polyfills for freedesktop/systemd. for wayland you don't want users in input group. 2024-07-13 17:16:31 also the users will need to be in the seat group 2024-07-13 17:58:14 invoked: You need to be in the hot seat...group 2024-07-14 04:17:38 how do I install latest mesa with NVK on Alpine? 2024-07-14 04:59:06 well, I ended up installing nvidia-src and it fails to compile make[3]: *** [/usr/src/linux-headers-6.9.9-0-edge/Makefile:1918: /var/lib/akms/6.9.9-0-edge/nvidia/535.86.05-r0/build/kernel-open] Error 2 2024-07-14 04:59:06 make[2]: Leaving directory '/usr/src/linux-headers-6.9.9-0-edge' 2024-07-14 04:59:06 make[2]: *** [Makefile:240: __sub-make] Error 2 2024-07-14 04:59:06 make[1]: *** [Makefile:82: modules] Error 2 2024-07-14 04:59:07 make[1]: Leaving directory '/var/lib/akms/6.9.9-0-edge/nvidia/535.86.05-r0/build/kernel-open' 2024-07-14 04:59:09 make: *** [Makefile:59: modules] Error 2 2024-07-14 04:59:11 akms: ERROR: Failed to build module nvidia/535.86.05-r0 for 6.9.9-0-edge 2024-07-14 04:59:13 akms: ERROR: examine /var/lib/akms/6.9.9-0-edge/nvidia/535.86.05-r0/build 2024-07-14 04:59:18 oops, I think I should've just used a paste service 2024-07-14 04:59:19 sorry 2024-07-14 06:03:56 back 2024-07-14 06:19:53 Ok so building alpine vm cloud images using the official build script repository requires Packer but packer is no longer supported on alpine.  2024-07-14 06:20:42 So it seems you can’t build these alpine images on alpine itself anymore because Packer license hyjinx 2024-07-14 06:43:04 GeneSharp[m]: yeah, that's unfortunate 2024-07-14 06:48:32 I also don’t see a suggested distribution to use for this so I wonder what the official alpine builds are using 2024-07-14 07:35:50 when I add nouveau to vulkan drivers, it won't let me compile mesa anymore: Run-time dependency syn found: NO (tried pkgconfig and cmake) 2024-07-14 07:35:50 Looking for a fallback subproject for the dependency syn 2024-07-14 07:35:50 src/nouveau/compiler/meson.build:28:10: ERROR: Automatic wrap-based subproject downloading is disabled 2024-07-14 07:37:04 nekofisher: Then you have to find out what provides syn 2024-07-14 07:37:26 how do I find that? 2024-07-14 07:37:40 when I do apk search syn there are a lot of packages 2024-07-14 07:37:50 apk search -qx pc:syn revleas nothing, so might be not packaged 2024-07-14 07:38:29 I wonder if I could grab the source code from that and compile it somewhere without a package 2024-07-14 07:38:34 but I don't know where is it from 2024-07-14 07:39:41 me neither 2024-07-14 07:39:54 https://crates.io/crates/syn 2024-07-14 07:39:58 I think it's from this 2024-07-14 07:41:06 seems unlikely 2024-07-14 07:41:52 I think I can understand why mesa doesn't come compiled with nvk ;_; 2024-07-14 11:54:03 Hello fellow alpine-rs 2024-07-14 11:55:26 I have a general question: I was recently put in charge of a project for building a server for best possible Http/PHP (ver 7) performance 2024-07-14 11:55:39 for an old webapp that needs it 2024-07-14 11:56:12 I am using latest alpine-stable on some pretty old (but rock-stable) Xeon v3 machines w 64GB of ram 2024-07-14 11:56:23 I am asking for some advice on the software to use 2024-07-14 11:56:47 lighttpd or Nginx ? php-fpm ? haproxy with some containers? 2024-07-14 11:58:12 I am thinking about haproxy+come containers with php-fpm "workers" and some lighttpd containers for static http 2024-07-14 12:01:29 but I'm open to any kind of software since I must start from scratch 2024-07-14 12:02:12 These are very subjective questions. In the end, it comes down to what your familiar with, or benchmarked to suite your usecase 2024-07-14 12:13:46 wyk72, you can still deliver what you want, then somebody else will actually have to *prove* it's not the best performance 2024-07-14 12:14:18 (or have to admit the requirement wasn't good) 2024-07-14 12:15:33 Ok thanks, I will do some testing then 2024-07-14 12:16:02 I am curious about this "caddy" server, written in GO 2024-07-14 12:23:17 "I am curious about this "caddy..." <- Will sure perform worse than Apache + mod_php as all the PHP stuff will go through expensive cgo wrappers. Or even CGI. 2024-07-14 12:24:43 I see, thanks 2024-07-14 12:24:53 that's moot with fastcgi (fpm) 2024-07-14 12:24:58 "for an old webapp that needs it" <- Chances are that PHP performance does not even matter. Two things do matter though: latency (from request to PHP and from PHP to network), and database performance (independent from PHP) 2024-07-14 12:25:34 ikke: Sort of - in process PHP in the web server will still perform better due to less IPC overhead 2024-07-14 12:26:00 regarding latency, mod_php probably wins, unless Apache adds so much of its own bloat on top that it cancels out 2024-07-14 12:26:41 Just, Apache sucks a lot in comparison for serving anything static 2024-07-14 12:27:38 I've found Apache to be noticeably slower than lighttpd or Nginx, and with something like 200% of the RAM. Also, I'm using fastgi with fpm-php7 not mod_cgi or any sort. Haproxy can also "speak" fastCGI directly to php-fpm so no http involved 2024-07-14 12:28:25 real problem is: PHP code sucks, it's written poorly and there's nothing I can do about it ^^ 2024-07-14 12:29:10 Caddy rocks :) 2024-07-14 12:29:20 chance is that the database or something else is the bottleneck anyway 2024-07-14 12:29:51 I just have to build this server which will be hosted at a 10GB fiber optic-connected farm 2024-07-14 12:31:05 database has to be mariadb 2024-07-14 12:31:21 don't know if it's worth to build a galera cluster for it or not 2024-07-14 12:32:54 clustering is also a can of worms on its own 2024-07-14 12:33:06 Ikke: agreed 2024-07-14 12:34:57 there is a way to avoid tcp stack completely to talk to a mariadb server ? i.e. using a socket between two containers or vms (running on the same machine of course) 2024-07-14 12:36:27 maybe gains are not so great, but it adds up 2024-07-14 14:07:54 That sounds more like micro optimization 2024-07-14 15:38:45 What would the process in Alpine be like when a package maintainer is not reachable anymore? I pinged one about an overflow-fix in the package being available upstream, but no reaction so far 2024-07-14 15:39:51 jemius: Do you have an MR for it? 2024-07-14 15:40:18 ikke, for what precisely? the overflow-fix? 2024-07-14 15:41:59 yes 2024-07-14 15:42:11 or the upgrade 2024-07-14 15:44:51 Upstream unfortunately ignores semantic versioning. The fix was just pushed there in February by the maintainer 2024-07-14 15:44:53 https://github.com/sctp/lksctp-tools/commits/master/ 2024-07-14 15:45:03 alpine package should be: https://pkgs.alpinelinux.org/package/edge/main/armhf/lksctp-tools-dev 2024-07-14 15:45:51 Package dependency requirement 'mate-desktop-2.0 >= 1.27.1' could not be satisfied. 2024-07-14 15:45:57 oops 2024-07-14 15:46:26 If you're up to it, feel free to make an MR to upgrade the package 2024-07-14 15:47:10 If there is no new release, you could include the commit that fixes it as a patch, if that's feasible 2024-07-14 15:48:04 hm. Maybe I should try to convince upstream first to go to semantic versioning 2024-07-14 15:48:51 alpine is already on 1.0.19, so just a new release would suffice 2024-07-14 15:58:35 1.0.19 is from 2021. A ton of fixes is on top of it. As I said, they ignore versioning. That's not good for packagers in any case 2024-07-14 19:21:12 how can i get X working as a non-root user? the screen just freezes, i cant even get into a tty. 2024-07-14 19:22:31 szigma: make sure your user is part of the input and video group 2024-07-14 19:23:40 its done, i also installed a lot of *input packages 2024-07-14 19:24:00 this distro seems great btw 2024-07-14 19:26:19 How are you starting x? 2024-07-14 19:27:44 via startx. I only have exec dwm in it. 2024-07-14 19:28:08 Did you relogin after you added you user to those groups? 2024-07-14 19:28:24 yea 2024-07-14 19:29:02 it seems like others have this issue too, but i didnt found a solution posted 2024-07-14 19:38:12 i have so many questions 2024-07-14 19:39:23 What's life? 2024-07-14 19:40:07 quinq also has questions 2024-07-14 19:42:16 I hope szigma comes back 2024-07-14 20:45:15 to be fair, I remember that happening to me too on a fresh install 2024-07-14 20:45:21 no idea how to fix it tho 2024-07-15 05:45:09 I'll answer what questions I can. I'm pretty new to Alpine myself but I'm an old Linux greybeard in general. 2024-07-15 05:47:43 hexapotomus: o/ 2024-07-15 05:57:03 ? 2024-07-15 06:31:18 hexapotomus: i think he means welcome. 2024-07-15 06:34:02 Gotcha well thanks. I'll be around. 2024-07-15 06:36:19 o/ is an emoticon for a person raising their hand and waving 2024-07-15 06:41:43 Yeah I get it now lol 2024-07-15 06:45:29 I've been really enjoying Alpine in containers. Haven't used the actual full OS yet. How's the state of development overall? I joined the gitlab and poked around. Kinda looks like y'all could use some more hands. I guess that's every distro though. 2024-07-15 06:58:33 hexapotomus: help is always welcome 2024-07-15 06:58:52 we try to keep up, sometimes personal live gives challenges. 2024-07-15 07:08:47 How do you set locale stuff in Alpine? Is it just setting the right variables in /etc/environment? 2024-07-15 07:11:56 (because it doesn't seem to do it for me) 2024-07-15 07:14:13 not sure if this is still valid: https://wiki.alpinelinux.org/wiki/Musl#Locale 2024-07-15 07:15:16 i set mine in ~/.profile 2024-07-15 07:15:52 Yeah I have musl-locales and it includes the required stuff for the language I want but locale just reports C.UTF-8... 2024-07-15 07:15:58 I rebooted as well, no dice 2024-07-15 07:34:45 Oh huh applications (Thunderbird in this case) do pick it up, just locale doesn't report the right variable 2024-07-16 03:16:48 . 2024-07-16 12:21:11 https://xtech.nikkei.com/atcl/nxt/column/18/02864/070900007/ 2024-07-16 12:21:27 sorry, mistake..... 2024-07-16 12:22:47 RIP VMware 2024-07-16 15:13:17 How nice of broadcome to push the industry from VMWare towards open source solutions :) 2024-07-16 15:17:17 if you double the price and lose less then 50% of your customers you made profit and save time 2024-07-16 15:20:28 and you kill all future onboarding 2024-07-16 15:20:39 nobody wants to make a new contract with a corp who does this 2024-07-16 15:21:30 are you sure that's true? 2024-07-16 15:28:14 Would Alpine become better if it were maintained by a commercial company like ubuntu or suse🤔 2024-07-16 15:28:56 It would become more like Ubuntu or Suse 2024-07-16 15:30:22 draining your laptops battery twice as fast 2024-07-16 16:00:18 Any in particular? I'm aware of the PyQt6 stuff already. 2024-07-16 19:56:09 is there a reason why cmake searches for ninja in sbin? I have a fresh container on alpine edge with build-base, cmake and samurai installed and `cmake -B build -G Ninja` fails 2024-07-16 19:56:43 > Running '/usr/sbin/ninja' '--version' failed with: no such file or directory 2024-07-17 09:24:24 Is there an overview of hardening methods Alpine uses out of the box? I wasn't able to find a page for it on the Wiki 2024-07-17 12:49:56 how to zoom in before executing setup-alpine in iso? 2024-07-17 12:50:13 move head closer to monitor 2024-07-17 12:50:45 smart 2024-07-17 12:51:14 if you're on a 4k display or something -- boot with video=1024x768, I guess 2024-07-17 12:52:28 how boot with something? can i modify boot arguments? 2024-07-17 12:52:55 either Tab or E at the bootloader menu 2024-07-17 13:00:22 no grub in alpine iso, i think i cant E 2024-07-17 13:03:25 if it uses syslinux, that generally wants Tab 2024-07-17 13:03:35 or right-arrow maybe 2024-07-17 13:07:14 what i need to do here? 2024-07-17 13:07:50 (all becoming small after openrc loads hardware drivers) 2024-07-17 13:08:36 add the option video=1024x768 at the end 2024-07-17 13:08:51 or video=800x600 or video=1280x1024 or whatever you think would work 2024-07-17 13:09:43 how add option in the end? 2024-07-17 13:09:52 here only boot: 2024-07-17 14:08:40 hi, i'm trying to debug haproxy on alpine, i'd like to extract the stack trace. apparently the package for it has debug symbols stripped. how would i get those debug symbols? do i have to recompile the package myself? 2024-07-17 14:11:33 There are `-dbg` versions of some packages, but it seems that haproxy doesn't have one 2024-07-17 14:22:29 i cant add something to boot: its not working 2024-07-17 14:27:10 i press esc and then boot: video=800x600 and it prompting "Loading video=800x600 failed no such file or directory" 2024-07-17 14:32:50 Guest460: prefix 'lts ' 2024-07-17 14:35:26 thank you 2024-07-17 15:08:15 arf, qutebrowser seems broken now with pyqt6: Error relocating /usr/lib/python3.12/site-packages/PyQt6/QtGui.abi3.so: _ZN5QFont11tagToStringEj: symbol not found 2024-07-17 15:08:24 what upgrade broke that? 2024-07-17 15:37:43 pyqt i would guess 2024-07-17 15:38:04 hello - it seems these instructions for kubernetes are out of date https://wiki.alpinelinux.org/wiki/K8s 2024-07-17 15:38:23 # modprobe br_netfilter <-- this command fails 2024-07-17 15:39:46 proxymoxie: just to rule out, did you recently upgrade your kernel without rebooting? 2024-07-17 15:40:15 well, let me reboot to be 100% sure 2024-07-17 15:41:54 just rebooted to confirm 2024-07-17 15:42:17 modprobe: FATAL: Module br_netfilter not found in directory /lib/modules/6.8.4-2-pve 2024-07-17 15:42:38 -pve 2024-07-17 15:42:45 that's not an alpine kernel 2024-07-17 15:42:59 oh interesting 2024-07-17 15:43:12 i'm running an alpine linux 3.18 lxc container on proxmox 2024-07-17 15:43:19 or at least, that's the template that proxmox provides 2024-07-17 15:43:28 if it doesn't use an alpine kernel that definitely muddles things 2024-07-17 15:43:44 proxymoxie: what does uname -r return? 2024-07-17 15:43:55 6.8.4-2-pve 2024-07-17 15:44:05 Yeah, looks like some custom kernel 2024-07-17 15:45:23 lxc uses the Kernel of the host-system, and that kernel you get shown ist exactly the actual Proxmox Kernel 2024-07-17 15:45:48 yeah, if it's a container, that's not a virtual machine 2024-07-17 15:46:16 so for this little homelab experiment sounds like I should be using VMs and not containers 2024-07-17 15:46:53 If you want to use the alpine kernel, yes 2024-07-17 15:48:27 And I think if you want to use kubernetes, it would be good anyway 2024-07-17 15:48:39 not sure how well kubernetes runs in containers 2024-07-17 15:49:44 Sorry, never used kubernetes ... I'm out on that site! 2024-07-17 15:50:36 kubernetes manages containers it self, and nesting containers inside containers is a can of works 2024-07-17 15:51:18 and also networking 2024-07-17 15:53:31 Tks, ikke! 2024-07-17 17:40:46 anyone here know how I might enable sound blaster 16 support on alpine with linux-lts? seems like the snd-sb16 module isn't there. I'm running Alpine Linux v3.20 with kernel version 6.6.37-1-lts but there is no snd-sb16.ko in /lib/modules/6.6.37-1-lts/kernel/sound/isa/sb/ 2024-07-17 17:44:05 installing linux-firmware-sb16 with apk doesn't fix the issue because it doesn't install the missing module, neither does upgrading linux-lts 2024-07-17 17:44:34 Apperently it requires CONFIG_SND_SB16 to be set 2024-07-17 17:44:47 Which it is not atm 2024-07-17 17:44:52 https://www.kernelconfig.io/config_snd_sb16 2024-07-17 17:45:13 oh god does that mean I have to recompile the kernel 2024-07-17 17:45:25 Or request it to be enabled 2024-07-17 17:45:31 where can I make that request 2024-07-17 17:46:57 is there someone or someplace in particular I should ask? 2024-07-17 17:47:52 https://gitlab.alpinelinux.org/alpine/aports 2024-07-17 17:50:00 should I just post an issue there? 2024-07-17 17:50:41 yes 2024-07-17 17:51:12 okay 2024-07-17 17:52:22 thanks for your help, i'll post the issue later today 2024-07-17 18:15:16 @ikke btw using VMs and the instructions from that wiki page work 2024-07-17 19:00:18 Hi everyone. I'm looking for some help getting pipewire set up on my system. Right now I'm just kicking the tires and have installed Alpine in a VirtualBox VM with the Intel HD emulated hardware. I installed the standard base image (no GUI) and then followed the wiki instructions. I'm able to start pipewire on the command line and everything appears to be running but `wpctl status` only lists the dummy device as a sink, whereas 2024-07-17 19:00:18 h `aplay -L` I'm able to see the Intel HD card in all its various output configurations. I've gone over the instructions many times but clearly I'm missing something. Does anyone in this channel have any experience with this or suggestions? 2024-07-17 19:32:54 Hi there, I'm trying to cross compile a Rust application of mine in an Alpine container. My host is x86_64, my target is aarch64. For linking I need compiler-rt libs for aarch64, which are only in the aarch64 version of the compiler-rt package. However, when I do apk --arch=aarch64 --no-cache add compiler-rt, it still installs theh x86_64 version of the package. How would I make it install the aarch64 2024-07-17 19:32:56 package? 2024-07-17 19:34:46 apk does not support installing packages for different architectures 2024-07-17 19:35:36 and trying to run an aarch64 binary on x86_64 would not work anyway (unless you have something like binft) 2024-07-17 19:35:40 binfmt 2024-07-17 19:36:24 I'm not trying to run it, just need the libraries in compiler-rt for linking with lld and clang 2024-07-17 19:36:39 They'll be statically linked in the cross-compiled binary 2024-07-17 19:39:44 Maybe ask in the #alpine-rust channel how this could be done 2024-07-17 19:40:25 but apk will not mix packages from different arches 2024-07-17 19:57:12 kisielk: you need to use virt image, which oncludes needed drivers 2024-07-17 19:57:41 kisielk: so install linux-virt and remove linux-lts 2024-07-17 19:57:52 let's see if that helps 2024-07-17 19:58:26 tbh you're not the first one with such issues on virtualbox, and I don't know the reason... 2024-07-17 20:17:10 alright! i posted that issue 2024-07-17 20:17:15 Noticed it 2024-07-17 20:17:15 about the sb16 2024-07-17 20:17:32 now all I gotta do is wait for a response... 2024-07-17 20:18:19 If it's for 3.20, maybe you want to mention that as well 2024-07-17 20:18:31 oop i should probably add that 2024-07-17 20:18:40 should I mention it in the description or title 2024-07-17 20:19:02 Description is fine 2024-07-17 20:19:25 kk 2024-07-17 20:19:57 edited 2024-07-17 20:29:41 kisielk: this may not be very helpful, but have you considered using pulseaudio instead? i use the same card in QEMU with alpine and after following the instructions to set up pulseaudio from the wiki it works fine 2024-07-17 21:15:25 @HunterBoy344: That is indeed not helpful, the whole point is to use pipewire :) 2024-07-17 21:15:31 @Ermine I will give linux-virt a try 2024-07-17 21:16:42 fwiw I can see the card in lspci and aplay sees it too.. but I'm willing to try whatever at this point 2024-07-17 21:27:01 i wish you luck 2024-07-17 22:40:52 @Ermine no change with linux-virt 2024-07-17 22:41:36 :( 2024-07-17 22:42:17 can you send pw-dump output here (through a pastebin service, of course) 2024-07-17 22:42:19 ? 2024-07-18 13:12:55 Did something happend to the `tree-sitter-vimdoc` package? 2024-07-18 13:13:04 I swear I remember it existing 2024-07-18 13:53:41 i got a weird thing going on, when i run setup-alpine i get a bunch of Error relocating /lib/libssl.so.3 SYMBOLXXXX: symbol not found. 2024-07-18 13:54:51 i feel like this is not normal, all i did was boot the iso from a usb, then login as root to try the setup-alpine script 2024-07-18 13:57:54 i'm also getting a ERROR: Failed to create /lib/libcrypto.so.3 2024-07-18 14:00:18 so i tried it on another machine and got the same results 2024-07-18 14:07:03 faulty medium or image download? 2024-07-18 14:07:57 maybe i'm going to try and verify the sha now 2024-07-18 14:12:24 tux /home/laptop/Downloads # cmp /home/laptop/Downloads/alpine-extended-3.20.1-x86_64.iso /dev/sda /home/laptop/Downloads/alpine-extended-3.20.1-x86_64.iso /dev/sda differ: byte 4284417, line 11384 2024-07-18 14:12:47 sha on the iso checks out, looks like my usb is junk 2024-07-18 15:28:50 I'm getting filtered super hard RN, when i try to setup my networking it gave me an option to manually edit, which was a mistake since idk how to use the editor or even how to get the heck out 2024-07-18 15:29:47 I assume you are in vim? 2024-07-18 15:30:03 what comes with the alpine iso? 2024-07-18 15:30:15 i thought vim but the instructions on google didnt work 2024-07-18 15:30:51 It would be vi, but those instructions are the same 2024-07-18 15:31:11 esc esc and q!? 2024-07-18 15:31:19 :q 2024-07-18 15:31:36 or with the exclemation mark if you made changes and don't want to keep them 2024-07-18 15:32:13 the exit vim meme is real... 2024-07-18 15:32:21 okay i'm just having issues then, at least i had the right idea 2024-07-18 15:36:49 whatever, have been trying to exit Emacs since 2012 2024-07-18 15:36:51 gave up and just learned it 2024-07-18 15:38:23 lol 2024-07-18 15:41:09 and since 2012 he's been learning it. still haven't exited. 2024-07-18 15:41:56 (i'm kidding btw, i use both emacs and vi myself) 2024-07-18 15:43:36 I mostly use emacs & mg, occasionally nano, vi/vim if that's all that's available 2024-07-18 15:48:24 i use codium and i dont even know how to exit it, i just shut the whole pc down 2024-07-18 15:48:42 nano is pretty solid for editing configs tho 2024-07-18 15:49:16 vis is nice too 2024-07-18 15:49:44 user87, that's the modern way anyway 2024-07-18 15:49:47 One computer for one task 2024-07-18 15:49:52 Like one computer for the web browser 2024-07-18 15:50:07 Another one for the web-browser-based editor 2024-07-18 15:50:17 Maybe another one for the web-browser media player 2024-07-18 15:52:26 actually yeah my dev machine is not my target hardware, its a decade older but the keyboard is superior to my newer laptop 2024-07-18 15:53:03 :) 2024-07-18 15:54:59 quinq: what's the difference between vi and vis? 2024-07-18 15:58:35 from user perspective - vis is scriptable using lua programming language. You configure if using lua language also (like neovim, but simpler). The plugins are also lua. 2024-07-18 15:59:39 It's a mix between vi and sam editors 2024-07-18 16:00:30 otherwise - the editor is written in C (like vi), but the codebase is smaller than viM. I do not know if the codebase is smaller than the other vi flavours... 2024-07-18 16:01:10 ^ exactly 2024-07-18 17:17:48 @Ermine here's my pw-dump output https://gist.github.com/kisielk/1993cc67fd0bcb321b42c7057e20c463 2024-07-18 17:23:29 kisielk: seems like both pw and wp are running as root. They should be run as your user 2024-07-18 17:23:40 do you run pipewire-launcher as root? 2024-07-18 17:27:58 I don't have any users other than root 2024-07-18 17:28:18 This is a VM just for testing pipewire capabilities so I never created any other users 2024-07-18 17:30:44 so you run your desktop environment as root? 2024-07-18 17:31:11 there is no desktop environment, this is going to run in a docker container on an IoT device 2024-07-18 17:32:07 I'm just starting dbus with `export $(dbus-launch)` 2024-07-18 17:36:44 Hm. I don't have any experience with such configuration and I'm not sure if it is even supposed to work 2024-07-18 17:37:11 Do you have eudev started btw? 2024-07-18 17:38:15 you don't even need dbus session bus in such a case. It's for DE's to configure PW/WP iirc 2024-07-18 17:39:12 yeah I tried without dbus as well, but that had the same problem so I re-enabled dbus 2024-07-18 17:39:37 https://wiki.alpinelinux.org/wiki/PipeWire had the export $(dus-launch) thing for non-GUI environments 2024-07-18 17:39:48 I don't have eudev, didn't see that mentioned in the docs 2024-07-18 17:41:03 I'll fix that Soon(tm) 2024-07-18 17:42:31 that article in general needs revision, since they've changed wp config format... 2024-07-18 17:44:41 ok I'll try installing eudev 2024-07-18 17:50:27 And if that didn't work, please try to grab PW and WP's output 2024-07-18 17:57:55 happy to report that the newly-enabled SB16 support in linux-virt works like a charm ;) 2024-07-18 18:05:20 @Ermine that seems to have done the trick! 2024-07-18 18:05:43 kisielk: nice to hear that 2024-07-18 18:06:56 Also, please consider creating a new account to run pipewire and wireplumber. Running stuff as root is usually a bad idea 2024-07-18 18:07:02 ooh I missed this, what ended up fixing it? 2024-07-18 18:07:12 eudev 2024-07-18 18:07:16 aha 2024-07-18 18:07:33 @Ermine it's going to run in a docker container as part of a Docker compose setup, there's no user accounts in any of the containers 2024-07-18 18:09:02 not my decision as I didn't design the system, but it's also not unusual when running containerized apps 2024-07-18 18:12:04 obviously if I was running this on a regular machine I'd have user accounts etc 2024-07-18 18:43:11 any idea why `ssh-add` could not work for non-default key? adding id_ed22519 works but literally any other key will still prompt for password 2024-07-18 18:51:44 surprisingly it works with sshfs, just not ssh client itself 2024-07-18 18:52:41 ashie: do ssh-add -L and check if all keys are in the agent 2024-07-18 18:53:06 keys with a custom name need to be specified explicitly to ssh-add to be added to the agent 2024-07-18 18:58:41 nero: yes they are, I also specify the key in the ssh command 2024-07-18 19:00:10 doing `ssh-add && ssh ... -i ` asks for password both times, doing the same for default keys (the ones named id_*) just works tho 2024-07-18 19:11:37 ashie: tell me when you find out why is that so 2024-07-18 19:11:43 surely interesting 2024-07-18 19:20:27 its not bothering me enough to run it in debugger and aside from that I have no clue anymore 2024-07-18 19:20:47 Ill hope it will fix itself one day when I update :P 2024-07-18 21:15:35 Hello 2024-07-18 23:23:17 Whats up heart 2024-07-18 23:46:06 h0bbl3s: downloading some stuff 2024-07-19 09:06:55 can tr or something replace 'col -b'? 2024-07-19 09:07:28 seem to have used col in a not-too-old script, almost didnt even find it in repo .. 2024-07-19 09:08:55 frag: what does that do? 2024-07-19 09:10:15 col filters out certain terminal control characters 2024-07-19 09:12:26 from man page: Do not output any backspaces, printing only the last character written to each column position. 2024-07-19 09:25:21 ill just instlal it :] 2024-07-19 09:25:36 heh ok 2024-07-19 09:25:51 I guess tr -d '\b' isn't quite enough 2024-07-19 14:11:43 Hey, I'm having some issues with running a python package with python3.12 . Is there a way I could install python3.11 alongside the latest version ? 2024-07-19 14:36:13 what's the intended way to replace udev rules? mdev.conf is not helping since mdev does not trigger uevents itself 2024-07-19 14:36:31 I feel that init script missing proper coldplug 2024-07-19 14:38:17 xaltsc: something like conda can make environments with earlier Python versions 2024-07-19 14:41:06 consus: what kind of udev rule? mdev.conf can exec commands 2024-07-19 14:46:30 create vrf on network interface 2024-07-19 14:46:41 *vf 2024-07-19 14:46:58 the problem is that mdev runs too late 2024-07-19 14:53:34 i dont know what you refer to with vf/vrf 2024-07-19 14:53:53 vf 2024-07-19 14:54:05 virtual function 2024-07-19 14:55:06 ah, it seems that module is loaded by initramfs 2024-07-19 14:55:37 so probably rebuilding initramfs to include new mdev.conf should help, right? 2024-07-19 14:57:28 iavf module? 2024-07-19 14:57:41 mlx5_core 2024-07-19 14:58:05 yeah, did the trick 2024-07-19 14:59:31 good that you found it out on your own 2024-07-19 14:59:59 it's actually minimal, he's DMing me 2024-07-19 15:00:07 LOL 2024-07-19 15:00:34 im not an nvidia user, so i did not know the "virtual function" terminology 2024-07-19 15:00:51 not SR-IOV user, either 2024-07-19 15:00:53 it's not nvidia 2024-07-19 15:01:01 it's connectx-5 2024-07-19 15:01:11 a network adapter capable of ROCEv2 2024-07-19 15:01:31 this is what i found https://docs.nvidia.com/networking/display/mlnxofedv461000/single+root+io+virtualization+(sr-iov)#src-12013542_safe-id-U2luZ2xlUm9vdElPVmlydHVhbGl6YXRpb24oU1JJT1YpLWN4LTMvcHJvc3ItaW92Y29uZmln 2024-07-19 15:01:55 refers to your kernel module, "virtual function", and connectx adapters 2024-07-19 15:02:49 yeah 2024-07-19 15:03:00 well it works fine 2024-07-19 15:03:12 I just wanted to automate the setup 2024-07-19 15:03:24 And it's not obvious that mdev running twice 2024-07-19 15:04:31 has to, as the initramfs environment does not have access to all kernel modules, but some are still required to be loaded 2024-07-19 15:05:06 udevd just triggers uevents the second time hence there is no need to rebuild initramfs to include the rules 2024-07-19 15:05:09 forgot about that 2024-07-19 15:06:07 do you actually need the networking stuff from the initramfs? 2024-07-19 15:06:30 It's just there because mlx5_core is in drivers/ethernet 2024-07-19 15:06:37 *drivers/net/ethernet 2024-07-19 15:09:34 xaltsc: alpine only supports one python version installed with the package manager 2024-07-19 15:25:29 ikke: is it risky/dangerous to rollback to python3.11 for some time ? 2024-07-19 15:26:41 You mean for Alpine Linux, or for yourself? 2024-07-19 15:27:34 xaltsc: I mean, technically we could have multiple python versions side-by-side, but that would duplicate the amount of work, all python packages would need to be duplicated as well for each python version 2024-07-19 15:28:46 ikke: for my alpine install, rollback to the alpine version that has python3.11 and rollback the packages that depend on it as well 2024-07-19 15:30:01 xaltsc: only 'risk' is that 3.19/community is no longer supported, so those packages don't receive security updates anymore. If you only use packages from main, or can live with that, it should be fine 2024-07-19 15:37:58 ikke: how can I print a list/tree of packages that depend on python ? 2024-07-19 15:38:12 apk info -R python3 2024-07-19 15:38:37 sorry, wrong command 2024-07-19 15:39:02 apk list -I -d python3 2024-07-19 15:48:22 ufw and certbot depend on python3... 2024-07-19 15:49:31 and fail2ban... 2024-07-19 15:49:37 ACTION doubts 2024-07-19 15:50:52 It's just a small personal use server after all, am I at huge risk ? While being able to use this python package would help me a lot with my productivity/mail mgmt. 2024-07-19 17:11:24 I have a question about Alpine and CVE-2023-6992 (zlib "Cloudflare" vuln). It has been my understanding that Alpine builds zlib from upstream sources and does not use the Cloudflare modifications which are affected by 6992. I just checked security.alpinelinux.org/vuln/CVE-2023-6992, and it shows that zlib 1.31.1-r0 is "possibly vulnerable". How do I interpret this; does this mean Alpine is using the 2024-07-19 17:11:30 zlib source? 2024-07-20 01:13:24 hello 2024-07-20 01:13:35 greets. 2024-07-20 01:15:10 howre you doing 2024-07-20 01:17:32 all good here; but lets keep the chit-chat in #alpine-offtopic 2024-07-20 01:17:47 ikke: we could potentially have a python3.11 package, and only the minimum stuff to build a virtualenv. 2024-07-20 01:18:19 E.g.: don't need to duplicate *all* python packages. 2024-07-20 01:24:34 "all good here; but lets keep the..." <- is that a public room, i only see #alpine-linux and #_oftc_#alpine-devel:matrix.org 2024-07-20 01:24:54 just join it 2024-07-20 01:24:57 it's there 2024-07-20 01:27:22 ¯\_(ツ)_/¯ doesnt seem to exist for my client 2024-07-20 02:58:09 chris123456789[m]: matrix bridge moment, I switched to just using irc directly maybe two weeks ago because of this kinda stuff 2024-07-20 08:32:42 wait, #alpine is bridged? 2024-07-20 08:32:58 the hell I'm doing on IRC then 2024-07-20 08:33:39 irc rocks 2024-07-20 08:33:43 lol no 2024-07-20 08:34:18 rocking since 1988 2024-07-20 08:34:27 it was great in 1988 2024-07-20 08:34:35 it's a UX nightmare in 2024 2024-07-20 08:34:57 hosting a VM with a bouncer to have a chat is... a meme basically 2024-07-20 08:35:21 and when i connect to a server it doesnt federate with 30 others to download a list of everyones names profile pictures and other metadata 2024-07-20 08:35:41 yeah, because there is no picture 2024-07-20 08:35:43 no video 2024-07-20 08:35:46 no code listing 2024-07-20 08:35:47 no threads 2024-07-20 08:35:50 nothing basically 2024-07-20 08:36:14 i find these things distracting and a negative to my ux experience 2024-07-20 08:36:26 but we off topic ig 2024-07-20 08:36:29 yeah 2024-07-20 08:51:31 musl actually has a pretty decent allocator nowadays 2024-07-20 08:52:13 qemu + musl vs qemu + jemalloc resutls only in ~5us latency increase in my application running inside that VM 2024-07-20 08:52:49 well, that's a ~25% increase in latency, but actually I've expected much worse 2024-07-20 08:57:34 I like in irc is that: no video, no code listing, no threads, nothing basically, just plain old text chat :) 2024-07-20 08:59:45 that works okay until you get to the subway 2024-07-20 08:59:52 now you lost all replies 2024-07-20 08:59:54 :DDDDD 2024-07-20 09:00:59 consus[m] what do you mean? 2024-07-20 09:01:06 no history lol 2024-07-20 09:02:14 persistent irc session with screen 2024-07-20 09:02:21 screen? 2024-07-20 09:02:23 what screen 2024-07-20 09:02:28 I'm on my phone 2024-07-20 09:02:32 No screen :DDDDDD 2024-07-20 09:02:59 well, screen runs on the server 2024-07-20 09:03:11 What server? Now I have to have a server to have a chat? 2024-07-20 09:03:25 but i have screen on my phone too, with postmarketos 2024-07-20 09:03:51 I don't want to do that anymore, I'm too old for that 2024-07-20 09:03:56 consus[m]: we always need a server in one form or another 2024-07-20 09:04:05 We already have 2024-07-20 09:04:07 IRC server 2024-07-20 09:04:14 :DDD 2024-07-20 09:05:00 mm but irc server relays the messages, not centrally store them like matrix 2024-07-20 09:05:25 Aha, that's why it a total UX nightmare for modern day and age 2024-07-20 09:05:54 BTW there are IRC servers with chathistory support 2024-07-20 09:06:18 anyone whos seen a noob struggle with matrix cross signing knows matrix not much better 2024-07-20 09:06:21 It just takes waaaaay too long to actually deploy all that stuff 2024-07-20 09:06:49 You mean "select these funny emojies"? 2024-07-20 09:07:02 Please, my mother can do that 2024-07-20 09:07:20 if you arent there to guide them its much harder 2024-07-20 09:07:40 It's much harder to host a virtual machine where you have to configure a bouncer 2024-07-20 09:07:48 My mother can't do that, I'm pretty sure of that 2024-07-20 09:07:52 and if you mess up the inital signing of your first two devices it leaves it in a broken state 2024-07-20 09:08:20 Non-issue, really 2024-07-20 09:08:37 Skill issue, I could even say 2024-07-20 09:09:07 With IRC it's straight up wasting time on admining a VM 2024-07-20 09:09:10 What a meme 2024-07-20 09:09:15 can you write your own matrix client? 2024-07-20 09:09:36 with enough adderall and self hatred you can 2024-07-20 09:09:51 cuz it's complication A.F. 2024-07-20 09:09:57 can you write your own irc client? 2024-07-20 09:10:01 no 2024-07-20 09:10:05 nobody written ircv3 client 2024-07-20 09:10:08 :DDDDDD 2024-07-20 09:10:20 there is basically no client that supports all ircv3 features as of now 2024-07-20 09:10:28 you can write irc client in plain shell too 2024-07-20 09:10:34 so? 2024-07-20 09:10:38 I still need a VM to use that 2024-07-20 09:10:40 :DDDD 2024-07-20 09:10:45 because it's simpler! 2024-07-20 09:10:47 And external services to hold logs 2024-07-20 09:11:01 So no 2024-07-20 09:11:10 lets talk in alpine-offtopic? 2024-07-20 09:11:14 Sure 2024-07-20 09:11:23 nah, I'm done with this 2024-07-20 09:11:24 hf 2024-07-20 09:11:44 I was under the impressions it was a polite way to ask us to stop flooding 2024-07-20 09:12:00 I'm not sure that Internet needs another IRC flame war 2024-07-20 09:12:26 you've started this one 2024-07-20 09:14:01 Pretty much 2024-07-20 10:03:53 tes 2024-07-20 10:04:03 rest 2024-07-20 13:11:59 Hello, when I run apk upgrade the response comes back as "OK: 773 MiB in 404 packages" and then I am returned right back to the prompt. How can I actually install these updates? 2024-07-20 13:15:04 afaik that's not the size of updates, it's the total of your current packages? 2024-07-20 13:16:03 ah. I have not run updates in over a year and there is nothing to update? 2024-07-20 13:16:05 if it only shows this messages, your packages are already up-to-date 2024-07-20 13:16:28 internetsurfer2000: which alpine version do you use? 2024-07-20 13:17:30 how do I check release version on alpine linux? 2024-07-20 13:18:40 oh yeah. 3.15. eol was last november 2024-07-20 13:18:44 I guess that's it 2024-07-20 13:19:11 does alpine do lts releases? 2024-07-20 13:20:41 No, ever release has the same time of support 2024-07-20 13:27:10 does anyone know what this means when using apk? "Unable to lock database: temporary error (try again later)" 2024-07-20 13:27:32 it happens with `apk update`, `apk add ...` 2024-07-20 13:27:44 i am running as root also 2024-07-20 13:31:01 It means another process is using the database, so it must wait for the lock to be released (the other process to stop using the db) before it can acquire the lock and start using it. 2024-07-20 13:31:12 Do you have another instance of apk running at the same time? 2024-07-20 13:32:27 i found another process running apk 2024-07-20 13:32:40 do i kill -9 it? 2024-07-20 13:33:40 i had tried to upgrade a package yesterday and i dont even know how that process is still alive 2024-07-20 13:34:43 the state for it is sleeping, ig ill try to kill it 2024-07-20 13:35:49 alr, that worked. thanks internetsurfer2000! 2024-07-20 13:37:50 no prob. cheers. 2024-07-20 14:01:35 i usually run alpine on latop in cli mode, is there a way to shut the dislay(lcd) from cmd line ? 2024-07-20 14:02:14 tried looking in /sys/devices/platform/backlight ... does not exist 2024-07-20 14:03:00 don't want to install xrandr 2024-07-20 14:06:46 Perhaps something like this? echo "4" > /sys/class/graphics/fb1/blank 2024-07-20 14:06:59 trying 2024-07-20 14:08:40 https://unix.stackexchange.com/questions/529583/is-there-a-better-way-than-this-to-force-monitor-blank-off 2024-07-20 14:21:28 nice works, i have fb0, did that from ssh login as doing from device itself is interpreted as keystroke and brings back from being blank 2024-07-20 14:21:57 thanks 2024-07-20 14:22:11 Greetings, I've compiled pam_cap.so and placed it in /usr/lib/security/pam_cap.so, added this: "auth optional pam_cap.so" to /etc/pam.d/su-l and created /etc/security/capability.conf with this in it "cap_net_admin @foo". I have a test code that creates a dummy nic and I run it as user goo like this: "su -lc "~igor/test" - foo" which is part of foo but I still get Operation not permitted error, any ideas why? 2024-07-20 15:32:46 \quit 2024-07-20 15:42:50 sleep 3 && echo 0 ... works 2024-07-20 16:27:57 is enyone elses discovery/ flatpak unable to connect to server 2024-07-20 16:28:21 it was on a fresh install too 2024-07-20 16:28:57 Hello, I am trying to install Alpine as Dom0 for Xen. I am following https://wiki.alpinelinux.org/wiki/Xen_Dom0 but I experience an error when running `apk add xen-hypervisor` 2024-07-20 16:29:13 The error is: /usr/bin/grub-probe error failed to get canonical path of tmpfs 2024-07-20 16:29:41 If I ignore this and reboot anyways, I then get the repeated error: can't open /dev/hvc0: No such file or directory 2024-07-20 18:22:55 It seems the installer is just not adding the grub boot menu entry for Xen 2024-07-20 18:23:51 use systemd boot instead 2024-07-20 19:46:09 btw whats yalls favourite desktop enviroment 2024-07-20 19:46:38 I don't use any 2024-07-20 19:47:02 Only a window manager 2024-07-20 19:47:33 once i growup i wannabe like ikke 2024-07-20 19:48:20 i bet he runs a raspberry pi zero 2024-07-20 19:48:40 Don't let your dreams be dreams 2024-07-20 19:49:04 BeutifullScience, rather an i586 2024-07-20 19:54:48 no i586 or rpi0 here 2024-07-20 20:04:40 BeutifullScience: how did you jump to that (amusing) conclusion 2024-07-20 20:06:26 you dont run a de is your hardware isnt to slow 2024-07-20 20:09:07 wm is for potato computers 2024-07-20 20:09:56 I don't run it because I don't need it 2024-07-20 20:10:03 less bloat 2024-07-20 20:10:07 why use a tiling wm term 2024-07-20 20:10:13 Because I use one 2024-07-20 20:10:19 just use a ty 2024-07-20 20:10:53 tilling wm are anoing 2024-07-20 20:11:10 You are free to use something else 2024-07-20 20:11:19 running a wm doesn't require having a potato computer 2024-07-20 20:11:35 plenty of people do it on powerful hardware because they prefer it 2024-07-20 20:11:36 yes but why would you? 2024-07-20 20:11:54 Because it doesn't get in the way 2024-07-20 20:13:02 i like to resize my windows to fit me preferences not having windows jumping around resizing on their own 2024-07-20 20:13:46 You can have both :) 2024-07-20 20:13:59 real programes would make a tiling terminal 2024-07-20 20:14:08 A window manager isn't necessarily tiling 2024-07-20 20:14:13 my friend 2024-07-20 20:14:17 do whatever you want 2024-07-20 20:14:26 :) 2024-07-20 20:14:26 and let others do the same 2024-07-20 20:15:06 a tilling term might make sence not using x 2024-07-20 20:15:28 that would be something 2024-07-20 20:15:32 That exists 2024-07-20 20:15:42 runing x to have a glorified terminal manger 2024-07-20 20:15:50 doent make sence 2024-07-20 20:15:53 You can look at dvtm for example if you're interested 2024-07-20 20:16:47 You can also run other type of windows than virtual terminals in your X session 2024-07-20 20:17:00 shocker 2024-07-20 20:17:11 Apparently :) 2024-07-20 20:18:44 you can run mpv on terminal without x 2024-07-20 20:19:33 noted 2024-07-20 20:21:06 why even run a display server 2024-07-20 20:22:22 such waste of resources 2024-07-20 20:22:23 You don't have to :) 2024-07-20 20:23:37 th e problem is most apps require it 2024-07-20 20:26:49 displays are such bloat 2024-07-20 20:26:56 bring back line printers 2024-07-20 20:28:39 now your just troling 2024-07-20 20:29:23 Contrary to “Why even run a display server such waste of resources” 2024-07-20 20:30:59 display servers running all the time are 2024-07-20 20:32:48 a term manager on the other hand would be pretty cool 2024-07-20 20:34:10 tilling trem manager 2024-07-20 20:34:13 TTM 2024-07-20 23:39:15 why can't i not connect to flatpak servers 2024-07-20 23:39:33 like its a fresh install and i obviosly have an internet connection 2024-07-21 00:07:00 i don't use flatpak, but i found these: https://github.com/flatpak/flatpak/issues/5740 and https://gitlab.alpinelinux.org/alpine/aports/-/issues/15915 2024-07-21 00:08:05 does flatpak use container for isolation? could be net namespace is unshared... 2024-07-21 00:32:53 great 2024-07-21 00:33:05 i may just switch to a different os 2024-07-21 00:43:15 it's a strange issue, as i've never seen it myself (i'm using flatpak on alpine) 2024-07-21 00:46:19 i've been curious about it, was starting to set it up to see if i run into same/similar issue 2024-07-21 00:46:50 now op is gone, i'm purging it from this box... don't really need it 2024-07-21 00:47:32 the only benign issue i've ever noticed is that it will sometimes lag returning to the prompt after updating (even though there were no updates) 2024-07-21 00:48:17 i've mainly used flatpak for stuff like steam. 2024-07-21 00:50:02 ah, right, makes sense. if i gamed more i would probably got that route. i mostly play descent (old 90's dos game ported) and that rarely anymore 2024-07-21 00:50:12 s/got/go 2024-07-21 00:52:02 i don't game much but it's interesting to see stuff work pretty much on par as they do in windows these days. 2024-07-21 00:52:49 nice 2024-07-21 02:23:46 I'm having a problem with the APK cache on a diskless system where, after running "apk cache -v sync", my system no longer finishes booting....just lands on.... 2024-07-21 02:23:58 can't run '/sbin/openrc': No such file or directory 2024-07-21 02:24:00 can't run '/sbin/openrc': No such file or directory 2024-07-21 02:24:01 can't run '/sbin/openrc': No such file or directory 2024-07-21 02:24:03 (none) login: 2024-07-21 06:16:07 kgmoney: sounds like packages missing from the cache directory on the boot media? 2024-07-21 11:28:51 Trying to connect a bluetooth keyboard, but bluez reports Failed to connect: org.bluez.Error.NotAvailable br-connection-profile-unavailable. Any idea how to fix that? 2024-07-21 11:44:33 @ikke ...ya missing packages seems like it would make sense, but if they are missing, it's because "apk cache sync" is deleting them because the system runs fine prior to that command. Also, if I boot without the "silent" option in syslinux.cfg, it says that package installation succeeds. 2024-07-21 11:48:09 Something I noticed with cache sync, is that it seems to be removing two APKINDEX files from the cache folder that seem like they would they would be needed. As I understand it, the first part of the sha1sum of the apk repo path is used in the name APKINDEX.xxxxxxxx.tar.gz, and the two that match my repos both get deleted by apk cache sync. 2024-07-21 12:36:17 ikke: maybe you need to modprobe uhid? 2024-07-21 12:39:20 hmm, I get a different error message now, so seems it helped with the missing profile 2024-07-21 12:41:07 Ok, it's connected now, but it's not registring any input yet 2024-07-21 12:52:47 WhyNotHugo: success, thanks! 2024-07-21 12:53:17 what was the other issue? 2024-07-21 12:53:59 1. I needed to connect while the keyboard is in pairing mode 2024-07-21 12:54:23 2. I needed to set ClassicBondedOnly=false in /etc/bluetooth/input.conf 2024-07-21 12:56:40 Not sure if #1 is always necessary 2024-07-21 12:59:38 I think so, at least for the first pairing. 2024-07-21 12:59:54 it was already paired 2024-07-21 13:00:11 but gives the authorized error while trying to connect while the keyboard is not in pairing mode 2024-07-21 13:00:27 Macmpi had similar errors (gh issue) 2024-07-21 13:00:52 https://github.com/bluez/bluez/issues/319 2024-07-21 16:50:25 hello. I was going to install alpine on a small box to be a file server but the hard drive is slow. Should I use diskless mode or sys mode? 2024-07-21 16:50:51 Or, they "data" mode? I'm still not sure how that would work 2024-07-21 16:52:59 I assume you want the files to be stored on the hard drive 2024-07-21 16:53:30 It would hardly matter what kind of setup you choose in that case 2024-07-21 16:53:38 either sys or data 2024-07-21 16:53:54 sys means everything is installed on disk 2024-07-21 16:54:11 data means run-from-ram, but some directory is mounted on the disk 2024-07-21 16:57:47 setup-xen-dom0 complain about xenqemu, as in https://wiki.alpinelinux.org/wiki/Xen_Dom0 ? 2024-07-21 16:59:15 for installing dom0, is "rc-update add xenqemu" needed ? 2024-07-21 17:15:00 hmm, that comes from xen-qemu pkg 2024-07-21 21:38:53 Hello. I'm unfamiliar with alpine; how would I go about installing libGLESv2.so.2? https://paste.debian.net/1323930/ 2024-07-21 22:02:37 try just mesa-gles 2024-07-21 22:09:07 Ah that worked. 2024-07-21 22:09:08 Thanks. 2024-07-21 22:32:06 ikke: Thank you for the response. Would I get better performance as "data" or "sys" on a spinning 5400rpm drive? This is meant to be a network file server and the files are intended to be on the drive. 2024-07-22 02:50:54 Is there a method to install two different versions of postgresql on Alpine so that one can migrate a database between versions? I'm having trouble figuring out if this is possible. 2024-07-22 02:58:31 I have installed postgresql16 postgresql16-contrib postgresql16-openrc & postgresql14 postgresql14-contrib postgresql14-openrc, but it appears to only let me start/stop psql version 16 and not 14 2024-07-22 02:59:49 um... I'm trying to compile a very simple program that uses curl and openssl for https, statically 2024-07-22 02:59:53 and i'm not sure how to do that 2024-07-22 03:10:16 https://paste.sr.ht/~runxiyu/78ce1beb9a82ee2600537109a0be507f6673a1ce 2024-07-22 03:22:14 same for cc -Wall -Wextra -pedantic -O3 -o sjauth sjauth.c /usr/lib/libcurl.a -Wl,--as-needed,-O1,--sort-common -L/lib -lcares -lnghttp2 -lidn2 -lpsl -lssl -lcrypto -lssl -lcrypto -lzstd -lbrotlidec -lz -static 2024-07-22 03:22:38 (everything after sjauth.c is just $(shell curl-config --static-libs)) 2024-07-22 04:23:18 How is anyone auto-mounting smb at boot on stable? Edge has pam_mount which I use, but I am not putting my wife on Edge as her first Linux experience 2024-07-22 04:39:11 runxiyu_: do you have brotli-static installed? 2024-07-22 04:56:07 runxiyu_: might be some more, at least libunistring-static for one 2024-07-22 04:57:44 also will probably need to add '-lunistring' 2024-07-22 05:00:07 it seems, '_kBrotliPrefixCodeRanges' is defined in libbrotlicommon 2024-07-22 05:00:47 same applies to BrotliDefaultFreeFunc, BrotliDefaultAllocFunc, _kBrotliContextLookupTable, etc 2024-07-22 05:01:45 so, I'd add -lbrotlicommon as well 2024-07-22 05:04:44 yep, though there might be a problem with the brotli static libs... readelf -s is not showing any of the missing symbols 2024-07-22 05:10:30 readelf -s works for me (older than v3.20) 2024-07-22 05:10:45 I mean, on brotlicommon.a 2024-07-22 05:11:18 ok, on edge here, will look at changes to the aport 2024-07-22 05:13:43 Saijin_Naib[m]: apparently if you have cifs-utils installed, you can add an entry to fstab? 2024-07-22 05:18:19 aron: ok, i wasn't paying attention well enough... brotli static libs compiled with lto, so needed --lto-syms and -X to see to full naming 2024-07-22 05:21:20 nice, so at least the libs aren't broken 2024-07-22 05:21:20 j_v: yes, brotli-static is installed 2024-07-22 05:21:47 "Saijin_Naib: apparently if..." <- Yeah, have that and the entry, but it looks like I cant use an auth file witjout pam_mount. doas mount -a works fine, but I cant expect her to do that every login 2024-07-22 05:23:13 runxiyu_: try to append -lunistring and -lbrotlicommon after $(curl-config) 2024-07-22 05:23:27 seems to work 2024-07-22 05:23:28 thanks! 2024-07-22 05:23:29 Her user is in every group mine is, she is just on latest-stable so no pam_mount, unlike my setup on edge 2024-07-22 05:23:51 I'd say curl-config is the one, which is broken 2024-07-22 05:24:59 a 5.6MB binary :o 2024-07-22 05:25:03 hehe 2024-07-22 05:25:04 We have the exact same laptops, so I just diff'd the apk world to get hers setup like mine, except GNOME instead of XFCE 2024-07-22 05:25:17 https://linuxconfig.org/how-to-mount-a-samba-shared-directory-at-boot 2024-07-22 05:25:29 (21.7 when unstripped) 2024-07-22 05:25:53 I'm not surprised 2024-07-22 05:26:41 are you sure, you want static curl? :P 2024-07-22 05:31:34 "https://linuxconfig.org/how-to-..." <- Yep, auth file does not work without pam_mount, only password in fstab does 2024-07-22 05:31:50 Would rather not do that 2024-07-22 05:32:33 Saijin_Naib[m]: so credentials=/path/to/file does not work? 2024-07-22 05:32:40 Not aware of any alternative 2024-07-22 05:35:00 ikke: Correct, not at boot from netmount service, only when invoked via doas mount -a, or on the twin edge machine with pam_mount 2024-07-22 05:35:45 ikke: I might have to ask pam_mount go Community... it seems like a big hurdle when missing 2024-07-22 05:36:38 I am now remembering chasing my tail over this here a year or more ago and finding some arcane post that said this all leans on pam_mount to just work 2024-07-22 06:16:03 how about pam_cifscreds and mount with -o multiuser 2024-07-22 06:16:41 wait, that credentials issue is really weird 2024-07-22 13:41:22 "how about pam_cifscreds and..." <- I can try, but isnt that package only in edge? 2024-07-22 18:46:36 is it possible to install/run desktop(xfce) in xen, as in https://wiki.alpinelinux.org/wiki/Create_Alpine_Linux_PV_DomU ? 2024-07-22 18:47:05 or need a different iso ? 2024-07-22 19:07:19 vnc view say "vm has no graphics display device", probably need to set up graphics card 2024-07-22 19:08:37 Hello, Anyone know why when installing Alpine, the keymap breaks? I'll set it up with "us, us" and when i reboot and go to login my username shows correctly but after i hit enter it starts doing some "^[a ^[b ^[c " entry on key press 2024-07-22 19:23:11 hmm, probably PV does not support gfx_passthrough or pci passthrough 2024-07-22 20:06:48 woohoo https://alpinelinux.org/posts/Alpine-3.17.9-3.18.8-3.19.3-released.html 2024-07-22 22:27:27 vkrishn: it's definitely possible. 2024-07-22 22:28:08 what are you hoping for in that configuration (as opposed to running a desktop normally)? 2024-07-22 22:30:34 that is, i'm not saying gpu passthrough is possible on pv, but it is possible on hvm 2024-07-23 01:19:27 So is there really any reason to use packer to build cloud images? I’ve just been using this guy’s script that uses qemu nbd mounts to create an image. I just took that and added my own stuff. Can add packages in build or in a first boot script.  2024-07-23 01:20:10 The official cloud image build repos on gitlab seem unnecessarily elaborate 2024-07-23 06:39:39 packer with qemu only makes sense if you need vnc typing 2024-07-23 08:40:07 hi 2024-07-23 08:42:05 Ive made an alpine linux router, however, I've got a problem with dhcpcd 2024-07-23 08:43:26 when my isp resets the internet session, dhcpcd won't get an ipv6 address 2024-07-23 08:44:52 I've put "service dhcpcd restart " in pppd ipv6-up script but it doesn't work 2024-07-23 08:45:38 because of restricted service command 2024-07-23 08:47:43 so the question is how do I make dhcpcd obtain an ipv6 address when ppp session resets 2024-07-23 09:20:12 Hi 2024-07-23 17:35:22 hi. can I update from 3.17 straight to 3.20? 2024-07-23 17:44:55 i think you can, but check https://wiki.alpinelinux.org/wiki/Upgrading_from_older_versions#Upgrading_Alpine_%E2%89%A4v3.19_to_v3.20.0 2024-07-23 20:26:48 hey all, is anyone familar with the LVM on LUKS guide on the wiki? I've tried a few times, and I'm not sure why, but root fails to mount for me after the install 2024-07-23 21:12:16 make sure your root fs (whatever it is) listed as a feature in mkinitfs 2024-07-23 21:12:53 there's a lot of ways this can go wrong, but people seem to make that mistake most often 2024-07-23 21:13:29 if you're using setup-alpine then it might be a bug. 2024-07-23 21:28:18 henlo 2024-07-23 21:30:52 im booting Alpine from a USB drive in diskless mode, only saving my changes with LBU. everything is perfect, but on some machines where /dev/sda is already mapped to the internal drive, my USB key (that at installation time was /dev/sda) get reassigned to sdb or sdc. When it happens, the apkoverlay is not loaded and the OS defaults to stock. Is there a way to force it reloading the overlay after booting or ideally to edit the usb key 2024-07-23 21:32:01 i have many things saved in the the overlay, i'd rather not start over because of the mount point issue. 2024-07-23 21:36:14 @invoked I'm following the guide on https://wiki.alpinelinux.org/wiki/LVM_on_LUKS, using the "optimized for security" Luks1 settings and UEFI 2024-07-23 21:37:59 if your UUID is wrong, if you type even one thing wrong, it can fail. i take no position on the wiki guides (for or against). alpine's setup tool will do it for you automatically one way, but there are many ways to do it. 2024-07-23 21:38:52 if you're willing and can't figure it out, you could try the alpine installer method 2024-07-23 21:40:48 I'm reinstalling again, I'll see if I can get it working. Otherwise, I'll try that and report back 2024-07-23 21:40:49 Thx 2024-07-23 22:13:15 Solved my issue by using labels 2024-07-23 22:13:18 hey guys, i had installed the font-material-icons on alpine linux and noticed it was missing icons due to google not making a new release for the font (last release was 2020 i think) 2024-07-23 22:13:33 it also seems to have a slight rebranding to "Material Symbols" instead of material icons 2024-07-23 22:13:45 if i were to make an update, do i make a new package or update the current one? 2024-07-23 22:14:15 (i already made an APK for my own use, might as well update/make a new one in the mainstream repo) 2024-07-23 22:17:34 tldr; update font-material-icons (google stalled and won't make a new release) or create new font-material-symbols with files from the master branch 2024-07-23 23:11:59 Since last update it seems, my ~/.lesskey isn't taken into account anymore 2024-07-23 23:12:05 Any idea what could be happening? 2024-07-23 23:12:31 less 661 (POSIX regular expressions) 2024-07-23 23:37:58 quinq: i normally don't use a .lesskey file, but whipped one up and less (less 661) seems to be using it 2024-07-23 23:38:14 Question to see if someone can point me in the right direction or give me some hints to dig deeper to try and figure this out. I am running 30 alpine linux VM's with the virt kernel. They are all up to date and running the current packages. I have been using kexec-tools and set it up like the wiki for the init.d files. From time to time it hangs at the kexec reboot for a minute or so before it actually reboots. Then once it does this I have to 2024-07-23 23:38:14 stop the kexec-load service and do a full reboot. It will run fine with normal with sub 9 seconds reboots using kexec, but then after a week or so it starts to hang for a minute or longer before I have to do the full reboot. I know lot of info here, but any insights of why would be awsome. Thanks 2024-07-24 02:05:33 something changed with zfs on 3.20? 2024-07-24 02:05:41 got a `modprobe: ERROR: could not insert 'zfs': Invalid argument 2024-07-24 02:10:50 seems like an `apk upgrade -al` + reboot did it huh 2024-07-24 02:16:53 :q 2024-07-24 05:27:47 j_v, that's my ~/.lesskey: https://clbin.com/6mDP4 2024-07-24 05:28:08 And yet when I press 't', it tells me “No next tag (press RETURN)” 2024-07-24 05:28:36 And with 's', “Input is not a pipe (press RETURN)” 2024-07-24 05:29:25 Nor can I quit with 'b', I have to press 'q', which should be deactivated 2024-07-24 05:35:43 quinq: busybox less versus non-bb less? 2024-07-24 05:38:43 It's using actual less 2024-07-24 05:39:01 $ command -v less 2024-07-24 05:39:01 /usr/bin/less 2024-07-24 05:39:06 $ apk info -W /usr/bin/less 2024-07-24 05:39:06 /usr/bin/less is owned by less-661-r0 2024-07-24 05:39:32 Also I straced it, is opens correctly m. ~/.lesskey, no other lesskey file 2024-07-24 05:39:41 s/m./my/ 2024-07-24 05:40:06 I run it with LESS=, and there's no other LESS* variable in my env 2024-07-24 09:08:39 konman: have you considered zfs instead? Takes a bit of learning but has many advantages . Zfs not so good if you need to alter disc sizes. 2024-07-24 10:44:29 /22/22 2024-07-24 13:54:32 I'm writing some custom startup scripts and trying to analyse the boot time, is there anything akin to systemd-analyze for OpenRC? 2024-07-24 15:56:21 you wish 2024-07-24 17:26:14 i can mount cdrom easily in dosbox-x but not in dosbox, would be nice to have it al if its stable 2024-07-24 17:35:58 i mean inside guest os 2024-07-25 03:44:15 hello 2024-07-25 08:13:22 hi 2024-07-25 08:49:36 hey 2024-07-25 09:08:41 Hello fello alpine-rs 2024-07-25 09:08:46 fellow 2024-07-25 09:09:21 I stumbled yesterday on an Alpine-based distro specifically built for Java named "Alpaquita" from BellSoft 2024-07-25 09:10:11 It claims it has fer better performance than Alpine standard because it uses a different Kernel, and even has a version built with gblibc 2024-07-25 09:10:15 glibc 2024-07-25 09:11:16 I tried it yesterday and found the claimed gains are truly ....minimal, compared to Alpine 3.20 2024-07-25 09:11:38 like less than 1% on a old Xeon 2676v3 haswell cpu 2024-07-25 09:13:13 https://bell-sw.com/assets/files/WP_Alpaquita_Linux.pdf#page=5 2024-07-25 09:13:55 they claim a lot of kernel "optimizations" I was wondering if is worth including them in the "official" Alpine ? 2024-07-25 09:14:17 (just wondering) 2024-07-25 09:14:58 sure 2024-07-25 09:16:28 we already have NO_HZ_FULL 2024-07-25 09:17:07 we have NUMA enabled for all the 64 bit arches 2024-07-25 09:18:05 "BellSoft musl is built with -O3 for internal, malloc, and string subsystems and -O2 for the rest (Alpine musl uses -Os for all subsystems" 2024-07-25 09:18:27 this is sort of interesting. We could probably build malloc and string funcs with -O3 as well 2024-07-25 09:18:43 and I think we build musl with -O2 already 2024-07-25 09:21:32 -O3 = https://i.imgur.com/Je9bfca.mp4 2024-07-25 09:23:36 I tried both glibc&musl "performance optimized" from them yesterday, differences were really close to none, I wonder where they got those charts 2024-07-25 09:24:42 I mean from standard Alpine 3.20 and the Alpaquita "stream" latest version (rolling I guess) 2024-07-25 09:32:06 default musl build actually sucks less than I thought even against jemalloc 2024-07-25 09:32:46 there is marginal difference in pure I/O latency even for multithreaded qemu 2024-07-25 11:56:58 Hi guys. Trying to sign in to https://gitlab.alpinelinux.org/alpine/aports/-/issues using gitlab.com authentication (already signed into the gitlab.com) I get ` Sign-in using GitLab.com auth failed ` after clicking authorize... 2024-07-25 11:59:22 frojnd: does it mention some kind of request id? 2024-07-25 12:01:09 ikke: in url I have state and callback code params 2024-07-25 12:01:35 But on the web page I just have https://imgur.com/a/yDQMDfh 2024-07-25 12:02:13 "Email has already been taken" 2024-07-25 12:02:35 Well let me think.. 2024-07-25 12:02:50 Was I so stupid, I registered with the same email I'm using for gitlab.com ... 2024-07-25 12:02:57 Instead of just using auth 2024-07-25 12:03:21 frojnd: github 2024-07-25 12:05:01 Oh man ur right ikke ! 2024-07-25 12:05:18 Thank you! 2024-07-25 12:05:24 You're welcome 2024-07-25 12:18:07 Can I disable mounting of debugfs on /sys/kernel/debug? 2024-07-25 12:21:50 It's mounted in /etc/init.d/sysfs, but apparently not optional 2024-07-25 15:06:02 which package provides ss(8) (or was it (1)) 2024-07-25 15:06:48 https://pkgs.alpinelinux.org/contents?file=ss&path=&name=&branch=edge&arch=aarch64 2024-07-25 15:07:26 You can also run: apk search -x cmd:ss 2024-07-25 15:08:07 aarch64/iproute2-ss 2024-07-25 15:08:10 ah, thanks 2024-07-25 15:09:41 hi. since: https://git.alpinelinux.org/aports/commit/?id=e166e939682aeae846e132579a3634c71a485f26 every few minutes.-> "page unresponsive" + ends with an "aw, snap". x86_64 2024-07-25 15:10:08 oleovit: I think ncopa ran into that as well 2024-07-25 15:12:40 @ikke yep, looks like he and me are the only ones that use chromium on alpine :) 2024-07-25 15:12:47 This also happened a few months ago. "revert to clang and compiled-in llvm libc++" fixed it, as far as I can remember. 2024-07-25 15:14:32 maybe, has something to do with all these patches. can't work anymore. freeze too often 2024-07-25 15:17:17 oleovit: Could you perhaps create an issue on gitlab? 2024-07-25 15:19:38 @ikke gitlab will freeze. i tried it already. downloading firefox and setting it up in order to create an issue... i am not happy about that... 2024-07-25 15:23:37 Well, again, nothing works, everything freezes after a while. Maybe I will download this fiefox alternative (what's the name...) and write an issue. 2024-07-25 15:24:42 oleovit: i switched to firefox today for this reason 2024-07-25 15:26:49 @ncopa hi, well i am trying librewolf. but only temporarily. have you written an issue or should i write an issue? 2024-07-25 15:28:00 i guess, lauren knows already... 2024-07-25 15:31:55 i would appreciate if you wrote, thanks 2024-07-25 16:06:29 It's been so long since I've been in your gtlab. Don't laugh, but "what project should I select for this issue", regarding Chromium? 2024-07-25 16:06:58 alpine/aports 2024-07-25 16:07:09 https://gitlab.alpinelinux.org/alpine/aports/-/issues/ 2024-07-25 16:08:10 thx for the reminder. the new browser feels awful. 2024-07-25 16:34:18 weird, i've tried to reproduce this since today morning and i didn't run into a single crash/hang yet... 2024-07-25 16:34:51 version is 127.0.6533.72 indeed 2024-07-25 16:42:00 oh. 2024-07-25 16:42:05 i had some system libraries not updated yet 2024-07-25 16:42:06 now it hangs 2024-07-25 16:42:13 i think it might not be chromium itself? 2024-07-25 16:42:56 https://tpaste.us/8M1R 2024-07-25 16:43:01 not sure which one of those could cause that 2024-07-25 16:50:37 okay, i managed to hang it in less than 2 minutes with libpulse-17.0-r2, after downgrading back to r1 i can't manage it again 2024-07-25 16:50:44 it *could* be a false negative 2024-07-25 16:50:50 but maybe worth a try? 2024-07-25 16:51:17 ncopa: do you wanna try rebuilding libpulse locally from aports sha cc2f22db47369930dbbaed5ec7fd4d31a1f00415 and see if that works? 2024-07-25 16:52:11 ( with !check, because they otherwise fail ) 2024-07-25 16:52:37 oh. spoke too soon, it's not libpulse; the changes didn't make sense anyway 2024-07-25 16:52:57 mesa..? 2024-07-25 16:53:49 no idea 2024-07-25 16:54:53 im trying to figure if I can use distrobuilder and incus to test tiny-cloud 2024-07-25 17:19:51 ok, done - with chromium. had luck + wish good luck to find/fix the issue 2024-07-25 17:39:52 podman on fresh alpine install, getting Error: cannot re-exec process to join the existing user namespace when doing podman ps -a 2024-07-25 17:40:09 have 3 other machines where podman is working fine, trying to narrow down what is not working 2024-07-25 17:41:41 strace shows: running `/usr/bin/newuidmap 4461 0 1000 1 1 200000 500001`: newuidmap: Coul 2024-07-25 17:41:43 d not set caps\n 2024-07-25 17:42:14 still debugging but if it rings a bell, appreciate any tips 2024-07-25 17:43:25 also: yes this is when trying it without root 2024-07-25 18:43:01 Hi all :-) 2024-07-25 18:43:29 o/ 2024-07-25 18:43:30 I'm trying to install armhf bootstrap on x86_64 2024-07-25 18:44:37 bootstrap.sh script generate the error : cannot find -latomic: No such file or directory 2024-07-25 18:44:40 any idea ? 2024-07-25 18:47:44 (I tried installing libatomic but still the same error) 2024-07-25 18:50:47 ((oups ... I may need libatomic for guest platform)) 2024-07-25 19:11:23 has zstd been audited? 2024-07-25 19:12:00 WhyNotHugo: in what context? 2024-07-25 19:39:21 ACTION waits for the punchline 2024-07-25 19:39:28 thought maybe it was setting up a facebook joke 2024-07-25 19:44:45 Sorry to disappoint. 2024-07-25 19:45:10 I meant it in the security audit sense. 2024-07-25 19:45:24 in general, or by alpine linux? 2024-07-25 19:45:29 Er, both? 2024-07-26 03:03:13 my podman error: newuidmap could not set caps, was actually the result of bad data in my .xdg directory *sigh* starting fresh fixed it. 2024-07-26 05:49:01 hello hello 2024-07-26 05:50:02 Hi! I'm running an Alpine Linux image on OCI, and I want to enable a password for the "alpine" user, or perhaps a different user, so that I can log in via the remote serial console they have for troubleshooting if SSH doesn't work 2024-07-26 05:50:16 but I don't want to allow SSH logins with a password, only with a key 2024-07-26 05:50:20 is that a thing i can do? 2024-07-26 05:55:17 sure, disable password logins in your SSH service 2024-07-26 06:00:22 ah so just need to look up how to do that in openssh then? 2024-07-26 06:00:29 i believe that's what this image runs 2024-07-26 06:03:26 okay one other question: how would I set an environment variable such as `JAVA_HOME` so that all applications use it? I know /etc/profile for interactive shells but I want to make sure that my Minecraft server launch script and such picks it up 2024-07-26 06:04:59 in short you wouldn't, I want to say 2024-07-26 06:05:15 environment for users and environment for services is somewhat separate 2024-07-26 06:06:35 I don't know openrc well, but afaik the config locations for those don't really overlap... if one service needs an envvar, probably best to set it in that service's launch script specifically 2024-07-26 06:12:35 or the service's config file in /etc/conf.d 2024-07-26 06:18:32 i see 2024-07-26 06:19:45 so for a minecraft server service I'd have it set JAVA_HOME in the service script/file, and if there are any environment variables specific to a specific server I'd have those in a file that the generic service would read? 2024-07-26 06:19:57 also do y'all know which firewall the alpine linux OCI images use? 2024-07-26 06:20:22 i've set up an ingress rule in my dashboard but i think the firewall is blocking me now 2024-07-26 06:21:37 I'd suggest to read the relevant openrc documentation 2024-07-26 06:22:30 it's not that long, but is a really great doc 2024-07-26 06:22:50 there are answers for a lot of your questions 2024-07-26 06:29:43 okay i ended up `cat`-ing /etc/init.d/networking in the hopes it would say something about a firewall but alas no dice 2024-07-26 06:29:49 i hope i don't have to restart the instance 2024-07-26 06:30:37 you did what? 2024-07-26 06:32:40 i thought it might have information on firewalls but it appears to be just actual networking 2024-07-26 06:32:48 i don't think there is a firewall installed 2024-07-26 06:33:05 i think i am going to restart the VM and hope it picks up the new rules I set for the virtual subnet 2024-07-26 06:33:59 ah, I see 2024-07-26 06:35:48 also Minecraft is complaining about only having 1GB of RAM but I have zero dollars in my bank account so who's really got it worse, huh? 2024-07-26 06:36:00 not even 1GB actually i had to limit to 512M lmao 2024-07-26 06:36:07 because 1GB is the total for this VM 2024-07-26 06:36:34 yeah, java loves memory, for breakfast, for lunch, for dinner :) 2024-07-26 06:39:10 oh wait i think i know what i did wrong 2024-07-26 06:39:28 looking at the security list (basically port forwarding) for this subnet 2024-07-26 06:39:47 the default/built-in SSH has All for source ports and 22 for destination ports 2024-07-26 06:39:54 I misinterpreted what source and dest mean here 2024-07-26 06:40:07 i think if i change source to all and leave dest as 25565 I should be able to connect 2024-07-26 06:40:51 source means the client's port, dest is the sshd's port 2024-07-26 06:41:01 yeah i see that now 2024-07-26 06:41:22 minecraft clients don't connect only from 25565, that's just the destination port 2024-07-26 06:41:24 and also 2024-07-26 06:41:30 LETS FUCKING GO IT'S WORKING 2024-07-26 06:41:35 *hacker voice* i'm in 2024-07-26 06:41:41 the minecraft server hosted on a sad sad VM 2024-07-26 06:42:09 this is what you wanted, I suppose 2024-07-26 06:43:53 yeah i gotta give oracle money for their fancy ARM VMs now 2024-07-26 06:44:07 their support forums are full of people saying they can't make Arm VMs anymore on free tier 2024-07-26 06:44:24 so i think they must have silently decided there were too many freeloaders 2024-07-26 06:44:30 aaaand oracle loves your money :P 2024-07-26 06:44:43 they're only getting two bucks a month if I can help it 2024-07-26 06:47:04 even if it's not free two bucks a month was the estimate for the still very small VM i need for this minecraft server 2024-07-26 06:49:40 also i scheduled a call with their sales so hopefully i can ask them how to pay as little as possible and the sales person will ironically probably be able to help with that 2024-07-26 06:50:15 lol 2024-07-26 06:52:11 anyway, gl with that 2024-07-26 07:24:12 tyty 2024-07-26 07:24:47 its 3am here and the testing (blowing up a bunch of shit up to see how it handled it. poorly, by the way, it handled it poorly) is complete so i'm going to bed 2024-07-26 08:19:32 I'm trying to build a release from crystal lang code: crystal build --release --static --no-debug src/big-data.cr and I get: /usr/lib/gcc/x86_64-alpine-linux-musl/13.2.1/../../../../x86_64-alpine-linux-musl/bin/ld: cannot find -lrdkafka (this usually means you need to install the development package for librdkafka): No such file or directory I searched in apks and I get 2024-07-26 08:19:34 librdkafka and librdkafka-dev but when I install either package or both, I still get the same error. Do I need librdkafka-libs something? 2024-07-26 08:22:40 Error: execution of command failed with exit status 1: cc "${@}" -o /home/frojnd/work/big-data/big-data -rdynamic -static -L/usr/bin/../lib/crystal -lrdkafka -lpthread -lrt -ldl -lm -L/lib -lz -lssl -lcrypto -ldl -pthread -lsasl2 -ldl -lresolv -lresolv -lzstd -pthread -llz4 -lz -lpcre2-8 -lgc -lpthread -ldl -levent 2024-07-26 08:23:50 I also get: cannot find -lsasl2 (this usually means you need to install the development package for libsasl2): No such file or directory 2024-07-26 08:25:08 Installing libsasl-2.1.28-r6 didin't help 2024-07-26 08:28:32 well, I see a -static there 2024-07-26 08:28:42 you need static libraries in this case 2024-07-26 08:28:54 which are not provided by the mentioned pkgs 2024-07-26 08:29:24 but the whole thing is odd 2024-07-26 08:29:30 -rdynamic and -static? 2024-07-26 10:23:00 Yeah not sure 2024-07-26 10:23:30 I'm building with static flag so when I deploy to e.g. debian all it has to be done is just run big-data binary file and kafka starts kickin 2024-07-26 11:04:50 aron so I would have to build a static package somehow? For alpine? 2024-07-26 11:05:11 or just compile as dynamic 2024-07-26 11:05:16 (remove -static) 2024-07-26 11:05:54 I wouldn't build static if it wasn't supplied originally 2024-07-26 11:06:17 Whole point of static is I don't need to install dependencies on other distros 2024-07-26 11:06:32 what other distros? 2024-07-26 11:07:00 (btw, "install dependencies": this is package manager is used for) 2024-07-26 11:07:29 Like debian 2024-07-26 11:07:45 that binary you build on alpine will NOT work on debian, whatsoever 2024-07-26 11:09:05 It should if built with --static flag. I have built other crystal apps with --static flag and just pushed binary to debian and ran it (and it worked) 2024-07-26 11:09:41 different distros (sometimes the same distro but different versions) are not binary compatible -- it MAY work, but don't take it as granted 2024-07-26 11:10:38 no, it not "should", it _may_ work 2024-07-26 11:11:00 if you are lucky, and kernel versions and kernel configurations match 2024-07-26 11:12:06 hmmm 2024-07-26 11:12:09 I discourage you to do this, but you can do what you want, ofc 2024-07-26 11:12:46 I've installed a bunch of apks today and would like to remove them. Like filter installed apks by date 2024-07-26 11:13:07 Was messing around with those static libs... 2024-07-26 11:14:12 sorry, I never faced this problem, I never do this on my system, I do it in other VMs/chroots/etc 2024-07-26 11:14:13 gtg 2024-07-26 11:15:46 ok ty 2024-07-26 11:21:55 this might be more of a question for -devel, but: what the heck is going wrong with my bootstrap here? https://termbin.com/rgh7 2024-07-26 12:10:23 did anyone reply while my isp decided not to supply internet for a bit? 2024-07-26 12:10:56 no 2024-07-26 12:43:36 What is 'enum-header'? I can't compile dwl window manager. 2024-07-26 18:06:56 hi, question, what should be the expected environment to have dracut and update-grub successfully do their duties? 2024-07-26 18:07:09 just a working and accurate /proc and /dev? 2024-07-26 18:28:05 is it possible to build a package dependency diagraph of only the already installed packages? 2024-07-26 18:30:56 apk dot --installed 2024-07-26 18:32:08 ah thanks, well im processing the dot output of all the packages from world and its taking ages to complete because its so huge 2024-07-26 19:12:37 Does the /usr merge solve any technical issues? Everything that I find online seems to indicate "it fixes hardcoded shebangs across platforms". 2024-07-26 19:13:20 I found this BS 2024-07-26 19:14:37 Hi, what's the doas equivalent of `sudo -sE`? I want to retain my invoking user's environment and just have root privs added. 2024-07-26 19:17:28 ForeverNoob[m]: I'm not aware of a CLI option, but there is a configuration setting to enable it 2024-07-26 19:17:32 keepenv 2024-07-26 19:18:06 ForeverNoob[m]: you want keepenv, from doas.conf 2024-07-26 19:18:46 If you're referring to `permit persist keepenv :wheel` and then do `doas -s me` that one doesn't seem to do the trick, but maybe I'm just misreading the man page. 2024-07-26 19:23:02 It does work for me 2024-07-26 19:23:24 Hmm, do I need to logout/login for this? 2024-07-26 19:23:38 no 2024-07-26 19:24:05 I just added keepenv to the config, ran doas -s (or doas -s -u me), en checked whether an environment variable I just declared was set 2024-07-26 19:25:05 Was your PS1 sourced like the invoking user? 2024-07-26 19:33:08 Is it even possible to compile dwl on Alpine? 2024-07-26 19:34:11 PS1 is a local variable, not an environment variable 2024-07-26 19:45:37 Hmm, so I guess it's not equivalent to `sudo -sE` then. If I have `permit persist keepenv :wheel` as the only thing in doas.conf, and then do either `doas -s` or `doas -s -u me`, then in the former none of the aliases get retained, and in the latter it doesn't have root privs. 2024-07-26 19:46:40 Doas only keeps environment variables. aliases are not environment variables 2024-07-26 19:47:36 doas is a lot simpeler by design, so it does not try to preserve your entire shell 'environment' 2024-07-26 19:50:32 Hmm, in that case I'll try to find a workaround for this and post updates if I do. 2024-07-26 20:09:32 Just sudo 2024-07-26 20:26:15 Ok, so suppose I want to run something in `/etc/local.d/` as a normal user, and have access to that user's environment and most importantly $PATH, what would the best way be to do that? 2024-07-26 20:45:25 su -l 2024-07-26 20:48:48 Doesn't that give me interactive login shell? How would that work if I want to automatically start something at boot? 2024-07-26 20:49:23 su can work as sufo 2024-07-26 20:49:46 Depends on how you pass the parameters 2024-07-26 20:52:54 You can also use runuser from util-linux 2024-07-26 20:56:08 Might try that because `su -l` does not load target user's $PATH 2024-07-26 21:02:04 Well that doesn't do it either. 2024-07-26 21:02:26 Man, I thought this would be simpler considering all things Alpine :P 2024-07-26 21:06:54 Ok, figured it out. Here it comes... 2024-07-26 21:08:35 As root: user_path=$(doas sh -c 'su - me -c "source ~/.profile ; source ~/.bashrc 2>/dev/null ; echo \$PATH"') ; doas su - me -c "PATH='$user_path' command_to_run" 2024-07-26 21:09:58 Well I sure am glad that's over with :D 2024-07-26 21:19:33 Probably pam does the setup on arch 2024-07-27 02:41:01 quit 2024-07-27 02:41:19 ACTION quits 2024-07-27 03:32:07 can 'busybox more' be made to pause at the end, even if the text fits the screen? so if a new terminal is launched to display text, one will see it 2024-07-27 03:38:46 frag: after looking at busybox's more.c, i think not. all options i saw that it accepts are ignored 2024-07-27 03:41:05 ah thanks j_v! im not even sure if other implementations allow it either, havent used more in decades... 2024-07-27 03:41:44 looking at manpage from util-linux, seems not, but i might be missing something 2024-07-27 03:42:03 this is something i would like also 2024-07-27 03:49:14 could use busybox's less instead 2024-07-27 03:55:01 yes have been using less, but wanted to get more used to posix, and more is suprisingly nice... 2024-07-27 05:14:17 Y'all think there would be interest in an Alpine Linux package of scripts for running Minecraft servers? I'm going to be writing this kind of thing anyways so I wonder if I could make it more generic to be able to be used by anyone 2024-07-27 05:17:26 I don't want it to be like, a whole panel. Maybe integrate into ACF to avoid having to write my own web app. But the intended use case would be on Alpine first and foremost and it's not meant to be a panel to be dropped in on any old distro 2024-07-27 05:18:14 but like, I want it to be able to handle managing Java versions, updates, backups, stuff like that 2024-07-27 05:19:38 maybe managing containerization if I go that route but I'm not certain 2024-07-27 05:20:06 i want to keep is simple 2024-07-27 05:33:50 if someone thinks of a good name for an Alpine Linux + ACF Minecraft server panel/server manager daemon leave it here and I'll see it when I peruse the logs later 2024-07-27 05:35:21 also how exactly is s6 used on Alpine? Like why is it used in addition to OpenRC for some things? 2024-07-27 05:54:28 hi, I'm trying to install alpine-sdk for the first time and I become the error "unable to select packages: alpine-sdk (no such package)" 2024-07-27 05:54:49 how can I proceed? I can't find any answer to this problem 2024-07-27 05:59:22 glitchapp: what happens if you do: apk update 2024-07-27 06:03:00 ok: 408 distinct packages available 2024-07-27 06:07:59 that's... not many packages heh 2024-07-27 06:08:33 (just for comparison, on my 3.20 system: OK: 24181 distinct packages available) 2024-07-27 06:09:06 mm I downloaded the extended alpine iso and installed on qemu, am I doing anything wrong? 2024-07-27 06:09:46 does your alpine system have a connection to the internet? (can you, eg, ping stuff? does dns work? etc) 2024-07-27 06:09:53 ah let me check 2024-07-27 06:15:11 no there is no network, ping returns error, I don't think the set up on qemu is the problem because I have the exact same network configuration on other virtual machines and they work, so apparently alpine has no network 2024-07-27 06:16:05 I have linux and windows vm with same network configuration and they all have internet 2024-07-27 06:18:14 glitchapp: maybe check https://wiki.alpinelinux.org/wiki/Configure_Networking 2024-07-27 06:18:26 ok it turns out I need to do alipine-setup first 2024-07-27 06:18:38 ideally setup-alpine takes care of it for you but if you did that and it didn't work, then you can see what's missing 2024-07-27 06:18:43 yes 2024-07-27 06:19:37 I'm doing it now, hopefully it works 2024-07-27 06:19:51 may the odds be ever in your favour 2024-07-27 06:23:01 now it works thanks 2024-07-27 06:24:06 awesome 2024-07-27 07:17:12 I'm stucked now in the abuild process, I become the error: rootbld: mapath ../.roobld-repositories does not exist" and sh: aportsdir: not found 2024-07-27 07:18:18 please keep your questions to a single channel 2024-07-27 07:19:56 allright 2024-07-27 07:32:46 Hello 2024-07-27 07:37:32 I would like to use clang-format version 18 in alpine linux, however it seems the default version is 17. Version 18 is packaged, but the package doesn't install itself on to the path. When will version 18 become the new default? 2024-07-27 07:46:18 mber: it is already the default on edge 2024-07-27 07:56:15 I'm just using the `alpine` docker image from dockerhub. How exactly does it all work? E.g. when would I expect that image to provide clang 18? 2024-07-27 07:56:59 that would be alpine:3.21 when it is released around november 2024-07-27 07:57:11 or alpine:edge now, but with the caveats of running edge 2024-07-27 07:57:28 (alpine:latest would be 3.21 when it's released) 2024-07-27 07:59:58 Ah okay. So I guess even if they start packaging a newer version of clang, they won't make it default until the next release of alpine. 2024-07-27 08:02:35 Not immediately no 2024-07-27 08:06:57 re: As root: user_path=$(doas sh -c 'su - me -c "source ~/.profile ; source ~/.bashrc 2>/dev/null ; echo \$PATH"') ; doas su - me -c "PATH='$user_path' command_to_run" 2024-07-27 08:07:24 How would I export a variable this way? 2024-07-27 08:08:11 My Bash $PS1 gets colored differently depending on a variable being set to 0 or 1, but I can not for the life of me figure out how I'd do that. 2024-07-27 08:09:08 I've never tried to run anything as another user on startup in Linux, but oh wow, I feel like this seems much more difficult than it should be :( 2024-07-27 08:11:55 ForeverNoob[m]: what is it that you're actually trying to do? 'cause I can't quite figure out what that command is doing 2024-07-27 08:13:03 well. I can kinda see what it's trying to do but it doesn't really make any sense to me hy 2024-07-27 08:13:06 er, why 2024-07-27 08:16:25 fission: Basically I want to start a tmux sesion as a normal user on boot, so I put all of that inside /etc/local.d/02_run_tmux.start and so far it works (after _many_ attempts to do it right), but the PS1 values inside tmux are not colored in a certain way I want, and those colors are dependent on a variable which I can not figure out how to set. 2024-07-27 08:17:38 which variable are you trying to set? 2024-07-27 08:18:57 $ISREMOTE 2024-07-27 08:19:47 And inside that script, $ISREMOTE is supposedly set to 1, but it's just not being reflected in my PS1, which defaults to a default value it seems. 2024-07-27 08:21:48 I mean, you can try just doing: export ISREMOTE 2024-07-27 08:21:59 I dunno whether it will really work (as I don't know enough about the script or what is launching what) 2024-07-27 08:26:58 btw... I'm a lazy person, so normally I just start user processes at boot using a cron @reboot job 2024-07-27 08:27:13 and then there's no fiddling around with doas/su/etc 2024-07-27 08:27:29 I tried adding that in like all points of that command, and the script is just a bunch of tmux commands that make a bunch of windows and panes. 2024-07-27 08:27:33 bb cron does not support @reboot 2024-07-27 08:27:54 I think I'm using cronie 2024-07-27 08:28:01 er 2024-07-27 08:28:12 ikke: are you sure? :) 2024-07-27 08:28:27 'cause I have @reboot jobs and they are definitely kicking off 2024-07-27 08:28:35 photon:~# grep cron /etc/apk/world 2024-07-27 08:28:36 photon:~# 2024-07-27 08:28:40 and I don't have any special crons on that box 2024-07-27 08:29:11 phttps://stackoverflow.com/questions/35282187/alternative-to-reboot-for-busybox-crond 2024-07-27 08:29:44 photon:~# grep -q @reboot /var/spool/cron/crontabs/root ; echo $? 2024-07-27 08:29:45 0 2024-07-27 08:29:52 it definitely works 2024-07-27 08:30:27 https://github.com/nawawi/busybox/blob/master/miscutils/crond.c it's a configurable feature 2024-07-27 08:30:42 //config:config FEATURE_CROND_SPECIAL_TIMES 2024-07-27 08:30:49 //config: bool "Support special times (@reboot, @daily, etc) in crontabs" 2024-07-27 08:31:13 //config: default y 2024-07-27 08:31:15 ok, TIL then 2024-07-27 08:31:57 CONFIG_FEATURE_CROND_SPECIAL_TIMES=y 2024-07-27 08:33:03 I suppose since dc6cfb7cebc64cdb278f02d32f897819bba5f983 2024-07-27 08:33:14 ForeverNoob[m]: maybe you can set it in the user's .profile or .bashrc or whatever 2024-07-27 08:33:58 cool 2024-07-27 08:42:57 Hello all ^^ 2024-07-27 08:43:14 I have a (possibly dumb) question, don't BBQ me 2024-07-27 08:43:38 For my workplace I'm tring to use ALpine as a Desktop OS 2024-07-27 08:44:05 it works beautifully but I really need (not me, PHP developers....) a program 2024-07-27 08:44:20 a terrible program that everybody needs 2024-07-27 08:44:32 Visual Studio Code 2024-07-27 08:44:36 (I know....) 2024-07-27 08:44:56 there's a REMOTE possibility to run it on Alpine XFCE desktop ? 2024-07-27 08:45:12 wyk72: there is codium 2024-07-27 08:45:58 I can obviously run it into a VM or a container, but I have no idea on how to make it work in a "flawless" way 2024-07-27 08:46:10 code-oss 2024-07-27 08:46:15 in testing 2024-07-27 08:46:22 I tried it 2024-07-27 08:46:27 code-oss 2024-07-27 08:46:40 but I have no idea on how to launch it 2024-07-27 08:46:57 `code-oss` 2024-07-27 08:47:18 zodiac3:~# code-oss -sh: code-oss: not found 2024-07-27 08:47:40 and you have code-oss installed? 2024-07-27 08:47:45 yup 2024-07-27 08:47:59 what does `which code-oss` return? 2024-07-27 08:48:41 if you installed vscodium you launch it via codium 2024-07-27 08:49:10 i dont recommend code-oss because its very barebones, codium tidies it up for proper consumption 2024-07-27 08:49:24 We don't have a package for vscodium 2024-07-27 08:49:49 fission: Hmm, I'm trying to keep those files uniform across multiple instances, but I'll take a look at that @reboot option, that might be a much more elegant solution as you've stated. 2024-07-27 08:49:59 ok 2024-07-27 08:50:16 it seems to work! I missed one package (electron) 2024-07-27 08:50:28 ForeverNoob[m]: cool, hope it works for you 2024-07-27 08:50:34 Hmm, you should not have to install that manually 2024-07-27 08:51:18 it's because I kept packages masked @edge @testing I think 2024-07-27 08:51:39 this ugly thing starts 2024-07-27 08:51:45 I use code-oss myself, but for what I need, it suffices 2024-07-27 08:52:59 thanks a lot ! 2024-07-27 08:55:01 how to install codium ? Is in edge repo ? testing ? 2024-07-27 08:56:27 Like I said, we don't have it as far as I know. We do have code-oss 2024-07-27 09:17:30 another dumb question: how to install node.js specific version without compiling it ? I need 20.12.2 .... 2024-07-27 09:17:49 latest Alpine 3.20 installs node 21.x 2024-07-27 09:17:51 We only provide 2 versions 2024-07-27 09:18:12 nodejs and nodejs-current 2024-07-27 09:20:19 Oh I see 2024-07-27 09:20:38 for some reason I installed -current 2024-07-27 09:20:49 now it's ok 2024-07-27 09:25:51 this nodejs stuff is a pain in the a** 2024-07-27 09:26:05 (/opinion) 2024-07-27 09:27:59 as much as I love Alpine (I use on all my servers) coping with systemd deps, nodejs absurdities and most of all, glibc, it's becoming a losing battle recently 2024-07-27 09:28:08 vm/container usage is not always viable 2024-07-27 09:28:26 dire situation overall 2024-07-27 09:29:15 I'm amazed ppl still uses ubuntu this much 2024-07-27 09:29:23 canned distros everywhere 2024-07-27 09:29:39 for the most part ubuntu got better 2024-07-27 09:29:47 apt is way spedier 2024-07-27 09:29:53 *faster 2024-07-27 09:29:54 lol 2024-07-27 09:30:49 still have the crazy amount of in-house package modifications and absurd subpackages, but it gets the job done 2024-07-27 09:30:50 yeah, but systemd horrors....bulkiness....snapd monstrosities everywhere...you need 200GB hdd to run 16k apps 2024-07-27 09:31:01 systemd horrors? 2024-07-27 09:31:30 also apt is good until..it doesn't 2024-07-27 09:32:18 I have to run apt --fix-broken-install a lot of times, because it just sucks overall 2024-07-27 09:33:38 anyway...still better than running Crowdstrike on windows (good move enterprises, good move...) 2024-07-27 09:34:12 I kinda don't get it 2024-07-27 09:34:24 The whole hysteria 2024-07-27 09:34:33 of what? 2024-07-27 09:34:35 Two-three days of outage? 2024-07-27 09:34:36 Meh 2024-07-27 09:35:08 People mising flights, hospitals needing to send people home? 2024-07-27 09:35:09 We had a world-wide lockdown three years back 2024-07-27 09:35:17 Millions of dead motherfuckers 2024-07-27 09:35:24 some 911 (US) services were down in some places 2024-07-27 09:35:29 ugh umm missed flight microsoft bad 2024-07-27 09:35:30 This becomes #alpine-offtopic 2024-07-27 09:35:34 yeah, true 2024-07-27 09:37:46 to stay on topic: lslogins doesn't show the last login day, given that utmps is up and running 2024-07-27 09:37:53 who and others work 2024-07-27 09:38:12 never heard of lslogins myself 2024-07-27 09:38:13 not only Microsoft, the whole "enterprise software environment" is pure shit 2024-07-27 09:38:19 it's part of util-liunx 2024-07-27 09:38:49 I benn working there: the only thing that matter is "how much money can I milk from it?" 2024-07-27 09:39:33 you either pay with your time or with your money 2024-07-27 09:39:35 so 2024-07-27 09:39:40 pick your poison :))) 2024-07-27 09:43:42 how long does it take graphviz to render the gigantic dot file that is the apk package dependency graph? 2024-07-27 09:44:41 Infinitely, since it generally runs out of memory 2024-07-27 09:44:55 xD 2024-07-27 09:44:55 consus: I do not agree at all, this is oversimplification 2024-07-27 09:45:58 ikke: oh come on, im genuinely wondering what it takes to render the 7.53mb monster 2024-07-27 09:46:37 I was only half joking 2024-07-27 09:47:46 Juest: I never managed to do it 2024-07-27 09:48:42 wyk72: enterprise usually has special needs, so the stack reflects that 2024-07-27 09:49:00 wyk72: i.e. the only thing that constantly crashes on my alpine box is nslcd for ldap 2024-07-27 09:49:39 not because it's that bad, but because nobody usually gets to test that, epsecially on musl 2024-07-27 09:50:14 consus[m]: wyk72 left 2024-07-27 09:50:18 oh 2024-07-27 09:50:25 well 2024-07-27 09:50:40 at least he's reminded me to go poke upstream for a release 2024-07-27 09:55:06 ikke: is it somehow possible to teach openrc not to corrupt logs on boot? 2024-07-27 09:55:26 https://bpa.st/YPGQ 2024-07-27 09:55:28 like this 2024-07-27 09:56:20 How is this corrupt? 2024-07-27 09:56:45 line breaks are messed up 2024-07-27 09:56:55 so it's a just giant trashy wall of text 2024-07-27 09:58:51 I don't think openrc is the cause of that 2024-07-27 09:59:15 it's rc.log 2024-07-27 09:59:17 that openrc writes 2024-07-27 09:59:26 AFAIR it's alright on the screen during the boot 2024-07-27 10:02:11 My guess is that fsck think it's writing to a tty and uses all this clever interactive escape codes 2024-07-27 10:03:04 yes, that would've been my guess as well 2024-07-27 10:03:53 Also something weird happens with listings 2024-07-27 10:03:57 * Starting networking ... * lo ... [ ok ]... (full message at ) 2024-07-27 10:04:10 like it seems that * lo should be on a separate line too 2024-07-27 10:07:26 ah, it should be einfo, not ebegin 2024-07-27 10:15:48 is 'seq 100|more' the right way to call the default pager? 2024-07-27 10:16:03 is 'seq 100|$PAGER' the right way to call the default pager? 2024-07-27 10:18:39 busybox has both 2024-07-27 10:18:41 less and more 2024-07-27 10:18:41 so 2024-07-27 10:19:42 PAGER=less is by default 2024-07-27 10:35:03 i mean, if you make a script, how do you call the default pager? 2024-07-27 10:38:44 might want have a default, something like `pager=${PAGER:-/bin/more}; some_action | $pager` 2024-07-27 10:41:46 assuming its a bourne shell script 2024-07-27 10:45:59 I would default to less by default, but it seems okay 2024-07-27 10:51:53 cool 2024-07-27 11:31:01 (2/221) Upgrading alpine-baselayout (3.6.5-r0 -> 3.6.6-r0) 2024-07-27 12:52:43 fission: Well the @reboot thing is definitely more elegant, but I'm still unable to truly export that variable. 2024-07-27 12:52:52 In my /etc/crontabs/me, only 2 lines: 2024-07-27 12:52:57 SHELL=/bin/bash 2024-07-27 12:53:03 @reboot source ~/.profile ; export ISREMOTE=1 ; /home/me/tmux_script.sh 2024-07-27 12:53:29 I'm starting to think this is like some anti-pattern in UNIX or something. 2024-07-27 12:53:50 ForeverNoob[m]: I have a feeling it's overwritten somewhere 2024-07-27 12:54:17 If you set it in .profile for example, then shell sessions in tmux would source that again 2024-07-27 12:57:13 Hmm, that might be the case. I'll dig around and try to simplify my test setup until I can find the culprit. 2024-07-27 13:31:26 any hint on how to play win95/win98 games on al ? 2024-07-27 13:32:20 wine 2024-07-27 13:32:22 I suppose 2024-07-27 13:34:09 tried, usually works, some of them cannot find cdrom or won't start 2024-07-27 13:34:50 even after setting emulattion to win97 in wine settings 2024-07-27 13:35:14 sorry win98^ 2024-07-27 13:36:19 ran win98 in dosbox-x and started ok 2024-07-27 13:41:29 will fresh install wine and retry 2024-07-27 13:43:37 Note that becuase Alpine does not support multilib, you need to run wine on an 32-bits system (or 32-bits chroot) to make it work with 32-bits applications / libraries 2024-07-27 13:51:40 32-bit chroot, example ? lxc ? 2024-07-27 13:51:46 or qemu 2024-07-27 13:52:26 trying now a general setup using i386 iso 2024-07-27 13:52:50 this will limit ram to 4gb, i guess that's ok to play those games 2024-07-27 16:17:47 in busybox grep, is it a bug that you have to add folder to the -r option? so 'grep -r fobar dir' doesnt work without the dir part, also seems to just work forever if there are pipes in the subdirs .. maybe the last part is normal.. 2024-07-27 16:18:15 frag: I think that's POSIX 2024-07-27 16:18:22 only gnu grep makes the dir optional 2024-07-27 16:18:44 grep from the grep package seems to handle pipes.. 2024-07-27 16:18:58 ikke oh i see 2024-07-27 16:20:46 been using grep from busybox as the daily driver a while now, not bad, but made a script to color output if needed, not too often needed tho 2024-07-27 16:22:04 hmm, this does not even list a -r option: https://pubs.opengroup.org/onlinepubs/9699919799/utilities/grep.html 2024-07-27 16:24:11 does anyone know what i should set resume to in grub config when i'm using lvm crypt on the swap partition? 2024-07-27 17:00:44 My internal microphone is detected as "unplugged" by pavucontrol in AL. This has never been the case with any other distro. Can someone help me? 2024-07-27 17:05:43 are other devices ok? 2024-07-27 17:07:29 Unfortunately, I don't have another known-working microphone to test out, but my internal speaker is working fine. Just like my external headphones 2024-07-27 17:11:09 And the internal speaker and microphone seem to use the same sound card, which makes me think that all the necessary drivers are installed 2024-07-27 17:15:33 are you using pipewire? 2024-07-27 17:16:32 Yes, in my .profile I first start pipewire, then wireplumber and then pipewire-pulse (not sure if important) 2024-07-27 17:18:21 Do you have eudev started? 2024-07-27 17:18:45 anon83928: ^ 2024-07-27 17:19:45 I think I'm using udevd 2024-07-27 17:20:52 this is eudev, yes 2024-07-27 17:21:14 can you please send output of pw-dump (through a pastebin) ? 2024-07-27 17:21:43 and also wpctl status 2024-07-27 17:22:15 Can you recommend a pastebin? 2024-07-27 17:22:51 tpaste.us 2024-07-27 17:23:43 Also, do you have dbus session bus started? (You can do it in your .profile: export $(dbus-launch) ) 2024-07-27 17:26:30 Here my "wpctl status": https://tpaste.us/RBNO 2024-07-27 17:27:04 And here my pw-dump: https://tpaste.us/9MpZ 2024-07-27 17:27:30 And yes I'm using "export $(dbus-launch)" in my .profile 2024-07-27 17:48:23 cron and env meh 2024-07-27 17:48:24 Just set the necessary env there 2024-07-27 17:49:33 I'm not sure if you mean me and if yes what you mean 2024-07-27 17:59:37 anon83928: can you please get me PW and WP logs also? 2024-07-27 18:01:15 You mean something like this?: $pipewire > logfile 2>&1 2024-07-27 18:01:30 yeah, that would do 2024-07-27 18:08:09 Here the logs from pipewire: https://tpaste.us/z5N1 2024-07-27 18:08:40 And here the logs from wireplumber: https://tpaste.us/XyQe 2024-07-27 18:11:14 anon83928: Is the capture on for that mic like so? https://0x0.st/Xf9X.png 2024-07-27 18:12:14 I'm not familiar with the tool 2024-07-27 18:26:01 well, logs were of no use... 2024-07-27 18:27:30 anon83928: btw, what is the name of device which was marked as unplugged? 2024-07-27 18:27:35 I'm not sure what it could be, but it must be specific to the default state of Alpine, because I never had this issue with any other distro 2024-07-27 18:27:54 anon83928: it's alsamixer 2024-07-27 18:28:12 I mean, what pavucontrol shows 2024-07-27 18:28:38 The name of the unplugged internal microphone is "Tiger Lake-LP Smart Sound Technology Audio Controller Stereo" 2024-07-27 18:32:56 ForeverNoob: On the default sound card it was at 50%, but when I selected the only sound card being sof-hda-dsp it was at 0% 2024-07-27 18:34:21 anon83928: sorry, may I ask you to alo send output of pactl status? 2024-07-27 18:37:41 Ermine: Of course, my "pactl stat" output is: https://tpaste.us/LRLW 2024-07-27 18:38:13 ohh, my bad, I meant pactl list 2024-07-27 18:39:44 No problem, here is the output of "pactl list": https://tpaste.us/Z5q6 2024-07-27 18:46:11 Maybe I should have said this earlier, but in pavucontrol a "monitor" mic also shows up and when used just mirrors my output. not sure if this is normal 2024-07-27 18:52:08 it's normal yes 2024-07-27 18:52:18 Oh, ok 2024-07-27 18:53:19 Not sure but could this unplugged problem have something to do with missing drivers or kernel modules, that are normally shipped on other distros? 2024-07-27 19:14:29 ACTION consults with pipewire devs 2024-07-27 19:16:32 I think I figured out which field in pw-dump output makes pavucontrol say it's unplugged. Now I need to figure out why pipewire decided to fill this field in the way it does 2024-07-27 19:21:37 Take your time, I'm just thankful that someones looking at this 2024-07-27 20:52:34 I have finished my quest for high quality bluetooth audio (aac and sbc-xq). Everything uses pipewire. 2024-07-27 20:52:34 Just created a PR with the necessary upstream patches for bluez-alsa. Please provide feedback and test if alpine works well with your bluetooth headphones 🙂 2024-07-27 20:56:09 Please test if you need to run `modprobe uinput` for for AVRCP to work. 2024-07-27 20:57:24 Please advise if I should add `touch /etc/modules-load.d/uinput.conf` to any package 2024-07-27 20:57:24 I am unsure, if this is releated to elogin.d-daemon, bluetooth or bluez-alsa. 2024-07-27 21:03:40 Why my pipewire doesn't work? 'wpctl status' prints "error relocating /usr/lib/libwireplumber-0.5.so.0: pw_log_set_level_string: symbol not found". 2024-07-27 21:03:40 i can't find anything from Google. 2024-07-27 21:07:00 Is wireplumber installed? `apk add wireplumber` 2024-07-27 21:09:41 and apk add `apk add wireplumber-libs` 2024-07-27 21:10:20 you can use apk-file to find out which packages provides the missing files. `apk-file /usr/lib/libwireplumber-0.5.so.0` 2024-07-27 21:23:15 On Linux, the quest for working audio is eternal. Doubly so for pro audio. 2024-07-27 21:38:02 Summary of the Bluez-alsa PR: Enabled A2DP, AAC, aptX, aptX HD, mSBC, LC3 SWB, UPower, and libfreeaptx. My goal is expanding codec support for better audio quality and device compat. https://gitlab.alpinelinux.org/alpine/aports/-/merge_requests/69824 2024-07-27 21:38:02 Feedback is welcome 🙂 2024-07-27 21:44:45 ForeverNoob[m]: my advice would be to run only one command in the @reboot, and make sure (by testing it "manually") that it behaves exactly as you expect first 2024-07-27 21:46:20 ForeverNoob[m]: just #! whatever script you do and chmod it appropriately 2024-07-27 21:54:26 I think I've figured out what was setting the ISREMOTE variable. I was trying to set it before and after it sourced ~/.bashrc, but apparently somewhere in a file it sourced was a condition based on the existence of $SSH_CLIENT (which is pretty dumb because apparently I tried to set it twice: Once in ~/.bashrc and another time in another file which got sourced) 2024-07-27 21:54:47 Oh well, another mystery solved. Thanks for the help all :) 2024-07-27 21:55:07 grand 2024-07-27 22:03:50 Hello 2024-07-27 22:10:34 Anyone have a working example of answers file for 3.20.2 ? When I set KEYMAPOPTS="us us", it keeps asking for flavor 2024-07-27 22:36:56 anon[m]: so far PW devs are not responding. Idk how much will it take 2024-07-27 22:52:11 Well if it doesn't pan out, thanks very much for all your effort anyways. I'm really at the end of my rope with this one 2024-07-27 22:55:22 I'll stick around though 2024-07-28 01:02:07 um, how am i supposed to upgrade to the next major version? sorry for asking 2024-07-28 01:16:42 "um, how am i supposed to upgrade..." <- Change /etc/apk/repositories to be targeting latest-stable instead of whatever version number, then doas apk update && doas apk upgrade -al 2024-07-28 01:22:12 thanks Saijin_Naib[m] 2024-07-28 01:34:51 hiya 2024-07-28 01:36:30 I'm planning to make an ACF module for a Minecraft server management tool. Basically something to start, stop, restart, and update servers. But I'm not planning to use an existing panel software, I want to use my ACF view as the panel. Do y'all think it makes sense to write the whole application (like, daemon/service parts) in Lua since I'll be interfacing with Lua a lot for the ACF module anyways? 2024-07-28 01:51:04 sure why not 2024-07-28 02:56:24 what is the support period for each version of alpine? 2024-07-28 03:02:26 Juest: https://alpinelinux.org/releases/ i found that information on the releases page 2024-07-28 05:01:27 @Juest any time. That tweak and apk-cron is how I have family setup 2024-07-28 05:34:34 Can ACF do websockets? Like, can I set up a tab/panel in my ACF model view and have it stream logs to a viewing box there over websockets? 2024-07-28 05:34:44 acf *view 2024-07-28 05:34:50 it's 1:30 am here i apologize 2024-07-28 05:37:31 I'm also not sure if my game server management app for ACF should be part of the ACF module itself or separate and the module just for talking to ACF 2024-07-28 05:37:37 or if it even makes a difference 2024-07-28 05:38:24 like i'm sure I can write the service/daemon part in whatever language I want but I was thinking I could use Lua since I can rely on it being there to begin with because of ACF which this would still be tightly coupled with 2024-07-28 05:38:52 i don't have a clear idea in my head of what the separation of concerns for this is, which I should probably work on 2024-07-28 05:39:14 would probably help if i were daily driving alpine instead of dual booting into it occasionally and running it on cloud servers 2024-07-28 05:41:28 chexo3: Doooo ittt 2024-07-28 05:42:33 Saijin_Naib[m]: Get off my ass and switch to Linux finally or get off my ass and make Minecraft server panel for ACF? 2024-07-28 05:43:03 Either way I lack a lot of the skills necessary and it's a bit overwhelming though I feel once I get started progress will likely be steady. 2024-07-28 07:55:52 Hi, I'm using the workstation setup guide at https://whynothugo.nl/journal/2023/11/19/setting-up-an-alpine-linux-workstation/ but having a bit of trouble with encrypted root, basically the root btrfs won't automatically mount after unlocking, but I can manually mount it under /sysroot from the initrd emergency console, then exit and boot normally 2024-07-28 07:57:13 tokyovigilante: any error when trying to mount? Did you add it to fstab? 2024-07-28 07:58:03 WhyNotHugo: thanks, just Mounting root: failed 2024-07-28 07:58:05 ah, possibly not 2024-07-28 07:58:27 hmm, it does seem to be in my fstab 2024-07-28 08:00:19 that error message could be more useful xD 2024-07-28 08:00:34 did you include the subvolume details in fstab? 2024-07-28 08:03:28 UUID=997d5f83-39e6-4dc4-8ce8-da04a0f39bb3/btrfsrw,relatime,ssd,space_cache=v2,subvolid=256,subvol=/root 0 1 2024-07-28 08:03:44 so looks like it should be there, possibly without the / in /root 2024-07-28 08:06:05 removing the / made no difference, at the emergency shell I'm also getting "mounting /dev/mapper/main on /sysroot failed: no such device" despite /dev/mapper/main (the btrfs volume) being present 2024-07-28 08:07:12 actually probably need to update my kernel hooks after editing fstab... 2024-07-28 08:10:15 hmm, no dice. 2024-07-28 08:32:44 sounds like initramfs issues 2024-07-28 08:33:52 might have to regenerate the initramfs 2024-07-28 08:34:43 Thanks, I do have that set up in my kernel hook as per the guide, but I must have misconfigured it somehow 2024-07-28 08:35:25 could be missing modules in the initramfs, see /etc/mkinitfs/mkinitfs.conf 2024-07-28 08:43:13 features="ata base ide scsi usb btrfs keymap kms usb cryptsetup cryptkey resume" 2024-07-28 08:43:13 disable_trigger=yes 2024-07-28 08:43:39 is my mkinitfs.conf, this is an old thinkpad x230 so no nvme, just a sata ssd 2024-07-28 08:50:26 huh, not sure, guess i was off base, sorry 2024-07-28 08:51:37 thought it would be more of a boot string issue as I can just "mount /dev/mapper/main -o subvol=root /sysroot" and be on my way 2024-07-28 08:51:55 from the emergency console 2024-07-28 08:56:07 yep, been a while since i got dumped to emergeny shell in the initramfs, so i am a bit stuck for being much help, but i'll look at the initramfs /init script to see if I can get a clue 2024-07-28 09:00:39 no worries, thanks! 2024-07-28 09:20:16 Hello 2024-07-28 09:40:29 im guessing some of you might be using supermicro hardware in here: https://www.supermicro.com/en/support/security_PKFAIL_Jul_2024 2024-07-28 09:48:54 Apachez: that's ... not pretty 2024-07-28 10:00:22 Hihttps://github.com/binarly-io/SupplyChainAttacks/blob/main/PKfail/ImpactedDevices.md 2024-07-28 10:04:54 ouch 2024-07-28 10:37:26 oh, that's worse :) 2024-07-28 10:48:57 any clever tricks to teach rsyslogd of openrc services? 2024-07-28 12:35:00 tokyovigilante: did you double-check the UUID? It needs to be the UUID of /dev/mapper/crypt 2024-07-28 12:35:12 (e.g.: not the UUID of the physical partition) 2024-07-28 12:35:23 /dev/mapper/main in your case maybe 2024-07-28 13:58:05 "Saijin_Naib: Get off my ass..." <- Haha, sorry, nothing that antagonistic. Trying to gently goad/encourage you to tey dailying Alpine 2024-07-28 14:34:34 why by default a branch is tracked instead of latest-stable? 2024-07-28 14:36:39 Juest: because unexpected release upgrades might break something 2024-07-28 14:36:59 but usually they dont break anything, right? 2024-07-28 14:37:01 sometimes there's breaking changes and we note them down in release notes so people can prepare 2024-07-28 14:37:04 *usually* no 2024-07-28 14:38:47 Juest: perhaps the idea is that people who use edge should be ready for things to break and be prepared to help figure it out. 2024-07-28 14:39:15 right, but like casual users can just use latest-stable instead of edge, yeah? 2024-07-28 14:39:16 or not even edge, but you know... upgrading 2024-07-28 14:39:29 i get that, 2024-07-28 14:39:39 thanks for the response 2024-07-28 15:29:05 Hello, I don't think this is a hardware issue since my laptop used to work just fine (used to have gentoo on it) , I decided to install alpine linux .. and idk why but I couldn't get it to work. 2024-07-28 15:29:09  aplay -l**** List of PLAYBACK Hardware Devices ****card 0: PCH [HDA Intel PCH], device 0: ALC256 Analog [ALC256 Analog] Subdevices: 1/1 Subdevice #0: subdevice #0card 0: PCH [HDA Intel PCH], device 3: HDMI 0 [HDMI 0] Subdevices: 1/1 Subdevice #0: subdevice #0card 0: PCH [HDA Intel PCH], device 7: HDMI 1 [HDMI 1] Subdevices: 1/1 Subdevice #0: subdevice #0card 0: PCH [HDA Intel PCH], device 8: HDMI 2 [HDMI 2] Subdevices: 1/1 2024-07-28 15:29:09 Subdevice #0: subdevice #0 2024-07-28 15:29:33 (sorry if that doesn't look good , but aplay detects the devices normally) 2024-07-28 15:30:06 I've also double checked with alsamixer and pactl , everything seems good (unmuted, volume 100%, etc) 2024-07-28 15:30:55 `lspci | grep Audio` -> `00:1f.3 Audio device: Intel Corporation Sunrise Point-LP HD Audio (rev 21)` 2024-07-28 15:32:04 not sure what I can share to make things helpful .. but I can't hear anything at all (tried aplay, mpv with -ao=alsa and -ao=pulse , and firefox .. all seem to thing they're playing (not showing any errors) while I can't hear anything) 2024-07-28 15:32:20 thank you for your time (whomever reads this) 2024-07-28 16:42:00 try to mute and unmute, and of course check that alsamixer is set to playback 2024-07-28 20:48:43 WhyNotHugo: I have a couple of times but will recheck, I am prompted to unlock the right partition (/dev/sda2 in my case) so I think at least that part is right 2024-07-28 20:49:08 Thanks for the workstation tutorial by the way, it's extremely helpful 2024-07-28 21:18:11 is there a way to see the additional packages that `apk add` would install before actually installing them? like what apt does by default? 2024-07-28 21:21:49 apk add --help 2024-07-28 21:24:27 Hello71: thanks! 2024-07-29 01:53:05 is it just me or is the latest chromium build on edge hosed/ 2024-07-29 02:20:40 "is it just me or is the latest..." <- What in specific? I only use it for BetaFlight and ESC Configurator 2024-07-29 02:20:56 qemu/libvirt is borked, though 😭 2024-07-29 02:28:09 browsing Amazon, tabs just freeze and eventually I need to close them when they eat shitloads of memory 2024-07-29 03:21:20 "browsing Amazon, tabs just..." <- insert obligatory "sounds like business as usual for chrome" joke here 2024-07-29 03:21:34 I'll try to use it a bit to see if it hits me. Do you use many extensions? 2024-07-29 06:01:25 WhyNotHugo: embarassed to admit, but it would seem that "rootfstype=brtfs" is not the same as "rootfstype=btrfs". Working a treat now anyway, thanks for the help! 2024-07-29 06:58:10 tokyovigilante: haha, I've had the exact same typo 2024-07-29 08:21:25 maybe choose a filesystem that is less likely to be mistyped :P 2024-07-29 08:21:53 but yeah, typos get everyone eventually 2024-07-29 08:32:59 Hello 2024-07-29 09:15:16 socksinspace: etx4 you mean? ;) 2024-07-29 09:17:57 zeefs would also be an option ;D 2024-07-29 09:18:15 or zedfs depending on where you learned english 2024-07-29 09:19:05 After cgdisk setup I type 'mkfs.vfat /dev/nvme0n1p1' and it prints 'mkfs.vfat: unable to open /dev/nvme0np1: no such file or directory". I wrote partitions on cgdisk and it shows the partitions fine. 2024-07-29 09:19:05 How is this even possible? 2024-07-29 09:20:21 where did the 1 disappear to? 2024-07-29 09:20:46 In my typo here and no editing option in neochat :D 2024-07-29 09:22:36 if you do "fdisk -l" you see the relevant partition? 2024-07-29 09:25:11 socksinspace: It shows the partitions but lists them only as a number not 'p1' and 'p2' 2024-07-29 09:29:22 wdym only as a number? 2024-07-29 09:31:36 Hello, I was using the wiki to add an ssid and passworf using wda_passphrase but it dors not work when restarting the networking service, any ideas? 2024-07-29 09:34:08 I did it with doas but still no permission, so I entered su mode and then it worked. I could also see how the conf file was updated 2024-07-29 09:34:22 socksinspace: For my usb stick fdisk -l prints totally another section listing the partitions with device names. For my nvme it prints just 'number' 1 for ESP and 2 for the 2nd partition. 2024-07-29 09:34:22 Cgdisk shows filesystem as vfat anyway so maybe I'll just keep going and see what happens. 2024-07-29 09:36:43 if they weren't there at all i would've thought that maybe cgdisk didn't properly reload the partition table, but idk, maybe someone else has an idea 2024-07-29 09:56:19 "if they weren't there at all i..." <- I rebooted and it works now lol :D 2024-07-29 09:56:48 ok, so it probably was that 2024-07-29 09:57:21 good old turning it off and on again, never fails unless it does :D 2024-07-29 13:30:48 in the recent windows/crowdstrike event, microsoft's official guidance was to keep rebooting windows until it works, saying that people had reported to them it required as many as 15 reboots. 2024-07-29 13:34:41 invoked: Triggering the boot failure repair mechanism over and over haha 2024-07-29 13:39:25 Does /etc/mkinitfs/mkinitfs.conf require dm-crypt with cryptsetup or is crypsetup enough? 2024-07-29 14:16:38 malmi70[m]: 'cryptsetup' is what you need 2024-07-29 14:17:29 'cryptkey' if you want to set up encrypted boot with grub but that's a whole other thing 2024-07-29 14:20:58 in other words, the 'cryptsetup' feature pulls in dm-crypt, and if dm-crypt weren't available in the kernel as it should be, it would error out 2024-07-29 14:59:07 "in other words, the 'cryptsetup'..." <- Thanks. That's what happened when I tried to install encrypted btrfs with guide. No errors during install, but dm-crypt wasn't found after encryption password was set. I gave up and went with default install. 2024-07-29 15:02:03 malmi70[m]: you will need 'btrfs' and 'lvm' (if you used lvm) 2024-07-29 15:04:39 malmi70[m]: did you use the stock kernel or roll your own? 2024-07-29 15:09:09 invoked: Btrfs was already in mkinktfs.conf but I didn't use lvm. I used stock kernel. 2024-07-29 15:16:05 malmi70[m]: not sure what went wrong in your case. just fyi, you can inspect the complete features available in /etc/mkinitfs/features.d/ and you can see what these features expect by looking at those files (look at /etc/mkinitfs/features.d/cryptsetup.modules for instance) 2024-07-29 15:18:35 Thanks, I'll take a look 2024-07-29 15:58:20 anon[m]: hello, may I ask you to post PW logs again, but this time with -vvv flag? 2024-07-29 15:59:37 anon[m]: also, which version of the kernel is used by distros which don't have that unpluggedness issue? 2024-07-29 16:08:42 Ermine: Of course, just give me a minute 2024-07-29 16:21:15 Ermine: Here are the logs from "pipewire -vvv": https://tpaste.us/MRZv 2024-07-29 16:22:51 Ermine: And I'm currently on the v3.20 branch with the 6.6.42-0-lts kernel 2024-07-29 16:41:01 anon[m]: may I ask you to try out linux-edge kernel? It is a possibility that kernel misreports port state to pipewire 2024-07-29 16:42:28 Ermine: Will do 2024-07-29 17:15:27 ergh, my lualatex isn't working on a edge's texlive-full installation 2024-07-29 17:15:36 luaotfload | load : × Failed to load "fontloader" module "basics-gen". 2024-07-29 17:15:47 and no writable cache path 2024-07-29 17:27:21 Ermine: I had tried out the linux-edge kernel a few weeks ago, but it didn't help so I switched back to linux-lts. Now that I have tried it out again it seemed to have fixed it. And surprisingly not just on linux-edge. I tried it again on linux-lts, where it now works too, even after I uninstalled linux-edge 2024-07-29 17:28:02 ACTION shrugs 2024-07-29 17:30:07 anon[m]: let me know if the issue reappears 2024-07-29 17:30:10 Ermine: I have to do a little more testing, so I might get back to you. But for now, thank you so much, I tried to get it working for weeks without any luck. I hope I didn't waste too much time 2024-07-29 17:58:31 anon[m]: no problem 2024-07-29 18:39:18 chexo4[m]: (Assuming you're the same as chex03) You can do anything in ACF that you want, including websockets, assuming you can figure out how to do it in Lua. 2024-07-29 18:39:44 But, I would not recommend putting any kind of service / daemon within ACF. 2024-07-29 18:39:46 I see 2024-07-29 18:39:58 Keep it separate then? 2024-07-29 18:40:39 ACF runs in response to a API request (usually an HTTP request) and does not have any persistent process. 2024-07-29 18:41:18 It is launched as a CGI script from a web server or is launched from a shell / script to run a single command. 2024-07-29 18:41:48 Generally, ACF is used to manage a service that runs in its own process. 2024-07-29 18:42:15 So, yes, keep it separate 2024-07-29 20:09:34 just a quick question is it considered a trivial situation in the auto install if i have a free partition im willing to install to instead of the whole drive 2024-07-29 20:14:28 queueq: Setup-alpine uses whole disk so you need to manually do the partitions. 2024-07-29 20:17:24 oh got it 2024-07-29 20:17:25 thanks 2024-07-29 20:17:58 if you take care of the bootloader stuuf yourself you can just mount the partition on /mnt (i think that was the mountpoint) and do "setup-disk /mnt" see the instructions regarding manual partitioning on the wiki, and the docs page 2024-07-29 20:18:21 it's a bit fiddly but not too hard 2024-07-29 20:18:32 hm 2024-07-29 20:18:33 ill see 2024-07-29 20:19:09 i did it recently where i wanted just one partition, no seperate boot or swap 2024-07-29 20:19:49 it only borked itself during an upgrade once :P 2024-07-29 20:41:39 "oh got it..." <- Do backups or clone the disk before you attempt to install anything else into it. 2024-07-30 06:46:29 unfortunately the fix for chromium doesn't work. please, i need this browser for work. firefox eats my cpu on my old laptop.. https://gitlab.alpinelinux.org/alpine/aports/-/issues/16315#note_424489 . thx 2024-07-30 07:51:15 "unfortunately the fix for..." <- I daily drive Qutebrowser 2024-07-30 07:59:10 hello 2024-07-30 09:26:05 @malmi70[m] i don't understand your post, sorry. well, i daily drive chromium, for good reasons. i know it's not popular to say this, but i don't care. i hope one can find the issue. i know, chromium as a big task and i am thankful for the work on it @ alpine. 2024-07-30 09:28:09 .. but what happened to ff/motilla over the last few years is unacceptable. 2024-07-30 09:28:45 mozilla. f. migraine again. 2024-07-30 09:30:42 therefore: \part. bye 2024-07-30 10:15:10 Hi, Im Good morning, 2024-07-30 10:15:10 The situation is best explained in this discussion https://github.com/pieroproietti/penguins-eggs/discussions/377 2024-07-30 10:15:10 i am Piero Proietti an italian svluppatore author of Penguins' eggs a remastering software, on the line of systemback, remastersys, etc that runs on Arch, Debian, Devuan, Manjaro, ubuntu and derivatives. 2024-07-30 10:15:10 I am trying to port a project of mine to Alpine linux and am well on my way. What I am missing, however, is the crucial step: an initramfs that booted from live, mounts my filesystem.squashfs and runs swich_root. 2024-07-30 10:15:12 I would like to thank all those who would like to bring help/advice and the authors of Alpine, a cheap but definitely interesting scopera. 2024-07-30 11:23:39 why is userfaultfd disabled? 2024-07-30 13:05:48 sewn: because nobody's had a need for it to be turned on? seems quite niche 2024-07-30 13:06:33 aawwosoomeee 2024-07-30 14:16:54 good day 2024-07-30 14:17:25 I am currently getting "ERROR: http://dl-2.alpinelinux.org/alpine/edge/main: UNTRUSTED signature" on one machine. If I replace the repo with another one, I get the same error message: "ERROR: https://mirror.ungleich.ch/mirror/packages/alpine/edge/main: UNTRUSTED signature" 2024-07-30 14:17:26 What might be the reason for this? 2024-07-30 14:17:39 Time is correct, Tue Jul 30 16:17:25 CEST 2024 seems to be fine 2024-07-30 14:24:58 telmich: check the contents of /etc/apk/keys 2024-07-30 14:25:23 that has: alpine-devel@lists.alpinelinux.org-4a6a0840.rsa.pub alpine-devel@lists.alpinelinux.org-5243ef4b.rsa.pub alpine-devel@lists.alpinelinux.org-5261cecb.rsa.pub 2024-07-30 14:26:12 wait 2024-07-30 14:26:15 Something is really broken 2024-07-30 14:26:16 WARNING: Ignoring http://dl-cdn.alpinelinux.org/alpine/edge/main: No such file or directory 2024-07-30 14:26:21 that is not true 2024-07-30 14:26:38 Like it says no such file or directory, but the url is a 200 2024-07-30 14:27:03 I'm on apk-tools 2.12.5 2024-07-30 14:28:35 apk-tools-2.14.4 here 2024-07-30 14:28:58 But that warning means it was not able to get the index 2024-07-30 14:29:16 Due to that error you posted earlier 2024-07-30 14:33:15 it's very confusing 2024-07-30 14:33:21 curl / wget for the url on the same machine works 2024-07-30 14:33:24 so it's not a cert issue 2024-07-30 14:38:57 This is about the apk signing keys, not tls certificates 2024-07-30 14:39:10 These are just rsa keys 2024-07-30 14:39:30 So check what's in /etc/apk/keys 2024-07-30 14:43:03 ikke: I pasted it above, or do you mean the content of the files? 2024-07-30 14:44:02 telmich: apk version alpine-keys 2024-07-30 14:44:26 alpine-keys-2.2-r0 2024-07-30 14:44:27 I suspect it's out of date 2024-07-30 14:44:30 Yes 2024-07-30 14:44:53 I guess this has not been updated for some time 2024-07-30 14:44:58 Is there a "standard procedure" to get them back up-to-date? 2024-07-30 14:45:19 ikke: Maybe a year or so - it's one vpn endpoint that is rarely touched 2024-07-30 14:47:07 https://wiki.alpinelinux.org/wiki/Release_Notes_for_Alpine_3.15.0 2024-07-30 14:52:50 i cant seem to get the wifi working, i tried installing linux-firmware and rebooting and some other things but... (full message at ) 2024-07-30 14:54:20 What wifi chipset do you have? 2024-07-30 14:54:38 Thanks a lot ikke, the apk add -X http://dl-cdn.alpinelinux.org/alpine/v3.14/main -u alpine-keys fixed it! 2024-07-30 14:56:19 Has been more than a year ago since the keys were rotated 2024-07-30 14:59:01 hehe - long running infrastructures :-) 2024-07-30 15:01:25 mandoc seems to conflict with man-db... but I need mandoc(1) 2024-07-30 15:03:22 eh whatever, compiling from source works for me 2024-07-30 15:05:11 "What wifi chipset do you have?" <- it seems to be a realtek one, how can i check? 2024-07-30 15:05:57 lshw for example 2024-07-30 15:40:21 what is the difference between xf86-video-nouveau and xf86-video-nv ? the wiki says the proprietary drivers are not supported 2024-07-30 17:42:34 hi, quetion. should i open a new issue for chromium-127.0.6533.72-r1 or was it enough to reopen the issue: https://gitlab.alpinelinux.org/alpine/aports/-/issues/16315#note_424041 2024-07-30 17:43:18 should be enough to reopen it 2024-07-30 17:43:33 thx ikke, like always 2024-07-30 17:56:27 I installed Emacs 29.x from v3.20 repo but it doesn't launch. Is the problem related to Emacs or Sway? Is the package non-compatible with Wayland? Terminal prints just "/bin/sh: emacs: not found" 2024-07-30 17:57:59 malmi70[m]: what does `ldd $(which emacs)` return? 2024-07-30 18:00:47 ikke: cannot load $(which emacs): no such file or directory. 2024-07-30 18:00:47 My app launcher shows 2 icons as expected. 2024-07-30 20:36:31 there is various versions of emacs, you have emacs-nox which has no GUI only runs in terminals, and you have various -x11 and gtk3 flavors, no idea if anyone ever added a wayland variant 2024-07-30 20:36:50 (i was the original dev/maintainer of emacs long ago) 2024-07-30 20:37:20 have a look at 'apk search emacs' and see if any of the packages has a name that that rings a bell for wayland users... 2024-07-30 20:40:50 malmi70[m]: ^^^^ 2024-07-30 21:28:41 yMVV_HsHcX0: emacs has two gtk3 flavors, -gtk3 and -pgtk, and the former is only meant to be run on x11 and the latter only on wayland which is pretty funny 2024-07-30 21:29:25 the -gtk3 uses some x11 stuff which prevents it from working on wayland the latter supposedly is "buggy" under x11 2024-07-30 22:59:16 anyone else had this issue where you couldnt modify files on mounted ntfs drives at all despite having sufficient rw perms 2024-07-30 22:59:58 seems to be an alpine specific issue since i could mkdir and stuff just fine on some random liveiso 2024-07-30 23:04:22 historically you needed to do "mount -t ntfs-3g /dev/sdb1 /mnt" dunno if that is still the case, but back in the days the default ntfs odule was ro 2024-07-30 23:12:44 hm put ntfs-3g in my fstab and it did fix it 2024-07-30 23:12:50 thanks for the help 2024-07-30 23:13:22 thought every semi-modern kernel version had basically pushed ntfs-3g to the side since it had stuff built in but i guess not in this case 2024-07-31 04:59:05 For the Raspberry Pi Zero 2W, is it still recommended to go for the armv7 image instead of aarch64? 2024-07-31 09:07:07 hi & sry. question from a non-pro: the chromium fix doesn't work. unfortunately. is chromium 127 "clang with stdc++" and a lot of amazing patches (which i do not understand, of course)? if so, why not revert again to "clang and compiled-in llvm libc++" minus some patches - like a few months ago: https://gitlab.alpinelinux.org/alpine/aports/-/commit/1e49682f69aa682081689a0bccd8b97896ac27d5 - until: 2024-07-31 09:07:14 chromium works with stdc++ and all these patches. until: chromium is tested by the maintainer before shipping it. until this: https://gitlab.alpinelinux.org/alpine/aports/-/issues/16315 does not happen again. before this mentioned revert, the same stuff happened. the revert "fixed" it. is that possible? can we revert all this stuff again, until it works? 2024-07-31 09:38:38 if some sort of revert is possible, this would also give the maintainer more time to work on the problem, and less stress to provide a fix for this particular problem, that the end user is now struggling with. just my toughts. thx and a nice day everyone. 2024-07-31 12:37:23 When does usually pkg gets from testing to community? 2024-07-31 12:38:31 Sorry, from edge to main 2024-07-31 12:39:11 Asking because I'm using openfortivpn a few hours a day for the last couple of months, but I had to build it myself since it's only in edge. 2024-07-31 12:42:11 It doesn't have to go to main. It will be automatically included in the next release if it's in either main or community 2024-07-31 12:49:30 It's just on edge 2024-07-31 12:57:46 is using edge for your system not an option? 2024-07-31 13:17:04 "is using edge for your system..." <- it works 2024-07-31 13:17:47 well, i know that, but does the other person 2024-07-31 15:26:24 socksinspace: not an option atm. 2024-07-31 20:30:49 Can I manually download an alpine package from the website? 2024-07-31 20:30:59 You can download it from a mirror 2024-07-31 20:31:11 https://dl-cdn.alpinelinux.org/alpine/ 2024-07-31 20:31:37 Thanks! 2024-07-31 23:12:13 what is up sigmas