2021-11-01 00:03:06 night clandmetere 2021-11-01 00:03:27 so who's out there helpfully rehosting the grsec patches 2021-11-01 00:11:37 you 2021-11-01 00:11:40 amirite 2021-11-01 00:14:43 freq: lmao, i just need brad's favor, i bet 2021-11-01 00:14:46 all twitter users are like that 2021-11-01 00:14:57 ah 2021-11-01 00:15:11 gotta hit him with some ego boosters and then ask pretty please 2021-11-01 00:15:21 man 2021-11-01 00:15:32 sounds like a lot to do 2021-11-01 00:15:40 i better get to it 2021-11-01 00:15:56 i've got this new power supply that i'm using 2021-11-01 00:16:04 it's from the pinestore. the 5v desktop power 2021-11-01 00:16:24 thing is unreal. has 5 usb ports and wireless charging on top 2021-11-01 00:17:20 one of the ports is capable of 65 watts 2021-11-01 00:17:48 god damn. usb 3 compatible for the 65watts thing? 2021-11-01 00:17:57 yeah it's PD65 2021-11-01 00:18:01 christ 2021-11-01 00:18:07 i think the spec is 100W max, right? 2021-11-01 00:18:12 maybe 2021-11-01 00:18:22 that pinesupply is cheap, right? 2021-11-01 00:18:26 yes very 2021-11-01 00:19:17 i was using a 3A phone charger and was getting brown outs for my rpi400 2021-11-01 00:19:34 i can now drive a usb mechanical drive 2021-11-01 00:19:43 *power 2021-11-01 00:20:20 hell yes 2021-11-01 00:21:01 OC'd to 2Ghz 2021-11-01 00:21:11 didn't wanna push it with the theoretical max of 2.1 2021-11-01 00:21:42 the hell is the rpi400? 2021-11-01 00:21:55 it's the rpi4 updated and built into a keyboard 2021-11-01 00:22:04 lmao, like the commodore 2021-11-01 00:22:08 yes 2021-11-01 00:22:28 how is 2.1 the theoretical max? 2021-11-01 00:22:44 lots of testing says that it becomes unstable past 2.1 2021-11-01 00:23:11 iiinteresting 2021-11-01 00:23:48 CPU: Quad Core ARMv7 v7l (-MCP-) speed/min/max: 1100/600/2000 MHz 2021-11-01 00:24:04 still 32 bit, but there are 64 bit distros 2021-11-01 00:25:11 i tested kali, but i can't seem to tweak the /boot/config.txt without kernel panics 2021-11-01 03:22:53 Hi there. Is anyone active here? 2021-11-01 03:44:35 Alpinepoo: your mom 2021-11-01 03:49:58 Are you able to provide some assistance with installing alpine? 2021-11-01 04:19:25 what issue do you have 2021-11-01 05:40:02 oic 2021-11-01 06:07:38 Everytime I attempted an install on a laptop with Luks on lvm, following the guide accordingly, after reboot, initfs was unable to mount /sysroot. It has to be done manually everytime. After checking the modules and rebuilding the ramdisk, still no luck 2021-11-01 06:21:54 hm 2021-11-01 06:22:00 all pine poo 2021-11-01 06:22:10 imagine the 2021-11-01 06:22:15 outcome 2021-11-01 06:27:06 Incredibly alpineless 2021-11-01 06:27:12 yuh 2021-11-01 06:27:23 i'm on debian until rpi400 supports more 2021-11-01 06:27:24 like 2021-11-01 06:27:33 i'm not going to move binaries over manually every update 2021-11-01 06:27:38 aint nobody got time fo day 2021-11-01 06:27:39 t 2021-11-01 06:28:05 I'm not sure if there's anything that needs to be edited in /etc/mkinitfs/mkinitfs.conf 2021-11-01 06:29:58 did you add the lvm and cryptsetup features? 2021-11-01 06:32:40 Yep, plus cryptkey for the keyfile 2021-11-01 06:33:36 and you set the cryptroot parameter in the bootloader? 2021-11-01 06:34:12 In the Grub config, you mean? 2021-11-01 06:34:22 yes 2021-11-01 06:35:00 It might not have been formatted correctly, as the wiki doesn't mention how that needs to be done 2021-11-01 06:35:25 are you usign a hand-written grub.cfg, or did you use grub-mkconfig? 2021-11-01 06:38:29 hekel 2021-11-01 06:51:18 After installing Grub, the guide asserts to edit /etc/default/grub with these details 2021-11-01 06:52:14 that's correct, just have to make sure that you rerun grub-mkconfig after editing that 2021-11-01 06:55:39 Yep, did that but still no luck 2021-11-01 06:56:02 is this LUKS on LVM, or LVM on LUKS? 2021-11-01 06:56:32 lvm on luks 2021-11-01 06:56:37 you said "Luks on lvm" but https://wiki.alpinelinux.org/wiki/LVM_on_LUKS is LVM on LUKS 2021-11-01 06:56:47 oh ok 2021-11-01 06:57:34 what exactly did you put in the GRUB_CMDLINE_LINUX_DEFAULT variable? 2021-11-01 07:02:38 My bad, that was a typo. It was entered as thus: GRUB_CMDLINE_LINUX_DEFAULT="cryptroot=UUID= cryptdm=lvmcrypt cryptkey" 2021-11-01 07:03:03 did you replace with the LUKS container's UUID? 2021-11-01 07:03:08 Is that the correct format, I was going with what the rest of the file looked like 2021-11-01 07:03:26 Yes, which came from the command mentioned 2021-11-01 07:03:36 without <> 2021-11-01 07:19:37 grub me 2021-11-01 07:36:48 Yes, without the <> 2021-11-01 07:37:07 Are you aware of any reason why this may have happened? 2021-11-01 07:37:26 I'm not familiar with mkinitfs, but someone else here may have ideas 2021-11-01 07:38:03 when you boot grub, can you use e to edit the boot parameters and check that the linux line specifies the right cryptroot? 2021-11-01 09:50:21 Hello world 2021-11-01 09:50:46 world hello 2021-11-01 12:52:50 Hey all, it is the second time I send an email to get my mirror into the official list but I get no answer, it is a deprecated way to do this? 2021-11-01 13:11:40 qontinuum: hi 2021-11-01 13:12:59 qontinuum: when did you sent it? 2021-11-01 13:30:03 Why is MariaDB package lack mysqladmin? 2021-11-01 13:31:00 mariadb-client has it? 2021-11-01 13:31:42 oh, thank you 2021-11-01 13:46:59 clandmeter: on the 23rd of october for the most recent one, months ago for the other one 2021-11-01 13:56:25 qontinuum: ah i see it 2021-11-01 13:56:32 it got covered in spam emails 2021-11-01 13:59:56 Damn, those spammer. 2021-11-01 14:02:13 Ah xD 2021-11-01 17:04:03 i think it was frag 2021-11-01 17:05:08 i got a blackmail yes 2021-11-01 18:09:11 hm 2021-11-01 18:21:32 what do you if you're compiling a program that needs the linux kernel headers 2021-11-01 18:22:12 do you add the location of them to CPATH? 2021-11-01 18:32:07 would this question be for the alpine-devel channel? 2021-11-01 18:32:24 you add linux-headers ? 2021-11-01 18:32:53 they are then in the standard include path under which is how programs import them anyway 2021-11-01 18:33:13 i have linus-lts-dev 2021-11-01 18:33:21 so there's a separate linux-headers package? 2021-11-01 18:33:30 *linux-lts-dev 2021-11-01 18:33:47 linux-headers is the name of the thing that puts them in the actual location 2021-11-01 18:34:05 ok 2021-11-01 18:34:12 thanks 2021-11-01 19:14:53 what is CPATH 2021-11-01 21:21:32 quick question: how do I "downgrade" from edge to stable (v3.14)? I changed /etc/repositories and ran apk update/upgrade, but nothing happened 2021-11-01 21:21:57 apk upgrade --update --available 2021-11-01 21:22:19 thanks! 2021-11-01 21:22:51 --available makes apk use the packages that are available in the selected repositories, and thus, allows it to downgrade packages 2021-11-01 21:55:51 how bad is #ifdef lol 2021-11-01 21:56:08 i was reading some C style guides that recommend against it 2021-11-01 22:31:41 kelvinfahrenheit43: it's a room? 2021-11-02 01:03:40 Dear all, do you have an estimation about the minimum disk/memory requirement for Alpine Linux standard version? I got a bit confused based these two webs: https://wiki.alpinelinux.org/wiki/Requirements and https://wiki.alpinelinux.org/wiki/Installation One said minimum ram is 100MB another said it is 512MB for x86_64, and there is not a concrete disk (persistent storage) requirement. 2021-11-02 01:55:29 jacky_: You can boot with 100MiB but I can't imagine you'll be doing much with it afterwards. What is the purpose of your question? What kind of system are you trying to run? 2021-11-02 01:56:24 I think the smallest I've run were the 256MiB ALIX boards. And that works well enough with some limited GUI even. 2021-11-02 01:57:37 As for disk, my smallest containers are around 27 MiB and that could probably shave a few megs off if I tried. But that's not "standard" by any stretch. Again, depends on what you want to do with it. 2021-11-02 02:49:16 Thanks. I am working on a research project. My goal is to create a KVM with small Linux (e.g., Alpine Linux). What I need to do is creating block devices and mount file systems (e.g., ext4, xfs) and do some operations for the mounted file systems. The reason I need minimum setting is that I need to take/restore a lot of KVM snapshot , which needs to be very fast. So, I think a virtual machine with less memory/disk space is faster to conduct 2021-11-02 05:30:00 tried to connect to a windows machine via remmina's rdp function; i'm getting /usr/lib/freerdp2/librdpsnd-client-fake.so not found errors, and it's in an infinite loop 2021-11-02 05:32:19 that's in freerdp-plugins package 2021-11-02 05:33:51 gonna have to wait until i fuck up my system real good with upgrade -U on edge 2021-11-02 05:34:08 i really gotta figure out how to set up my fontconfig 2021-11-02 05:34:19 why isn't there a cli provided BY fontconfig for preference 2021-11-02 05:35:11 psykose: Looks like your black magic fixed it. Perhaps the package needs to be fixed to pull in freerdp-plugins? 2021-11-02 05:35:49 there is probably some usage without the plugins 2021-11-02 05:35:58 that's why they would be separate in the first place i guess 2021-11-02 05:36:06 mmm, i see, i see. 2021-11-02 05:36:12 i don't see what a fontconfig cli would accomplish 2021-11-02 05:36:55 lessen the user's requirement to type out xml 2021-11-02 05:39:38 it's mostly some copypaste xml though that you put in config/fontconfig/fonts.conf 2021-11-02 05:39:56 Oh, it'll look in there? Is that where xdg_config_home should be? 2021-11-02 05:39:59 gentoo basically has a 'fontconfig cli' in the sense of you can use the eselect commands to enable/disable the specific files that have all the options you would pick anyway 2021-11-02 05:40:02 but it's not really much better 2021-11-02 05:40:27 yes, you can see the definition at /etc/fonts/conf.d/50-user.conf 2021-11-02 05:40:59 i can do system-wide config so that fonts are properly selected when i run stuff under superuser just by going into the /etc/fontconfig thing, right? 2021-11-02 05:41:02 somewhere in there 2021-11-02 05:41:22 also, can you explain to me when this [].d/[numbers]-[name].conf shit came into being 2021-11-02 05:41:28 what is the actual reasoning behind it 2021-11-02 05:41:29 51-local.conf load a conf.d/local.conf too 2021-11-02 05:41:38 for system 2021-11-02 05:41:47 uhh the name is just old standards, not really sure 2021-11-02 05:41:55 the old numbers-name defined ordering, lower numbers get loaded first 2021-11-02 05:41:58 how'd the 50 become significant 2021-11-02 05:42:00 oh, interesting 2021-11-02 05:42:02 no clue 2021-11-02 05:42:06 and the x.d implied a conf dir 2021-11-02 05:43:47 also the readme inside the conf.d dir tell you what the numbers are supposed to mean for this specifically 2021-11-02 05:44:03 oh hell yeah, i could have just looked 2021-11-02 05:44:26 i've taught myself to ask before looking on alpine because of the occasional elision of docs, etc 2021-11-02 05:44:29 should get back into the habit 2021-11-02 05:44:40 no harm in asking 2021-11-02 07:14:57 psykose: What's setting your XDG_HOME and XDG_CONFIG_HOME variables on your system? 2021-11-02 07:15:02 I'm pretty certain I'm not setting mine 2021-11-02 07:18:45 they are not set by default 2021-11-02 07:19:10 and they don't have to be, the defaults are ~ ~/.config 2021-11-02 07:19:46 oh, good 2021-11-02 07:19:54 is the same true for XDG_RUNTIME_DIR? 2021-11-02 07:24:04 no, that has to either be made yourself or is set by having linux-pam and logging in 2021-11-02 07:24:15 or maybe some other combination of things 2021-11-02 07:24:42 Damn; interesting. I cheated for pipewire and just made it ~ in my xinitrc 2021-11-02 07:24:46 wonder if that's fucking some other stuff up 2021-11-02 07:24:56 it's not supposed to be persistent 2021-11-02 07:25:10 and that frequently breaks a lot of things when people don't read what it's supposed to be :p 2021-11-02 07:27:32 the cheapest close-enough way without just using pam is something like this https://img.ayaya.dev/qIXMG90FnHn7 , though this doesn't clear it on logout, but that rarely breaks anything 2021-11-02 07:27:41 the spec is in https://specifications.freedesktop.org/basedir-spec/basedir-spec-latest.html , just search for xdg_runtime_dir 2021-11-02 08:43:15 jacky_: In that case I'd suggest several things: 1) sounds like busybox has all the executables you need. Together with musl libc it's about 1.5MiB which should give you functional userspace. You will likely want to write init/linuxrc scripts yourself anyway. 2021-11-02 08:46:31 jacky_: 2) There is probably very little persistent mutable state you need to have. Kernel (and initrd) can be passed to qemu as external files. Root filesystem can also be read-only an shared. So what remains to be persisted is content of RAM and the writable part of filesystem. 2021-11-02 08:49:14 jacky_: 3) Compile own kernel configuration without modules. This will save you tremendous amount (relatively) of RAM wasted on drivers you won't ever need. Get rid of initramfs unless you want to use it for your mutable data. 2021-11-02 08:51:06 jacky_: 4) Consider using 9p passthrough that qemu has built-in, it allows you to expose a directory on the host as a filesystem in the guest without having to go through creating an image from it. If you do, be wary of which caching mode do you pick. 2021-11-02 08:53:29 jacky_: 5) Note that if you'll be using those for any kind of performance measurement the amount of available RAM will likely affect the I/O operations, especially when it's so tightly restricted. 2021-11-02 11:31:17 hi, i have a question about alpine edge 2021-11-02 11:31:25 is this the right place to ask? 2021-11-02 11:33:04 Yes, it's the place to ask 2021-11-02 11:33:17 I'm not sure where else you would ask 2021-11-02 11:33:50 ok here goes 2021-11-02 11:34:31 if i get hired as a sysadmin, and i run the servers on alpine edge 2021-11-02 11:34:42 would that make someone an edgelord? 2021-11-02 11:38:53 Ok, this is not the right place to ask that 2021-11-02 11:42:23 I don't even get it 2021-11-02 11:44:09 They are trolling/memeing 2021-11-02 11:44:19 leah: Nah, at that point you're barely edgeop. It's when you get commit access to the aports tree you become a true edgelord. 2021-11-02 11:44:37 ok, thanks 2021-11-02 11:45:20 PureTryOut (matrix.org) , I need their braincells ASAP 2021-11-02 11:45:51 i'm considering trying alpine 2021-11-02 11:47:03 but not edge 2021-11-02 11:47:29 It's fairly reasonable IME. I've been running Void and Alpine boxes for a while (though not very hands-on). I generally prefer alpine little bit better if I don't need glibc. 2021-11-02 11:48:01 My main dev driver is still Gentoo though. 2021-11-02 11:49:25 i'm setting up a separate machine just for testing stuff, and i'm planning to set up arch/parabola. but alpine is also interesting 2021-11-02 11:50:31 non-gnu free os is a novelty that i've yet to experience, outside of openbsd 2021-11-02 11:50:45 leah: Any specific considerations or worries? 2021-11-02 11:51:37 none so far. i'm familiar with busybox already because i've set up a few openwrt routers (from cli) before 2021-11-02 11:52:17 i guess i do have one question 2021-11-02 11:52:24 serious this time, not trolling :) 2021-11-02 11:53:05 i understand the purpose of alpine is to be lightweight and efficient, so stuff like gnu is thrown out, but: 2021-11-02 11:53:18 s/thrown out/optional/ 2021-11-02 11:53:29 You can still install it if you want 2021-11-02 11:53:40 hypothetically, if i did run into some stuff that needed parts of gnu. e.g. glibc, bash, the gnu version of m4 (which is superior) 2021-11-02 11:53:51 could i have the best of both worlds 2021-11-02 11:54:00 glibc is a system lib, you cannot combine glibc with musl 2021-11-02 11:54:11 You can just install bash 2021-11-02 11:54:23 yeah what about on some containerized thing though 2021-11-02 11:54:33 apk add bash 2021-11-02 11:54:35 like, is it common for people to do stuff like that in alpine 2021-11-02 11:54:38 Yeah, it does mix and match fairly well. Busybox utils are default until you install something else. 2021-11-02 11:54:56 well the busyboxness of it is specifically what appeals to me 2021-11-02 11:55:52 i've been seeking to throw out as much bloat as possible from my computer. my motivation is simple: as a kid, i had a computer with 256MB RAM, and that was high-end and plenty of ram back then. with a 20gb hdd. and i want to have a modern system that runs well on such specs, even though i have much better hardware nowadays 2021-11-02 11:56:36 Libc is obviously libc, that's hard to swap out; but the utilities themselves you can just install desired version. Look in the package db for whatever you might want https://pkgs.alpinelinux.org/ 2021-11-02 11:56:55 anyway, gnu+linux is "safe" hence arch/parabola, but if mix/match of gnu things is feasible in alpine, i might tread uncharted waters and make alpine my main/only OS 2021-11-02 11:57:11 i guess glibc isn't really a requirement in most cases 2021-11-02 11:57:45 glibc is only rlly a requirement for proprietary software 2021-11-02 11:57:46 i guess it's easier than when i was switching from windows back in the mid 2000s 2021-11-02 11:58:07 back then it was, well, dual-boot or bust. with alpine it's easier if you can mix and match like that on the same system 2021-11-02 11:58:20 so if you're using proprietary software, you'll have difficulties (use a chroot) 2021-11-02 11:58:21 Alpine is generally set up wit "reasonable minimalism" in mind. To have more control over optional features you would have to go for source-based distro and tweak compile-time flags. 2021-11-02 11:58:32 eris[m]: the only proprietary software that i use is what's on my hdd's firmware, webcam's firmware, cpu microcode etc 2021-11-02 11:58:43 plus i guess web javascript. alpine is a perfect fit for me 2021-11-02 11:58:44 well, thats not software ;) 2021-11-02 11:58:51 do you use nvidia? 2021-11-02 11:58:59 no, intel. i915 driver 2021-11-02 11:59:14 cantiga platform. 2008 gen hardware 2021-11-02 11:59:15 perfect hw for alpine then 2021-11-02 11:59:42 yeah well, i have interest in alpine for one of my projects too. so it's on my radar 2021-11-02 11:59:54 you'll rarely if ever need gnu stuff - maintainers have that fixed for you 2021-11-02 12:00:05 yeah i saw alpine at first and i was like 2021-11-02 12:00:08 does not compute 2021-11-02 12:00:14 i wouldn't run edge on anything for 'production' but it's fine for desktop use, if this isn't actual sysadmin :) 2021-11-02 12:00:14 linux without gnu 2021-11-02 12:00:33 but ariadne has been slowly selling me on the idea of ditching gnu software ;) 2021-11-02 12:00:57 alpine still uses the good parts of gnu software 2021-11-02 12:01:01 I mean, it's still gcc compiling it. 2021-11-02 12:01:05 i thought busybox was just for embedded systems 2021-11-02 12:01:09 or like 2021-11-02 12:01:22 Wasn't the original Sabotage musl-based? 2021-11-02 12:01:23 that annoying busybox shell you get dropped into in debian, when you mess up the linux command in grub.cfg 2021-11-02 12:01:52 Both busybox/toybox and musl have grown slightly in the past years. ;-) 2021-11-02 12:02:01 sabotage was the first musl iirc 2021-11-02 12:02:02 busybox on a full workstation environment is what "did not compute" until i saw alpine. and i'm intrigued 2021-11-02 12:02:06 they say if you use busybox on a non-embedded system you get visited by ghosts 2021-11-02 12:02:08 i wouldn't try it 2021-11-02 12:02:09 but the creator of sabotage is a dick 2021-11-02 12:02:25 psykose: have you met these ghosts? 2021-11-02 12:02:33 some of them are pretty cute 2021-11-02 12:02:39 then i will try it 2021-11-02 12:03:04 gnu free linux is pretty hard 2021-11-02 12:03:15 but u dont need glibc or coreutils 2021-11-02 12:03:18 Ariadne: what parts? 2021-11-02 12:03:29 you said good parts. which parts, off the top of your head? 2021-11-02 12:03:41 ..gcc? 2021-11-02 12:03:42 gcc would be the first thing on mine 2021-11-02 12:03:49 bison? gmake? 2021-11-02 12:04:13 actually, i would consider bison to be in the "not good" part of gnu software 2021-11-02 12:04:14 maybe not good and moreso 'no benefit to avoiding them' 2021-11-02 12:04:16 but we're stuck with it anyway 2021-11-02 12:05:20 (i read some paper comparing bison to byacc and was taken away that berkeley yacc generates more reliable parsers, but i forget the details) 2021-11-02 12:05:25 yea, thats what i thought 2021-11-02 12:06:14 so, gcc basically 2021-11-02 12:06:19 gmake i would consider quite awful 2021-11-02 12:06:28 but, again, stuck with it 2021-11-02 12:06:56 2005. older than i thought 2021-11-02 12:07:02 i just read the alpine wikipedia page 2021-11-02 12:07:59 tbh any version of make is awful. i try to avoid writing makefiles 2021-11-02 12:08:16 the syntax is ugly and unintuitive 2021-11-02 12:08:41 hmm, posix make is not bad. 2021-11-02 12:10:05 i often just write custom stuff in bash 2021-11-02 12:10:05 Posix make is pretty bad in that it's a templating language for shell, while shell is already a macrosubstitution engine by itself. Each with slightly different syntax. 2021-11-02 12:10:17 but even then, posix shell would do nicely 2021-11-02 12:10:26 make make used only for small parts 2021-11-02 12:10:50 I like how mk did throw out that part completely and just lets shell do it's own thing. 2021-11-02 12:11:36 cmake is nice 2021-11-02 12:13:21 cmake is configuration system, not a build system; granted the distinction is not often clear 2021-11-02 12:13:45 is there a nice way to communicate with normies from a linux computer, is mail the only way? 2021-11-02 12:13:59 how do they communicate themselves 2021-11-02 12:14:07 whatsapp 2021-11-02 12:14:10 snap 2021-11-02 12:14:21 i guess? 2021-11-02 12:14:36 well those are mobile apps :p 2021-11-02 12:14:45 but maybe they are willing to install something :p 2021-11-02 12:14:45 whatsapp runs in a browser. 2021-11-02 12:14:47 telegram is nice. proprietary server, but there is a free software client. and there's a nice mobile app that's intuitive for normies 2021-11-02 12:14:56 WA is annoying. I've had to set up Android emulator in qemu the other day. 2021-11-02 12:15:13 leah: ah yes, i guess thats good 2021-11-02 12:15:13 Don't use Telegram, it's just another proprietary walled garden 2021-11-02 12:15:22 we need a cli telegram client for alpine tho 2021-11-02 12:15:24 to be clear, the telegram service is non-free, but with a free client available 2021-11-02 12:15:34 And the non-free part is the problem 2021-11-02 12:15:37 what PureTryOut said. but frag said: normies 2021-11-02 12:15:44 Telegram does code dumps occasionally and they're not great, but there's active FOSS fork. 2021-11-02 12:15:54 normies enjoy being chained and gagged behind walled gardens that tell them how to think while spying on them 2021-11-02 12:16:07 normies can use e.g. Matrix just fine, just let them install a mobile client like FluffyChat which behaves more like what they are used too from WhatsApp and stuff than Element does 2021-11-02 12:16:15 oooooh yeah matrix is great 2021-11-02 12:16:59 I like XMPP myself but since App Store is anti-GPL tough luck getting good client there (otherwise Gajim and Conversations are as feature-complete as it gets) 2021-11-02 12:17:04 frag: whatsapp web 2021-11-02 12:17:05 my girlfriend is a "normie" and uses Matrix with me just fine. My parents and sister also use it just fine, but I wouldn't call them normies 2021-11-02 12:17:20 i use matrix 2021-11-02 12:17:55 email is still a tried and true method though 2021-11-02 12:18:12 I think Matrix is architecturally unsound. Federation is wonky at best. Their docs are post-hoc and make non-web implementations rather hard. 2021-11-02 12:18:26 and setting up a mail server isn't difficult 2021-11-02 12:18:32 It's a silo that tries to pretend it isn't IME. 2021-11-02 12:18:39 IMO? 2021-11-02 12:18:42 IME? 2021-11-02 12:18:51 wym 'non web implementations' 2021-11-02 12:18:53 like 2021-11-02 12:19:00 synapse is in python 2021-11-02 12:19:04 setting up a mail server is one of the harder things to set up correctly 2021-11-02 12:19:13 fluffychat is flutter 2021-11-02 12:19:19 theres a bunch of qt clients 2021-11-02 12:19:32 Do they have interoperable e2ee? 2021-11-02 12:19:36 email and irc is all you need really. but normies like to use phones a lot nowadays, and irc clients for phones aren't that great 2021-11-02 12:19:38 ...yes? 2021-11-02 12:19:56 Good. It's been a nightmare when I left. 2021-11-02 12:20:10 im not sure how its a silo either 2021-11-02 12:20:23 considering im running a synapse server 2021-11-02 12:20:24 irc is actually pretty great for private communications, if you set up your own instance. and there are nice solutions for that 2021-11-02 12:20:29 It still relies on centralized identity FWIU. 2021-11-02 12:21:09 hey, how do I set up software RAID1? I"m getting a hetzner server soon and it has dual 1TB SSDs, I think they set RAID up for you on their preinstalled OS but i'm installing alpine manually 2021-11-02 12:21:42 matrix makes you ban homeservers on a per-room basis 2021-11-02 12:21:55 I know they have some smart folks working on it though. So given some love it might evolve into something competitive with XMPP in few years. Though when I left my impression was that from-scratch redesign would get there way faster. Also it leaks metadata everywhere. 2021-11-02 12:22:12 spiralw: mdadm? 2021-11-02 12:22:17 spiralw: set up the raid by hand, format everything, mount the filesystems as you want them, and then do `setup-disk -m sys /mnt` 2021-11-02 12:22:31 you use mdadm, eg format two disks identically then mdadm --create /dev/md0 --level=mirror --raid-devices=2 /dev/diskpart1 /dev/diskpart2 2021-11-02 12:22:55 I've used mdadm before, but how does the boot partition work? 2021-11-02 12:23:56 is there an android client thats easy/install-and-use? i guess it would be a bit weird to have a default sever... 2021-11-02 12:24:03 for itc 2021-11-02 12:24:04 irc 2021-11-02 12:25:17 CcxWrk: are you saying XMPP doesn't rely on a centralized identity? 2021-11-02 12:25:20 i will say this 2021-11-02 12:25:42 i will too 2021-11-02 12:25:43 this 2021-11-02 12:25:53 So will I! This 2021-11-02 12:25:58 i like busybox's site 2021-11-02 12:26:22 it's nothing fancy. html pages, git repo and a mailing list. just what you need 2021-11-02 12:26:22 PureTryOut: Of course not. But you also don't get mobility between federated instances (you can set up forwarding though). 2021-11-02 12:26:57 I'm not sure about the "of course" but ok 😛 2021-11-02 12:27:15 Anyways that's a thing they have expressed interest in in the past and I'm sure it'll happen at some point in the future 2021-11-02 12:27:34 frag: ive not seen a good irc client on android 2021-11-02 12:27:47 XMPP is nice too, but it seems it has never caught on. It did in the past, and then Google and Facebook stopped using it and split off their own walled gardens 2021-11-02 12:28:19 eris[m]: damn 2021-11-02 12:29:20 do you think the govs/corps are spying on/selling irc user dataaaaaaaa? 2021-11-02 12:29:27 -aaaaaaaaaaa 2021-11-02 12:29:33 too much effort 2021-11-02 12:29:36 revolution irc is kinda okay 2021-11-02 12:29:56 i just use a webclient on my vps 2021-11-02 12:30:15 PureTryOut: You don't see XMPP as much because you don't get to see private communications (Doh). Matrix is popular because they managed to haggle unlimited connections with Freenode/Libera, something XMPP bridges never managed to get and thus those get rate limited and can't compete in the "free bouncer" market. 2021-11-02 12:31:06 Not every public Matrix room is bridged with IRC though 😉 Although yeah a lot 2021-11-02 12:31:20 omg https://busybox.net/live_bbox/live_bbox.html 2021-11-02 12:32:59 that's funny 2021-11-02 12:33:03 People keep negative impression of XMPP that was true 15 years ago when it was ICQ/AOL clone. But the very point of the protocol was to add on stuff as needed. So nowadays you get everything from low-power polling for mobile devices to Signal-grade encryption. 2021-11-02 12:46:12 at least XMPP you can say "i'm not going to federate with kiwifarms.cc" and have that actually mean something 2021-11-02 12:46:24 with matrix, you have to do it for every single room 2021-11-02 12:46:42 like how do you come up with something more broken than IRC in that regard 2021-11-02 12:46:49 IRC is really good at banning entire things 2021-11-02 12:46:51 :D 2021-11-02 12:54:18 /kline Ariadne@* 2021-11-02 12:54:36 "improve your IRC experience with this one trick!" 2021-11-02 12:58:07 Discord is surprisingly good at handling trolls/spammers 2021-11-02 12:58:19 with things like "You can't talk for 15min if you're joining for the first time" 2021-11-02 13:03:35 Tried Discord. "Join in, you don't even need account". Cool. "Oh, you want to post anywhere? Register with email." Ah well, those are easy to make. "We locked your account for security. Put in all the personal information and a phone number to verify." 2021-11-02 13:03:54 No effin way if you are going to bait and switch me like that. 2021-11-02 13:04:22 I hear that folks over in US don't experience that. But it's been fairly consistent whenever I tried. 2021-11-02 13:04:37 i had a discord account for maybe 5-6 years without it asking me for a phone number 2021-11-02 13:04:43 then the other day it did suddenly :p 2021-11-02 13:05:18 but it is separate from the regular add a phone number flow.. so the phone number field is still blank 2021-11-02 13:05:18 funny 2021-11-02 13:05:40 jvoisin: yes, which is why i begrudgingly set up a discord 2021-11-02 13:05:57 granted though, i don't think there is any other 'all the personal information', it's just email+phone 2021-11-02 13:05:58 maybe matrix will eventually be usable :) 2021-11-02 13:06:44 also, it is possible for the moderators of a discord community to set "user has to pass phone verification to join" 2021-11-02 13:06:59 which is really effective at uhh, stopping driveby trolling from say kiwifarms :) 2021-11-02 13:07:28 though i admit is unfortunate for those wishing to remain anonymous 2021-11-02 13:07:36 ^ 2021-11-02 13:07:43 I have hopes for matrix 2021-11-02 13:08:03 i wonder how people do things online just throwing around their real info everywhere 2021-11-02 13:08:05 it seems so strange 2021-11-02 13:08:11 If they were upfront about phone number requirement I'd consider it. If they are straight up lying to my face? No way. 2021-11-02 13:08:48 Ariadne: huh, I swear I saw the option to blacklist certain servers on a homeserver-wide level at some point 🤔 2021-11-02 13:08:53 sure, i'm just saying that you may have hit the phone number requirement as a side effect of trying to join a discord server which required it 2021-11-02 13:09:13 "server" 2021-11-02 13:09:17 god, i hate that they call them that 2021-11-02 13:09:19 they are not 2021-11-02 13:09:29 No. It hard-locked the account and from what I've talked with the admin they mentioned no such setting. 2021-11-02 13:09:43 Ariadne: they're called Guilds internally at Discord I think 2021-11-02 13:09:46 ah, well, there is such a setting for some discord "guilds" 2021-11-02 13:09:53 I guess I'm "east" enough to have scary IP address block or something. 2021-11-02 13:09:54 precisely because it's FUCKING CONFUSING 2021-11-02 13:09:58 it hard-locked for me too, there's 2 versions of it 2021-11-02 13:10:30 https://usercontent.irccloud-cdn.com/file/0RPYob65/Screen%20Shot%202021-11-02%20at%208.10.25%20AM.png 2021-11-02 13:10:49 i switch between high and highest depending on what level of fuckery is afoot 2021-11-02 13:12:27 at any rate, discord solves a problem that the current free alternatives do not: abuse mitigation 2021-11-02 13:12:54 it's on Matrix' roadmap 2021-11-02 13:13:03 to be fair to some extent, almost nothing solves abuse mitigation in general 2021-11-02 13:13:10 it's far too easy to be a dickwad online to people 2021-11-02 13:13:23 when twilio launched SMS support, i proposed adding support to atheme for it, to send registration codes to a mobile number 2021-11-02 13:13:43 but people got mad 2021-11-02 13:13:57 even though having to use a phone does introduce a lot of friction that stops drive-by abuse 2021-11-02 13:14:00 No open system does for sure. At best they can make it costly, but dedicated trolls still gonna get in. 2021-11-02 13:14:13 right, but there can be value in making it costly :) 2021-11-02 13:14:27 reputation/based, captchas, … 2021-11-02 13:15:18 unrealircd has a new mode where it gives a point to every IP that has been connected for 5 minutes 2021-11-02 13:15:28 and as you get more points, it reduces the restrictions 2021-11-02 13:15:39 i guess something like that could work 2021-11-02 13:16:27 Reputation systems can fail hard. ISTR the old Freenet mailing system had a "report spam" feature which people mass-abused as "I disagree with your opinion" leaving everyone in small bubbles. 2021-11-02 13:17:20 And good automated reputation systems are nonexistent (otherwise they'd be all the rage in the blockchain world; so far I've only seen theoretical work) 2021-11-02 13:17:58 Ariadne: looking at the default config for Synapse, it does allow blacklisting of ip ranges and url's and I also see issues on the Github repo related to existing blacklisting functionality 2021-11-02 13:18:23 thats for clients 2021-11-02 13:18:50 the advice i got for "how do i stop kiwifarms from joining my rooms" *from* matrix devs was "use this undocumented ACL feature to ban the entire homeserver on a per-room basis" 2021-11-02 13:19:03 which is just ridiculous 2021-11-02 13:19:17 No I believe it's for servers as well 2021-11-02 13:20:29 it has to be on the per-room ACL, because of how the DAG works 2021-11-02 13:20:31 Like there are issues like this https://github.com/matrix-org/synapse/issues/7297 which imply there is a blacklist for servers 2021-11-02 13:20:49 if you block a server from directly contacting yours, they can still participate in the room 2021-11-02 13:23:55 Oh you mean you want to block them entirely from, even indirectly from other servers? 2021-11-02 13:27:16 directly and indirectly 2021-11-02 13:27:52 we talked about this elsewhere, there isn't any way to do so unless you craft your settings via API 2021-11-02 13:27:59 which is dumb 2021-11-02 13:29:02 from Matrix docs: "Setting server ACLs on a room is a relative rare operation, and Element does not currently provide a user interface for doing so. Instead, these can be set either by script or by manually creating a server ACL state event via Element/Web’s developer tools UI." 2021-11-02 13:31:17 IMO it's fault of Matrix not going for federated nor fully server-agnostic controls (like Signal where their servers don't even see groups). As the result the control structure isn't sound nor consistent. (Had seen kicked people reappearing after some random homeserver rejoined partition and such) 2021-11-02 13:32:26 PureTryOut: yes, the goal would be to avoid having that server's users interacting *at all* with said community, for hopefully obvious reasons (if not obvious, googling them will surely be enlightening) 2021-11-02 13:33:08 Haha no it's obvious, I just didn't understand properly what you meant 2021-11-02 13:33:30 Comparable to the wider Mastodon community banning Gab from federating with them 2021-11-02 13:34:34 When Fediverse was getting attention I was telling myself "let's see if they get popular enough to get Eris Wars of their own" 2021-11-02 13:35:05 yes, i'm skeptical that a fully open federated network can be safe 2021-11-02 13:35:38 in part because "safe" is subjective to begin with 2021-11-02 13:35:51 but also because trustless architecture is incredibly hard to get right 2021-11-02 13:37:05 We don't even have good PKI for that to begin with. (I do have ideas about how could you implement it, but definitely not enough manpower) 2021-11-02 13:37:26 Ariadne: if trolls really want to attack you, they will. my own experience indicates this, when one of my projects came under heavy troll attack a few years ago 2021-11-02 13:37:33 they will find ways to get around any restrictions 2021-11-02 13:37:47 yes, but having zero restrictions is not the answer 2021-11-02 13:38:07 there is value in being able to say, proverbially, "you must be this tall to ride" 2021-11-02 13:38:16 i once muted an entire irc channel, and only allowed registered people to join, it was that bad 2021-11-02 13:38:23 even then, people still found ways to spam the channel 2021-11-02 13:38:39 with join/quit messages perfectly well-synced 2021-11-02 13:39:20 in the end, it was actually quite impressive so i just gave up and let it run its course 2021-11-02 13:39:50 anyway, for that project, the requirements were: actually good mobile client, and some sort of reasonable moderation story 2021-11-02 13:40:04 with the FOSS solutions, it's basically pick one or the other 2021-11-02 13:40:17 matrix has nice clients, irc has reasonable moderation 2021-11-02 13:40:47 what ultimately wins in the end is strength in numbers. enough teenagers with unlimited free time will completely crush an irc channel. i learned this the hard way 2021-11-02 13:40:53 xmpp has okay clients and reasonable moderation, 2021-11-02 13:40:59 but nobody wishes to use xmpp :D 2021-11-02 13:41:32 xmpp strong 2021-11-02 13:41:49 The onboarding story is still slightly confusing. (There's Quicksy, but still) 2021-11-02 13:43:01 I'm surprised Movim doesn't get more love. 2021-11-02 13:43:19 other human beings can be extremely resourceful with enough boredom/time. for example, i just discovered that solder, depending on how its packed for sale, can behave like mr slinky 2021-11-02 13:43:51 https://imgaz2.staticbg.com/thumb/large/oaupload/banggood/images/F0/38/845ec842-b712-4b34-8089-cb60446b3fff.jpg 2021-11-02 13:44:24 why are you buying solder on banggood 2021-11-02 13:44:31 i'm not 2021-11-02 13:44:35 oh 2021-11-02 13:44:35 that's google image search 2021-11-02 13:44:48 but what i have looks like that, and also this: 2021-11-02 13:44:51 i was gonna say, RoHS does not prohibit people like you from buying and using lead solder 2021-11-02 13:44:53 https://en.wikipedia.org/wiki/Slinky 2021-11-02 13:45:02 just big companies 2021-11-02 13:45:21 yeah but in practise that makes it hard to buy 2021-11-02 13:45:28 wuh? 2021-11-02 13:45:38 i can go into any shop i want and buy lead solder 2021-11-02 13:45:45 you live in america 2021-11-02 13:46:01 i live on the continent where rohs was invented 2021-11-02 13:46:15 very quickly i will not 2021-11-02 13:46:19 hmm 2021-11-02 13:46:23 arbitrage opportunity 2021-11-02 13:46:29 give your rohs solder to a kid. they will use it like mr slinky 2021-11-02 13:46:42 ACTION plans to pack a few lbs of lead solder 2021-11-02 13:47:12 i buy leaded stuff, but it's a pain in the ass. sellers usually only stock lead-free, because people demand it because it's "safer" 2021-11-02 13:47:36 if you want large quantity here, you usually have to import from america/asia 2021-11-02 13:49:08 i mean lead-free is fine. just crank up the temperature 2021-11-02 13:50:32 i want muh freedom solder 2021-11-02 13:50:48 hmm, this became #alpine-offtopic long ago 2021-11-02 13:51:38 maybe, but nothing else is happening here anyway 2021-11-02 13:52:27 Ariadne: #alpine-slinky 2021-11-02 15:35:18 If I'm modifying the APKBUILD for tumbler for Thunar to include more builddeps so it builds more preview/file handler plugins, how do I go about troubleshooting why it won't enable a particular plugin? 2021-11-02 15:35:48 I've got it building everything except the RAW handler:... (full message at https://matrix.org/_matrix/media/r0/download/matrix.org/yXjyqqzcAiZHUefTojymcSSW) 2021-11-02 15:36:01 Which I made the libopenraw aport for, specifically for this reason 2021-11-02 15:36:10 Did I do something stupid like name the package wrong? 2021-11-02 15:36:23 I have libopenraw-dev in the builddeps 2021-11-02 15:54:44 the pkgconf is invalid for libopenraw 2021-11-02 15:54:46 or rather 2021-11-02 15:55:53 the ac script thecks for libopenraw-gnome-0.1, then libopenraw-gnome-1.0, but libopenraw in repo is -gnome-0.3 and -0.3 (i don't know if this works with those values) 2021-11-02 15:56:18 but then also, the -gnome-0.3 .pc file depends on -0.1, which doesn't exist, there are only 0.3 files 2021-11-02 15:56:55 you can read it yourself at /usr/lib/pkgconfig/libopenraw-gnome-0.3.pc, it needs -0.1 that doesn't exist/get installed 2021-11-02 15:57:15 i can't tell you why it works this way or what's happening :) but i assume something needs to be changed in the libopenraw apkbuild 2021-11-02 15:58:26 https://libopenraw.freedesktop.org/libopenraw/ 2021-11-02 15:58:32 running pkgconf --libs libopenraw-gnome-0.3 will give you the error, and assuming it gets called like that by ac 0.3 isn't even valid for it, but idk how ac works 2021-11-02 15:58:35 So should I target the 0.1.x series in my APKBUILD? 2021-11-02 15:59:32 Or is it better to ask XFCE to update their autoconfigure script to allow usage of the more current libopenraw? 2021-11-02 16:07:06 i don't have the slightest idea 2021-11-02 16:07:15 reading this stuff is a nightmare 2021-11-02 16:07:48 i suppose you can build 0.1 locally, and then see if it works 2021-11-02 16:10:53 Hello any used alpineterm??? 2021-11-02 17:30:11 Is this some Android thing? 2021-11-02 17:30:20 Oh, he left 2021-11-02 21:36:26 a bit buried but this mr is ready to be re-reviewed: https://gitlab.alpinelinux.org/alpine/aports/-/merge_requests/26434 2021-11-02 22:19:34 psykose: looks like Arch already carries a patch specifically for building with libopenraw 0.3.0 2021-11-02 22:19:35 https://github.com/archlinux/svntogit-packages/blob/dc572beb30ff31cc154a15b2e1d3c3fa88cbfa68/trunk/PKGBUILD 2021-11-03 02:03:51 Attempting to install the virtual image onto a Vultr VPS and after using the setup-alpine script to install to disk it doesn't appear to be working 2021-11-03 02:04:05 nothing sticks after a reboot 2021-11-03 02:09:59 The disk is detected and appears as an option for install when using the setup script 2021-11-03 02:10:21 /dev/vda 2021-11-03 02:11:17 Not experiencing the same issue with the Extended image 2021-11-03 02:16:27 after reboot kernel can't see the disk 2021-11-03 02:17:34 Virtual image optimized for virtual machines but not QEMU? 2021-11-03 02:18:58 in particular the virtio disk driver 2021-11-03 02:55:46 seems to be a bug with the latest version of alpine 2021-11-03 02:56:09 have had no problem installing alpine on a vultr VPS before 2021-11-03 02:56:34 not working on the extended image either 2021-11-03 02:58:00 somehow the virtual disk "dev/vda" is detected by the setup script and it'll write to it but it silently fails. running df doesn't pick up the disk though... 2021-11-03 03:02:41 after more research it appears i had it previously working prior to the latest release of Alpine- possibly a problem with my provider 2021-11-03 03:04:29 correction: following the latest release of alpine 2021-11-03 04:38:43 I wasn't removing the install media/ISO ): 2021-11-03 04:39:14 Disregard all previous messages lol 2021-11-03 04:40:16 man 2021-11-03 04:40:53 lol 2021-11-03 04:42:39 big dummy hours 2021-11-03 04:44:55 o'rite 2021-11-03 09:37:35 Is running wine on alpine a fool's errand 2021-11-03 09:37:37 ?** 2021-11-03 09:38:12 is that a romantic way to ask if it will work? 2021-11-03 09:38:16 does wine work with musl? 2021-11-03 09:39:01 wine is in the repos so i assume someone has it working 2021-11-03 09:39:28 why would wine not work with musl? 2021-11-03 09:40:26 i didn't know if it was glibc only 2021-11-03 09:41:04 wine is build from source afaik, so should just work™️ 2021-11-03 09:43:10 Wine works fine, but due to Alpine lacking multilib support you can only run programs that are 64-bit on x86_64. And since a lot of Windows programs are still (partially) 32-bit, it won't run a lot 2021-11-03 09:44:09 Hopefully that will change soon, due to MS deprecating 32bit Windows 2021-11-03 09:47:37 deprecate 32bit version windows i guess 2021-11-03 09:48:04 not 32bit compat i would assume. 2021-11-03 10:11:09 puretryout: really? any program that still has 32bit support at all isn't going to work? 2021-11-03 10:11:56 only full 64-bit programs will work on x86_64, and only full 32-bit will work on x86 2021-11-03 10:26:39 maddening 2021-11-03 10:26:55 wonder why noita is x86 2021-11-03 10:29:37 i'm gonna have to actually learn how to virtualize 2021-11-03 10:29:40 and go fast as fuck 2021-11-03 10:29:45 need the qemu cheatcodes 2021-11-03 10:41:25 you could also use flatpak wine 2021-11-03 10:43:34 that's kinda interesting 2021-11-03 10:44:11 hmmmm. 2021-11-03 10:44:16 or create a chroot of something with multilib manually 2021-11-03 10:44:47 afaik the alpine kernel does have the 32bit compat stuff enabled, just no multilib userspace 2021-11-03 10:47:10 clandmeter: yes, just like Ubuntu drops 32bit distribution but software will still run 2021-11-03 13:29:18 shlpo: where from did you buy Noita? 2021-11-03 13:30:25 you might be able to use the Minigalaxy (for GoG games), Steam, or Lutris Flatpak packages 2021-11-03 13:41:49 shlpo: Just make chroot with apk --root whatever --arch x86; no virtualization overhead that way 2021-11-03 13:46:30 That's good to know that there's a way to create Alpinelinux Bootstrap, whew 2021-11-03 13:47:10 Yeah, you can even download statically linked version of apk if you need to do that from non-Alpine system. 2021-11-03 13:47:50 (and don't want to compile apk yourself) 2021-11-03 13:48:02 good thing to know 2021-11-03 13:48:19 By the way alpine just gets better by day to the day, right? 2021-11-03 13:51:37 Yonle: what do you mean? 2021-11-03 13:52:15 Alpinelinux just gets more better. 2021-11-03 13:52:27 especially at GUI, Packages, and etc 2021-11-03 13:52:55 it does indeed get more betterer 2021-11-03 13:54:48 yeah 2021-11-03 13:54:59 hope the wiki gets more better soon 2021-11-03 14:48:39 hope will not improve the wiki ;-) 2021-11-03 15:50:26 Only blackmagic at this point 2021-11-03 18:47:00 has anyone recently managed to run a x86_64 VM using virt-manager with libvirt and qemu-kvm? Right now it fails to start an existing Alpine vm: `Error starting domain: internal error: Failed to start QEMU binary /usr/bin/qemu-system-x86_64 for probing: ** 2021-11-03 18:47:00 ERROR:../accel/accel-softmmu.c:82:accel_init_ops_interfaces: assertion failed: (ops != NULL)` 2021-11-03 18:58:20 Actually, even just running qemu-system-x86_64 directly causes that error 2021-11-03 19:01:22 yep, just by itself 2021-11-03 19:03:39 so you can reproduce it? 2021-11-03 19:03:54 same error as you just by running it with no args, yeah 2021-11-03 19:04:07 Oh ok glad I'm not alone at least then 2021-11-03 19:04:11 are you not group kvm? 2021-11-03 19:04:24 both kvm and qemu 2021-11-03 19:04:34 in both* 2021-11-03 19:05:39 leah2: I am 2021-11-03 19:07:08 have qemu-accel-tcg-x86_64 installed? 2021-11-03 19:07:50 i guess with -enable-kvm it could also avoid it maybe 2021-11-03 19:09:07 with -enable-kvm it doesn't crash 2021-11-03 19:09:39 adding the qemu-accel also fixes it with or without enable kvm 2021-11-03 19:09:45 strange 2021-11-03 19:10:01 makes sense 2021-11-03 19:10:06 the non-kvm backend is missing :) 2021-11-03 19:12:04 successfully debugged :) 2021-11-03 19:12:26 https://www.mail-archive.com/qemu-devel@nongnu.org/msg826119.html was key 2021-11-03 19:12:38 Hey folks; checked ~users list archive, see nothing. So, `apk upgrade` is suddenly failing: ERROR: unable to select packages: libpq-14.0-r6: masked in: @edgemain // satisfies: postgresql-client-13.4-r0[so:libpq.so.5] qt5-qtbase-postgresql-5.15.3_git20210406-r0[so:libpq.so.5] 2021-11-03 19:13:17 Mixing repositories is not supported 2021-11-03 19:13:20 this sounds like an issue of mixing edge-main with stable 2021-11-03 19:13:23 My /etc/apk/world has postgresql-client and I only have the edge repos in with @edge* prefices. 2021-11-03 19:13:59 you are trying to upgrade libpq from edge to 14 when the other things in that output are not on edge, and are 13.4 2021-11-03 19:14:09 So, terminology question then: I only have stable without @tags, I have @edgemain defined, plus the other two, so I only pull in if explicitly requested, not by default. Is this still "mixing"? 2021-11-03 19:14:19 yes 2021-11-03 19:14:37 i don't know why you are trying to pull in edge libpq though 2021-11-03 19:14:56 it's a "postgres dep", so either pull in all of postgres from edge, or none of it 2021-11-03 19:15:01 if you need to ask, you shouldn't do it. in fact you probably shouldn't do it in general 2021-11-03 19:15:30 I'm not pulling in any of postgres from edge. 2021-11-03 19:15:46 libpq is postgres 2021-11-03 19:16:25 grep -E '^@' /etc/apk/world 2021-11-03 19:16:33 grep -E '@' /etc/apk/world 2021-11-03 19:16:38 Right, but I'm not pulling it from edge, so I don't know why edge is being considered. `grep '@' world | xargs` shows: direnv@edgetesting gnupg@edgemain go@edgecommunity kubectl@edgetesting libcurl@pt libgpg-error@edgemain libgpg-error-dev@edgemain yodl@edgetesting 2021-11-03 19:18:23 leah2: interesting, that does work, but then I wonder why that wasn't necessary in the past and how I make my VM in virt-manager pass that argument 2021-11-03 19:18:25 AFAICT, libpq is currenly installed by v3.14, as is everything else postgres, and now suddenly apk wants to swap it to edge. 2021-11-03 19:18:26 Guys, anyone knows why my rpi with alpinelinux hangs at "EXT4-fs (mmcblk0p2): mounted filesystem with ordered data mode. Opts: (null)" when mounting it with "noatime"? https://pastebin.com/AaZqdGAG 2021-11-03 19:19:08 syscomet: It seems like it needs a newer version that is only on edge, most likely as a dependency from something you installed from edge 2021-11-03 19:20:42 `apk info -r libpq` will tell you what pulls it 2021-11-03 19:20:45 `apk info --all libpq` says: "libpq-13.4-r0 is required by: postgresql-client-13.4-r0 qt5-qtbase-postgresql-5.15.3_git20210406-r0" and this system has no X/qt, the only one I have is postgresql-client. How can I find what else might, in the upgrade, be trying to depend upon it and messing up the SAT stuff? 2021-11-03 19:22:52 neither of those should want libpq-14, so no idea tbh 2021-11-03 19:28:18 any idea? 2021-11-03 19:29:57 you should put defaults,noatime 2021-11-03 19:31:01 it is what I had before, does the same 2021-11-03 19:31:08 just noatime is an invalid fstab, the defaults are not implied, you need at minimum 'rw' manually specified otherwise 2021-11-03 19:31:14 ah, does it boot without noatime 2021-11-03 19:31:36 didn't try, but it boots without fstab 2021-11-03 19:32:34 I will do what you suggested and see how it goes 2021-11-03 19:32:54 weird thing is that fstab boots fine with raspbian 2021-11-03 19:33:04 your 5th field should be 0 not 1 2021-11-03 19:33:58 1 1 ? 2021-11-03 19:34:02 but i dunno, strange 2021-11-03 19:35:17 I mean perhaps there is something wrong with that fstab and OpenRC might be strickter than systemd when parsing the fstab 2021-11-03 19:36:13 the only thing that was wrong was no defaults, but aside from that the issue is not here 2021-11-03 19:36:29 could it be kernel related then? 2021-11-03 19:37:51 well the "Mounting root: ok." output indicates / is mounted ok, so I'd guess either (a) its having problems mount /boot or (b) the problem is unrelated to fstab or mounting in general 2021-11-03 19:40:58 minimal, this is what I thought, but I can guarantee the partition is not mounted and init stalls there 2021-11-03 19:41:15 as for having trouble mounting /boot, don't know. Will have to check 2021-11-03 19:41:27 well you could post the full log 2021-11-03 19:41:35 Latrina: how do you guarantee that? 2021-11-03 19:41:36 as I said, the strange thing is that it boots fine without fstab 2021-11-03 19:42:30 minimal, I am connected to the serial, there is no ping, no prompt, boot process stuck there 2021-11-03 19:42:46 Latrina: is this Alpine using run-from-Ram mode or sys/disc install mode? 2021-11-03 19:43:04 runs from sd 2021-11-03 19:43:24 Latrina: none of what you said guarantees that the partition is not mounted though, only that the boot is hung 2021-11-03 19:43:30 ok 2021-11-03 19:44:54 anyways, I lost the full boot logs. Will try again and if it gets stuck I will share it here 2021-11-03 19:44:58 thanks guys :) 2021-11-03 19:44:59 Latrina: I'm assuming that you mean "disk" mode. So you could try looking at the logfiles in /var/log/ to see if there's any useful info there 2021-11-03 19:45:10 yes disk mode 2021-11-03 19:45:17 ok 2021-11-03 23:39:47 "how bad is #ifdef lol" <- #ifdef is not inherently bad—as any other feature of C—but it should be used with great care 2021-11-04 03:35:26 If I'd like to make a MR to update Ristretto, and the notes indicate this:... (full message at https://matrix.org/_matrix/media/r0/download/matrix.org/oTtjXwjBmvfUyHtcEWPWcKxU) 2021-11-04 12:20:25 where would i recommend that https://pkgs.alpinelinux.org/package/edge/testing/x86_64/bitwarden_rs be renamed to it's new name "vaultwarden" 2021-11-04 12:20:29 > loudspeaker Note: This project was known as Bitwarden_RS and has been renamed to separate itself from the official Bitwarden server in the hopes of avoiding confusion and trademark/branding issues. Please see #1642 for more explanation. 2021-11-04 12:21:02 is this an issue for bugs.alpinelinux.org? 2021-11-04 12:21:16 algitbot mistake lol 2021-11-04 12:25:36 oh it's already been filed https://gitlab.alpinelinux.org/alpine/aports/-/issues/13134 2021-11-04 12:39:03 Hi, does anyone know "what" writes '/var/log/dmesg'? I'm just thinking about #13156 2021-11-04 12:47:41 How should I make my alpine vm? Q35 or I440FX and BIOS or UEFI? 2021-11-04 12:53:29 host and bios 2021-11-04 13:35:20 donoban: it sounds to me like this guy has never heard of syslog 2021-11-04 13:36:03 /var/log/dmesg is written in the awful /etc/init.d/bootmisc 2021-11-04 13:49:40 oh thanks Hello71 2021-11-04 13:51:59 so it's dumped after boot and never wrote again? 2021-11-04 14:11:48 unless you configure syslog 2021-11-04 14:26:43 dmesg - print or control the kernel ring buffer 2021-11-04 14:26:49 so yeah reboot and dmesg is gone 2021-11-04 14:26:58 unless syslogged into disk or onto som syslogserver 2021-11-04 14:35:19 ... 2021-11-04 14:43:22 iwd keeps segfaulting. Somehow it works fine when running with valgrind. What can I do? 2021-11-04 14:45:27 Apachez: you can set "previous_dmesg=yes" in /etc/conf.d/bootmisc so that on boot the previous dmesg file is renamed to dmesg.old 2021-11-04 14:46:18 or enable klogd and throw out bootmisc 2021-11-04 14:55:41 +1. Just use syslog and be done with it. 2021-11-04 18:05:36 . 2021-11-04 18:19:44 hello 2021-11-05 04:05:02 endigma3: 2021-11-05 04:26:00 is there a release for rpi400? 2021-11-05 06:30:09 isnt there a rpi4b inside rpi400? 2021-11-05 06:31:48 hmm seems not 2021-11-05 08:38:27 plugging in any new usb devices isn't giving me a block device 2021-11-05 08:38:28 i'm mad 2021-11-05 08:45:18 I DON'T WANNA REBOOOOOOT 2021-11-05 08:45:27 02:45:19 up 9 days, 2:00, 0 users, load average: 1.64, 1.48, 1.48 2021-11-05 08:49:30 7 days, u win 2021-11-05 10:28:36 Hey, I installed latest postgresql from repos and it seems like it's missing files for pgcrypto extension. is that something I can get from a separate package? 2021-11-05 10:29:30 postgresql-contrib probably 2021-11-05 10:29:39 though you can always also post the actual error too 2021-11-05 10:30:02 yes, postgresql-contrib fixed it 2021-11-05 10:30:05 the error was > ERROR: could not open extension control file "/usr/share/postgresql/extension/pgcrypto.control": No such file or directory 2021-11-05 10:30:17 yep 2021-11-05 10:30:28 for future reference you can always check like this too https://pkgs.alpinelinux.org/contents?file=pgcrypto.control&path=&name=&branch=edge&arch=x86_64 2021-11-05 10:30:57 oh that's really helpful, thanks 2021-11-05 14:08:50 Hi there! I'm tying to post to the infra mailing list about a security vulnerability in the wiki. I am currently subscribed to the list but unable to post. I've tried a few things but I can't seem to get my mail delivered, either to the mailing list, the help mailing list, or the postmaster. What is the intended route for me to send mail about this? Happy to provide more details if someone is willing to help :) 2021-11-05 14:11:25 edd_: if it's not an issue, you could create a confidential issue here: https://gitlab.alpinelinux.org/alpine/infra/infra/-/issues 2021-11-05 14:13:58 I'm happy to put an issue on the tracker, sure. How do I make the issue confidential? I can't seem to find an option for it at "New Issue" 2021-11-05 14:14:26 Not under the description bex? 2021-11-05 14:14:32 box* 2021-11-05 14:14:53 For me, the only thing under the description box are the "Create issue" and "Cancel" buttons 2021-11-05 14:14:56 I'd also like to know how to post to the mailing lists because that seems useful in general 2021-11-05 14:16:20 edd_: as it received quite some spam lately, we made it read-only, as it was hardly used 2021-11-05 14:17:47 Ah that's a shame. Maybe we can update the wiki to reflect this. 2021-11-05 14:18:05 edd_: what is your username on our gitlab instance? 2021-11-05 14:18:25 @eddsalkield 2021-11-05 14:18:57 I've added you as a reporter 2021-11-05 14:19:36 Thanks! 2021-11-05 14:19:55 As an aside, when my mail didn't get delivered, I was recommended to reply to Alpine Infra with the delivery report, which promptly responded with "Please remove your message/delivery-status attachments and send again." 2021-11-05 14:22:40 I've filed my issue confidentially over at the issue tracker 2021-11-05 14:23:20 Which mailing lists should I currently be able to post to? 2021-11-05 14:25:19 ~alpine/devel 2021-11-05 14:25:26 ~alpine/users 2021-11-05 14:25:32 ~alpine/aports 2021-11-05 14:26:40 Thanks 2021-11-05 14:27:31 I take it that information about the help mailing list as detailed at https://wiki.alpinelinux.org/wiki/Alpine_Linux:Mailing_lists is out of date? 2021-11-05 14:28:14 I'm happy to help get this updated as necessary to save future users some time 2021-11-05 14:36:07 is this about exiftool 2021-11-05 14:38:43 no 2021-11-05 17:12:59 has anyone had issues on edge with FF screensharing on sway? 2021-11-05 17:13:14 and running pipewire 2021-11-05 17:18:07 i have xdg-desktop-portal-wlr installed, and pipewire-media-session is running 2021-11-05 17:18:44 but when i try to screenshare via webrtc, it doesn't actually show a screen to share, and just says "Use operating system settings", which doesn't actually display anything 2021-11-05 17:27:24 jpsamaroo: I'm getting the same thing, but I haven't really looked into setting this up yet. Happy to test things if that's helpful 2021-11-05 18:14:39 installing zsh-calendar, seems only caledar_add is executable in /usr/share/zsh/5.8/functions/Calendar, havent used this function before, shouldnt all the files there be execuable? 2021-11-05 18:20:18 frag: Not really, I think they are supposed to be autoloaded. They aren't supposed to spawn separate shell so no need to mark them executable. 2021-11-05 18:21:20 By autoloaded I mean the 'autoload' built-in command in zsh that searches $fpath for given function name an then defines such function with body matching the file content. 2021-11-05 18:26:57 CcxWrk: so how do i make them available in path? not too familiar with autoload 2021-11-05 18:30:13 autoload -Uz Calendar # should make function "Calendar" available to you 2021-11-05 18:31:26 fpath should be set given you are in correct version of zsh 2021-11-05 18:33:19 CcxWrk: ah nice, was trying zshcalsys.... 2021-11-05 18:34:04 Surely zsh-calendar has some kind of documentation somewhere? 2021-11-05 18:35:14 man zshall seem to call it zshcalsys 2021-11-05 18:37:57 actually it seems not to work 2021-11-05 18:38:20 i guess this is why i gave up last time i tried haha 2021-11-05 18:38:32 calcurse seems a bit interesting 2021-11-05 18:39:35 Ask in #zsh over at Libera. But certainly there's something that you're supposed to call that autoloads the relevant functions for you similarly to compinit and promptinit. 2021-11-05 18:46:02 frag: fpath+=( /usr/share/zsh/*/functions/Calendar ); autoload -Uz calendar calendar_{add,edit,showdate,sort} 2021-11-05 18:47:35 CcxWrk: ah yes, seems i can run without errors at least :) 2021-11-05 18:47:42 This "works" for me but I don't really understand what it's supposed to do. 2021-11-05 18:48:20 tbh im just checking too 2021-11-05 18:48:30 but i think agenda and probably more 2021-11-05 21:58:01 hello there! could someone please help me install Alpine Linux with dual boot with another Linux? I'm new to Alpine. I followed Alpine wiki's page on that, but I didn't get it. I got some problems setting up the /boot :( . Both Alpine and Artix show up on GRUB, but Alpine won't boot if selected. On Artix, I mount /dev/sda1 on to /boot/efi. I mounted /dev/sda1 on /mnt/boot/efi during Alpine setup, but it didn't seem to work; I have no 2021-11-05 21:58:01 clue on how to set up /boot. If someone can help me please... 2021-11-05 22:40:29 if you use artix grub then you need to patch it to detect alpine initramfs or somesuch 2021-11-06 01:30:33 > if you use artix grub then you need to patch it to detect alpine initramfs or somesuch 2021-11-06 01:30:33 well, i think it is detecting alpine because it shows up on GRUB. Or am I wrong? 2021-11-06 01:33:55 iirc it assumes no initramfs required so you get unknown-block(0,0) 2021-11-06 02:50:04 sorry, i dont understand what it means 2021-11-06 02:50:31 can you explain it? 2021-11-06 11:40:36 Hi. I have this weird issue here. I'm trying to install gnome and gdm just refuses to start 2021-11-06 11:41:09 "Cannot open virtual console 1 (Permission denied) 2021-11-06 11:42:48 Any ideas? I have a user with uid >= 1000 so that's not the issue. i have the graphics driver (amdgpu) installed too. weirdly enough, gdm tries to load the "ati" module instead of amdgpu but can't, because that is not installed 2021-11-06 11:43:37 how are you starting gdm? 2021-11-06 11:43:43 and is elogind installed? 2021-11-06 11:45:06 with openrc and elogind is installed 2021-11-06 11:46:23 i just ran "apk add gnome" it includes elogind apparently. didn't include udev though, i had to install it after the fact 2021-11-06 11:48:10 i just figured out that xf86-video-amdgpu was not installed, could that have been the issue? I'm currently using irssi on this machine so i can't check without rebooting 2021-11-06 11:48:26 and thus leaving the irc 2021-11-06 11:50:04 i'll try, maybe i'll join again later. bye 2021-11-06 12:46:10 good afternoon, I am trying to pxeboot alpinelinux on my rpi 4b using the /boot files and kernel provided by the aplinelinux rpi aarch64 image however I am getting a "/dev/nfs: Can't open blockdev" error 2021-11-06 12:46:24 does this mean the kernel does not have the nfs driver to mount the network drive? 2021-11-06 13:22:26 bhayek[m]: loading alpine (and most distros) requires multiple files. if only some are detected then it will not work 2021-11-06 13:22:37 why don't you tell your actual error isntead of "Alpine won't boot if selected" 2021-11-06 13:24:12 I never liked the menu generator GRUB2 uses. I just write my own config that actually works instead. 2021-11-06 13:26:04 By the way, How many `bs` is required for booting into AlpineLinux Live? If i try `512`, It seems like it refuse to work & returning `./init` error 2021-11-06 13:26:23 For `4M`, It works fine 2021-11-06 13:31:14 sounds like you have a fraudulent storage device 2021-11-06 13:31:30 or just failing 2021-11-06 14:09:28 > why don't you tell your actual error isntead of "Alpine won't boot if selected" 2021-11-06 14:09:28 ok, I'll take a picture of the error and send it here 2021-11-06 14:14:12 ACTION uploaded an image: (77KiB) < https://matrix.org/_matrix/media/r0/download/matrix.org/lncxYpBlpYPscBVqWZigGtcB/photo1636207982.jpeg > 2021-11-06 14:14:58 after selecting the Alpine on GRUB, it takes a few seconds then display this error. If I press any key it goes back to GRUB again 2021-11-06 14:17:57 I didn't try to install GRUB while setting up Alpine because GRUB is already installed on Artix, so I thought it would be unnecessary. Am I wrong? Do I need to have 2 GRUBs, one in each OS ? 2021-11-06 14:19:02 Maybe wrong Partition UID. 2021-11-06 14:20:09 Boot into another distro (if available), Copy your /dev/sdX (at where you install alpine) UUID that shows in `blkid` and fix the wrong UUID on grub config 2021-11-06 14:23:17 I guess you are correct Yonle. Here is the output of `blkid /dev/sda4`: 2021-11-06 14:23:17 /dev/sda4: LABEL="ALPINE" UUID="38ed1799-7740-404d-8855-8635f804ef9d" BLOCK_SIZE="4096" TYPE="ext4" PARTUUID="e78b7fbd-bf4e-0144-811a-ca71f5e0e311" 2021-11-06 14:23:46 That explain your issue as well! 2021-11-06 14:23:50 but in `grub.cfg` it seems to have the wrong number 2021-11-06 14:24:31 Then replace it :^) 2021-11-06 14:24:49 ACTION sent a code block: https://matrix.org/_matrix/media/r0/download/matrix.org/lCUNoFDqkPPnLcuVHkWmHVPu 2021-11-06 14:25:20 Won't it cause trouble if I modify `/etc/grub/grub.cfg` directly? 2021-11-06 14:25:41 I guess the only line I have to change is the `uuid` right ? 2021-11-06 14:25:50 Yeah 2021-11-06 14:27:19 should I replace 2c083173-6c83-4627-9eb6-06ed36728f25 with e78b7fbd-bf4e-0144-811a-ca71f5e0e311 ( PARTUUID) or with 38ed1799-7740-404d-8855-8635f804ef9d (UUID) ? 2021-11-06 14:28:37 I fear messing up GRUB .... 2021-11-06 14:29:10 I think, The non-PARTUUID first 2021-11-06 14:32:54 bhayek[m]: at worst, you’ll break an already non-working menuentry. 2021-11-06 17:08:58 Is possible to use 'need dns' with OpenRC? 2021-11-06 17:09:35 why not? 2021-11-06 17:10:17 When I use 'need dns', a warning is returned saying that the service needs a non-existent service 2021-11-06 17:10:47 probably because you don't have a service with provide dns 2021-11-06 17:11:44 Should dnsd provide 'dns'? 2021-11-06 17:13:04 what do you need exactly? normally you don't need a dns daemon - any dns lookups are handled by musl and consult /etc/resolv.conf to figure out how to do them 2021-11-06 17:13:11 based on what it seems to do, yes, but since i don't think anyone has used it it's missing from the file 2021-11-06 17:13:32 you can add provide dns to init.d/dnsd i guess, or use dnsmasq 2021-11-06 17:15:14 dnsd doesn't appear to be enabled in Busybox so that won't work 2021-11-06 17:15:32 GenericNick2: why do you need dns service though? 2021-11-06 17:15:43 hah, didn't expect that 2021-11-06 17:22:48 libc handles dns? 2021-11-06 17:22:54 oh 2021-11-06 17:23:21 sshd uses 'use dns' in it's service 2021-11-06 17:23:37 s/it's/its 2021-11-06 17:23:37 GenericNick2 meant to say: sshd uses 'use dns' in its service 2021-11-06 17:24:27 I thought the 'dns' service was required as a service for dns lookups 2021-11-06 17:24:35 GenericNick2: do you actually need a DNS server on the machine? 2021-11-06 17:24:56 nope, the C library provides lookups via /etc/resolv.conf 2021-11-06 17:25:03 what's in your resolv.conf? 2021-11-06 17:25:41 I don't need a DNS server, no. All I wanted was to resolve localhost 2021-11-06 17:26:06 Is the 'dns' service supposed to be a dns server or resolver? 2021-11-06 17:27:14 GenericNick2: "localhost" is typically defined in /etc/hosts and the C libs resolver consults /etc/hosts file 2021-11-06 17:27:26 Ok, thanks 2021-11-06 17:28:57 GenericNick2: even if you did install a DNS server on the machine if you wanted programs on the machine to actually use it you would still need to add a "server 127.0.0.1" line into /etc/resolv.conf to use it 2021-11-06 17:30:39 I still don't understand why /etc/init.d/sshd has 'use dns' though 2021-11-06 17:35:01 "UseDNS Specifies whether sshd(8) should look up the remote host name and check that the resolved host name for the remote IP address maps back to the very same IP address. 2021-11-06 17:35:01 Enabling this makes access from a location without proper (forward and reverse) DNS generate a warning in the logs." 2021-11-06 17:35:57 nickersonm: he's referring to "use dns" in the init.d, not the sshd_config "UseDNS" setting 2021-11-06 17:36:28 GenericNick2: I assume it is there as the upstream OpenRC put it there (for some undetermined reason) ;-) 2021-11-06 17:36:30 minimal: oh, thanks. I misread the init.d path 2021-11-06 17:40:48 minimal: Yeah, seems so, Gentoo also has 'use dns' in it's sshd script 2021-11-06 17:41:06 s/it's/its 2021-11-06 17:41:06 GenericNick2 meant to say: minimal: Yeah, seems so, Gentoo also has 'use dns' in its sshd script 2021-11-06 17:41:10 it doesn't do anything being there 2021-11-06 17:41:13 you can just disregard it 2021-11-06 17:41:57 GenericNick2: oh, sshd, sorry that comes from openssh. A lot of the other init.d script come from openrc 2021-11-06 17:42:47 that *is* an Alpine specific file: https://git.alpinelinux.org/aports/tree/main/openssh/sshd.initd 2021-11-06 17:43:20 anyway its a "use" rather than a "need", they are trated differently by OpenRC 2021-11-06 17:44:18 GenericNick2: so I don't understand why you were having "localhost" problems 2021-11-06 17:45:20 I wasn't having any problems on my machine 2021-11-06 17:45:26 I was just paranoid if not using 'need dns' would lead to the service starting before domains can be resolved 2021-11-06 17:47:23 Last question, if I use a pidfile, should I use 'after bootmisc'? 2021-11-06 18:06:31 Hello, does alpine support the rootfs to be booted from nfs ? 2021-11-06 18:07:31 doesn't quite make sense the way I said it, but I am trying to PXE boot alpine on a rpi4b and I am getting this '/dev/nfs: Can't open blockdev 2021-11-06 18:07:39 any idea? 2021-11-06 18:25:13 "Yeah" <- that didnt work. I got some error like "kernel panic" 2021-11-06 18:25:27 ACTION uploaded an image: (162KiB) < https://matrix.org/_matrix/media/r0/download/matrix.org/tWMNlOLYYVtJXCXZdRyrhkOl/photo1636223054.jpeg > 2021-11-06 18:26:02 looks like an error you get if you're missing the initrd line 2021-11-06 18:28:26 Yesterday, when I tried to set up Alpine, i followed this guy https://wiki.alpinelinux.org/wiki/Installing_Alpine_on_HDD_dualbooting , but I didn't get the last section "Configuring the bootloader". I didn't really understand what to do. I thought I should reboot the machine, then go to Artix and try add Alpine to Artix's grub config.... 2021-11-06 18:28:37 s/guy/guide/ 2021-11-06 18:28:37 bhayek[m] meant to say: Yesterday, when I tried to set up Alpine, i followed this guide https://wiki.alpinelinux.org/wiki/Installing_Alpine_on_HDD_dualbooting , but I didn't get the last section "Configuring the bootloader". I didn't really understand what to do. I thought I should reboot the machine, then go to Artix and try add Alpine to Artix's grub config.... 2021-11-06 18:29:55 so, after running `setup-disk -m sys /mnt`, i reboot the system, and while doing so I already got some errors as I remember 2021-11-06 18:42:26 Latrina: without details of how you have set things up its hard for anyone to help 2021-11-06 19:02:05 psykose: which is what i said yesterday 2021-11-06 19:02:20 indeed :) 2021-11-06 19:03:23 bhayek[m]: as i said, if you want to use artix grub then you need to patch grub.d or make grub.cfg manually 2021-11-06 19:03:49 "Latrina: without details of..." <- I ran these commands... (full message at https://matrix.org/_matrix/media/r0/download/matrix.org/hfzHQJYOpvjPFacdWgfhOkdL) 2021-11-06 19:04:12 e.g. https://wiki.archlinux.org/title/User:Eschwartz/Grub 2021-11-06 19:04:31 https://wiki.archlinux.org/title/User:Eschwartz/Grub#Write_the_configuration_from_scratch 2021-11-06 19:04:39 Hello71: How can I do that? I thought enabling `os-prober` would be enough.. 2021-11-06 19:06:54 Hello71: thanks, I think this will be of great help! 2021-11-06 20:23:02 Hi. I'm trying to get gnome going on my system but gdm cannot start "Cannot open virtual console 1 (Permission denied)" 2021-11-06 20:23:02 Sounds like something isn't configured correctly, any ideas? 2021-11-06 20:23:39 add yourself to video/audio/input groups 2021-11-06 20:24:18 Keep in mind that gdm runs as the dedicated gdm user 2021-11-06 20:25:13 oh right gdm not gnome 2021-11-06 20:25:31 I think it might be because it shows the text login before it even gets to start gdm. You can see "Starting display manager [ok]" before the text login appears... 2021-11-06 20:26:27 My wild guess would be that the text login and graphical login (read gdm/xorg) are competing for tty1 2021-11-06 21:21:44 bhayek[m], the rpi boots the kernel from tfp, and upon the init system tries to mount the rootfs this is when I get that error 2021-11-06 21:22:42 rootfs folder is exported through NFSv3 from 172.12.1.57, exports here > /zfs/netboot/node0 *(rw,sync,no_subtree_check,no_root_squash) 2021-11-06 21:23:31 and this is the cmdline.txt syntax > console=serial0,115200 console=tty1 root=/dev/nfs nfsroot=172.12.1.57:/zfs/netboot/node0,vers=3,proto=tcp rw ip=dhcp rootwait cgroup_memory=1 cgroup_enable=memory 2021-11-06 21:24:37 nfs and nfsmount rc services are added to default 2021-11-06 21:24:52 nfs-utils is also installed on the rootfs 2021-11-06 21:25:24 if you need me to be more detailed I will need to write something down 2021-11-06 21:28:39 btw, how can I request to registed an account for the mailing list? 2021-11-06 21:28:50 *on the mailing list 2021-11-06 21:45:10 I really hope someone can help me out here. Or is there any other place where I could ask? 2021-11-06 21:45:36 mailing list thing is sorted.. 2021-11-06 21:47:11 Gaming4LifeDE, what's the problem? 2021-11-06 21:47:36 keep in mind that is almost 10pm and Saturday night 2021-11-06 21:47:54 not many people around at this time of the day 2021-11-06 21:48:04 and during weekend 2021-11-06 21:48:20 Latrina gdm won't start. Cannot open virtual console 1 (Permission denied) 2021-11-06 21:48:36 I'm trying to get gnome running on my system 2021-11-06 21:49:01 Sounds like the system won't transfer ownership of tty1 to the gdm user 2021-11-06 21:49:02 why would you want to use a graphical login manager? 2021-11-06 21:49:07 especially GDM. 2021-11-06 21:49:21 Because I'm trying to get gnome going 2021-11-06 21:49:40 can't you just bring it up using startx ? 2021-11-06 21:49:49 or sddm for instance? 2021-11-06 21:50:03 I could but I really don't want to 2021-11-06 21:50:16 anyways, I would need to investigate it. Never really faced this issye 2021-11-06 21:50:26 okay, good luck then 2021-11-06 21:50:49 Sounds weird, I really don't know how I'd approach this 2021-11-06 21:51:56 Interesting, the boot log says "starting display manager [ok]" but then the text login shows up. I'm guessing it's a conflict between text login and gdm 2021-11-06 21:52:11 no 2021-11-06 21:52:20 thats not the issue 2021-11-06 21:52:32 What is it then? 2021-11-06 21:52:38 well I dont know? 2021-11-06 21:52:47 is the user part of the video group? 2021-11-06 21:53:14 it is definitely a permission issue that's for sure. Not entirely sure whether specifically related to GDM or X rather 2021-11-06 21:53:38 X actually. It's in the xorg log 2021-11-06 21:54:43 you need to provide details in order for somebody to be able to troubleshoot it 2021-11-06 21:55:44 What kind of details? I'd give you the log but pasting on a terminal without a desktop isn't the easiest thing 2021-11-06 21:55:48 you should put the gdm user in the groups and make sure elogind is running 2021-11-06 21:55:58 pasting is simple enough 2021-11-06 21:56:10 nc termbin.com 9999 < logfile 2021-11-06 21:56:59 btw, since when Gnome works under Alpine? 2021-11-06 21:57:05 a long time 2021-11-06 21:57:08 cool 2021-11-06 21:57:21 Huh ok. It'll take a bit for me to get to my PC though, I'm on my phone right now 2021-11-06 21:57:24 always thought it was off limits 2021-11-06 21:57:54 I wonder why the gdm user wouldn't be in the correct groups from the start 2021-11-06 21:58:57 And what group would the gdm user need to be in 2021-11-06 21:59:36 some time ago it wasn't required as the whatever x stuff was all suid 2021-11-06 21:59:41 gnome 2021-11-06 22:00:01 normally thats the group 2021-11-06 22:00:11 video/audio/input, though i'm not sure if you need them with elogind 2021-11-06 22:00:16 i've always done both anyway 2021-11-06 22:00:24 psykose what exactly changed with x? 2021-11-06 22:00:45 not related to X whatsoever 2021-11-06 22:00:51 Latrina can't be in this case, since the x server cannot start 2021-11-06 22:01:02 well it's not suid anymore, so things need actual permissions 2021-11-06 22:01:03 oh doesn't it? 2021-11-06 22:01:07 Isee 2021-11-06 22:01:12 you can also post the x log 2021-11-06 22:01:44 Just need to figure out how to get a link copied in here. Hold on 2021-11-06 22:02:15 you can use sprunge.us 2021-11-06 22:02:18 you can type it, it's like 4 characters 2021-11-06 22:02:37 cat logfile | sprunge.us.url.whatver 2021-11-06 22:03:03 http://sprunge.us/ 2021-11-06 22:03:09 https://termbin.com/mtia 2021-11-06 22:03:42 psykose did that 2021-11-06 22:05:20 Does that even help? 2021-11-06 22:06:53 if you chmod u+s /usr/bin/X it should be fixed for now 2021-11-06 22:07:13 i can see you're on 3.14 and the rootless x stuff was still a bit broken 2021-11-06 22:07:41 Isn't 3.14 the latest release? I just downloaded the iso today 2021-11-06 22:07:56 as a stable release, yes 2021-11-06 22:09:10 psykose so you're implying I should rather use edge instead? 2021-11-06 22:09:48 i'm not implying anything 2021-11-06 22:10:05 it should work with u+s, then if it does it will work until 3.15 comes out and then it should be fixed 2021-11-06 22:10:15 i believe this is fixed on edge, but i didn't check gdm specifically 2021-11-06 22:10:24 Doesn't work :/ 2021-11-06 22:10:53 how do you test it after the changes 2021-11-06 22:10:59 Same log actually, nothing changed. Should I just try edge? 2021-11-06 22:11:02 Rebooted 2021-11-06 22:11:17 can you try to run startx instead? 2021-11-06 22:11:19 unfortunate 2021-11-06 22:11:20 plain startx 2021-11-06 22:11:42 should start into minimal twm if it works.. 2021-11-06 22:12:23 that to rule out X being the issue 2021-11-06 22:12:49 the primary problem I mean 2021-11-06 22:12:53 Doesn't work 2021-11-06 22:12:57 okay 2021-11-06 22:13:00 so X is the problem 2021-11-06 22:14:00 I am falling asleep. Good night all 2021-11-06 22:14:27 Bye and thanks for helping out 2021-11-06 22:14:40 np 2021-11-06 22:15:18 psykose any clues? Maybe just update to edge and try again? Although I don't think edge is exactly stable, is it? 2021-11-06 22:15:28 i suppose you can try edge 2021-11-06 22:15:49 edge is stable enough, i run it on desktops, then you can swap back to 3.15 in a month or so :) 2021-11-06 22:19:15 What's the release cycle for alpine? Haven't even thought about it much although I use it quite a bit 2021-11-06 22:19:43 https://alpinelinux.org/releases/ 2021-11-06 22:19:55 you can't keep beating me to these 2021-11-06 22:20:53 :P 2021-11-06 22:21:03 Funny how they say "each may and November" but that hasn't happened for the last 2 years at least 2021-11-06 22:21:20 3.12 was may :p 2021-11-06 22:21:30 Upgraded to edge, let's hope it works now 2021-11-06 22:22:47 Now it's completely borked, yay... 2021-11-06 22:23:22 It's at "loading initial ramdisk" and the screen just became stuck. There wasn't even any modest happening 2021-11-06 22:23:48 *modeset 2021-11-06 22:26:00 remove quiet 2021-11-06 22:26:22 I did notice my keyboard lighting to to out and reenable after a few seconds, maybe x just can't show the screen?? 2021-11-06 22:26:47 Hello71 I don't think I have a quiet parameter in the cmdline but I'll check 2021-11-06 22:26:58 Ffs can't even switch ttys anymore 2021-11-06 22:27:10 But Ctrl alt del works 2021-11-06 22:27:57 Quiet wasn't set in the first place 2021-11-06 22:30:54 I can reach it via ssh but my password won't work 2021-11-06 22:34:28 I don't know anymore but I'll head to bed now. Bye and thanks for your help guys. I'll try to check in tomorrow 2021-11-06 22:34:55 Just thought about it, no root login over ssh by default. Oh well, Ill deal with it tomorrow 2021-11-06 22:46:21 "e.g. https://wiki.archlinux.org..." <- i followed the guide but the custom entry doesn't show up on grub menu.. 2021-11-07 06:55:47 is there a browser in repo that works well with javascript switched on? 2021-11-07 08:23:13 anyone documented bare(cli) alpinelinux install on android-phones ? 2021-11-07 08:28:50 frag, firefox 2021-11-07 08:31:07 Seems like sddm and plasma desktop does not wanted to boot with Mesa DRI Intel(R) 945GM 2021-11-07 08:31:22 however, xfce4 and lightdm is savior 2021-11-07 09:00:40 lopid: aww ok 2021-11-07 09:14:13 Yonle: hi 2021-11-07 09:14:38 Heyya 2021-11-07 09:14:53 what is plasma doing 2021-11-07 09:16:04 Well, a plasma desktop, basically a DE 2021-11-07 09:16:16 While i start it, I got splash error, then plasma error 2021-11-07 09:16:31 can you start with startx? 2021-11-07 09:21:56 freq, Same thing happened. 2021-11-07 09:24:16 i'd say recompile it, but that's a big DE 2021-11-07 09:39:35 That's why 2021-11-07 09:40:13 What's the Plasma OpenGL requirement? For my CPU, it's 2.0 2021-11-07 10:28:45 It should support 2.0 just fine 2021-11-07 10:29:03 it has it as an option for the kwin rendering backend anyway 2021-11-07 10:29:39 vkrishn: most people just install postmarketOS. You can use either the none UI or the console UI for a barebones CLI installation, none UI being as barebones as you can get 2021-11-07 10:30:07 Yonle: idk minimum requirements for sddm, but I can't imagine it not supporting OpenGL 2.0 2021-11-07 10:30:17 could you post some logs? 2021-11-07 10:44:05 SDDM log file only shows that it exit without some of verbose error 2021-11-07 10:44:28 And yes. It should support OpenGL 2.0 since i try it on debian and it works on that opengl version 2021-11-07 10:51:36 please actually post the log 2021-11-07 10:58:29 PureTryOut: tried the postmarketOS scripts, but the device is quiet old (2013-14) 2021-11-07 10:58:54 so? 2021-11-07 10:59:11 currently just customized recovery, and piggy-backed on old kernel 2021-11-07 10:59:50 configured wifi, added alpine-repositories, and kinda working 2021-11-07 11:00:20 seems better to just install pmOS proper rather than making a frankenstein recovery? 2021-11-07 11:00:51 yes, I would want to 2021-11-07 11:01:32 I cannot figure out how to use internet via user account, works in root though 2021-11-07 11:01:45 that is in pmOS? with networkmanager? 2021-11-07 11:04:35 this install is just recovery, with minimal properties (init.rc), then at later stage a scripts takes over 2021-11-07 11:05:09 and and does cleanups and unmounts /system and enables wpa_supplicant 2021-11-07 11:05:55 kinda works link alpine install from cdrom (almost ;)) 2021-11-07 11:06:02 like^ 2021-11-07 11:06:56 practically all cli based progs work, even ssh, php, irssi 2021-11-07 11:07:12 nginx, apache 2021-11-07 11:08:10 but need to keep constant watch on it 2021-11-07 11:10:44 I was just wondering if any thoughts/work has been gone in re-using android init, with (get|set)prop 2021-11-07 11:10:47 ? 2021-11-07 11:11:01 lol no, nobody should want that 2021-11-07 11:11:08 :)) 2021-11-07 11:12:06 kinda works, one can add postboot-hooks-scripts 2021-11-07 11:14:06 can be made to not use /system or any other oem partitions 2021-11-07 11:14:43 adb reboot, just boots back android 2021-11-07 11:17:12 it all sounds more like hacks than a proper solution 2021-11-07 11:17:51 well if the .rc files are properly structured it might not 2021-11-07 11:18:05 putting in normalized way of adding hooks 2021-11-07 11:19:08 it would just be linux-boot, sans android (clean slate) 2021-11-07 11:20:28 more like multiboot 2021-11-07 11:20:39 hmmm... dual-boot 2021-11-07 11:21:35 piggy-backing on twrp is also quite possible 2021-11-07 11:22:23 which would then give a powerfull recovery install (ui+cli) 2021-11-07 11:26:47 whenever I use android, it feels like "Hardware as a Service" 2021-11-07 11:27:03 even if I own the device 2021-11-07 11:27:08 :) 2021-11-07 11:42:08 what would become hack is enabling/scripting GSM/HSPA without proper authorization channel 2021-11-07 11:42:31 till then everyting is yet another tiny-mini computer 2021-11-07 11:44:07 I would love to see modular mobile(hardware), specially with removalble mobile(radio) dongle 2021-11-07 11:45:30 removalble mobile(radio), gsm/hspa only ;) 2021-11-07 12:54:25 uhM, boot hungs on "Loading initial ramdisk ..." after upgrade and reboot 2021-11-07 13:28:33 I finally booted installing linux-edge 2021-11-07 14:04:34 Heyya! Mind asking how can i add watchdog to openrc autostart service? 2021-11-07 14:04:45 I afraid that i messing something up, 2021-11-07 14:05:03 By the way, i'm doing `rc-update add watchdog boot`. Is that command ok? 2021-11-07 14:15:53 `rc-update` will show you what's what 2021-11-07 14:50:39 Yes i know what rc-update does, My question is that ok to put `watchdog` to `boot`? 2021-11-07 14:53:40 i guess so 2021-11-07 14:54:02 `man openrc` describes each 2021-11-07 15:01:30 I asked a question yesterday, but couldn't stick around for the answer because bad internet :P 2021-11-07 15:01:36 Anyway, if I use pidfile, should I have 'after bootmisc' in depend? 2021-11-07 15:14:46 I installed alpine in a laptop last night and was able to use apk to install packages. Today I get 'Bus error' if I use adk add or apk update. 2021-11-07 15:17:12 is everything mounted ok? 2021-11-07 15:19:44 Yes I am logged into xfce, using furefox, etc... 2021-11-07 15:20:03 firefox 2021-11-07 15:21:32 could you run the same command again and give us the dmesg log? 2021-11-07 15:24:00 Will do, but away from laptop now, be back in 20 mins. 2021-11-07 15:49:13 Hi folks! I am trying to install alpine on my desktop. The installation works fine in itself, however the SSD on which alpine is installed does not appear in the boot list order. I tried to install Ubuntu instead and it works fine. My guess is that something prevent the firmware to recognize the EFI partition installed alongside alpine. If that 2021-11-07 15:49:14 kind of error rings a bell to someone, I would greatly appreciate to have some advice. 2021-11-07 15:51:02 Looked at dmesg, looks like I have a hardware disk problem, thanks. 2021-11-07 15:51:58 Leo40: do you run alpine alongside ubuntu? 2021-11-07 15:55:09 @caskd, No Alpine was erased when I installed Ubuntu. 2021-11-07 15:56:07 Leo40: did you use setup-alpine to install the system? 2021-11-07 16:00:32 @caskd Yes, I followed the instructions from the wiki. I did a sys install on /dev/sda. The installation completed successfully, but after rebooting I could no longer find the SSD on the boot list. 2021-11-07 16:02:22 I am looking into it, meanwhile, do you happen to have secure boot enabled? 2021-11-07 16:03:34 2nd, can you enable CSM and if yes, does it show up in CSM mode? 2021-11-07 16:06:10 or actually better, can you show the output of lsblk -o NAME,FSTYPE,SIZE,MOUNTPOINTS /dev/{DEVICE} (lsblk is included in util-linux) 2021-11-07 16:09:01 GenericNick2: It's possible, but I don't think it's required. I've never had issues with services that use pidfiles without declaring after bootmisc 2021-11-07 16:09:25 Ah ok 2021-11-07 16:09:27 Thanks 2021-11-07 16:12:48 @caskd Secure boot is disabled, CSM is disabled. If I enable CSM, the drive shows up and I can boot on it, but I end up in grub rescue. 2021-11-07 16:12:59 NAME FSTYPE SIZE MOUNTPOINT 2021-11-07 16:13:00 sda 119.2G 2021-11-07 16:13:00 ├─sda1 vfat 512M 2021-11-07 16:13:01 ├─sda2 swap 4G [SWAP] 2021-11-07 16:13:01 └─sda3 ext4 114.8G 2021-11-07 16:13:50 seems like you did a BIOS setup on a EFI system 2021-11-07 16:15:10 grub rescue is another weird thing, it means that the partiton where the stage 2 grub is installed is also not recognized 2021-11-07 16:18:39 I tried to mount /dev/sda1 and there is an EFI directory in it. So it looks like an EFI partition. 2021-11-07 16:21:06 There is two files in /dev/sda1. EFI/alpine/grubx64.efi and EFI/boot/bootx64.efi 2021-11-07 16:22:28 and the drive doesn't even show up in the boot menu when booting with EFI? 2021-11-07 16:23:09 No, it does not show. 2021-11-07 16:24:56 can you show the output of fdisk -l /dev/sda 2021-11-07 16:26:55 Disk /dev/sda: 119.25 GiB, 128035676160 bytes, 250069680 sectors 2021-11-07 16:26:56 Disk model: SAMSUNG SSD 830 2021-11-07 16:26:56 Units: sectors of 1 * 512 = 512 bytes 2021-11-07 16:26:57 Sector size (logical/physical): 512 bytes / 512 bytes 2021-11-07 16:26:57 I/O size (minimum/optimal): 512 bytes / 512 bytes 2021-11-07 16:26:58 Disklabel type: gpt 2021-11-07 16:26:58 Disk identifier: 01412E3A-F2CB-7543-ACB8-E27FC5B0BAE2 2021-11-07 16:26:59 Device Start End Sectors Size Type 2021-11-07 16:27:22 Uploaded file: https://uploads.kiwiirc.com/files/5b538d716e705a02312fdcdd0eee4d1f/pasted.txt 2021-11-07 16:27:56 copy/paste did some weird things, i uploaded a file instead. 2021-11-07 16:28:05 that's very odd, everything looks normal 2021-11-07 16:29:06 my best guess is there's some bios setting that disables this drive from showing up in EFI mode 2021-11-07 16:30:22 if you try writing the ISO to this drive, does it boot from it? 2021-11-07 16:30:41 I believe this might be a bios issue. However, everythiing works file when the EFI partition is created by Ubuntu installation. 2021-11-07 16:31:00 it might be either a BIOS issue or EFI quirk 2021-11-07 16:31:17 but as long as the setup didn't throw any error, this should work just fine 2021-11-07 16:42:46 I wrote the ISO on /dev/sda with dd. I get the exact same behavior, the drive does not show up in the list. 2021-11-07 16:43:43 At this point I am pretty sure this is a bios issue. I might have to change my motherboard. 2021-11-07 16:45:34 I would double check if ubuntu boots fine in EFI mode, there seems to be some kind of setting similar to secure boot that only lets the drive boot when the bootloader is signed 2021-11-07 16:45:56 and that setting doesn't affect peripherals 2021-11-07 17:11:12 There is no other setting similar to secure boot. However there is some kind of advance menu alongside secure boot to manage custom keys, but these settings are only available if secure boot is enabled. 2021-11-07 17:16:48 Leo40: you wrote which ISO on /dev/sda using "dd"? 2021-11-07 17:17:33 In anycase, thank you for your help. Knowing that the issue comes from the mb firmware is a good thing. I will try on another setup in as soon as I can get one. 2021-11-07 17:18:54 @minimal It was the alpine ISO from the website, but this was some kind of desperate attempt to get something working. 2021-11-07 17:20:53 Leo40: right, that didn't make sense. 2021-11-07 17:24:25 Leo40: so you don't have the Alpine install anymore to check more settings? 2021-11-07 17:25:31 @minimal However, I can boot on /dev/sda (only with CSM enabled) after copying the ISO with dd. But for some reason I can't install alpine on /dev/sda with the setup-alpine script, the disk does not show in the list. 2021-11-07 17:26:47 Leo40: does "dmesg" show the SSD? 2021-11-07 17:28:03 @minimal, I still have the USB key with the ISO, but I am out of ideas. And at this point, I am not sure if it is worth spending more time on it. 2021-11-07 17:29:13 Leo40: well that depends on whether you want to get Alpine working on it or not 2021-11-07 17:36:46 If you have any ideas, they are all welcome. 2021-11-07 17:38:18 Leo40: well if you try again if /dev/sda does not show up in the setup-alpine list then you should check if there actually is a /dev/sda device/file and if the SSD show up in "dmesg" output 2021-11-07 17:39:37 Leo40: this is Alpine 3.14.2 right? 2021-11-07 17:43:52 Yes this is 3.14.2. But the SSD not showing is because I dd the ISO in the SSD directly, this is not a clean install. When I install Alpine from the USB key I don't have that issue. 2021-11-07 17:45:10 Leo40: so the issue when installing from USB key is what exactly? 2021-11-07 17:46:31 The installation is successful but the SSD does not appear in the boot list after having reboot the computer. 2021-11-07 17:47:16 The BIOS is configured without CSM. But Alpine is installed with an EFI partition and should be bootable in such conditions. 2021-11-07 17:48:03 Secure Boot is disabled. 2021-11-07 17:48:27 If I installed Ubuntu in place of Alpine everything works fine 2021-11-07 17:49:44 The conclusion we had is that something in the BIOS prevent the SSD from showing in the boot list. 2021-11-07 17:53:37 Leo40: or that there is something about the ESP partition that the machine's UEFI does not like... 2021-11-07 17:55:14 @minimal Yes, it might be due t the way the partition is recognized by the bios. 2021-11-07 17:56:23 Leo40: did the machine eventually drop into the UEFI command line when it tried to boot? 2021-11-07 18:02:09 No, I did not see any UEFI command line. 2021-11-07 18:02:33 many vendors don't provide efi shell 2021-11-07 18:03:01 it could be ignoring efivars 2021-11-07 18:03:19 Leo40: doh! if the shell appears then you'd expect to see FS0: listed as the 1st FAT partition (i.e. ESP/EFI) one the disk that it spotted 2021-11-07 18:10:24 Hello71: shouldn't it still boot regardless of efivar settings as there is the "fallback" file inside the ESP partition? 2021-11-07 18:10:43 grub doesn't install one unless --removable 2021-11-07 18:13:07 Hello71: ah, think I've always used that myself routinely 2021-11-07 18:14:58 Hello71: looking at setup-disk quickly, for EFI it calls grub-install with --no-nvram so that won't set up efivars at all, right? 2021-11-07 18:18:51 Hello71: though setup-disk *does* create a fallback file manually after that 2021-11-07 18:25:42 mm. 2021-11-07 18:41:57 gnome-keyring-doc manpages are borked 2021-11-07 19:38:01 Just installed Alpine through chroot on a new machine and the the boot process is getting stuck on "EFI stub: Loaded initrd from command line option". What did I miss? 2021-11-07 19:39:18 There are no logs even without "quiet" flag 2021-11-07 19:41:43 alandiwix: yes I'm suspecting there's something wrong with UEFI - I've build 2 VMs, one using BIOS, other other using EFI and the EFI one stops when loading/running the initramfs where the BIOS one boots as expected 2021-11-07 19:41:52 I'm assuming you're using Alpine Edge also? 2021-11-07 19:42:24 Yes, I do 2021-11-07 19:43:50 you're using EFISTUB, I'm using Grub but we both appear to see the same behaviour. I'm seeing this with linux-lts 5.15.0 kernel, haven't tested linux-lts 5.15.1 yet but don't expect to see a change in behaviour 2021-11-07 19:44:01 have you tested linux-edge instead to see if that works? 2021-11-07 19:44:46 not yet 2021-11-07 19:46:40 Did you try debug_init? 2021-11-07 23:05:33 apk fix and apk [add or del] outputs a list of packages and their conflicts/satisfies 2021-11-07 23:06:06 I'm stuck unable to manage the system. How can I break out of this? 2021-11-08 00:35:27 bovis, does this work? /sbin/apk --force-broken-world [add, del] 2021-11-08 01:19:08 also did you try to sacrifice a pigeon or two? 2021-11-08 01:21:12 Hi, I've got a few USB drives with slow storage that I'd like to use for a diskless install. I was able to follow the steps at https://wiki.alpinelinux.org/wiki/Create_a_Bootable_Device#Manually_copying_Alpine_files but am receiving a `Mounting boot media: failed` message. Enabling debug and debug_init in the cmdline shows that the USB drive isn't brought up before nlplug-findfs times out. 2021-11-08 01:22:21 https://wiki.alpinelinux.org/wiki/Create_a_Bootable_Device#Slow_USB_Devices suggests adding waitusb=3 to cmdline, but it doesn't appear to work. Looking through nlplug-findfs.c, it appears the USB timeout is hard-coded. Are there any other kernel parameters I should use to boot Alpine on this slow USB drive? 2021-11-08 01:40:32 is it a bad idea to install `linux-lts` from edge but everything else from stable 2021-11-08 01:41:32 dthusian: yes 2021-11-08 01:42:21 if I need the latest kernel, what better options do I have? 2021-11-08 01:42:38 you will break things doing that 2021-11-08 01:43:56 I've been using edge as my daily driver for over a year 2021-11-08 01:44:02 option would be to install Edge if you want the kernel in Edge. However there appear to be issues with the current kernel in Edge - some people are seeing boot problems with it. 2021-11-08 01:44:23 yes, I am seeing boot problems with the LTS kernel in edge, so I use the edge kernel in edge 2021-11-08 01:44:56 Edge's linux-lts now uses compressed kernel modules, if you install it on non-Edge then the rest of your system (i.e. mkinitfs) wouldn't be able to handle that 2021-11-08 01:45:42 valerius: yes, I have recently added a note to the issue tracking that 2021-11-08 01:46:33 thanks for your help, I'll consider whether Edge is good for my usecase 2021-11-08 01:46:50 what's your use case? 2021-11-08 01:47:06 home server 2021-11-08 01:47:21 yeah, for that I'd advise using stable 2021-11-08 01:48:00 dthusian: if linux-edge ends up enabling kernel module compression then you will at that point also have problems with it on non-Edge 2021-11-08 01:58:15 lzsaver: That allowed me to install an arbitrary package, but the list of conflicts/satisfies remains. What am I looking at when this happens? 2021-11-08 02:00:44 Apachez: the sacrifices haven't helped a bit 2021-11-08 02:05:15 I'm also seeing now the error '/bin/ash: task: not found' 2021-11-08 02:29:24 bovis, just resolve conflicts (like gentoo guys do). look your /etc/apk/world 2021-11-08 02:30:19 bovis, probably you have some packages what do not like each other 2021-11-08 02:37:02 lzsaver: What would be unusual there? It's a list of what I've installed. I do see though that [firefox, task, a few other programs] are in the file, but they were deleted and I can't uninstall them. 2021-11-08 02:38:19 The best I get now from "--force-broken-world add" is the "OK" output. 2021-11-08 02:38:46 *can't reinstall them 2021-11-08 02:44:08 bovis, backup this file. try to remove deleted packages from the file. feel free to add them back using apk add 2021-11-08 02:44:54 (I'm not alpine pro. just gentoo guy. someone here can help better) 2021-11-08 02:46:49 the thing you need to do is that your world should have no conflicts 2021-11-08 02:47:59 probably you may want to check that apk upgrade does work right without errors 2021-11-08 02:48:20 what's your /etc/apk/repositories by the way 2021-11-08 02:49:38 Repositories are all edge for main, community, testing 2021-11-08 02:50:10 ah.. so you should be okay with broken system 2021-11-08 02:50:27 if you're on edge - you're the beta-tester 2021-11-08 02:51:16 you may try to replace all lines from it with these two lines if you do not want to be on edge anymore: 2021-11-08 02:51:16 https://mirror.yandex.ru/mirrors/alpine/latest-stable/community 2021-11-08 02:51:16 https://mirror.yandex.ru/mirrors/alpine/latest-stable/main 2021-11-08 02:51:34 (replace mirror.yandex.ru with your fav mirror) 2021-11-08 02:52:11 Editing the world file (remove mpv), then trying to apk add mpv didn't do anything different. Still just the "OK". 2021-11-08 02:53:30 if you need just few packages from edge you may add something like this (alpine guys may correct my words) 2021-11-08 02:53:30 @edge https://mirror.yandex.ru/mirrors/alpine/edge/community 2021-11-08 02:53:30 after that you can install some app from edge like this: apk add mpv@edge 2021-11-08 02:53:30 @edge https://mirror.yandex.ru/mirrors/alpine/edge/testing 2021-11-08 02:53:30 @edge https://mirror.yandex.ru/mirrors/alpine/edge/main 2021-11-08 02:53:30 all other packages will be from stable branch 2021-11-08 02:54:51 lzsaver: I know they're bleeding edge packages, but this situation feels like more that I screwed up the system, not that it's a broken package problem. 2021-11-08 02:55:54 bovis, did you do some alias for apk? try this: alias apk='/sbin/apk' 2021-11-08 02:56:05 yeah, probably your apk is just broken somehow 2021-11-08 02:57:38 Haha, switching the repositories to main, then trying to upgrade made it a lot worse. Output is more crowded with a list of packages, breaks, satisfies, but still no ability to install them. 2021-11-08 02:58:41 lzsaver: Using /sbin/apk doesn't affect the result 2021-11-08 02:59:48 try to reinstall apk-tools from main 2021-11-08 03:02:44 lzsaver: I tried force broken world with one of the packages there without a conflict listed. It's basically updating the system now. So: progress. I'll let you know when it's done. 2021-11-08 03:04:03 okay. perhaps it will update apk-tools too 2021-11-08 03:05:00 do you have important files on this system? 2021-11-08 03:07:04 keep in mind that you can lose them if your system does not work correctly 2021-11-08 03:32:37 lzsaver: Some progress. Now with apk I'm only seeing output like "surf (no such package): required by world[surf]" 2021-11-08 03:34:09 That's for about a dozen packages. Apk thinks I should have them, but they're not installed. Is /etc/apk/world the only place it keeps track of what I *should* have? 2021-11-08 03:37:58 yep, I think so 2021-11-08 03:39:54 it seems that surf is available in testing repo. you should install each of those packages with (for example) apk add surf@edge 2021-11-08 03:39:54 replace @edge with the repo name as you defined it in /etc/apk/repositories 2021-11-08 03:40:22 or you may just remove them if you do not need them 2021-11-08 03:44:05 the idea is to use the latest-stable repo for the system and add packages from other repos using apk add packagename@reponame 2021-11-08 03:46:21 if you used gentoo it may be usual way for you ( with ATOMs like "www-servers/caddy::gentoo" in portage, where gentoo is a repo name ) 2021-11-08 04:20:38 > each of those packages 2021-11-08 04:20:38 I mean packages that are not available in latest-stable 2021-11-08 04:30:17 lzsaver: Thanks for the tip about using testing. I'll be doing that in the future. 2021-11-08 04:31:02 Unsure how exactly, but I organized the database and everything looks to be back to normal. 2021-11-08 04:32:09 nice. good job 2021-11-08 05:55:31 is the mips64 iso little or big endian? 2021-11-08 05:59:47 google makes me think its little 2021-11-08 06:00:34 https://wiki.alpinelinux.org/wiki/MIPS64/Known_Issues says the MIPS64 port is big 2021-11-08 06:01:18 oh interesting 2021-11-08 06:02:12 the search result i found was that the main mips64 page says it should work on the Loongson boards and wikipedia says all the Loongson boards are le 2021-11-08 06:05:48 looked again and the wikipedia article says citation needed 2021-11-08 06:05:56 so the link you posted is prolly right 2021-11-08 06:06:23 I just extracted the busybox package and its bin/busybox is big endian 2021-11-08 06:06:51 thanks :) 2021-11-08 06:07:46 i started work recently on building a ci and wanted to make sure i identified the target correctly 2021-11-08 07:00:30 hello? 2021-11-08 07:00:43 hello 2021-11-08 07:01:44 Alpine/xfce4/ how could this computer be hacked when connected to wifi? 2021-11-08 07:02:37 thats not really a question thats specific to either alpine or xfce 2021-11-08 07:10:29 If I use alpine, and connect to internet via GUI on xfce4, what vulnerabilities could an opposing force on the same network be able to exploit. To add further context, if I have artistic intellectual property on my hard drive in the form of mp3, or .txt. This is the real world scenario I am facing. 2021-11-08 07:14:11 Is it possible to harden Alpine/xfce4 to be impossible to hack. Or should I focus on having a strict gap, and keep top secret information on computers that are entirely offline 2021-11-08 09:10:17 heh! 2021-11-08 13:28:07 hello 2021-11-08 13:29:00 I have installed alpine on a testing virtual machine, alpine is installed and booting from /dev/vda 2021-11-08 13:29:19 how to make my changes persistent after a reboot ? 2021-11-08 13:31:20 How did you install alpinelinux? 2021-11-08 13:31:54 If changes are not persisted, it means you did a run from ram install 2021-11-08 13:33:09 @ikke I booted from iso downloaded from alpine website, and then I ran setup-alpine and setup-disk 2021-11-08 13:33:21 I have choosen sys install mode 2021-11-08 13:37:29 I have removed the iso/cdrom and it now boots from the local disk, but I don't know how to use lbu or make packages that I have installed persistent on the disk 2021-11-08 13:40:11 on the disk /dev/vda1 is /boot, vda2 is swap, vda3 is root filesystem 2021-11-08 13:40:19 If you did a sys install, that is not necessary 2021-11-08 13:40:33 but It appears not mounting these directory when booting 2021-11-08 13:42:29 ikke, I found, was my failure, I have removed cd-rom from virt-manager but I forgot to power it off, so even cd-rom was not appearing anymore, it was booting from cd-rom / iso 2021-11-08 13:44:28 Aha 2021-11-08 15:53:44 Trying to build opencv-4.5.4 from source and no matter what I do it seems to not found v4l/v4l2 support. I have installed linux-headers v4l-utils-dev v4l-utils-libs, am I missing any other packages? 2021-11-08 15:59:46 I'm looking for information on an updated package for ncruses that resolves CVE-2021-39537. Does anyone know if an update is being worked on for 3.14.2? 2021-11-08 16:00:30 acidtonic_: do the makedepends/depends_dev from https://git.alpinelinux.org/aports/tree/community/opencv/APKBUILD work? 2021-11-08 16:00:41 oops, that ncurese 2021-11-08 16:00:50 :P 2021-11-08 16:04:03 I guess I'm having all sorts of typing problems today :-) 2021-11-08 16:06:27 Ideally, I would like to find a packages that is compiled and in the standard repo so we don't have to update our current processes. 2021-11-08 16:31:15 I'm trying to access my laptop /dev/sda from a rescue disk. It needs to be decrypted. pvdisplay "fails to find the physical volume" and pvscan/vdscan verbose output shows the drive is "Not processing filtered" 2021-11-08 16:31:52 This is the default lvm.conf. Why can't I access the disk? 2021-11-08 16:34:19 you probably need to cryptsetup luksopen it first? 2021-11-08 16:38:45 if he got LVM on LUKS then ye but some people doing LUKS on LVM 2021-11-08 16:48:58 psykose: Thank you. That worked. But one followup: why does mounting error with "invalid argument" even when I include the type ext4? 2021-11-08 16:49:14 mounting what 2021-11-08 16:55:04 psykose: mount -t ext4 /dev/mapper/lvmdrive /mnt/on/rescuedisk 2021-11-08 16:55:38 if lvmdrive is what luksopen outputs then that isn't an ext4 partition, but the lvm root that you still have to mount 2021-11-08 16:56:01 dmesg would give you another error too 2021-11-08 17:00:48 psykose: I'm trying to mount that volume. How to I find out the type? 2021-11-08 17:01:35 if you are using lvm you need to use the lvm commands to get the lvm volumes to show up, don't you 2021-11-08 18:23:02 good evening folks. I am unable to mount the rootfs from NFS during a PXE boot on my rpi 4b. Everything is detailed here https://lists.alpinelinux.org/~alpine/users/%3Cb444fbf4-c2f0-9ab9-9a5a-8c68725381ca%40googlemail.com%3E 2021-11-08 18:23:09 thanks in advance 2021-11-08 18:32:32 Latrina: I didn't notice any sign of the Pi's network interface initialising in the kernel output you posted 2021-11-08 18:37:45 Latrina: for example on a RPI3 when the ethernet is active dmesg shows "[ 31.781267] smsc95xx 1-1.1:1.0 eth0: Link is Up - 100Mbps/Full - flow control off". If the network interface is not active then obviously NFS cannot work 2021-11-08 18:37:53 don't you need rootfstype=nfs or is that optional 2021-11-08 18:40:20 psykose: not sure, the "rootfstype=" option is used to trigger a kernel module load for the relevant module but NFS is build into the RPIs kernel so that's not an issue 2021-11-08 18:40:32 ah, i see 2021-11-08 18:40:34 thanks for the info :) 2021-11-08 18:42:17 psykose: actually it does also pass the value of "rootfstype" to the mount command so I guess it is needed to be present for that 2021-11-08 18:43:32 i dunno, i've personally always had mounts work without it for regular install with ext4/xfs/btrfs/whatever, but maybe nfs is special case 2021-11-08 18:43:58 Latrina: try adding "rootfstype=nfs" to your /var/tftpboot/node0/cmdline.txt file 2021-11-08 18:43:59 just every single nfs copypaste guide out there has it it seems 2021-11-08 18:44:11 minimal, that I didn't notice, the interface not being initialised 2021-11-08 18:44:20 I mean it booted the kernel 2021-11-08 18:44:56 though, thats a different process. I am confused 2021-11-08 18:45:34 psykose, I don't understand. Do you mean I don't need rootfstype=nfs? 2021-11-08 18:46:12 i'm not saying either that you need it or don't, merely a passing thought that maybe you do 2021-11-08 18:46:28 psykose: mount inside the initramfs (and also in "normal" Alpine unless you install util-linux-misc for "full" mount) is not as smart as usual mount, I've found for cloud-init to work with ISO/VFAT devices/partitions I had to create a /etc/filesystems file to get it to automatically handle those 2021-11-08 18:46:47 as its Busybox's stripped down mount command 2021-11-08 18:47:04 makes sense, i've seen plenty of cases of people having issues missing -t type 2021-11-08 18:47:07 I don't remember ever using it before, and it always worked under debian. I will add it 2021-11-08 18:47:25 Latrina: but this is not Debian... 2021-11-08 18:47:31 sure :) 2021-11-08 18:47:38 hey I am glad if that's the issue 2021-11-08 18:47:51 this is technically a case where the full-mount functionality is better :p 2021-11-08 18:49:32 booting now 2021-11-08 18:50:22 lets see 2021-11-08 18:51:12 no go 2021-11-08 18:51:57 minimal, are you saying that perhaps the network service is not enabled? 2021-11-08 18:52:09 psykose: checked my notes, the issue with Busybox "mount" was to do with "-t auto" - works fine with util-linux mount but does not with Busybox mount unless you create a /etc/filesystems file containing a linue referencing the kernel module to be loaded for the filesystem type 2021-11-08 18:52:10 this is before any servi es 2021-11-08 18:52:15 services* 2021-11-08 18:52:35 minimal: yep, exactly what i had in my thoughts 2021-11-08 18:52:56 Latrina: no, I'm saying you should add "rootfstype=nfs" to cmdline.txt and try that 2021-11-08 18:53:02 the network rc service cannot be disabled as tcp/ip works on the mmc and I am currently sshing into it 2021-11-08 18:53:14 the rootfs on the nas its a mere copy of the mmc 2021-11-08 18:53:18 psykose: util-linux's mount is clever enough to modprobe the right kernel module itself 2021-11-08 18:53:24 indeed 2021-11-08 18:53:26 minimal, I just did. It didnt work 2021-11-08 18:53:44 Latrina: ok, so did anything change in the output you saw? 2021-11-08 18:54:08 checking 2021-11-08 18:55:53 yes 2021-11-08 18:55:56 Latrina: does your /boot/initramfs contain udhcpc? (needed for the initramfs' init to get an IP prior to mounting the NFS route) 2021-11-08 18:56:01 the output seems different 2021-11-08 18:56:22 Latrina: what is different? 2021-11-08 18:56:53 minimal, I am not sure. If udhcpc is in the initramfs provided with the rpi aarch64 image then yes, otherwise I don't 2021-11-08 18:57:08 the oputput looks indeed different 2021-11-08 18:58:07 https://pastebin.com/QUbiaLSR 2021-11-08 18:58:43 root fails but the output is different 2021-11-08 18:58:58 root mount 2021-11-08 19:00:52 it just says mounting root failed. Doesn't really say why 2021-11-08 19:01:34 Latrina: I believe udhcpc is not present in your initramfs and so the initramfs' init cannot do DHCP to bring up the ethernet and so therefore obviously it cannot mount the NFS fs 2021-11-08 19:02:26 minimal, would it make a difference if I sent the interface a static ip? 2021-11-08 19:02:31 *set 2021-11-08 19:03:00 I really don't need dhcp 2021-11-08 19:03:16 although it is what I am using in this case 2021-11-08 19:03:47 might need network in the mkinitfs.conf 2021-11-08 19:03:52 Latrina: you have "ip-dhcp" in your cmdline.txt file. If you want static IP then change this accordingly 2021-11-08 19:04:03 ah 2021-11-08 19:04:07 grr 2021-11-08 19:04:17 and dhcp in mkinitfs for dhcp 2021-11-08 19:04:32 is that in the initramfs? 2021-11-08 19:04:47 .conf 2021-11-08 19:05:00 psykose: yes, though that only adds the udhcpc script, don't see where the udhcpc binary gets added 2021-11-08 19:05:18 fsck lets make it all static 2021-11-08 19:05:51 Latrina: basically the standard initramfs supplied for RPI is not designed for network booting, so typically if you want that you need to reconfigure and rebuild it 2021-11-08 19:06:10 minimal, makes sense 2021-11-08 19:06:19 although I wish they had it in mind 2021-11-08 19:06:28 that someone could have wanted to PXE boot it 2021-11-08 19:06:35 Latrina: there are Alpine netboot image, just not for RPI 2021-11-08 19:06:46 I see 2021-11-08 19:07:04 Latrina: RPI does not PXE boot, it may network boot. "PXE" is an x86/x86_64 specific term 2021-11-08 19:07:05 if the workaround is adding a static ip to cmdline.txt than I don't really care 2021-11-08 19:07:07 minimal: udhcpc is part of busybox, it can't not be added 2021-11-08 19:07:22 minimal, yes it is netboot in fact 2021-11-08 19:07:40 psykose: doh! of source, and initramfs's init will create the softlink to it during startup 2021-11-08 19:07:58 s/source/course/ 2021-11-08 19:07:58 minimal meant to say: psykose: doh! of course, and initramfs's init will create the softlink to it during startup 2021-11-08 19:08:20 Latrina: the "ip=" entry for static has a specific format 2021-11-08 19:08:37 and with ip is that the NFS ip server? 2021-11-08 19:08:54 isnt that already provided with nfsroot? :S 2021-11-08 19:09:04 or is it the pi IP rather? 2021-11-08 19:09:08 but that file is needed for ip=dhcp to work, at least if this is what gets run https://github.com/alpinelinux/mkinitfs/blob/1bffc0c7e7563ba9ce6471979cfe18616d36423e/initramfs-init.in#L193 2021-11-08 19:09:17 "ip=client-ip:server-ip:gw-ip:netmask:hostname:device:autoconf::dns0-ip:dns1-ip:ntp0-ip" 2021-11-08 19:10:15 psykose, what is that? 2021-11-08 19:10:31 i would try to modify the mkinitfs first with a dhcp in the features= and apk fix mkinitfs just to be sure :) 2021-11-08 19:10:54 cool, I'll leave that for later 2021-11-08 19:11:11 but it'd indeed be cool to have dhcp to work 2021-11-08 19:11:32 minimal, I don't think I understood you 2021-11-08 19:11:35 Latrina: the line I posted above is the syntax of the "ip=" option for static IP use 2021-11-08 19:11:49 Latrina: it's the init= file the initrd runs 2021-11-08 19:11:55 oh I need to declare the whole shebang? 2021-11-08 19:12:25 LAtrina: have you read this? https://www.kernel.org/doc/Documentation/filesystems/nfs/nfsroot.txt 2021-11-08 19:12:34 no 2021-11-08 19:12:44 it documents all the relevant cmdline options including the "ip=" one 2021-11-08 19:13:56 I will give it a good read indeed, thanks 2021-11-08 19:29:53 I am not sure the IP address has got anything to do with my problem 2021-11-08 19:29:58 I declared all static 2021-11-08 19:30:01 same error 2021-11-08 19:30:24 adding rootype=nfs definitely help 2021-11-08 19:30:46 Latrina: did you modify the mkinitfs config and rebuild the inirfamfs? 2021-11-08 19:30:51 but still still mounting the rootfs fails 2021-11-08 19:31:01 minimal, no I didn't 2021-11-08 19:31:15 I didn't even know of its existance up until 30 minutes ago 2021-11-08 19:31:19 maybe less 2021-11-08 19:31:32 look at /etc/mkinitfs/mkinitfs.conf, what is currently in the features list? 2021-11-08 19:32:02 what do you mean with features list? 2021-11-08 19:32:09 the line inside that file 2021-11-08 19:32:23 let me see 2021-11-08 19:32:53 there is no such mkinitfs folder under /etc 2021-11-08 19:33:08 ok, then "apk add mkinitfs" 2021-11-08 19:33:37 let me boot it up 2021-11-08 19:36:28 added 2021-11-08 19:37:32 features="ata base cdrom ext4 keymap kms mmc nvme raid scsi usb virtio" 2021-11-08 19:37:42 these are the defaults 2021-11-08 19:38:05 so you should add "dhcp" to that list 2021-11-08 19:38:22 where do I find the features list? 2021-11-08 19:38:35 do I need to add anything related to nfs? 2021-11-08 19:39:09 oh and "network" just to be safe 2021-11-08 19:39:59 then you need to rebuild the /boot/initramfs by running "mkinitfs" 2021-11-08 19:40:14 added 2021-11-08 19:40:15 then copy that onto your NFS location 2021-11-08 19:40:37 minimal, will it just create a initramfs? 2021-11-08 19:40:48 what do you mean? 2021-11-08 19:41:26 what does this tool exactly create? just an initramfs? 2021-11-08 19:41:31 or the kernel as well? 2021-11-08 19:41:47 yes, that's why its called "mkinitfs" for "make initramfs" 2021-11-08 19:41:57 lmao, I am so lame 2021-11-08 19:41:58 sorry 2021-11-08 19:41:59 it will replace the /boot/initramfs file with a new one 2021-11-08 19:42:04 cool 2021-11-08 19:42:35 then copy that file to the right place on your NFS server and try booting again 2021-11-08 19:43:09 minimal, here they also add nfs https://wiki.alpinelinux.org/wiki/Talk:PXE_boot 2021-11-08 19:43:16 should I add it as well? 2021-11-08 19:46:26 minimal, ? 2021-11-08 19:49:41 Latrina: most of the NFS stuff is compiled into the kernel (apart from NFS server) so you shouldn't need to do that 2021-11-08 19:49:51 s/the kernel/the RPI kernel/ 2021-11-08 19:49:51 minimal meant to say: Latrina: most of the NFS stuff is compiled into the RPI kernel (apart from NFS server) so you shouldn't need to do that 2021-11-08 19:50:11 thanks man 2021-11-08 19:57:06 Latrina: is it working now? 2021-11-08 19:57:42 just built the new initramfs 2021-11-08 19:58:00 had to install the kernel apparently as the kernel modules weren't there 2021-11-08 19:58:32 I will copy the new image over and try 2021-11-08 19:58:39 will take few more minutes :) 2021-11-08 19:59:58 minimal, how come you happen to know so much? Are you a alpine dev? 2021-11-08 20:07:23 ah of course, yes the initramfs puts kernel modules into initramfs so of course it need them. Strange they were not present as I'd expect RPI to need at least some of them to boot 2021-11-08 20:08:01 the alpine rpi aarch64 image only provides the bootfile 2021-11-08 20:08:11 then I had to create an overlay of the installer 2021-11-08 20:08:19 to have an actual userland to boot into 2021-11-08 20:08:45 Latrina: I'm a Alpine package maintainer and I also build my own custom Alpine disk images including for RPIs 2021-11-08 20:08:49 and once booted the persistant userland was installed through apk 2021-11-08 20:09:03 sweet 2021-11-08 20:09:47 I have been using alpine for the last few years and it is my defacto favourite distro 2021-11-08 20:10:05 but I admit I never had to create an initramfs 2021-11-08 20:10:38 and there are still few underground areas I am surely not aware of 2021-11-08 20:10:56 but so far I love it. kudo to you and everyone else 2021-11-08 20:12:42 I have done quite a bit of embedded with gentoo though 2021-11-08 20:16:32 alpine works like a charm :) 2021-11-08 20:16:42 just wish that some packages should be updated and have better defaults 2021-11-08 20:16:46 it does indeed 2021-11-08 20:16:49 for example munin monitoring 2021-11-08 20:16:59 ACTION moment of the truth 2021-11-08 20:19:49 boot still fails 2021-11-08 20:19:50 haha 2021-11-08 20:19:54 I am loving it 2021-11-08 20:20:12 Wouldn’t have been fun otherwise. ;) 2021-11-08 20:20:23 eh 2021-11-08 20:20:24 ACTION stays silent but follows the all adventure. 2021-11-08 20:20:59 Latrina: any different in output during boot? what is you "ip=" setting? 2021-11-08 20:21:19 logs here https://pastebin.com/raw/n0r91UqY 2021-11-08 20:21:44 I have dhcp as ip= 2021-11-08 20:21:58 we just added dhcp to initramfs 2021-11-08 20:22:51 I wonder if I should have added nfs to the mkinitfs features.. 2021-11-08 20:24:39 anyways, will try again tomorrow I guess. The never ending saga... 2021-11-08 20:24:54 Thanks minimal and psykose 2021-11-08 20:25:36 i didn't do anything :) 2021-11-08 20:25:49 if anybody has got any idea as to why this thing refuses to mount the rootfs via nfs, please reply here :) https://lists.alpinelinux.org/~alpine/users/%3Cb444fbf4-c2f0-9ab9-9a5a-8c68725381ca%40googlemail.com%3E 2021-11-08 20:26:02 psykose, you did confuse me further than I already was 2021-11-08 20:26:03 :D 2021-11-08 20:26:34 good night 2021-11-08 20:26:36 sleep well 2021-11-08 20:26:44 ty 2021-11-08 20:53:48 Heuj 2021-11-08 20:53:51 Een Peetz0r 2021-11-08 20:54:03 ? 2021-11-08 20:54:09 barada, nikto 2021-11-08 20:55:04 That sounds suspiciously dutch 2021-11-08 20:57:45 Yes :) sorry for that 2021-11-08 20:58:07 (ikke ikke ikke en de rest kan stikke?) 2021-11-08 20:59:58 :) 2021-11-08 21:02:23 we already have a Dutch Peter in here 2021-11-08 21:02:25 quota is filled 2021-11-08 21:02:27 shoo 2021-11-08 21:02:36 (hi Peetz0r) 2021-11-08 21:02:52 Oh lol, hi Habbie 2021-11-08 21:02:57 hi :D 2021-11-08 21:03:27 Oh hello fellow totally not Dutch totally not Peter :) 2021-11-08 21:03:49 fun fact, I am sitting at RevSpace and another Dutch Peter just entered here too :p 2021-11-08 21:05:26 this has to stop 2021-11-08 21:06:10 (which one?) 2021-11-08 21:17:02 It's pbx 2021-11-08 21:20:48 ah, no bells 2021-11-08 21:29:52 Anyway, I have a legit alpine question. Trying to make some machine run X on a machine with only 2GB storage. So I installed Alpine, and a bunch of packages including whatever setup-xorg-base installs and then lightdm, but it fails to actually start. 2021-11-08 21:30:32 are you member of the video / input group? 2021-11-08 21:30:39 Also when I do startx on the tty it fails too, with no real errors, just "[ 3689.082] (II) Server terminated successfully (0). Closing log file." or similar 2021-11-08 21:31:05 eh, I tried things as root to rule out permission issues. should I try a normal user in those groups instead? 2021-11-08 21:31:58 did you check the xorg log? 2021-11-08 21:32:57 yes, that quoted line is the last line in there. No errors, just a few warnings that seem unrelated 2021-11-08 21:38:03 nvm, fixed it 2021-11-08 21:38:12 lightdm was installed, but with no greeters 2021-11-08 21:38:18 fixed that, and now it starts fine :) 2021-11-08 21:54:27 has anyone tried to use alpine as a firewall/router system? 2021-11-08 22:04:57 jhjacobs81: suits well for this task 2021-11-08 22:19:37 hi, I've installed alpine on a laptop and I'm trying to set up an X desktop and xfce but when I try and start x it appears to load as normal, but then fails with (EE) Caught signal 6 (Aborted) 2021-11-08 22:19:58 let me get the entire x11 log. This is from running startx as a user I created who is in the wheel and video groups 2021-11-08 22:20:00 wait 2021-11-08 22:20:57 Probably need to install mesa-egl or something related 2021-11-08 22:21:21 Do you see something in the logs about libegl? 2021-11-08 22:22:09 OK, fixed it by installing mesa-egl and actually adding exec startxfce4 to my xinitrc 2021-11-08 22:22:23 but now when xfce starts, it seems to start fine but no input works. No keyboard or mouse input at all 2021-11-08 22:22:27 just frozen 2021-11-08 22:22:36 Need to be in the input group as well 2021-11-08 22:23:05 ah 2021-11-08 22:23:13 k 2021-11-08 22:23:27 also, even though I'm in the wheel group, `doas` keeps saying I don't have permission to run things as root 2021-11-08 22:23:31 is there another group I'm missing? 2021-11-08 22:23:43 input and video should do it 2021-11-08 22:23:53 maybe audio as well for sound 2021-11-08 22:25:47 OK I'm in the input group but my mouse and keyboard input in X are both still not working 2021-11-08 22:26:17 I think I'm in those groups anyway, I did `adduser nico input` 2021-11-08 22:26:27 did you login again? 2021-11-08 22:27:10 oh 2021-11-08 22:27:12 lol 2021-11-08 22:44:37 now it all works! I have a basic desktop 2021-11-08 22:45:24 next question: as this is on a laptop I want to be able to connect and disconnect to/from different wifis easily. On my desktop I use networkmanager for this. Does networkmanager run on alpine? Is there a better alternative? 2021-11-08 22:47:01 Yes, it works on alpine. 2021-11-08 22:47:19 k 2021-11-08 22:47:49 Isn’t there netifrc on alpine, too? 2021-11-08 22:48:04 I recommend using iwd as the backend though. YMMV. 2021-11-08 22:48:26 ok 2021-11-08 22:48:44 i use wpa supplicant but nm + iwd is probably wisest if you need to connect to many different networks as opposed to a set amount of networks 2021-11-08 22:48:46 that you move between 2021-11-08 22:48:56 yeah I connect to a lot of different nets 2021-11-08 22:48:59 I'll go with nm 2021-11-08 22:49:00 shouldnt be that hard to write an equivalent script for wpa supp, though 2021-11-08 22:49:06 if you're logging in often 2021-11-08 22:49:22 i get a weird bug with wpa supplicant though... 2021-11-08 22:49:48 hmm 2021-11-08 22:49:51 idk 2021-11-08 22:50:05 I could never get wpa supplicant to auto connect.... Dunno what the deal with that is. No issues manually connecting though. 2021-11-08 22:50:39 wpa supp autoconnects fine for me on my laptop 2021-11-08 22:50:40 odd 2021-11-08 22:51:01 but when i suspend (by closing screen), and move to a different network, it'll connect without internet 2021-11-08 22:51:09 so i have a keybind to restart the networking 2021-11-08 23:00:44 ok, I have networkmanager installed now, and iwd, nm is configured to use iwd, but in the menu it's just saying "device not managed" and not letting me change network things 2021-11-08 23:00:59 the network is connected but seems to be outside of nm, how do I make it inside of nm 2021-11-08 23:01:26 maybe restart a few services? 2021-11-08 23:01:52 I'm not sure what services are involved in network stuff, like what is it that brings the networks up at boot 2021-11-08 23:01:55 on the default install 2021-11-08 23:02:11 `networking` 2021-11-08 23:06:23 ? 2021-11-08 23:06:29 like, it's just called networking 2021-11-08 23:06:42 ACTION is sorry to disappear and reappear, the alpine machine is in another room 2021-11-08 23:07:41 yes 2021-11-08 23:07:43 it is networking 2021-11-08 23:07:53 ok 2021-11-08 23:08:13 so should I restart that? disable that? or do what with it, if I'm using nm 2021-11-08 23:08:58 i think that iwd and nm might've gotten confused about coexisting, so restarting it should restart both of them in the proper order 2021-11-08 23:09:07 if you started them in a different order 2021-11-08 23:12:14 restarting it and nm still says "device not managed" in the applet 2021-11-08 23:12:23 hmm 2021-11-08 23:14:04 I might enable iwd and networkmanager and reboot 2021-11-08 23:14:05 see what happens 2021-11-08 23:21:23 tried that, same thing. I'm connected to wifi but through the connection that happens at boot, networkmanager says device not managed 2021-11-08 23:30:56 ah, I had to remove them from my /etc/network/interfaces for networkmanager to see them 2021-11-09 00:21:36 @MY-R i think so too.. but im very new to alpine 2021-11-09 00:22:32 Oh.. well.. 2021-11-09 07:26:31 Hi, how do I use bootchartd? I tried pressing 'e' in the GRUB menu and adding chart=yes to the cmdline, but it doesn't generate a /var/log/bootchartd.tgz. In dmesg I do see 'starting bootchartd' (or something similar I can't remember). I also notice there isn't any /etc/bootchartd.conf file. 2021-11-09 09:40:14 what font-pkg is needed for all emojis the kids use these days? 2021-11-09 09:40:52 font-noto-emoji 2021-11-09 09:41:34 though nerd-fonts would be an appropriate name 2021-11-09 09:42:45 hmm i think i installed that and it "didnt work" so i uninnsalled it, should work out of the box i guess (even on st terminal) 2021-11-09 09:43:12 You probably need to set some fontconfig 2021-11-09 09:44:01 oh i see 2021-11-09 12:39:21 hmm, "convert in.jpg -rotat 180 out.jpg", gives me "convert: unrecognized option `-rotat' @ error/convert.c/ConvertImageCommand/2682." 2021-11-09 12:40:10 "-rotate"? 2021-11-09 12:40:55 yes, to rotate the image (180 degrees) 2021-11-09 12:41:20 its defined in the manpage for convert 2021-11-09 12:41:42 (and ive used it for a long time, not 100% sure ive used it on alpine tho) 2021-11-09 12:41:57 from imagemagick btw 2021-11-09 12:42:06 I think what lopid means is you put -rotat instead of -rotate 2021-11-09 12:43:32 oh :$ 2021-11-09 13:26:45 good afternoon 2021-11-09 13:42:03 hmz 2021-11-09 18:56:56 i guess there are a few norwegians here ... got a email from my bank, it used 0 instead of ø and a instead of å, didnt see any æ, is this some kind of standard? 2021-11-09 18:57:00 kinda cool imo 2021-11-09 18:57:50 ø is usually oe 2021-11-09 18:57:54 never seen it as 0 2021-11-09 18:58:40 yeah i thought it was mblaze that did it, might be i guess... 2021-11-09 18:59:09 i checked my webmail and its the same there, but might have changed and synced up i guess.. 2021-11-09 18:59:12 what, mblaze can't just do utf-8? :^) 2021-11-09 19:00:16 i would guess the bank sent it wrong 2021-11-09 19:00:18 usually the case 2021-11-09 19:00:52 > mblaze is written in portable C, using only POSIX functions (apart from a tiny Linux-only optimization), and has no external dependencies. It sup‐ ports MIME and more than 7-bit messages (everything the host iconv(3) can decode). It assumes you work in a UTF-8 environment. mblaze works well with other Unix utilities such as mairix(1), mu(1), or offlineimap(1). 2021-11-09 19:01:08 yep 2021-11-09 19:01:52 i seem to remember 7bit and i guess i thought 7bit ascii ... 2021-11-09 19:05:11 in my experience it is always the sender that does weird conversions 2021-11-09 19:06:02 and while oe is most correct just 'o' is the most common i've had to experience shipping things, as well as å->a 2021-11-09 19:06:14 thankfully the postal services don't really care and figure it out 2021-11-09 19:07:02 i think most web searches dont differentiate between those also 2021-11-09 19:07:09 yep 2021-11-09 19:21:47 hi, I have set up networkmanager on my alpine install but I'm having an issue with wireguard. I have a wireguard config file that works when I use it with `wg-quick tablet` but if I use it in networkmanger (via `nmcli c import type wireguard file /etc/wireguard/tablet.conf`) then I can't view anything on the internet while it's active 2021-11-09 19:23:15 and I'd like to use it with networkmanager 2021-11-09 20:51:06 also apk seems to have just disappeared, it's not in /bin and not in /usr/bin and I am very confused 2021-11-09 20:51:39 apk lives in /sbin 2021-11-09 20:51:58 OK it's there 2021-11-09 20:51:59 phew 2021-11-09 20:52:03 I think this is just a PATH problem then 2021-11-09 20:53:14 I installed fish and it doesn't seem to have /sbin in its path 2021-11-09 20:53:55 how come it's in the path usually but not when I have fish as my login shell? odd 2021-11-09 20:55:12 set by .profile likely 2021-11-09 20:55:15 or well 2021-11-09 20:55:19 etc/profile 2021-11-09 20:55:55 I thougt /etc/profile worked whatever shell you are using 2021-11-09 20:55:56 hmm 2021-11-09 20:56:07 yes 2021-11-09 20:56:10 you have to use fish -l 2021-11-09 20:56:20 with the exception of stinky shells 2021-11-09 20:56:23 like fish 2021-11-09 20:56:33 does fish have a bash parser yet 2021-11-09 20:56:36 so I set my login shell to fish -l rather than just fish? 2021-11-09 20:56:37 it's the same in every other shell 2021-11-09 20:56:44 without -l it won't source /etc/profile 2021-11-09 20:56:46 with chsh 2021-11-09 20:57:01 fish cant source /etc/profile 2021-11-09 20:57:25 bcos profile is meant to be vaguely posix compatible 2021-11-09 20:57:46 I like fish because it has nice completions and shit without me putting in effort and idk a posixy shell that also has that 2021-11-09 20:57:55 zsh 2021-11-09 20:57:57 unless anybody has recommendations, because I am lazy and need my tab suggestions 2021-11-09 20:58:16 I used zsh before fish but fish's completions are wayy better 2021-11-09 20:58:25 or maybe I just had a bad zsh config idk. I was using grml zsh 2021-11-09 20:58:51 fush completions are probably better 2021-11-09 20:59:22 im like 200% sure passwd supplies -l to the user's shell 2021-11-09 20:59:55 bcos on unix.lgbt only fish users have path issues, remedied in /etc/profile 2021-11-09 21:00:07 ah ok 2021-11-09 21:00:09 and none of our scripts add the -l flag 2021-11-09 21:00:13 so fish is being dumb ig 2021-11-09 21:00:26 ah, indeed it is 2021-11-09 21:00:35 fish chose to be not posix compatible and do literally nothing with that 2021-11-09 21:00:37 might be time to move back to zsh 2021-11-09 21:00:45 because it's actually like, compatible with shit 2021-11-09 21:01:06 if you put effort in you can make shit compatible 2021-11-09 21:01:35 i toyed w running rc for a little but i missed `~` for `$HOME` 2021-11-09 21:01:40 I don't want to put effort in 2021-11-09 21:02:00 then why are you running alpine? :p 2021-11-09 21:02:11 well, I don't want to put effort into making shit things good 2021-11-09 21:02:14 I want to use good things 2021-11-09 21:02:19 that's why I'm running alpine 2021-11-09 21:02:58 what makes fish shit? 2021-11-09 21:03:53 if it doesn't load my path 2021-11-09 21:04:08 although if I can just add the -l flag it's fine 2021-11-09 21:04:10 I guess 2021-11-09 21:04:26 I like fish as a user interface 2021-11-09 21:04:31 i forgot i had a `bass . /etc/profile` in config.fish for this reason :p 2021-11-09 21:04:52 that being a random python script from https://github.com/edc/bass/tree/master/functions 2021-11-09 21:05:00 nihilazo: as stated, -l doesnt fix it 2021-11-09 21:05:00 but yes, it doesn't read profile at all even with -l, bad memory 2021-11-09 21:05:22 oh 2021-11-09 21:05:25 that's uhh 2021-11-09 21:05:30 annoying 2021-11-09 21:06:18 I don't want to have to use some python script to load profile, nor do I want to have to deal with doing my path more myself 2021-11-09 21:06:35 so I guess I'm switching back to zsh(?) 2021-11-09 21:06:41 I mean, if zsh sucks I can go back to fish 2021-11-09 21:11:48 ok, zsh actually honors profile. Good to know 2021-11-09 21:55:13 hello, suppose the time of an Alpine instance inside a VM (usually because of a freeze / pauze or something like that) runs out of sync. What is the best method to resync it? I know in Ubuntu I have to do `systemctl restart systemd-timesyncd`, but I'm unsure how to best approach this with Alpine. 2021-11-09 21:55:35 depends on the ntp client that you use 2021-11-09 21:55:46 not sure I use any :p 2021-11-09 21:55:49 With chrony, you can run chronyc and execute makestep 2021-11-09 21:55:52 it's just a default install 2021-11-09 21:55:59 The default asks you to select one 2021-11-09 21:56:04 the installer* 2021-11-09 21:56:17 oh, hmm... 2021-11-09 21:56:26 rc-status 2021-11-09 21:56:46 crond 2021-11-09 21:57:00 that's a task scheduler 2021-11-09 21:58:42 this is the full output: https://paste.debian.net/plain/1218852 2021-11-09 21:59:18 hmm, no ntp related service there 2021-11-09 21:59:31 `find /etc/periodic -type f` 2021-11-09 22:00:27 returns nothing 2021-11-09 22:00:38 I'm using alpine-virt-3.14.0-x86_64.iso ... maybe that's relevant? 2021-11-09 22:01:56 easiest would be to just do: apk add chrony; rc-update add chronyd; rc-service chronyd start 2021-11-09 23:17:58 If I've noticed a ton of python3/sphinx packages are backlevel, is the preffered thing to mark them out of date on the pkgs.alpinelinux.org site so the Maintainer gets emailed, or should I open MRs for each one that I can update? 2021-11-10 00:09:15 Am I confused? Got the openjdk version 12 to 17 packages being removed from the 3.14 community repository lately? 2021-11-10 00:32:22 Maybe they were in the testing repo, but now it seems they are available in the edge repo only https://pkgs.alpinelinux.org/packages?name=openjdk1*jre*headless*&branch=edge&repo=community 2021-11-10 01:44:10 is it possible to get mandb and docs (which depends on mandoc) installed at the same time? 2021-11-10 02:13:45 AlpineForFun: considering that openjdk 17 was released after alpine 3.14, it seems unlikely that it was in 3.14 community 2021-11-10 02:15:13 asdflkj: hm, it should possibly depend on cmd:man. not sure how that works with upgrades and priority though 2021-11-10 02:50:36 well, openjdk 17 is not in alpine v3.14 yet 2021-11-10 03:03:59 ... 2021-11-10 03:04:31 ikke: setup-ntp is skipped for qemu 2021-11-10 03:23:03 Can I pin a locally-compiled untrusted package so that apk upgrade -al won't replace it with repo packages and I can get other packages back on track? 2021-11-10 08:25:38 Hi, were there recently changes regarding the naming of python packages? apparently one of our builds does not find "python" any more but specfying "python3" works fine. somebody else seems to have the same problem but that fix doesn't work for them https://stackoverflow.com/questions/69884391/docker-alpine-unable-to-select-packages-python-no-such-package-while-buildin 2021-11-10 08:26:09 herndlm: not anything recent 2021-11-10 08:27:44 strange, is there a concept of package aliases or something like that? was the "python" alias dropped maybe? I'm pretty sure our build wants python3, so I now how to fix it. But wondering why this happened 2021-11-10 08:28:33 python2 provides /usr/bin/python still 2021-11-10 08:29:11 Except for a short period a couple of years ago, you always had to explicitly refer to python3 2021-11-10 08:31:33 ah ok thx, than the stackoverflow person might need python2 2021-11-10 08:32:43 in a Dockerfile we have "apk add --no-cache --virtual python" and I wonder how this was working until recently 2021-11-10 08:33:49 without the --virtual basically, bad copy/paste of a multiline command.. 2021-11-10 08:37:59 Hello everyone. Is there anyone who knows a few things about gnome desktop on alpine who can help me out? 2021-11-10 08:43:39 hi Gaming4LifeDE 2021-11-10 08:43:53 it's basically apk add gnome and enabling dbus, gdm :) 2021-11-10 08:45:11 markand_: Well, kind of... Gnome software is bugged. I wanted to try playing around with gnome-initial-setup but it breaks when trying to enter anything into the "Full Name" field when creating a new user. gnome-control-center segfaults at the exact same point too btw 2021-11-10 08:46:25 markand_ and gnome-software cannot update the system "access for operation UpdatePackage denied" 2021-11-10 08:53:42 ACTION didn't even know we had gnome-software  2021-11-10 08:54:05 gnome-software is probably the most bugged GNOME application ever made 2021-11-10 08:54:22 lol idk but it received a huge revamp with gnome 41 2021-11-10 08:54:25 but, packagekit is as well so you have two layers of broken stuff 2021-11-10 08:54:39 isn't it appstream now? 2021-11-10 08:55:28 I think appstream is only there to describe available software, not how to install (which was why packagekit was made) AFAIK 2021-11-10 08:59:06 i don't really know. i just want it to work (mostly for flatpak but i want to do system updates...) 2021-11-10 08:59:53 markand do you think you can help out? 2021-11-10 09:00:12 I'd just say to not use gnome-software 2021-11-10 09:00:26 that's not a solution though, just a workaround 2021-11-10 09:00:35 on alpine I think only flatpak is supported with it (it explicitly disable packagekit as there are no apk packagekit backend) 2021-11-10 09:00:42 if it's in the repo, it should at least work 2021-11-10 09:00:53 there is though... 2021-11-10 09:01:31 https://pkgs.alpinelinux.org/package/edge/community/x86_64/gnome-software-plugin-apk 2021-11-10 09:04:01 ah yes 2021-11-10 09:28:42 markand do you know anything about the gnome packages or are you doing something else or just a community member? I'm just trying to find someone who can help me really :/ 2021-11-10 09:30:47 I've used GNOME on alpine yes, but never gnome-software 2021-11-10 09:34:28 markand have you ever tried gnome-initial-setup? also, does gnome-control-center segfault when trying to add a new user? 2021-11-10 13:19:26 Hey, there is one little thing what prevents me to sleep well ;/ maybe two now ... 2021-11-10 13:20:23 #medical 2021-11-10 13:21:03 first is regarind wlroots windows mangers - sway and now river too, it is build with eudev, so eudev-libs is like runtime dependency too, however it does not start with it, it requires libudev-zero, after installing and replacing eudev-libs it start well 2021-11-10 13:21:07 lopid: ;/ 2021-11-10 13:21:55 second thing is, after last update of mkinitfs - edge, I adds like 5 seconds extra boot time, before openrc even start - blank screen, without 'quiet' no error 2021-11-10 13:22:51 also one extra think regarding wiki - efi stub boot 2021-11-10 13:23:11 the last change, there was removed loading modules https://wiki.alpinelinux.org/w/index.php?title=Bootloaders&type=revision&diff=20230&oldid=19577 2021-11-10 13:23:38 which fail to boot, bc. it can't boot '/' 2021-11-10 13:24:27 after dropped to shell, if manualy 'mount -t ext4 /dev/sda3 to sysroot and continue boot, it boots 2021-11-10 13:24:49 so tell me, how can I sleep well now 2021-11-10 13:25:23 in all cases I am on edge 2021-11-10 13:27:13 so regarding build wlroots/sway/river with eudev, after build/install manually repalce it with libudev-zero is it something I miss or is it well know 'bug' 2021-11-10 13:28:22 it works for me with eudev 2021-11-10 13:28:47 you have to remember to add the udev-* services yourself though 2021-11-10 13:34:59 like run setup-udev or whatever is it? btw. I am using mdev 2021-11-10 13:35:03 so that can be an issue? 2021-11-10 13:35:27 Gaming4LifeDE: are you using a user with ID 1000 or more and in wheel group? 2021-11-10 13:35:27 I had no trouble installing and updating like that 2021-11-10 13:35:31 if you are not using eudev then yes you want libudev-zero 2021-11-10 13:35:38 don't see where the bug is 2021-11-10 13:35:39 you should use libudev-zero thebmn 2021-11-10 13:35:48 I also had to run it as root at least once, I think to have it make a cache directory 2021-11-10 13:36:08 mdevd is even better then mdev 2021-11-10 13:38:13 oh now I see it in wiki for sway to run setup-udev, okay 2021-11-10 13:38:39 it's still weird, that it does not need it for running as mdev is enough 2021-11-10 13:40:02 eudev is deprecated 2021-11-10 13:40:24 alandiwix: Not really. iirc, Ariadne took over maintenance 2021-11-10 13:41:55 well, mdev works well for me, just have to manually install libudev-zero for sway/river 2021-11-10 13:42:28 ikke: heard void guys were going to maintain it themselves also 2021-11-10 13:42:28 I was just not sure what I fucked up, that it always complains 2021-11-10 13:49:09 Saijin_Naib[m] yes and no. gnome-initial-setup/gdm is suppose to create a user "gnome-initial-setup" for first run 2021-11-10 13:50:30 Ah, I never went through that flow since the wiki cautions that GDM won't start without an existing user 1000 2021-11-10 13:50:33 so now two things left, efi stub boot / wiki and mkinitfs boot time ;/ 2021-11-10 13:51:56 But once I was in and my user had some extra groups to it, gnome-software-center worked happily (and I sometimes still use it) 2021-11-10 13:54:36 Well, abuild-keygen is using `sudo`. For doas user, You may need to link `/usr/bin/doas` to `/usr/bin/sudo`. 2021-11-10 13:55:08 SUDO=doas works as well 2021-11-10 13:56:15 In case if some program really need the real sudo 2021-11-10 13:56:26 Well, At that case, i would like to use `SUDO="su -c"` 2021-11-10 13:59:16 is there a reason why is it not changed to doas in that script? 2021-11-10 14:03:39 *shrug* 2021-11-10 14:04:02 su -c should be a better option 2021-11-10 14:04:25 However, su -c only work if root account is not disabled 2021-11-10 14:04:33 Saijin_Naib[m] what groups did you need to put your user in for gnome-software? 2021-11-10 14:04:34 So.......... How can we deal with it? 2021-11-10 14:06:09 oho, you are too far my friend, first wee need to boot ;/ 2021-11-10 14:06:41 first wee is best wee 2021-11-10 14:07:47 Weeeeeeeeeeeeeeeeeeeeeee 2021-11-10 14:09:11 ACTION installing dpkg package in alpine 2021-11-10 14:09:44 ACTION got architecture error due to unmatching `amd64` with current host `musl-linux-amd64` 2021-11-10 14:10:12 Yonle what are you trying to do?+# 2021-11-10 14:10:46 or maybe the issue is, I checked wiki after long time 2021-11-10 14:10:49 Installing some deb package like normal 2021-11-10 14:10:54 you can't do that 2021-11-10 14:10:56 it's probably not usefull to go there 2021-11-10 14:11:04 Yonle but why? 2021-11-10 14:11:18 Experiment :^) 2021-11-10 14:11:19 you can make a debian chroot with debootstrap, but you can't use any debian packages in alpine 2021-11-10 14:11:42 So dpkg is only for debootstrap? 2021-11-10 14:11:51 there is no dpkg 2021-11-10 14:11:56 psykose would be happy enough just for a desktop to work properly :P 2021-11-10 14:12:03 "Saijin_Naib what groups did..." <- I'm in a lot of groups right now, and I can't remember which one(s) I had added only to get gnome-software-center working, but I believe it was likely wheel that was needed 2021-11-10 14:12:18 psykose, Then why it's on apk repo? 2021-11-10 14:12:41 Saijin_Naib[m] just call the groups command, it should tell you all groups you're in 2021-11-10 14:13:10 ah, that's what you mean 2021-11-10 14:13:48 yes, it's only for bootstrapping 2021-11-10 14:13:49 Ah yet, Time to compile dpkg (not apt) from skratch(scratch) 2021-11-10 14:13:55 debian doesn't build anything for musl to my knowledge 2021-11-10 14:14:14 psykose, Isn't musl has glibc compatibility? 2021-11-10 14:14:33 That compatibility is available when you install `gcompat` 2021-11-10 14:14:59 if you want to install debian packages just make a debian chroot 2021-11-10 14:15:53 hm, is Pi Zero support decent enough? I have only run on a Pi 3 2021-11-10 14:17:43 ...still hoping someone knows they're stuff about gnome, maybe the gnome maintainer or so? not sure if anyone else can really help... Kinda getting hopeless by now :/ 2021-11-10 14:18:18 The person maintaining gnome for alpine is a bit busy at the moment 2021-11-10 14:18:44 ikke do you know when the maintainer will have time for stuff like that? 2021-11-10 14:19:15 nope 2021-11-10 14:19:42 too bad. I also can't really help myself much anymore. Can't really deal with segfaults 2021-11-10 14:20:01 Gaming4LifeDE: could you open an issue on gitlab.a.o? 2021-11-10 14:20:15 ... or the nuances of gnome-software 2021-11-10 14:20:28 there was never a better time to embrace tiling window management 2021-11-10 14:20:47 there is a gnome-software-dbg if you want to try and get a backtrace 2021-11-10 14:21:00 i already flagged gnome-software out of date at least and for gdm there's an issue already #12413 2021-11-10 14:21:43 And for gnome-control-center i don't really know how i can provide more info. Do you know how i can get logs for a segfault? maybe some debug package/command or something? 2021-11-10 14:22:43 that issue is not about segfaults 2021-11-10 14:22:52 you can install the -dbg for it and start it in gdb and have it segfault there 2021-11-10 14:24:51 ikke i know, gnome-control-center segfaults when i try to type or paste into the full name field 2021-11-10 14:25:53 how do i start a program with debug stuff? 2021-11-10 14:30:30 this is all i got https://www.toptal.com/developers/hastebin/ufawupureq.yaml 2021-11-10 14:34:53 Gaming4LifeDE: do you know how to get a backtrace with gdb? 2021-11-10 14:35:16 no 2021-11-10 14:35:47 apk add gnome-control-center-dbg 2021-11-10 14:35:53 gdb gnome-control-center 2021-11-10 14:35:57 r 2021-11-10 14:36:00 (reproduce crash) 2021-11-10 14:36:02 bt 2021-11-10 14:38:19 Gaming4LifeDE:... (full message at https://matrix.org/_matrix/media/r0/download/matrix.org/gSWCNcpvkFkzJjRqTDNFqhEn) 2021-11-10 14:39:10 Saijin_Naib[m] those groups don't really seem related 2021-11-10 14:39:35 And I was happily using GNOME 40 for the better part of a year on Alpine as a DE, so don't give up hope. It is eminently possible and I believe the maintainer will get it all sorted 2021-11-10 14:39:46 Like I said, I think wheel is what I needed specifically for gnome-software-center 2021-11-10 14:40:20 i vers much hope so. it would be cool to just apk add gnome and it'll handle everything else, including running gnome-initial-setup if you didn't create a user 2021-11-10 14:40:41 What maintainer? :^) 2021-11-10 14:40:43 i'm calling my sudo group sudo but my user is in sudoers 2021-11-10 14:43:57 Newbyte... You, I thought haha. I can certainly be mistaken! 2021-11-10 14:46:23 The thing is that GNOME doesn't really have a maintainer in Alpine at the moment since Thomsen can't use Alpine at the moment for university reasons, so maintenance is not really guaranteed 2021-11-10 14:46:41 Unfortunately 2021-11-10 14:47:39 Newbyte are you ready to maintain? :D I'd want to but I'm not that good with alpine yet 2021-11-10 14:48:31 Ahh, that's unfortunate 2021-11-10 14:48:54 It does look like quite a huge piece of interwoven stuff to maintain... but I guess any large DE stack would be (KDE I assume is also hard) 2021-11-10 14:49:07 is anyone here good with maintaining a software stack? 2021-11-10 14:49:24 I can't really promise I'll be able to devout enough time to do it, especially since I don't use Alpine on my main desktop, so no 2021-11-10 14:49:33 I'm sure plenty of folks are, but likely already are at or beyond their limits 2021-11-10 14:49:37 Saijin_Naib[m] the Solus team took a long time to get kde working properly so yeah 2021-11-10 14:49:52 probably 2021-11-10 14:50:03 afaik kde is far better maintained in alpine 2021-11-10 14:50:34 Yeah but it doesn't have a very good bus factor either 2021-11-10 14:51:06 that goes for just about everything 2021-11-10 14:51:16 That is true 2021-11-10 14:51:24 Where;s that XKCD comic... 2021-11-10 14:51:45 https://xkcd.com/2347/ 2021-11-10 14:51:45 Dependency | Alt-text: Someday ImageMagick will finally break for good and we'll have a long period of scrambling as we try to reassemble civilization from the rubble. 2021-11-10 14:52:13 nebraska is not the most fun place 2021-11-10 14:52:28 I guess that is why so much maintenance happens from there? 2021-11-10 15:18:48 Newbyte i got a trace but i don't find anything weird in there: https://www.toptal.com/developers/hastebin/genukupimi.sql 2021-11-10 15:25:20 weird.... gnome-initial-setup has neither debug symbols nor does a dbg package exist 2021-11-10 15:30:15 well, it's a one-time script 2021-11-10 15:30:30 still segfaults though... 2021-11-10 15:31:44 also that bt looks normal to me, something in cc-add-user-dialog calls strlen wrong 2021-11-10 15:32:08 psykose not much i can do about that i think 2021-11-10 15:32:16 but where should i open an issue for this 2021-11-10 15:32:25 aports issues 2021-11-10 15:32:49 sounds a lot like a gnome issue though... on the other hand it works fine on other distros 2021-11-10 15:34:26 (gnome-initial-setup:3300): GLib-CRITICAL **: 16:33:25.550: g_ascii_strdown: assertion 'str != NULL' failed 2021-11-10 15:34:26 Segmentation fault 2021-11-10 15:34:39 same issue with gnome-initial-setup and gnome-control-center 2021-11-10 15:34:50 can anyone reproduce the issue with gnome-control-center? 2021-11-10 15:37:22 can't, as you are the only user of it ;/ 2021-11-10 15:37:45 i bet theres anyone here using gnome :P :D 2021-11-10 15:38:02 yes maybe some archlinux user here ;/ 2021-11-10 15:38:08 lol 2021-11-10 15:40:18 Now i hope you didn't say that "I use arch btw" thingy 2021-11-10 15:40:36 "i use alpine btw" lol 2021-11-10 15:40:39 hah I am 100% alpine 2021-11-10 15:40:50 I am 100% Alpine, too 2021-11-10 15:40:56 Using XFCE :^) 2021-11-10 15:41:01 not many use it on main machine for whatever reason, that makes a lot of undiscovered bugs 2021-11-10 15:41:07 i'm not. i'm running a mix of everything 2021-11-10 15:42:19 that does not count, you are second class citizen ;/ 2021-11-10 15:42:38 that's absolutely no issue for me 2021-11-10 15:42:39 My laptop is Alpine Edge with XFCE now, but was GNOME until GDM broke a while back 2021-11-10 15:42:45 Tha'ts a daily 2021-11-10 15:43:08 Gaming4LifeDE: debug symbol packages need to manually be added on a per-package basis 2021-11-10 15:43:21 How come you didn't switch back to GNOME btw? 2021-11-10 15:43:26 can't it run without gdm with something else? or from tty? 2021-11-10 15:43:42 gdm works now anyway 2021-11-10 15:43:46 can anyone of you dump alpine with gnome into a vm and try to reproduce my issue? I'll need to look for a solution if i'm the only one 2021-11-10 15:43:59 and gnome-session itself had issues 2021-11-10 15:44:07 Gaming4LifeDE: what's your issue again? 2021-11-10 15:44:11 Newbyte there's no dbg package for gnome-initial-setup 2021-11-10 15:44:11 who will voluntary polute system with gnome things? 2021-11-10 15:44:34 Newbyte: GNOME 40 DE was about 0.3 - 0.4 load completely passive, so it was killing my battery life when untethered. XFCE is like 0.1 tops 2021-11-10 15:44:38 helby: me 2021-11-10 15:44:39 Newbyte gnome-control-center and gnome-initial-setup both crash when you try to type into the full name field 2021-11-10 15:44:55 helby me too 2021-11-10 15:44:58 what full name field? could you send a screenshot? 2021-11-10 15:45:14 helby: Me, as well. especially on desktop 2021-11-10 15:45:33 Despite my issues with some of their changes, GNOME's HIG does honestly make for pretty nice to use software 2021-11-10 15:46:03 you are [m] it does not count ;/ 2021-11-10 15:46:12 I don't understand 2021-11-10 15:46:43 don't take me seriously, I don't understand myself too 2021-11-10 15:46:48 ah yes, more discrimination because I use Matrix 2021-11-10 15:46:54 Oh... Okay 2021-11-10 15:47:15 Uploaded file: https://uploads.kiwiirc.com/files/08f86ef2bebd0e641bafd8756af09148/Screenshot%202021-11-10%20164652.png 2021-11-10 15:47:17 A religious thing, haha. Yeah, sorry, I missed IRC since I didn't have a computer or internet back then 🙂 2021-11-10 15:47:17 just trying matrix, bc. signed for shell account and they use it 2021-11-10 15:47:19 omg 2021-11-10 15:48:13 Saijin_Naib[m]: back then? It's never gone away ;) 2021-11-10 15:48:14 I used to use IRC, but it's just so much more convenient to use Matrix than having a bouncer set up and all that 2021-11-10 15:48:28 dude just had to upgrade RAM to 128GB bc. of matrix ;/ 2021-11-10 15:48:56 That's not to say I think Matrix is amazing … but more bearable than IRC. Either way, off-topic 2021-11-10 15:49:07 yes without a bouncer or constatly running client it's not so fun 2021-11-10 15:49:24 nickersonm: Indeed not, but I wasn't steeped in it and the culture so it is just a 🤷 from me. About the only hill I'll die on is that I miss multi-protocol chat clients and I hate having forty different chat apps that are all proprietary 2021-11-10 15:49:37 thelounge is an easy fix if you want irc. you need a server for that though 2021-11-10 15:53:01 i can reproduce the segfault myself i guess 2021-11-10 15:53:14 also funnily i installed gnome and it worked first try after 15 seconds :p 2021-11-10 15:53:16 Gaming4LifeDE: Is your IRC Client OK? Your message is sent twice 2021-11-10 15:53:20 let me see if i can find it in the code 2021-11-10 15:53:28 which one? 2021-11-10 15:53:35 Yonle: the messages are not sent twice 2021-11-10 15:53:51 Saijin_Naib[m]: yeah, annoying to have 4 separate clients these days 2021-11-10 15:54:11 Well then my client is dwonk 2021-11-10 15:54:20 Let me reconnect... Oh yeah bois here we go 2021-11-10 15:54:30 hah 2021-11-10 15:54:36 nickersonm: Some folks only use Facetime/iChat, so it's like... I guess we'll never speak again 🤣 2021-11-10 15:54:45 HURRAH 2021-11-10 15:55:06 psykose another thing is that i haven't figured out what's missing so the unlocking in gnome-control-center would work. it doesn't quite yet 2021-11-10 15:58:35 well probably good starting point would be to check arch wiki and all their dependencies for gnome, to check out what u maybe need 2021-11-10 15:59:25 wau, if I try apk add -i gnome, it shows ~ 1GB of fancy things ;/ 2021-11-10 15:59:42 not big screen enough for everything displayed ;/ 2021-11-10 15:59:54 Saijin_Naib[m]: good choice 2021-11-10 16:00:25 what can go wrong with 1000 packages ;/ 2021-11-10 16:00:36 Well, gnome-apps-core 2021-11-10 16:00:40 Gaming4LifeDE: unlocking the usersettings works fine for me 2021-11-10 16:00:45 By the way, gnome-apps-extras is broken 2021-11-10 16:01:59 Yonle well i didn't want gnome-core, just gnome. gnome-core ships a lot of unnecessary stuff 2021-11-10 16:02:21 Ok 2021-11-10 16:09:28 think i found the issue 2021-11-10 16:10:52 g_convert_with_fallback in generate_username_choices() returns null, the actual `name` is what you type into the box (the first character you put) 2021-11-10 16:11:06 as for why it fails no idea, but the issue should be there 2021-11-10 16:11:18 so, as the debug output says in the error 2021-11-10 16:20:02 psykose where do you think i can report that? gnome or on aports? 2021-11-10 16:22:37 fucking captcha on wiki 2021-11-10 16:22:54 can't register that crap is annoying ;/ 2021-11-10 16:23:22 Given the number five hundred eighteen million three hundred twenty-four thousand two hundred fifty-seven, what is the ninth digit? 2021-11-10 16:24:05 what is the answer? ;/ 2021-11-10 16:24:47 from left or right, good question 2021-11-10 16:24:50 Gaming4LifeDE: i would make an aports issue, with the backtrace you have 2021-11-10 16:25:01 ok psykose, thanks 2021-11-10 16:25:06 you can add gtk+-dbg or whatever to fill in the other blanks i think, but it doesn't matter 2021-11-10 16:25:12 and the issue is in the function call i mentioned 2021-11-10 16:25:13 Gaming4LifeDE: I think at this point it's easier to get used to sway/KDE/etc. than fix gnome 2021-11-10 16:25:37 as we all know the way things get fixed is by never fixing anything and moving to something else 2021-11-10 16:25:51 oh wait, that's not how it works :p 2021-11-10 16:26:17 I agree, it would be nice to have gnome in the long term 2021-11-10 16:27:20 also it works, this is just the fullname field in one place 2021-11-10 16:27:20 alandiwix and i'm pushing it now 2021-11-10 16:27:28 it fails for the same reason in both places probably 2021-11-10 16:29:13 is it reproducable though? 2021-11-10 16:29:20 can someone try? 2021-11-10 16:29:29 i'd open an issue if i know howto reproduce 2021-11-10 16:32:31 okay, no way I can register in alpine wiki 2021-11-10 16:32:47 that shit is like gnome 2021-11-10 16:32:58 maybe even worse, bc. it does not work 2021-11-10 16:33:40 Gaming4LifeDE: i already did 2021-11-10 16:33:50 reproduce it * 2021-11-10 16:37:03 psykose can you give instructions on how to reproduce it with minimal effort? I'd include that in the report 2021-11-10 16:40:37 okay so last update of mkinitfs fixed that boot delay 2021-11-10 16:41:09 Yeah, ncopa fixed the issue 2021-11-10 16:43:59 yes, just saw that commit 2021-11-10 16:46:30 hm 2021-11-10 16:48:02 hm 2021-11-10 16:51:29 there is still one extra issue regarding foot terminal 2021-11-10 16:52:17 that tiny shit bc. of ncurses-terminfo went from 500KiB to 8000KiB it's heavier now than alacritty 2021-11-10 16:53:10 ncurses-terminfo-base is 216kb 2021-11-10 16:53:40 alacritty depends on that 2021-11-10 16:53:46 foot depends on ncurses-terminfo 2021-11-10 16:54:21 yes 2021-11-10 16:54:36 and that's 7000Kb 2021-11-10 16:54:59 for 400kb terminal it's somehow a lot 2021-11-10 16:57:51 also there is not much packages that requires it, so the chance that user already need/have it is almost zero 2021-11-10 17:04:08 foot is wayland only terminal and there is no any wayland package that even needs it for wayland 2021-11-10 17:08:15 and foot provides a way to install 'own' foot terminfo without ncurses nonsense 2021-11-10 17:13:43 ikke: thanks for the chronyc recommendation. As a test I paused the VM before taking a nap and then unpaused it. After skimming the manpage I thought `chronyc makestep` (as root) would sync the time, but that didn't work, so I just had to `rc-service chronyd restart` to get it synced. 2021-11-10 17:14:08 Guest3462: I would expect makestep to synchronize the time 2021-11-10 17:15:23 Guest3462: for hypervisor sync it depends on which hypervisor you are using, e.g. I think Virtualbox provides stuff to sync with host. For KVM/QEMU you can use chrony with PHC to sync 2021-11-10 17:15:59 it _did_ give me a "200 OK" message, but didn't end up syncing it for some reason. I'll try and test it out later. I'm using virt-manager with KVM/QEMU. 2021-11-10 17:16:21 PHC with chrony will be more accurate than using NTP 2021-11-10 17:18:18 Guest3462: more info here: https://access.redhat.com/documentation/en-us/red_hat_enterprise_linux/7/html/virtualization_deployment_and_administration_guide/chap-kvm_guest_timing_management 2021-11-10 17:19:07 thanks, I'm somewhat familiar with NTP but never heard of PHC before, so I'll take a look at that. 2021-11-10 17:43:59 hmm, seems that chronyd doesn't like to have `refclock PHC /dev/ptp0 poll 2` in `/etc/chrony/chrony.conf` 2021-11-10 17:47:59 Gaming4LifeDE: you just run it, go to users, unlock, add user, type anything into full name 2021-11-10 18:02:54 Guest3462: let me know how you get on with PHC as am intended to incorporate it into my own scripts 2021-11-10 18:09:36 helby: ever heard of wayst? 2021-11-10 18:10:21 minimal: no more rng warnings with HZ=1000 on rpi4 :P because "random: crng init done" is done before eudev :) 2021-11-10 18:18:10 alandiwix: never ;/ 2021-11-10 18:18:15 seems to be nice opengl-accd terminal, but you have to compile it yourself 2021-11-10 18:19:31 well, in that case I would abuild with changed apkbuild foot ;/ 2021-11-10 18:19:47 but I m not much fan of building, even I do it sometime 2021-11-10 18:20:45 alandiwix: do u have apkbuild for it? 2021-11-10 18:21:40 anyway foot is great, I did not have one issue with it for whole time I use it and it's tiny compard to other choices, even not many for wayland in alpine 2021-11-10 18:21:59 but that ncurses crap is killing my during nights ;/ 2021-11-10 18:24:48 could docs be made to depend on cmd:man? it being uninstallable with mandb is quite irritating, especially since mandoc seems to be missing apropos and whatis (no idea why, both are present with mandoc on OpenBSD) 2021-11-10 18:29:17 there is mandoc-apropos 2021-11-10 18:33:27 ok, seems that I had to add the module and then reboot. 2021-11-10 18:40:02 asdflkj: you should make an issue for it 2021-11-10 18:40:25 i also agree, i would like to have man-db with docs for auto-docs on everything 2021-11-10 18:51:25 MY-R: which kernel version package/version is that with? 2021-11-10 18:51:58 Guest3462: kvm-clock module? 2021-11-10 18:57:08 minimal: /etc/modules-load.d/ptp_kvm.conf 2021-11-10 18:57:28 it had to have ptp_kvm in it 2021-11-10 18:59:47 Guest3462: right - there's another module if you're using VMware (from memory). So is it now working as intended? 2021-11-10 19:02:27 I don't know yet, I'll have to pause the VM and then try and see whether `chronyc makestep` works. If it doesn't, then it won't really bother me since I can just restart the service to resync the clock, but if `chronyc makestep` is intended to do that, then perhaps it would warrant a deeper dive to figure out why it wouldn't. 2021-11-10 19:03:37 helby: thanks, that helps (should have tried apk search apropos, but am way too lazy and tired of tapping searches into my pmOS phone when I'd just apt install foo on my laptop) 2021-11-10 19:05:15 asdflkj: oh u use pmos on phone? how is it? 2021-11-10 19:08:32 Guest3462: what's the "makestep" line that you have in your chrony.conf? 2021-11-10 19:08:39 minimal: after kernel recompile with HZ=1000, that was enough, kernel version doesnt matter but crng init done is in 2nd second of boot instead of 4th second with Alpine default HZ=250 2021-11-10 19:09:43 so I cut by half time of that, what isnt suprising since kernel ticks fire up much faster and entropy is higher I think 2021-11-10 19:10:22 minimal: currently there is no makestep line, it now just looks like this: https://paste.debian.net/plain/1219029 2021-11-10 19:10:26 MY-R: I think mps recently increased the HZ value in his linux-edge/linux-edge4virt kernels 2021-11-10 19:10:57 minimal: Im using stable so not playing with -edge stuff 2021-11-10 19:11:58 minimal: but you was right cus it was for sure eudev which was expecting more entropy 2021-11-10 19:12:08 Guest3462: quoting from https://chrony.tuxfamily.org/doc/3.5/chrony.conf.html re makestep, "This is particularly useful when using reference clocks, because the initstepslew directive works only with NTP sources" and that shows an example config file entry of "makestep 0.1 3" 2021-11-10 19:12:36 entropy/rng whatever 2021-11-10 19:13:24 Guest3462: so try adding something like that to your chrony.conf 2021-11-10 19:15:35 thanks, I'll take a look 2021-11-10 19:16:02 helby: mostly awesome, only thing I hate is alpine (stable) :P - it's missing packages for some things I really miss, especially xcape and keynav (just got my pinephone and still trying to get all their sources and compile them locally, and might try to package them later, and might in keynav as iiuc it's in edge), and bash-completion doesn't work on many commands (I miss it on 2021-11-10 19:16:05 tmux and apk) 2021-11-10 19:16:36 s/in keynav/p%/ 2021-11-10 19:16:36 asdflkj meant to say: helby: mostly awesome, only thing I hate is alpine (stable) :P - it's missing packages for some things I really miss, especially xcape and keynav (just got my pinephone and still trying to get all their sources and compile them locally, and might try to package them later, and might p% as iiuc it's in edge), and bash-completion doesn't work on many commands (I miss it on 2021-11-10 19:16:58 oops, meant & not % :P 2021-11-10 19:18:56 Guest3462: you can adjust the values for the makestep to ensure the VM clock re-syncs quickly 2021-11-10 19:19:57 the one other thing I'd like is glibc, and I could run sxmo on alarm (arch linux arm) to get it, but systemd is the one thing I'd hate more than any other distro issue. maybe I should try getting sxmo on artix, but am too ~~lazy~~ busy to try 2021-11-10 19:29:13 oh, and I should have mentioned that I'm not using my phone as a phone atm :P - I'm switching carriers and accidentally activated the pinephone with a new number when I meant to transfer my old locked flip phone's №, so I had to deactivate it and am carrying both phones until I get shipped a new sim card 2021-11-10 19:30:43 I think most poeple don't use their phone as phones ;/ more like app launcher with fb google crap 2021-11-10 19:32:03 How I miss FireFoxOS 2021-11-10 19:32:09 About 5 years too early 2021-11-10 19:32:28 helby: too true. and I use irc (sshing into a shell server running weechat in tmux) a lot like they use whatsap and discord 2021-11-10 19:33:07 whatsapp* 2021-11-10 19:37:13 Saijin_Naib[m]: nah, maemo was better. and maemo wasn't even too early, M$ just gave up on msw mobile too late 2021-11-10 19:38:06 Missed maemo, unfortunately. First smartphone was WebOS, then into FireFoxOS, then WP7 and now android... Maybe /e/OS soon 2021-11-10 19:38:41 What maemo device was your favorite? 2021-11-10 19:39:24 idk, I deliberately missed all smartphones until the latest gen of GNU ones came out 2021-11-10 19:40:56 I was just a kid when maemo was made and murdered 2021-11-10 19:42:21 My hand was forced, unfortunately. Needed smartphones for contract work. I'd much rather have a T9/QWERTY dumbphone I have to charge maybe once a week and doesn't distract me 2021-11-10 19:42:41 The PinePhone Pro keyboard looks quite nice, speaking of 2021-11-10 19:45:14 I wonder if any small keyboards fit nicely on the Librem 5, it seems like a more distinctive design compared to the pinephone's generic cheap android feel 2021-11-10 19:46:13 asdflkj: yes I m using irc the similar way, irc client running in shell server inside of tmux, simple and works great. I don't need care about any client in my laptop 2021-11-10 19:49:49 helby: I actually originally meant for this to be a temporary solution while I set up a good log backup system locally, but it's so convenient I've never left 2021-11-10 19:49:51 it's "cloud computing" with a small, friendly, and rather non-evil "cloud" 2021-11-10 19:51:20 also not too reliable, this shell server goes down ~4× a month :P 2021-11-10 19:54:10 asdflkj: hah similar for me, I was searching for kind of bouncer or at least irc logger, then found it easier and reliable to run my client in shell account ;/ 2021-11-10 19:54:17 and works great 2021-11-10 19:59:45 ircnow's free bouncer actually worked pretty well for me, I just lost all my local logs when my diskless system crashed and yak shaved / had analysis paralysis looking for an overengineered cp/rsync for my ~/.weechat that'd deduplicate and compress and have rollback and was not borg backup because that was too big a package and etc, etc, etc, until I got bored and gave up 2021-11-10 20:01:21 minimalism strikes again 2021-11-10 20:04:52 psykose: ta da! :-) 2021-11-10 20:05:10 :p 2021-11-10 20:12:21 helby: No, I wish I had patience to cook an apkbuild for it. I just 'make' it. 2021-11-10 20:15:24 alandiwix: nah I tried, I got some erros, so deleted everything ;/ 2021-11-10 20:15:43 alandiwix: I don't have patience to deal with 'making' 2021-11-10 20:16:43 alandiwix: of course I have always ton of erros, bc. my system is to tiny, that I miss practically everything needed ;/ 2021-11-10 20:20:34 the makefile for wayst doesn't even work as it uses a `ldconfig -p` to check for a library that is invalid 2021-11-10 20:20:57 if they tag an actual release and make it 'not alpha software' and fix the makefile i will make an apkbuild 2021-11-10 20:24:49 is there any doc/man how to use ethtool in ifupdown-ng? 2021-11-10 20:25:16 MY-R: what do you need it to set? 2021-11-10 20:25:25 there is also 'havoc' but that's probably so tiny that it has many issues 2021-11-10 20:25:46 so practically now just bloated terminals for wayland 2021-11-10 20:26:12 minimal: few different things and doing it with "pre-up ethtool" on physical interface result in segfault, was working with vlan interfaces.... 2021-11-10 20:27:31 minimal: I wanna set up nic with options like: ethtool -G ... ethtool -K ... ethtool -C ... 2021-11-10 20:27:35 MY-R: if you mean you want to create vlan interfaces, I've done that on a few machines, can check their interfaces file 2021-11-10 20:28:05 minimal: problem is that with vlan all working fine, but not anymore with physical devices like eth0 in rpi4 2021-11-10 20:28:40 MY-R: behaviour has changed between linux-rpi kernel versions? 2021-11-10 20:29:03 minimal: nope 2021-11-10 20:29:24 ah, you'd said "not anymore" which implies it worked before 2021-11-10 20:29:58 when used vlans then all pre-up were fine, but when doing it directly with physical devices then segfaults 2021-11-10 20:30:02 RPI ethernet may not support it - e.g. you certainly can't use an MTU larger than 1500 with RPI 2021-11-10 20:30:51 minimal: it is all about when "boot" after boot I can set them as much want but not during boot 2021-11-10 20:31:21 MY-R: ah, what about post-up then? 2021-11-10 20:32:27 example when got vlan "lan" I could set "pre-up ethtool --set-eee eth0 eee off" but when wanna use it again "iface eth0 inet static" then it segfault 2021-11-10 20:33:42 in short, used rpi4 as a router with so called "router on a stick" 1 physical eth with managed switch and all was working fine 2021-11-10 20:34:13 then I add extra usb ethernet and use normal switch and nothing work fine anymore 2021-11-10 20:34:44 even built-in eth0 in raspberry pi 2021-11-10 20:36:16 wanna set ethtool things before/during nics will go/are up but looks like I cant 2021-11-10 20:36:18 MY-R: strange. you're sure that RPI's network port actually supports stuff like this? is "ethtool -k" the option to show that? 2021-11-10 20:36:42 minimal: it support all those options cus using them already 2021-11-10 20:36:55 MY-R: ah, "ethtool --show-eee eth0" 2021-11-10 20:37:25 what does that show? 2021-11-10 20:38:07 minimal: all those stuff working but not working with ifupdown-ng script when using physical interfaces 2021-11-10 20:38:31 during boot 2021-11-10 20:39:01 with usb adapter I expect some delay etc but didnt expect same with builtin eth0 on in raspberry that I need some weird delays or whatever 2021-11-10 20:39:20 running all those thing after boot manyaly or by later running script working 2021-11-10 20:39:59 at some point I see that probably will disable "networking" service if cant deal with those things properly 2021-11-10 20:41:11 running "ethtool --show-eee eth0" on a RPI3 here I get "netlink error: Not supported" 2021-11-10 20:41:45 already not using eth1 usb adapter with ifupdown-ng because there is a delay to bring that interface up and ifupdown-ng give up with error so replaced that with dhcpcd daemon... but still have eth0 one... 2021-11-10 20:42:02 minimal: on rpi4 those changes working just fine 2021-11-10 20:43:50 MY-R: yeah the RPI4 (and 3B+) have different networking chips/phy 2021-11-10 20:43:52 minimal: and be sure you using ethtool from pkg and not the one in busybox 2021-11-10 20:44:40 or it was "ip" dont remember but behavior of ifupdown-ng is weird af 2021-11-10 20:44:46 yes this is "normal" ethtool, I think the older RPIs (100Mb, not 1Gb) probably do not support things like EEE 2021-11-10 20:48:08 MY-R: for ifupdown-ng stuff best to talk to ariadne 2021-11-10 20:48:24 best to ask in #ifupdown-ng 2021-11-10 20:48:54 I see there are ethtool things but thought maye there is some man.doc for it, Im tired to dig in all those scripts/codes to figure out what I can and what can not use 2021-11-10 20:50:00 however, 2021-11-10 20:50:06 man interfaces-ethtool 2021-11-10 20:50:15 should give you options 2021-11-10 20:50:53 TIL: #ifupdown-ng 2021-11-10 20:51:10 thank you Ariadne, but any online page with it? 2021-11-10 20:52:02 oh, apparently we do not have an interfaces-ethtool(7) yet 2021-11-10 20:52:25 ye I will put all things together and post at some point to make all things clear, I just have got hope that if will add extra "nic" then will have less "problems" to use it 2021-11-10 20:53:30 one thing you might try doing is 2021-11-10 20:53:52 iface eth1 2021-11-10 20:53:52 pre-up sleep 5 2021-11-10 20:53:52 ... 2021-11-10 20:54:14 Ariadne: that is what I would rly want to avoid, hate using sleep 2021-11-10 20:54:17 if the problem is that you're waiting for the USB nic to come up 2021-11-10 20:54:29 all of this is supposed to get solved by `ifmond` 2021-11-10 20:54:31 I know it will help at some point but is too hacky 2021-11-10 20:54:46 which will wait for events from the kernel, and bring up NICs based on that 2021-11-10 20:54:54 Ariadne: the problem isnt for now with USB one but the one built in raspberry pi 2021-11-10 20:55:12 well, what exactly is the problem then 2021-11-10 20:55:40 Ariadne the ethernet port on the pi is connected to the soc via usb 2021-11-10 20:56:12 okay, but that should be dealt with far before ifupdown-ng comes into play 2021-11-10 20:56:29 i agree but how would you do that? 2021-11-10 20:56:37 I dont see how should use ethtool stuff in "interfaces" because "pre-up ethtool ..." not working 2021-11-10 20:56:47 again 2021-11-10 20:56:51 i don't understand the issue 2021-11-10 20:57:30 i can't suggest a solution if i do not know what needs to be solved 2021-11-10 20:58:41 I have "iface eth0 inet static" and using "pre-up ethtool --set-eee eth0 eee off" return basicaly segfault 2021-11-10 20:59:04 tried with "pre-up ethtool -K eth0 sg off rx off tx off gso off gro off" and is same 2021-11-10 21:01:12 all was fine when used vlans like "iface lan0" 2021-11-10 21:01:48 pretty sure gso doesn't mean anything on vlan 2021-11-10 21:02:02 whatever, doesnt matter, it was example which was working with vlans 2021-11-10 21:03:24 with vlans probably ifupdown-ng waiting until physical interface will be up and somehow it doesnt do it when using directly physical interface 2021-11-10 21:10:11 MY-R: do you have `vlan` package installed? if so, remove it. 2021-11-10 21:10:40 almost all problems with vlans are relating to users using the `vlan` package rather than ifupdown-ng's built-in implementation 2021-11-10 21:11:56 Ariadne: nah, isnt installed, I was all doing on fresh 3.14.2 install on rpi4 in sys mode 2021-11-10 21:12:13 are you sure :) 2021-11-10 21:12:37 The following NEW packages will be installed: 2021-11-10 21:12:39 busybox-ifupdown vlan 2021-11-10 21:12:43 when trying install vlan 2021-11-10 21:12:47 ok 2021-11-10 21:12:54 maybe ask on #ifupdown-ng 2021-11-10 21:13:03 Barbarossa probably can be more helpful on this 2021-11-10 21:13:39 Ariadne: yeah, for sure I will do it and do it in more readable way ;P 2021-11-10 21:14:13 I was just missing ethtoo ifupdown-ng man/doc stuff, that is why asked on irc 2021-11-10 21:14:32 probably will install "man" stuff and try first 2021-11-10 21:17:09 is just weird that probably have to read some man/docs to figure out how to put ethtool stuff in a ifupdown-ng way 2021-11-10 21:30:48 Does anyone know how to install alpine on btrfs? i have no experience with btrfs at all and i want to take this as a learning experience 2021-11-10 21:33:48 Gaming4LifeDE: it is not very difficult 2021-11-10 21:34:02 still got no idea though 2021-11-10 21:34:24 the same way as you normally do, but make the fs btrfs 2021-11-10 21:34:49 maybe you want to don't use default volume for root if you want separated @ROOT, @HOME, etc... 2021-11-10 21:34:56 i suppose you can't use setup-alpine all the way, but instead you do formatting yourself and setup-disk on the mounted partition instead :) 2021-11-10 21:35:01 but anyway you can "fix" it later 2021-11-10 21:36:08 my end goal is to get timeshift running but for that i need a @ and @HOME volumes 2021-11-10 21:37:08 psykose i did run arch before but only on ext4/xfs 2021-11-10 21:38:50 also you have to take in account the efi/boot partitions 2021-11-10 21:39:17 but well, basically do setup-alpine, skip the installation step 2021-11-10 21:40:49 btw are there hooks for apk (not tied to a specific package). I'd want to use it to automatically do a snapshot with timeshift before running any apk operation 2021-11-10 21:41:10 then add btrfs-progs, mkfs.btrfs /dev/sdXY, then mount it somewhere, create @ROOT subvolume, mount that subvolume in other place 2021-11-10 21:41:19 and finally run, setup-disk /that/place 2021-11-10 21:43:32 maybe you have to add btrfs to mkinitfs 2021-11-10 21:43:58 probably. next issue is that timeshift needs grub instead of syslinux 2021-11-10 21:44:23 Doesn't Alpine default to GRUB on EFI systems? 2021-11-10 21:44:25 well you can run grub, setup-disk on sys-install mode installs grub by default (AFAIK) 2021-11-10 21:44:55 I'm not sure what is default but you can use grub 2021-11-10 21:45:02 which does exist for alpine afaik but idk how to do an installation like this because btrfs AND grub together is probably pretty advanced 2021-11-10 21:45:29 donoban syslinux is the default 2021-11-10 21:46:15 issue is the setup scripts don't really account for wanting other bootloaders... or fs types but you can just mount the targets where you want them 2021-11-10 21:47:28 well, if you have a global idea just try to do it and ask if something fails 2021-11-10 21:48:20 you can change some things with env vars 2021-11-10 21:49:26 uhm 2021-11-10 21:49:36 Gaming4LifeDE: maybe you can just use ROOTFS=btrfs 2021-11-10 21:49:42 and let setup-disk to do all 2021-11-10 21:50:04 read 'setup-disk --help' 2021-11-10 21:52:37 donoban just run the command like setup-disk ROOTFS=btrfs or like an env var? 2021-11-10 21:54:31 *sigh* 2021-11-10 21:55:15 uhM, theorically it's an env var 2021-11-10 21:56:33 ok i'll try 2021-11-10 21:58:40 good luck, I'm quitting now see you tomorrow 2021-11-11 00:37:17 exit 2021-11-11 02:06:16 i'm considering switching to alpine as my base system on the server and then running my "dev server" in a VM, mostly for security reasons. right now i'm a software dev (read: not a security concious administrator). can anyone recommend articles i should be aware of? 2021-11-11 02:08:00 big topic. though, if i could suggest one thing... anytime you read one of those articles and they're recommending changes... be sure to take the time to understand what those changes actually do. 2021-11-11 02:08:13 ("dev server" being pretty loose. it's my home system where i mostly just hack on software and don't run mission critical stuff on it really but as you hopefully understand, i do increasingly run some services on it such as pi-hole, a minimal webserver, etc). 2021-11-11 02:08:32 the inverse of that is, don't make changes you don't understand 2021-11-11 02:10:23 invoked: definitely. :) but i am hoping alpine has some sort of guideline. want to have a fairly secure "core" base system, so that i can have my VM's logically separated in terms of role. 2021-11-11 02:11:23 i realise this isn't "ideal" - they should be physically separate machines on separate networks, but i don't have that, gotta make do with what i have and make sure it's fit for purpose. this is still ultimately a personal system, afterall. 2021-11-11 02:11:43 the security concepts are generally the same across the linux ecosystem. alpine starts with a minimal base system, so there's less for you to unpack/understand when you get started 2021-11-11 02:12:21 excellent, thats what i'm after :) 2021-11-11 02:12:43 arahael2: also depends on what you're using to run the VMs, I'm guessing libvirtd? 2021-11-11 02:13:02 i keep forgetting i'm going to nick highlight minimal when i write minimal 2021-11-11 02:13:03 minimal: yep! 2021-11-11 02:13:56 invoked: I was biting my lip at your mention of a "minimal base system" ;-) 2021-11-11 02:14:08 that's on me 2021-11-11 02:14:43 see, the scenario is that i have been playing with various services and opening them up to the internet, but i've realised that a lot of the services i am running had no thought to security - eg, i am not confident they won't leak files outside of their docroot, that sort of thing. 2021-11-11 02:15:07 and i've recently been reading this document: https://madaidans-insecurities.github.io/guides/linux-hardening.html 2021-11-11 02:15:44 (thats way in excess of what i intend to do - i am *not* compiling my own kernel, as an example, but it is an interesting and illuminating document) 2021-11-11 02:15:59 arahael2: yeah I've done some of those things, still working my way through hardening 2021-11-11 02:17:00 i currently use, and like nixos, but for various reasons i am not sure it's a great fit for a public facing server. has other awesome features for some aspects of software dev though. 2021-11-11 02:17:36 arahael2: even for a "minimal" distro there's still quite a bit of work to harden/lock things down 2021-11-11 02:17:56 you're not going to learn it all in a day or a month, so when you're putting it together, understand the changes you make and take notes as you go along. 2021-11-11 02:18:03 i am hoping that with alpine, i'll at least be ahead of the pack. 2021-11-11 02:18:41 and you can still shoot yourself in the foot with alpine :) 2021-11-11 02:18:58 just kind of... uh, be methodical 2021-11-11 02:19:30 invoked: i've been using linux for close to two decades now, in various capacities, so i am not really starting as a newbie. :) that said, my working understanding of linux is still from the 2.4 days, and i am not sure they had namespaces back them in the same way we do today. 2021-11-11 02:19:49 yeah, seccomp and everything is kind of new, i guess 2021-11-11 02:20:21 arahael2: I've gone down the root of hardening a base image but also using Ansible run on a regular basis to ensure the hardening stays in place 2021-11-11 02:20:26 s/root/route/ 2021-11-11 02:20:26 minimal meant to say: arahael2: I've gone down the route of hardening a base image but also using Ansible run on a regular basis to ensure the hardening stays in place 2021-11-11 02:21:27 alpine-meetbot: curious, that might be interesting to take a peek at? i dont think i'll be going to quite that level. 2021-11-11 02:22:11 i also need to use glibc, etc, but i think my workaround for that will be to do glibc crap in a vm. 2021-11-11 02:23:18 arahael2: you'll still have to think about logging/monitoring/metrics/backups etc for the host OS 2021-11-11 02:24:31 yeah, i archive everything to a second hard drive, which happens to be internal, and i duplicate it all wih my laptop which is almost always with me. rsync is awesome. 2021-11-11 02:43:04 whats the alpine release process? is it a rolling release? (i don't like rolling releases, except for security updates) 2021-11-11 02:48:28 alpine has releases and then 'edge' which is basically your rolling release 2021-11-11 02:49:31 "latest-stable" will match whatever the most recent stable version is, otherwise you can specify the release number. 2021-11-11 02:52:11 https://docs.alpinelinux.org/user-handbook/0.1a/Working/apk.html 2021-11-11 02:53:09 awesome, thanks! 2021-11-11 03:34:30 there's a user handbook? :O 2021-11-11 03:34:35 so what's the wiki for then? 2021-11-11 03:37:06 online observation 2021-11-11 04:07:09 ha 2021-11-11 04:38:56 i thought i heard a rumor that the wiki might be going away. i know void is killing their wiki at some point 2021-11-11 04:42:27 Anyone can help me spot what I'm doing wrong with my APKBUILD? 2021-11-11 04:42:28 https://gist.github.com/Saijin-Naib/b81d7d619d17660bc7d4048a581c9e7c 2021-11-11 04:42:44 ACTION sent a code block: https://matrix.org/_matrix/media/r0/download/matrix.org/QEcsFtzRfAxSPWTlAPyxCzwh 2021-11-11 05:03:53 hi, how can I replace OpenRC with runit? There's a guide on gentoo wiki and a discussion on dev.alpinelinux.org about it, but i'm still confused on how to get it working 2021-11-11 05:04:22 i'm currently on void linux and i want to use runit on alpine 2021-11-11 10:13:50 Hi 2021-11-11 10:14:16 Is it possible to create virtual packages, kind of like sets on Gentoo? 2021-11-11 10:22:36 Forza, if i recall correctly, alpine-sdk might be interesting for you to look at 2021-11-11 10:23:43 2 options: manually with apk add -t [dependency..] 2021-11-11 10:23:53 2nd is a metapackage, which has dependencies but no files 2021-11-11 10:32:35 Ah ok. Was thinking it would be easy to manage virtuals when testing many packages. If I want to remove i simply remove the virtual. Does that make sense? 2021-11-11 10:33:35 Then apk add -t (--virtual) would make most sense 2021-11-11 10:33:53 That's what abuild uses to install dependencies and then remove them again after building 2021-11-11 10:34:10 Thanks! 2021-11-11 10:46:37 ikke are there plans to implement package groups too? i mean stuff like gnome should be a metapackage because they're all essential but if you install gnome-apps-core and then decide you don't need one of the applications it installs, it wipes everything, which won't happen with a package group 2021-11-11 10:47:22 afaik not any plans at the moment 2021-11-11 10:48:04 but gnome is meta package ;/ 2021-11-11 10:53:04 how can I know from a script if kernel can mount squashfs file ? 2021-11-11 10:53:16 like from some info in /sys or /dev 2021-11-11 10:58:38 vkrishn: you could look at /proc/filesystems, but squashfs is a module which might not have been loaded yet 2021-11-11 10:59:43 Hi again.how do i see what package an existing file belongs to 2021-11-11 11:00:19 Info -W 2021-11-11 11:00:22 yes 2021-11-11 11:00:22 :) 2021-11-11 11:00:26 quack 2021-11-11 11:01:24 I need to send mails through an external smtp. Could ssmtp be a good alternative? It's netdata monitoring that uses local sendmail 2021-11-11 11:08:47 thanks, info here /proc/filesystems would show all kernel inbuild ones, irrespectively ? 2021-11-11 11:10:17 If it's built-in, it should show up there 2021-11-11 11:10:31 ok, thanks 2021-11-11 11:11:09 I think what ikke is implying is that if you have to load a module, it won't be in there. 2021-11-11 11:11:41 correct 2021-11-11 11:18:07 The util-linux version of mount will automatically insert the correct module for example 2021-11-11 11:18:27 bb mount will if you specify the fs type 2021-11-11 11:22:31 if its a module build with -m flags 2021-11-11 11:23:19 =m, yes 2021-11-11 11:23:40 here /proc/filesystems shows btrfs, which I have never used, so its inbuild 2021-11-11 11:25:16 https://gitlab.alpinelinux.org/alpine/aports/-/blob/master/main/linux-lts/config-lts.x86_64#L8415 2021-11-11 11:25:20 it's a module 2021-11-11 11:30:04 CONFIG_F2FS_FS_LZO=y, while CONFIG_F2FS_FS=m, will this not bloat the kernel ? 2021-11-11 11:30:32 or guessing CONFIG_F2FS_FS_LZO is a more generic module 2021-11-11 11:32:33 ok got it build the module ith LZO supprt, right? 2021-11-11 11:41:36 The packaged netdata doesn't seem to provide the mysql plugin. Do you know if its available in some other package? 2021-11-11 11:42:13 I don't think so 2021-11-11 11:44:32 Oh I see. The netdata-python has it. 2021-11-11 11:44:43 ah, ok 2021-11-11 11:45:18 But it complains for missing MySQLdb or PyMySQL missing 2021-11-11 11:48:35 Community repo has py3—mysqlclient 2021-11-11 11:48:39 Nice 2021-11-11 11:51:04 Ok, so that should probably be added as a dependency then 2021-11-11 11:51:10 Does it work with that package? 2021-11-11 12:30:23 I'm doing some cleanup on my laptop and noticed that I have /etc/init.d/asterisk but I don't have asterisk or asterisk-openrc installed, how can this happen? 2021-11-11 12:31:35 donoban: an apk info -W /etc/init.d/asterisk yields nothing? 2021-11-11 12:31:52 ERROR: /etc/init.d/asterisk: Could not find owner package 2021-11-11 12:32:29 https://pkgs.alpinelinux.org/contents?file=&path=&name=asterisk-openrc&branch=edge&repo=main&arch=x86_64 2021-11-11 12:32:48 I have /etc/init.d/asterisk but no /etc/conf.d/asterisk 2021-11-11 12:35:07 if you edit a file it's not removed in uninstall 2021-11-11 12:35:38 uhM 2021-11-11 12:36:03 nor replaced on install afterward :) 2021-11-11 12:36:21 try it with anything that installs anything in /etc 2021-11-11 12:36:34 add/del will add/remove it, edit it and it will stay forever 2021-11-11 12:36:43 I copied it and installed the package 2021-11-11 12:36:51 in fact they are different but I didn't remeber edit it 2021-11-11 12:37:48 probably due to unmerged updates between versions, but i dunno 2021-11-11 12:38:45 well, I will just remove it :D 2021-11-11 13:08:27 ikke: no, the kernel loads the module via modprobe upcall 2021-11-11 13:09:09 Oh 2021-11-11 13:09:40 e.g. modinfo vfat -> alias fs-vfat 2021-11-11 13:18:41 Hello71: And I suppose bb mount does not do fs detection so does not ask the kernel to do that? 2021-11-11 13:19:07 yes, you need to provide correct fs type to kernel 2021-11-11 13:19:17 right 2021-11-11 13:19:35 util-linux calls libblkid. busybox has blkid but doesn't use it for this purpose for some reason 2021-11-11 13:20:05 it just tries everything in /proc/filesystems one at a time 2021-11-11 13:20:59 (busybox blkid has fs detection) 2021-11-11 13:25:25 Hi again. I have a fixed ip on a machine, but I'm not getting a broadcast address set according to ifconfig. In /etc/network/interfaces I have "address 192.168.20.0/24" set. I'd assume that 192.168.20.255 would be set as broadcast? 2021-11-11 13:27:21 I tried adding broadcast 192...255 but doesn't seem to set 2021-11-11 13:27:44 seems like landley would have you manually copy all filesystem names to /etc/filesystems 2021-11-11 13:28:09 which is annoying and also because mount checks /etc/filesystems first, would cause all filesystems to be modprobed whenever you do a mount 2021-11-11 13:28:29 Forza: ifconfig is obsolete, use ip 2021-11-11 13:29:21 Yes true. ip is also not listing brd 2021-11-11 13:31:13 Why do you care about the broadcast address? Normally that should just work 2021-11-11 13:32:54 I just saw broadcast missing from ip/ifconfig output 2021-11-11 13:33:57 Isn't this a little weird? 2021-11-11 13:34:08 I don't have a need for broadcast though. 2021-11-11 13:35:18 https://wiki.alpinelinux.org/wiki/Configure_Networking doesn't list broadcast, while https://wiki.alpinelinux.org/wiki/Comparison_with_other_distros does 2021-11-11 13:41:15 you can add broadcast into the config if you want it, no? 2021-11-11 13:41:35 psykose: yes but it didn't help 2021-11-11 13:43:32 weird, also doesn't work for me if i use static and put it there 2021-11-11 13:43:34 with dhcp it works 2021-11-11 13:43:44 it doesn't do anything useful though 2021-11-11 13:44:46 the brd is ignored when broadcasting? 2021-11-11 13:45:10 I can ping the. 255 address and get responses from other ips 2021-11-11 13:45:41 But you are right. When using dhcp I do get a broadcast address set 2021-11-11 13:45:59 Maybe the networking scripts have a bug and forgets to set it? 2021-11-11 14:10:54 Should I report this as a bug somewhere? 2021-11-11 14:14:29 Unless there is a real issue, I would not bother 2021-11-11 14:16:44 OK. :) my OCD doesn't like it though haha 2021-11-11 14:22:08 that is not what ocd is 2021-11-11 14:22:20 what's going on? 2021-11-11 14:22:32 classic networking :) 2021-11-11 14:23:46 does anyone know how to inspect the difference between terminfo files? there is a difference between the foot-terminfo in ncurses-terminfo and the terminfo if i build foot myself 2021-11-11 14:24:05 the ncurses-terminfo is broken for foot in the sense that whatever sets the terminal title doesn't actually work, but it does if i use my own built terminfo 2021-11-11 14:28:03 psykose: I know. But it is an expression. :/ sorry if I was using it without considering 2021-11-11 14:28:07 the irc client i use sets it to via ncurses, with ncurses-terminfo the set is ignored/does nothing so it's just the default title of or whatever is set by the shell itself 2021-11-11 14:28:26 Forza: don't worry about it :) 2021-11-11 14:30:07 psykose: I think foot's own terminfo has some extra functionality which is not in ncurses, at least something like I read on foot's git about terminfo nonsense 2021-11-11 14:30:45 the majority of the terminfo discussion in foot was about ncurses upstreaming foot's terminfo into itself, like it has for most terminals, and dealing with install issues 2021-11-11 14:30:49 unless you can point to something specific 2021-11-11 14:31:16 granted, that was about 2 months ago and i haven't checked since 2021-11-11 14:32:04 if it really is an incomplete terminfo file i think we should bring foot-terminfo back 2021-11-11 14:32:24 it was removed in favor of ncurses-terminfo after the merge 2021-11-11 14:34:09 https://codeberg.org/dnkl/foot/src/branch/master/INSTALL.md#user-content-other 2021-11-11 14:38:55 How can I replace OpenRC with runit? I'm currently using Void and I'd like to use runit as my init system and a service supervisor 2021-11-11 14:38:57 ah 2021-11-11 14:39:11 i guess one of the missing features is... setting the terminal title then? that doesn't really seem correct 2021-11-11 14:39:24 that is probably a very standard feature going back a million years, unless something is different here 2021-11-11 14:40:32 cottnn: nobody on alpine uses runit for init+supervisor, the package is there mostly just for supervision (e.g. in container if wanted), so nobody has really done this i don't think aside from maybe one person somewhere with a lot of time on their hands 2021-11-11 14:40:49 it's not a 'supported' configuration or something, so you would have to trial-and-error into getting it to run yourself 2021-11-11 14:41:38 psykose: i see, would it be easy enough to somehow move the runit configuration and service files from Void to Alpine? 2021-11-11 14:41:55 i have no idea 2021-11-11 14:42:19 i would assume frankensteining 80% of void's setup from a musl install would work, but you really have to know what you're doing for the rest 2021-11-11 14:42:52 and it might just break at any time for whatever reason; if you don't actually know how all these pieces work to get started yourself without needing someone to tell you i would not recommend doing it 2021-11-11 14:43:19 but it's a fun learning experience, so go for it :) 2021-11-11 14:43:24 psykose: i see, so for someone that wants to use runit but also likes alpine's philosophy, you'd recommend sticking with void-musl? 2021-11-11 14:43:52 i don't recommend anything :) you can do what your heart seeks 2021-11-11 14:45:55 ok, thanks for the help 2021-11-11 14:47:12 i tried out alpine just today in a vm and it seems to be so similar to void in both install size and RAM usage 2021-11-11 14:47:56 like, they both use about 1g of disk space for a fresh install and about 50mb of ram 2021-11-11 15:16:28 Has anyone experienced rfkill showing your wireless lan hard blocked when the physical switch for wifi is open? What did you do about it? 2021-11-11 15:19:08 For me, I see at seemingly random times, "rfkill list" output "phy0 wireless LAN - hard blocked: yes". Sometimes flipping the switch off/on helps. Sometimes not. But it shouldn't happen in the first place. 2021-11-11 15:19:57 this sounds like more of a computer issue 2021-11-11 15:20:32 i do remember once having hard-blocked and being confused about it, but in my case there was a setting that blocks it when ethernet is connected at the firmware level 2021-11-11 15:21:54 I have seen the same on an hp laptop - toggling the physical switch worked for me 2021-11-11 15:22:39 I have it randomly happening on my (non-alpine) laptop as well 2021-11-11 15:30:00 helby: i made !27363 i guess, that should fix it :) 2021-11-11 15:32:21 psykose: depends="ncurses-terminfo" 2021-11-11 15:32:52 Ikke: any fix that has worked consistently? 2021-11-11 15:33:00 no 2021-11-11 15:33:04 that is correct, if you read what is written 2021-11-11 15:33:09 I just enable it again 2021-11-11 15:34:23 psykose: this way it works, that I have to install foot-terminfo and it will replace ncurser-... 2021-11-11 15:34:39 yes 2021-11-11 15:36:37 cool, I was just 'bitching' about that huge ncurses for foot yesterday ;/ 2021-11-11 15:57:46 ikke nangel: If I were to delete /dev/rfkill, would the device show up normally after a restart? This action was suggested in a forum. Have you tried this? 2021-11-11 15:58:25 Not sure why it would make sense, unless you expect something on your system messing with rfkill 2021-11-11 16:00:03 afaik doesn't hard blocked mean not related to userspace 2021-11-11 16:01:33 I don't know what to expect really. I thought my wifi card stopped working at first. At best any fix feels like random chance. 2021-11-11 16:02:10 I'd love to have something consistent to point to, but I don't yet. 2021-11-11 16:03:53 bovis: I had times where the entire wireless nic just stopped working / disappeared. I had to remove it on a PCI level in linux and rediscover it again to get network back 2021-11-11 16:04:11 but that was a while ago since that last happened, so it seems to be fixed 2021-11-11 16:04:18 Could be a firmware or kernel bug 2021-11-11 16:15:28 Recommend a way to easily track the on/off state of wifi? If I could see when it happens alongside other events, that could be really helpful. 2021-11-11 16:18:58 dmesg should have something that prints about it disconnecting 2021-11-11 16:19:10 if it's related kernel events at the same time they will be around it 2021-11-11 17:05:31 *sigh* i have aVPS that only allows installing Debian or Ubuntu or but no Alpine.. SO, i figured, i install one, make a new partition, unzip the minimalrootfs to it, add it to grub and me done with it. Clearly i was wrong. i still only boot ubuntu :| Does anyone know of a way to "bootstrap" AlpineLinux from within a running system? 2021-11-11 17:05:31 i tried http://sdomi.pl/weblog/12-bootstrapping-alpine-on-oraclecloud/ but that doesnt work either 2021-11-11 17:08:57 jhjacobs81: I thought there was a Wiki article on this but can't seem to find it 2021-11-11 17:09:25 there is, indeed, wifi stuff, but its either outdated or im doing something terribly wrong 2021-11-11 17:10:06 im dead set on using Alpine, because i want to base my firewall on it too :) 2021-11-11 17:10:16 awall seems rather nice to use, for a change 2021-11-11 17:11:09 jhjacobs81: if it was a cloud provider that provided means to import disk images then you'd have more options 2021-11-11 17:12:23 yeah, unfortunatly :( 2021-11-11 17:12:34 they added it as a suggestion for consideration :D 2021-11-11 17:12:48 okay, well.. i'llkeep playing some more :) thanks though! 2021-11-11 17:12:56 jhjacobs81: so what happens when it boots? does Grub menu show Alpine option? if so, then when you select it what happens next? 2021-11-11 17:13:04 next monday im running Alpine 2021-11-11 17:13:40 well, when i added it manually i got a kernel panic.. when i use update-grub it just moves on to ubuntu 2021-11-11 17:13:55 and now i just reinstalled the whole vps, but i will ask when i see new errors :) 2021-11-11 17:15:07 jhjacobs81: so update-grub wrote a modified grub.cfg with both ubuntu and alpine entries? 2021-11-11 17:15:52 change vps 2021-11-11 17:16:42 minimal yes :) 2021-11-11 17:16:56 helby if it was possible, i would have done that already :P 2021-11-11 17:21:43 jhjacobs81: this is the article I was thinking of: https://wiki.alpinelinux.org/wiki/Replacing_non-Alpine_Linux_with_Alpine_remotely 2021-11-11 17:22:25 thats a new one 2021-11-11 17:23:28 in a few minutes i will try :) 2021-11-11 18:20:52 jhjacobs81: smells like xen 2021-11-11 19:03:37 could very well be :) i have no access to the underlying hypervisor :) 2021-11-11 19:03:43 in any case.. im almost there :) 2021-11-11 19:20:18 jhjacobs81: do they have an option to boot a custom iso/rescue system? I usually use that to install Alpine on my VPS's 2021-11-11 19:20:44 no :) if they did i would boot the iso indeed 2021-11-11 19:20:51 BUT! i managed to make it work :) 2021-11-11 19:33:37 i really do not like someone say "sorry, that wont work" 2021-11-11 19:33:39 :) 2021-11-11 20:12:34 jhjacobs81: many things in history have been achieved with bloodymindedness after being told "that won't work!" :-) 2021-11-11 20:43:28 :D 2021-11-11 20:43:34 i got my install! 2021-11-11 20:43:50 the url i posted earlier was made for EFI systems.. D'OH! 2021-11-11 20:44:09 changed some stuff, booted into alpine, installed the alpine kernel etc etc :) 2021-11-11 20:44:16 now im fullblood alpine :) 2021-11-11 20:44:37 been a debian user since 1999 so it'll be very different 2021-11-11 20:47:26 Welcome and enjoy! 2021-11-11 20:49:13 jhjacobs81: good job! 2021-11-11 21:12:02 hey, I was wondering what git command you guys run to generate the commit statistics here https://alpinelinux.org/posts/Alpine-3.14.0-released.html 2021-11-11 21:12:11 I want to include it for a project I run 2021-11-11 21:12:21 Idk how to do it based on git tags 2021-11-11 21:12:38 git shortlog 2021-11-11 21:15:00 oh derp, I hadnt fetched the tags 2021-11-11 21:15:04 thanks ikke 2021-11-11 21:15:16 thats why shortlog wasnt working.... 2021-11-11 21:15:50 heh 2021-11-11 22:01:26 hello 2021-11-11 22:02:21 I need some help in using alpine-extendex in DATA (run in ram) mode 2021-11-11 22:02:46 when I update the kernel via the update-kernel script, I lose zfs modules & support 2021-11-11 22:03:31 if I force them into the initram via parameters to update-kernel 2021-11-11 22:04:02 the usb bootable disk fails to boot and hangs at boot "mounting boot media: failed" 2021-11-11 22:04:14 like it's not loading the usb modules (?) 2021-11-11 22:04:26 it's very weird indeed 2021-11-11 22:52:44 if only i could fix the boot proces :( 2021-11-11 23:09:30 jhjacobs81: what's the boot issue? 2021-11-11 23:27:41 well, i tried installing syslinux but i keep running into problems 2021-11-11 23:28:41 jhjacobs81: installing the syslinux package or setting up syslinux? What sort of problems? 2021-11-11 23:28:42 but oh well, thats for tomorrow :) 2021-11-11 23:29:17 https://wiki.alpinelinux.org/wiki/Bootloaders following this guide it keeps saying cant find os 2021-11-11 23:29:34 jhjacobs81: does the VPS Provider support Syslinux? maybe they only support Grub 2021-11-11 23:29:36 the old debian os is on /dev/vda2 with grub 2021-11-11 23:29:55 hmm.. good poit 2021-11-11 23:31:35 i hate this provider :) 2021-11-11 23:39:47 jhjacobs81: so install grub and overwrite the existing bootsector 2021-11-11 23:40:49 :-) 2021-11-11 23:58:44 Alpine would make a good switch os too yes? 2021-11-12 00:09:28 if you can boot it ;/ 2021-11-12 02:52:17 psykose: re wayst terminal, the problem with ldconfig is that Alpine uses a bare bones shell script for ldconfig which doesn't support the -p option. 2021-11-12 07:47:06 hey, I'm trying to install a kernel with efistub but 2021-11-12 07:47:27 why does copy-paste send the contents of the chatbox 2021-11-12 07:47:30 >.< 2021-11-12 07:47:56 anyway i'm trying to install a kernel with efistub but it never gets past "loaded initrd from command line option" 2021-11-12 07:48:30 not sure what i'm doing wrong, i checked efibootmgr -v and everything looks correct 2021-11-12 11:12:24 spiralw: how do u set it up? from wiki? 2021-11-12 11:33:09 ktprograms: yes, but i don't see the point of checking at all, they should add a configure script to let you --enable it if it's optional (it is), or just -l it like GL and fail if it's not there 2021-11-12 11:33:14 it's trying to be smart and auto it :) 2021-11-12 11:33:40 the terminal also doesn't select fonts properly, so alpha software it is indeed 2021-11-12 12:16:10 helby: yes, "Bootloaders" in wiki 2021-11-12 12:17:19 spiralw: use this previous entry https://wiki.alpinelinux.org/w/index.php?title=Bootloaders&oldid=19577 2021-11-12 12:17:38 spiralw: for whatever reason somebody deleted loading modules 2021-11-12 12:17:46 spiralw: paste your efibootmgr -v too 2021-11-12 12:19:10 spiralw: so what u need, it's just format /boot as vfat, mount it, also in fstab, run that script once 2021-11-12 12:19:13 efibootmgr: Boot0001* LinuxHD(1,GPT,686c1a2b-831d-c54c-bc9c-9d5fc0253d7c,0x800,0x100000)/File(\vmlinuz-edge)r.o.o.t.=.P.A.R.T.U.U.I.D.=.d.c.d.8.f.a.e.1.-.0.7.5.9.-.4.5.a.3.-.a.0.f.4.-.5.0.3.a.9.8.a.1.9.b.c.f. 2021-11-12 12:19:31 no clue what happened with the dots there, it was correct when i typed it in 2021-11-12 12:19:51 that is how it displays 2021-11-12 12:19:56 ah ok 2021-11-12 12:20:15 you seems to be missing everything in that entry, that only specifies root 2021-11-12 12:20:17 it's okay, just use previous script from the link with load modules 2021-11-12 12:20:30 psykose: yeah, I was trying a few things. i'll remove it and start again 2021-11-12 12:20:47 lemme reboot into alpine rq 2021-11-12 12:21:13 but yeah just add the initrd= lines for the initrd and ucode as per the wiki link, and the modules you need to mount root, and then it should be fine :) 2021-11-12 12:32:19 tfw typo 2021-11-12 12:32:32 luckily i can have both efistub and grub working at the same time 2021-11-12 12:33:26 oh sick, works now 2021-11-12 12:36:23 how do I add my own custom initramfs to this? I want to set up luks, and the "setting up a laptop" guide on wiki assumes root is unencrypted 2021-11-12 12:40:56 initrd= is the initramfs 2021-11-12 12:41:22 but those aren't custom for luks, you edit /etc/mkinitfs stuff to enable the things you need for luks 2021-11-12 12:42:21 ah ok, i will take a look at that 2021-11-12 13:13:24 afaik you want to add cryptsetup to features= and then having cryptroot= stuff on the cmdline will trigger the rest, initramfs init script will see it and use cryptsetup to unlock 2021-11-12 13:13:45 not sure how up to date the guides or whatever are 2021-11-12 13:46:58 hmm, how does alpine kernel get built? I tried building my own kernel and replacing alpine's one for efistub, and it didn't work 2021-11-12 13:48:47 https://gitlab.alpinelinux.org/alpine/aports/-/blob/master/main/linux-lts/APKBUILD 2021-11-12 13:49:44 nice thanks 2021-11-12 14:23:42 for some reason bluetooth works on -edge but not -lts for me, despite dmesg having identical output except for 1 line, and loading the same firmware, and having the same relevant kernel options... 2021-11-12 14:23:58 https://termbin.com/qzbh only difference is `hci0: No device address configured` for lts 2021-11-12 14:24:30 it's even the same kernel version currently, so it must be a config difference 2021-11-12 14:24:58 the card is an ax210 and uses btusb afaik 2021-11-12 14:28:45 psykose: Just wondering what you mean by wayst doesn't select fonts properly? If I set font = "DejaVu Sans Mono" in its config it seems to work? 2021-11-12 14:33:36 ktprograms: https://termbin.com/f08i 2021-11-12 14:33:56 i can pass any style-regular style-bold or whatever and it ignores all of them as they are in the same file; it can only use the default in the .ttc 2021-11-12 14:34:18 it doesn't actually use any list of fonts but depends on it being individual files 2021-11-12 14:34:41 it works if you have a .ttf per font, of course, but every other terminal i've used supports this fine, so :p 2021-11-12 14:35:21 or that's my assumption for what is happening; i have no actual idea what it's trying to do, but it certainly is not loading the style specified in the config 2021-11-12 14:36:29 for contrast in foot you can specify the entire font string as one, i.e. `Iosevka Curly Semibold:size=14` and it loads exactly that, using fontconfig correctly 2021-11-12 14:37:10 by fc-list that file has 162 variants 2021-11-12 14:48:53 okay, now I'm really confused. I re-built the kernel as per the link @ikke sent, and replaced the existing one with it, and it still gets stuck at "loaded initrd from command line option" 2021-11-12 14:49:14 nothing else different. kernel shipped with iso download works, kernel built myself does not work 2021-11-12 14:49:34 did you rebuild the initramfs after copying the kernel 2021-11-12 14:49:59 no, but I shouldn't have to do that since I replaced the initramfs completely 2021-11-12 14:50:19 literally any initramfs works correctly with the shipped kernel 2021-11-12 14:51:08 what did you 'replace' it with 2021-11-12 14:52:12 i replaced it with an otherwise empty cpio image with a small program in /init that prints text to the console and then waits forever 2021-11-12 14:52:28 it boots correctly with the iso kernel 2021-11-12 14:52:44 i also tried the iso's initramfs with the alpine kernel i built 2021-11-12 14:53:08 how did you build the kernel 2021-11-12 14:54:20 downloaded the sources / config files listed in the apkbuild, then ran the steps by hand 2021-11-12 14:54:29 i don't have an alpine dev machine to actually run abuild on 2021-11-12 14:55:24 cp config-lts.x86_64 .config && make ? 2021-11-12 14:55:29 or similar 2021-11-12 14:55:33 pretty much, yes 2021-11-12 14:55:45 if it was literally that then it should work 2021-11-12 14:55:49 if it was anything else, perhaps not 2021-11-12 14:56:22 it's using a patch kernel version so i applied 5.15.1 patches first, but other than that it's literally what you said 2021-11-12 14:56:27 including modules_install of course 2021-11-12 14:56:39 well 2021-11-12 14:56:45 I missed that part 2021-11-12 14:56:57 then you had a non built kernel 2021-11-12 14:57:01 what 2021-11-12 14:57:05 just add that part and it should work :) 2021-11-12 14:57:07 okay 2021-11-12 15:00:35 what is DTBS_PATH? i'm unsure how to adapt that to my build environment 2021-11-12 15:02:55 place to output devicetree overlays 2021-11-12 15:03:11 only arm* archs call dtbs_install though 2021-11-12 15:03:32 ah i'm on x86_64 so not important then 2021-11-12 15:03:44 let's see if this works 2021-11-12 15:04:13 even if you need it it's just set to /boot/dtbs-lts or etc 2021-11-12 15:04:19 ah ok 2021-11-12 15:04:53 nope, does not work :( 2021-11-12 15:05:39 what exactly did you run 2021-11-12 15:05:57 make -j1 modules_install ARCH=x86_64 2021-11-12 15:07:17 do i have to rebuild kernel after that? it seems this gets ran afterwards 2021-11-12 15:07:25 well, you don't need the arch in there 2021-11-12 15:07:27 no, that is after the build 2021-11-12 15:07:52 modules_install should make something in /lib/modules 2021-11-12 15:08:13 well, that's useless since the only thing i'm copying over from dev machine to my laptop is bzImage 2021-11-12 15:08:25 hahaha 2021-11-12 15:08:36 yes, missing a lot of files there :p 2021-11-12 15:08:45 aaa 2021-11-12 15:09:00 okay uh 2021-11-12 15:09:15 can I build those modules directly into the kernel somehow? 2021-11-12 15:09:29 only by spending 2 hours in the kconfig and knowing exactly what you need 2021-11-12 15:09:35 fun. 2021-11-12 15:09:45 i'll skip that for now then 2021-11-12 15:09:49 or you just copy the /lib/modules/kernel folder 2021-11-12 15:10:00 i don't know why you are building your own kernel in the first place though 2021-11-12 15:11:08 i don't *really* need a custom kernel now that i can boot arbitrary initrds with alpine kernel 2021-11-12 15:11:32 but i guess because I like knowing how to do stuff myself 2021-11-12 15:11:55 originally I thought that the kernel in efistub had to be really really small, but I was wrong 2021-11-12 15:12:19 yeah, it can be anything 2021-11-12 15:12:31 well, i would start by building it on the actual laptop so you don't have a bunch of the parts missing 2021-11-12 15:13:17 does it scan my hardware to see what it needs, or something? 2021-11-12 15:13:28 no 2021-11-12 15:13:40 but it's a lot more convenient than manually copying a bunch of files you seem to have forgotten each time 2021-11-12 15:13:49 for sure 2021-11-12 15:14:57 well, thanks for your help! i'll stick with alpine kernel for now and spend some time looking through kconfig at some point in the future 2021-11-12 15:16:06 when the kernel boots it loads modules from /lib/modules/kernelversion-localversion 2021-11-12 15:16:26 the localversion in alpine is set to -pkgrel-flavor, that's why you see e.g. -3-lts 2021-11-12 15:16:37 in the kconfig you can specify your own localversion so it doesn't conflict 2021-11-12 15:16:49 what you name it under /boot doesn't matter, feel free to make it whatever and load it as such 2021-11-12 15:18:05 that localversion is set in the abuild here https://gitlab.alpinelinux.org/alpine/aports/-/blob/master/main/linux-lts/APKBUILD#L110 , but if you are building yourself with make it will be empty; use the kernel config things e.g. `make nconfig` and edit the localversion somewhere in the menu there 2021-11-12 15:18:14 or directly in the .config if you know what you're doing :) 2021-11-12 16:05:27 psykose: hm, i just thought of something. you mentioned the kernel finds modules in /lib/modules if they are not compiled in, but that is not accessible if i'm using my custom initramfs 2021-11-12 16:05:31 but it still works 2021-11-12 16:05:49 so the required modules must be compiled directly into the kernel 2021-11-12 16:06:06 the modules you need are copied to the initramfs 2021-11-12 16:06:20 that is part of what mkinitfs does normally, in normal usage 2021-11-12 16:06:50 right, but my custom initramfs is just an empty directory with a single init binary, not built with mkinitfs 2021-11-12 16:06:56 yes, of course 2021-11-12 16:07:02 i'm just telling you how it normally goes 2021-11-12 16:07:10 if you want to do it yourself, you can do anything; no convention 2021-11-12 16:07:10 ah 2021-11-12 16:07:24 on gentoo i had most things built-in in my own config and i didn't use an initramfs at all 2021-11-12 16:07:39 that makes sense 2021-11-12 16:07:44 I'm just really confused as to why it works is all 2021-11-12 16:17:56 oh well that explains it, the kernel I built doesn't have fbdev module builtin 2021-11-12 16:18:04 of course it doesn't work 2021-11-12 16:44:42 it works... just the display doesn't :p 2021-11-12 16:55:45 omni: btw now with the last one river does not create any example 2021-11-12 17:00:49 omni: nevermind, I just see it install only if doc package is installed 2021-11-12 17:02:14 no idea how user get that something can be and maybe is there but ... 2021-11-12 22:05:49 Hello I just upgraded to 3.14.3, I now experience slow connection speeds to the internet. ANy has this problem? How can I revert to 3.14.2? 2021-11-12 22:08:55 I went for a 300mbit/s internet connection to a 30mbit/s.... The other node which is still om 3.14.0 has no problem at all 2021-11-12 22:11:29 revert and stay outdated with security holes forever? 2021-11-12 22:11:39 Marc0-1231: which protocols? HTTP? HTTPs? is the machine wireless or ethernet connected? does ping show any packet loss? 2021-11-12 22:11:49 you can probably if you find outdate repository I assume 2021-11-12 22:11:55 there's not enough information to go on 2021-11-12 22:12:20 I tested this with the CLI ookla speedtest. 2021-11-12 22:12:56 The machines are hardwired on a giga-bit ethernet 2021-11-12 22:13:21 I can only think of the kernel version making a difference 2021-11-12 22:13:43 yes, I would try change to edge kernel for test 2021-11-12 22:13:54 I like to revet to 3.14.2 or 3.14.1.... how can I do this? 2021-11-12 22:15:08 We do not keep packages for minor versions, the only thing that remains are the iso's 2021-11-12 22:15:39 you can somehow just try to find outdate mirror ;/ 2021-11-12 22:15:56 and stay unsecure, unupdatable until you die 2021-11-12 22:16:14 Marc0-1231: you should try add a kernel from edge and see if that changes it 2021-11-12 22:17:07 edge or revert to older kernel, depends on usage 2021-11-12 22:19:10 I like to revert to older kernel , but I'm not complete sure how to do this 2021-11-12 22:19:32 but yeah, as things are updated on mirrors, finding that older kernel is quite hard 2021-11-12 22:19:53 what are we now, 5.10.41 ? 2021-11-12 22:19:57 I look voor the edge-option then 2021-11-12 22:20:04 5.10.26 on 3.14 2021-11-12 22:20:05 it is easier to add a kernel from edge and see if changing only that will fix it 2021-11-12 22:20:05 Thanks for your help! 2021-11-12 22:20:16 if it does, you should make an issue 2021-11-12 22:20:24 Thanks 2021-11-12 22:33:14 Ok I change to EDGE 3.15-rc1 .... speed problems are gone 2021-11-12 22:33:30 interesting 2021-11-12 22:33:35 SO there is a provlem in 3.14.3 2021-11-12 22:34:22 so u did not change just kernel? but everything? 2021-11-12 22:35:00 helby... I am not so handy... I just change completly to 3.15-rc1 2021-11-12 22:36:29 Marc0-1231: what kind nics do you have? 2021-11-12 22:37:58 Weird... I do second speed test ...back to 30MB/s instead of 300mb/s.... a third linux machine not running Alpine... stays steady on 300mb/s 2021-11-12 22:39:04 I would try to connect from the oter machine wire 2021-11-12 22:40:24 Then both wires of of both alpines are broken? 2021-11-12 22:41:34 oh u experience this on two machines 2021-11-12 22:42:44 ikke's question is quite good to have, are both machines having same nic ? 2021-11-12 22:42:49 Yes... I have 3 linux machines here.... all on the same subnet.... same switch... same router... same ISP... 2 are alpine 1 is a different linux... both alpine run slow.... the other linux is fast 2021-11-12 22:43:30 No.. a is a intel NUC7 the other is a intel NUC8 2021-11-12 22:46:55 but those have same NIC chips, right? (haven't worked with NUC's for a while) 2021-11-12 22:49:15 they are intel nic chipsets on both 2021-11-13 09:46:48 Can't get network up on my laptop. iwd connects, obtains the lease, assigns ip address and creates the default route, but I can't even ping the router after that. 2021-11-13 09:51:13 The only thing that is different from the other laptop (with the same software setup) is that there is no "proto dhcp" in "ip route list", can it be the cause or symptom of smth? 2021-11-13 09:52:55 Do you have a default route? 2021-11-13 09:54:23 yes, but compared to the other laptop it lacks "proto dhcp" in its properties 2021-11-13 09:57:46 can you ping it? 2021-11-13 10:01:42 well, the issue just resolved itself, after leaving laptop idle for 30min :/ 2021-11-13 10:02:11 nothing changed 2021-11-13 10:02:27 thats creepy 2021-11-13 10:08:12 and the ping to the router in 33ms compared to 5ms from the other old crappy laptop with some ath chip onboard :/ 2021-11-13 10:09:19 I guess I should really replace this mediatek crap 2021-11-13 12:13:53 Hi everybody 2021-11-13 12:15:04 I'm trying to install wayland and dwl. I installed libinput, the problem is libinput list-devices show none devices 2021-11-13 12:15:55 libinput debug-events says Expected device events added... Maybe you don't have the right permissions. 2021-11-13 12:16:17 what do you thing it is a problem with eudev ?? 2021-11-13 12:26:22 captainchris: sudo libinput list-devices? 2021-11-13 12:35:49 vyivel no that does'nt work 2021-11-13 12:36:28 I'm trying to build kernel and set EVDEV_SUPPORT not in module but in fixe 2021-11-13 13:02:15 How can I set doas to use a different umask than the default umask of my regular user? I've set my regular user to use a umask of 027 but when I use doas to create files, those end up getting created with a umask of 027 as well. 2021-11-13 13:14:50 nvm, got it. I had to edit /etc/pam.d/doas and configure it to use pam_umask to set the umask to 022. 2021-11-13 14:17:34 do you have a range/mean time a package is usually in edge before getting to stable? 2021-11-13 14:18:23 frag: it will be in the next stable release as soon as it's moved to community or main 2021-11-13 14:18:59 ikke: okok 2021-11-13 14:19:35 So everything that is currently in main or community will be part of v3.15 2021-11-13 14:42:19 GoodAfter. Yesterday I came on here telling my both Alpine-boxes loosing bandwidth and the other nodes on my switch didn't . My thinking is was the both alpine-boxes which had problems... but in fact it was the switch. 2021-11-13 14:42:59 The switch seemed to have problems with the alpine boxes and not with the other nodes.... I replaced the switch and all is fine now. 2021-11-13 14:43:12 Just came on here to let you guys now. 2021-11-13 14:43:28 Thankks for al the support yesterday! It was appreciated!! 2021-11-13 14:44:27 The fact is strange that ONLY the alpine boxes were affected bij the faulty switch.... even when I switched ports on the switch... 2021-11-13 14:44:58 The other nodes were having 100% bandwidth... the alpine boxes were trottled 2021-11-13 14:45:16 I factory defaulted the switch... and stilll the alpines were trolled. 2021-11-13 14:45:25 **throttled 2021-11-13 14:45:42 The switch is replaced now.... al problems gone 2021-11-13 14:46:19 Thanks again! 2021-11-13 14:50:04 Marc0-1231: Nice that you found it 2021-11-13 14:50:08 and a relieve :) 2021-11-13 14:51:19 Yes Ikke. The things give me headaches,,, Because I want and like to know the problems... I had the drag my self to bed, because I wanted to fix it, but I needed the sleep.... this morning I figured it out. 2021-11-13 14:51:45 The working "none-alpines" gave me the wring clues. 2021-11-13 14:51:50 **wrong 2021-11-13 14:52:32 The didn't suspect the switch because of the other Linus nodes were fine, and I did swapped ports also on that switch 2021-11-13 14:55:09 I makes you wonder why the alpines nodes (which are 2 intel NUC) have the problem only... I guess a combination also with the Intel-hardware.... I don't thing it's alpine related 2021-11-13 14:55:35 **think 2021-11-13 14:56:55 Anyway thanks for the support yesterday as a group.... and have a wonderful day you all!! 2021-11-13 17:17:34 Hi everyone. Ariadne: I just noticed you pushed #12413 to the 3.15 milestone but you haven't reopened it. Was that intentional? 2021-11-13 18:04:28 Gaming4LifeDE: oopsie 2021-11-13 18:04:37 Gaming4LifeDE: btw have you considered sddm as a workaround 2021-11-13 18:39:15 do there exist some kind of policy/guideline for how new packages are supposed to be put together? Im thinking if I as maintainer should keep the default files of the original code or if I should apply some hardening (making life easier for users installing the package)? 2021-11-13 19:41:04 Hi everybody 2021-11-13 19:41:32 I'm trying to install wayland. but liibinput doesn't show devices 2021-11-13 19:42:32 libinput debug-events says Expected device events added ... Maybe you don't have the right permissions 2021-11-13 19:42:42 It is udev problem 2021-11-13 19:42:52 is udev running? 2021-11-13 19:42:58 yes 2021-11-13 19:43:13 try running udevadm trigger 2021-11-13 19:43:18 I recompile my kernel to set input event support 2021-11-13 19:44:18 udevadm trigger returns nothing 2021-11-13 19:44:48 does lininput find the devices now? 2021-11-13 19:45:40 nope 2021-11-13 19:46:00 I recompiled ^^ 2021-11-13 19:46:11 was evdev compiled as a module? 2021-11-13 19:46:32 I installed evdev witk apk 2021-11-13 19:47:07 if you were using linux-lts, it included evdev as a module 2021-11-13 19:48:00 Ok so it is not necessary to install it ? 2021-11-13 19:48:11 do you think it is the problem 2021-11-13 19:48:28 you do not need to install evdev separately. You do need eudev, which is different 2021-11-13 19:48:59 did you add your user to the input group? if you didn't then you should do that and log out 2021-11-13 19:50:55 yes my user is input group and seatd group^ 2021-11-13 19:51:14 do I remove eudev pkg from my system 2021-11-13 19:51:17 ?? 2021-11-13 19:51:21 no, keep eudev 2021-11-13 19:51:45 is seatd runnig? if not, are you using seatd-launch to start your wayland compositor? 2021-11-13 19:52:46 no I try with dwl, but I doesn't want to start it says libinput don't reconize devices ... 2021-11-13 19:53:11 and sway same issues 2021-11-13 19:54:48 did you run setup-udev ? 2021-11-13 19:54:49 raven523: remember to add user to seat group, not seatd 2021-11-13 19:56:21 I run mdev instead of eudev , so installed libudev-zero 2021-11-13 19:58:05 raven no i don't setup-udev 2021-11-13 19:59:01 I supose that i need to reboot 2021-11-13 19:59:04 ?? 2021-11-13 19:59:13 try rebooting 2021-11-13 19:59:23 Ok i come back 2021-11-13 19:59:34 well setup-udev remove 2 service from busybox and add eudev 2021-11-13 20:00:14 hwdriwers and mdev 2021-11-13 22:11:43 Would someone happen to know how I can install alpine linux using libvirt? In particular, how to configure the cdrom installer so that it uses the serial console? 2021-11-13 22:12:00 (As I don't have access to x11 right now) 2021-11-13 22:13:36 pretty sure alpine-virt comes with serial by default 2021-11-13 22:17:23 Hmm, maybe I should try downloading that one. 2021-11-13 22:20:38 Yep, that definitely works. 2021-11-13 22:31:11 Well, the live system is very nice. Will figure out libvirt networking and disks later. 2021-11-14 00:29:21 Is it normal for alpine linux to still tell root users about setup-alpine even after installation, or did I accidentally restart from the cdrom? 2021-11-14 00:31:50 arahael3: how is it telling you? 2021-11-14 00:32:07 When I log in, it says: You can setup the system with the command: setup-alpine 2021-11-14 00:32:25 And that I can edit motd to change that message. 2021-11-14 00:32:26 arahael3: scripts setup-alpine and others stays after installtion however it will not tells you about it, you are probably booted into cd 2021-11-14 00:33:01 Hmm, curious. I did power it off, and it seems to accept the new password when I power it back on. 2021-11-14 00:33:05 oh maybe it's in motd, don't remember 2021-11-14 00:33:22 Yep, that's what the motd has. 2021-11-14 00:34:09 oh yes, default motd has that text 2021-11-14 00:34:19 so don't worry about, you can type there whatever u like ;/ 2021-11-14 00:48:07 Ah, interesting, and makes sense, actually. 2021-11-14 00:48:39 I'm currently running nixos, and want to run a more "secure" host, but for now just got alpine in a VM. 2021-11-14 00:48:51 Eventually, I'll likely swap them around somehow 2021-11-14 03:56:43 Oh... Yikes... 'apk search' is damned fast. 2021-11-14 03:57:08 yes apk is speedy ;/ 2021-11-14 03:57:16 anything with apk is in fraction of sec 2021-11-14 03:57:47 only issie with apk search is, that's case sensitive ;/ if you use apk search -d /description/ 2021-11-14 03:58:07 Ah, good to know - most packages would be in lowercase, though, I'd expect. 2021-11-14 03:58:17 Hmm. no ghc in the package system. 2021-11-14 03:58:31 for regular search it does not matter as packages are lowercase 2021-11-14 03:59:08 ghc is there 2021-11-14 03:59:34 Hmm? Maybe my package selection is somehow limited... 2021-11-14 03:59:42 yes regular search is okay, but with flag '-d' you can search in description of packages, not only names, there it's an issue to be case sensitive 2021-11-14 04:00:11 bc. many things are Upper, like IRC or if it begans with that, like 'File manager' 2021-11-14 04:00:20 I might have only 'main'. 2021-11-14 04:00:27 did u uncomment community or maybe even testing? 2021-11-14 04:00:28 Ah, interesting. 2021-11-14 04:00:30 oh okay 2021-11-14 04:00:50 there is nothing in main practically ;/ so community is a must ;/ I assume 2021-11-14 04:04:20 Yep, ghc is indeed there. 2021-11-14 04:04:38 Just added community. Can't get over how *fast* it is. 2021-11-14 04:08:17 yes, I used before openbsd with pkg_add what here is done withing like 3 seconds there takes a minute 2021-11-14 04:08:52 are u on edge or 3.14? 2021-11-14 04:08:59 3.14 2021-11-14 04:09:13 oh okay, bc. on edge there is some crap with cabal and it does not work well 2021-11-14 04:09:29 I imagine that would be a temporary situation. 2021-11-14 04:09:45 edge is basically the bleeding edge, right? 2021-11-14 04:09:57 yes, kind of rolling 2021-11-14 04:10:39 Oh, do I need 'community' in order to run vm's and stuff? (Ie, how do I check if 'main' has everything I need for libvert, qemu, and the like? 2021-11-14 04:10:44 *libvirt 2021-11-14 04:10:51 Now that I'm on community, anyway. 2021-11-14 04:12:03 well if package is in main then also all of it's required dependencies should be there 2021-11-14 04:13:30 that crap will be all in community I assume 2021-11-14 04:14:01 Oh, awesome. My wife just found my old RPi 3. That could make a temporary server whilst I migrate my main server. 2021-11-14 04:14:40 Just need to get Alpine onto it! 2021-11-14 04:15:53 of course, put alpine everywhere, even visit your neighbour ;/ 2021-11-14 04:18:55 Yeah, the difficulty is how do I physically put alpine on it :D 2021-11-14 04:19:15 Need to find an sd card reader. 2021-11-14 04:21:25 or run from ram ;/ 2021-11-14 04:23:50 helby: Gotta get it on there first! 2021-11-14 04:23:59 Ok, found an old, ancient chinese sdcard reader. 2021-11-14 04:25:06 u have probably a lot of things ;/ u found rpi 3, sdcard reader just looking around ;/ 2021-11-14 04:26:01 Well, the SD Card reader is like, ,15 years old. 2021-11-14 04:27:15 Hmm. Wonder how I can get this onto the sd card from *macos*. 2021-11-14 04:27:29 Do we have 'pre-installed' alpine images for rpi? 2021-11-14 04:27:53 I mean, such that I can just 'cp' it over to the sdcard and Done! It's installed? 2021-11-14 04:32:09 search wiki, there is probably some fuckery for rpi 2021-11-14 04:47:06 That's a good idea. 2021-11-14 04:48:39 Ok, looks like I'll have to this from my current linux system, which is _also_ headless. Another day, but will put it with that system for now. 2021-11-14 04:52:45 Heh, the original instructions came from someone called "sodface". 2021-11-14 04:55:27 he is still active in wiki 2021-11-14 04:56:28 I was looking for him yesterday on irc, but did not find him ;/ 2021-11-14 05:29:59 arahael3: if the RPI is intended as a server you could use my script ot create an image ready to be "dd"ed onto a SDcard: https://github.com/dermotbradley/create-alpine-disk-image 2021-11-14 05:33:57 minimal: I'd prefer to avoid it, but it's certainly an option - does your script require root? Just giving it another readover. 2021-11-14 05:34:28 I suppose I could run it inside the new alpine VM I have. 2021-11-14 05:35:25 arahael3: the "main" script can be run as any user as all it does it create a "secondary" script that does the real work. The "secondary" script does need root access as it manipulates loop devices, creates and mounts disk images, etc 2021-11-14 05:36:12 That secondary script runs within the chroot, from the looks of it? 2021-11-14 05:36:25 I notice you're using dropbear in it - any particular reason why? 2021-11-14 05:36:34 arahael3: yes that would work - always a good idea not to run "random" scripts from the Internet without some form of restricting them :-) 2021-11-14 05:36:53 minimal: You understand why I'm so reluctant. :D 2021-11-14 05:37:34 arahael3: the secondary script doesn't run within chroot, it creates a chroot env and then runs commands within that 2021-11-14 05:38:01 Ah, I see. 2021-11-14 05:38:11 Why'd you do it all in sh? 2021-11-14 05:38:40 arahael3: re dropbear, out-of-the-box it uses openssh, I started work to support dropbear and tinyssh as alternatives but that's a work-in-progress as cloud-init only supports openssh so I'm wtill working on patches to it for the others 2021-11-14 05:39:09 Ah, fair enough. 2021-11-14 05:39:14 arahael3: its in shell to make it "portable" - my previous version was all in Ansible 2021-11-14 05:39:26 Ah, fair enough. 2021-11-14 05:39:58 I do prefer sh to python - used to be a major python fanboy when I was a kid. Now I can barely stand the language. 2021-11-14 05:40:42 arahael3: its overdue another push to sync it up with my current local version 2021-11-14 05:41:58 Nice. :) 2021-11-14 05:42:16 Testing a script like this must be painful. I notice you use shellcheck - very nice. 2021-11-14 05:42:25 Nice haskell utility, if I remember correct. :) 2021-11-14 05:42:51 ACTION is back in 5 min... 2021-11-14 05:44:23 arahael3: yeah shellcheck will only pick up errors in the "outer" portion of the script, not the "script-within-a-script" portions. Plus I'm testing against an ever growing set of physical/VMs and filesystems and other options too 2021-11-14 05:50:09 minimal: Very nice. :D 2021-11-14 05:51:18 arahael3: I find it useful but the main script is getting soooo long as I add more and more features 2021-11-14 05:52:39 I build RPI images, and x86_64 QEMU and Virtualbox images (BIOS & UEFI) on a regular basis and physical PC images less often 2021-11-14 05:53:11 Physical PC's are a pain, anyway. 2021-11-14 05:55:39 in what way? As I'm using cloud-init with a small partition for its YAML config files I just need to "dd" the image onto a disk and edit the YAML files to set hostname, IP, timezone, etc before 1st boot and then it all get configured 2021-11-14 05:55:54 same for RPIs 2021-11-14 05:59:17 Well, I'm more a software guy, really, not an admin. 2021-11-14 10:44:29 how can i change home folder for my user? just change it in /etc/passwd? 2021-11-14 11:02:43 hi, am I right with my assumption that clawsmail 4.x (currently in edge ) is in the next release? 2021-11-14 18:25:01 frag: correct 2021-11-14 18:25:28 yes it was.. 2021-11-14 18:25:45 ah actually i installed shadow 2021-11-14 18:25:49 frag: there is no any other utiliy in base for that 2021-11-14 18:26:01 well yes, that's option 2021-11-14 18:26:27 but practically it does the same 2021-11-14 18:27:23 ah ok 2021-11-14 18:27:41 prob wont move it again too soon :p 2021-11-14 18:27:45 bc it's not mentioned on other place in system, just in passwd file 2021-11-14 18:27:57 like login shell ... 2021-11-14 18:28:08 mhmh 2021-11-14 19:46:44 hm. i just wrote my (second) APKBUILD - everything goes well, post-install gets executed, but it seems that "rootpkg failed".. how could i debug this? 2021-11-14 19:49:02 dekenevs: can you show the log output? 2021-11-14 19:54:10 ikke: ah it seems that i chown'ed smth non-existing. btw, how are directory conflicts solved? i.e: is it okay if I package /var/spool/news, that directory gets populated, and later on an update, will installing the package complain that the directory already exists? 2021-11-14 19:54:29 no 2021-11-14 19:54:36 perfect ^^ 2021-11-14 19:54:48 otherwise only one package could install something in /usr/lib or /usr/bin for example 2021-11-14 19:54:58 true 2021-11-14 19:55:12 one thing i like about apkbuilds is that they're almost identical to arch's pkgbuilds 2021-11-14 20:05:18 because apkbuild was plagiarized^Winspired by pkgbuild 2021-11-14 20:07:52 (nothing wrong with that) 2021-11-14 20:08:43 i honestly like it. plus other things about alpine ofc 2021-11-14 21:28:29 Hey. Just wondering about the status of a package. So kopano-core is in Edge testing repo. But, what's the state of it for getting to main repo and release? As in, is it expected to make it for 3.15, and if so, will it remain in testing repo then? Or will it not be ready in time but perhaps in 3.16? Or is it one of those packages that will forever be stuck in testing like on Ubuntu where it has been stuck on 8.7 for ever and ever >_< 2021-11-14 21:28:56 things in testing don't make it into releases 2021-11-14 21:29:12 so unless something moves it to community and it gets merged 2021-11-14 21:31:50 Oh so only edge even has a testing repo? 2021-11-14 21:31:56 yes 2021-11-14 21:32:18 Ah right. That actually explains a lot. Thanks. 2021-11-15 00:58:42 psykose: any idea why your changes to foot/terminfo was not merged to foot? 2021-11-15 00:58:56 because nobody got around to merging it 2021-11-15 00:59:31 nah, but now foot got updated to next version without it :( 2021-11-15 01:12:57 sure, not much i can do about that though except rebase the changes 2021-11-15 01:13:55 so maintainer or whoever take care of packages don't check merge request for packages he/she manage? 2021-11-15 01:14:11 it's like just a luck that maybe he find it? 2021-11-15 01:14:38 the maintainer can't merge the package anyway 2021-11-15 01:16:27 but whoever can and do it should not check request or issues or whatever before he does something? 2021-11-15 01:16:57 no idea how that works or does not work 2021-11-15 01:18:16 there's hundreds of merge requests, so if the thing you mr is not maintained by someone with merge access, or isn't seen by someone with merge access, or you don't ask someone to take a look at it, it takes a while 2021-11-15 01:18:32 i don't like bothering people when i run the whatever changed patched version locally anyway, so no difference for me 2021-11-15 01:19:56 yes I build few crap localy too, but it takes out 'simplicity' what alpine offers 2021-11-15 01:20:37 bc. I can imagine to build something if want extra feature, but if want to remove crap ... 2021-11-15 01:22:25 bc. something like that will get lost forever no? or somebody somehow discover it later ? 2021-11-15 01:36:27 I assume that 'Maintainer' in APKBUILD for that package and most of other is probably leftover 2021-11-15 01:36:54 I feel there is kind of anarchy 2021-11-15 01:36:57 ;/ 2021-11-15 08:44:04 Hmm, I followed the instructions at https://wiki.alpinelinux.org/wiki/Raspberry_Pi_-_Headless_Installation for my rpi3, but it doesn't work. 2021-11-15 08:50:57 I wonder if the 64-bit image doesn't work. 2021-11-15 09:05:07 arahael3, what do you mean by doesn't work? 2021-11-15 09:05:14 I personnally use the minirootfs for my Pi images 2021-11-15 09:06:51 markand: Unfortunately I lack the means to troubleshoot it. I followed those instructions, then put the SD card in, plugged it into ethernet and then power, and... 2021-11-15 09:07:05 markand: The 'PWR' light is on. That's all I can tell. 2021-11-15 09:13:29 markand: I stuffed it up, let me try fixing it! 2021-11-15 09:23:05 Ok, it's better now, not on the LAN yet, though, but it's getting there. 2021-11-15 09:37:28 All good. Alpine is very sweet :D 2021-11-15 09:38:05 As for the mistake I made, well, I figured I should just re-check the installation, so popped the sdcard back into the other system and mounted the partition. 2021-11-15 09:38:09 It was... Empty... 2021-11-15 09:38:28 I had written the files to a random other partition. :( 2021-11-15 16:26:13 wol 2021-11-15 16:49:08 trying out Alpine Linux "edge" on my Dell Inspiron 3558 laptop, it instantly experiences "hard lock-up" when attempting to enter the GUI via "startx", kernel version "5.15.2-0-lts" 2021-11-15 16:54:33 https://gitlab.alpinelinux.org/alpine/aports/-/issues/13163 2021-11-15 16:54:34 Might be related to what I 2021-11-15 16:54:38 experience here 2021-11-15 16:55:28 Maybe try side-loading the linux-lts package from Alpine 3.14 (5.10.78-r0) 2021-11-15 16:55:29 https://dl-cdn.alpinelinux.org/alpine/latest-stable/main/x86_64/linux-lts-5.10.78-r0.apk 2021-11-15 16:55:52 I personally can't boot reliably with anything newer than 5.11.22 on my Celeron N3450 laptop 2021-11-15 16:56:13 IDK how related it is since it's an experience from a different distro, but Debian "sid" with the 5.14.x kernel worked just fine on this laptop, perfectly stable over a 8+ hour boot period 2021-11-15 16:56:32 also stressed with things like compiling lots of packages and "heavy" (read: Electron) apps like Discord 2021-11-15 16:57:01 so IDK if it's an Alpine-specific issue or if 5.15.x introduced a new regression that b0rked its support for my laptop's Intel graphics chip 2021-11-15 16:57:04 Can you diff the loaded modules or the kernel config? 2021-11-15 16:57:24 in both cases I have used the modesetting driver, not the intel one (which by now has tonnes of issues) 2021-11-15 16:57:25 Saijin_Naib: does "can't boot" mean it stops at the "Loading initramfs" part? 2021-11-15 16:57:57 I guess I could test with Arch, Debian, et al (that have 5.15 kernels available) and see if they also crash 2021-11-15 16:58:04 minimal: No, it'll get past that 99% of the time and then just hang/crash in DE from 0s-3min, or in console session somewhere around that long (up to 10min) 2021-11-15 16:58:26 Saijin_Naib[m]: my console sessions don't crash, just the GUI 2021-11-15 16:58:46 Ah, then I guess your intuition that it is the graphics config seems correct 2021-11-15 16:58:54 what's your CPU family? 2021-11-15 16:58:57 My issue must be different. Sorry 2021-11-15 16:58:57 mine is Broadwell 2021-11-15 16:59:21 Apollo Lake/Silvermont/Broxton 2021-11-15 16:59:24 MrTripleT: Alpine "Edge" is by its nature not guaranteed to be stable and there have been a few recent issues since the update to 5.15.x kernel and the introduction of kernel module compression 2021-11-15 16:59:58 the fact that other distros with 5.14 ran just fine on it leads me to think that it might be another Intel graphics kernel driver screw-up 2021-11-15 17:00:07 like I experienced between 5.4.x and 5.7.x on various distros 2021-11-15 17:01:16 hmm, I see from other distros' bugtrackers that 5.15 might not be all it's cracked up to be 2021-11-15 17:01:48 I guess I might want to go back to the previous kernel LTS (5.10.xx IIRC) until things get closer to "normal" with 5.15 or 5.16 2021-11-15 17:02:13 MrTripleT: there's a couple of issues related to initram not including firmware files that I raised MRs for over the weekend, one of them has only been in the past hour and the other is not yet merged. 2021-11-15 17:02:47 do you think that kernel "sideloading", like Saijin mentioned, would work on "edge"? 2021-11-15 17:02:48 s/only been/only been merged/ 2021-11-15 17:02:49 minimal meant to say: MrTripleT: there's a couple of issues related to initram not including firmware files that I raised MRs for over the weekend, one of them has only been merged in the past hour and the other is not yet merged. 2021-11-15 17:03:08 or should I go back to 3.14.3 (i'm only trying "edge" due to needing newer versions of some apps) 2021-11-15 17:03:38 MrTriple: the issues I'm referring to are not kernel issues, they're issues with mkinitfs and alpine-conf 2021-11-15 17:03:55 I've been sideloading my 5.11.22 compile for the past week or so. Biggest issue is that issuing an apk upgrade -al will overwrite it with the 5.15.x releases, and then I just load again. Doesn't seem to break any dependencies 2021-11-15 17:04:13 is it possible to "pin" package versions, like you can do in other package mangers? 2021-11-15 17:04:19 s/mangers/managers 2021-11-15 17:04:19 MrTripleT meant to say: is it possible to "pin" package versions, like you can do in other package managers? 2021-11-15 17:05:28 yes, I belive its something like "apk add linux-lts:5.11.22-r0 (or whatever the package version is) 2021-11-15 17:06:00 https://wiki.alpinelinux.org/wiki/Alpine_Linux_package_management#Repository_pinning 2021-11-15 17:06:34 I don't think it applies to local packages, though. Or I'm doing it wrong, which is likely 2021-11-15 17:06:46 also just noticed the CAPS LOCK light isn't working, lol 2021-11-15 17:10:00 what the hell, it immediately freezes up on 5.10 as well 2021-11-15 17:10:13 which it doesn't on other distros 2021-11-15 17:10:26 maybe it's not the kernel, maybe there's a driver fsck-up somewhere 2021-11-15 17:10:26 Do you have anything custom in your GRUB cfg? 2021-11-15 17:10:31 no, it's "stock" 2021-11-15 17:10:35 other than the downgrade to 5.10 2021-11-15 17:10:54 Yeah, I'm well beyond my familiarity at this point then. I'm sorry ☹️ 2021-11-15 17:10:55 I only ever mess with the config if using LILO on my other non-Alpine systems 2021-11-15 17:11:56 libinput is installed, as is the driver based upon it 2021-11-15 17:13:28 MrTripleT: freezes during boot or once booted? 2021-11-15 17:13:44 neither, when attempting to start XOrg and WindowMaker via "startx" 2021-11-15 17:13:56 my .xinitrc is literally just "exec wmaker" 2021-11-15 17:14:16 if you reboot is there anything in the Xorg logfile showing what's happening? 2021-11-15 17:15:03 MrTripleT: LILO? haven't seen that in decades lol 2021-11-15 17:16:06 it's available in the Gentoo repos (Gentoo is the other distro I use) 2021-11-15 17:16:40 anyways, here's a line that might be salient to this discussion: "(EE) xf86OpenSerial: Cannot open device /dev/input/event0 Permission denied." 2021-11-15 17:17:14 Can you add your user to video/audio/input? 2021-11-15 17:17:43 they're in video and audio 2021-11-15 17:17:50 assumed input was a default group, like in other distros 2021-11-15 17:17:56 Although, that might not help... 2021-11-15 17:18:01 Is this the rootless Xorg issue thing? 2021-11-15 17:18:07 IDK 2021-11-15 17:18:07 Where X can't get access to the TTY? 2021-11-15 17:18:15 Shoot, I think mps just patched this 2021-11-15 17:18:28 what would I look for in Xorg.log if it's a TTY fsck-up 2021-11-15 17:18:41 😬 2021-11-15 17:19:08 cuz i'd be relieved if it's just not being in the input group (i'd prefer it involving me being a dumb-dumb than the kernel driver being broken again) 2021-11-15 17:19:17 or rootless Xorg 2021-11-15 17:19:47 O thank $DEITY, the input group thing fixed it 2021-11-15 17:20:44 now i'm gonna play around for a couple hours just to make SURE there's no kernel driver issues 2021-11-15 17:22:04 on another topic, i'm wondering if Alpine would be a good choice to resurrect my mom's circa-2005 laptop 2021-11-15 17:22:12 with 1GB of RAM and ~80GB HDD space 2021-11-15 17:22:44 quite possibly 2021-11-15 17:22:56 is she wont be using any web browser then why not 2021-11-15 17:23:03 s/is/if 2021-11-15 17:23:03 MY-R meant to say: if she wont be using any web browser then why not 2021-11-15 17:23:19 she likes to browse the web, so it might be a browser like NetSurf 2021-11-15 17:23:31 although last time I used NetSurf for anything involving JS it rapidly segfaulted 2021-11-15 17:24:24 and she likes to play old games that were WinXP-only, and IDK if 1GB of RAM is enough for Wine to work smoothly 2021-11-15 17:24:41 even with a hyper-minimal distro (in terms of resources) like Alpine 2021-11-15 17:30:04 wine uses hardly any ram by itself. "wine is not an emulator" 2021-11-15 18:14:55 MrTripleT: 2005 is too new, I am using T420, it's from 2001 and it flies with alpine ;/ 2021-11-15 18:16:04 with few tabs I am under 1GB of RAM usage with firefox btw 2021-11-15 18:16:24 but 1 is probably limiting, 2 would be okay 2021-11-15 18:17:01 yes, netsurf is practically not usable, there is btw. new netsurf kid 'visurf' ;/ 2021-11-15 18:19:38 my mum's not really used to keyboard-based UIs anymore (even though she grew up with DOS, it's prob been 16+ years since she's used a TBUI in any meaningful capacity), so IDK if visurf would really 'work out' for her 2021-11-15 18:20:18 no it will not, I just mentioned it as just tried it yesterday 2021-11-15 18:20:35 but netsurf does not work for 'regular' modern browsing 2021-11-15 18:22:00 but still I play youtube music in firefox, opened few tabs and all system usage is under 700mb of RAM 2021-11-15 18:22:24 so maybe se can use it without problems 2021-11-15 18:23:46 not sure how webkit based browsers performs, but for me it's not usable, it's somehow very slow, no idea why 2021-11-15 19:16:59 so, i came across an article about vulnerability's i busybox, anyone know? 2021-11-15 19:18:08 https://heimdalsecurity.com/blog/14-security-vulnerabilities-found-in-busybox/ oa this one 2021-11-15 19:18:12 jhjacobs81: I think if you check release notes for 3.14.3 there is mentioned fix 2021-11-15 19:19:04 https://alpinelinux.org/posts/Alpine-3.14.3-released.html 2021-11-15 19:19:52 not sure why so late if these were fixed in busybox long ago, but I have no idea how all that security cvs nonsense works 2021-11-15 19:24:59 > why so late 2021-11-15 19:24:59 as fast as possible I think. articles about those vulnerabilities appeared last week. some systems are not fixed yet, alpine is fixed. 2021-11-15 19:28:42 alpine team added all needed patches. it took about 1 day. and patched alpine was released after that. 2021-11-15 19:28:49 ahhh okay :) 2021-11-15 19:29:08 article? but busybox got fixed like a year ago 2021-11-15 19:29:16 helby: but no backports 2021-11-15 19:29:22 helby ok thanks :) im not surewhich releasenotes i was supposed to read 2021-11-15 19:29:28 We had to backport it to older versions ourselves 2021-11-15 19:29:38 helby check the link i posted, article apeared siz days ago 2021-11-15 19:29:57 jhjacobs81: but article is post, somebody write about, busybox knew and fixed it long time ago 2021-11-15 19:30:00 https://alpinelinux.org/posts/Alpine-3.14.3-released.html 2021-11-15 19:30:09 it does not matter what somebody write on the internet 2021-11-15 19:30:23 helby: before, the vulnerabilities were not disclosed 2021-11-15 19:30:27 its not the only site :) 2021-11-15 19:30:37 there are more sites that wrote about it in the past days :) 2021-11-15 19:30:48 This is the source: https://jfrog.com/blog/unboxing-busybox-14-new-vulnerabilities-uncovered-by-claroty-and-jfrog/ 2021-11-15 19:30:53 hence me asking :) 2021-11-15 19:31:01 keep in mind that there may be some kind of Disclosure Policy 2021-11-15 19:31:13 i wouldn't ask it if it was some person's personal opinion about busybox ;) 2021-11-15 19:32:09 thanks ikke :) i follow multiple security sites on my RSS reader :) i saw more sites popping up 2021-11-15 19:32:59 i was a bit preoccupied with learning Rust, so i forgot about it :) 2021-11-15 19:33:35 so busybox fixed it a year ago and ... world was keep vulnareable one year ... cool how it works 2021-11-15 19:34:09 well, if it was fixed a year ago, and Alpine included the fixes, then i care not about the rest ;-) 2021-11-15 19:34:11 that's how disclosure works 2021-11-15 19:34:28 > All vulnerabilities were privately disclosed and fixed by BusyBox in version 1.34.0, which was released Aug. 19. (c) from jfrog.com 2021-11-15 19:34:47 but then i wonder.. how did Alpine get these fixes? 2021-11-15 19:35:03 and it took time while fixed version spreads. after some time info became public 2021-11-15 19:35:21 ofcourse im just curious, because what do i know :) 2021-11-15 19:38:04 oh well, i read its all fixed 2021-11-15 19:38:28 jhjacobs81: What happened: busybox learned (privately I suppose) about vulenaribilties, and fixed then in the latest version 2021-11-15 19:38:36 We upgraded in edge to that version 2021-11-15 19:38:53 oh I understood it wrong, that it was fixed long time ago, but it was in 2021, not 2019 ;/ 2021-11-15 19:39:11 later, jrog disclosed the vulnerabilities 2021-11-15 19:39:35 We heard about it, and ncopa figured out what commits these fixes were done in and backported to the relevant versions in alpine linux 2021-11-15 19:40:02 And we made new releases to include these fixes (mostly because our docker images are based on releases) 2021-11-15 19:40:18 ahhh :) 2021-11-15 19:40:25 here're backports 2021-11-15 19:40:25 https://git.alpinelinux.org/aports/log/main/busybox?h=3.12-stable 2021-11-15 19:40:25 https://git.alpinelinux.org/aports/log/main/busybox?h=3.13-stable 2021-11-15 19:40:25 https://git.alpinelinux.org/aports/log/main/busybox?h=3.14-stable 2021-11-15 19:40:35 and you are of the programmer team? 2021-11-15 19:40:49 Who is you/ 2021-11-15 19:40:51 ? 2021-11-15 19:40:57 you and ncopa? 2021-11-15 19:41:04 because you say "we" 2021-11-15 19:41:05 :D 2021-11-15 19:41:10 We as alpinelinux 2021-11-15 19:41:15 ohh okay 2021-11-15 19:41:29 im very new to Alpine, so im still learning who is who :) 2021-11-15 19:43:18 ncopa started alpinelinux 2021-11-15 19:43:26 awesome :) 2021-11-15 19:44:04 im always so jealous of people who start great things :) i can barely start breakfast :O 2021-11-15 20:52:38 Don’t underestimate a good breakfast:) 2021-11-15 20:53:30 especially if you wake up at 7pm 2021-11-15 21:19:35 :D 2021-11-15 22:14:18 hello 2021-11-15 22:24:59 bye 2021-11-15 22:26:14 ;/ 2021-11-15 22:32:33 What differences are there between alpine-make-vm-image and create-alpine-disk-image? Has this question already been answered somewhere? I vaguely remember seeing this discussed in this channel a while back. 2021-11-15 23:07:26 xordspar0: alpine-make-vm-image is an official Alpine script whereas create-alpine-disk-image is my own personally-authored unofficial script 2021-11-15 23:09:34 Sure, but there must be some cool tricks your script can do that make-vm-image can't. :) 2021-11-15 23:10:29 or opposite ;/ 2021-11-15 23:10:38 xordspar0: yupe, its just not for creating VM images, also for physical machines, its got options for filesystem types, encryption, LVM, etc 2021-11-15 23:11:11 its using cloud-init to do initial OS configuration etc 2021-11-15 23:23:23 hello all 2021-11-15 23:23:48 someone can help me to guide me install alpine linux 2021-11-15 23:34:52 Ganonk: some issue with installer? 2021-11-15 23:36:27 Is it possible install alpine-standard.ISO from directly HDD ? 2021-11-15 23:37:18 currently i'm using Devuan linux 2021-11-15 23:37:30 dont have usb that want to complicate things? 2021-11-15 23:38:12 i want install to NAT vps , but ucdhcp failed 2021-11-15 23:39:32 any solutions?.. 2021-11-15 23:47:23 I think you need to elaborate so that people better understand what you're trying to do 2021-11-15 23:59:54 this is result i'm install alpine via Netboot.xyz on NAT vps https://paste.pics/EY9MD 2021-11-16 00:47:33 hey, I just fresh installed Alpine and I'm running into an issue with LXD. I think cgroups have been disabled, is that a new default? how can I enable them, using rc.conf? 2021-11-16 00:50:14 am I better off installing Alpine 3.14 and then using LXD through @edge repository or directly installing edge? I've previously installed the edge edition when I was using Alpine in August and I didn't encounter any issues. 2021-11-16 01:38:12 https://www.youtube.com/watch?v=ykT_g0-bgZo&t=44s 2021-11-16 05:21:02 So is there a good SIP softphone available via apk? 2021-11-16 05:21:57 LinPhone seems like the expected default but I get a "symbol not found" error when executing linphonec. 2021-11-16 07:12:24 "Docs" link on the home page is going to 404 at: https://docs.alpinelinux.org/ 2021-11-16 07:13:54 how to join mailing list? 2021-11-16 07:15:02 LInk to the mailing list can be found on this page: https://www.alpinelinux.org/community/ but it points to https://lists.alpinelinux.org/lists/%7Ealpine which is 404 too 2021-11-16 07:15:47 Registration is currently closed 2021-11-16 07:15:54 very sad 2021-11-16 07:16:13 not sure, looks like a server down issue to me at a guess. 2021-11-16 07:16:27 all of those load for me 2021-11-16 07:16:33 i'm working fine 2021-11-16 07:17:11 really i just get "404 Not Found nginx" 2021-11-16 07:17:15 mailing list subscription doesn't need 'registration' 2021-11-16 07:17:37 the subscribe button on the lists is just a mailto: 2021-11-16 07:17:39 Ganonk: You need to subscribe to the list 2021-11-16 07:17:40 afaik you can send whatever 2021-11-16 07:17:49 You do not need an account 2021-11-16 07:21:08 btw, anyone can help me : how to boot alpine-standard.ISO from hdd directly? 2021-11-16 07:21:41 is it possible ? 2021-11-16 07:26:37 ncopa , ikke : can you help me 🙏 2021-11-16 07:27:19 good morning Ganonk. everything is possible. the impossible just takes a bit more time 2021-11-16 07:27:34 Ganok: what do you need help with? 2021-11-16 07:28:08 You should be able to boot from the ISO on your HDD if you flash it there as it should act like any other boot device, but why? 2021-11-16 07:29:33 yes, you can dd if=alpine-standard.iso of=/dev/sdaX and that should boot 2021-11-16 07:29:52 but it would still be iso9660 filesystem, which is readonly 2021-11-16 07:30:31 do you boot with EFI or legacy bios? 2021-11-16 07:31:13 i will install alpine-standard on NAT vps (kvm) 2021-11-16 07:33:17 you should be able to specify the iso image as a disk 2021-11-16 07:34:36 wait 2021-11-16 07:35:30 this is result install alpine via Netboot.xyz on NAT vps https://paste.pics/EY9MD 2021-11-16 07:35:56 what must i do? 2021-11-16 07:40:51 looks like it fails to get any ip address 2021-11-16 07:41:02 i have never used netboot.xyz 2021-11-16 07:41:13 do you have a dhcp server set up? 2021-11-16 07:41:42 I guess you could also try https://boot.alpinelinux.org/ if you need net boot 2021-11-16 07:42:04 that's why i tried to boot alpine-standard.ISO directly from hdd 2021-11-16 07:42:27 should work to boot alpine-standard.iso from hdd 2021-11-16 07:43:36 hey anyone know if cgroups are disabled by default now? 2021-11-16 07:43:48 but I don't really understand the steps to boot an iso from a direct hdd 2021-11-16 07:44:08 DanielYrovas[m]: yes they are. you need to `rc-update add cgroups boot` to enable them 2021-11-16 07:44:23 ahh ok 2021-11-16 07:44:52 Ganonk: that depends on the machine you are booting. if its qemu, vmware, or a physical machine 2021-11-16 07:45:05 qemu 2021-11-16 07:45:26 do you think I am better of using 3.14 and using @edge repository to install LXD and linux-edge, or just direct install alpine edge? 2021-11-16 07:45:41 qemu-system-x86_64 -enable-kvm -hda /path/to/alpine.iso 2021-11-16 07:46:40 DanielYrovas[m]: depends on what you are using the machine for. It is generally not recommended to mix 3.14 with edge repositories though 2021-11-16 07:47:00 alpine standard OR alpine extended ? 2021-11-16 07:47:10 Ganonk: both should work 2021-11-16 07:47:33 however, im testing it here and the kernel module for the hdd does not autoload for some reason 2021-11-16 07:47:48 yeah ok, the LXD wiki entry recommends enabling them on an app by app basis but it didn't seem like the greatest idea to me. Thanks for an amazing operating system btw, I've loved running it as a host os for quite a few months now. 2021-11-16 07:51:30 hum.. alpine-extended.iso does not boot with the iso as disk 2021-11-16 07:54:08 I've run into a few issues with bricking my install recently, once by enabling lxcfs at runlevel default, and it would hang the system on startup, and another by running grub-mkconfig. Is there a recommended way to regenerate initramfs, like mkinitcpio from arch? 2021-11-16 07:55:21 ah, qemu-system-x86_64 -hda alpine.iso does work. just need enough ram. 2021-11-16 07:55:33 apparently it does not boot with only 128MB 2021-11-16 07:55:49 DanielYrovas[m]: mkinitfs 2021-11-16 08:49:02 what is a safe way to ask grub to update? I have set GRUB_DISABLE_OS_PROBER=false. 2021-11-16 11:12:48 Truely bizarre. I can ssh to my rpi service from my linux server, but not from my macos laptop. It's as if it's no-longer even reaching the box. 2021-11-16 11:14:01 Weirder: I can ssh from one terminal tab on the macos system, but not another! (This is clearly a macos issue, then) 2021-11-16 11:23:30 *sigh*. Was a clear case of PEBKAC. 2021-11-16 11:32:46 !ping 2021-11-16 11:43:13 ncopa : hello Sir, are you still online? 2021-11-16 11:43:26 Ganonk: just ask your questions 2021-11-16 11:44:53 ikke : this is "qemu-system-x86_64 -enable-kvm -hda /path/to/alpine.iso" syntax command , or what? 2021-11-16 11:45:07 i'm newbie 2021-11-16 11:49:03 yes it is the emu syntax. but I think you also need -m 768 to give the vm a bit more than the 128MB default ram 2021-11-16 11:51:49 where do i start? 2021-11-16 11:51:57 i hope you are willing to guide me 2021-11-16 11:53:04 im not sure what you are trying to do 2021-11-16 12:01:03 currently i'm using Devuan linux. on panel vps only netboot.xyz. 2021-11-16 12:35:21 Ganonk: there is also https://boot.alpinelinux.org/ 2021-11-16 12:35:30 not sure it has the latest version 2021-11-16 14:11:49 I'm running "Linux raspi 5.10.61-0-rpi2 #1-Alpine SMP Thu Aug 26 22:16:53 UTC 2021 armv7l GNU/Linux" (alpine version 3.14-2 on a raspi 3 B). After an upgrade to alpine 3.14-3 (see upgrade log on https://dpaste.org/ROjO) my raspi didn't boot anymore. I've to restore to alpinelinux 3.14-2 from my backup to boot again. 2021-11-16 14:14:18 Any suggestion what is going wrong here ? 2021-11-16 15:24:16 what does happen at boot? 2021-11-16 15:26:29 something went wrong with the kernel likely 2021-11-16 16:15:48 lopid: The green led just flashed for a second and then nothing more. As this raspi is a headless one, I can't say something about the boot process. 2021-11-16 16:16:09 is that a sys install 2021-11-16 16:20:54 Yes it is. But my /var/log is mounted on tmpfs to take care of my sdcard. 2021-11-16 16:21:44 seeing as how it doesn't boot and it's headless, then you need to check that you're getting the kernel you think you're getting at boot 2021-11-16 16:22:09 mount the sdcard, etc 2021-11-16 16:25:17 kernel + firmware, i meant to write. 2021-11-16 16:25:56 rpi sys installs seem to break in this way with some people 2021-11-16 16:33:49 Although with me it is not a new rpi sys installation but only an update? 2021-11-16 16:35:44 Restoring my backup (alpine version 3.14-2) on the sdcard works flawlessly. 2021-11-16 16:51:34 freeit: is your /boot mounted as fat? 2021-11-16 16:52:31 freeit: maybe after your upgrade, look at config.txt on the sdcard and see what's there, that gets overwritten on upgrades 2021-11-16 16:53:10 make sure it matches what you've got 2021-11-16 16:54:21 invoked: does config.txt get modified? From memory there are no Alpine packages that create/modify it, it is only created by the script to build RPI images 2021-11-16 16:54:37 yes. 2021-11-16 16:55:07 the default file says so ("do not modify this file as it will be overwritten on upgrade") 2021-11-16 16:55:58 invoked: just double-checked, raspberrypi-usbboot is the only package with a config.txt and its not relevant for "normal" RPI deployments 2021-11-16 16:56:18 invoked: so what updates config.txt then? 2021-11-16 16:56:38 hm. i dunno, i was just poking around for things they could check 2021-11-16 16:56:51 upgrading diskless setup when using generated images :P 2021-11-16 16:57:33 but ye on sys install my config.txt was never touched 2021-11-16 16:58:49 there was probably a firmware update as part of it, correct? 2021-11-16 16:59:05 (i don't have an alpine rpi to look at) 2021-11-16 16:59:13 setup-disk creates config.txt 2021-11-16 16:59:31 I didnt notice "raspberrypi-bootloader" package upgrade in freeit logs 2021-11-16 16:59:55 if he did sys install manualy from wiki then that package isnt installed at all 2021-11-16 17:02:33 MY-R: Typing "mount" on my crrently running alpine 3.14-2 raspi: /dev/mmcblk0p1 on /boot type vfat (rw,relatime,fmask=0022,dmask=0022,codepage=437,iocharset=ascii,shortname=mixed,errors=remount-ro) 2021-11-16 17:05:45 freeit: does "apk info -ve raspberrypi-bootloader" return something? 2021-11-16 17:05:55 ah. man this is done so many ways. i've seen it that way, or /boot symlink to /media/mmcblk0p1 or other ways. 2021-11-16 17:06:08 invoked: ye exactly 2021-11-16 17:06:57 if it's the symlink some people will link /boot to /mount/boot/ and everyone gets different results. 2021-11-16 17:07:18 imo there should be one prescribed way to do sys on rpis 2021-11-16 17:07:32 good that alpine-setup support sys install already so setups can be more consistent but many people did that in old wiki way 2021-11-16 17:08:14 MY-R: "apk info -ve raspberrypi-bootloader" says nothing 2021-11-16 17:08:42 freeit: linux-rpi2 in Alpine 3.14.x was recently updated to 5.10.78 four days ago, perhaps that newer kernel causes an issue 2021-11-16 17:09:19 freeit: so you tried run newer kernel with dunno how old raspberry pi boot files, it should still work but who knows 2021-11-16 17:09:50 minimal: That is also my fear 2021-11-16 17:10:26 freeit: check what that package provide, install it, reboot 2021-11-16 17:10:29 freeit: https://pkgs.alpinelinux.org/contents?branch=v3.14&name=raspberrypi-bootloader&arch=armhf&repo=main 2021-11-16 17:10:39 freeit: RPI kernels on Edge went from 5.10.73 straight to 5.15.x 2021-11-16 17:11:09 I have done upgrad lately like freeit and all went smooth but it was on pi4 2021-11-16 17:12:02 MY-R: In past 4 days? 2021-11-16 17:12:28 ye 2021-11-16 17:12:54 MY-R: OK. I try to install raspberrypi-bootloader package and then do an "apk upgrade" which installs the new kernel and then I reboot. Is that correct ? 2021-11-16 17:13:18 freeit: ye 2021-11-16 17:13:48 MY-R: ok, so yes lack of updated bootloader files or firmware files might be the issue then 2021-11-16 17:14:48 freeit: also show what you got in /boot/cmdline.txt and in /boot/config.txt 2021-11-16 17:14:58 As I'm running this irc-session on my raspi (weechat), I will quit in a few moments and come back later. 2021-11-16 17:15:13 freeit: so show those 2 files first pls :) 2021-11-16 17:16:31 freeit: wait! 2021-11-16 17:16:41 remember to use "-a upgrade" 2021-11-16 17:16:49 OK. 2021-11-16 17:17:06 apk update; apk -i -a upgrade 2021-11-16 17:17:23 OK. 2021-11-16 17:17:27 gl! 2021-11-16 17:26:35 `apk -Uia upgrade` will update too. 2021-11-16 17:36:53 MY-R: the desired files after the upgrade and before the reboot: cmdline.txt: https://dpaste.org/NmTU config.txt: https://dpaste.org/6MHE 2021-11-16 17:37:45 freeit: after? so now all working? 2021-11-16 17:38:14 No I didn't boot yet. 2021-11-16 17:38:19 kernel=boot/vmlinuz-rpi2 2021-11-16 17:38:21 initramfs boot/initramfs-rpi2 2021-11-16 17:39:07 Only installed the raspberrypi-bootloader package and made a "apk update; apk -i -a upgrade" 2021-11-16 17:39:09 I dont like those paths, if you mount /dev/mmc*1 directly to /boot which is on mmc*2 then should remove "boot/" 2021-11-16 17:39:59 freeit: remove both "boot/" from paths in [pi2] in config.txt 2021-11-16 17:40:39 should be just "kernel=boot/vmlinuz-rpi2" and "initramfs boot/initramfs-rpi2" 2021-11-16 17:40:43 sorry 2021-11-16 17:40:49 should be just "kernel=vmlinuz-rpi2" and "initramfs initramfs-rpi2" 2021-11-16 17:41:24 freeit: did you add the "iomem-relaxed" and "rootflags=comit=3000" options to the cmdline.txt file or did the installer put them there? 2021-11-16 17:43:44 monimal: Puhh, my initial installation was good a 1 year ago. But I think it''s from me. 2021-11-16 17:44:14 freeit: ok, just wondered why you have those options 2021-11-16 17:44:39 It's all to take care of the sdcard (minmal writes). 2021-11-16 17:45:05 in old installs people got symlink /boot/boot but fat doesnt support it right? 2021-11-16 17:45:23 It works great. No problems with my sdcrad and broken filsystems anymore. 2021-11-16 17:46:05 free: ah, the iomem=relax is for GPIO access? 2021-11-16 17:46:32 MY-R: mount gives for /dev/mmcblk0p1: /media/mmcblk0p1 type vfat (rw,relatime,fmask=0022,dmask=0022,codepage=437,iocharset=ascii,shortname=mixed,errors=remount-ro) 2021-11-16 17:47:07 freeit: ls -al /media/mmcblk0p1 2021-11-16 17:47:07 MY-R: and /dev/mmcblk0p1 on /boot type vfat (rw,relatime,fmask=0022,dmask=0022,codepage=437,iocharset=ascii,shortname=mixed,errors=remount-ro) 2021-11-16 17:47:45 It's mounted on two places. 2021-11-16 17:48:19 it is enough to mount /dev/mmcblk0p1 to /boot 2021-11-16 17:48:42 freeit: yeah I did the same in the past using a bindmount to ensure when the bootloader package updated files in /boot/ that they would update them on the SDcard 2021-11-16 17:48:59 freeit: so inside of /boot you have vmlinuz-rpi2 initramfs-rpi2 files, correct? 2021-11-16 17:49:12 ls -al /media/mmcblk0p1 : https://dpaste.org/EeCH 2021-11-16 17:49:37 nope, you dont have it 2021-11-16 17:49:53 but have extra boot dir 2021-11-16 17:51:11 MY-R: ls -al /boot : https://dpaste.org/rKRv 2021-11-16 17:51:11 so there you have some weird mounted point or symlinks 2021-11-16 17:51:47 MY-R: Maybe, but until now it worked. 2021-11-16 17:51:54 freeit: so you modified the vmlinux-rpi2 and initramfs references in config.txt to remove the "boot/" prefix? what's in your usercfg.txt file? 2021-11-16 17:53:24 minimal: usercfg.txt : https://dpaste.org/aw0w 2021-11-16 17:53:33 freeit: ye but since alpine-setup support sys installs on rpi* then it keep kernel/initramsf directly on vfat partition together with all bootloader files, without any extra boot dirs 2021-11-16 17:54:49 MY-R: there was a change in either Alpinr 3.14.x or 3.13 in terms of how this all worked so I'm guessing freeit installed his RPI before then 2021-11-16 17:55:07 MY-R: OK. I will try this too. 2021-11-16 17:55:40 freeit: so in your fstan should have basicaly 2 mount points: /dev/mmcblk0p1 /boot vfat blabla 0 2 and /dev/mmcblk0p2 / ext4 rw,noatime 0 1 2021-11-16 17:55:50 s/fstan/fstab 2021-11-16 17:55:50 MY-R meant to say: freeit: so in your fstab should have basicaly 2 mount points: /dev/mmcblk0p1 /boot vfat blabla 0 2 and /dev/mmcblk0p2 / ext4 rw,noatime 0 1 2021-11-16 17:56:28 minimal: I think my initial install was with alpine 3.12 2021-11-16 17:57:43 MY-R: fstab : https://dpaste.org/mQRK 2021-11-16 17:57:49 minimal: so ye, that path changed 2021-11-16 17:58:20 MY-R: yes, that's why in past I used a 2nd bind-mount 2021-11-16 17:59:47 freeit: ye kind of messy but shouldnt do it like that anymore with multiple mount points 2021-11-16 18:00:28 MY-R: This configuration was not made (manually) by me. 2021-11-16 18:00:40 I guess all what you have inside "/media/mmcblk0p1/boot" should copy directly in to /dev/mmcblk0p1 2021-11-16 18:00:56 without boot dir 2021-11-16 18:01:43 MY-R: I do my own custom RPI images so haven't kept track what the official stuff does but certainly in the past there was a disconnect between the layout of the downloadable RPI images and where the related packages put stuff 2021-11-16 18:01:57 minimal: ye 2021-11-16 18:02:31 but now with sys install dont need do any such a tricks and all install during upgrade smooth in correct paths without any weird links 2021-11-16 18:03:16 freeit: then remove those both lines starting from "/dev/mmcblk0p1" and "/media/mmcblk0p1/boot" 2021-11-16 18:03:17 MY-R: the raspberrypi-bootloader installs/updates stuff in /boot so if SDcard is not mounted there there updates won't happen on SDcard 2021-11-16 18:03:39 setup-alpine add automaticly to fstab vfat partition 2021-11-16 18:04:31 O.K. until then, thank you very much for your help, suggestions and advices. It's time for action :) I'm rebooting. It will take some time until I'm back in IRC. 2021-11-16 18:04:36 freeit: and add that: /dev/mmcblk0p1 /boot vfat defaults 0 2 2021-11-16 18:04:46 O.K. 2021-11-16 18:05:29 freeit: just be sure that got all bootloader stuff and kernel/initramfs directly in /dev/mmcblk0p1 partition and then rpi will ready it nicly 2021-11-16 18:05:35 ah 2021-11-16 18:06:07 setup-disk still has some references to "initramfs boot/initramfs-rpi" 2021-11-16 18:06:31 minimal: probably for diskless mode where images got boot directory 2021-11-16 18:07:35 MY-R: sees to be only for armhf pi0 and pi0w for some reason 2021-11-16 18:09:12 MY-R: https://gitlab.alpinelinux.org/alpine/alpine-conf/-/blob/master/setup-disk.in#L412 2021-11-16 18:09:54 ye, weird 2021-11-16 18:22:13 i got pulled away from this but i suspected a path issue 2021-11-16 18:28:33 My reboot was unfortunately without success. I have just installed weechat on another system, so I can chat while my pi is offline. 2021-11-16 18:29:16 if you have irc, then you are okay, don't need more 2021-11-16 18:32:28 freeit: well, one pretty fast thing you can do is put fresh Alpine image on your card and try out "sys" install so after you can compare the result with your current/backup one 2021-11-16 18:33:38 and even if that wont work then something is wrong 2021-11-16 18:38:58 I will now try one or the other and will report. 2021-11-16 18:45:57 Speaking of RPi, has anyone here successfully emulated one with QEMU, running Alpine? I extracted 'alpine-rpi-3.14.2-armhf.tar.gz' to the first partition (mkfs.vfat -F32) of an 8GB image, but am unable to get any output from QEMU when invoked with: `qemu-system-arm -sd "rpi.qcow" -machine type=raspi0 -no-reboot -vnc :0` Extracting the kernel and initramfs from the alpine release and 2021-11-16 18:45:57 passing to QEMU with `-kernel ./vmlinuz-rpi -initrd ./initramfs-rpi` doesn't help. 2021-11-16 18:49:10 but, like... why? :p 2021-11-16 18:50:35 Hello71: I prefer to get my configurations perfected in VMs before deploying. Especially if I have to otherwise keep modifying an SD card. 2021-11-16 18:51:48 i think most of the testable things should be arm generic, no? e.g. i doubt qemu precisely emulates the videocore 2021-11-16 18:57:00 Running headless, ideally 2021-11-16 19:06:15 nickersonm: whilst QEMU has some RPI support I don't believe it can fully emulate it 2021-11-16 19:07:46 minimal: I don't mind if the emulation isn't perfect, but I'm clearly doing something wrong with ARM emulation as I can't even boot it (or `-machine type=versatilepb` either). x86_64 works just fine with `qemu-system-x86_64 -machine pc -m 1024 -no-reboot -cdrom alpine-virt-3.14.3-x86_64.iso -vnc :0` 2021-11-16 19:12:23 nickersonm: I've just QEMU to run "generic" aarch64 Alpine images but not tried RPI, I had the impression that even if I got it working it would not be a realistic emulation 2021-11-16 19:12:34 nickersonm: maybe look at this: https://github.com/dhruvvyas90/qemu-rpi-kernel 2021-11-16 19:13:27 nickersonm: do I build and test in VMs for non-RPI-specific stuff 2021-11-16 19:15:17 s/do I/I do/ 2021-11-16 19:15:17 minimal meant to say: nickersonm: I do build and test in VMs for non-RPI-specific stuff 2021-11-16 19:15:42 nickersonm: this gives a step-by-step method to use the above kernel: https://linuxconfig.org/how-to-run-the-raspberry-pi-os-in-a-virtual-machine-with-qemu-and-kvm 2021-11-16 19:16:53 minimal: Yeah, I saw the debian kernels there and tried one instead of the Alpine kernel extracted from 'alpine-rpi-3.14.2-armhf.tar.gz', to no avail. I've not yet tried emulating rasbian instead of alpine. 2021-11-16 19:22:00 nickersonm: the Alpine RPI kernels may not have some stuff enabled that versatilepb needs 2021-11-16 19:28:13 minimal: I guess so; perhaps I'll try on hardware next. I feel like I'm missing something obvious about ARM emulation, though. 2021-11-16 19:29:49 nickersonm: I don't think its ARM emulation in general, rather RPI specific emulation 2021-11-16 19:30:18 I've no problem using QEMU with my "generic" aarch64 Alpine images 2021-11-16 19:30:43 they're using linux-virt aarch64 kernel package 2021-11-16 19:31:27 minimal: thanks, I'll try using that 2021-11-16 19:32:48 nickersonm: for those I've used: qemu-system-aarch64 -machine type=virt-3.1 with virtio devices 2021-11-16 19:33:44 but then as its not RPI-compatible you could just as well use a x86_64 VM also as it would be faster (on a x86_64 desktop/laptop) 2021-11-16 19:34:02 as long as you remember to enable kvm 2021-11-16 19:36:09 minimal: indeed, and I already have LXD for x86_64 stuff 2021-11-16 19:56:52 minimal: yes, both `-machine type=raspi*` and Alpine rpi kernel separately prevent QEMU from booting with my configuration. Arch-for-rpi or Alpine-for-arm both work on `type=virt` or `type=versatilepb`. I guess I need to use hardware to configure my RasPi Zero W. 2021-11-16 19:59:04 nickersonm: well the RPI kernel is expecting Device Tree stuff for RPI's hardware. I think one of those articles mentioned passing an option to QEMU for that but not sure it would work too well 2021-11-16 20:00:08 minimal: Yeah, tried that with the dtb for rpi0 in the alpine image, didn't help. I didn't try beyond that simple test, though. 2021-11-16 20:00:45 I just use the same old small SDcard for image testing, and then new-ish SDcards once I have a working config. If the "test" cards dies some day its not the end of the world lol 2021-11-16 20:01:27 Yes, it would probably have been faster for me to start with hardware in the first place! But that violates my standard practices ;) 2021-11-16 20:04:17 nickersonm: your standard doesn't fit with your results? invent a new standard, that's what everyone else does :-) 2021-11-16 22:21:22 :-) 2021-11-17 00:38:43 MY-R: minimal: I took a new sdcard, created two partitions. On the first one I made an new "sys" install. On the second I copied my old root system. After some changes in fstab my sytem booted up with with the newest kernel as provided by alpine 3.14-3 :) Again: Thank you very much for your help. 2021-11-17 01:43:18 can nginx proxy handle 'git://' request? maybe need to install a module? 2021-11-17 01:44:11 no 2021-11-17 01:45:36 no and why would you 2021-11-17 01:45:38 there is also not much use for git:// , you are better off setting up an http git frontend with git-http-backend (packaged in git-daemon) 2021-11-17 01:45:56 plenty of examples of how to pass fastcgi to it in nginx online you can copy paste if you want 2021-11-17 01:53:33 thanks 2021-11-17 12:44:12 could I do some workaround for !13200 ? 2021-11-17 12:44:16 ups 2021-11-17 12:44:21 could I do some workaround for #13200 ? 2021-11-17 12:44:35 I tried apk add py3-aiorpcx=0.19 2021-11-17 12:46:18 maybe forcing the version on electrum APKBUILD? 2021-11-17 13:16:05 py3-aiorpcx-0.22.1-r0: 2021-11-17 13:16:07 breaks: .makedepends-electrum-20211117.131504[py3-aiorpcx=0.19] 2021-11-17 13:24:56 electrum commiter here. let's check the other way too 2021-11-17 13:24:56 sed -i 's~required: 0.18.7<=ver<0.19~required: 0.18.7<=ver<0.23~' /usr/bin/electrum /usr/lib/python3.9/site-packages/electrum/electrum 2021-11-17 13:24:56 does this work for you? 2021-11-17 13:24:56 sed -i 's~aiorpcx._version < (0, 19)~aiorpcx._version < (0, 23)~' /usr/bin/electrum /usr/lib/python3.9/site-packages/electrum/electrum 2021-11-17 13:32:17 lzsaver: re your "label" issue on 1st boot from a couple of days ago, did you notice this #13194, specifically "busybox implementation does not set label correctly, at least according to dosfstools fsck" 2021-11-17 13:32:56 minimal, yes. I read that. I'll retest with right symlink later 2021-11-17 13:34:00 donoban, seems gentoo forced aiorpcX 0.19*. probably new version is not compatible. need to think 2021-11-17 13:38:03 uhM, lzsaver I don't it if its a good idea 2021-11-17 13:38:48 upstream doesn't support 0.22 yet because there are several api changes 2021-11-17 13:38:56 thanks. I just dont have alpine with gui for now, so cant test 2021-11-17 13:51:32 and is there some way for force the package to downgrade? 2021-11-17 14:00:43 you're on edge right? 2021-11-17 14:04:31 you may add 3.14 repo to /etc/apk/repositories this way 2021-11-17 14:04:31 @3.14 https://dl-cdn.alpinelinux.org/alpine/v3.14/main 2021-11-17 14:04:31 @3.14 https://dl-cdn.alpinelinux.org/alpine/v3.14/community 2021-11-17 14:04:31 after that you can install the package from that repo 2021-11-17 14:04:31 apk add py3-aiorpcx@3.14 2021-11-17 14:04:33 also you can force specific version 2021-11-17 14:04:33 apk add py3-aiorpcx~0.18.7 2021-11-17 14:04:35 or both 2021-11-17 14:04:35 py3-aiorpcx@3.14~0.18.7 2021-11-17 14:08:03 ok let me try 2021-11-17 14:09:51 WARNING: The repository tag for world dependency 'py3-aiorpcx@3.14' does not exist 2021-11-17 14:10:05 you need to edit /etc/apk/repositories 2021-11-17 14:10:12 I did 2021-11-17 14:10:18 # Fix Electrum 2021-11-17 14:10:20 http://alpine.42.fr/v3.14/community 2021-11-17 14:10:25 ahh 2021-11-17 14:10:29 didn't see the @3.14 2021-11-17 14:10:42 yeah 2021-11-17 14:10:58 it's a label for repos 2021-11-17 14:11:00 (1/1) Downgrading py3-aiorpcx@3.14 (0.22.1-r0 -> 0.18.7-r0) 2021-11-17 14:11:13 nice, it works again 2021-11-17 14:13:12 good. now we need to know why py3-aiorpcx was bumped. was there some reason for some package or not 2021-11-17 14:13:38 I think that andypost upgraded it directly, I didn't see any MR for that 2021-11-17 14:14:05 commit a024c11d588c8b1efaebd022a339aacea672f93a 2021-11-17 14:16:57 if there's no packages with dep for 0.22.x, it may be reverted back to working version 2021-11-17 14:17:03 would be good to know 2021-11-17 14:29:22 donoban, if you'll want to remove @3.14 from repos in the future without breaking the world file, you'll need to force specific version before that using 2021-11-17 14:29:22 apk add py3-aiorpcx~0.18.7 2021-11-17 14:29:22 it means that 'apk upgrade' will keep this version for the package. 2021-11-17 14:29:56 if you're okay with @3.14 in your repos you are fine, and no actions needed 2021-11-17 14:49:44 ah thanks I will remove it better 2021-11-17 14:50:19 lzsaver: it seems that electrum is the only package that requires it https://pkgs.alpinelinux.org/package/edge/community/x86_64/py3-aiorpcx 2021-11-17 15:20:16 donoban, if apk says that electrum needs it. try "apk add py3-aiorpcx@3.14~0.18.7" and after that - "apk add py3-aiorpcx~0.18.7". or simply edit /etc/apk/world: py3-aiorpcx@3.14 -> py3-aiorpcx~0.18.7 2021-11-17 15:20:36 after that @3.14 may be removed 2021-11-17 15:24:33 you'll may need to remove "~0.18.7" from the world file to upgrade to next electrum version later if it will require new aiorpcx 2021-11-17 15:32:11 ok 2021-11-17 15:33:10 (4/11) Replacing py3-aiorpcx (0.18.7-r0 -> 0.18.7-r0) 2021-11-17 15:33:41 all right, thanks lzsaver 2021-11-17 15:33:50 you're welcome 2021-11-17 16:02:32 !26746 2021-11-17 16:03:51 wow, how did you found it ikke ? 2021-11-17 16:08:17 I think that py3-websockets has no relation with aiorpcx: https://github.com/websocket-client/websocket-client/search?q=aiorpcx&type= 2021-11-17 16:17:25 donoban: on the top of the commit page in gitlab is a reference to the MR that merged it in 2021-11-17 16:18:40 oh nice, ty 2021-11-17 16:18:51 donoban: depends="python3 py3-websockets" 2021-11-17 16:19:01 There is a relation 2021-11-17 16:19:19 ah but on the reverse direction 2021-11-17 16:19:25 uhM 2021-11-17 16:19:32 maybe current py3-websockets fails with older version? 2021-11-17 16:26:51 uhM 2021-11-17 16:27:02 https://github.com/kyuupichan/aiorpcX/pull/20 2021-11-17 16:35:23 running into this issue? https://github.com/kyuupichan/aiorpcX/issues/17 2021-11-17 16:35:38 no, that's about 0.17.0 2021-11-17 16:36:04 running into this https://github.com/spesmilo/electrum/issues/7446 2021-11-17 18:45:36 Per Alpine Wiki, I'm supposed to Install these 2 pkgs, virtualbox-guest-additions virtualbox-guest-modules-virt BUT the 2nd pkg virtualbox-guest-modules-virt is not in repo. Any 1 has a clue what to do? 2021-11-17 18:46:02 do you need them? 2021-11-17 18:46:41 the wiki is wrong 2021-11-17 18:47:06 sorry this is for VBox Guest addtions 2021-11-17 18:47:27 so you mean I don't need the 2nd package?? -virt pkgs? 2021-11-17 18:47:29 iirc it only provides shared "folder" support 2021-11-17 18:47:52 I see. thanks. However, I'd LOVe to use the shared folders :) 2021-11-17 18:48:14 any other way to make shared folder to work if needed? 2021-11-17 18:48:16 there are many types of network mount technologies 2021-11-17 18:48:29 k. thx. I'lll look into it 2021-11-17 18:48:33 thx for the help 2021-11-17 18:53:05 the virtualbox kernel modules were upstreamed 2021-11-17 20:22:13 Hello folks, git pulling aport is super long. Is there a way to speed it up ? 2021-11-17 20:22:48 libcap package does not install header files. Is that expected? In which package I can find sys/capability.h? 2021-11-17 20:23:42 libcap-dev: https://pkgs.alpinelinux.org/contents?file=capability.h&path=&name=&branch=v3.14&arch=x86_64 2021-11-17 20:27:05 Opps, not familiar with package names. Shouldn't `apk info --who-owns /usr/include/sys/capability.h` command show that result? 2021-11-17 20:27:57 staceee: which one 2021-11-17 20:28:07 Biswa96: no, apk info -W is only for installed packages 2021-11-17 20:28:35 all linux package managers work this way. otherwise you must download a very large file database 2021-11-17 20:29:15 Hello71: git://git.alpinelinux.org/aports 2021-11-17 20:29:36 but maybe I should just use the gitlab one 2021-11-17 20:29:53 git: is not a good protocol and also the main repository is on gitlab now 2021-11-17 20:30:08 you can also download from github if you have a faster connection to that 2021-11-17 20:30:16 then git remote set-url 2021-11-17 20:30:25 Thank you very much :) 2021-11-17 20:52:43 Biswa96: you need libcap-dev (see also https://pkgs.alpinelinux.org/contents?file=capability.h ) 2021-11-17 20:53:10 all linux package managers work this way. otherwise you must download a very large file database ← gentoo’s portage uses rsync I think. 2021-11-17 20:53:18 wat 2021-11-17 20:53:31 actually maximum wat 2021-11-17 20:54:28 that's like, three steps away from being remotely relevant 2021-11-17 20:55:59 rsync is for ebuilds, not installed files. portage (and any primarily source-based distro) fundamentally can't provide a comprehensive file database. you could theoretically sync the file database with rsync but that's no different from the sync package database 2021-11-17 20:56:18 liske: lopid already said that half an hour ago 2021-11-17 20:57:18 Oh right, I tend to think binary and source repository work alike. 2021-11-17 20:59:55 'yum whatprovides …' can work on not installed packages, but might do a remote lookup 2021-11-17 21:00:47 Hello71: I did reboot my pc and forgot that my irc log had a gap *facepalm* 2021-11-17 21:03:35 lopid: sure but (i hope) nobody is dumb enough to download whole database by default 2021-11-17 21:04:23 it could only be a few MiBs compressed, surely… 2021-11-17 21:06:41 ~30 MB for arch, plus what is your policy for updating them? normally package database is downloaded in full on each update, but people will be annoyed to download 30 MB just to check for updates, so you will need to implement some kind of diff update 2021-11-17 21:07:11 theoretically it is possible but the cost-benefit is not good 2021-11-17 21:08:41 it wouldn't surprise me of redhat 2021-11-17 21:09:54 i guess 2021-11-17 21:10:34 i would expect it more from ubuntu or something. i think cnf database is downloaded by default, but pretty sure it only contains binary names, not all files 2021-11-17 23:35:11 Is 112mb of ram used with xorg reasonable or should it be lower? I am just running a wm and st nothing else 2021-11-17 23:39:59 how are you measuring "ram used" 2021-11-17 23:43:49 I use htop and free 2021-11-17 23:44:55 ... 2021-11-17 23:46:48 htop does say how much ram is being used 2021-11-17 23:52:59 when asking smart questions, it is important to be specific. "i used free" is barely more useful than "i used the terminal". say something like "i ran free and compared the Mem: used value to the Mem: total value", or better, post the results. if it's more than two lines long, use a pastebin. 2021-11-17 23:56:06 htop reports that 112mb of ram is being used. Is that a normal amount or should/could it be lower? 2021-11-18 06:43:53 I'm attempting to make a custom service log the output to a file. HOw do I do this? I thought all I would have to do is set output_log and error_log to the corresponding files in /var/log, however, I get a message at startup saying 'permission denied'. 2021-11-18 06:44:24 How should I be configuring logging? 2021-11-18 06:44:36 (Ideally with a log size limit as well) 2021-11-18 06:45:33 you need to make the files for _log and ensure they have the correct ownership/permissions 2021-11-18 06:46:31 as for size you would use logrotate i guess to cycle logs over time 2021-11-18 06:47:08 Ah, that should work. What's the correct ownership/permissions? 2021-11-18 06:47:33 I thought it should be owned by root, etc, as it's the init system that writes to them? 2021-11-18 06:49:23 whatever is command_user should be fine 2021-11-18 06:49:41 Ah, interesting. 2021-11-18 06:51:30 Works a charm! 2021-11-18 06:51:34 Thanks for that psykose :) 2021-11-18 06:51:42 :) 2021-11-18 06:52:27 afaik this is only for supervise-daemon, i have no idea how logs get redirected with the default start-stop-daemon behaviour aside from configuring the service itself to log things somewhere 2021-11-18 06:55:24 arent stdout / stderr detached when running a process in the background? 2021-11-18 06:55:52 (which happens with start-stop-daemon, without supervisor) 2021-11-18 06:56:31 most likely, hence the latter part 2021-11-18 06:57:10 from what i can see every s-s-d service is either going directly to syslog, has a --flag to go to syslog, or has something in its own /etc configuration to write to /var/log and the init.d service just makes sure the files get made 2021-11-18 06:57:47 (or a --flag to write to /var/log, but same difference) 2021-11-18 07:22:20 ikke: Seems to be working perfectly, though. 2021-11-18 07:22:49 ikke: I mean, my service doesn't detach - it relies on openRC to background it. 2021-11-18 07:25:20 Is it possible to configure the service so that it restricts the amount of memory it can use? 2021-11-18 07:45:00 arahael3, yes, that's what cgroup does 2021-11-18 07:45:36 fcolista: How do I learn about it? 2021-11-18 07:45:51 I already bubblewrap it. 2021-11-18 07:45:55 giyf :) 2021-11-18 07:46:22 arahael3, take a look here, this might help in your case: 2021-11-18 07:46:23 https://unix.stackexchange.com/questions/44985/limit-memory-usage-for-a-single-linux-process 2021-11-18 07:46:46 i've just googled it.. 2021-11-18 07:47:10 it seems like youtube doesn't work on all the web browsers using webkit2gtk, do I need to install some codec or something? 2021-11-18 07:47:51 fcolista: Nice - it's that simple. I'll check it out. 2021-11-18 07:47:54 generally speaking, cgroups is one of the kernel feature that allows container to work, along the namespaces. So if you search for this information, i'll find tons of documentation 2021-11-18 07:48:28 under the hood, all containers technologies relies upon ns/cgroups 2021-11-18 07:49:05 iquit: yes, some gst-plugins-something and gst-libav I think 2021-11-18 07:49:18 like -base good and maybe even ugly ;/ 2021-11-18 07:49:24 I see, thanks! 2021-11-18 08:06:23 :o 2021-11-18 08:53:31 fcolista: Yeah, I'm using bubblewrap, I think I have to set up a cgroup, and then tell bubblewrap *not* to create it's own. But I'll do that another evening. 2021-11-18 10:42:03 Greetings. 2021-11-18 10:42:39 I've joined this channel because I am installing Alpine Linux and have a question. 2021-11-18 10:42:59 I've burned the ISO to a CD-R and booted from there to start the installation process. 2021-11-18 10:43:30 However, during the OpenRC booting sequence, it has failed to verify the signature of modloop 2021-11-18 10:43:51 I am not connected to the Internet during this process. 2021-11-18 10:44:47 I'm running the latest version, which is 3.14.3 for x86_64 architecture 2021-11-18 10:45:44 My question is what to do with this. Is this something to worry about? Should I connect to the Internet in order for the modloop signature to be verified? 2021-11-18 11:13:11 does it continue booting? 2021-11-18 11:37:01 smv https://gitlab.alpinelinux.org/alpine/aports/-/issues/11136 somehow happening? 2021-11-18 11:55:37 Yes, it continues booting. 2021-11-18 11:55:41 jhjacobs81 2021-11-18 11:56:00 im reading the article you linked, artok 2021-11-18 11:57:11 it could be that the disks are not clean from previous linux installations 2021-11-18 11:57:33 you see, i've tried to install 2 or 3 before (though I believed I wiped the disks before this installation) 2021-11-18 11:57:44 i'll wipe the disks and try again 2021-11-18 11:57:44 brb 2021-11-18 11:58:01 (and with linux installations, i mean different versions) 2021-11-18 12:20:56 Wiping the disk is not the solution here. 2021-11-18 12:21:14 Maybe it's because of the installation medium (CD-R). 2021-11-18 12:21:32 I'll see if i can continue installing and check with an USB later 2021-11-18 12:22:03 (meaning if the asme error is displayed while installing from USB) 2021-11-18 12:32:38 oh wait, i never wiped the disks properly 2021-11-18 12:32:44 so that might be a solution, anyway 2021-11-18 12:32:52 there were still disk signatures present, i just found out 2021-11-18 12:33:04 i'm pretty new at this :-) 2021-11-18 18:37:29 Are stats available from package builder pipelines? I'm curious to see RAM use of certain builds. 2021-11-18 18:52:15 khm: No, we do not have any stats about that 2021-11-18 21:37:45 how can i see versions before answering yes to an upgrade? adding verbosity to `apk -ia upgrade` does nothing 2021-11-18 21:43:19 lopid: probably -s but that's extra step before actual upgrade 2021-11-18 21:44:11 that could work 2021-11-18 21:45:32 well as alias first apk upgrade -s && apk upgrade -ia 2021-11-18 21:45:57 nod 2021-11-18 21:46:37 i usually do `apk -ai upgrade && update-conf`. i'll just add `apk -as upgrade` beforehand 2021-11-18 21:47:54 hah never used update-conf ;/ 2021-11-18 21:48:14 lopid: after "apk update" you can run "apk version" 2021-11-18 21:48:15 I see I have apk update before, probably for nothing 2021-11-18 21:48:46 oh 2021-11-18 21:51:40 `apk -U version` might suffice 2021-11-18 23:11:02 how would i go about running two instances of the same service at the same time, with different conf.d configs? 2021-11-18 23:12:49 is it possible to contribute to alpine-conf/setup-bootable.in through a github PR, or is it preferred to make a MR on the gitlab instead? currently it fails to install syslinux due to the intel and amd ucodes, it doesn't split on comma like it should 2021-11-18 23:28:04 krystianch, copy-paste the service? 2021-11-18 23:29:31 lzsaver: right, thanks, that's what i thought but wanted to make sure that's the way 2021-11-18 23:30:08 but be sure that your services do not use the same files 2021-11-18 23:30:50 I mean they should not write to the same file at the same time for example 2021-11-18 23:31:25 yup, thanks 2021-11-18 23:39:38 should not a package which create shitty user/group also delete it during uninstall? 2021-11-18 23:41:41 lzsaver: init.d symlink seems to work and copy with different config in conf.d, so one less file to think about 2021-11-19 00:27:07 krystianch, good 2021-11-19 08:06:38 ncopa: i know it's 4 years old now but i was shown this talk you did in 2017: https://vid.puffyan.us/watch?v=sIG2P9k6EjA - thank you for the amazing talk. i actually learned a lot about alpine because of it. i'm planning to switch to it for all my things 2021-11-19 10:01:08 leah: I don't watch videos but what was the decisive point that makes you use alpine specifically? 2021-11-19 10:01:33 well, i already knew that musl and busybox are extremely light 2021-11-19 10:01:43 For me, alpine seems simple, clean, and very very well thought out - rare in a time where everyone seems to aim for complexity. 2021-11-19 10:01:45 but i never knew they were suitable for workstation/laptop OS 2021-11-19 10:01:52 yeah, that too 2021-11-19 10:02:06 i've been studying the project, looking at its infrastructure, seeing how it does things, and i like what i see 2021-11-19 10:02:14 it has far less integrated packages than, say, debian 2021-11-19 10:02:20 but i'm ok compiling stuff 2021-11-19 10:02:40 i started off when i got sick of extreme bloat in modern gnu+linux distributions 2021-11-19 10:02:51 so i began searching for lightweight solutions. and i found alpine 2021-11-19 10:02:57 i'm currently running ubuntu with lxde 2021-11-19 10:03:05 but not for long. i'm switching all my stuff to alpine 2021-11-19 10:03:19 servers too. my server hosting is in need of an overhaul 2021-11-19 10:03:38 i've not used alpine yet but i already know it will be good. i've been talking to people 2021-11-19 10:03:40 plus 2021-11-19 10:03:49 i already know openwrt very well, and openwrt is very similar to alpine 2021-11-19 10:04:00 (uses busybox. similar package management) 2021-11-19 10:04:32 there is another reason 2021-11-19 10:04:49 Plus, one of the security staff is very cool, I read Ariadne's twitter. :) 2021-11-19 10:04:49 She's a good rabbit. 2021-11-19 10:05:05 i currently maintain a build system for one of my projects, written in bash. however, i wish to make it portable, compliant with posix shell 2021-11-19 10:05:13 and my understanding is that busybox is quite strict in that regard 2021-11-19 10:05:28 yeah ariadne is awesome. i read her twitter too 2021-11-19 10:06:33 i was planning on switching all my stuff back to archlinux. i had used that for years. alpine has couped this plan, and i'm going to install that instead 2021-11-19 10:27:09 minimalism is key if you want something maintainable 2021-11-19 10:27:29 hopefully this distros stays minimal instead of falling for the temptation to become everything to everyone 2021-11-19 10:28:22 leah: I’m glad you like it 2021-11-19 10:40:10 Yeah. I find myself flip-flopping between three distros as late. I'm a *long* time debian guy, but I use macos at work and now home for a long time, too. Sadly there's no alternative to macos if you're doing iOS dev. 2021-11-19 10:40:27 But when it comes to linux, I really like different aspects of alpine, debian, and nixos. 2021-11-19 10:41:06 (nixos is both *awful* and *awesome* - good for software devs, I think) 2021-11-19 11:24:10 I am finding myself using Alpine, OpenBSD, and macOS in that order 2021-11-19 11:24:33 OpenBSD is great for software devs as well because it forces you to write very portable code 2021-11-19 11:44:06 valerius: That's probably true only for C. 2021-11-19 11:44:10 I tend to avoid C. 2021-11-19 16:38:43 fyi: alpine 3.15 rc4 for rpi image, the files initramfs-rpi and initramfs-rpi4 have mode 600 instead of mode 644 like the rest of the files in /boot. Causes some minor issues when tftpserver does not run as root 2021-11-19 18:58:32 strange, I upgraded the linux-lts kernel yesterday (latest-stable) and `veth` is no longer an available kernel module 2021-11-19 18:59:00 did you reboot 2021-11-19 18:59:10 Unsure, I will try 2021-11-19 19:05:00 That was it, thanks 2021-11-19 19:05:34 upgrading the kernel deletes the old /lib/modules directory for the previous one, so you can't add any new modules until you load the new one 2021-11-19 20:25:47 If you enable package caching, you can still extract the old kernel modules from the package 2021-11-20 00:15:22 I'm attempting to run the aarch64 version of alpine in qemu on x86, does anyone know the command to use to boot this up using the serial console? I've been using "qemu-system-aarch64 -M raspi3b alpine-virt-3.14.3-x86_64.iso -nographic" and it "seems to be doing something" but I'm not sure. 2021-11-20 00:16:50 -serial stdio 2021-11-20 00:17:07 but that does the same as -nographic 2021-11-20 00:17:13 so if you see nothing, it's not doing anything 2021-11-20 00:17:42 psykose: I see nothing, but it's using more than 330% CPU, so it's doing _something_. 2021-11-20 00:18:40 -append "console=ttyS0" will send kernel output to the serial 2021-11-20 00:19:03 Hmm, I can't seem to kill it. 2021-11-20 00:19:10 Should I try to kill -9 it? 2021-11-20 00:19:18 well it obviously won't work if you give the x86_64 iso to qemu-aarch64 2021-11-20 00:19:23 that's not very surprising 2021-11-20 00:19:43 Hello71: But I'm not doin... Oh... 2021-11-20 00:20:55 Ok, fixed. Same behaviour. 2021-11-20 00:21:21 Now using qemu-system-aarch64 -M raspi3b alpine-virt-3.14.3-aarch64.iso -display none 2021-11-20 00:21:26 I'll give psykose's tip a go... 2021-11-20 00:22:10 -append doesn't do anything if you aren't passing -kernel 2021-11-20 00:22:24 Yeah, indeed, it requires -kernel. 2021-11-20 00:22:33 So that won't work for me, I don't know what to put in for -kernel. 2021-11-20 00:22:34 or it just makes it fail, not sure which 2021-11-20 00:22:39 It just fails yeah. 2021-11-20 00:22:44 i'm not sure how arm boot process works but i doubt generic arm iso will work with rpi machine type 2021-11-20 00:22:50 pretty sure you need -M virt 2021-11-20 00:23:39 arahael3: I've never tried serial console but used "-machine type=virt -display curses - vga std -drive file=alpine.qcow2,if=virtio,format=qcow2" fine in past 2021-11-20 00:24:00 arahael3: that's for booting a pre-prepared Alpine disk image 2021-11-20 00:24:21 minimal: Pretty sure virt shouldn't work, though... 2021-11-20 00:24:30 minimal: I have to emulate this, not run it in virtualisation. 2021-11-20 00:24:48 arahael3: why not if various virtio stuff is setup? 2021-11-20 00:25:00 minimal: Because my CPU can't run aarch64? 2021-11-20 00:25:15 e.g. also using "-device virtio-keyboard-pci -device virtio-net-pci" etc 2021-11-20 00:26:09 -display curses just clears the screen, so I don't see anythign at all. Interesting. 2021-11-20 00:26:38 arahael3: yes its virtualisation is software not in hardware, didn't think KVM on x86_64 could do hardware virtualisation of aarch64 2021-11-20 00:27:00 minimal: It can't, and again, that's why I'm specifically not doing virt. 2021-11-20 00:27:14 minimal: I'm trying to get this to emulate the rpi. 2021-11-20 00:27:49 arahael3: the "virt" machine simply means its emulated a virtio-targeted "machine" rather emulating a real hardware machine 2021-11-20 00:28:01 minimal: Right... So i think that won't work? 2021-11-20 00:28:18 arahael3: yeah but QEMU does not emulate everything needs for a RPI 2021-11-20 00:28:38 00:24 minimal: Pretty sure virt shouldn't work, though... 2021-11-20 00:28:40 00:24 minimal: I have to emulate this, not run it in virtualisation. 2021-11-20 00:29:09 as minimal is trying to explain, -M virt is completely unrelated to whether you are using KVM acceleration 2021-11-20 00:29:25 arahael3: you mention alpine-virt-3.14.3-aarch64.iso earlier that's why I thought you wanted to emulate a generic aarch64 server 2021-11-20 00:29:51 -M rpi vs -M virt specifies the firmware and platform devices available to the guest 2021-11-20 00:29:53 minimal: The sole reason I'm using that particular .iso is because it defaults to using the serial console. 2021-11-20 00:30:17 enabling kvm is done by -M machine,accel=kvm or -enable-kvm 2021-11-20 00:30:23 or possibly -accel kvm i think 2021-11-20 00:30:49 Hello71: you can use either option - but that won't work for aarch64 on x86_64 2021-11-20 00:30:54 yes 2021-11-20 00:31:08 but my point is -M virt is completely unrelated to kvm 2021-11-20 00:31:19 Hmm, I was reading https://www.qemu.org/docs/master/system/target-arm.html and it says: 2021-11-20 00:31:23 For QEMU’s Arm system emulation, you must specify which board model you want to use with the -M or --machine option; there is no default. 2021-11-20 00:31:50 theoretically i think it could possibly work if you ran qemu-aarch64 in qemu-x86_64 on aarch64 2021-11-20 00:32:00 but i'm sure it's not actually hooked up 2021-11-20 00:32:29 arahael3: yes so if you want QEMU to pretend its a rpi then use "-M rpi", if you want QEMU to pretend its a versatilepb board then use "-M versatilepb" 2021-11-20 00:33:06 but as I already explained, AFAIK QEMU cannot 100% emulate a RPI currently, some stuff just does not work 2021-11-20 00:34:45 minimal: Sorry if I'm a bit slow, just trying to reconsole all the information. That 'rpi' type there is not an output of 'qemu-system-aarch64 -machine help' so shouldn't work at all? 2021-11-20 00:35:22 arahael3: depends on the version of QEMU I believe 2021-11-20 00:35:35 Ah, that could be it. 2021-11-20 00:35:45 Interesting: virt QEMU 6.0 ARM Virtual Machine (alias of virt-6.0) 2021-11-20 00:36:11 'virt' doesn't work, though, it's still just a blank screen when using the -display curses 2021-11-20 00:36:15 https://www.qemu.org/docs/master/system/arm/raspi.html 2021-11-20 00:37:03 note the "Missing Devices" section 2021-11-20 00:37:23 minimal: I don't think I need those devices, though? 2021-11-20 00:38:23 I just want this to boot, and then run my compiler. 2021-11-20 00:38:29 arahael3: depends on if the kernel you are using tried to access them I guess 2021-11-20 00:38:32 (For cross compilation, really) 2021-11-20 00:38:40 minimal: Ah. Shouldn't it print something first, though? 2021-11-20 00:39:01 if it's for cross compilation you don't need to emulate an rpi unless you want to use -march=native 2021-11-20 00:39:13 arahael3: perhaps, or it might crash as they are supposed to be present in a real RPI. Don't know 2021-11-20 00:39:22 minimal: even a crash would be something! 2021-11-20 00:40:26 arahael3: you might also need to supply DTB info to QEMU 2021-11-20 00:40:27 psykose: The language I'm using is notoriously complex when setting it up for cross compilation, so I think I do need it to emualte and run a real aarm64 system. 2021-11-20 00:40:37 minimal: How do I do that? 2021-11-20 00:40:58 "-dtb" but you need to provide a suitable file 2021-11-20 00:41:33 Maybe I should just buy parallels and run a VM on my mac, this is sounding really complex. :( 2021-11-20 00:42:23 (My mac is an arm64 system) 2021-11-20 00:42:33 arahael3: and does Parallels emulate a RPI? 2021-11-20 00:42:48 minimal: Doesn't really matter, it virtualises aarch64. 2021-11-20 00:43:10 So it should be able to run the same alpine linux iso. 2021-11-20 00:43:31 right, but that's why I asked you earier why you were wanting to emulate a RPI rather than a "virt" (i.e. virtio-based) armv7 or aarch64 server 2021-11-20 00:44:22 Ah, and I was confused by that question because I didn't realise that "virt", the name of the driver, includes the ARM 64-bit CPU's 2021-11-20 00:45:18 Looking at the qemu page for virt, it does say this: Note that the default is cortex-a15, so for an AArch64 guest you must specify a CPU type. 2021-11-20 00:45:34 But it seems to start (and burn CPU) if I don't specify the CPU type... 2021-11-20 00:47:31 Hello71: Thanks for your clarification earlier, it makes even more sense, I was somehow still interpreting that as "ok, so not related to kvm, but I'm not using that either..." - it's such a bad name on qemu's part! 2021-11-20 00:47:33 arahael3: rightm and for a RPI3/4 you'd use "cortex-a72" 2021-11-20 00:47:43 no it isn't 2021-11-20 00:49:30 minimal: Ok, that does something different: Burns 100% of teh CPU, instead of 340%, but that's the only diference I can see. 2021-11-20 00:49:50 I'm now using this command: qemu-system-aarch64 -M virt alpine-virt-3.14.3-aarch64.iso -display curses -cpu cortex-a72 2021-11-20 00:50:12 arahael3: so what about specifying things like virtio devices? 2021-11-20 00:50:14 Hello71: I must've understood the virtual machine vs emulated machine distinction incorrectly 2021-11-20 00:50:17 minimal: I don't know how! 2021-11-20 00:51:13 arahael3: as I mentioned earlier by using options like "-device virtio-keyboard-pci" 2021-11-20 00:51:24 "-device virtio-net-pci" 2021-11-20 00:51:41 minimal: I'm literally copying those over to the command line now... 2021-11-20 00:52:37 Ok, and it's back to burning 100% CPU. (I really don't know what I'm doign with this. Once it boots and gives me a shell, I'll be back in my element) 2021-11-20 00:52:38 try reading the QEMU docs first rather than just randomly trying things. The principle is the same as if you were trying to run a virtio-based x86_64 VM 2021-11-20 00:55:00 minimal: There are a lot of docs, but the "System Emulation" -> "Quick Start" seemed like a good start, which suggests just "qemu-system-x86_64 linux.img", and that's where I went from... That one's obviously x86 specific, but there's a later section "QEMU System Emulator Targets" which made me try the -M settings. 2021-11-20 00:57:56 minimal: I'm not sure how to navigate the docs otherwise... 2021-11-20 00:58:19 This is the page I'm referring to, incidentally: https://www.qemu.org/docs/master/system/index.html 2021-11-20 00:59:17 I've read: Quick Start, Invocation, QEMU System Emulator Targets, and finally Arm System emulator 2021-11-20 01:01:33 arahael3: for example I don't remember if the "virt" aarch64 machine supports BIOS, I doubt it and so it is likely UEFI-based booting, so you need to give QEMU the AAVMF files to use to emulate UEFI also 2021-11-20 01:01:49 minimal: Let me google what "AAVMF" is... 2021-11-20 01:02:22 BIOS is basically a x86 & x86_64 specific thing. Have you ever booted a UEFI x86_64 machine before? 2021-11-20 01:02:59 My current server is UEFI-based. Ah, so you're suggesting it might not be able to boot the .iso file directly? 2021-11-20 01:03:13 s/machine/virtual machine/ 2021-11-20 01:03:13 minimal meant to say: BIOS is basically a x86 & x86_64 specific thing. Have you ever booted a UEFI x86_64 virtual machine before? 2021-11-20 01:03:30 So I probably should prepare an image in the same way I was preparing it for the rpi itself? I thought that was just literally because I was using it as an embedded system and I didn't have a CD Drive for it. 2021-11-20 01:03:35 have you ever used QEMU to boot a UEFI-based x86_64 VM? 2021-11-20 01:03:51 minimal: I wouldn't know, I've used qemu to boot an x86_64 VM. 2021-11-20 01:04:25 minimal: Using libvirt, though, not directly, using virt-install and all that. 2021-11-20 01:04:30 arahael3: right but you need to use different options for QEMU when booting BIOS (the default) versus UEFI 2021-11-20 01:04:57 Ah... Then I guess it's probably using the BIOS. 2021-11-20 01:05:25 so try getting a UEFI x86_64 VM booting before then trying UEFI aarch64 as the techniques are similar 2021-11-20 01:06:07 though x86_64 uses OVMF files whereas aarch64 uses AAVMF files 2021-11-20 01:07:33 note: Alpine does not correctly package AAVMF files presently, you can extract them from the Alpine "ovmf" *aarch64* package 2021-11-20 01:09:21 Well, "qemu-system-x86_64 -display curses alpine-standard-3.14.3-x86_64.iso" just works, but you're saying that's going to be defaulting to the BIOS. So now I need to see how to specify UEFI explicitly. 2021-11-20 01:10:25 Nothing on https://www.qemu.org/docs/master/system/invocation.html?highlight=uefi seems to suggest how to do that? 2021-11-20 01:11:28 The only mention of uefi anywhere in the docs seems to be the -smbios argument, but that's i386 only... 2021-11-20 01:13:28 I'm not trying to be difficult, I really don't know what I'm doing here, and I'm thinking that qemu is really the wrong tool for the job for me. 2021-11-20 01:13:30 arahael3: BIOS has a text console (so "-display curses" works), UEFI is completely graphics based so for a console you need something like "-vga std" instead 2021-11-20 01:14:01 a good summary: https://www.kraxel.org/blog/2019/09/display-devices-in-qemu/ 2021-11-20 01:14:10 minimal: Ah, so that's why it's just a blank screen? I guess I'll have to use -display none if I persist with qemu 2021-11-20 01:16:48 arahael3: well if you look at the end of that article, for UEFI server it recommends: "-vga none -nographic" - that would require a serial instead 2021-11-20 01:17:37 Yeah, I want to use a serial. 2021-11-20 01:18:21 But I think unlike the x86_64 system, which is evidently using teh bios, it's not booting the iso image in the first place so I think the serial device hasn't been connected yet. 2021-11-20 01:18:57 so you need to specific a virtio serial device to qemu and also add "console=ttyS0" or "console=ttyAMA0" (from memory) to the boot cmdline 2021-11-20 01:20:10 as U said, if it requires UEFI to be present then you need to pass those OVF/AAVF files on the QEMU commad line or it will never get as far as typing to loasd the kernel 2021-11-20 01:20:19 minimal: From my understanding, the "virtual" alpine releases should set up the console automatically, I think - it's why I'd picked those. 2021-11-20 01:21:29 arahael3: I though that was removed a while ago and the automatic serial console interfered with some hypervisors, could be wrong (maybe it was removed for physical machines rather than VMs) 2021-11-20 01:21:44 s/and the/as the/ 2021-11-20 01:21:45 minimal meant to say: arahael3: I though that was removed a while ago as the automatic serial console interfered with some hypervisors, could be wrong (maybe it was removed for physical machines rather than VMs) 2021-11-20 01:21:49 minimal: it's definitely still there for x86_64. 2021-11-20 01:23:21 Hmm, interesting... I just tried parallels, and it doesn't support alpine's "standard" .iso image at all. 2021-11-20 01:23:37 arahael3: so there's several things I've pointed out that you need to do. Best to get all this working wih QEMU for a x86_64 VM before then trying aarch64 2021-11-20 01:23:41 if you want to use a gui, utm is better 2021-11-20 01:25:22 Hello71: It's looking like I am going to have to run this on my mac, instead of my (headless) x86_64 system. 2021-11-20 01:25:29 Hello71: Taking a look at utm now... 2021-11-20 01:25:51 Is it this? https://getutm.app/faq/ 2021-11-20 01:25:59 yes, but https://mac.getutm.app 2021-11-20 01:28:20 running natively on mac m1 is likely to be faster than qemu tcg on any consumer x86_64 2021-11-20 01:29:20 Hello71: It works absolutely perfectly, imported the .iso image as a drive - and launched it. And I'm already in. 2021-11-20 01:30:18 arahael3: so is that running via BIOS or UEFI? 2021-11-20 01:30:31 minimal: Gosh, let me check... 2021-11-20 01:31:08 arahael3: does /sys/firmware/efi directory exist and have files inside? 2021-11-20 01:31:16 Hello71: Ooh! And it exports the 'qemu' command! 2021-11-20 01:31:34 minimal: well you said it yourself, there is no bios on arm 2021-11-20 01:31:35 Hello71: So hypothetically I could work backwards from that. 2021-11-20 01:32:02 Hello71: oh this is ARM? I thought he ws running UTM on an Intel Mac 2021-11-20 01:33:03 minimal: My headless system is x86_64, my mac is an M1 laptop, so it can run ARM64 instructions directly. 2021-11-20 01:33:23 And yes, /sys/firmware/efi exists and has files inside it. 2021-11-20 01:34:06 so the qemu command it "exports", does it specify the "-bios" option? 2021-11-20 01:35:02 minimal: Doesn't seem to. 2021-11-20 01:37:01 arahael3: sorry a "-drive if=pflash" option? This is how I specify a OVMF file for a x86_64 VM: "-drive if=pflash,format=raw,readonly,file=/usr/share/OVMF/OVMF_CODE.fd" 2021-11-20 01:38:13 Hello71: Incidentially, given how good UTM seems to be, I think I might delete Parallels! 2021-11-20 01:38:36 Hello71: It'd be virtualising this, in my case? (arm on arm) 2021-11-20 01:38:40 yes, qemu is very good, only with terrible interface and docs 2021-11-20 01:38:52 i assume utm uses hvf on mac by default 2021-11-20 01:39:01 Very nice. 2021-11-20 01:39:09 pretty easy to test, just run any cpu benchmark 2021-11-20 01:39:42 Heh, I'll probably just compile the app I want to build, a bit later, and see if it gives me a chance go grab a coffee :) 2021-11-20 01:39:46 arahael3: was that option present? 2021-11-20 01:40:12 minimal: Give me a moment, I have to export it fully, then bring up that file, etc. 2021-11-20 01:40:24 /bin/bash -c "for i in {1..100000}; do :; done" 2021-11-20 01:40:36 minimal: you don't really need to ask; utm is open source: https://github.com/utmapp/UTM/search?q=pflash 2021-11-20 01:41:07 i guess you don't know for sure that that code path is taken here but it's a reasonable guess 2021-11-20 01:41:36 arahael3: I'd also expect a 2nd option to be present also: "-drive if=pflash,format=raw,file=./uefi-config/OVMF_VARS.fd" 2021-11-20 01:41:50 minimal: Yes, I see an efi_vars.fd file. There are multiple pflash options. 2021-11-20 01:42:06 minimal: No OVMF_VARS.fd 2021-11-20 01:42:41 Hello71: had assumed UTM was commercial, and am not a Mac user. Anyway yes its source mentions both those lines: https://github.com/utmapp/UTM/blob/7fd0f08c63f68d129415f0c1d3b4ebd6d3a11981/Managers/UTMQemuSystem.m#L181 2021-11-20 01:43:35 arahael3: the 1st pflash line I mentioned points to the UEFI code for QEMU to use, the 2nd pflash line points to a file you create *per VM* to store the EFI variables (which are stored on flash in a real machine) 2021-11-20 01:43:38 yes 2021-11-20 01:43:40 Hello71: Thanks for your help, I think for now, I should just use UTM on the laptop. I'd prefer this to be on the server (as a build machine, really), but this will work. 2021-11-20 01:44:35 arahael3: for running a aarch64 UEFI machine on x86_64 you'd replace those OVMF references with appropriate AAVF references (assuming you have to AAVF code installed) 2021-11-20 01:45:13 UTM will certainly be faster (assuming M1 has "HV" CPU flag and UTM supports it) 2021-11-20 01:45:35 minimal: I think when it comes down to it, when I need this going, perhaps I'll just pay you or Hello71 for whatever magical incantation does the trick, but for now, the UTM works fine. 2021-11-20 01:46:17 minimal: pretty sure you need os support to use arm hv 2021-11-20 01:46:31 you can't just go setting up shadow page tables yourself 2021-11-20 01:46:33 Hello71: doesn't MacOS provide that? 2021-11-20 01:46:36 yes 2021-11-20 01:46:51 I thought he is running UTM on MacOS 2021-11-20 01:46:56 minimal: MacOS, from what little I understand, pretty much insists that applications uses MacOS's hypervisor framework. 2021-11-20 01:47:39 arahael3: I haven't basically used Mac system the System 7 days :-) 2021-11-20 01:47:48 s/ the/since the/ 2021-11-20 01:47:48 minimal meant to say: arahael3: I haven't basically used Mac systemsince the System 7 days :-) 2021-11-20 01:48:25 minimal: It's changed... A bit... 2021-11-20 01:48:31 though I did use NextSTEP extensively before it became OS-X 2021-11-20 01:48:42 minimal: It's changed a lot. 2021-11-20 01:50:10 QEMU is very flexible but its got so many bells and whistles it take time to get it configured. Once I figure stuff out I put them in scripts and never look at them again lol 2021-11-20 01:51:20 minimal: Yeah, well, I'd rather have someone else figure that crap out. :( 2021-11-20 01:52:07 arahael3: doh! after all that I just realised that I have notes written for running aarch64 on qemu: https://github.com/dermotbradley/create-alpine-disk-image/blob/main/Virtual/QEMU.md in the "Using a disk image" section 2021-11-20 01:52:08 minimal: It's a distraction from my real task: Build my program for aarm64. 2021-11-20 01:53:19 if you wrote your program in a portable way, normally you can just use a cross-compiler and test with qemu-user 2021-11-20 01:53:22 minimal: I think I'll *definitely* give that a go next time I look at running this on the server! And wow, those are long commands. 2021-11-20 01:53:36 Hello71: Haskell is notoriously difficult to set up as a cross compiler. :( 2021-11-20 01:53:40 cross-compiling will certainly be vastly faster than compiling under qemu tcg 2021-11-20 01:53:47 hm, maybe 2021-11-20 01:55:00 Hello71: The root problem is the way it allows developers to "splice" compile-time programs into the source which then gets compiled. These expect to run on the target system. 2021-11-20 01:55:09 mm. 2021-11-20 01:55:23 Hello71: (Ie, what they call "template haskell". It's a mess in an otherwise wonderful dev environment) 2021-11-20 01:55:52 arahael3: I'll try and find some time to re-test that command as its been a while since I actually ran a aarch64 machine with qemu. You'll see further down that page the 2 different commands to run a BIOS-based x86_64 VM and UEFI-based x86_64 VM 2021-11-20 01:56:35 minimal: Let me know how it goes, if you're successful, I'll definitely (eventually) run your scripts! 2021-11-20 01:56:54 I had no idea figuring out qemu was this painful. 2021-11-20 01:59:42 Hello71: it's imho definitely a wart in Haskell. It's one reason, I wonder, that their javascript target basically requires a docker image that needs 16GB of memory (ghcjs). People have gotten cross compilation from x86 to arm working, but it requires a complex nix-based setup. 2021-11-20 01:59:57 mmhmm 2021-11-20 02:00:04 But for building things natively... It's a dream. 2021-11-20 02:00:10 I hope they eventually fix it. :( 2021-11-20 02:28:26 arahael3: ok, tested an old aarch64 image I had lying around with qemu: the only thing I had to tweak from the command in the doc I pointed you to was to add "-serial stdio" 2021-11-20 02:45:07 busybox pgrep -a is only showing command names, not full command lines for me 2021-11-20 02:45:26 did I find a bug? 2021-11-20 02:49:33 busybox pgrep has no -a 2021-11-20 02:50:31 linux pgrep -a does one thing, bsd pgrep -a does something else, busybox pgrep just doesn't have it 2021-11-20 02:52:38 then why does -h claim it does‽ 2021-11-20 02:53:49 s/-h/busybox's --help/ 2021-11-20 02:53:49 asdflkj_sh meant to say: then why does busybox's --help claim it does‽ 2021-11-20 03:01:23 no idea. busybox's documentation elides it, but the code is there. make sure your search regex is fewer than 15 characters. 2021-11-20 03:05:31 asdflkj_sh: which Alpine version are you on? running "pgrep -a sshd" on Edge gives me "1792 sshd.pam: /usr/sbin/sshd.pam [listener] 0 of 10-100 startups" 2021-11-20 03:07:45 asdflkj_sh: actually you're right, its giving some more info for sshd and dhcpcd but not command line options 2021-11-20 03:13:25 minimal: nice! 2021-11-20 03:14:07 I'm on postmarketOS stable, perhaps I should have asked on #postmarketos 2021-11-20 03:15:09 busybox v1.33.1 2021-11-20 06:35:50 hello all, good morning 2021-11-20 08:50:51 ikke : are you online Sir? 2021-11-20 08:51:51 i want some help Sir.. 🙏 2021-11-20 09:00:45 Ganonk want food. feed Ganonk. 2021-11-20 11:28:55 Hi all, I would like to update postgresql from version 13 to 14 with my databases and tablespaces getting moved automatically. Any way to achieve this? 2021-11-20 14:48:41 hmm have one dp and one hdmi, the turned off the dp for a few days, and now when i turn on it just show the screen half a sec and goes back to black... 2021-11-20 14:48:55 i might have booted with it off.. 2021-11-20 14:49:07 do i have to reboot? 2021-11-20 15:05:33 it would have been quicker by now to just reboot 2021-11-20 15:07:27 didnt even help to reboot, could the monitor be broken if it shows the screen for a small sec only? 2021-11-20 15:07:44 i should never have turned it off! 2021-11-20 15:13:33 frag: happened to me few times already when screen stuck in sleep mode but wasnt fully in standby cus of led was green but somehow backlight was really off 2021-11-20 15:14:58 frag: reboot didnt help but had to disconnect screen from power for at least minute 2021-11-20 15:14:59 MY-R: how did u fix? 2021-11-20 15:29:41 ah 2021-11-20 15:33:46 can u turn it on while not connected ? if yes, some kind of reset setting can help if it has some crap menu 2021-11-20 15:37:57 helby: thats seems to have done the trick :]] 2021-11-20 15:41:18 once I had something similar, it inherited some shitty settings from OS and then did not function properly 2021-11-20 23:52:02 Hi, is there a way to run multimon-ng on a alpine machine? 2021-11-20 23:52:10 https://github.com/EliasOenal/multimon-ng 2021-11-20 23:57:40 did you try downloading and compiling it 2021-11-20 23:58:05 it claims that "So far multimon-ng has been successfully built on Arch Linux, Debian, Gentoo, Kali Linux, Ubuntu, OS X, Windows and FreeBSD." so i assume it's at least reasonably portable 2021-11-20 23:58:34 Hi everyone, I am on alpine edge and recently a kernel upgrade broke my system's video (something with radeon gpu firmware). I was surprised to see that the old kernel was uninstalled. Is there a way to keep old kernels while installing new ones in alpine? 2021-11-21 00:00:00 there is not currently a built-in method 2021-11-21 00:00:18 you can enable apk cache 2021-11-21 00:01:50 i am missing some libraries i need - how do i find a package that includes it? 2021-11-21 00:02:17 trying to get haskell installed for arm. 2021-11-21 00:04:12 https://pkgs.alpinelinux.org/contents 2021-11-21 00:10:02 that seems to have no results for even "libt". 2021-11-21 00:10:20 So you think it is a chance get it to run on alpine? 2021-11-21 00:11:22 Hello71: specifically, i am looking for libtinfo.so.6 and libnuma.so.1 2021-11-21 00:11:43 "Use * and ? as wildcards" 2021-11-21 00:12:14 I never compiled on Alpine, is there a Place were I'm able to find out what I need to compile basicly? 2021-11-21 00:12:20 i have some skepticism that ghc will work unmodified on musl 2021-11-21 00:12:23 ghc for glibc 2021-11-21 00:14:59 hmm. libtinfo* doesnt seem to exist 2021-11-21 00:15:48 Hello71: yeah, alpine doesn't package ghc for arm, it seems. hoping the upstream works but yeah it's not promising. 2021-11-21 00:16:04 i might end up needing to cross compile properly. 2021-11-21 00:16:25 feuerteufel: https://wiki.alpinelinux.org/wiki/Creating_an_Alpine_package 2021-11-21 00:16:49 aarch64 is probably possible but needs bootstrap 2021-11-21 00:17:00 why don't you just use a glibc distro 2021-11-21 00:17:11 or just make the user compile it themselves 2021-11-21 00:18:24 Hello71: because i am the user, and i havent actually written this server yet, but i *want* to use haskell on it. 2021-11-21 00:19:34 sounds optimistic 2021-11-21 00:21:48 probably. i might well pick a different language but lets see. 2021-11-21 00:23:40 looks like musl is supported. and aarch64 is supported. but not yet both together. i might be able to compile it. 2021-11-21 00:33:34 ikke: I think that should be at least the second step, first I just want to try if the tool is running on alpine. Do you think, it will be enough just to install "build-base"? 2021-11-21 00:33:57 Thats 2021-11-21 00:34:13 That gets you a compiler and other related tools 2021-11-21 00:36:39 So that's what I need for a first try? 2021-11-21 00:40:11 Ok, it looks like I have to go the "hard way" --> trial-and-error method :-) 2021-11-21 00:42:16 I'll report what I find out ... 2021-11-21 01:12:04 OK, basically I'm able to compile. But now I have to solve dependence's ... But it's enough for tonight 2021-11-21 01:12:49 Tks for help! I'll report further on ... 2021-11-21 05:15:54 it seems aarch64 has neither haskell nor rust? 2021-11-21 05:29:19 it has rust 2021-11-21 05:31:46 Just found it - in community repos - that works! 2021-11-21 10:15:22 is you copy a line with tabs into primary and paste it with shift-ins or mouse, tabs are removed?? 2021-11-21 10:15:45 xsel pasts with tabs 2021-11-21 10:22:00 copy it with what 2021-11-21 10:24:09 From a terminal? 2021-11-21 10:24:30 It's typically your terminal that converts tabs to spaces 2021-11-21 10:25:03 i use a c-code, but i guess xsel is similar 2021-11-21 10:25:20 got same result in st andxterm 2021-11-21 10:25:25 got same result in st and xterm 2021-11-21 10:26:07 also same result in irssi (maybe thats the same as terminal not sure) 2021-11-21 10:26:15 yes 2021-11-21 10:26:29 and in qutebrowser 2021-11-21 10:29:10 both types of paste paste with tabs preserved in my terminal 2021-11-21 10:30:35 c-s-v, shift-ins for 'primary', wl-paste all have the tabs in the output, former two pasted into vim 2021-11-21 10:31:03 psykose: what terminal do you use? 2021-11-21 10:31:05 foot 2021-11-21 10:31:31 never heard of 2021-11-21 10:31:32 you are testing by pasting into vim or something right? the shell can reject tabs if it wants, or at least fish strips them when you paste into it 2021-11-21 10:32:38 if xsel paste has the tabs then either the shell or the terminal eats them after the hotkey press 2021-11-21 10:33:27 in testing with pasting in terminal/irssi/qutebrowser 2021-11-21 10:33:51 *im 2021-11-21 10:34:54 got the same in sh 2021-11-21 10:35:05 (normally use zsh) 2021-11-21 10:35:22 maybe st then... jaja 2021-11-21 10:35:47 st strips them too it seems 2021-11-21 10:37:19 qutebrowser : menu pastes them correctly 2021-11-21 10:37:23 websites convert them to spaces 2021-11-21 10:37:35 seems pretty normal overall 2021-11-21 10:39:00 hmm i use this in qb: config.bind('', 'insert-text {primary}', mode='insert') 2021-11-21 10:40:31 my irc client seems to get them correctly and send them correctly too, tabs on receiving end 2021-11-21 10:40:41 mhmhm 2021-11-21 10:40:55 can prob modify st 2021-11-21 10:41:35 the qb test was just pressing C-v in : 2021-11-21 10:41:59 c-v is clipboard not primary no? 2021-11-21 10:42:12 that shouldn't make a difference 2021-11-21 10:42:52 yep, works the same with middle click on the : tab :) 2021-11-21 10:43:01 ah maybe not, i seem to get same result, not used to using clipboard tho 2021-11-21 10:43:32 ah middle click in qb works correctly 2021-11-21 10:43:46 jaja 2021-11-21 10:45:29 psykose: do you have clipboard manager btw? 2021-11-21 10:45:33 nope 2021-11-21 10:45:36 okok 2021-11-21 10:45:53 i would /hope/ clipboard managers don't try to filter anything put into them, but it's possible 2021-11-21 11:36:20 good morning 2021-11-21 11:36:37 I am still unable to netboot alpine aarch64 on the rpi 4b https://lists.alpinelinux.org/~alpine/users/%3Cb444fbf4-c2f0-9ab9-9a5a-8c68725381ca%40googlemail.com%3E 2021-11-21 11:37:17 despite having added the udhcp features to mkinitfs and rebuilt the initramfs 2021-11-21 11:37:59 the error has now changed from /dev/nfs: Can't open blockdev to Mounting root: failed. 2021-11-21 11:38:25 anybody here managed to netboot alpine before? 2021-11-21 11:46:12 uhm, this guy uses a pxe bootloader https://blog.skunkw0rks.io/blog/2018/09/03/pxe-boot-alpine-linux-pt.-1/ 2021-11-21 11:46:20 interesting 2021-11-21 11:47:01 many times, but not with nfs 2021-11-21 11:51:53 clandmeter, how then? 2021-11-21 11:52:05 I am not convinced as to the need of a pxebootloader 2021-11-21 11:52:33 https://boot.alpinelinux.org/ 2021-11-21 11:53:09 im not sure this will work on rpi 2021-11-21 11:56:40 apkvol would be the rootfs ? 2021-11-21 11:56:50 I am confused 2021-11-21 11:57:02 root will be in ram 2021-11-21 11:57:28 I want root to be on storage instead 2021-11-21 11:57:34 as persistant install 2021-11-21 11:57:39 that was not your question :) 2021-11-21 11:57:43 ah okay 2021-11-21 11:57:45 true 2021-11-21 11:57:47 :) 2021-11-21 11:58:00 thats why i added, not with nfs 2021-11-21 11:58:37 but i guess you could still mount things via fstab 2021-11-21 11:59:48 it could be we are missing something in the kernel to mount nfs properly, but i never played with it. 2021-11-21 12:00:20 I think some ppl have done it before, so i would assume it should somehow work. 2021-11-21 12:01:16 do you end up in emergency shell? 2021-11-21 12:01:39 or the kernel just hangs 2021-11-21 15:09:36 hm, what would be the way to rename the source file in APKBUILDs? (except rewriting a build step) 2021-11-21 15:11:31 kitzman: :: 2021-11-21 15:11:33 in source 2021-11-21 15:11:42 nice thank you 2021-11-21 15:12:06 $pkgname-$pkgver.tar.gz::https://github.com/project/name/archive/v$pkgver.tar.gz 2021-11-21 15:12:14 Latrina: https://git.disroot.org/kitzman/pxe-utils/src/branch/master/pxe-root/grub_example.cfg this is how I successfully mounted an NFS share. but it's a different kernel. also i know i had issues with UID/GID mapping so I disabled it for root. 2021-11-21 15:16:24 kitzman: they are using a RPI so there's no Grub involved 2021-11-21 15:17:28 true, i missed that; i don't know if the kernel option would be the same - on my rpi i didn't use the upstream kernel. 2021-11-21 15:32:49 Latrina: I think the problem is that CONFIG_IP_PNP and CONFIG_IP_PNP_DHCP are not enabled in the RPI kernels 2021-11-21 15:34:39 do you use firefox? is it "ok" for privacy? (or is there a guide to follow to make it more private?) 2021-11-21 15:36:44 frag: that's not really an Alpine question, isn't there a firefox irc channel? 2021-11-21 15:38:13 well afaiu its one of the very few browser that works well on alpine/musl 2021-11-21 15:41:28 frag: agreed but detailed Firefox configuration discussion isn't really Alpine-specific. There's plenty of info online at places like ghacks 2021-11-21 16:49:22 when trying to upgrade I get a conflict between emacs-nox-27.2-r1 and emacs-x11-27.2-r1. I have emacs-x11-27.2-r1 installed at the moment. Anyone seen that before? I'm on edge and using `apk upgrade -a` 2021-11-21 16:52:15 telmich: what is the message you get? 2021-11-21 16:52:32 and what repostiries do you have? 2021-11-21 16:56:29 ikke: https://www.nico.schottelius.org/temp/apk-upgrade-emacs-2021-11-21.txt 2021-11-21 16:57:12 interesting 2021-11-21 16:59:30 telmich: grep emacs /etc/apk/world 2021-11-21 17:00:00 telmich: btw, you probably know this already, but you can directly use https on dl-cdn.a.o nowadays 2021-11-21 17:12:30 hey kitzman, I'll take a look, thanks 2021-11-21 17:12:49 minimal, how come? 2021-11-21 17:13:13 I have also tried static but got the same result 2021-11-21 17:13:24 either way, I am okay to use a different kernel 2021-11-21 17:19:39 Latrina: how come what? 2021-11-21 17:20:02 home come the alpine kernel doesnt have CONFIG_IP_PNP and CONFIG_IP_PNP_DHCP ? 2021-11-21 17:20:28 I mean I guess I can build a new kernel if thats whats needed 2021-11-21 17:20:55 Latrina: because it doesn't 2021-11-21 17:27:27 Latrina: have you tried adding "debug_init" to the cmdline options to see exactly that the initramfs is doing? 2021-11-21 17:28:42 minimal, nope. Good one. let me see 2021-11-21 17:48:30 ikke: I've added the grep to https://www.nico.schottelius.org/temp/apk-upgrade-emacs-2021-11-21.txt and it contained a "funky" line: 2021-11-21 17:48:37 notmuch-emacs> not sure how this made it into world 2021-11-21 17:49:09 Ok, can you try to remove that and run apk fix? 2021-11-21 17:49:27 removing it -> apk upgrade works directly 2021-11-21 18:12:37 it seems the latest ugprade made xorg dpi aware / changed the dpi of my screen -> finally menus are rendering normal :-) 2021-11-21 18:13:28 heh 2021-11-21 18:13:31 yes 2021-11-21 18:13:36 for nmeum, the opposite happened 2021-11-21 18:14:56 https://lists.x.org/archives/xorg/2021-October/060799.html 2021-11-21 18:15:40 ^ this xorg >= 21.1.0 has some sort of revised DPI handling 2021-11-21 19:31:47 Hi, what is the right path for auto disabling and enabling wlan0 when I connect a wire to eth0? 2021-11-21 19:35:40 route metric 2021-11-21 19:40:11 uhM 2021-11-21 19:44:00 It means that it will prefer routes with the lower metric 2021-11-21 19:44:22 and routes from wired interfaces should have a lower metric as routes from wireless interfaces by default 2021-11-21 19:46:01 yes but if I disconnect the cable 2021-11-21 19:46:09 the route via eth0 is not removed 2021-11-21 19:47:16 also if I'm gonna stay hours with the wired connection, it seems better to just disable wireless 2021-11-21 19:48:24 the route should be removed 2021-11-21 19:48:32 because the interface goes down 2021-11-21 19:49:00 uhM 2021-11-21 19:49:34 dmesg shows [14285.246833] r8169 0000:04:00.0 eth0: Link is Down 2021-11-21 19:49:45 but I can confirm that the route was not removed 2021-11-21 19:49:48 let's check again 2021-11-21 19:50:41 same 2021-11-21 19:50:53 the route was removed after manually "ifconfig eth0 down" 2021-11-21 19:52:59 i think it requires network management program to remove, otherwise it would be lost when link comes back 2021-11-21 19:53:08 but iirc kernel should skip routes with link down? 2021-11-21 19:53:34 it seems that not 2021-11-21 19:54:48 I just lose connection 2021-11-21 19:56:18 apparently it needs sysctl ignore_routes_with_linkdown 2021-11-21 19:56:36 Why would it make sense to not ignore routes that are down? 2021-11-21 19:56:47 You cannot send traffic there anyway 2021-11-21 19:57:39 historical reason i guess 2021-11-21 19:58:04 So it prefers to drop traffic rather than route it to the interface that is up? 2021-11-21 19:58:36 yes 2021-11-21 19:58:46 yeah 2021-11-21 19:58:49 now works 2021-11-21 19:58:54 🤔 2021-11-21 19:59:01 probably same reason why you need CAP_NET_ADMIN to use SO_BINDTODEVICE 2021-11-21 19:59:43 because admin may be using routing table for access control and will not expect packets to go out an "unexpected" interface 2021-11-21 20:00:15 ignore_routes_with_linkdown is apparently only in kernel since 2015, before that you needed to use userspace daemon to manage route status 2021-11-21 23:56:29 Latrina: i just checked, it seems you need also CONFIG_ROOT_NFS which depends on IP_PNP 2021-11-22 00:15:27 Latrina: i asked ncopa to apply this: https://tpaste.us/41mO lets see if he agrees on it. 2021-11-22 01:34:08 jonathon: xanadu 2021-11-22 02:13:00 jonathon: xanadu 2021-11-22 08:06:55 good morning 2021-11-22 08:08:07 can that nfs patch be sent to upstream rpi kernel? https://github.com/raspberrypi/linux 2021-11-22 08:55:37 morning 2021-11-22 08:55:42 not sure its a feature or a bug? 2021-11-22 08:57:52 looks like its already enabled, weird. 2021-11-22 09:21:04 ncopa: https://git.alpinelinux.org/mkinitfs/commit/?id=a3671aa297db852c9b5fedd64a713bc6d7fb9d4f 2021-11-22 09:24:03 Latrina: can you try to boot as described in this commit msg? 2021-11-22 10:45:31 somethign broke my bluetooth since last week. bluetoothctl no longer list any adapters. 2021-11-22 10:45:50 bluetoothctl devices gives error: No default controller available 2021-11-22 12:05:27 The last user of bluetooth discovered! 2021-11-22 12:09:21 it was kernel 5.15.3 that broke bluetooth. it works again with 5.15.4 2021-11-22 12:15:37 bt not popular? 2021-11-22 12:16:00 i use it for my k380 kb, but im a bit worried about health risks tbh 2021-11-22 12:16:37 i guess bt headphones are pop 2021-11-22 12:37:31 Hello71: is there something like UTM for android devices, or I can just use qemu ? 2021-11-22 12:38:28 dunno 2021-11-22 13:22:43 clandmeter: the mkinitfs init script currently looks different from that 2013 MR and does not contain some of those specific lines. It is also not clear if nlplug-findfs supports mounting a NFS rootfs 2021-11-22 14:58:08 can 'bb/date' or other basic tool display milliseconds since 1970? 2021-11-22 14:58:36 heh, just times by a thousand 2021-11-22 15:00:23 date +%s%N should give you nanoseconds since the epoch. 2021-11-22 15:00:49 Get rid of the last six characters and you should have your milliseconds. 2021-11-22 15:02:01 seems busybox date doesnt have %N, but i have $EPOCHREALTIME and zsh/strftime, should work :] 2021-11-22 15:07:30 There are advantages to having the kitchen sink of zsh available. ;-) 2021-11-22 15:12:44 very useful :] 2021-11-22 17:21:13 minimal: ok, i have no clue, i never used nfs root. 2021-11-22 17:21:25 would be nice to know if its actually still supported 2021-11-22 17:23:26 people still use nfs? sorry. 2021-11-22 17:23:31 after doing 'ifdown wlan0' its related wpa_suplicant process shouldn't be killed? 2021-11-22 17:25:02 uhM , it seems that ifup/ifdown doesnt' handle it 2021-11-22 17:26:46 9p is significantly saner than nfs, I don't think it's particularly more practical though. 2021-11-22 17:29:40 whats wrong with nfs? 2021-11-22 17:30:16 clandmeter: not sure if Alpine ever supported it - the "netboot" download for example is for a run-from-ram 2021-11-22 17:30:41 jhjacobs81: someone is having problems setting up/using a RPI with a NFS rootfs 2021-11-22 17:31:01 yes, i see that :) 2021-11-22 17:31:43 i was refering to invoked's message of "people still use nfs? sorry." 2021-11-22 17:31:50 sounds as though there are better options/ 2021-11-22 17:33:48 jhjacobs81: apart from NFS and SMB/CIFS I don't think there are really any other general purpose "industry standard" network filesystems. Block setups like iSCSI and NBD could be used for booting but they're not filesystems. 2021-11-22 17:34:23 there's lots wrong with nfs, and there's nothing better available for linux. both can be true 2021-11-22 17:34:26 yeah, i was half-shitposting. sorry x2 2021-11-22 17:39:16 :) 2021-11-22 17:39:42 okay, just making sure :) i would hate to setup NFS only to discover there's something better :D 2021-11-22 17:43:55 this will be fun https://github.com/rust-lang/team/pull/671 2021-11-22 17:44:02 jhjacobs81: I did mention SMB/CIFS as an alternative possibility 2021-11-22 17:44:11 *sad. This will be sad 2021-11-22 17:47:20 wait.. wut?? i am just in the process of learning Rust! 2021-11-22 17:47:23 and Alpine! 2021-11-22 17:50:13 the whole thing reads like: "under the advice of my attorney, i cannot say anything specific but..." 2021-11-22 17:50:51 this stuff seems way too corporate. 2021-11-22 17:52:06 about the same shit as with freenode.. 2021-11-22 17:52:26 or maybe not.. but to me it looks like the same :) 2021-11-22 17:53:38 it's a lot of words saying nothing. how are you supposed to have a constructive conversation about governance if you don't even explain what went wrong 2021-11-22 17:54:10 btw. does freenode still exist? few days ago I tried join to check it and it failed 2021-11-22 17:54:51 then read about some 'kingdom' on their website ... 2021-11-22 17:55:06 if nobody cares whether it exists, does it exist 2021-11-22 17:55:46 just curious if some projects are still there or it becomes just group of some new people 2021-11-22 17:56:24 even their webclient did not connect 2021-11-22 17:56:51 once word got around that it was taken over by a cocaine-addled wannabe korean aristocrat, it stopped existing 2021-11-22 17:57:36 even fans of cocaine don't want to be there 2021-11-22 18:00:55 on other hand discovered ergo servers and it has cool things 2021-11-22 18:12:13 *sigh* my alpine booted to grub rescue again.. 2021-11-22 18:12:28 and really none of the normal howto's seem to work 2021-11-22 18:14:19 "unknown command linux" 2021-11-22 18:22:26 what's wrong with your alpine? ;/ 2021-11-22 18:24:35 "*sigh* my alpine booted to..." <- What's the problem? 2021-11-22 18:25:49 well, i seem to be the problem :) 2021-11-22 18:26:07 i now installed grub-bios, at least i got to a initramfs emergency shell 2021-11-22 18:26:28 the problem is grub obviously 2021-11-22 18:26:33 true 2021-11-22 18:27:03 what i did, was create a vps with debian on it. then i bootstrapped alpine to a secundairy partition. this works and i can boot off it 2021-11-22 18:27:30 but obviously i then want to delete the first partition and have grub boot from the alpine partition 2021-11-22 18:27:50 so, i start cfdisk, make vda2 bootable, delete vda1 and then run grub-install 2021-11-22 18:27:52 helby: Not fact 2021-11-22 18:28:10 that says "grub-install ran without problems" 2021-11-22 18:28:38 i can reboot, select alpine which then boots into initramfs, complains about sysroot not beeing able to mount and drops me to an emergency shell 2021-11-22 18:28:40 of course, 9/10 sane people don't recommend grub OS over another OS 2021-11-22 18:28:54 well, call me unsne then 2021-11-22 18:28:57 unsane 2021-11-22 18:29:05 that's insane ;/ 2021-11-22 18:29:10 but it got me further then syslinux 2021-11-22 18:29:15 jhjacobs81: but did you ensure you /etc/default/grub file had appropriate settings before running grub-install? 2021-11-22 18:29:42 i thought so, but obviously i did not :) 2021-11-22 18:29:56 im now restoring a backup, for retry 1938111 2021-11-22 18:29:58 :D 2021-11-22 18:30:13 jhjacobs81: and did you check the resultant /boot/grub/grub.cfg contents to see that they made sense before rebooting? 2021-11-22 18:30:41 no.. 2021-11-22 18:30:47 will do that this time :) 2021-11-22 18:33:47 well, /etc/default/grub is nothing special.. timeout settings and so. grub.cfg looks good 2021-11-22 18:35:17 now looking at https://wiki.alpinelinux.org/wiki/Create_A_VirtualBox_Guest_with_Grub_and_XFS 2021-11-22 18:43:17 follow wiki is way to disaster ;/ 2021-11-22 18:43:34 first command on that page 'apk add nano' that tells a lot ;/ 2021-11-22 19:05:06 Ariadne would probably disagree with you 2021-11-22 19:05:35 i don't want to speak for anybody though. 2021-11-22 19:06:26 whats wrong with nano? 2021-11-22 19:06:57 absolutely nothing 2021-11-22 19:07:00 It’s overfeatured. 2021-11-22 19:08:04 heh 2021-11-22 19:09:07 nano's fine. not my jam, not my jelly, not my peanut butter, but Ariadne's post about nano made me reconsider some things. 2021-11-22 19:09:33 invoked: would you have a link to that post? I’m curious. 2021-11-22 19:10:06 nano seems like a perfectly reasonable editor to suggest to somebody who is new to the UNIX environment 2021-11-22 19:10:24 true 2021-11-22 19:10:43 hoel: https://ariadne.space/2021/08/13/gnu-nano-is-my-editor-of-choice/ 2021-11-22 19:10:44 or would you suggest that we suggest people use vim, and forget to mention `:x` or `:wq!` like every other tutorial involving vim does? 2021-11-22 19:10:59 maybe we can go wild, and suggest Emacs 2021-11-22 19:11:08 let's use ed 2021-11-22 19:11:14 ed obviously, that's the standard editor ;/ 2021-11-22 19:11:15 sam -d 2021-11-22 19:11:31 but yes, nano is relatively small compared to vim monster 2021-11-22 19:11:56 clearly, the user would be served by learning to type `C-x C-c (kill-terminal)` 2021-11-22 19:11:58 to quit 2021-11-22 19:13:54 I think I never quit emacs properly, usually killed window 2021-11-22 19:14:06 I still use joe :-) 2021-11-22 19:14:19 the only good thing about emacs is org-mode 2021-11-22 19:14:22 ACTION throws gas on the fire 2021-11-22 19:14:29 Wordstar keystrokes for the win! 2021-11-22 19:14:31 I usually quit emacs by rebooting into an installer for a better OS 2021-11-22 19:14:43 anyway, back to working on my EDLIN.COM port 2021-11-22 19:15:11 what's that? 2021-11-22 19:15:30 the editor that ed(1) ripped off. 2021-11-22 19:16:18 OH! 2021-11-22 19:16:26 somehow grub thinks its botting from ext2 2021-11-22 19:16:50 so when i drop to the emergency shell, and i manually mount /dev/vda2 as ext4 on /sysroot then it works 2021-11-22 19:19:53 maybe you did not load ext4 module? 2021-11-22 19:20:22 im very new to Alpine :) 2021-11-22 19:21:19 google tells me insmod ext2 is for ext2, ext3 and ext4 2021-11-22 19:22:51 jhjacobs81: you mean the Grub ext2 module? (as in "insmod ext2" in grub.cfg? that is correct as ext2 module seems to handle ext2/3/4 2021-11-22 19:23:31 the grub menu for alpine is correct, including UUID and all 2021-11-22 19:23:50 so why then wont it boot untill i manually mount /sysroot :( 2021-11-22 19:24:06 you can mount an ext4 fs as ext2, but you probably should not 2021-11-22 19:25:17 minimal: what other options are specified on the "linux" line in grub.cfg? 2021-11-22 19:25:31 s/minimal/jhjacobs81/ 2021-11-22 19:25:31 minimal meant to say: jhjacobs81: what other options are specified on the "linux" line in grub.cfg? 2021-11-22 19:25:41 hanh on :) 2021-11-22 19:26:19 linux /boot/vmlinuz-virt root=UUID=7bff1984-7c8e-4459-a3d3-85bd7e4d69a0 ro 2021-11-22 19:26:35 7bff is indeed the right partition 2021-11-22 19:26:48 maybe i should use vanilla kernel? 2021-11-22 19:27:29 truth be told, i'd much rather move to syslinux, but i simply cant get it to work 2021-11-22 19:28:19 jhjacobs81: I'd have expected additional options, i.e. "rootfstype=ext4 console=tty0" (though the console one is probably optional) 2021-11-22 19:29:08 jhjacobs81: this is BIOS boot, right? or is it UEFI? 2021-11-22 19:29:23 bios indeed 2021-11-22 19:29:35 on debian i am used to other options as well 2021-11-22 19:29:55 i'll take another shot at syslinux again :) 2021-11-22 19:30:35 jhjacobs81: the other options would typically be defined in /etc/default/grub as I mentioned earlier 2021-11-22 19:32:09 i.e. add GRUB_CMDLINE_LINUX_DEFAULT="rootfstype=ext4 console=tty0" to /etc/default/grub before you run grub-install 2021-11-22 19:32:54 did not you xfs instead of ext4? 2021-11-22 19:33:27 funny.. now syslinux works without a hitch 2021-11-22 19:33:39 i wouldnt dream of using xfs 2021-11-22 19:33:51 havent been using anythong other then ext4 or btrfs for years 2021-11-22 19:34:59 oh it was not u who mentioned xfs? 2021-11-22 19:35:19 helby: he mentioned a wiki article that mentioned xfs 2021-11-22 19:35:41 I remember somebody wanted xfs today ;/ 2021-11-22 19:35:59 well.. call me stupid 2021-11-22 19:36:05 but syslinux works now 2021-11-22 19:36:16 and i have no clue what i did "special" to make it work 2021-11-22 19:36:45 jhjacobs81: both syslinux and grub should be straightforward to setup 2021-11-22 19:37:01 yes.. well.. 2021-11-22 19:37:19 in any case, now it is fixed, i can template the vps :D 2021-11-22 19:39:38 now i love Alpine 2021-11-22 20:00:24 and awall 2021-11-22 20:09:47 minimal: yes i know, but i think it worked before as else the commit doesn't make sense. 2021-11-22 20:12:05 clandmeter: if you're referring to mkinitfs, it looks like the code has changed a bit since that old commit and some of the stuff added then is no longer in place 2021-11-22 20:13:05 clandmeter: I've been meaning to test NFS boot (on x86_64) for some time so I may dig into it in the next week or two 2021-11-22 20:13:09 im just referring to your comment that it probably never worked. 2021-11-22 20:13:28 yes, that would be nice. 2021-11-22 20:13:49 i also need to update boot.a.o 2021-11-22 20:13:57 or anyone else who wants to :) 2021-11-22 20:52:45 Could u build an accespoint with Alpine? *thinking out loud* 2021-11-22 20:53:14 It's certainly possible 2021-11-22 20:53:47 But you would not advise it? :) 2021-11-22 20:54:39 i used to like Ubiquiti, but im running into much issues lately 2021-11-22 20:55:32 the new ubiquiti gear is a whole other creature from what it used to be 2021-11-22 20:55:37 sadly 2021-11-22 20:56:01 Indeed 2021-11-22 20:57:51 Im looking into mikrotik atm, 2021-11-22 20:58:16 i was about to mention mikrotik :) 2021-11-22 20:58:44 they're still good i think. you might be able to find smaller makes from smallnetbuilder forum 2021-11-22 20:59:26 I have one but i think im switching back to ubiquity 2021-11-22 21:00:14 i have feeling the performance is less and i have no time to figure out on how to improve it. 2021-11-22 21:00:17 Why is that? 2021-11-22 21:00:21 ahh :) 2021-11-22 21:00:36 it should be plug and play for me. 2021-11-22 21:00:45 but if you are into it, why not make one with alpine :) 2021-11-22 21:01:56 Yeah :) i have one of those small 4 ports router thingy’s without OS 2021-11-22 21:02:14 plan to put Alpine on it once im confident enough with awall :) 2021-11-22 21:02:23 maybe even build a simple web gui for it 2021-11-22 21:02:48 small 4 port router, thats not really specific 2021-11-22 21:03:08 I know 2021-11-22 21:04:01 https://www.reddit.com/r/AlpineLinux/comments/qztzd0/viability_of_using_alpine_as_a_daily_driver_linux/ 2021-11-22 21:04:01 [REDDIT] Viability of using alpine as a daily driver linux desktop distro? (self.AlpineLinux) | 1 point (03100.0%) | 0 comments | Posted by QuantP1 | Created at 2021-11-22 - 19:47:37UTC 2021-11-22 21:05:16 Ofcourse now i cant find the link to it… pff.. 2021-11-22 21:07:44 A pc engines apu 2021-11-22 21:10:35 Owell, i’ll try installing alpine on it tomorrow :) 2021-11-22 21:24:48 ah apu, i have lots of them :) 2021-11-22 21:24:56 but 3 ports 2021-11-22 21:26:31 I've had Alpine as daily driver on laptop for years now. Granted, I live in the terminal and do lot of low-level fiddling anyway. 2021-11-22 21:32:24 from ubuntu lts to gentoo how up to date are the alpine repos normally 2021-11-22 21:33:33 stable releases are similar to ubuntu releases, typically not going beyond the minor version that was started with 2021-11-22 21:33:38 edge is kept up-to-date 2021-11-22 21:34:00 (which sometimes causes breakages) 2021-11-22 21:34:33 like gentoo up-to-date or arch up-to-date? 2021-11-22 21:34:51 also where are the edge isos? 2021-11-22 21:34:59 Hey, would anyone have any idea why gcc would be failing to compile with linker errors but clang would pass correctly? I'm only linking to libretls at the moment on alpine-edge. 2021-11-22 21:35:08 s/isos/iso/ 2021-11-22 21:35:08 Qixt[m] meant to say: also where are the edge iso? 2021-11-22 21:36:01 Qixt[m]: I'm not familiar with gentoo that much, so I cannot compare to Gentoo 2021-11-22 21:36:28 Arch generally puts packages in testing first before pushing it out to the main repos 2021-11-22 21:36:39 edge is where our testing happens 2021-11-22 21:37:23 ah ok 2021-11-22 21:38:14 how easy is it to setup a custom mirror if I wanted to create my own base-system package? 2021-11-22 21:38:24 quite easy 2021-11-22 21:38:52 You just need a dir with packages + an APKINDEX and a static file webserver 2021-11-22 21:39:04 oh nice 2021-11-22 21:39:11 abuild already creates valid repositories 2021-11-22 21:39:37 how's wayland+pipewire on alpine? I normally use sway on void and I want to keep my desktop 2021-11-22 21:40:06 No personal expierence, but quite some people use it 2021-11-22 21:40:45 also can't remember if I asked but where are the edge isos, I can only seem to find the regular ones 2021-11-22 21:40:59 We don't have isos for edge 2021-11-22 21:41:09 easiest is to use a stable release iso and upgrade it to edge 2021-11-22 21:41:26 ah, how do you do that? is there a wiki page 2021-11-22 21:42:27 change the repos in /etc/apk/repositories from the version to edge, and then run apk upgrade --update --available 2021-11-22 22:05:54 Hmz.. 2021-11-23 07:03:20 wake up people! 2021-11-23 07:03:46 For what 2021-11-23 07:05:23 whatever you like ;/ 2021-11-23 07:31:29 hmmm 2021-11-23 07:46:41 helby: you got lost? 2021-11-23 07:48:40 clandmeter: "if mountain will not come to helby, helby must go to mountain ;/ 2021-11-23 07:49:34 clandmeter: I am on OpenBSD right now, don't tell anybody 2021-11-23 07:50:06 I will fix it tonight ;/ 2021-11-23 07:50:14 as long as i dont have to rescue you from the mountain, all should be fine. 2021-11-23 07:52:52 clandmeter: just have permanent connection from irc client on the server, so can fuck around more servers even shitty one like oftc ;/ 2021-11-23 07:56:19 also it does not require google catpcha more so 2021-11-23 09:04:12 good morning 2021-11-23 09:07:57 so did you boot again ;/ 2021-11-23 09:08:52 oh, yes :) the VPS is fixed with syslinux now 2021-11-23 09:09:08 today im gonna try the pcengines apu thingy 2021-11-23 09:51:52 hi everyone, is it possible to use php8 with composer? it seems like the alpine package depends on php7 2021-11-23 10:57:02 tinybird: it has php8 as dependency 2021-11-23 10:57:43 in edge, not in 3.14 2021-11-23 11:33:23 thanks helby - I wondered about the edge repository, I'll look into that! 2021-11-23 11:34:10 tinybird: I assume that will become release soon with v 3.15 so 2021-11-23 18:14:41 Hmz 2021-11-23 18:16:10 hold my zombie 2021-11-23 18:16:45 IN YOURRR HEAAAAAAADDDDDD IN YOUR HEAAAAAD! ZOOMMBBIEEEEE! *lalala* 2021-11-23 19:30:47 Hmz 2021-11-23 19:47:47 is there an Alpine rss feed? 2021-11-23 19:48:10 About what? 2021-11-23 19:48:16 https://alpinelinux.org/atom.xml 2021-11-23 19:48:18 news feed 2021-11-23 19:50:34 thanks :) 2021-11-24 03:25:53 Don't suppose anybody here runs qbittorrent-nox on Alpine? I've (stupidly) run an apk upgrade and broken qbit... rebuilt my VM from scratch and now can't install qbittorrent-nox, it's complaining about libcrypto.so.3 (no such package) upon attempting to `apk add qbittorrent-nox` :( 2021-11-24 03:27:52 that is very strange, considering that exists and is provided by libcrypto3 2021-11-24 03:28:30 this it just needs to be rebuilt to use 1.1 again 2021-11-24 03:28:46 Agreed, this is on a fresh install too... just changed apk repos to include testing etc... just falls over every time on a fresh install from the 13.4.3 iso 2021-11-24 03:30:01 Do you know of any quick wizardry that might get me back up and running? I'm tempted to resort to using qbittorrent-nox-static but I'll have to manually setup init.d and that doesn't sound enjoyable :P 2021-11-24 03:30:51 not sure what -static is 2021-11-24 03:31:31 if you mean get a static build from somewhere, the most you have to change in init.d is like one line to the executable 2021-11-24 03:31:35 Prebuilt binaries, I can grab that and just inject it on my build... but I won't have all the init.d scripts etc because I can't install -nox by itself 2021-11-24 03:31:54 you can install -openrc for the init script by itself 2021-11-24 03:32:04 ACTION tries 2021-11-24 03:32:28 yeah, -openrc installs fine :D 2021-11-24 03:32:37 ACTION grabs -static and tests 2021-11-24 03:33:16 and yeah, rebuilding qbittorrent-nox again fixes it, so someone has to bump it in aports 2021-11-24 03:33:33 it would be nice to update it and libtorrent-rasterbar but the latter is quite finicky and has failing tests on everything 2021-11-24 03:33:41 and nobody cares particularly much for getting it to work 2021-11-24 03:34:23 ya, it seems to be always behind by one minor version (eg. 4.3.8 vs 4.3.9 currently) 2021-11-24 03:34:52 A few months ago I think it was still on 4.3.2 where the latest was 4.3.6 or something... I'm not a developer though, so I can't help out unfortunately :( 2021-11-24 03:37:56 think bumping that minor should be fine 2021-11-24 03:37:59 let me see if it starts at least 2021-11-24 03:47:37 -static seems to work fine, just need to figure out where it stores the config file :P 2021-11-24 03:50:13 i bumped it, when it gets merged you can just install the repo version again 2021-11-24 03:50:24 these issues usually happen for things in testing as they get neglected a lot 2021-11-24 03:51:17 You're a legend :) I'll keep my eye on it 2021-11-24 03:51:28 I get scared of the testing repo for this reason, haha 2021-11-24 03:51:36 such is life in edge-testing 2021-11-24 03:51:37 hah 2021-11-24 03:51:40 I just forgot to snapshot my system before hitting the upgrade... won't make that mistake again :P 2021-11-24 04:31:27 xMPLR: merged, should be in dl-cdn, bit more time for other mirrors to sync :) 2021-11-24 04:42:14 That's great, thanks psykose :) 2021-11-24 07:18:25 is it possible to have xorg without opengl on alpine? 2021-11-24 07:18:49 everything is possible. the impossible just takes a bit more time 2021-11-24 07:21:02 just curious, why does xorg in alpine use mesa instead of xf86 drivers? 2021-11-24 07:25:13 GenericNick2: I found that if you don't install mesa-dri-gallium and mesa-egl then you won't be able to have OpenGL acceleration (at least in a QEMU VM) 2021-11-24 07:27:33 I was wondering about the mesa-gl and mesa-gbm packages 2021-11-24 07:27:46 xorg-server depends on them 2021-11-24 07:27:54 Im beginning to think i should run a desktop on Alpine :) small, secure, privacy friendly :) 2021-11-24 07:28:14 just so long as you have all of the packages that you need 2021-11-24 07:36:24 :) 2021-11-24 07:38:59 GenericNick2: are the xf86 drivers not legacy drivers for cards that don't support the drm or whatever directly 2021-11-24 07:39:07 you can still install them and configure xorg to use them yourself 2021-11-24 07:40:00 xf86-video-intel seems to depend on mesa too 2021-11-24 07:41:36 i'm not sure how you get anything to work without mesa 2021-11-24 07:56:07 ktprograms: Well you definitely need a graphics driver to get stuff done like psykose mentioned. 2021-11-24 07:56:18 mesa-dri-gallium is a software renderer though and will be quite slow. 2021-11-24 07:57:17 It's better to try and use the guest additions graphics drivers, but I have not tried to get those working on Alpine. 2021-11-24 07:57:22 https://wiki.alpinelinux.org/wiki/QEMU 2021-11-24 07:57:29 The wiki page has some decent info though. 2021-11-24 08:09:33 boomanaiden154: It's ok, it seems the things I need in order to get non-software rendering are: xf86-video-fbdev, mesa-dri-gallium and mesa-egl (using QEMU with virgl, so using -device virtio-ramfb-gl) 2021-11-24 08:13:41 ktprograms: I' 2021-11-24 08:14:39 boomanaiden154: What does I' mean? 2021-11-24 08:14:52 wouldn't it be possible to use xf86-video-fbdev standalone? 2021-11-24 08:16:31 ktprograms: Hit enter too soon. That will probably get you non-software rendering, but it might also fall back to swrst. 2021-11-24 08:17:53 boomanaiden154: What's swrst? 2021-11-24 08:18:33 It's the software renderer. 2021-11-24 08:18:57 boomanaiden154: Well it seems to be fast enough so I'm happy with it. 2021-11-24 08:19:49 It's not terribly slow depending upon what you're doing. And it might not be using swrst. I use swrst for some tests in some Aports CI pipelines and it isn't too slow. 2021-11-24 08:19:58 But it's definitely not hardware acceleration like in terms of performance. 2021-11-24 10:25:14 Does anyone have any experience with running a PKI on Alpine? im looking into: https://smallstep.com 2021-11-24 10:38:05 step is packaged in alpine and a few people use it afaik 2021-11-24 10:50:11 Hm. Im trying to use alpine on an sbc. I downloaded the rootfs, ran setup-alpine, anf installed dhcp and ssh. I then enabled the sshd. But it seems that the default runlevel is not ran at all 2021-11-24 10:50:18 nice :) 2021-11-24 10:50:33 What did I do wrong? 2021-11-24 10:51:37 Hm actually Im using an ubuntu kernel for now but should that run /sbin/init anyway? 2021-11-24 10:51:46 Shouldn't 2021-11-24 10:52:14 i think not 2021-11-24 10:52:25 but honestly, im fairly new to Alpine :) 2021-11-24 10:52:37 but what happens if you install the alpine kernel and boot from it?> 2021-11-24 10:53:32 Ill try that. Might have to compile it but ill just adjust the init system and see if it works 2021-11-24 10:54:04 if you can boot into alpine, then you can install it with apk add? 2021-11-24 10:54:25 https://wiki.alpinelinux.org/wiki/Kernels 2021-11-24 11:03:03 kitzman: you mean you downloaded and used the "mini rootfs"? That's meant for containers so it doesn't have any init 2021-11-24 11:04:12 jhjacobs81: i would like to use the vendor's kernel, then Ill try to use the one from the repos. 2021-11-24 11:05:01 minimal: hm i did install openrc tho. i'll check later if it also contains the init bin 2021-11-24 11:06:12 kitzman: installing openrc will not enable the various init.d services though. The "mini root" is not intended for use on a physical machine. It also will not have any firmware etc installed which a physical machine may need. 2021-11-24 11:06:48 kitzman: as its description says: "For use in containers and minimal chroots" 2021-11-24 11:49:18 im really starting to like Alpine :) 2021-11-24 11:52:12 it's ok 2021-11-24 11:52:37 whats ok? 2021-11-24 11:54:10 jhjacobs81: I'm the maintainer for step-cli and step-certificates packages. I've been intended to write a setup-ca script to cover various forms of CA for while but unfortunately other things keep stealing my time 2021-11-24 12:15:01 :-) 2021-11-24 12:15:06 well, maybe i can help :) 2021-11-24 12:38:54 jhjacobs81: certainly welcome to any suggestions. I thought of initially of providing config for 3 things: (1) offline root CA, (2) intermediate CA using offline root CA, and (3) root and intermediate CAs 2021-11-24 13:21:21 sounds doable :) 2021-11-24 13:22:30 *makes note for RALPF* 2021-11-24 13:22:57 RALPF (Silent P) is my project for a firewall... >:) Rust ALPine Firewall 2021-11-24 13:23:19 would be nice if it integrates a small pki for the smb :) 2021-11-24 13:29:26 jhjacobs81: MR opened for latest step-cli version and will open MR for step-certificates shortly 2021-11-24 13:33:27 MR? 2021-11-24 13:33:31 waddat? 2021-11-24 13:34:00 Merge Request. 2021-11-24 13:35:01 ahh :) 2021-11-24 14:59:22 hi all 2021-11-24 14:59:48 after upgrading to 3.15, libvirtd won't start properly anymore ... no virbr0 ever appears. can somebody help me debug this? 2021-11-24 15:00:04 rapha: did you reboot? 2021-11-24 15:00:36 ikke: no, not yet. scared of doing so tbh. 2021-11-24 15:00:44 also, https://i.imgur.com/Hj5yP5r.png ... why are some of them not executable? 2021-11-24 15:01:06 rapha: then the kernel modules can probably not be loaded 2021-11-24 15:01:13 hmm ok makes sense 2021-11-24 15:01:27 lemme see if i can get to the out-of-band controller from here 2021-11-24 15:02:05 ikke: and those files in /etc/init.d not being +x, is that normal? 2021-11-24 15:02:59 No, it's not 2021-11-24 15:03:42 hmm, let's +x them and try a reboot 2021-11-24 15:03:59 i hope grub is alright, but looks good 2021-11-24 15:11:02 sometimes i wish Dell's "lifecycle controller" could be disabled altogether 2021-11-24 15:11:05 ACTION yawns a bit 2021-11-24 15:12:38 hmm, postfix and avahi failed to start, otherwise it came back up. i can live with that. 2021-11-24 15:13:51 YIKES! no VM is working! 2021-11-24 15:13:58 Hello, I have a question or two on licensing: In this issue: https://gitlab.alpinelinux.org/alpine/aports/-/issues/7423 It is said "As per distribution policy, we only ship software licensed under... (full message at https://matrix.org/_matrix/media/r0/download/matrix.org/RPFYaVzSTcogHdxvXVYqvJFm) 2021-11-24 15:14:07 ikke: does this tell you anything? https://gist.github.com/sixtyfive/e826c85a45ca5ca0adac0f7895cd2d4f 2021-11-24 15:14:20 something to do with the upgrade from python3.7 to 3.9 perhaps? 2021-11-24 15:15:06 similar message when trying to edit a vm. 2021-11-24 15:15:50 rapha: #13225 2021-11-24 15:16:50 phew, booting again! thanks ikke, you just prevented me from getting a heart attack :P 2021-11-24 15:18:16 hmm, 3d accelleration not supported? must be another missing module then... 2021-11-24 15:19:17 ah, qemu-hw-display-virtio-gpu-pci, there we go 2021-11-24 15:21:42 is there any yaml<->json tool ? 2021-11-24 15:22:39 yaml can be styled as valid json 2021-11-24 15:24:17 yaml.safe_dump(json.load(sys.stdin), sys.stdout, default_flow_style=False, allow_unicode=True) 2021-11-24 15:24:58 I use this Python one-liner for "pretty-printing" ^ vkrishn 2021-11-24 15:26:50 I was trying to avoid python/per, but is ok if can be done easily 2021-11-24 15:27:01 a ruby one-liner would work, too 2021-11-24 15:28:12 import yaml;import json;import sys 2021-11-24 15:28:29 So not a "true" one-liner but close enough 2021-11-24 15:29:06 Yeah, this is the trivial use-case for the libraries, so anything that has decent JSON and YAML library should be equivalently easy. 2021-11-24 15:31:02 ruby -ryaml -rjson -e "puts YAML.load_file('my_conf.yaml').to_json" 2021-11-24 15:31:06 vkrishn: yq is one such tool 2021-11-24 15:31:16 that would be the yaml => json direction, vkrishn. true one-liner :) 2021-11-24 15:31:46 yes, thanks, trying yq 2021-11-24 15:35:45 yc looks interesting but is not currently packaged for alpine 2021-11-24 15:36:19 yaml, yajl libs/bins are there, found https://github.com/sjmulder/json-yaml 2021-11-24 15:37:14 yc, url pls 2021-11-24 15:40:37 https://github.com/kellyjonbrazil/jc 2021-11-24 15:49:34 ikke: all working again now 100% and 3.15 seems like a very nice release - congrats and thanks very much! 2021-11-24 15:56:31 rapha: glad to hear 2021-11-24 16:15:25 Hmz 2021-11-24 16:16:00 i only have two hands 2021-11-24 16:18:29 the human condition 2021-11-24 16:27:26 congratz on the release ♥ 2021-11-24 17:17:04 Are apk exit codes documented anywhere? Don't see anything in apk-add(8) or apk(8) 2021-11-24 17:18:50 probably not 2021-11-24 17:19:14 Hello71: I see you activated your magic 8-ball :P 2021-11-24 17:19:31 definitely so 2021-11-24 17:32:13 rapha: ruby -ryaml -rjson -e "puts JSON.load_file('releases.json').to_yaml" would work ? not quite used ruby 2021-11-24 17:41:02 found it https://gist.github.com/lwoodson/a9421897b9ab85e395117b90269848da 2021-11-24 17:42:44 vkrishn: can 2021-11-24 17:42:52 vkrishn: can't remember exactly, but looks good 2021-11-24 17:43:08 hmm, according to https://rubyapi.org/3.0/o/json#method-i-load_file , yes, it's exactly correct 2021-11-24 17:43:40 ruby -ryaml -rjson -e "File.write('releases.yaml', JSON.load_file('releases.json').to_yaml)" 2021-11-24 17:43:48 that would write it into releases.yaml right away 2021-11-24 17:45:08 if you put `ruby -ryaml -rjson -e "File.write(ARGV[1], JSON.load_file(ARGV[0]).to_yaml)" $1 $2` into json2yaml.sh and made it executable, i think you could call that with ./yaml2json.sh releases.json releases.yaml 2021-11-24 17:46:04 ofc at _that_ point just add a `#!/usr/bin/env ruby` shebang at the top of a json2yaml.rb and make it three lines: require 'yaml' ... require 'json' ... File.write(ARGV[1], JSON.load_file(ARGV[0]).to_yaml) 2021-11-24 17:46:27 or call it json2yaml and stuff it into ~/.local/bin and put that into your PATH :-P 2021-11-24 17:49:10 thanks! 2021-11-24 18:04:04 Hi, I have an issue with X on alpine in a KVM. When running `xinit` or `startx` as root X with openbox will start. If running as user I get xf86EnableIOPorts: failed to set IOPL for I/O permission denied. My user is in tty, input, wheel and video group. I tried to change some default udev rules to get read and write permission for tty when in tty group. I even tried to set the sticky bit on /usr/bin/xinit. Can I somehow fix this or drop the 2021-11-24 18:04:04 X-servers rights to my user? 2021-11-24 18:35:42 Forgot to mention: If running `doas xinit .xinitrc` X works as expected but as root user. 2021-11-24 18:37:11 this is shortest from various sources, https://tpaste.us/xnOw 2021-11-24 19:05:18 hi all, how would one upgrade between postgres 13 and postgres14 2021-11-24 19:05:25 it seems I need the postgres13 binarys 2021-11-24 19:05:43 we have good news for you 2021-11-24 19:05:51 I installed the postgresql13 binaries but I cant find them. Keep them in mind, postgresql14 is also installed 2021-11-24 19:06:16 apk info -L 2021-11-24 19:06:17 https://wiki.alpinelinux.org/wiki/Release_Notes_for_Alpine_3.15.0#PostgreSQL_multiple_major_versions 2021-11-24 19:06:51 ikke: sorry, I guess I shouldve read the release notes more carefully 2021-11-24 19:07:01 I got really excited. This has been a long awaited release 2021-11-24 19:17:44 Worked like a charm, I love the new packaging. Thanks for all the hard work! 2021-11-24 20:21:04 question w.r.t. 3.15: gzip kernel ko files, any reason not to go with zstd? 2021-11-24 20:21:58 kernel needs to have it implemented 2021-11-24 20:22:03 im not aware of the kernel supporting zstd 2021-11-24 20:22:38 if you check the make menuconfig UI in the kernel tree, you can see the algorithms that are potentially available 2021-11-24 20:23:42 oh, zstd got indeed added! and its enabled in alpine 2021-11-24 20:23:44 zstd been in-tree for some time, including the versions that alpine uses 2021-11-24 20:24:23 ahh but busybox modprobe does not support it 2021-11-24 20:28:57 https://gitlab.alpinelinux.org/alpine/aports/-/issues/12857 2021-11-24 20:29:05 this has been discussed previously 2021-11-24 20:34:44 Naib: modprobe was switched to kmod's one, zstd was implemented but not found worthwhile, compression was changed to gz but kmod's busybox still remains for now, kmod likely to be removed to switch back to Busybox's modprobe shortly which does support gz (but not zstd) 2021-11-24 20:38:16 yup thanks, i found that Issue, it was more some comment on phoronix 2021-11-24 20:38:56 there is a patch for bb to support zstd http://lists.busybox.net/pipermail/busybox/2021-September/089179.html 2021-11-24 20:45:39 oook, time to update Alpine on pi4, backup done... 2021-11-24 20:45:48 im scared :) 2021-11-24 20:46:45 me, too, kind of afraid new kernel... 2021-11-24 20:46:57 yup 2021-11-24 20:47:11 especaily as the LTS that is part of 3.14 won't boot on my headless ... 2021-11-24 20:47:26 im running community-edge for the kernel due to this 2021-11-24 20:47:35 so concerned flipping back to LTS 2021-11-24 20:48:04 Naib: huh? do you have raspberry too? 2021-11-24 20:48:13 no 2021-11-24 20:48:36 oh 2021-11-24 20:48:44 i have a via with a via_velocity nic 2021-11-24 20:48:52 the only thing that happened for me after the upgrade was on a laptop where /dev/dri/card0 became card1, for whatever reason. 2021-11-24 20:49:04 all other alpine installs went fine 2021-11-24 20:49:39 something up with either alpine building or that series that meant a kernel segfault once it is modprobed 2021-11-24 20:50:18 Naib: ahh, ye nics and wifi things always can destroy the mood when not working :/ 2021-11-24 20:50:24 but... brb, hope so! 2021-11-24 20:50:36 yup :) 2021-11-24 20:51:38 i had run gentoo on this headless for years, but a lot of braindead have crept in and the dev's did a major depreciation ... so much so that a 2month old box would not upgrade... 2021-11-24 20:52:26 I have had alpine in containers for quite some time, but this was it so I rebuilt with alpine \o/ (found some holes in my backup :( ) and then had issues when booting from the ssd while the usb was fine... 2021-11-24 20:52:56 I think i re-installed about 10times, each time panicking until I saw the error and gambled with edge 2021-11-24 21:13:27 oook, upgrade to 3.15 went smooth on pi4 :) 2021-11-24 21:19:12 i suppose moving a Debian based Nextcloud environment with 300 users to Alpine isnt a good idea for the friday night hmm 2021-11-24 21:21:36 it's wednesday 2021-11-24 21:22:23 At night, everyday is friday. 2021-11-24 21:23:45 its wednesday NOW.. but when you blink your eye.. tadaaaa! friday! and just when you want to start drinking your first beer.. 2021-11-24 21:23:56 :D 2021-11-24 21:24:48 deploy to prod on a friday afternoon is a well established best practice 2021-11-24 21:25:19 don't mess with tradition 2021-11-24 21:25:22 :D 2021-11-24 21:25:57 well, in my theory its very simple.. since its a php application anyway 2021-11-24 21:26:30 but seriously if it were me, if it's not broken, i'd leave it alone, and do the alpine switcheroo later when something else needs to be done 2021-11-24 21:26:54 but some things never break :( 2021-11-24 21:27:04 that's not a bad thing 2021-11-24 21:27:17 its an installation i inherited of someone else :) 2021-11-24 21:27:49 surely you can find something better to spend your time on, than fixing things that aren't broken. i'm not making a pro-debian argument, just saying. 2021-11-24 21:28:40 hmm no.. not friday night i think :) 2021-11-24 22:01:12 what's wrong with sudo? i don't see a reason for the change to doas in https://gitlab.alpinelinux.org/alpine/tsc/-/issues/1 2021-11-24 22:03:11 LoC size, hell of a configuration 2021-11-24 22:04:21 did it suddenly get worse? 2021-11-24 22:04:55 not really, but sometimes better alternatives appear 2021-11-24 22:05:21 hmm 2021-11-24 22:06:33 imho, I support that decision, I've been using doas for a long time and also run into issue where abuild-keygen wanted to use sudo but I didn't have it :D 2021-11-24 22:07:07 the config file(s) are very neat 2021-11-24 22:08:28 that's fair. i was just wondering if it was a case of not-broken-don't-fix-it 2021-11-24 22:09:02 sometimes it's worth breaking things to make life easier long-term :) 2021-11-24 22:13:27 some discussion about it here: https://news.ycombinator.com/item?id=29330394 2021-11-24 22:14:01 i shouldn't have asked so late. will read tomorrow 2021-11-24 22:14:23 Somtimes you have to break things yourself in the name of progress :) 2021-11-24 22:16:48 I recommend not reading that, but that's my opinion :P there are comments from people who don't know what they are talking about and some from people who do 2021-11-24 22:18:30 Thats why im reading it :) makes for fun 2021-11-24 22:23:31 I don’t know doas. Would one of you care to explain to me how it’s different from sudo or point me to relevant documentation? 2021-11-24 22:24:11 https://github.com/Duncaen/OpenDoas 2021-11-24 22:24:26 tl;dr it's a Linux port of OpenBSDs doas 2021-11-24 22:25:08 it does absolute minimum required thing, so you won't have things like sudo -e or visudo 2021-11-24 22:25:31 Less code usually means less attack surface :) 2021-11-24 22:25:55 and less code to review ;) 2021-11-24 22:26:56 That too :) makes it even easier to spot any bugs 2021-11-24 22:27:20 from the gitlab.a.o TSC issue: > Keep in mind that we are comparing it with sudo, about 14,500 LoC if I don’t count bundled libs and plugins and horrifying 147,000 LoC if I count everything! 2021-11-24 22:27:37 while doas is ~2.500 LoC 2021-11-24 22:29:42 :-) 2021-11-24 22:33:38 Thanks, will read. :) 2021-11-24 23:40:58 Did alpine 3.15 lose unrar package? Is there any replacement? 2021-11-25 00:02:13 graywolf: https://gitlab.alpinelinux.org/alpine/aports/-/merge_requests/26234 2021-11-25 00:02:16 graywolf: https://gitlab.alpinelinux.org/alpine/tsc/-/issues/23 2021-11-25 00:04:48 Ah, I should have check the aports repo. Sorry and thanks. 2021-11-25 00:06:24 graywolf: no reason to sorry, it isnt easy to track all changes 2021-11-25 00:07:12 to say* 2021-11-25 01:35:08 hello, is it to be expected that when we boot the alpine-standard-3.15.0-x86_64.iso without working ethernet/wifi that we cannot complete setup-alpine due to missing dependencies, like dosfstools, grub-efi, sfdisk? 2021-11-25 01:35:26 i.e. alpine can only be setup with internet access? 2021-11-25 01:36:31 wondering why i didn't do a netboot instead :) 2021-11-25 01:37:17 busybox itself has fdisk and mkfs.vfat to an extent that lets you do at least something 2021-11-25 01:37:44 but yes, you need the extended image if you have no internet i think :) 2021-11-25 01:38:03 ah, i see. i thought it might be a bug at first. 2021-11-25 01:38:16 i.e. nobody noticed or something ;) 2021-11-25 01:38:59 i'm gonna dd some of my generic preinstalled images to the disk now as i just took it out of the pc anyway 2021-11-25 01:39:11 no, the standard image if you mount it and look at apks/ is just apk, and a bunch of networking, pretty much 2021-11-25 01:40:14 right, i was gonna ask some more details around that, but haven't prepared my questions yet. something about firmware, that folder got very big suddenly 2021-11-25 01:40:46 is the 'data' style installation *very* uncommon btw? 2021-11-25 01:41:12 cause that might explain some of my seen weirdnesses, if yes, then i'll try sys first to compare :) 2021-11-25 01:41:24 plenty of people use it, but it's not very well documented 2021-11-25 01:41:35 and sys is by far the easiest way to run a system, so 2021-11-25 01:42:02 i'm trying to run some stuff without state, so to say 2021-11-25 01:42:40 i'd update in my house inside some VM to test, then send the resulting lbu over to the user 2021-11-25 01:43:03 and possibly if they fuck it up send a new flash drive via post 2021-11-25 01:43:41 i like the idea of running from RAM. it's unused anyway :) 2021-11-25 01:44:31 doing a cat on everything in the system to cache it in memory works basically the same if the cache never gets cleared due to low memory 2021-11-25 01:46:24 extended worked. 2021-11-25 01:48:49 i suppose i can just rsync everything into the /boot and / filesystems 2021-11-25 01:49:02 even if i send them a stick... 2021-11-25 01:49:25 i can also backup the partition table and recreate the filesystems remotely, as long as they boot the stick 2021-11-25 01:49:44 ah, i guess i can install sys to stick, too, i never thought of that 2021-11-25 01:50:15 of course there's too many ways to do this, earlier i thought i don't have enough creativity to find a sys solution, now i have too many. 2021-11-25 01:50:19 gonna sleep instead 2021-11-25 01:50:27 thanks and good night 2021-11-25 01:50:38 sleep well :) 2021-11-25 02:11:22 is it possible to specify kernel params for the installer iso before it boots? 2021-11-25 02:25:57 hiro_: the downloads page does say "STANDARD: Network connection is required." 2021-11-25 02:26:13 bandali: yes, just type it in 2021-11-25 02:32:06 Hello71, where exactly? i've started looking into syslinux/isolinux but so far as far as i can tell i can only type 'lts' as a valid string in that prompt 2021-11-25 02:32:36 say i want to add console=hvc0 to the boot params for the dfeault lts config for the live iso 2021-11-25 02:32:44 i think it is "lts console=hvc0" 2021-11-25 02:33:01 iirc linux-virt comes with console=hvc0 by default 2021-11-25 02:33:58 ah great, i think you're right, thanks! i'd tried 'console=hvc0 lts' a la classic environment variables but that didn't work 2021-11-25 02:34:13 also thanks for the pointer about linux-virt, i'll check it out :) 2021-11-25 02:34:46 also it's generally faster to use host kernel loading but maybe that's too complicated for your case 2021-11-25 02:37:24 right, yea not sure how feasible it would be for my use-case here 2021-11-25 04:32:38 how does mandoc-apropos-1.14.5-r5.trigger manage to be so hecking slow‽ it often seems to take multiple times longer than the entire rest of an apk add 2021-11-25 04:33:48 installs instantaneous here 2021-11-25 04:33:52 what mirror? 2021-11-25 04:34:05 the trigger, not the download time 2021-11-25 04:34:10 oh, right 2021-11-25 04:34:19 it does the whole setup for apropos to work 2021-11-25 04:34:23 which parses like every manpage 2021-11-25 04:34:45 I guess it depends on how many doc packages you have installed 2021-11-25 04:34:45 if you have `docs` you have hundreds of them, and it's not exactly built to be fast 2021-11-25 04:34:57 if you have 5 manpages it probably takes under a second 2021-11-25 04:35:49 yes 2021-11-25 04:55:06 Weird. 2021-11-25 04:57:04 it's okay, there are other weird people here 2021-11-25 04:58:03 Trying to boot my machine leads to a black screen. I unplugged the whole thing and have it just down to KBM and monitor, and it's still a black screen. It momentarily flashes "EFI Stub: Loaded initrd from command line option" and then does absolutely nothing after that. 2021-11-25 04:59:32 I don't think that I had upgraded before this reboot that has caused it, but I have no clue. 2021-11-25 05:00:50 i'm assuming you have no serial console or anything to know what's going on 2021-11-25 05:01:05 in that case i would take the drive/media/whatever and mount it on another running system 2021-11-25 05:01:19 I don't think this board has serial output on it. Maybe I'll boot some alpine live media. 2021-11-25 05:01:44 or boot into live image to poke around 2021-11-25 05:01:45 What's the process for chrooting from live media to a drive to then run another update? 2021-11-25 05:02:02 it's in the docs 2021-11-25 05:07:10 you mount partitions like you would in fstab, then proc, sys, dev, then make /etc/resolv.conf or copy from host, then chroot 2021-11-25 05:07:57 https://wiki.alpinelinux.org/wiki/Alpine_Linux_in_a_chroot like here 2021-11-25 05:08:06 interesting 2021-11-25 05:08:09 just the chroot section 2021-11-25 05:08:28 also, weird behavior on my system's part: i have rgb in the keyboard and mouse, and after that efistub message plays out, the lights go out. 2021-11-25 05:10:32 can't tell if this old ass mac keyboard is trolling me right now 2021-11-25 05:10:50 really don't wanna swap the keyboard i'm using to type here into the machine and back and forth ad nauseam 2021-11-25 05:10:55 um** 2021-11-25 05:16:28 wtf this system is hexed 2021-11-25 05:16:59 it'll read my keyboard for booting and now that i'm booted into alpine (live) it doesn't care about the inputs 2021-11-25 05:18:35 wizardry 2021-11-25 05:19:22 so much for "universal" in usb am i right ladies 2021-11-25 05:20:08 seems like that's the same thing that causes the mouse to turn off 2021-11-25 05:20:13 the kernel boots and the usb stuff breaks 2021-11-25 05:20:28 i guess replugging does nothing 2021-11-25 05:22:08 nah, a normal keyboard is fine here 2021-11-25 05:22:17 it was the old mac bastard being old and bastard 2021-11-25 05:23:54 nice 2021-11-25 05:24:34 wait a second; if i just wanna boot into my alpine install 2021-11-25 05:24:47 oh nvm, /dev and /proc and /sys aren't going to be updated with anything from apk, right 2021-11-25 05:24:48 so i'm stupid 2021-11-25 05:26:18 Does all of the software that's on my system become available the moment I chroot? Does all of the linux-kernel-necessary stuff become available? I dunno if btrfs is supported by the live media 2021-11-25 05:26:34 you probably have to modprobe btrfs to mount your root 2021-11-25 05:26:39 aside from that, yes 2021-11-25 05:26:49 i am hacking 2021-11-25 05:27:01 psykose, the police are on the way to arrest you for abetting a stone cold hacker 2021-11-25 05:27:04 we did it it's mounted 2021-11-25 05:27:18 big cquestion 2021-11-25 05:27:48 if i mount proc and sys and dev all to /mnt/ba/{proc,sys,dev} and THEN mount my filesystem to /mnt/ba, am I overriding the mounts i previously did? 2021-11-25 05:28:04 no idea, so do it the other way 2021-11-25 05:28:18 i would assume yes 2021-11-25 05:28:30 because you are mounting something new up the root 2021-11-25 05:28:41 everything below it is gone until unmount, in a sense 2021-11-25 05:29:17 i'm also entirely stupid and need to read how i mount this ish in the fstab because just mounting the partition leads to a bunch of balogna 2021-11-25 05:29:34 probably some btrfs fun 2021-11-25 05:30:19 is there an easy way to tell mount to just read the fstab (by pointing to it) to get it to properly mount? 2021-11-25 05:31:07 the options line is just -o 2021-11-25 05:31:20 aside from that the issue would be somewhere else 2021-11-25 05:32:02 ah, btrfs subvolumes are special 2021-11-25 05:32:20 i've got everything listed as UUIDs 2021-11-25 05:32:33 blkid should tell you what is what for those 2021-11-25 05:32:52 do i just type that to mount where i'd normally put the /dev file? 2021-11-25 05:33:03 or does it need to be UUID="" 2021-11-25 05:33:17 /dev is just dev 2021-11-25 05:33:19 not sure what you mean 2021-11-25 05:34:07 for btrfs i think you need -o subvol=root or something for the root one, then afterward you mount the others normally and /dev and etc like in the link before 2021-11-25 05:34:37 so if my line in fstab is something like "UUID=vomit / btrfs noatime,subvol=volume/root 0 0", is the mount-equiv command going to be `mount vomit -o noatime,subvol=volume/root /` ? 2021-11-25 05:34:49 or is "vomit" in command 2 going to be UUID=vomit 2021-11-25 05:35:17 UUID= 2021-11-25 05:35:34 and a -t btrfs if you want to make it complete 2021-11-25 05:36:31 big schweet 2021-11-25 05:36:41 we are about to type for eight hours everyone 2021-11-25 05:43:20 here goes 2021-11-25 05:43:30 oh neat wtf lol i'm not plugged in 2021-11-25 05:43:36 but apk is still upgrading soething of course big brain 2021-11-25 05:49:26 what's the best way to find my mainboard's name from ash 2021-11-25 05:49:51 3 fooking ethernet ports on this board and i never remember which one is which 2021-11-25 05:50:25 hyperplexing: inxi -M 2021-11-25 05:50:54 seemed unfamiliar, because it is; that's not installed by default on alpine, aye? 2021-11-25 05:50:56 nor in livedisk? 2021-11-25 05:51:49 nope, need repos 2021-11-25 05:52:26 THERE it is 2021-11-25 05:52:29 dmidecode -t 2 2021-11-25 05:55:35 figures 2021-11-25 05:55:40 i'm in the damned ipmi port as usual 2021-11-25 05:57:03 finally some PINGS 2021-11-25 06:02:36 we are now blindly rebooting 2021-11-25 06:02:40 everyone join hands 2021-11-25 06:02:45 GIVE ME YOUR ENERGY 2021-11-25 06:03:04 yeah that didn't work at all, lmao 2021-11-25 06:04:10 the curious case of wtf mane 2021-11-25 06:05:27 it made the initramfs, so i wonder 2021-11-25 06:05:32 big wondering 2021-11-25 06:06:03 guess i could try to interrupt grub and boot from there, but grub might not have btrfs already enabled, aye? 2021-11-25 06:06:59 the true joys of not using grub on your system are that when these things change you can't tell where the bios ends and the machine begins amirite ladies 2021-11-25 06:34:16 funny story, i'm doing this so i can ssh in and irc 2021-11-25 06:34:20 from some trash windows machine 2021-11-25 06:34:24 because i don't wanna config hexchat 2021-11-25 06:36:27 hmm 2021-11-25 06:41:07 what the hell 2021-11-25 06:41:33 what's the "i fucked up get me off the ride" button for canceling startx 2021-11-25 06:41:47 control c works sometimes 2021-11-25 06:41:55 no dice 2021-11-25 06:41:58 damnit 2021-11-25 06:41:59 otherwise, alt f2, login, pkill x 2021-11-25 06:42:08 noper doper 2021-11-25 06:42:13 er 2021-11-25 06:42:16 was it ctrl alt f2 2021-11-25 06:42:22 i wonder if this has something to do with the fact that i'm not on displayport 2021-11-25 06:42:29 no 2021-11-25 06:42:42 i've got two monitors, and usually the main one is the displayport monitor 2021-11-25 06:42:44 ports are agnostic to server software 2021-11-25 06:43:03 so if i had some config where the primary monitor was normally over dp it wouldn't bug out? 2021-11-25 06:43:04 they just take the signal 2021-11-25 06:43:33 if you have matching ports on both ends of the wire, you're good 2021-11-25 06:44:35 ports are more like microcontrollers 2021-11-25 06:44:40 they do one thing 2021-11-25 06:46:00 aye, though i thought i remembered setting something up with xrandr to make each monitor matter in the x script, no clue if that fux anything 2021-11-25 06:46:08 regardless, i don't need to actually start x to diagnose this 2021-11-25 06:46:19 was kinda perplexed that efibootmgr wasn't installed 2021-11-25 06:46:26 so that might be part of it 2021-11-25 06:46:30 i don't even try writing xrandr stuff 2021-11-25 06:46:33 i'll have to return in a minute; buddies wanna game 2021-11-25 06:46:37 touch me lad 2021-11-25 06:46:44 freq: touch me gran 2021-11-25 06:46:50 i want cookies 2021-11-25 06:48:58 who wants to teach me how to use the ipmi thing 2021-11-25 06:49:24 does it let me type and stuff 2021-11-25 06:49:56 dunno man. i'm running a rpi400 2021-11-25 06:57:17 isn't caskd brew 2021-11-25 06:57:30 :o 2021-11-25 06:57:40 homebrew 2021-11-25 06:57:56 i chose this nickname before i knew about it 2021-11-25 06:58:00 oic 2021-11-25 08:11:05 hello, is the `srm` command anyway available in latest version (3.14) ? 2021-11-25 08:11:38 apk search -e cmd:srm 2021-11-25 08:12:26 will search in repos that you've configured in /etc/apk/repositories 2021-11-25 08:14:33 yep, already did that, also crawled on search engine on alpine.org; and also lurked in github ( https://github.com/alpinelinux/aports/commit/a7850c7f1df3627cdc06e2d8bd85cf26aaedc5ba ) 2021-11-25 08:15:03 I m asking to other other people for confirmation 2021-11-25 08:15:11 Does any alternative exists ? 2021-11-25 08:15:27 shred is present but it does not work with folder 2021-11-25 08:16:05 so it's a secure file remover 2021-11-25 10:19:06 hmm seems bb awk cant search specific column? awk '$3 == 001' /B/s/mem/vref 2021-11-25 10:22:58 works for me 2021-11-25 10:23:05 seems to kinda work, but sometimes i seem to get random extra line 2021-11-25 10:25:12 eg https://x0.at/JpR-.txt 2021-11-25 10:26:04 guess it's broke 2021-11-25 10:26:51 brokebox :/ 2021-11-25 10:41:31 Any volunteer to help update the description of alpine? https://distrowatch.com/table.php?distribution=alpine the current text was true 10+ years ago 2021-11-25 10:45:38 still looks corect to me? 2021-11-25 10:49:08 arches missing 2021-11-25 10:49:12 i don't think pax is still a thing 2021-11-25 10:49:18 latest is 3.15.0 2021-11-25 10:49:22 (3.14 is missing) 2021-11-25 10:49:32 bug tracker is now gitlab.alpinelinux.org 2021-11-25 10:50:24 documentation should probably point to docs.a.o 2021-11-25 10:57:14 jhjacobs81: we no longer use pax (because its not public available anymore) 2021-11-25 10:58:24 we are also more a general purpose distro than router/vpn/voip specific nowdays 2021-11-25 10:58:41 could include containers as well, which probably is the biggest use case 2021-11-25 10:58:47 gnome. kde desktops available 2021-11-25 10:58:59 how do you even update distrowatch o-o 2021-11-25 10:59:34 oh, email them 2021-11-25 11:07:46 yeah, send them an email. I think I did it last time, when we switched from uclibc to musl libc 2021-11-25 11:08:10 yeah, i'm just setting up a new domain, then i'll have a look 2021-11-25 11:08:29 thank you! 2021-11-25 11:12:27 time to learn javascript again, bahah 2021-11-25 11:28:29 i dont even know what pax is :) 2021-11-25 11:39:23 Can I send to alpine-users PSA on how to resolve bug cause by update to 3.15 (everything was rendered super huge and it took me ~hour to find the solution on the internets) or is that offtopic/spam unless asked by someone first? Not sure what CoC/policy is on that list. 2021-11-25 11:42:56 that will be fixed in xorg-server 21.1.2 and probably backported 2021-11-25 11:43:34 if you mean having to set the dpi value 2021-11-25 11:44:53 Yeah, my fix was xrandr --dpi 96 in xinitrc; Took bit long to find out, turns out "xorg everything is huge" is not the most accurate query :/ 2021-11-25 11:45:02 lmao 2021-11-25 11:45:17 oh no xrandr 2021-11-25 11:46:29 graywolf: you could also set it in .Xresources 2021-11-25 11:46:33 yes, the 21.1 release fixed the ancient thing where every monitor was '96' dpi by default, and got the correct value instead 2021-11-25 11:46:49 but since everything assumes it is always 96, that breaks most things, so it got reverted for next release 2021-11-25 11:46:51 Xft.dpi 2021-11-25 11:46:56 Yeah, and now my i3-statusbar was across half of my tv 2021-11-25 11:47:05 yep 2021-11-25 11:47:12 peer: what's fron with xrandr? 2021-11-25 11:47:15 wrong* 2021-11-25 11:47:42 nothing - i work for some tech confrences and.. have some _very_ interesting problems :P 2021-11-25 11:48:10 the phrase "does anyone know xrandr?" is frequent 2021-11-25 11:48:17 graywolf: We did think about adding something to the release notes about it, but we got the idea it was reverted already 2021-11-25 11:48:35 !27599 2021-11-25 11:55:08 ikke: ah, yeah, release notes was the first place I've checked 2021-11-25 11:55:31 Oh, btw, I like the new postgresql approach in 3.15. Makes the upgrade much easier. 2021-11-25 11:58:05 ikke: the revert is not in 21.1.1, the revert is removing the probed* and putting dcc* variables back, the sample mr is actually the opposite of a revert 2021-11-25 11:59:33 i think bart put the original patch ontop of the non-reverted 21.1.1 tag, saw the reject due to it already existing, and thought that meant it was already reverted? the patch in the mr is the original change from a year ago 2021-11-25 12:00:23 for reference https://gitlab.freedesktop.org/xorg/xserver/-/commit/35af1299e73483eaf93d913a960e1d1738bc7de6 2021-11-25 12:01:02 and the original https://gitlab.freedesktop.org/xorg/xserver/-/commit/05b3c681ea2f478c0cb941c2f8279919cf78de6d 2021-11-25 12:04:52 I put the reverted patch on top of 21.1.1 2021-11-25 12:04:59 took it from the Arch package 2021-11-25 12:05:27 but that failed indeed 2021-11-25 12:05:45 that patch is already in 21.1.1 is what i mean, it's the original broken change from the link above 2021-11-25 12:05:52 i am getting confused now with all these reverts, haha 2021-11-25 12:06:09 I know. I wanted to revert the dpi change, but that didn't apply on 21.1.1 2021-11-25 12:06:15 ahh 2021-11-25 12:06:56 do you mean revert to old behaviour, or revert the revert of the dpi change 2021-11-25 12:07:11 revert the dpi change like I said, so revert to old behaviour 2021-11-25 12:07:28 then the patch in the mr is the opposite of a revert, it readds the new bahaviour after a revert to the old one 2021-11-25 12:20:09 the arch patch is ran with -R 2021-11-25 13:33:25 What is the equivalent of systemd-analyze for OpenRC? 2021-11-25 13:33:48 there is no equivalent thing 2021-11-25 13:33:50 what does that do? 2021-11-25 13:35:13 lopid: gives boot time statistics 2021-11-25 13:35:56 So is there any OpenRC setting to make it log service startup time? 2021-11-25 13:36:27 boothviz 2021-11-25 13:37:04 hmm 2021-11-25 13:41:48 ktprograms: what do you mean by “startup time”? 2021-11-25 13:42:44 hoel: Like see which services are slow 2021-11-25 13:43:14 in my experience all the services take 0.1s 2021-11-25 13:43:25 except for networking, or chronyd which blocks for time 2021-11-25 13:44:11 i wish it was possible to 'background' networking for a desktop install 2021-11-25 13:45:17 I think I found that spice-webdavd takes at least half a second 2021-11-25 13:45:44 possible 2021-11-25 14:21:52 Hm, apk fetch does not support the @repo syntax? 2021-11-25 14:35:56 ktprograms: there is bootchart but when I tried enabling it a few months ago it didn't work and I haven't had time to dig into why 2021-11-25 15:07:51 'cut' cant output latter field before former field? 'print 1 2 3|cut -d" " -f3,1' is the same as -f1,3 2021-11-25 15:08:05 Hi! I want to install the Monero daemon, and I see that there is a package in testing. When I try to build it myself, it fails because some of the dependencies are not available in stable. Any ideas? Any way to install a package from testing without building it? 2021-11-25 15:12:53 stian_: use edge 2021-11-25 15:13:34 it doesn't install on edge either 2021-11-25 15:13:39 frag: indeed it cannot 2021-11-25 15:14:15 frag: for literally that example you can use awk i guess, { print $3, $1 } 2021-11-25 15:14:19 darn, guess I'll have to use a Docker container. 2021-11-25 15:20:04 psykose: yeah, was just suprised :] 2021-11-25 15:57:07 stian_, psykose: seems the Monero package hasn't been touched in quite a while and it needs a rebuild against the latest boost before it'll install from the testing repo 2021-11-25 15:57:26 that's probably also why it's still in testing, nobody is caring enough about it to maintain it and move to community 2021-11-25 15:58:44 I'll fixup the package for now 2021-11-25 17:18:15 Hello71: oops, you're right haha, i only read the first line somehow... 2021-11-25 17:18:59 if it's easy enough to generate my own iso images, that might also be an alternative to the 'data' mechanism 2021-11-25 17:19:13 probably it's just adding the packages to the ISO :) 2021-11-25 17:19:42 (definitely i have too many ways to do this now ;)) 2021-11-25 17:26:55 time to see if my headless explodes with 3.15 upgrade 2021-11-25 18:26:04 is 'busybox grep -r test' supposed to work? (it seems to just pause forever) 2021-11-25 18:26:33 . 2021-11-25 18:26:49 The path is requried 2021-11-25 18:26:51 required* 2021-11-25 18:27:04 If not passed, it expects data on stdin 2021-11-25 18:27:07 which is why it waits 2021-11-25 18:27:21 ah i see 2021-11-25 19:35:35 can someone help me with doas config? 'permit setenv { DOCKER_HOST=$DOCKER_HOST } :wheel cmd /usr/bin/go' yields "doas: syntax error" and I have no idea why 2021-11-25 19:36:49 wow... tried uninstalling gnu grep, one of my script went from running in 0.2 seconds, to 2 seconds... 2021-11-25 19:40:33 sbase/grep used 1.5 sec 2021-11-25 19:44:16 9base/grep 0.6 sec 2021-11-25 19:44:27 jaja.. 2021-11-25 20:01:27 /usr/bin/abuild-keygen: line 63: sudo: not found 2021-11-25 20:01:48 if you wanna switch to doas, then maybe it's better not to have scripts like this :) 2021-11-25 20:02:56 i prefer how things are separated e.g. in makefiles, where make never needs root and make install always needs root 2021-11-25 20:03:15 SUDU=doas abuild-keygen :P 2021-11-25 20:03:21 but yes, we should switch the default 2021-11-25 20:04:01 i'd prefer changing the instructions, so that you tell people to run abuild-keygen -a; doas abuild-keygen -i 2021-11-25 20:04:08 like make; doas make install 2021-11-25 20:04:24 and then jus tremove all the sudo logic from abuild-keygen 2021-11-25 20:07:14 hmm, abuild-doc has no abuild(1) man page? 2021-11-25 20:07:30 i see only two files inside, 2021-11-25 20:07:31 usr/share/man/man5/APKBUILD.5.gz 2021-11-25 20:07:31 usr/share/man/man1/newapkbuild.1.gz 2021-11-25 20:08:05 APKBUILD(5) refering to abuild(1), which doesn't exist 2021-11-25 20:08:32 it does not exist 2021-11-25 20:08:37 not in the source 2021-11-25 20:11:47 nrpe.cfg wants to have the log file at /var/nrpe.log 2021-11-25 20:12:48 That does not sound good 2021-11-25 20:13:25 i don't know if that's new or not, just noticed the update-conf diff to what i already have (/var/log/nrpe.cfg) 2021-11-25 20:14:18 .cfg? 2021-11-25 20:14:54 .log, sorry. today i woke up in typo city :( 2021-11-25 20:15:05 heh 2021-11-25 20:15:46 The sample config logs to syslog and the logdir is commented out 2021-11-25 20:16:29 #log_file=/var/nrpe.log 2021-11-25 20:17:20 sure 2021-11-25 20:17:44 Need to supply --with-logdir to the configure script 2021-11-25 20:18:20 i don't see it in https://git.alpinelinux.org/aports/tree/main/nrpe?h=3.15-stable so maybe it's like that upstream? 2021-11-25 20:19:19 Yes, it seems to default to /var 2021-11-25 20:20:26 https://tpaste.us/KMmY 2021-11-25 20:20:45 nice 2021-11-25 20:21:59 indeed, i see "logdir=${logdir="$localstatedir"}" in the source 2021-11-25 20:29:18 i saw "ERROR: gvim-8.2.3567-r0: trying to overwrite usr/bin/vimdiff owned by vimdiff-8.2.3437-r0." during upgrade, but it doesn't appear to have had any bad effect 2021-11-25 20:29:34 is somebody here Legale ? 2021-11-25 20:29:45 Not aware of anyone with that name 2021-11-25 20:30:41 the person behind this wiki change: https://wiki.alpinelinux.org/w/index.php?title=Connecting_to_a_wireless_access_point&type=revision&diff=20172&oldid=19851 2021-11-25 20:30:45 i don't think it makes sense :) 2021-11-25 20:30:56 wpa_supplicant knows just fine how to reconnect without this addition 2021-11-25 20:31:16 i don't know which specific hardware requires "disable_scan_offload=1 2021-11-25 20:31:32 but it surely is out of context in there 2021-11-25 20:31:54 without ap_scan=1 it scans anyway. 2021-11-25 20:36:11 "apk -v fix mkinitfs" says i have one error. how do i see what it is? 2021-11-25 20:44:32 2apk fix 2021-11-25 20:44:40 but that's probably due to that conflict 2021-11-25 20:44:52 lopid: maybe you can open an issue about it 2021-11-25 20:45:47 2apk fix? 2021-11-25 20:45:59 without 2 2021-11-25 20:46:29 oh, that gvim error popped up again 2021-11-25 20:47:23 That needs to be fixed 2021-11-25 20:48:38 i need to leave in a few mins, don't know if i'll have time to create an issue about it until tomorrow 2021-11-25 21:17:06 I wonder what keeps some packages in edge (for example, blender) from being included in releases 2021-11-25 21:29:55 minimal: long due and possibly uninteresting reply. Can t manage to boot the iso. Spi flashed a bootloader, also my u-boot can t seem to do it for some reason. The rootfs worked fine in the end, just had to adjust some things 2021-11-25 21:39:45 curiousgay: it's in testing 2021-11-25 21:40:26 for blender the reason why it's in testing is because it stinks 2021-11-25 21:59:56 ikke, https://gitlab.alpinelinux.org/alpine/aports/-/issues/13234 2021-11-26 00:15:40 does the resume= kernel parameter also inform the kernel where to dump memory when hibernating or must you actually swapon the partition to which you want to dump to? 2021-11-26 04:50:33 any recommendations for how to best swapon a swap partition that's lvm on luks but on a different physical volume than the root one? i can automatically unlock the corresponding physical volume using a /etc/conf.d/dmcrypt entry, and lvm later picks up, but it seems by that point it's too late and it doesn't get automatically turned on despite the entry in fstab 2021-11-26 04:51:03 so as of now i have to manually run swapon on each boot. i was wondering what would be the best way to mitigate this 2021-11-26 05:11:19 nvm, figured it out :) 2021-11-26 08:07:31 ladida ladida! 2021-11-26 08:15:33 there's no decent forum for Alpine, is there? 2021-11-26 09:10:24 Hey,After upgrading to 3.15.0 from 3.14.3 my wifi card stopped working I found that 3.15 is missing the iwlwifi-cc-a0-59.ucode firmware file, or an equivalent (among other iwlwifi files that i personally don't need) Installing linux-firmware-other provides a working version, cc-a0-63, and an aditional 185 MiB of files This seems a bit overkill, so i'd like to know if this was an intentional change 2021-11-26 09:13:59 that ucode file was in -other in 3.14 and 3.13 too 2021-11-26 09:14:30 what ucode does the fw loader load currently, from dmesg? 2021-11-26 09:16:12 from what i can see all the iwlwifi* ucodes have always been in -other 2021-11-26 09:16:28 iwlwifi-cc-a0-63, from -firmware-other. Without that package the device simply won't be enabled 2021-11-26 09:16:49 i assume because alpine doesn't ship with the -cc-a0 family of firmware anymore 2021-11-26 09:19:38 that file has always only existed in -other, so i dunno 2021-11-26 09:21:19 the problem is that it's not in alpine _without_ linux-firmware-other anymore 2021-11-26 09:22:50 and im curious if that was an intended change 2021-11-26 09:33:00 i just rechecked and alpine-3.14.3.iso does include the single iwlwifi-cc-a0 firmware file in boot/modloop-lts/modules/firmware/ 2021-11-26 09:34:20 thanks for the quick response btw 2021-11-26 09:35:56 ah, perhaps it does 2021-11-26 09:36:06 but that is the modloop in the iso, which is separate from anything on the system with packages 2021-11-26 09:36:06 hmm 2021-11-26 10:09:23 Is there a way to configure abuild to use distcc? 2021-11-26 10:12:41 if you invoke it with something like `distcc gcc ..` then you can just set CC= 2021-11-26 10:12:50 and CXX, etc 2021-11-26 10:13:57 the distcc docs seem to do exactly that, CC="distcc gcc", and DISTCC_HOSTS 2021-11-26 10:14:07 you can export things in /etc/abuild.conf 2021-11-26 10:45:11 Nice thank you I'll try that 2021-11-26 13:49:42 Hm, so it seems like I cannot install vim in 3.15? https://paste.vpsfree.cz/csH:LP34/raw/ Any advice on what am I doing wrong? 2021-11-26 13:50:59 maybe nothing - https://gitlab.alpinelinux.org/alpine/aports/-/issues/13234 2021-11-26 13:51:19 lopid: this problem is a result of that 2021-11-26 17:11:45 I'm trying to cleanup firmware packages but brcm/brcmfmac43455-sdio.bin fails to load (regardless it is already installed) 2021-11-26 17:12:01 it loads fine if I install all the firmware packages 2021-11-26 17:12:11 is there some hidden depedency? 2021-11-26 17:24:14 uhM 2021-11-26 17:24:34 it's a broken symlink 2021-11-26 17:24:39 brcmfmac43455-sdio.bin -> ../cypress/cyfmac43455-sdio.bin 2021-11-26 17:27:08 [ 14.129325] IPv6: ADDRCONF(NETDEV_CHANGE): wlan0: link becomes ready nice 2021-11-26 18:26:03 ping: permission denied (are you root?) ... Any idea? If I create new user, that has no issue 2021-11-26 18:33:46 got it ;/ no idea if that should happen 2021-11-26 18:34:01 it happens only if user is in wheel group 2021-11-26 18:36:04 you changed user primary group 2021-11-26 18:36:34 during creation I set up -G wheel 2021-11-26 18:36:39 so no 'user' group 2021-11-26 18:36:46 that's the issue probably 2021-11-26 18:37:02 does user need own group? 2021-11-26 18:38:39 normally if I set up user without -G it creates his group as user's name 2021-11-26 18:38:54 as login/primary group 2021-11-26 18:57:33 wheel is not valid primary group 2021-11-26 19:10:57 hell(o) 2021-11-26 19:11:02 so 2021-11-26 19:11:05 a question 2021-11-26 19:11:14 how can I get a list of manually installed packages 2021-11-26 19:11:17 ? 2021-11-26 19:11:45 if that's possible 2021-11-26 19:12:19 `apk list -I` will list all the packages 2021-11-26 19:12:26 I'm not sure about "manually installed ones" 2021-11-26 19:12:45 guessssst: cat /etc/apk/world 2021-11-26 19:13:18 nickersonm: dude, thanks a lot! 2021-11-26 19:13:32 couldn't find any mention of that file in the docs for some reason 2021-11-26 19:14:16 um 2021-11-26 19:14:18 actually 2021-11-26 19:14:34 I just found it in apk-add(8) 2021-11-26 19:14:39 >_< 2021-11-26 19:15:18 anyway, thanks and sorry for stupid question 2021-11-26 19:17:06 np, and not a stupid question. 2021-11-26 19:18:33 Hello71: thanks 2021-11-26 19:20:15 but even I have still primary wheel, while also in user's group then it works 2021-11-26 19:38:16 what treats wheel as 'special' so it cannot be a primary group? 2021-11-26 20:33:30 well nothing in particular, it's just illogical 2021-11-26 20:35:15 specifically, your egid needs to be in the range defined by net.ipv4.ping_group_range 2021-11-26 20:35:24 which is set to 999 59999 on alpine by default 2021-11-26 21:05:03 do you have problem with firefox sometimes seem the window doesnt update before you put it in the background and cycle back to it? i got the same problem with telegram-desktop, but not on qutebrowser... 2021-11-26 21:56:38 Hello71: I can't find what's wrong about blender that makes it so worse than rust, firefox, chromium, gnome, kde and libreoffice it's stuck in testing for years 2021-11-26 21:58:45 all the crappy dependencies that nobody wants to touch. actually there is already MR to move chromium to testing because nobody is taking care of it 2021-11-26 21:59:00 you can ask blender maintainer if they want to move it to community 2021-11-26 22:02:06 So I was running deluge 2.0.3 as a headless server on 3.14 using edge community, and i upgraded to 3.15 because deluge 2.0.3 was moved from edge to the normal release branch, and now deluged wont start. complaining it cant find libtorrent or something to that effect. but i have libtorrent installed according to apk 2021-11-26 22:04:17 always post the full error, not a bad paraphrase 2021-11-26 22:04:44 File "/usr/lib/python3.9/site-packages/deluge/_libtorrent.py", line 23, in 2021-11-26 22:04:44 ModuleNotFoundError: No module named 'deluge.libtorrent' 2021-11-26 22:04:44 import deluge.libtorrent as lt 2021-11-26 22:05:55 During handling of the above exception, another exception occurred:... (full message at https://matrix.org/_matrix/media/r0/download/matrix.org/oLUsmSOqLKvMLLCdtetTjrbt) 2021-11-26 22:08:06 according to the package database deluge requires libtorrent-rasterbar and py3-libtorrent-rasterbar, both of which I have installed 2021-11-26 22:12:10 ACTION hoped qtwebengine was also going to testing 2021-11-26 22:23:25 oh I was curious... 2021-11-26 22:23:27 >>> firefox: Build complete at Tue, 02 Nov 2021 22:27:12 +0000 elapsed time 1h 14m 30s 2021-11-26 22:23:38 >>> chromium: Build complete at Wed, 27 Oct 2021 08:43:48 +0000 elapsed time 3h 42m 37s 2021-11-26 22:25:06 :'( 2021-11-26 22:25:31 those are from Alpine builder logs 2021-11-26 22:27:37 kind of weird because on my gentoo box with old i5 4cores 3200 MHz cpu firefox 94 build in 1h 6m using 3 cores 2021-11-26 22:27:48 or Alpine build machine was busy enough 2021-11-26 22:29:12 I remember emerging xorg in the mid 2000s on my 500mhz ibook g3. that took 3 days I think. 2021-11-26 22:30:10 modern browsers are heavy from what I understand, so on a machine that's doing a ton of build jobs I believe it could take a while. 2021-11-26 22:30:49 omg, really? I was using gentoo in 2000s with celeron 800 and was pretty quick 2021-11-26 22:32:38 hmm maybe it was KDE and not xorg itself. its been a while 2021-11-26 22:33:04 that machine now runs void in CLI only. its actually fairly snapy, considering 2021-11-26 22:33:30 pretty sure an actual 80386 can compile xorg-server in less than 3 days 2021-11-26 22:34:41 yeah, say it's running at 33 MHz, that's 100 times slower than 3.3 GHz, so if it takes 5 minutes now, then it should take about 8 hours on the 80386 2021-11-26 22:35:03 this was years ago. I know it was something related to the desktop environment, but im note sure exactly what 2021-11-26 22:35:29 it might have been the entire desktop package, actually 2021-11-26 22:38:17 yes 2021-11-26 22:38:25 huh 2021-11-26 22:38:42 i thought gentoo had stopped its 32 bit ppc release 2021-11-26 22:39:46 ok never mind that, I still need ideas to fix my deluge install haha. I wouldn't mind starting from scratch if I have to, it doesn't take long, I just want to avoid that. 2021-11-26 22:39:52 i don't think gentoo officially kills anything, just has a bunch of unmaintained crap 2021-11-26 22:40:01 try apk upgrade -a 2021-11-26 22:40:41 same error 2021-11-26 22:41:03 works fine for me 2021-11-26 22:41:21 possibly you have some disk corruption, or maybe custom installed packages 2021-11-26 22:41:23 what do you have in world 2021-11-26 22:41:31 how do i check that 2021-11-26 22:42:08 mtm84[m]: I know it isnt advice at all but I had better "luck" with qbittorrent and transmission than deluge 2021-11-26 22:42:43 everything ive installed was from a repository as far as I know. i just moved from edge community to 3.15 community 2021-11-26 22:43:21 if you ised upgrade -a then should be fine 2021-11-26 22:43:26 let me clearify that. my repos had been 3.14 main, 3.14 community, and edge community, and now they are just 3.15 main and 3.15 community 2021-11-26 22:44:13 i did upgrade, i might have missed the -a switch originally 2021-11-26 22:44:31 that could be first mistake 2021-11-26 22:44:53 is something jump out when run "apk version" ? 2021-11-26 22:45:00 i saw it update a lot of packages. now that commnad with the -a switch does nothing 2021-11-26 22:45:29 Installed: Available: 2021-11-26 22:45:29 deluge-2.0.3-r9 = 2.0.3-r9 2021-11-26 22:45:58 did you try remove deluge and install again? 2021-11-26 22:46:16 yeah 2021-11-26 22:46:42 ill do it again 2021-11-26 22:46:49 there is some dependency which block that upgrade probably 2021-11-26 22:47:14 hmm 2021-11-26 22:47:29 well i must have done something stupid 2021-11-26 22:47:38 package version is same but could be build with different dependencies 2021-11-26 22:48:03 What are the advantages of this distro compared to Void? Both can use musl and are minimal 2021-11-26 22:48:10 removing and installing it just now worked, but didnt when i tried it last night lol 2021-11-26 22:48:28 alpine uses busybox which is super light 2021-11-26 22:48:36 mtm84[m]: so apk version return nothing now, right? 2021-11-26 22:49:02 with no other options, yes its blank 2021-11-26 22:49:22 well it works now so thanks 2021-11-26 22:49:23 lol 2021-11-26 22:49:31 mtm84[m]: so all fine, keep i mind to run apk -Uia upgrade everytime to avoid such a issues 2021-11-26 22:50:16 ok will do, wasnt sure the exact switches needed for a clean upgrade, i dont think i saw that in the examples i found so ill make note, thank you 2021-11-26 22:50:26 -a is important, I basically using it all time 2021-11-26 22:50:34 ok 2021-11-26 22:52:02 sometimes some downgrades happen in Alpine and without -a it wont replace a package which was build with some other dependencies in that time 2021-11-26 22:52:26 ok, after reboot the service running deluged is working fine, thank you 2021-11-26 22:52:50 Im glad it working, you are welcome :) 2021-11-26 23:42:45 does anybody know if 3.15 supports the pi zero 2 (rpi armhf)? there are some files that lead me believe so (i.e. bcm2710-rpi-zero-2-w.dtb), but I can't get past the rainbow screen of death 2021-11-26 23:59:44 lol rainbow screen of death 2021-11-27 00:00:12 this is all that i know: https://wiki.alpinelinux.org/wiki/Raspberry_Pi 2021-11-27 01:15:22 Does alpine have any package with a default Alpine logo/background or such? Equivalent to Ubuntu's ex. ubuntu-wallpapers-focal package 2021-11-27 01:56:04 Actually, https://alpinelinux.org/banner.jpg / https://gitlab.alpinelinux.org/alpine/infra/alpine-mksite/-/blob/master/_static/banner.jpg would work really well; if it's not packaged I might just wget it and be done 2021-11-27 02:42:43 ... on which note: What's the license situation on https://gitlab.alpinelinux.org/alpine/infra/alpine-mksite/-/blob/master/_static/banner.jpg ? I don't see anything discussing license/ownership on https://gitlab.alpinelinux.org/alpine/infra/alpine-mksite at all 2021-11-27 02:48:15 https://unsplash.com/@roylandnye?photo=dfZbts6B4yw 2021-11-27 02:49:02 it's mentioned in https://gitlab.alpinelinux.org/alpine/infra/alpine-mksite/-/merge_requests/3#note_82215 2021-11-27 03:46:26 Ah, thanks:) 2021-11-27 03:56:11 While there's talk about Alpine images, what's the licensing on https://gitlab.alpinelinux.org/alpine/infra/alpine-mksite/-/blob/master/_static/alpine-logo.png? (Can it be added to https://github.com/utmapp/UTM/tree/master/Icons?) 2021-11-27 04:24:41 afaik it's similar to aports: *mumblemumble* 2021-11-27 04:26:32 Hello71: In what way? 2021-11-27 05:28:36 ... I was wondering why aports, and several other repos in the alpine gitlab, don't seem to have licensing info on them. That seems... a little odd? Like, maybe some of it's not copyrightable ("download tarball from $URL, then configure;make;make install" isn't creative so AFAIK not copyrightable), but some of it certainly is. 2021-11-27 05:29:08 Is this just a historical-accident-that-would-be-really-awkward-to-revisit situation? 2021-11-27 05:31:05 Or at least, so much work that nobody wants to deal with it 2021-11-27 05:52:18 it's a mix of most of them being uncopyrightable, history, community practice, and nobody wants to bother 2021-11-27 05:55:28 there's arguably also an implied license when contributing, same as how emailed patches and github PRs (until recently) are technically all-rights-reserved but everybody understands them to be contributing back under the current license 2021-11-27 06:22:58 Where should i go to report a bug with a package? 2021-11-27 06:45:26 zola: https://gitlab.alpinelinux.org/alpine/aports/-/issues 2021-11-27 08:26:40 Hello! 2021-11-27 08:27:17 Anybody using BTRFS on Alpine ? 2021-11-27 08:27:54 I'm actually trying it out, for use in a snapshot/recovery use case. 2021-11-27 08:28:21 Used ZFS for four years, now I'm trying BTRFS for a change. 2021-11-27 08:28:53 It has some ..strange quirks still 2021-11-27 08:31:11 sometimes spews out crc errors without an apparent reason, on a SUperMicro server, ssd hdd (brand new and tested), like "BTRFS warning (device sdb): csum failed root 271 ino 257 off 74 73131520 csum 0xd5ae9caa" 2021-11-27 08:31:38 it's kinda weird 2021-11-27 12:36:00 erentar, your connection is dropping 2021-11-27 12:42:12 ddevault's contribution of crypt to disk setup is great! does swap also get encrypted? from diffs it doesn't look like that, but haven't been looking too much yet... 2021-11-27 12:43:20 km: it does not make any special provisions for encrypted swap, no 2021-11-27 12:44:41 i don't think it blocks using a swapfile inside the encrypted volume, which is good enough 2021-11-27 12:44:49 though encrypted swap isn't very fast i don't think 2021-11-27 12:57:07 psykose: never noticed any particular slowness with encrypted swap - though I'm sufficient ram that swap isn't really used in general :-) 2021-11-27 12:57:22 s/I'm/I've/ 2021-11-27 12:57:22 minimal meant to say: psykose: never noticed any particular slowness with encrypted swap - though I've sufficient ram that swap isn't really used in general :-) 2021-11-27 12:57:27 well if you never actually use the swap.. 2021-11-27 12:57:46 afaik linux always uses swap, even with plenty of ram 2021-11-27 12:58:20 ie, swap is not just used as emergency memory 2021-11-27 12:58:26 unless you set vm.swappiness=0 like a fool :p 2021-11-27 12:58:32 ikke: depends on "swappiness" setting and other things - "htop" currently shows me 0K swap used 2021-11-27 12:58:45 but yes, i frequently have like a G in there as soon as i reach 50% mem 2021-11-27 12:59:49 swap is like disk cache, except for pages that are created by programs and not related to any files 2021-11-27 13:00:36 so if you have a process that is stopped for a long time, its likely has been written out to swap 2021-11-27 13:09:09 well I'd assume if you're encrypting your filesystems including rootfs that you'd also want to encrypt swap for the same reason as an unencrypted swap might "leak" keys, passwords, etc 2021-11-27 13:09:26 assuming you use swap in the first place 2021-11-27 13:09:43 so its a performance vs security trade-off 2021-11-27 13:34:19 Hi. I am trying to move ext4 partition in gparted and it complains that it needs e2fsprogs but I have this package already installed. 2021-11-27 13:34:39 Maybe I need to load some kernel module? 2021-11-27 13:38:53 gera: e2fsprogs-extra 2021-11-27 13:39:46 thank you, works :) 2021-11-27 13:40:10 gera: Glad it helped you 2021-11-27 14:28:22 Hello, I just wanted to ask, is there an ETA for fixing the vim package? Inability to get vim into a container is... unpleasant 2021-11-27 14:29:32 Soon™ 2021-11-27 14:42:53 graywolf: To be honest, I use this as an opportunity to make these fixes easier in the future, that's why it's taking a bit longer 2021-11-27 14:45:22 ikke: no problem, I did not want to come across as demanding, I really appreciate the work everybody does on alpine and was honestly just curious if there is an estimate 2021-11-27 14:45:36 Basically today 2021-11-27 14:45:44 And thanks for the explanation :) 2021-11-27 14:56:01 anyone know why xxd keeps having a bad signature on edge? 2021-11-27 14:56:09 currently on armv7 2021-11-27 14:56:37 Danct12: yes 2021-11-27 14:59:10 #13241 2021-11-27 14:59:39 ah ok, i'll wait then 2021-11-27 14:59:49 thanks for your response! 2021-11-27 15:01:30 strangely it's fine for me on dl-cdn. and uk. on edge x86_64 2021-11-27 15:01:37 funny how it's only randomly broken 2021-11-27 15:01:49 It's not random 2021-11-27 15:02:00 and this only affects dl-cdn 2021-11-27 15:02:23 It's a known 'issue' 2021-11-27 15:06:44 Hi, anyone getting bad signature for vim and xxd in 3.15? 2021-11-27 15:06:55 #13241 2021-11-27 15:07:52 Thanks 2021-11-27 15:45:16 I tend to always encrypt swap, even when not encrypting the storage 2021-11-27 15:46:29 thanks for confirming my observation re. crypt and swap, ddevault 2021-11-27 15:49:56 graywolf: issue is fixed now 2021-11-27 15:53:59 ikke: can confirm, thank you very much of the fix :) 2021-11-27 15:54:03 for* 2021-11-27 15:56:24 and I know have a tool to do this more quickly in the future :) 2021-11-27 15:56:47 ./repo fastly purge pkg --origin vim --release v3.15 2021-11-27 16:03:39 sorry if this is stupid question, but why would not pkgrel bump and letting the normal propagation take care of it work as well? 2021-11-27 16:04:40 because dl-cdn caches the packages 2021-11-27 16:04:45 and there was a revert being done 2021-11-27 16:05:13 So yes, a pkgrel bump would've fixed it as well 2021-11-27 16:06:03 We do need to have a tool to purge latest-stable after each release anyway 2021-11-27 16:06:12 I now expanded this tool to be able to purge individual packages 2021-11-27 16:07:23 ah, makes sense, thx 2021-11-27 16:35:07 Hi, is there a program that can do text templating (with constructs such as if/for, not just variable substitution like envsubst) that is shipped in Alpine packages? Other options I think of are doing a simple Python program that uses Jinja2, or write shell script (but I want to avoid this one since I want to get templates from a 3rd party source). Any idea? 2021-11-27 16:37:03 jinja2 is probably the easiest choice 2021-11-27 16:38:02 but jinja2 doesn't come with a standalone tool and I want to avoid fetching 3rd party packages (out of Alpine packages) 2021-11-27 16:38:06 there is also https://github.com/jirutka/esh 2021-11-27 16:38:11 m4 2021-11-27 16:38:36 [])])])]])) 2021-11-27 16:39:12 and jinja2 does have a 'standalone' tool and is packaged, there is jinja2-cli 2021-11-27 16:40:13 I'm not sure esh supports for loops 2021-11-27 16:41:44 psykose: hmm it's in edge only, but it should work out 2021-11-27 16:44:59 in any case there's nothing wrong with just writing python scripts with jinja, if you're doing some Serious Templating 2021-11-27 16:45:36 I just wanted to save time, it's for something simple 2021-11-27 16:47:44 If Python is too much then I wrote templating engine in AWK. But the unless you are aching for space or need to restrict yourself to POSIX-only binaries then Jinja2 is fine. 2021-11-27 16:49:36 hmm it's in testing, then I should really write that simple Jinja2 script 2021-11-27 16:49:41 thanks anyway 2021-11-27 16:50:15 Exagone313: m4? 2021-11-27 16:52:19 m4 doesn't seem easy to understand 2021-11-27 16:53:08 Exagone313: gomplate? 2021-11-27 16:54:36 helm ? 2021-11-27 16:54:58 minimal: yes gomplate seems to be an option of what I'm looking for 2021-11-27 16:55:03 thanks 2021-11-27 18:59:47 is there a remote chance of geting php81 backported or upgraded in 3.15? 2021-11-27 19:00:50 slim 2021-11-27 19:55:58 minimal: some operating systems encrypt SWAP even if you don't encrypt your disk, for example, openbsd splits swap into chunks and each chunk is encrypted with different key, dragonflybsd has an option to encrypt entire swap with different key at each boot, but it comes at cost of being unable to debug kernel when it crashes 2021-11-27 19:57:28 ok. However I was making the point that if someone encrypts their normal filesystems then I'd expect them to ALSO encrypt swap for the same reason (security/privacy) 2021-11-27 20:03:50 ah, I see, it uses luks without lvm 2021-11-27 20:05:15 you can either use LUKS on its own or layer LUKS on top of LVM (never tried that, prefer LVM-on-LUKS) 2021-11-27 20:08:06 ddevault: if you have some basic knowledge about cryptography, you better read this https://sockpuppet.org/blog/2014/04/30/you-dont-want-xts/ 2021-11-27 20:09:27 it says to use xts 2021-11-27 20:09:27 I'd rather be happy with installer that supports ZFS' native encryption https://wiki.alpinelinux.org/wiki/Root_on_ZFS_with_native_encryption 2021-11-27 20:12:18 at least I can't find any other file system that states in documentation it uses GCM or other authenticated mode 2021-11-27 20:20:22 Hello71: I think they say that sarcastically because afaik only ZFS offers authenticated encryption 2021-11-27 20:21:29 and at the time that article was written, ZFS couldn't do encryption on linux and freebsd 2021-11-27 20:22:20 authentication doesn't matter for swap unless you are using remote storage or something 2021-11-27 20:23:03 also dm-integrity is a thing 2021-11-27 20:23:11 and bcachefs 2021-11-27 20:28:28 bcachefs sounds good to me, unfortunately wikipedia doesn't mention it https://en.wikipedia.org/wiki/List_of_cryptographic_file_systems 2021-11-27 20:28:28 [WIKIPEDIA] List of cryptographic file systems | "This is a list of filesystems with support for filesystem-level encryption. Not to be confused with full-disk encryption." 2021-11-27 20:28:40 ACTION freaked out 2021-11-27 20:29:19 for a moment I thought the bot was gonna ban me 2021-11-27 20:35:52 and because of the license it has the chance to land in kernel, I didn't like zfs anyway 2021-11-27 20:36:26 s/did/do/ 2021-11-27 20:36:27 curiousgay meant to say: and because of the license it has the chance to land in kernel, I don't like zfs anyway 2021-11-27 20:37:03 whoever implemented ed commands into the bot is a genius 2021-11-27 20:40:07 still unfortunately bcachefs is not packaged in alpine unlike zfs 2021-11-27 20:44:19 you can also use LUKS with "--integrity" apparently, never tried it 2021-11-27 20:54:31 like authenticate entire disk and error once there's hardware failure? 2021-11-27 21:01:27 anyway, that article about XTS appeared before bcachefs, so back then you had to use solaris/illumos in order to have authenticated filesystem-level encryption 2021-11-27 21:02:48 and solaris/illumos is not what end-user wants because it has less programs than *bsd 2021-11-27 21:23:31 Hello71: you know, I think there's a potential replacement for blender - wings, it only needs erlang, erlang-wx, erlang-cl (opencl binding) and libigl (geometry processing), I'll be happy to create aports for erlang-cl, libigl and wings after I install alpine 2021-11-27 21:24:47 I call it potential only because it doesn't support rigging 2021-11-27 22:03:31 hey, my wifi stops working after my laptop wakes up from suspend 2021-11-27 22:03:45 im using wpa supplicant and i have to restart the networking service each time 2021-11-27 22:22:38 curiousgay: the article you linked literally says to use xts 2021-11-27 22:28:57 Hey, is there a way to make Alpine _save_ coredumps? I have gotten a few segfaults lately but no coredumps 2021-11-27 22:29:07 ulimit -c unlimited 2021-11-27 22:29:12 Thinking of something like coredumpctl but I don't know if it's necessary 2021-11-27 22:29:22 ^ 2021-11-27 22:29:23 ikke: ulimit already says unlimited though 2021-11-27 22:29:45 hmm, it defaults to 0 for me 2021-11-27 22:30:14 Oh, ulimit -c says 0 2021-11-27 22:30:18 What is `-c` anyway? 2021-11-27 22:30:33 And what do I do to enable it by default? Do I just put it in an rc.local? 2021-11-27 22:30:57 psykose: I think that was said sarcastically because in 2014 the only option for end users to have authenticated filesystem-level encryption was to use solaris/illumos, even freebsd didn't support zfs' encryption back then, and we both know that solaris/illumos not only has less programs than linux, but also less programs than *bsd and that's a red flag for end users 2021-11-27 22:31:11 Nulo: ulimit -c sets the max size of coredumps 2021-11-27 22:31:40 Nulo: if it's for a service, you should add it to /etc/conf.d/ rc_ulimit 2021-11-27 22:31:55 if it's for a user, /etc/profile(.d) or ~/.profile 2021-11-27 22:32:04 .profile it is, thanks! 2021-11-27 22:32:33 but now both freebsd 13 and zol support it, and there's bcachefs that has potential to land into linux kernel 2021-11-27 22:33:23 Is it just me or netsurf-gtk3's Preferences menu shows generic strings? 2021-11-27 22:33:51 (of course on freebsd 13 you must do it manually because installer still prefers geli) 2021-11-27 22:34:27 ScrumpyJack: netsurf-gtk3 shows generic strings in the preferences menu, any clues? 2021-11-27 22:41:16 curiousgay: it's very much not sarcastic, but what exactly are you proposing instead? to only allow people to run the setup-disk-encryption script to use zfs? 2021-11-27 22:55:53 that's one option, another option is to use bcachefs, one more option is to use layered FS if it offers authentication (currently I'm not aware of any, if there is one I blame the lack of documentation) 2021-11-27 23:08:40 I can't seem to get mesa to use modesetting/_not_ mesa-dri-intel on my i5-3320M, any clue? 2021-11-27 23:08:51 It goes straight to llvmpipe 2021-11-27 23:18:13 modesetting is an xorg driver, it is unrelated to mesa opengl driver 2021-11-27 23:23:51 holy, wifi setting is ... just firmware-other is 188MB and then fuckery with iwd and dbus 2021-11-28 01:03:26 Hello71, I guess I mean the gallium driver 2021-11-28 01:03:59 huh? 2021-11-28 02:16:12 Hello71, sorry, I confused some things. It seems the i915 is the right (only?) driver for this board. I wanted to switch away from it because the `mesa-dri-intel` package is deprecated, and there's newer drivers like Iris 2021-11-28 02:16:59 Oh wait, forcing the crocus driver with MESA_LOADER_DRIVER_OVERRIDE seems to work. I guess I will try that 2021-11-28 02:17:47 https://www.phoronix.com/scan.php?page=news_item&px=Intel-Crocus-April-2021-State 2021-11-28 02:20:17 Nulo: mesa-dri-classic is much smaller packag than gallium and works for i965 2021-11-28 02:20:54 helby, yes, but as I understand it, it will be removed in the (near?) future and the newer driver has better performance 2021-11-28 02:21:24 https://www.phoronix.com/scan.php?page=news_item&px=i965-Mesa-Classic-To-2022 2021-11-28 02:21:32 > Deleting classic Mesa didn't happen for Mesa 21.3 as was discussed as a possible target before, but it looks like it will be removed in 2022. 2021-11-28 02:21:59 not sure about a future, but right now, for wayland gallium does not work for me, I need classic 2021-11-28 02:22:13 fori 2520M 2021-11-28 02:23:36 That's too old, crocus supports Gen4 upwards and yours is 3rd gen. I believe the i915g driver is the "future" driver for that one. 2021-11-28 02:23:58 yes, I am not too young too ;/ 2021-11-28 02:33:32 "To old", he says, as I struggle to get a rage 128 working in 2021 2021-11-28 04:03:11 Hello! What's the rust situation on alpine-linux? Are pre-built bootstrap binaries used? 2021-11-28 04:03:25 Alpine Linux, pardon. 2021-11-28 04:06:52 ttps://git.alpinelinux.org/aports/tree/community/rust/APKBUILD 2021-11-28 04:07:01 it's built from source 2021-11-28 04:07:05 * https://git.alpinelinux.org/aports/tree/community/rust/APKBUILD 2021-11-28 04:17:25 I'll read it, but as a spoiler, does it cross-build to other architectures (from the mrustc-built x86_64) ? 2021-11-28 04:19:55 I'm not sure, but the packages appear to be built with native rust 2021-11-28 04:21:08 I'm not sure how it was originally bootstrapped 2021-11-28 04:22:13 I thought most distros don't care and simply use fresh official binaries for bootstrapping 2021-11-28 04:22:20 ah, OK. yeah, I see: _bootver=1.55.0, and then wonder where that bootstrap is gotten from 2021-11-28 04:22:42 Debian does it at least, but they care (they'd rather bootstrap it from source if it was possible) 2021-11-28 04:23:04 Guix bootstrap via mrustc from sources (starting with the rust 1.39.0 sources) and building all the way up to 1.52 currently 2021-11-28 04:23:13 but that's only supported on x86_64, which sucks 2021-11-28 04:23:55 rust-bootstrap=>$_bootver is provided by a previous version of the package 2021-11-28 04:24:13 you mean debian doesn't download official binaries from mozilla in order to build rust from source? 2021-11-28 04:24:23 apparently it's possible to use that x86_64 rustc build and cross-build itself with it for other architectures... but so far I've yet to find an example. 2021-11-28 04:24:59 curiousgay: they do, but they'd rather not (see: https://salsa.debian.org/rust-team/rust/-/blob/debian/sid/debian/README.source#L86) 2021-11-28 04:29:49 apteryx: that's what every other distro (except guix) does 2021-11-28 04:30:45 because rust is a bloated shit and depends on its previous release features 2021-11-28 04:36:34 just because its source code is available doesn't mean it has no backdoors on source level, afaik linux kernel, ubuntu, firefox, chromium and musescore contain backdoors 2021-11-28 04:37:53 s/previous release/last previous release/ 2021-11-28 04:37:53 curiousgay meant to say: because rust is a bloated shit and depends on its last previous release features 2021-11-28 04:44:00 now tell us how you really feel 2021-11-28 04:45:37 hey boys, someone wrote on Reddit about vim and xxd having bad signatures and that seems to be correct. anyone who can pass it on to the maintainers? 2021-11-28 04:47:28 Lampomaniac: it's https://gitlab.alpinelinux.org/alpine/aports/issues/13241 2021-11-28 04:48:03 ah nice, this community works quick :) 2021-11-28 15:18:37 what's the runit stage 1 equivalent in alpine ? (a.k.a. where to put my oneshot (systemd-speak) scripts i need to run on boot)? 2021-11-28 15:26:43 piraty: look at /etc/local.d/README 2021-11-28 15:27:44 might possibly be what you need 2021-11-28 15:28:11 thanks 2021-11-28 15:28:23 looks right, although it's a hack when a 'service' does this ;) 2021-11-28 16:06:12 piraty: you can start with 'init --help' then move forward ;/ 2021-11-28 16:06:34 local seems fine for now 2021-11-28 16:15:32 it's pretty simple, runlevel is just a directory and service is symlink to init.d 2021-11-28 19:30:36 I installed linux-lts, which upgraded to 5.15.5-0(on edge), I also installed linux-edge. When booting into lts version, boot just locks, but if I try edge kernel, it will say loading modules and errors saying it can't cd into 5.15.5-edge-0. Curiously if I chroot into the partition, PATH isn't set so I have to export PATHS=/bin:/usr/bin:/sbin in order to use any commands. I am suspecting maybe PATHS isn't being set at boot possibly? 2021-11-28 19:43:32 well it's PATH, not PATHS, and chroot is an unrelated issue 2021-11-28 19:43:40 probably you are using wrong initramfs 2021-11-28 19:51:59 ndowens: sounds similar to #13218 2021-11-28 19:57:20 ndowes: are you sure the linux-edge error is a cd failure for "5.15.0-edge-0" and not "5.15.5-0-edge"? 2021-11-28 19:58:18 that's exactly what i had as well with upstream kernel 2021-11-28 20:03:06 caskd: upstream kernel? 2021-11-28 20:03:24 yeah 2021-11-28 20:03:39 i mean edge 2021-11-28 20:04:40 caskd: so you had the "5.15.5-edge-0" (not "5.15.5-0-edge") error? 2021-11-28 20:05:22 no, just the framebuffer error 2021-11-28 20:05:29 sorry if i misunderstood 2021-11-28 20:05:42 just looked at the linked issue and wanted to say i was also affected 2021-11-28 20:05:45 ^^' 2021-11-28 20:06:43 caskd: you mean you have a problem with linux-lts (on Alpine Edge) but not with linux-edge (on Alpine Edge), right? 2021-11-28 20:07:52 it was on edge actually for me, but it hung at exactly the same spot 2021-11-28 20:08:02 and now lts is the edge version i tried 2021-11-28 20:08:08 or 2 patches behind of what lts is now 2021-11-28 20:08:20 i haven't tried latest yet 2021-11-28 20:10:19 caskd: confused, are you confirming that what I just asked is the case or not? 2021-11-28 20:10:49 caskd: linux-lts package on Alpine Edge? on Alpine 3.15.0? 2021-11-28 20:11:13 it was on linux edge 5.15.1 2021-11-28 20:11:49 so the issue likely happened on both configurations 2021-11-28 20:12:01 caskd: ok, and what's your usual graphics card driver? i915? amdgpu? radeon? 2021-11-28 20:12:08 amdgpu 2021-11-28 20:14:51 caskd: and how did you fix it / resolve the problem? 2021-11-28 20:15:17 i just fell back to a previous version 2021-11-28 20:15:28 i didn't find a resolution yet 2021-11-28 20:16:01 caskd: was this a sys (disk install) or run-from-ram? 2021-11-28 20:16:11 sys 2021-11-28 20:17:41 caskd: ok. I guess if you were to try it again one thing to check after upgrading, but before rebooting, would be to see if the initramfs (which should be updating when the new kernel is installed) includes the amdgpu firmware 2021-11-28 20:18:59 s/updating/updated/ 2021-11-28 20:18:59 minimal meant to say: caskd: ok. I guess if you were to try it again one thing to check after upgrading, but before rebooting, would be to see if the initramfs (which should be updated when the new kernel is installed) includes the amdgpu firmware 2021-11-28 20:21:21 Wouldn't i have to write a amdgpu.files for mkinitfs? 2021-11-28 20:21:26 so it gets included 2021-11-28 20:31:28 you need to add `kms` to the features list i think 2021-11-28 20:33:50 i don't get any gpu firmware without it in there, because the way the firmware works is it includes firmware based on the included module modinfo, and kms feature includes all the gpu modules 2021-11-28 20:34:43 if you have `linux-firmware` installed this will place every single firmware file for every gpu in there pretty sure, so you want to only have the linux-firmware-* of what you actually use i guess 2021-11-28 21:14:38 caskd: nope, mkinitfs does a "modinfo | grep firmware" on all the modules it includes to check what firmware files they need, and the amdgpu module declares a list of files so those *should* be automatically be included in the initramfs *if* the linux-firmware-amdgpu package is installed 2021-11-28 21:15:15 ah, didn't see psykose already explained :-) 2021-11-28 23:12:17 I updated system and used newer mkinitfs.conf file and now it works fine. 2021-11-28 23:41:25 Is there a good tui packages manager for alpine? I know void has fuzzypkg and a website to index packages and I didn't know if something existed for alpine 2021-11-28 23:47:54 why 2021-11-28 23:48:08 more precisely, what would it do 2021-11-28 23:56:30 I found it, the packages tab on the website 2021-11-29 00:26:42 ndowens: "I updated system and used newer mkinitfs.conf file" - do you mean you modified mkinitfs.conf yourself (before running mkinitfs) or that you updated via something like "apk upgrade -a"? 2021-11-29 00:27:42 i added kms initially and that didn't work 100% and i upgraded with apk upgrade and seen new conf and so i mv'd new conf to replace older one 2021-11-29 00:29:20 ndowens: hmm, what was different between old mkinitfs.conf and new one (apart from "kms")? 2021-11-29 00:30:33 not competely sure, i suspect it was mostly due to older pkgs causing the issues more so than conf 2021-11-29 00:33:37 ndowens: right, the feature files (/etc/mkinitfs/features.d/*.modules) of older versions of mkinitfs package referenced "*.ko" which wouldn't place compressed kernel modules into initramfs. What this an upgrade you did from Alpine 3.14 to 3.15? 2021-11-29 00:34:35 nah i use edge 2021-11-29 00:36:32 ndowens: same issue would apply for Edge if you have the older mkinitfs package (before compressed mod support) and that package didn't get updated when you installed linux-lts. did you do just a "apk upgrade" or a "apk -Ua upgrade"? The 2nd version should ensure all packages are up-to-date 2021-11-29 00:37:05 just apk upgade, i will do 2nd part to ensure 2021-11-29 00:37:32 well that gave a few extra updates 2021-11-29 03:18:33 looks like thunderbird in edge needs rebuild as it gives "unable to select packages: so:libffi.so.7 (no such package)" 2021-11-29 08:49:07 Hi all 2021-11-29 08:50:41 I want to have your opinion here as I think I did a mistake 2021-11-29 08:52:54 I've added SDLPop to testing repository, but it's legality is debatable. Should I remote it quickly ? (clandmeter, ikke, ncopa, programmerq) 2021-11-29 08:53:12 https://news.ycombinator.com/item?id=29316058 2021-11-29 09:01:47 I personally wouldn't want anything "legally debatable" in official repos, unless those repos physically resided in a jurisdiction with looser legal issues. 2021-11-29 09:02:19 but thats just me 2021-11-29 09:03:34 Any idea why 'eiwd' - "Internet Wireless Daemon without dbus" needs dbus? ;/ 2021-11-29 09:03:48 mtm84[m]: Yes, that's a good point, I think I wont wait longer and remove it 2021-11-29 11:02:12 Hello! How do i run a service as a specific user? 2021-11-29 11:08:39 durak: using OpenRC? Set the user_command variable to the desired user. 2021-11-29 11:24:21 command_user( 2021-11-29 11:24:23 * 2021-11-29 12:40:03 hi all 2021-11-29 12:40:46 on one box, apk update doesn't work anymore ... says "untrusted signature" ... i think it still has 3.14 and now because 3.15 is out maybe it's lacking a new signature? 2021-11-29 12:45:49 oh, okay, manually going to v3.14 then v3.15 helped 2021-11-29 12:48:32 buuuut can't make a new initramfs ... yikes o_O 2021-11-29 12:49:04 because the running kernel is not the installed kernel anymore. any way around this? or no choice but to poweroff and boot from an older USB stick? 2021-11-29 12:50:36 should mkinitfs use the last kernel/modules ? 2021-11-29 12:51:45 otherwise it will fail in any kernel upgrade 2021-11-29 12:52:17 have you tried 'apk fix mkinitfs' ? 2021-11-29 12:52:34 aha! 2021-11-29 12:52:47 apk add cpio, and THEN apk fix mkinitfs, that fixed it. thank you, donoban :) 2021-11-29 12:53:05 apparently you need GNU cpio, not busybox cpio. 2021-11-29 12:53:20 ahm, ok great 2021-11-29 12:53:44 I don't have it installed 2021-11-29 12:53:50 pretty sure you don't 2021-11-29 12:54:14 posting the error message would probably be more useful than 'it didnt work' 2021-11-29 12:55:11 it's now working 2021-11-29 12:55:30 fwiw, the error message was "cpio: unrecognized option: renumber-inodes" 2021-11-29 12:55:52 busybox cpio simply doesn't have that, so it made sense 2021-11-29 12:58:53 `busybox cpio -h` and you will see --renumber-inodes at the bottom 2021-11-29 13:06:57 rapha: https://alpinelinux.org/posts/Alpine-3.15.0-released.html see upgrade notes 2021-11-29 13:07:33 https://alpinelinux.org/posts/Alpine-3.15.0-released.html#upgrade_notes 2021-11-29 13:07:46 oh. i forgot --available again :( 2021-11-29 13:33:44 maybe default available will be better, what is the purpose of that option? Avoid reinstall same packages that just rebuilded or a mirror change? 2021-11-29 14:31:03 is anyone running an aarch64 alpine vm on a x86_64 host and can help me with the setup? I can't figure out the magic incantations from what's available online 2021-11-29 14:38:40 c7s: using which hypervisor? qemu? 2021-11-29 14:38:51 yeah 2021-11-29 14:39:48 I gather there's something to do with EFI you need to account for 2021-11-29 14:41:07 c7s: this was talked about on here about 10 days ago. Let me dig up the qemu cmdline I use 2021-11-29 14:45:08 c7s: qemu-system-aarch64 -machine virt -cpu cortex-a53 -display none -serial stdio -drive if=pflash,format=raw,readonly,file=/usr/share/AAVMF/AAVMF_CODE.fd -drive if=pflash,format=raw,file=./machine-AAVMF_VARS.fd -device virtio-scsi-device,id=scsi -drive file=alpine.qcow2,if=scsi,format=qcow2,id=hd0 -device scsi-hd,drive=hd0 2021-11-29 14:45:34 minimal: I'm using some more options for virtio net, rng, etc 2021-11-29 14:45:36 thanks 2021-11-29 14:47:25 minimal: what package do the AAVMF files come from? 2021-11-29 14:49:44 c7s: the AAVMF_CODE.fd is Arm UEFI stuff. In Alpine its not currently packaged "correctly" as the x86_64 "ovmf" package correctly has the x86_64 equivalent files whereas the aarch64 "ovmf" package has a "QEMU_EFI.fd" file which is actually the "AAVMF_CODE.fd" file plus you cannot install the aarch64 package on a x86_64 machine to get the right files - both these packages should be changed so both sets of files can be installed on any type 2021-11-29 14:49:44 of aarch 2021-11-29 14:50:35 so where do you get it 2021-11-29 14:50:36 I've just grabbed the AAVMF files from the Debian qemu-efi-aarch64 package :-) 2021-11-29 14:50:39 ah 2021-11-29 14:50:52 that is what arch does as well 2021-11-29 14:51:06 https://aur.archlinux.org/cgit/aur.git/tree/PKGBUILD?h=aavmf-bin 2021-11-29 14:51:07 hehe 2021-11-29 14:51:17 minimal: ugh, guess I'll dig in some packages then 2021-11-29 14:51:25 I've been meaning to look at fixing the Alpine ovmf package... 2021-11-29 14:51:46 c7s: you can follow that arch package right now, it's just unpacking a .deb and you have it 2021-11-29 14:51:48 you just add -cdrom [...].iso during setup right? no other funny business 2021-11-29 14:52:36 wait it actually depends on a debian package 2021-11-29 14:52:38 that's amazing 2021-11-29 14:52:43 this is on my list of "things I'd like to fix in Alpine before the death of the universe" :-) 2021-11-29 14:57:39 c7s: alternative you could always download the Alpine ovmf *aarch64* package file and unpack from that and rename the files... 2021-11-29 15:01:16 I've already got the debian ones, gtg now but I'll try in an hour or so 2021-11-29 15:02:19 c7s: interested to know how you get on - I've only been testing my own Alpine aarch64 disk images this way 2021-11-29 15:07:40 I suspect forgetting the "-a/--available" option for "apk upgrade" is why some people are having 3.15/Edge "boot hang" issues (due to combination of compressed kernel modules and switch to Simpledrm) 2021-11-29 15:11:11 i wish i had an amdgpu machine to see myself if i can reproduce it with that 2021-11-29 15:11:15 In release 3.15 the package ucspi-tcp6 seems to have a conflict with coreutils. Whenever I install coreutils, the package ucspi-tcp6 is purged. How can I solve this issue? 2021-11-29 15:13:19 midasi: you report it on issues, someone has to fix ucspi-tcp6 2021-11-29 15:13:51 ok thx 2021-11-29 15:14:37 ucspi-tcp6 seems a bit broken though, the conflicts are in it providing e.g. `date@` and that command can't even be run 2021-11-29 15:17:37 depending on what you are using it for (i assume just tcpserver) you could also use s6-tcpserver from s6-networking 2021-11-29 15:19:20 at least that's what i do on my servers for tinysshd 2021-11-29 15:20:45 psykose: I have an old AMD machine here which uses either radeon or amdgpu that I plan to scrub and put Alpine on so I might get a chance to test that 2021-11-29 15:21:33 amdgpu is anything newer than like.. hd7750 2021-11-29 15:21:56 although those are more using the -si compat of amdgpu driver 2021-11-29 15:23:35 and that compat is disabled on edge and enabled on lts, so you want volcanic islands or newer 2021-11-29 15:23:38 from here https://wiki.gentoo.org/wiki/AMDGPU 2021-11-29 15:37:03 psykose: ucspi-tcp6 is quite old and should be upgraded. In other distros the updated version seems to run fine 2021-11-29 15:46:21 Hey folks, I'm trying to create a custom mini-repo in a container using Alpine and `apk fetch` everything has been going well except I can't seem to figure out include `libcrypto1.1`. Says it's missing on my client and is obviously required by a lot of other things. 2021-11-29 15:46:21 psykose: yeah I have an old single core Sempron box that I use for testing stuff :-) 2021-11-29 15:49:35 ShredGhoul: did you try to include libcrypto1.1? 2021-11-29 15:54:56 Apparently not... 2021-11-29 15:55:01 ...thanks! 2021-11-29 16:12:14 I am not quite using go, but fould nice set of tools, projectdiscovery.io 2021-11-29 16:35:41 what's the best way to make a global keybind (eg media keys) in the kernel console? (is there a way less hacky than a tmux keybind, I guess) 2021-11-29 16:39:38 not sure what is hacky about a tmux keybinding 2021-11-29 16:40:25 it doesn't work everywhere 2021-11-29 16:40:35 minimal: works great, thanks 2021-11-29 16:40:42 I guess that was phrased inelegantly though, 2021-11-29 16:40:58 "something that isn't so hacky the tmux keybinding is preferable despite it not working everywhere" ? 2021-11-29 16:41:22 the least hacky way to bind things would be to start a graphical session and not use the console 2021-11-29 16:42:32 the machine is 10+ years old, so I'm trying to have a minimal memory footprint 2021-11-29 17:29:38 hello, how do i start a service as a specific user? 2021-11-29 17:30:03 command_user in the init.d 2021-11-29 17:30:59 e.g. command_user=myuser:mygroup 2021-11-29 17:35:09 thanks i'll have a look 2021-11-29 18:49:31 Hi 2021-11-29 18:51:22 I've been enjoying Alpine in my VMs so I was thinking to attempt a storage server on bare metal. It's there anything that would speak against this use case in favour of another distro? 2021-11-29 18:52:41 I'd run iscsi target, nfs and samba servers, as well as rsync. Maybe netdata abs a few smaller services 2021-11-29 18:54:36 Hi, official EC2 images have link alpine.linux.org/cloud, should be alpinelinux.org/cloud; not sure where exactly to report this 2021-11-29 18:56:04 Either on #alpine-cloud, or https://gitlab.alpinelinux.org/alpine/cloud/alpine-ec2-ami 2021-11-29 18:57:02 thx 2021-11-29 19:06:21 Forza: Alpine should be able to do that 2021-11-29 19:06:41 :) 2021-11-29 19:07:01 I'll have a go at it. Thanks 2021-11-29 20:40:26 Hi, wondering if there is any reason why I got an error message with setup-alpine from virt iso when trying to use/setup a vlan interface (ERROR: unable to select packages vlan-2.2-r2 breaks: ifupdown-ng-0.11.3-r0[!vlan] satisfies: world[vlan]) the interface is working fine but I am wondering why this message 2021-11-29 20:45:13 because the vlan package conflicts with the ifupdown-ng package 2021-11-29 20:47:38 my curiosity is more to understand if there is something to be done to fix that on the setup-alpine script 2021-11-29 20:49:44 searI assume it should be ifxed in a package if alpine deciced to use ifupdown-ng instead of busybox and some package depends on busybox-ifupdown 2021-11-29 20:59:50 blt[m]: can you file an issue at https://gitlab.alpinelinux.org/alpine/alpine-conf/-/issues 2021-11-29 22:48:27 oof, on i686 I'm getting a SIGILL from ripgrep 2021-11-29 22:48:55 (need to investigate more, not sure if this is a ripgrep bug or a packaging flags bug) 2021-11-29 22:51:07 or bad ram 2021-11-29 22:51:55 it's reproducible on every run 2021-11-29 22:52:07 gdb says its dying on an sse4.1 instruction 2021-11-29 22:52:14 and yep my CPU does not have that... 2021-11-29 22:53:02 I guess, is this a packaging bug? what're the minimum cpu flags i686 attempts to support 2021-11-29 23:01:29 sse 2021-11-29 23:02:04 i think 2021-11-29 23:02:41 cmov? 2021-11-29 23:04:34 ah, yes, 486 has cmpxchg, not cmov 2021-11-29 23:04:44 but you're not actually using a pentium pro, are you 2021-11-29 23:13:56 Never assume people aren’t running old hardware >.> 2021-11-29 23:46:50 no, T2400 2021-11-29 23:46:56 it has sse2, but no sse4 2021-11-29 23:47:44 the insn is pextrd 2021-11-30 00:08:16 mtm84[m]: there's the one guy with a vortex86 but i think that's because of no-intel 2021-11-30 00:12:02 Impressive. Wonder why he didn’t go for a via chip. I wasn’t aware anyone else had an x86 license. 2021-11-30 00:36:09 Hey guys, need help for running eww on musl alpine linux. It's related to https://github.com/elkowar/eww/issues/253 but I can't figure out how to avoid compiling gtk3 for a working eww bin. 2021-11-30 00:57:36 what is a 'shared package' instead of rebuilt? the gtk3 in alpine repos is built on alpine builders 2021-11-30 00:57:48 Goodnight. After a few days using another system, I couldn't resist I went back to alpilinux. Very fast in this version. 2021-11-30 00:57:59 psykose: huh? 2021-11-30 00:58:09 Hello71: in the linked issue at the bottom 2021-11-30 00:58:12 oh 2021-11-30 00:58:22 i missed that line 2021-11-30 00:59:32 Is it possible to install Torbrowser? 2021-11-30 01:04:17 if you build it, sure 2021-11-30 01:04:20 it's not packaged though 2021-11-30 01:04:49 i don't think it builds 2021-11-30 01:04:53 else: flatpak -> flathub -> torbrowser-launcher 2021-11-30 01:08:20 all i can get for "eww" is that like most rust binaries compiled with the rustup toolchains for me it sigsegvs with 0 information 2021-11-30 01:08:34 or, well 2021-11-30 01:08:39 this one has at least some 2021-11-30 01:09:32 appears to segfault on gtk::init for some reason 2021-11-30 01:11:30 I found that there is no an lxappearance package. How would I change the gtk theme on my window manager then? 2021-11-30 01:14:07 gsettings set org.gnome.desktop.interface gtk-theme $theme 2021-11-30 01:14:09 pretty sure 2021-11-30 01:14:41 I don't use gnome 2021-11-30 01:14:51 jason123onirc: it is in testing repo 2021-11-30 01:15:39 lxappearance is in the testing repo? 2021-11-30 01:15:45 not sure what gsettings has to do with gnome 2021-11-30 01:15:59 i also export GTK_THEME and that also works 2021-11-30 01:16:08 yes, in testing repository 2021-11-30 01:16:38 and config/gtk/settings.ini gtk-theme-name = $theme 2021-11-30 01:16:45 but from what i remember these don't all work the same 2021-11-30 01:17:57 is the testing repository recommended to install or is it only mostly for testing? 2021-11-30 01:18:14 highly recommended! ;/ 2021-11-30 01:19:14 does not that crap use still gtk2? 2021-11-30 01:22:19 I think it still uses gtk2 but it can set gtk3 themes 2021-11-30 01:24:59 its the only way to set the gtk theme from the gui that works independently of the desktop environment. 2021-11-30 01:25:11 if add --enable-gtk3 then it will use gtk3 2021-11-30 01:27:47 psykose: So you confirm that eww segfault on your musl system too ? And no way to debug that ? 2021-11-30 01:28:00 i have no idea what this library code does 2021-11-30 01:28:15 I can try to set the theme with gsettings first and then try lxapperance if it does not work 2021-11-30 01:28:24 https://termbin.com/wdw6 2021-11-30 01:28:38 it segfaults somewhere here 2021-11-30 01:28:56 in the gtk crate specifically, rt.rs 2021-11-30 01:30:06 but how it can be related to musl ? 2021-11-30 01:30:14 jason123onirc: if dont wanna install/compile LXAppearance to generate two config files then can use this: https://tpaste.us/PreQ 2021-11-30 01:30:37 better link https://github.com/gtk-rs/gtk3-rs/blob/0.14.3/gtk/src/rt.rs#L115 2021-11-30 01:30:42 but i don't care to debug whatever this is doing 2021-11-30 01:30:52 or just install, set up and uninstall I assume it shoudl work 2021-11-30 01:31:45 arch wiki has info about gtk nonsense, u can check it there 2021-11-30 01:32:57 yeah I looked at the arch wiki 2021-11-30 01:34:33 but if u have good connection you can also install it, it does not have dependencies like gnome ;/ 2021-11-30 01:35:36 yeah I could just install it 2021-11-30 01:35:48 did not know it was in the testing repos 2021-11-30 01:46:24 kxzaon: all i can guess is that `&mut 1` in the code is a raw pointer to literally 0x1, and it gets passed to https://docs.gtk.org/gtk3/func.init_check.html which says it will modify it... so 2021-11-30 01:47:02 i don't know why this wouldn't break on glibc though, or why they don't pass a fake argc or something 2021-11-30 01:51:22 alright... Too tricky for me. I appreciate the effort. Thx. 2021-11-30 01:54:09 somewhere down the chain of that gtk function it calls g_option_context_parse, which does a `*argc -= k` which would 'obviously' segfault with a &0x1, but idk if that's what really happens 2021-11-30 01:55:35 it dereferences *argc before that even, who knows, it would take more setup i think to get a backtrace past the rust code 2021-11-30 01:55:44 and why glibc doesn't fail on this i can't tell you either 2021-11-30 02:04:31 gosh... I'm drowning here. I was hoping that a similar issue in the past with gtk3 that can help me with eww 2021-11-30 02:29:39 rust pointers are *, not & 2021-11-30 02:30:20 i'm not familiar with the syntax &mut 1 but i'm pretty sure it means to construct a temporary int 1 and then pass the address 2021-11-30 02:31:22 ah, true 2021-11-30 02:31:32 that should be fine then 2021-11-30 02:33:49 “Should be” 2021-11-30 02:42:33 are you noodling 2021-11-30 02:42:37 i will eat more hotdogs 2021-11-30 03:20:47 Greetings 2021-11-30 03:21:32 recently upgraded my diskless setup to alpine 3.15, unfortunately there's a change in the kernel that broke vfio/amdgpu passthrough 2021-11-30 03:21:47 from what I read online this has since been fixed in kernels 5.13+ 2021-11-30 03:22:32 apologies, I meant 5.15.3+ 2021-11-30 03:23:44 and so I wish to know about bumping my kernel 2021-11-30 03:23:59 I don't think a minor jump would cause many issues would it? 2021-11-30 03:26:51 I use update-kernel when I wish to include modules that are not included by default. Am I right in assuming this can be used in my case too? 2021-11-30 03:28:34 namely, to update the kernel that shipped with alpine 3.15 to, say, version 5.15.5 2021-11-30 03:29:15 the kernel in 3.15 is 5.15.5 2021-11-30 03:32:45 recentlyupgrade: you have to wait for Alpine 3.15.1 images or try build them by self 2021-11-30 03:32:53 oh dear, indeed it is 2021-11-30 03:32:56 pretty sure you can just run upgrade -U 2021-11-30 03:33:08 psykose: he using diskless mode 2021-11-30 03:33:35 does diskless not let you modify the boot image in some fashion 2021-11-30 03:34:25 yes I can modify it 2021-11-30 03:34:50 but I must have been looking in the wrong place because 3.15 is running 5.15.5 already.. 2021-11-30 03:35:03 kind of, depends how much hate writing on any media :) 2021-11-30 03:35:16 did you check with uname -a 2021-11-30 03:35:22 yes 2021-11-30 03:35:39 you aren't the only person that reported vfio breaking for them 2021-11-30 03:35:50 so update-kernel done the job 2021-11-30 03:35:53 iirc you need to blacklist gpu stuff or some shit for it 2021-11-30 03:36:05 and in 3.15 there is simpledrm now and maybe affected it 2021-11-30 03:36:39 and someone with the same issue unblacklisted everything and it fixed it for them, dunno 2021-11-30 03:36:40 exact vfio setup works perfectly with alpine 3.14, in recent kernels AMD pushed some changes that makes the driver grab the gpu (after vfio grabbed it) 2021-11-30 03:37:34 ah, seems you know more than me then 2021-11-30 03:38:22 psykose: I doubt it :), just wondering, wouldn't putting /etc/modprobe.d/drm.conf softdep pre: vfio vfio_pci be sufficient ? 2021-11-30 03:39:00 iirc that is how you normally have to do vfio 2021-11-30 03:39:22 i don't know how diskless works, do you still have the initramfs doing stuff before the actual boot 2021-11-30 03:39:35 because if so the initrd doesn't have vfio modules in it i don't think 2021-11-30 03:40:36 and iirc the new simpledrm changes force drm to activate and load gpu firmware inside the initrd sequence 2021-11-30 03:40:54 no nothing special, stock alpine. I simply populate /etc/modprobe.d to make it so vfio_pci loads before amdgpu (and with gpu IDs specified, it grabs them), dmesg shows vfio grabbing them.. But then amdgpu still takes the gpu (lspci -k shows amdgpu) 2021-11-30 03:41:10 hmm 2021-11-30 03:41:23 psykose: I see.. would I need to have vfio built in then? 2021-11-30 03:42:09 or possibly disable simpledrm ? 2021-11-30 03:43:13 to do that you would have to build your own kernel, and as for vfio edit the mkinitfs and make a new one 2021-11-30 03:47:52 to do the latter.. you would need to make a vfio.modules with something like https://termbin.com/o9y0 in /etc/mkinitfs/features.d and add vfio to mkinitfs.conf features, and rebuild a new one for the diskless image 2021-11-30 03:48:00 but i have no idea if any of this works as i have never tried vfio :) 2021-11-30 03:51:00 actually just "kernel/drivers/vfio" should work i think, no need for the two lines 2021-11-30 03:52:21 psykose: so inside vfio.modules I would put kernel/drivers/vfio ? 2021-11-30 03:52:28 yep 2021-11-30 03:53:17 while at it, I also wish to pass a kernel command line (of vfio ids so it binds before anything else gets a chance) 2021-11-30 03:53:37 if you are building the image you can put that in the bootloader 2021-11-30 03:53:44 this also needs to be added to mkinitfs.conf isnt it? 2021-11-30 03:54:09 no, bootloader options 2021-11-30 03:54:57 that would be grub.cfg then 2021-11-30 03:55:00 yep 2021-11-30 03:55:33 if you generate the grub config with grub-mkconfig -o grub.cfg then it reads from /etc/default/grub GRUB_CMDLINE_LINUX_DEFAULT 2021-11-30 04:00:20 psykose: what if, on my diskless media, I -oremount,rw and just put the generated file into /media/usb/boot/grub/grub.cfg 2021-11-30 04:00:22 sane? 2021-11-30 04:01:01 maybe? i don't know if it's possible to just rw diskless and edit it that way 2021-11-30 04:01:07 i don't think it is? 2021-11-30 04:01:29 since the boot is from a usb/whatever iso, you need to actually write a new one 2021-11-30 04:01:49 this goes for all the other changes as well 2021-11-30 04:03:18 psykose: many thanks for the tips, I will experiment 2021-11-30 04:04:56 see you later 2021-11-30 04:51:46 units_cur seems broken in stable: https://termbin.com/8l11 2021-11-30 05:20:54 !27976 should fix it 2021-11-30 05:22:45 i'll make one for 3.15 also 2021-11-30 05:27:17 thanks! 2021-11-30 05:29:50 Open source is action :) 2021-11-30 07:51:25 psykose: hey tried it but I think I'm doing something wrong 2021-11-30 07:51:32 I use: 2021-11-30 07:52:01 update-kernel /media/usb/ (after modifying mkinitfs.conf of course to include vfio) 2021-11-30 07:53:20 I tried to look for documentation on how to run mkinitfs directly but I'm not sure I'm doing it right (given diskless setup) 2021-11-30 07:54:10 `apk fix mkinitfs` will run it, but i'm not sure about diskless 2021-11-30 07:55:13 psykose: doesn't that just reinstall it? One needs to generate a new initramfs with it right? 2021-11-30 07:55:27 it does that on install 2021-11-30 07:55:33 it runs the trigger 2021-11-30 07:55:36 https://git.alpinelinux.org/aports/tree/main/mkinitfs/mkinitfs.trigger this specifically 2021-11-30 07:55:39 ah 2021-11-30 07:57:15 I will try remounting rw and see how it goes 2021-11-30 07:58:11 that doesn't work on diskless, but update-kernel runs mkinitfs for you 2021-11-30 07:58:38 psykose: I just remounted my diskless as rw, and did: mkinitfs -o /media/usb/boot/initramfs-lts 2021-11-30 07:58:45 ah, that's what you mean 2021-11-30 07:58:55 you can just run update-kernel, it does everything for you 2021-11-30 07:58:58 no errors, is everything contained in that file (initramfs-lts) or does it have other stuff? 2021-11-30 07:59:05 but i think that works too 2021-11-30 07:59:14 if you run mkinitfs -l it should show the vfio in there 2021-11-30 07:59:48 psykose: I see vfio now 2021-11-30 08:00:01 neat 2021-11-30 08:00:06 no idea if that actually fixes the issue 2021-11-30 08:01:13 psykose: am I right in assuming vfio is built-in now and so it will grab the gpu before drm/simpledrm gets a chance? 2021-11-30 08:01:26 no, the modules are just there 2021-11-30 08:01:36 which means the softdep you added should actually mean something, assuming it is respected 2021-11-30 08:04:29 softdep worked before (with 3.14) without needing to do any mkinitfs changes, I suspect the new simpledrm is to blame 2021-11-30 08:04:33 also you need `softdep simpledrm pre: vfio vfio_pci` 2021-11-30 08:04:43 ah, that I did not try 2021-11-30 08:09:56 brb 2021-11-30 08:42:43 asdflkj_sh: should be available on dl-cdn 3.15 2021-11-30 09:58:04 psykose: finally found the problem, it was kms! 2021-11-30 09:58:17 :) 2021-11-30 09:59:09 many thanks for the help. It wasn't simpledrm, kms grabs the gpu and vfio fails, then amdgpu sees an unintialized gpu and grabs it.. 2021-11-30 09:59:40 preventing kms (rather, removing it from mkinitfs) allowed the softdep pre stuff to do its work 2021-11-30 10:00:12 Cheers, have a good one 2021-11-30 10:00:41 removing it removes the firmware files for the gpu from the initfs which i suppose can be a way of getting that to work 2021-11-30 10:00:43 nice that it works :) 2021-11-30 10:01:15 it does indeed :) 2021-11-30 11:19:44 hey folks, it seems that after an update, my /etc/init.d/networking service file has disappeared and my server is left without a working network following a reboot -- could someone upload their /etc/init.d/networking service file somewhere so I can recreate it? thanks! 2021-11-30 11:20:16 `apk fix openrc` should put it back 2021-11-30 11:20:53 but sure here's the default from edge https://git.alpinelinux.org/aports/tree/main/openrc/networking.initd 2021-11-30 11:22:49 thanks, apk fix did the trick! 2021-11-30 11:45:57 curious how did you apk fix without a network ;/ 2021-11-30 11:47:03 you don't need init.d/networking for network, but it was probably relied on to start it automatically? 2021-11-30 18:13:24 Hi, I am interested in this OS but I am seeing some licensing discrepancies; is there a good place to take that up? I haven't gotten an answer here 2021-11-30 18:13:59 biovoid: what discrepancies? 2021-11-30 18:16:37 In this issue: https://gitlab.alpinelinux.org/alpine/aports/-/issues/7423 It is said "As per distribution policy, we only ship software licensed under OSI-approved licenses that conform to the open source definition: https://opensource.org/osd in the main, community and testing repositories" 2021-11-30 18:16:37 But I see `linux-firmware-*` in `main`, so that can't be true. 2021-11-30 18:17:09 I can't find anything in the way of a published "distribution policy" beyond this 2021-11-30 18:20:24 biovoid: I don't believe there is a published policy, its more like its passed by word of mouth :-) 2021-11-30 18:22:15 basically the people who decide the policy may be too busy to actually write down the policy. Hopefully that will improve in time with the recent creation of the Alpine TSC and Council 2021-11-30 18:23:08 biovoid: as for the linux firmware packages, I believe the issue around this has been highlighted previously 2021-11-30 18:23:51 Oh neat—I did not know about this recent structure; that is promising :) 2021-11-30 18:23:51 highlighted, it more than the issue I linked? I have been searching for more resources on it 2021-11-30 18:24:03 s/it/in/ 2021-11-30 18:24:03 biovoid[m] meant to say: highlighted, in more than the issue I linked? I have been searching for more resources on it 2021-11-30 18:24:43 biovoid: https://gitlab.alpinelinux.org/alpine/tsc/-/issues/6 2021-11-30 18:25:38 ope I was two clicks away... thank you for this 2021-11-30 18:27:24 biovold: Ariadne mentioned about firmware licenses in September 2021-11-30 18:28:03 I am reading some good discussion in issue #23 2021-11-30 18:28:18 er, not that one 2021-11-30 18:28:36 https://gitlab.alpinelinux.org/alpine/tsc/-/issues/23 2021-11-30 18:28:49 yes thank you :P 2021-11-30 18:46:09 I switched to sway and wanted to use rtc (jitsi), /usr/libexec/xdg-desktop-portal-wlr must run for that. it being in libexec suggested that it is automatically started via dbus or so. This does not happen. I now start it in sway-config with `exec /usr/libexec/xdg-desktop-portal-wlr`. Is it intended like that? 2021-11-30 20:28:35 did you start dbus 2021-11-30 21:11:33 lol 2021-11-30 21:12:11 how do you launch sway? 2021-11-30 21:12:49 it should be launched akin to this: dbus-run-session sway $@ 2021-11-30 21:13:03 use a slingshot 2021-11-30 21:13:07 :) 2021-11-30 21:13:26 well only if one need dbus shit 2021-11-30 21:13:29 I prefer trebuchet but each to their own 2021-11-30 21:13:51 helby: very true, but mostly it is due to some 3rd party needing some IPC 2021-11-30 21:14:10 in my instance pipewire... the SINGULAR runtime... for the heaviest of heavy IPC. 2021-11-30 21:14:21 Playing: Obscura - A Valediction-3150488592.mp3 2021-11-30 21:14:37 ^ excellent choice. 2021-11-30 21:14:38 yes, without it for example firefox is not able to open tab/win from other app, it complains about another instnace running 2021-11-30 21:15:08 thus launch sway via dbus-run-session 2021-11-30 21:15:23 its the simplest method. sure you could do some if [[ $ ... but why bother 2021-11-30 21:16:01 you should have some script or something else with all the wayland env being set so it can be used there 2021-11-30 21:16:05 Hello71: yes, both the system and the user session. 2021-11-30 21:16:17 dbus-launch --exit-with-session seatd-launch sway 2021-11-30 21:16:19 6wayland is too new for me 2021-11-30 21:16:24 4i just startx 2021-11-30 21:16:40 yes new to me too. 2021-11-30 21:16:54 then it's time to try 'river' 2021-11-30 21:17:03 7dunno what river is 2021-11-30 21:17:10 please do not abuse colors 2021-11-30 21:17:18 lighter wayland wm 2021-11-30 21:17:26 ok 2021-11-30 21:17:29 .. 2021-11-30 21:18:09 *_* 2021-11-30 21:21:58 freq: Lovechild of DWM and BSPWM, but a wlroots compositor 2021-11-30 21:22:11 oic 2021-11-30 21:22:11 i use dwm 2021-11-30 21:22:23 i have a kit for distros 2021-11-30 21:22:38 https://www.mercurial-scm.org/wiki/freq?action=AttachFile&do=get&target=301_Linux.tar.gz 2021-11-30 21:23:26 works on alpine, arch, debian, etc.. 2021-11-30 21:23:44 just set your console font if you get weird issues in stterm 2021-11-30 21:23:51 usually, people have this set by default 2021-11-30 21:31:14 I prefer runtime configuration over editing source code 2021-11-30 22:09:42 rivers looks interesting but still too early. labwc looks nice though 2021-11-30 22:32:49 Has anyone ever tried to run graalvm native-image on Alpine? 2021-11-30 22:35:27 Oh, wow. Someone was asking me about a segfault he was seeing when trying to use graalvm on Alpine, but this is not a normal Alpine system. It's an AdoptOpenJDK Docker image that somehow has glibc installed? D: 2021-11-30 22:48:11 oh dear :) 2021-11-30 22:48:40 https://ariadne.space/2021/08/26/there-is-no-such-thing-as-a-glibc-based-alpine-image/ 2021-11-30 22:51:11 You would be surprised what someone who is bored can cook up 2021-11-30 22:51:33 breaking bad sort of shows that 2021-11-30 22:52:22 this is like trying to cobble in py3 onto a RH6 system... 2021-11-30 22:52:34 sure you can "sort of" get it in, but DAMN! 2021-11-30 22:55:03 ACTION uploaded an image: (110KiB) < https://matrix.org/_matrix/media/r0/download/uc.neviani.fr/vDURcUfGlmhwLmdctIMWMOHr/Screenshot_20211130-235451_Bromite.png > 2021-11-30 22:55:14 Sorry! 2021-11-30 22:55:18 Look I run void linux ppc on an iBook g3 500. People do dumb things lol 2021-11-30 22:56:13 You can only put up so many “don’t do this” signs before you have to wash your hands of it. 2021-11-30 23:08:29 I can't some specific example of cloud-init with Alpine Linux after the official support brought from 3.13, any example / article you'd share ? 2021-11-30 23:11:48 (Just found this https://git.alpinelinux.org/aports/tree/community/cloud-init/README.Alpine) 2021-11-30 23:21:59 blt[m]: If you have any more questions, I believe minimal is our local cloud-init expert.