2022-03-01 06:21:55 ACTION who the fuck need gui at server ... . just go from here ... . https://cryptobug.wordpress.com/2020/10/04/ssh-tunneling-xming-with-putty-opening-gui-application-with-command-line/ , is from client with ssh forwarder .... . 2022-03-01 14:13:22 Hello, I've been meaning to install system mode with a LUKS encrypted partition yet for some reason setup-alpine defaults to installing GRUB, I thought syslinux was the default one since GRUB doesn't quite support encryption, is there a parameter I'm missing? 2022-03-01 14:32:23 bosco1: looks like `BOOTLOADER=syslinux setup-alpine` (or `BOOTLOADER=syslinux setup-disk`): https://git.alpinelinux.org/alpine-conf/tree/setup-disk.in#n628 2022-03-01 14:34:33 that's, uh, nonsense? 2022-03-01 14:35:07 grub "doesn't quite support encryption" in the sense that it doesn't support some luks modes, which is only useful for encrypted /boot which is of dubious value in the first place 2022-03-01 14:35:16 but by that metric, syslinux "doesn't support encryption" at all 2022-03-01 14:36:50 I was wrong, it does support it 2022-03-01 14:37:14 But setup-alpine doesn't make the grub configs to have encryption working by default 2022-03-01 14:38:30 it does: https://git.alpinelinux.org/alpine-conf/tree/setup-disk.in#n616 2022-03-01 14:39:12 it gets passed then to setup_grub 2022-03-01 14:39:16 well it possibly doesn't support encrypted /boot, but it also doesn't create encrypted /boot so there is no problem 2022-03-01 15:08:37 Turns out the problem was with grub and nvme drives 2022-03-01 18:48:00 What kind of shell is available with openrc-run? I'm trying to do for loops in the /etc/init.d/hostname script but I can't figure out if bash, sh, or something else is doing the execution. 2022-03-01 18:50:19 just write POSIX-compliant shell scripts, that will do 2022-03-01 18:50:56 sh 2022-03-01 18:50:59 better not to depend on what shell running your script 2022-03-01 18:51:02 ash on alpine 2022-03-01 18:53:10 openrc-run execs into /lib/rc/sh/openrc-run.sh , which is being executed with /bin/sh 2022-03-01 18:54:45 thanks 2022-03-01 18:55:04 I thought I was POSIX-compliant, but I guess not. 2022-03-01 18:57:33 Best approach is to make POSIX-compliant scripts, as aron stated 2022-03-01 18:59:58 Works for me. Too used to bash. :) gotta relearn what for loops look like 2022-03-01 19:00:11 shellcheck exists and is useful 2022-03-01 19:01:12 Also checkbashisms 2022-03-01 19:03:06 TIL. Those look super helpful. Thanks guys! 2022-03-01 19:25:03 only C-style for loops are bash-specific, and i don't know why you would need those 2022-03-01 20:13:19 does "use logger" in openrc work by default with syslogd on alpine? i tried that, but the service doesn't log anywhere, while i see stuff like cron and sshd using it properly 2022-03-01 20:14:54 "use is a soft dependency - if dns, logger or netmount is in this runlevel start it before, but we don't care if it's not in this runlevel." 2022-03-01 20:14:56 https://github.com/OpenRC/openrc/blob/master/service-script-guide.md#the-depend-function 2022-03-01 20:15:28 So it's only for ordering, it does not matter whether the services uses it or not 2022-03-01 20:15:51 (sorry if I misunderstood your question) 2022-03-01 20:18:05 ah, that answers it, thanks 2022-03-02 06:12:00 hi, I install the wget package 2022-03-02 06:12:12 when I run wget -h, I still get busybox wget 2022-03-02 06:12:16 any idea whats wrong? 2022-03-02 06:13:11 nvm, I just had to apk fix wget 2022-03-02 06:37:50 ACTION sent a code block: https://matrix.org/_matrix/media/r0/download/matrix.org/MIRsTMEGRlPbqQbNVkCTcUEf 2022-03-02 06:38:23 Working except for when I build with blake2 (libb2-dev), using any of the blake2 digests will crash GTKHash with an Illegal Instruction console message 2022-03-02 06:38:44 Is that a GTKHash issue, a libb2-dev issue, a platform issue (do I not have the kernel module?)? How can I tell? 2022-03-02 06:39:13 If I keep it as above, there are still two blake2 digests included in GTKHash that work properly. They must be provided by another library 2022-03-02 07:06:14 morning, i just installed the virtual image in my virtualbox vm and installed gnome, then noticed the system faces huge lags from time to time 2022-03-02 07:07:00 basically the whole system is unresponsive in that case and also receives multiple keyboard inputs even when i hit a key only once 2022-03-02 07:07:36 that happens more often when the system is under stress for example while compiling stuff with multiple threads or even when opening a bigger website in firefox 2022-03-02 07:08:04 any ideas what could cause this? 2022-03-02 07:08:07 its really annoing 2022-03-02 07:44:10 now even the background and top bar went black lol 2022-03-02 08:15:17 and now everything is black 2022-03-02 08:15:48 seems like i either missed a step or gnome is not usable on alpine linux in a vm 2022-03-02 09:23:35 Hi, I'd need some help with a productive server running alpine and some strange behaviour, maybe somebody has seen something similar.. 2022-03-02 09:24:13 If I call netstat it will sometimes only show one row of output or so and display "Illegal instrucion" 2022-03-02 09:24:38 As I check dmesg it says: "[ 1470.190608] traps: netstat[18085] trap invalid opcode ip:55e6810ce797 sp:7fff53a1aa20 error:0 in netstat[55e6810c9000+b000]" for each time I call netstat 2022-03-02 09:25:34 At the same time not only netstat seems to be affected, but also some network-services as for example a gatling http that is running there, that won't accept connections 2022-03-02 09:26:26 As i reboot this seems to vanish just to appear again shortly after 2022-03-02 09:27:19 user12344566: Can you run it in gdb to see what's the illegal instruction? 2022-03-02 09:28:02 Is it busybox netstat? 2022-03-02 09:28:13 yes 2022-03-02 09:28:29 What arch? 2022-03-02 09:28:34 amd64 2022-03-02 09:28:48 @ktprograms I'll check 2022-03-02 09:31:12 I might want to add that the system is a KVM VM 2022-03-02 09:34:03 I'm not familiar with gdb, can I do something like gdb netstat or is there something to set up first? 2022-03-02 09:34:20 user12344566: That should work 2022-03-02 09:36:25 right now the error does not show, I'll wait until it's there again an check with gdb. Its super sporadic and very weird as it vanishes on reboot and than seems to appear more frequently as the uptime gets longer 2022-03-02 09:36:41 *then 2022-03-02 09:42:50 its now its happening again, gdb drops me to some kind of "(gdb)" prompt, anything I should try here? 2022-03-02 09:43:06 type: "r" 2022-03-02 09:43:24 ^ user12344566 2022-03-02 09:43:46 Program received signal SIGILL, Illegal instruction. 0x000055555555f797 in ?? () 2022-03-02 09:44:24 user12344566: x/i $pc 2022-03-02 09:44:43 => 0x55555555f797: ud2 2022-03-02 09:45:01 ooooh, I know what is that, haha 2022-03-02 09:45:06 happened to me 2022-03-02 09:45:13 Nice 2022-03-02 09:45:19 this is the outcome of a new gcc thingy 2022-03-02 09:45:42 UB? 2022-03-02 09:45:57 I tracked this down to a "possible buffer overflow" bug in the software 2022-03-02 09:46:16 Ouch 2022-03-02 09:46:25 gcc will generate ud2 if he things there will be a buffer overflow 2022-03-02 09:46:33 i.e. strncpy() 2022-03-02 09:46:45 (or equivalent on different arch) 2022-03-02 09:47:13 s/things/thinks/ 2022-03-02 09:47:14 aron meant to say: gcc will generate ud2 if he thinks there will be a buffer overflow 2022-03-02 09:47:46 Maybe run with valgrind as well 2022-03-02 09:48:00 I'd say it is pointless at this point 2022-03-02 09:48:05 Ook 2022-03-02 09:48:07 Ok 2022-03-02 09:48:09 because the generated code will jump to ud2 2022-03-02 09:48:36 do, the buffer overflow will not be triggered 2022-03-02 09:48:39 **se 2022-03-02 09:48:41 **so 2022-03-02 09:48:41 And it does not do out of bounds checking 2022-03-02 09:48:49 (sry bunch of typo today) 2022-03-02 09:49:35 you might want to know where it comes from 2022-03-02 09:50:02 user12344566: install musl-dbg, and rerun the program 2022-03-02 09:50:13 after it crashes again, type: "bt" 2022-03-02 09:50:58 this will probably tell something what we need to look for in netstat's source 2022-03-02 09:51:08 Thanks for helping, I will let you know 2022-03-02 09:52:40 Hi All! I really love Alpine linux, and I've been trying to use it in my laptop. However, my touchpad is not working 😦 I have had a look and found there is this merge request: https://gitlab.alpinelinux.org/alpine/abuild/-/issues/10043 which has been merged already which is precisely for my touchpad (Hantick 5288). The device does appear in /sys/bus/i2c/devices, but when I try to bind it to the driv 2022-03-02 09:52:40 er doing echo "i2c-HTIX:5288:00" > /sys/bus/i2c/drivers/i2c-hdi-acpi/bind it is not working, saying resource temporarily unavailable. Maybe I'm doing the wrong thing, any hints, pretty please? 2022-03-02 09:52:57 You may need bb symbols as well, but they are not built atm 2022-03-02 09:53:17 It is working perfectly fine for Xubuntu and Fedora XFCE, so it is definitely something I have to configure, somewhere. I just don't know how to query Xubuntu to see what it's doing to set up my touchpad correctly 2022-03-02 09:53:22 I can bind and unbind the driver in Xubuntu without issues 2022-03-02 09:54:26 ikke: yes, that would help a lot, but maybe it is enough what libc call it does 2022-03-02 09:55:23 alpinenoobie: please tell what kind of laptop you have 2022-03-02 09:56:14 It is a Geobook GE013, with an intel N4000 processor 2022-03-02 09:56:25 aron: I get "No stack" as I type bt into gdb 2022-03-02 09:56:44 It's a cheap chinese laptop, also known by the brand Fivetech 2022-03-02 09:56:55 user12344566: "r" -> wait for crash -> then "bt" 2022-03-02 09:57:26 that make sense thanks 2022-03-02 09:57:28 #0 0x000055555555f797 in ?? () #1 0x000055555555c950 in ?? () #2 0x000055555555ce03 in ?? () #3 0x000055555555ad10 in ?? () #4 0x00007ffff7f83a03 in libc_start_main_stage2 (main=0x55555555a5c8, argc=1, argv=0x7fffffffec58) at src/env/__libc_start_main.c:94 #5 0x000055555555b000 in ?? () #6 0x0000000000000001 in ?? () #7 0x00007fffffffee70 in ?? () #8 0x0000000000000000 in ?? () 2022-03-02 09:58:20 ouch 2022-03-02 09:58:37 are you sure musl-dbg is installed? 2022-03-02 09:59:00 yup -> apk add musl-dbg OK: 460 MiB in 283 packages 2022-03-02 09:59:45 :( 2022-03-02 10:00:18 then busybox did this 2022-03-02 10:00:52 this will be much harder now 2022-03-02 10:01:15 especially with missing bb syms 2022-03-02 10:02:14 can you build bb on your system? 2022-03-02 10:02:25 we need those pesky symbols :) 2022-03-02 10:02:40 yes I will give it a try 2022-03-02 10:02:54 thanks for your work your're potentially saving my ass 2022-03-02 10:03:40 Make sure SKIP_STRIP=y is set in the environment 2022-03-02 10:04:22 alpinenoobie: what "uname -a" prints? 2022-03-02 10:04:47 ikke: are these any option to build with "-g"? 2022-03-02 10:05:02 build bb, ofc 2022-03-02 10:06:04 if you use abuild, add $pkgname-dbg as first subpackage 2022-03-02 10:08:00 where to get source from? What program or part of program is bb exactly? I'm sorry if this sould be known 2022-03-02 10:08:27 busybox is its own project 2022-03-02 10:08:37 aaron: Linux localhost 5.15.24-2-lts #3-Alpine SMP Mon, 21 Feb 2022 11:26:44 +0000 x86_64 Linux 2022-03-02 10:08:43 You could clone aports 2022-03-02 10:08:43 ohh bb is busybox I see, sorry that was dumb 2022-03-02 10:09:02 No worry 2022-03-02 10:11:07 I was using the v3.15 kernel, but just to discard if it was the kernel, I've upgraded to the edge kernel 2022-03-02 10:11:20 I mean alpine v3.15 kernel 2022-03-02 10:12:54 maybe if I could see how Xubuntu sets my touchpad, it would be much easier, but I'm too noobie to find that out. If you have any idea how I can do that, I can try to extract the information from Xubuntu 2022-03-02 10:24:34 alpinenoobie: if I understand correctly, MR failed 2022-03-02 10:24:59 therefore, you need to build your own kernel with the changes mentioned here: https://gitlab.alpinelinux.org/alpine/aports/-/merge_requests/25016 2022-03-02 10:25:12 but correct me, if I'm wrong 2022-03-02 10:26:26 I thought it was merged already to master? Here is the closed merge request: https://gitlab.alpinelinux.org/alpine/aports/-/commit/5f912a2ca510a89824fdbdc00cb2813cd3c08b93 2022-03-02 10:26:39 I'm using the aports git and set export SKIP_STRIP=y in abuild.conf, so far so good? 2022-03-02 10:27:36 aaron: so the original merge request was not accepted but Natael created his own 2022-03-02 10:27:55 oh, okay, thanks 2022-03-02 10:28:55 alpinenoobie: do you see the "broxton" thing in lsmod? I would expect that to be there 2022-03-02 10:30:11 Now, that might be misleading.. because I have a N4000 processor, I don't think it is a broxton 2022-03-02 10:30:29 So maybe that merge request fixes it for a certain laptop and not mine 2022-03-02 10:30:52 ikke: Is SKIP_STRIP still needed if adding $pkgname-dbg to subpackages in user12344566 's case? 2022-03-02 10:31:37 I'll have a look in lsmod 2022-03-02 10:32:33 so it was not there, but I have done a modprobe pinctrl-broxton before and it hasn't worked. I just did it now and still the same 2022-03-02 10:32:59 ktprograms: yes, otherwise the Makefile will still strip it 2022-03-02 10:33:20 I see. 2022-03-02 10:33:42 user12344566: In the APKBUILD for busybox, add $pkgname-dbg to the subpackages variable. 2022-03-02 10:33:43 okay, I found it: 2022-03-02 10:33:43 # CONFIG_PINCTRL_GEMINILAKE is not set 2022-03-02 10:33:49 ^ alpinenoobie 2022-03-02 10:33:54 N4000 is geminilake 2022-03-02 10:34:06 aha, so I would need to compile my own kernel, I guess? :( 2022-03-02 10:34:27 yes 2022-03-02 10:35:01 (N4000 is a decent cpu, btw) 2022-03-02 10:35:06 why aren't all of these flags set? What determines if this should go in master? 2022-03-02 10:35:21 dunno 2022-03-02 10:35:22 alpinenoobie: mostly user requests 2022-03-02 10:35:33 So please open an issue on gitlab 2022-03-02 10:36:11 (I always build my own kernel for arm, I need to change a lot of things) 2022-03-02 10:36:30 Yeah, it is a cheap laptop, but the cool thing is that it consumes only 6W and it looks kind of premium: https://business.currys.co.uk/catalogue/computing/laptops/windows-laptop/geo-book3-13-3-intel-celeron-laptop-32-gb-emmc-silver/N239150W 2022-03-02 10:36:52 ktprograms: thanks, currently making the APKBUILD 2022-03-02 10:36:59 anyways, I'll try to see how to open the merge request 2022-03-02 10:37:07 alpinenoobie: yep, low power consumption is good :) 2022-03-02 10:37:10 thank you all for withstanding my noobness :D 2022-03-02 10:37:25 no prob, yw 2022-03-02 10:38:00 how often does the kernel gets compiled for edge? 2022-03-02 10:38:19 so if my MR is accepted, I guess I wouldn't have to wait too long till it appears in edge 2022-03-02 10:38:47 alpinenoobie: As soon as it's merged it'll be added to the list of things for the builders to build 2022-03-02 10:39:08 yeiiii! I'll get started ASAP, thank you all 2022-03-02 10:39:48 If I understand it correctly, enabling new modules/features should be done in the edge kernel first (linux-edge aport) 2022-03-02 10:40:10 I have to learn how to compile the kernel :D it shouldn't be so complicated but I guess it is time consuming 2022-03-02 10:40:31 alpinenoobie: if you use aports it's quite simple 2022-03-02 10:40:44 I do this all the time 2022-03-02 10:40:54 but I won't argue with the time consuming part :P 2022-03-02 10:41:19 I meant compilation time, how long does it take to compile for you? 2022-03-02 10:41:22 get aports from git, change your config, and "abuild" 2022-03-02 10:41:37 yes, compilation time, ofc 2022-03-02 10:41:43 depends 2022-03-02 10:41:53 if you enable a ton of modules it took longer 2022-03-02 10:41:54 alpinenoobie: for the kernel, issues are more effective 2022-03-02 10:41:54 :P 2022-03-02 10:42:20 aron: Do you think the issue Saijin_Naib was having with GTKHash/libb2 illegal instructions is related to this gcc thing? 2022-03-02 10:42:44 I actually have the very same laptop with a N5000 processor, fortunately it is Gemini Lake as well :D 2022-03-02 10:43:01 ktprograms: I know very little about it, but it is likely 2022-03-02 10:43:26 Is there any compile time flag to set to see these possible buffer overflows? 2022-03-02 10:43:51 alpinenoobie: cool :) 2022-03-02 10:44:06 ktprograms: it is the default now 2022-03-02 10:44:21 displayed in warnings 2022-03-02 10:44:34 and appropriate(?) code is generated 2022-03-02 10:45:04 I mean, your program might function with that overflow, but it is better to crash at the point where it occured 2022-03-02 10:48:31 aron: I don't see anything related to it in the busybox build log (if I'm not mistaken that's where the illegal instruction seems to be in user12344566 's case, right?) https://build.alpinelinux.org/buildlogs/build-edge-x86_64/main/busybox/busybox-1.35.0-r4.log 2022-03-02 10:53:09 hmm, so it is not at an obvious location 2022-03-02 10:54:05 seems like busybox.net is down atm and abuild is trying to fetch https://busybox.net/downloads/busybox-1.35.0.tar.bz2 I'll have to try again later 2022-03-02 10:54:18 sorry, no warning is emitted for all of these, only for the obvious ones 2022-03-02 10:54:28 let me check stg 2022-03-02 10:55:53 > Option -Wstringop-overflow=2 is enabled by default. 2022-03-02 10:56:06 you can have up to 4 2022-03-02 10:57:28 user12344566: works for me, but it took some time to connect 2022-03-02 10:59:14 Now it worked for me too, currently building 2022-03-02 11:00:03 abuild finished successfully 2022-03-02 11:05:51 should I build only busybox or everything? 2022-03-02 11:14:59 I'm sorry, have been logged out, anything new since I've build busybox? 2022-03-02 11:17:28 user12344566: Run "doas apk add --repository /home/USER/packages/main busybox" (replace USER with your username) 2022-03-02 11:20:31 done 2022-03-02 11:20:59 Now repeat the gdb steps you did last time 2022-03-02 11:21:32 Program received signal SIGILL, Illegal instruction. 0x000055555555f797 in ?? () (gdb) bt #0 0x000055555555f797 in ?? () #1 0x000055555555c950 in ?? () #2 0x000055555555ce03 in ?? () #3 0x000055555555ad10 in ?? () #4 0x00007ffff7f83a03 in libc_start_main_stage2 (main=0x55555555a5c8, argc=1, argv=0x7fffffffec18) at src/env/__libc_start_main.c:94 #5 0x000055555555b000 in ?? () #6 0x0000000000000001 in ?? () #7 0x00007fffffffee43 2022-03-02 11:24:03 Oops, also do that apk command but replace busybox with busybox-dbg at the end 2022-03-02 11:24:09 then gdb again 2022-03-02 11:28:31 add --repository this just adds to /etc/repositories, right? Somehow it just terminates and does not add anything. If I add it manually to /etc/repositories and do apk upgrade it says "Ignoring /home/test/packages/main busybox-dbg: No such file or directory" but there are .apk and APKINDEX.tar.gz 2022-03-02 11:28:52 user12344566: no, it's a one-off thing 2022-03-02 11:29:02 user12344566: it will just make it use the repo for that invocation 2022-03-02 11:29:24 user12344566: the package name should not be in /etc/apk/repositories 2022-03-02 11:31:19 apk add --repository /home/test/packages/main busybox-dbg works but gdb output seems the same 2022-03-02 11:31:39 ikke: i see, thanks 2022-03-02 11:32:05 user12344566: hwo did you build busybox? 2022-03-02 11:32:11 With the APKBUILD? 2022-03-02 11:32:20 s/With/From 2022-03-02 11:32:20 ikke meant to say: From the APKBUILD? 2022-03-02 11:34:50 with abuild -r in the dir with the modified APKBUILD 2022-03-02 11:37:39 busybox-dbg.apk is present in the packages dir 2022-03-02 11:39:08 Did you also add `export SKIP_STRIP=y` somewhere? 2022-03-02 11:43:05 In abuild.conf but it seems like the user uses a different file 2022-03-02 11:43:19 I'll try again 2022-03-02 11:50:44 now added to ~/.abuild/abuild.conf instead of /etc/abuild.conf 2022-03-02 11:51:46 should not matter 2022-03-02 11:52:37 still does not produce any other output from gdb 2022-03-02 11:53:25 used 'apk add --repository /home/test/packages/main busybox-dbg' in between of course 2022-03-02 12:00:39 should I try downgrading the system from edge to v3.14? 2022-03-02 12:04:13 didnt do nothing, upgrading again 2022-03-02 12:07:55 user12344566: did you installed both busybox and busybox-dbg? 2022-03-02 12:09:26 by installing you mean apk add --repository /home/test/packages/main busybox-dbg? Or apk add --upgrade busybox? 2022-03-02 12:09:51 apk add --repository /home/test/packages/main busybox busybox-dbg 2022-03-02 12:10:41 yes 2022-03-02 12:10:45 o/ 2022-03-02 12:11:34 I wrote a minimal init script using supervise-daemon, I was surpised to see that the pidfile it generates it's the pid of supervise-daemon itself rather than the child process Oo 2022-03-02 12:11:40 did it overwrite anything? 2022-03-02 12:12:33 aron: I had no output, only "OK: 533 MiB in 298 packages" 2022-03-02 12:12:37 I mean: you have the same version number that is installed originally? 2022-03-02 12:12:53 I'll check 2022-03-02 12:13:19 I saw overwriting packages with the same version number is not that trivial 2022-03-02 12:13:39 (because normally they are the same, so no actions needed) 2022-03-02 12:14:04 could I do apk add .apk to overwrite? 2022-03-02 12:15:07 git it a try 2022-03-02 12:15:23 s/git/give/ 2022-03-02 12:15:23 aron meant to say: give it a try 2022-03-02 12:17:17 this has done something but still same output from gdb 2022-03-02 12:18:06 what grep busybox /etc/apk/world yields? 2022-03-02 12:19:30 busybox> nm /bin/busybox | head -n 1 2022-03-02 12:22:04 (if you built it with -g, and successfully installed, this must emit symbols) 2022-03-02 12:22:33 what tool is nm? 2022-03-02 12:22:56 built with -g does mean abuild -g? 2022-03-02 12:23:08 nm - list symbols from object files 2022-03-02 12:23:42 nm: /bin/busybox: no symbols 2022-03-02 12:23:55 no, if gcc got -g flag, but ikke said if dbg pkg is built (first subpackge!), then everything should be fine 2022-03-02 12:24:29 yes dbg is the first sub package in the APKBUILD 2022-03-02 12:24:48 like this 2022-03-02 12:24:49 subpackages="$pkgname-dbg $pkgname-static $pkgname-doc $pkgname-suid $pkgname-extras ssl_client $pkgname-ifupdown" 2022-03-02 12:25:03 this should be fine then 2022-03-02 12:25:14 in this case nm should emit something useful 2022-03-02 12:29:28 or check the date of /bin/busybox :) 2022-03-02 12:29:35 ls -l /bin/busybox 2022-03-02 12:30:49 date seems to match the 'apk add --upgrade' command but no symbols 2022-03-02 12:30:55 i think busybox strips itself 2022-03-02 12:32:02 Hello71: SKIP_STRIP=y should prevent that 2022-03-02 12:32:09 it worked for me when I debugged bb 2022-03-02 12:32:32 hm 2022-03-02 12:33:23 https://github.com/mirror/busybox/blob/master/Makefile#L723 2022-03-02 12:33:31 also nm doesn't search for split debug 2022-03-02 12:33:34 user12344566: build again with explicitly: export SKIP_STRIP=y before abuild 2022-03-02 12:33:44 ikke: mmhmm 2022-03-02 12:34:01 already tried export SKIP_STRIP=y ; abuild -r 2022-03-02 12:34:32 When you start gdb, do you see it finding the busbox symbols? 2022-03-02 12:34:36 It should mention that at the beginning 2022-03-02 12:34:47 gdb busybox netstat 2022-03-02 12:34:49 or 2022-03-02 12:34:53 gdb --args busybox netstat 2022-03-02 12:36:01 Type "apropos word" to search for commands related to "word"... Reading symbols from busybox... Reading symbols from /usr/lib/debug//bin/busybox.debug... 2022-03-02 12:36:07 if i run the second example 2022-03-02 12:36:09 first one fails 2022-03-02 12:36:18 yes 2022-03-02 12:36:24 but now netstat runs fine again 2022-03-02 12:36:25 ok, so it does find the busybox symbols 2022-03-02 12:36:36 in gdb i mean 2022-03-02 12:37:52 If I run gdb netstat without your example netstat won't run (like before) 2022-03-02 12:38:42 what did you run before? 2022-03-02 12:38:55 gdb netstat 2022-03-02 12:39:00 ok, then do that 2022-03-02 12:39:08 and check if it finds the symbols 2022-03-02 12:39:25 nope still same as before 2022-03-02 12:39:31 #0 0x000055555555f77d in ?? () #1 0x000055555555cad0 in ?? () #2 0x000055555555cf5b in ?? () #3 0x000055555555ad28 in ?? () #4 0x00007ffff7f83a03 in libc_start_main_stage2 (main=0x55555555a5e0, argc=1, argv=0x7fffffffec08) at src/env/__libc_start_main.c:94 #5 0x000055555555b018 in ?? () #6 0x0000000000000001 in ?? () #7 0x00007fffffffee3b in ?? () #8 0x0000000000000000 in ?? () 2022-03-02 12:39:46 but at load 2022-03-02 12:39:57 do you see it mentioning loading busybox.debug? 2022-03-02 12:40:37 only "Reading symbols from netstat... (No debugging symbols found in netstat)" if I run only gdb netstat 2022-03-02 12:41:11 what does apk info -W (which netstat) return? 2022-03-02 12:42:02 ikke: i think you dropped this: $ 2022-03-02 12:42:08 yes 2022-03-02 12:42:12 what does apk info -W $(which netstat) return? 2022-03-02 12:42:29 which -> /bin/netstat, apk info -W netstat -> ERROR: netstat: Could not find owner package 2022-03-02 12:42:36 ah sorry 2022-03-02 12:43:07 "/bin/netstat is owned by net-tools-1.60_git20140218-r2" 2022-03-02 12:43:15 ok, so _not_ busybox netstat 2022-03-02 12:43:51 so you'd need to debug net-utils, not busybox 2022-03-02 12:44:10 s/net-utils/net-tools/ 2022-03-02 12:44:10 ikke meant to say: so you'd need to debug net-tools, not busybox 2022-03-02 12:44:23 And debug symbols are available 2022-03-02 12:44:44 user12344566: `apk add net-tools-dbg` 2022-03-02 12:45:37 done 2022-03-02 12:45:55 sorry I really thought this could only be busybox netstat 2022-03-02 12:46:24 memcpy (__n=, __os=, __od=) at /usr/include/fortify/string.h:50 warning: Source file is more recent than executable. 50 __builtin_trap(); 2022-03-02 12:46:37 (gdb) bt #0 memcpy (__n=, __os=, __od=) at /usr/include/fortify/string.h:50 #1 INET6_getsock (sap=, bufp=) at inet6.c:166 #2 INET6_input (type=, bufp=0x7fffffffc9ea "::ffff:212.18.192.131", sap=) at inet6.c:174 #3 0x000055555555cad0 in tcp_do_one (lnr=lnr@entry=6, line=line@entry=0x7fffffffcba8 " 5: ", '0' user12344566: can you paste that somewher? https://tpaste.us for example 2022-03-02 12:48:42 https://tpaste.us/wYZv 2022-03-02 13:00:38 user12344566: how often does it happen for you? 2022-03-02 13:05:30 user12344566: can you show the output of "i r" too? 2022-03-02 13:06:20 After i reboot it does not happen for about 1 minute and after that sometimes pretty consitent and sometimes rather sporadic 2022-03-02 13:06:39 aron: sure on sec 2022-03-02 13:07:09 s/on/one/ 2022-03-02 13:07:10 user12344566 meant to say: arone: sure on sec 2022-03-02 13:07:26 ^^ 2022-03-02 13:07:29 hehe 2022-03-02 13:07:33 top 10 regex fails 2022-03-02 13:07:47 actually there's been more funny ones before 2022-03-02 13:08:07 wish I could see them :) 2022-03-02 13:08:41 aron: https://irclogs.alpinelinux.org 2022-03-02 13:09:18 https://tpaste.us/8boK 2022-03-02 13:10:10 ktprograms: that's a huge WOT :D 2022-03-02 13:10:45 aron: yes, unfortunately. I can't remember any specific cases. 2022-03-02 13:22:47 user12344566: please upload the output of x/120i INET6_getsock (or until that function ends) 2022-03-02 13:23:03 I don't see in the source code what is going on 2022-03-02 13:24:18 "Argument required (starting display address)." 2022-03-02 13:24:47 "INET6_getsock" is not good enough? 2022-03-02 13:25:20 0x55555555f702 then 2022-03-02 13:33:03 why don't you use disas 2022-03-02 13:35:28 whatever tool you use, just get the asm instructions ;) 2022-03-02 13:36:57 aron: I think Hello71 meant disas in gdb 2022-03-02 13:56:17 I know, but still, the result is the only thing I'm interested (personally I use x/Ni for small windows on very slow terminals) 2022-03-02 13:56:49 (and I stick to it, less tools to remember :P) 2022-03-02 14:02:09 aron: I see, makes sense 2022-03-02 14:33:48 https://tpaste.us/rndD here you go, sry I was occupied by something else 2022-03-02 14:35:04 no prob 2022-03-02 14:43:14 O.M.F.G. 2022-03-02 14:43:36 look at this 2022-03-02 14:43:37 https://gitweb.gentoo.org/proj/net-tools.git/tree/lib/inet6.c#n166 2022-03-02 14:44:33 (it is interesting, tho, your disassembly points into net-tools, not busybox) 2022-03-02 14:44:59 aron: isn't it net-tools? 2022-03-02 14:45:15 aron: they thought it was busybox, but we found out it's not 2022-03-02 14:46:02 key points from the disassembly: 2022-03-02 14:46:06 : cmp %rax,%r12 2022-03-02 14:46:10 : jb 0x55555555f77d 2022-03-02 14:46:13 : ud2 2022-03-02 14:47:09 and what it had in C: 2022-03-02 14:47:14 if (p != bufp) 2022-03-02 14:47:18 memcpy(bufp, p, strlen(p)+1); 2022-03-02 14:47:42 the memcpy is inlined, btw 2022-03-02 14:48:53 if I understand correctly, fix_v4_address will return an address which will be >=bufp 2022-03-02 14:49:16 this is 'p' 2022-03-02 14:49:37 and it will be part of 'bufp' 2022-03-02 14:49:48 however: memcpy(): The memory areas must not overlap. 2022-03-02 14:49:55 so yes, UB here 2022-03-02 14:50:07 for that, ud2 is appropriate 2022-03-02 14:50:38 and yes, this is the new gcc check-thingy what I was told at the beginning ;) 2022-03-02 14:50:54 (we've had it since 2013) 2022-03-02 14:51:17 was it enabled by default all the time? 2022-03-02 14:51:22 ys 2022-03-02 14:51:24 yes* 2022-03-02 14:51:40 anyway, the source is buggy AF, fix it 2022-03-02 14:52:46 i'm sure that, given a report to the appropriate upstream, they will fix it 2022-03-02 14:53:03 i have no personal desire to rewrite netstat at this time 2022-03-02 14:53:22 me neither :) 2022-03-02 14:53:54 there is iproute-ss already 2022-03-02 14:54:07 iproute2-ss** 2022-03-02 14:58:28 (well, the question was why this thing crashes, and there is the answer, anything else is up to you :P) 2022-03-02 14:58:42 why don't /you/ fix it :p 2022-03-02 14:59:14 actually it's an easy fix: s/memcpy/memmove/ 2022-03-02 14:59:18 yes 2022-03-02 15:20:29 anyways I thank you guys and gals for your work, I'll just use busybox netstat from now on and am glad that there's nothing fundamentaly wrong with my system 2022-03-02 15:26:28 So will anyone follow up? 2022-03-02 15:26:46 at least send an bug report upstream? 2022-03-02 15:28:28 just join #gentoo-dev on libera and demand they fix their shit obviously 2022-03-02 15:28:42 just kidding... don't actually do that 2022-03-02 15:28:53 i'll send them a patch later today 2022-03-02 15:28:57 i have a meeting in 2 minutes :) 2022-03-02 15:31:16 i would send one but i did not find the issue so it would be stolen military valor 2022-03-02 15:34:20 send it if you like, I don't mind 2022-03-02 15:38:18 (it was a fun project :P) 2022-03-02 21:35:31 I'm trying to use the apk binary on a not-alpine distro, but nearly every command gives me the error 'unable to read database state: no such file or directory' 2022-03-02 21:35:40 I'm guessing that it 2022-03-02 21:35:57 ...is looking for the APKINDEX.tar.gz file 2022-03-02 21:36:08 you can use "apk add --initdb" 2022-03-02 21:36:15 thanks! 2022-03-02 21:37:10 also, iirc it's not about apkindex, but /lib/apk/db 2022-03-02 21:38:24 gotcha. now it finds the database it created, but the db is empty...how do I populate it? 2022-03-02 21:38:42 when I did apk add nano, it said 'no such package' 2022-03-02 21:39:39 you need a /etc/apk/repositories 2022-03-02 21:39:49 or to add a literal `apk add something.apk` you downloaded 2022-03-02 21:40:05 if the distro is glibc most things will not be able to be added as they depend on musl 2022-03-02 21:40:39 apk add -X https://dl-cdn.alpinelinux.org/alpine/edge/main --initdb .. 2022-03-02 21:40:53 yep, that populates repositories for you 2022-03-02 21:41:48 is that the literal command I should run? It gave me an error about untrusted repositories 2022-03-02 21:42:04 untrusted signature, my bad 2022-03-02 21:42:15 does the initdb create the keys too 2022-03-02 21:42:15 hm 2022-03-02 21:42:28 also, thanks! that looks like what I need. 2022-03-02 21:42:53 the reason it's untrusted is because /etc/apk/keys is missing the signing keys, that's all 2022-03-02 21:42:53 If I have a local copy of the apkindex, can I force that to be read? 2022-03-02 21:43:08 psykose: they are pulled in if you install alpine-base 2022-03-02 21:43:24 but the first time you need --allow-untrusted 2022-03-02 21:43:29 ah, makes sense 2022-03-02 21:43:47 QDX45: to use a local apkindex just make sure it's at a path in /etc/apk/repositories 2022-03-02 21:43:49 you can also explicitly install alpine-keys if you want 2022-03-02 21:43:57 i.e. /home/you/packages/main/APKINDEX.tar.gz 2022-03-02 21:44:03 along with the .apks you have in that folder 2022-03-02 21:44:12 `apk index` can generate an index from .apk's too 2022-03-02 21:44:16 and abuild-sign can sign it 2022-03-02 21:44:48 and `SUDO=doas abuild-keygen -ai` can generate the key before you do that, drop the SUDO if you have sudo :) 2022-03-02 21:45:02 all I'm trying to do is parse dependencies and download packages 2022-03-02 21:45:03 but you also need to add abuild for those 2022-03-02 21:45:28 sure, you just need to `apk add --allow-untrusted alpine-keys` or alpine-base 2022-03-02 21:45:33 I think that works 2022-03-02 21:45:36 then fetch and regular stuff should work 2022-03-02 21:45:50 Thanks guys, that really helped 2022-03-02 21:45:55 I think this works now 2022-03-02 23:06:22 I pulled a copy of apk.static from https://gitlab.alpinelinux.org/api/v4/projects/5/packages/generic//v2.12.9/x86_64/apk.static, but some of the commands return ssl certificate verification errors 2022-03-02 23:06:39 I don't have that problem with the apk from my os repo 2022-03-02 23:07:12 It would be nice to use the more generic static apk, but I'm not sure what's up here 2022-03-02 23:07:33 do you have /etc/ssl1.1 2022-03-02 23:08:08 not that specific file, no 2022-03-02 23:08:15 symlink it to /etc/ssl or make a copy 2022-03-02 23:08:32 ok 2022-03-02 23:08:48 we were doing a failed migration to openssl3, and in that process openssl1.1 was using the openssl3 dir for some stuff and failed 2022-03-02 23:09:01 so made openssl1.1 use /etc/ssl1.1 instead of /etc/ssl, unlike anyone else 2022-03-02 23:09:11 and it doesn't actually exist, so 2022-03-02 23:09:21 alternatively you can place some specific file in /etc/apk i am forgetting 2022-03-02 23:10:01 /etc/apk/ca.pem 2022-03-02 23:10:10 from the system ca cert thing, the main one or something 2022-03-02 23:10:13 but either works 2022-03-02 23:10:54 the statically linked openssl into apk.static is the openssl-static from alpine repos, and that one is build to look at ssl1.1 instead of elsewhere, which is why it fails 2022-03-02 23:10:55 sounds good. the synlink got rid of the error 2022-03-02 23:11:11 yeah 2022-03-02 23:11:16 to be clear we should also revert that 2022-03-02 23:11:27 i haven't gotten around to it 2022-03-02 23:11:31 lol 2022-03-02 23:11:42 but it would break openssl3/1.1 mixing in some cases 2022-03-02 23:11:46 and we do ship openssl3 as well 2022-03-02 23:11:56 i don't know what a better solution would be for apk.static that others use 2022-03-02 23:12:27 well I'm scripting this so it's easy to do the workaround 2022-03-02 23:12:33 yeah 2022-03-02 23:12:33 I'm more worried that I broke my system doing experiments with apk package installs 2022-03-02 23:12:39 maybe 2022-03-02 23:12:40 :p 2022-03-02 23:12:48 :) 2022-03-02 23:12:56 later. gotta reboot & pray that helps 2022-03-03 07:39:23 after I searched where to report the found crash yesterday, I found this: https://sourceforge.net/p/net-tools/code/ci/ebe295a366086c03e5bbb843e38d3b3b774a88a4/ 2022-03-03 07:39:43 so they actually fixed, but never released it 2022-03-03 07:39:45 bummer 2022-03-03 15:10:40 there is a 2.1.0 2022-03-03 15:10:44 2.10* 2022-03-03 15:11:14 it contains the fix 2022-03-03 15:11:19 needs some config updates 2022-03-03 15:20:10 !31611 2022-03-03 18:00:16 I can't get dbus to work. I followed this https://wiki.alpinelinux.org/wiki/PipeWire#D-Bus but pipewire does not work and mako (notification deamon) does not automatically start when an application sends a notification. Any hints? 2022-03-03 18:02:53 what wm are you using 2022-03-03 18:04:52 sway 2022-03-03 18:05:09 `dbus-run-session -- sway` should work then 2022-03-03 18:05:26 hm 2022-03-03 18:05:51 what does pipewire log when you start it 2022-03-03 18:07:14 and what is DBUS_SESSION_BUS_ADDRESS and XDG_RUNTIME_DIR set to 2022-03-03 18:09:16 psykose: I tried both `dbus-run-session` and `dbus-launch --exit-with-session`. when i run pipewire i get a bunch of warnings but it does not exit and there are no errors. 2022-03-03 18:09:27 psykose: DBUS_SESSION_BUS_ADDRESS unix:abstract=/tmp/dbus-vi3JUKHApL,guid=d7304dd59a1ee92be7041ac4622103e8 2022-03-03 18:09:40 XDG_RUNTIME_DIR /tmp/1000-runtime-dir 2022-03-03 18:10:02 do you start pipewire by just running `pipewire` 2022-03-03 18:10:11 without editing any config files 2022-03-03 18:10:13 that's what i tried just now yes 2022-03-03 18:10:14 yes 2022-03-03 18:10:26 that doesn't work with the default config- it doesn't start the session manager of pipewire-pulse 2022-03-03 18:10:31 or * 2022-03-03 18:11:01 yeah but mako also does not work which makes me think i dont have dbus properly set up 2022-03-03 18:11:16 if you start it with `mako` then `notify-send asd asd` does it work 2022-03-03 18:11:23 yes 2022-03-03 18:11:26 so it works 2022-03-03 18:11:35 notify send would fail otherwise 2022-03-03 18:11:49 as for pipewire there is a /usr/libexec/pipewire-launcher 2022-03-03 18:11:49 hmm yes but on my arch install mako is started by dbus 2022-03-03 18:12:01 or you can add the lines to a pipewire override 2022-03-03 18:12:02 so notfy-send works even when i dont start mako manually 2022-03-03 18:12:25 i can't tell you why the dbus activation doesn't work, as it has never worked for me either :p 2022-03-03 18:12:29 and i don't find it very useful 2022-03-03 18:12:56 isn't it also needed to make webrtc screen sharing work? 2022-03-03 18:13:19 no, that works 2022-03-03 18:13:34 not sure how pipewire screen sharing relates to dbus activation 2022-03-03 18:15:23 i somehow assumed that xdg-desktop-portal-wlr is started by dbus 2022-03-03 18:16:13 it usually is 2022-03-03 18:19:07 psykose ah nice pipewire audio stuff works when i run `/usr/libexec/pipewire-launcher` 2022-03-03 18:19:27 :) 2022-03-03 18:20:05 the latest pipewire version now allows merged config overrides, so i think it would be better to ship a /etc/pipewire/pipewire.conf with overriden values instead of the launcher 2022-03-03 18:20:09 then just `pipewire` would work 2022-03-03 18:20:51 psykose yes i saw that in the pipwire changelog. can't i just add a file to `/etc/pipewire/pipewire.conf.d/` now? 2022-03-03 18:20:57 yea 2022-03-03 18:21:00 most likely 2022-03-03 18:21:06 forget the paths 2022-03-03 18:21:10 but sure 2022-03-03 18:21:22 the pipewire.conf has the correct things commented out already 2022-03-03 18:21:28 though i replace media-session with wireplumber 2022-03-03 18:24:12 ok and what do i need to do to get screen sharing to work? 2022-03-03 18:27:03 in my case i just had the portals installed, pipewire+wireplumber, and.. it worked pretty sure 2022-03-03 18:27:09 maybe i had elogind too, but i don't remember 2022-03-03 18:28:19 hmm ok I'll try again when i have pipewire properly configured 2022-03-03 18:28:24 thanks for the help :) 2022-03-03 21:56:39 is there anyone that uses mdev here? can it run a command on laptop lid open/close? 2022-03-03 22:09:33 j`ey: mdev is default in alpine. But... Isn't this related to acpid? 2022-03-03 22:09:48 this is a machine without acpid 2022-03-03 22:09:51 *acpi 2022-03-04 05:00:28 Hi, does anyone know why using tee to write to /proc/acpi/ibm/fan fails with an IO error (strace shows EINVAL), but just redirecting works? (I'm testing this in a root shell so I _can_ use redirections, but echo ... | doas tee /proc/acpi/ibm/fan has the same EINVAL/IO error). I have enabled fan_control in the thinkpad_acpi module, and since redirecting works that can't be the problem. 2022-03-04 05:02:28 no perms is EACCES 2022-03-04 05:02:37 EINVAL means the thing you are passing is wrong 2022-03-04 05:02:43 idk why it would work in one case and not the other though 2022-03-04 05:03:05 psykose: That's the thing. I'm not sure if tee is messing something up 2022-03-04 05:03:13 what are you passing specifically 2022-03-04 05:03:18 and what does a cat of the file give you 2022-03-04 05:05:00 writing "level 7". Cat of /proc/acpi/ibm/fan seems correct for allowing control: https://tpaste.us/LekV 2022-03-04 05:05:17 https://askubuntu.com/questions/1048379/tee-proc-acpi-ibm-fan-invalid-argument 2022-03-04 05:06:00 psykose: I saw that, but like I said redirection works and cat output seems correct 2022-03-04 05:06:14 the redirection does work 2022-03-04 05:06:46 psykose: what do you mean? 2022-03-04 05:08:52 as in as you say the echo works 2022-03-04 05:09:48 psykose: yes, echo level 7 >/proc/acpi/ibm/fan works (only in a root shell ofc) 2022-03-04 05:11:58 if you strace the two of them you will see the write calls are not the same 2022-03-04 05:13:01 echo passes two iov vectors of the content and the newline, tee passes an empty one and then one with content+newline 2022-03-04 05:13:06 or something 2022-03-04 05:13:27 roughly compared https://img.ayaya.dev/uJbEPPfvJ9nd.png 2022-03-04 05:13:37 top is echo 2022-03-04 05:13:48 so i assume, in the tee case 2022-03-04 05:13:54 the first iov vector is invalid, because it's just empty 2022-03-04 05:14:03 or something 2022-03-04 05:14:18 psykose: Yes, I see that too. 2022-03-04 05:14:22 (to pass to the kernel fd, in normal cases this doesn't matter) 2022-03-04 05:14:28 so is this a tee bug? 2022-03-04 05:14:32 i don't think so 2022-03-04 05:14:36 there is probably a reason for it 2022-03-04 05:14:37 well 2022-03-04 05:14:39 idk :) 2022-03-04 05:14:43 but you can just use echo 2022-03-04 05:15:43 psykose: it's a different syntax since you have to do "doas ash -c ..." instead of "echo ... | doas tee ..." 2022-03-04 05:16:38 sure, it is 2022-03-04 05:16:45 but it's just a minor syntax difference 2022-03-04 05:16:48 btw I just tested with coreutils tee and it also writes an empty iov vector 2022-03-04 05:16:53 yeah 2022-03-04 05:17:02 you also don't have to oneline this 2022-03-04 05:17:16 echo val; doas echo val > .. 2022-03-04 05:17:48 psykose: I don't need to have it echoed to stdout, it's that doas doesn't apply to the redirection so tee is needed 2022-03-04 05:17:58 ah right 2022-03-04 05:17:58 echo ... | doas dd of=/proc/acpi/ibm/fan 2022-03-04 05:18:00 but yes 2022-03-04 05:18:04 forgot that part 2022-03-04 05:19:27 Hello71: That works, thanks 2022-03-04 05:19:50 linux api for some virtual files requires single writes to contain the full value. this isn't normally guaranteed in shell, but stars normally align with glibc 2022-03-04 05:20:34 musl uses a more efficient implementation that unfortunately makes these writes typically misaligned 2022-03-04 05:21:15 the "correct" solution is to use dd which is guaranteed to write in the specified block size 2022-03-04 05:22:14 Hello71: I see, thanks for explaining 2022-03-04 07:07:35 Hello71: my workaround was /bin/echo, it worked for me 2022-03-04 07:25:27 aron: To fix this tee einval problem? 2022-03-04 07:25:40 yes 2022-03-04 07:26:10 /bin/echo input >/sys/class/gpio/gpio6/direction 2022-03-04 07:26:27 in my case 2022-03-04 07:26:42 Doesn't work with doas (the redirection), right? 2022-03-04 07:27:16 Hello. I see that "vaultwarden" package is available under edge/testing. I suppose this is just the back-end of the server, vaultwarden-web (the web gui frontend) is not yet available under the same branch/repos? 2022-03-04 07:27:18 didn't tried 2022-03-04 07:28:26 gopher22: apparently not it seems 2022-03-04 07:29:53 ok, cool 2022-03-04 07:30:19 psykose: this reminds me: x2goserver is available, but no x2goclient -- will it be packaged? 2022-03-04 07:30:58 if you want to package it 2022-03-04 07:32:27 hmm, well, I succeeded with the sdcc package (to program STM8s), I think I'll give it a try 2022-03-04 07:33:00 how about xpaint? same thing applies? 2022-03-04 07:33:20 aron: There's a sdcc package? Where? 2022-03-04 07:33:42 ktprograms: are you interested in? 2022-03-04 07:34:05 it's in my local svn repository :P 2022-03-04 07:35:23 aron: oh I see. I don't have a use for it right now, but if/when I get into STM8 programming it might be useful 2022-03-04 07:35:41 yes, definitely :) 2022-03-04 07:36:08 Can you post the APKBUILD somewhere? 2022-03-04 07:37:02 sure, if you like 2022-03-04 07:37:37 can I upload to the official site? 2022-03-04 07:37:51 aron: which official site? 2022-03-04 07:37:58 tpaste.us would be fine 2022-03-04 07:39:47 where the official APKBUILDs are stored 2022-03-04 07:40:12 aron: If you want to make a MR to aports you're welcome to :) 2022-03-04 07:40:45 do I need to register for that? 2022-03-04 07:40:49 yes 2022-03-04 07:44:52 https://tpaste.us/Z4Q9 2022-03-04 07:45:19 it might be not complete :P 2022-03-04 07:45:27 check seems to be missing 2022-03-04 07:45:39 anyway, STM8 part is fully functional 2022-03-04 07:46:10 aron: cool 2022-03-04 07:47:35 yw 2022-03-04 07:47:48 sadly x2goclient has way more dependencies 2022-03-04 10:15:29 How to get apk to log the upgrades it performs? I have installed rsyslog with default config but it doesn't appear to log anything from apk. 2022-03-04 10:26:04 mallory: apk2 does not have support for logging 2022-03-04 10:26:09 it has been added to apk3 2022-03-04 12:24:52 ikke: does it log via syslog? 2022-03-04 12:25:09 Ermine: Not sure 2022-03-04 12:25:27 But my suspicion is that it would log to a logfile 2022-03-04 12:41:49 ikke: ok thanks. Where can I find out about apk 3? Is apk 3 the master branch in https://gitlab.alpinelinux.org/alpine/apk-tools? 2022-03-04 12:42:00 Yes 2022-03-04 12:42:28 It's not released yet 2022-03-04 12:46:53 Ermine: appears to be this commit https://gitlab.alpinelinux.org/alpine/apk-tools/-/commit/646c834492a419a96b4032c230e842d27f87e997, it logs directly to a file 2022-03-04 12:53:49 ikke, mallory: thank you 2022-03-04 20:03:03 I would like to setup alpine mirror. Is there any wiki for that kind of thing? 2022-03-04 20:04:06 frojnd: this has instructions: https://wiki.alpinelinux.org/wiki/How_to_setup_a_Alpine_Linux_mirror 2022-03-04 20:04:44 Thank you 2022-03-05 21:01:53 hey all :) just wondering if I need to do more than update the repo data and then run 'apk update/upgrade' if I want to upgrade from 3.10 to 3.15? 2022-03-05 21:02:22 https://wiki.alpinelinux.org/wiki/Release_Notes_for_Alpine_3.15.0#New_package_signing_keys 2022-03-05 21:02:22 3.14 first for the new signing keys pretty sure 2022-03-05 21:02:32 or manually import them 2022-03-05 21:02:34 You can just install thse keys 2022-03-05 21:02:36 ye 2022-03-05 21:02:44 make sure you provide --available to the upgrade command 2022-03-05 21:03:10 apk add -X https://dl-cdn.alpinelinux.org/alpine/v3.14/main -u alpine-keys 2022-03-05 21:04:54 fab! thanks for the info :) 2022-03-05 21:05:09 silverwillow: also read the other releases notes 2022-03-05 21:06:24 Cool. Will go and RTFM right now. 2022-03-05 21:09:40 Think 'in the end' I'll add a spare SSD to my HP Microserver as root drive and just remount the btrfs filestore 'as-is'. Less hassle. Thanks again :) 2022-03-06 05:56:54 ACTION hey yah all , even you code in assembly there is a layer for it ... . unless you code it using machine language ... . which is really that electricity ... . is the source code of stdio.h 2022-03-07 03:06:21 hello, I'm trying to get my mouse position accurate in Alpine on qemu, Manjaro host. It always has a negative y offset, I've tried starting qemu with usb-tablet device to no avail 2022-03-07 03:06:37 does anyone have some pointers? Or is this strictly a qemu/host issue? 2022-03-07 03:07:49 host on x11, Alpine guest on Wayland if that matters 2022-03-07 04:25:00 manjaro sucks but that's probably not the proximate cause 2022-03-07 04:25:06 what is your qemu command line 2022-03-07 04:34:07 yeah that's the motivation for test driving Alpine 2022-03-07 04:36:27 currently qemu-system-x86_64 -m 2048 -nic user -hda alpine.qcow2-display sdl,gl=on -device virtio-vga-gl,xres=930,yres=620-chardev qemu-vdagent,id=ch1,name=vdagent,clipboard=on,mouse=on-device virtio-serial-pci-device virtserialport,chardev=ch1,id=ch1,name=com.redhat.spice.0 2022-03-07 04:38:06 I replaced -usb -device usb-tablet with the -chardev ..... based on a blog post https://www.kraxel.org/blog/2021/05/qemu-cut-paste/ 2022-03-07 12:38:58 if I want to make a fully readonly / I need to put /run and /tmp as tmpfs I think? but I see that /run has various subdirectories, are they reconstructed per services? 2022-03-07 12:39:59 and /var/log I guess too 2022-03-07 12:43:40 by default you should have /run as tmpfs, you could do /var/log or /var tmpfs too but don't you want to save that data? 2022-03-07 12:44:31 readonly / doesn't mean you can't have a persistent /var partition 2022-03-07 12:44:32 I think that you need a writable /var that survives reboots, at least for some apps/services 2022-03-07 12:45:02 actually I've put /var/log, /run and /tmp and it boots fine :)) 2022-03-07 12:45:13 my app does not need to write anything 2022-03-07 12:46:56 do you see too much benefit with that setup compared to iso boot+overlay / or data install? 2022-03-07 12:47:04 we have another appliance that needs to write a unique file (a SQLite database) and I'm struggling to create a dedicated rw partition just for that 2022-03-07 12:47:56 donoban, we use alpine on raspberry pi (and soon compute module) so basically I automate the task to create partitions, tar the minirootfs and add our own packages 2022-03-07 12:52:39 well, I never tried a read-only root, I don't know if the benefits outweigh the effort 2022-03-07 14:58:13 Hello, I am experiencing an issue installing alpine (alpine-standard-3.15.0-x86_64) with encryption using it's installer setup-alpine. I've used the setup with the following data on a Dell XPS 15 7000 Series-7590 Laptop with a USB Ethernet Adapter (omitted other fields). interface: eth0, Mirror: f (detect), Mode crypt and sys on nvme0n1 setup reported success, rebooted machine, got past grub menu and was not presented with password prompt 2022-03-07 14:58:32 mount: mounting UUID=da03afb-27b0-4c34-bf16-9fb3cbdcb35a on /sysroot failed: no such file or directory. Mounting root failed. initramfs emergency recovery shell launched. Type 'exit' to continue to boot sh: can't access tty: job control turned off / # exit BusyBox v1.34.1 (2022-02-02 18:21:20) multi-call binary Usage: switch_root [-c CONSOLE_DEV] NEW_ROOT NEW_INIT [ARGS] 2022-03-07 14:58:42 Free initramfs and switch to another rootfs: chroot to NEW_ROOT, delete all in /, move NEW_ROOT to /, execute NEW_INIT. PID must be 1. NEW_ROOT must be mountpoint. -c DEV reopen stdio to DEV after switch 2022-03-07 14:59:08 However, doing the installation without crypt works absolutely fine 2022-03-07 15:01:06 Did I forget something during the installation? 2022-03-07 15:35:50 I think that crypt setup is not very tested on differents schemes/fs/uefi/etc... combinations 2022-03-07 15:37:18 SirMetathyst: I would try again and check mkinitfs.conf and kernel parameters 2022-03-07 15:38:36 if it didn't ask for a password it's likely to miss cryptsetup on the initfs or cryptroot on kernel cmdline 2022-03-07 15:58:56 donoban: tbh I'm not very familiar with linux. I have installed arch before but that was because of the great documentation. Is there a resource you can point me at to help me with this issue? 2022-03-07 16:07:34 This is a guide for FDE with alpine https://battlepenguin.com/tech/alpine-linux-with-full-disk-encryption/ , theorically the setup-disk script should achieve something similar 2022-03-07 16:08:24 The /etc/mkinitfs/mkinitfs.conf should have cryptsetup and nvme 2022-03-07 16:08:42 and # /etc/default/grub 2022-03-07 16:09:05 should have a line like GRUB_CMDLINE_LINUX_DEFAULT="cryptroot=UUID=xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxx ..." 2022-03-07 16:09:50 you could install some irc client, enter here and paste that files before rebooting 2022-03-07 16:29:38 thank you for your help donoban, I'll look into what you sent me. 2022-03-07 16:31:58 crypt on nvme is broken 2022-03-07 16:32:42 https://gitlab.alpinelinux.org/alpine/alpine-conf/-/issues/10498 2022-03-07 16:32:45 ouch 2022-03-07 16:33:39 it seems another [3.8] curl: Multiple vulnerabilities (CVE-2019-5435, CVE-2019-5436) 2022-03-07 16:33:53 ahh, alpine-conf 2022-03-07 16:35:07 Hello71: so does it just need to add nvme to initfs? 2022-03-07 16:35:14 yes 2022-03-07 16:35:45 ok 2022-03-07 17:15:28 I've installed openssh-server-pam, should I add /etc/pam.d/sshd ? 2022-03-07 17:27:18 uhM, do you shouldn't need to ask it if you want to ose it :P 2022-03-07 17:27:35 s/to ose/to use/ 2022-03-07 17:27:35 donoban meant to say: uhM, do you shouldn't need to ask it if you want to use it :P 2022-03-07 17:28:20 what do you want to achieve with pam? If there is no 'sshd' config it will rely on 'others' which should be basic pam-unix (password based auth) 2022-03-07 17:30:21 Okay, basic auth (and key-based auth) is all I need 2022-03-07 17:30:58 so better use openssh-server, it has both built-in support 2022-03-07 17:36:29 ... but via pam because of cloud-init 2022-03-07 17:38:16 uhM, it's a requeriment for it? 2022-03-07 17:42:06 there is an issue: cloud-init will password lock users, and openssh will refuse key-based logins 2022-03-07 17:42:19 hello 2022-03-07 17:42:46 as per /usr/share/doc/README.Alpine from cloud-init-doc 2022-03-07 17:46:53 ah, I see, so it's some kind of workaround 2022-03-07 17:46:56 it is a openssh decision to treat accounts with locked passords as locked in general unless PAM is enabled 2022-03-07 17:47:32 and can't cloud-init stop disabling them? 2022-03-07 17:48:10 it could, but the default config is for only key-based access 2022-03-07 17:50:16 Ermine: are you having specific problems? 2022-03-07 17:50:31 why doesn't it use invalid password instead of locked 2022-03-07 17:53:12 minimal: no, I'm just making sure everything is configured correctly 2022-03-07 21:12:46 I'm following this, as I'd like to rename interfaces according to "inside,outside,wifi,etc" and to ensure that a usb nic always gets the same dev name: https://wiki.alpinelinux.org/wiki/Custom_network_interface_names nameif -s works and renames the interfaces accordingly, I did update the initramfs, but interfaces are not renamed on reboot. I'm using the standard version of 2022-03-07 21:12:48 alpine, it's running "on the metal" of this server. 2022-03-07 21:39:43 Paruza: so you have an /etc/mactab file? 2022-03-07 21:40:02 yeah 2022-03-07 21:40:21 And it's in the initramfs? 2022-03-07 21:40:22 and nameif -s when manually run, does rename the interfaces accordingly 2022-03-07 21:41:12 well i ran mkinitfs and it reports "==> initramfs: creating /boot/initramfs-lts" 2022-03-07 21:41:23 i havent tried to explicitly mount it and look though 2022-03-07 21:41:41 bsdtar tvf /boot/initramfs-lts 2022-03-07 21:41:46 That'll list the files 2022-03-07 21:42:00 `bsdtar tvf /boot/initramfs-lts /etc/mactab` to check for that one file specifically 2022-03-07 21:42:12 Note: you want bsdtar specifically 2022-03-07 21:42:30 what package is that in? 2022-03-07 21:42:42 nm 2022-03-07 21:42:43 libarchive? 2022-03-07 21:42:47 yeah 2022-03-07 21:43:21 hmm bsdtar isnt anywhee and libarchive exists 2022-03-07 21:43:32 I could just extract it into a folder though 2022-03-07 21:45:40 hmm 2022-03-07 21:50:07 ok I copied it into a different folder, gunzip, cpio -i, and I do not see that file 2022-03-07 21:51:11 Because mdev is configured to call nameif, I'm guessing it's doing so in the initramfs env and lacks the mactab file 2022-03-07 21:53:31 hmm it looks like that file is not inside of base.files 2022-03-07 21:54:21 brb, rebooting 2022-03-07 22:09:03 Did it work? 2022-03-07 22:09:19 well, I added /etc/mactab to /etc/mkinitfs/features.d/base.file, reran mkinitfs, verified it did in fact get added to the initfs, but it didnt rename the interfaces on boot. However I also tried manually renaming them and though according to ip a they are renamed... udhcp claims it cant find the interface 2022-03-07 22:10:38 :/ 2022-03-07 22:11:00 udchp gives that error /after/ the interfaces are renamed? 2022-03-07 22:12:00 yeah 2022-03-07 22:13:13 didnt dig too much into it though 2022-03-07 22:13:24 so I wouldnt give that itself too much thought 2022-03-07 22:13:32 Paruza: the /sbin/nameif isn't packed into initramfs, and tbh it shouldn't even be 2022-03-07 22:13:41 that should be done in userspace by the mdev service 2022-03-07 22:14:03 do you have that service enabled in the sysinit/boot runlevel? 2022-03-07 22:14:44 you should be able to see that by running `rc-update` 2022-03-07 22:21:31 mdev is not 2022-03-07 22:22:32 hm, should be on a default alpine install 2022-03-07 22:22:41 udev is... 2022-03-07 22:22:43 you can add it with `rc-update add mdev sysinit` 2022-03-07 22:23:09 this way it will run before networking, which is usually in the boot runlevel 2022-03-07 22:23:16 ok brb ill reboot quick 2022-03-07 22:26:53 did not see to do anything 2022-03-07 22:27:01 seem* 2022-03-07 22:29:37 does `service mdev status` report that it started? 2022-03-07 22:29:49 yes 2022-03-07 22:30:57 do you have net-tools installed? 2022-03-07 22:31:10 ah, dumb question 2022-03-07 22:31:16 you've run the exec before 2022-03-07 22:31:20 hm 2022-03-07 22:31:45 um well I just ran apk add net-tools and it installed stuff... 2022-03-07 22:31:58 yeah, that should be fine 2022-03-07 22:39:54 brb 2022-03-07 22:44:05 can you grep your /var/log/messages for "nameif"? maybe it errors out somehow 2022-03-07 22:45:05 a few messages from manual attempts over an hour ago 2022-03-07 22:45:14 nothing from recent reboots 2022-03-07 22:45:49 hm 2022-03-07 22:46:15 i guess you can try using udev instead 2022-03-07 22:47:04 and put this file into /etc/udev/rules.d, eg. '10-nameif.rules': https://tpaste.us/5kvX 2022-03-07 22:48:43 I may wind up doing that, but I need to go for now. Thanks for your help o/ 2022-03-08 03:24:28 Hi, does anyone here have experience with waydroid? Is the "waydroid session start" command supposed to exit or does it keep running? I'm getting a "Failed to get service waydroidplatform, trying again..." error when running "waydroid show-full-ui", but "doas waydroid shell" seems to work. 2022-03-08 04:14:29 when I reboot my alpine nas, it doesn't mount my zpool 2022-03-08 04:14:40 I have to manually run modprobe zfs and zpool import 2022-03-08 04:14:47 not sure when this started 2022-03-08 04:15:23 adding zfs to /etc/modules should make sure the first part is done 2022-03-08 04:15:26 not sure for the latter 2022-03-08 04:15:39 I've looked in dmesg and /var/log/messages, don't see anything 2022-03-08 04:19:54 hi, I'm trying to get clipboard sharing working with an Alpine guest in qemu, and using spice-vdagent and spice-vdagentd. When I try to start spice-vdagent after the vdagentd service is running, I get an error where /dev/uinput doesn't exist 2022-03-08 04:20:05 is this a red herring or do I need that kernel module somehow? 2022-03-08 04:20:13 I'm using alpine virt img 2022-03-08 04:21:12 mouse and keyboard work in the guest, so not sure if this is important 2022-03-08 04:21:46 wakyct: uinput isn't built in the virt kernel because the virt kernel is meant for headless cloud VMs 2022-03-08 04:22:01 ah I have the wrong kernel then 2022-03-08 04:22:17 or do you think I need it for my purpose? 2022-03-08 04:22:45 Just install the lts kernel, it should work fine in qemu. FYI add linux-firmware-none to your "apk add" command so it doesn't install all the firmware (not needed in qemu) 2022-03-08 04:22:58 OK thanks I'll try that 2022-03-08 04:23:17 If you want copy/paste, then you'll need spice-vdagent, which needs uinput, which needs the lts kernel. 2022-03-08 04:23:35 OK makes sense 2022-03-08 04:49:29 thanks ktprograms fixed that error, still working on clipboard sharing host<->guest, I feel like most of the pieces are there, must be missing something 2022-03-08 04:50:07 I have spice-vdagent and vdagentd, my qemu CL looks right 2022-03-08 04:50:28 does vdagent not work with Wayland maybe? I have Xwayland 2022-03-08 04:50:47 I haven't tested it with wayland 2022-03-08 04:51:15 OK I should try with X, xfce good to try? 2022-03-08 05:11:11 wakyct: I have had success with clipboard sharing with XFCE/X/QEMU 2022-03-08 05:15:59 Could I get some help with amove, or an example of an APKBUILD that uses it for a binary that doesnt have an install target in its makefile? 2022-03-08 05:16:17 I got something built fine, but I'm struggling with packaging step with amove (as I assume that's what I need) 2022-03-08 05:21:29 if it's not installed then there is nothing to amove 2022-03-08 05:21:34 amove just moves things already in pkgdir 2022-03-08 05:25:20 Ahh, okay. So if it is just a binary hanging out in the $srcdir, what do I do? Manually move it? 2022-03-08 05:25:38 sure 2022-03-08 05:28:48 use install or cp 2022-03-08 05:28:53 otherwise it will not be owned by root 2022-03-08 05:29:38 Ahh, thank you. No install target in this makefile, so I used cp. I was wondering about permissions, as I've seen some APKBUILDS manually specifiy them 2022-03-08 05:29:43 ACTION sent a code block: https://matrix.org/_matrix/media/r0/download/matrix.org/cDTbnsNxvMwGNfzDjLLZfTCZ 2022-03-08 05:29:51 This built, packaged, installed, and ran locally. 2022-03-08 05:29:59 Does it seem out of sorts? 2022-03-08 05:30:56 you can just use a single install 2022-03-08 05:31:29 install -Dm755 ./mp3val -t "$pkgdir"/usr/bin 2022-03-08 05:31:53 is that install command provided by abuild? 2022-03-08 05:31:58 busybox 2022-03-08 05:32:49 Oooh, thank you 2022-03-08 05:33:03 Other than that and cleaning up the spurious commented out options line, is it okay? 2022-03-08 05:33:13 Does my check actually count, or should I just skip check? 2022-03-08 05:34:05 it's better than nothing, i'm mostly indifferent 2022-03-08 05:34:18 Alright, thank you 2022-03-08 05:34:26 I suppose the rest will be covered in MR review 🙂 2022-03-08 05:34:40 Best to flag all as Draft first, correct? 2022-03-08 05:38:59 Thanks, as always, psykose & Hello 🙂 2022-03-08 05:43:36 no, draft means you yourself think it's not ready and you will make more changes 2022-03-08 05:44:13 Oooh, okay. I've misunderstood that 2022-03-08 05:44:37 aside from the typo on $"(), i would merge it already 2022-03-08 05:45:31 Alright, just fighting the linter first haha 2022-03-08 05:45:42 that's what it caught :p 2022-03-08 05:47:12 Yes, and I clearly did not get the memo the first fix haha 2022-03-08 05:47:20 Waiting on platforms now... 2022-03-08 05:47:35 I don't think anything should fail since it's just plain C, but I don't know 2022-03-08 05:48:01 looks fine to me 2022-03-08 05:48:09 Next up is wxmp3val/wxmp3gain front-ends since someone beat me to mp3gain already 🙂 2022-03-08 05:48:53 Removed draft 2022-03-08 05:49:27 This is such a huge part of my like, daily personal workflow I can now do under Alpine that I'm really quite pleased 😀 2022-03-08 05:49:36 :) 2022-03-08 05:50:23 Hopefully other folks benefit, too 2022-03-08 05:50:38 This tool has done it's work over the past like, 20 years I've been using it haha 2022-03-08 05:50:49 damn 2022-03-08 05:50:53 i could barely walk 20 years ago 2022-03-08 05:51:01 😛 2022-03-08 05:51:21 Yeah, I'm aging myself here. Remember DAPs? I was obsessed 2022-03-08 05:51:30 i definitely do not remember those 2022-03-08 05:51:31 :p 2022-03-08 05:51:43 Digital Audio Players, pre-Apple even 😛 2022-03-08 05:51:47 ahh 2022-03-08 05:52:27 I'm still oldschool. Buy my media on CD, rip/archive and then playback via *sonic, mpd, or local players 2022-03-08 05:52:33 I do remember those early mp3 players 2022-03-08 05:52:34 So having these tools is huge for me 2022-03-08 05:52:41 32MB storage :D 2022-03-08 05:52:43 ikke: What did you have?! 2022-03-08 05:52:58 What was your favorite, or holy-grail one you never got? 2022-03-08 05:53:03 I don't remember the brands anymore 2022-03-08 05:53:11 Ahhh haha 2022-03-08 05:53:30 My holy grail was the Rio Karma, which I did finally get 2022-03-08 05:53:32 At that time they were dime in a dozen 2022-03-08 05:53:42 My favorite is the MPIO HD300, which I still have, and converted to SSD 2022-03-08 05:53:54 Yeah, I was a tester/writer for DAPReview.net back then haha 2022-03-08 05:54:09 I would work all summer, save up, and then buy DAPs to test 2022-03-08 05:54:13 Good times, good times 2022-03-08 05:54:33 these kinds of form-factors: https://i5.walmartimages.com/asr/5fb3f74f-e7f4-44bd-8eca-59e4b77190d7_1.a4916c20b16abe7816e81d405978fbc9.png 2022-03-08 05:55:43 oh i remember these 2022-03-08 05:55:43 Ahh, yeah, stick ones were fun. I had an MPIO/iBullDog IMP-90 with a whopping 256MB storage and a AAA powercell 2022-03-08 05:55:45 lmaooo 2022-03-08 05:55:57 i used one as a usb stick as a teenager 2022-03-08 05:55:59 just flash storage 2022-03-08 05:56:08 still no sharing with xfce unfortunately, maybe a qemu config issue then 2022-03-08 05:56:47 wakyct: What qemu-* packages do you have installed? 2022-03-08 05:56:56 on host side? 2022-03-08 05:57:37 Yeah 2022-03-08 05:57:44 I have these... (full message at https://matrix.org/_matrix/media/r0/download/matrix.org/IeZJsRfAlQEjmNOBNaaMiqLb) 2022-03-08 05:59:14 psykose: Jealous. My first USB drive was 512MB from my University store and it cost me $60. 2022-03-08 05:59:28 👴 2022-03-08 06:03:55 ha yeah Saijin_Naib, I pretty much have just qemu 2022-03-08 06:04:23 thanks for the list, I'll see if I'm missing something 2022-03-08 06:04:40 wakyct: Basically for copy/paste to work, you'll need to use spice in the qemu argument, and use the spice viewer for interacting with the VM 2022-03-08 06:09:15 I have seen that, and haven't tried it, I should do that next. I thought I was good based on new stuff in qemu 6 2022-03-08 06:09:40 https://www.kraxel.org/blog/2021/05/qemu-cut-paste/ 2022-03-08 06:11:01 wakyct: what's your qemu command? 2022-03-08 06:13:50 https://gist.github.com/ergodicbreak/6c9113c4c33a543ee059efba5b5b0a6c 2022-03-08 06:16:17 spice-vdagent is working, but just showing clipboard events in the guest 2022-03-08 06:16:27 I wonder if it's because I'm using the sdl ui 2022-03-08 06:16:29 rather than gtk 2022-03-08 06:16:35 in qemu 2022-03-08 06:29:42 holy moley gtk works 2022-03-08 06:41:34 not working in wayland though, bummer 2022-03-08 06:59:46 Does anyone have insight on this, I've figured out that my spice-vdagent in the guest is seeing clipboard events in the host, but the text is not transmitting from host to guest 2022-03-08 07:00:23 basically I get a new event in var/log/messages when I select in the host 2022-03-08 07:00:52 something specific to how clipboard works in sway maybe? 2022-03-08 07:01:14 in var/log in the guest I should say 2022-03-08 07:32:45 I was wrong about text not sending from host to guest, it does send, but the paste key combo in the guest doesn't paste it 2022-03-08 07:32:58 it only pastes what's copied in the guest 2022-03-08 07:33:19 but if I xclip -o, I see in the guest what was copied in the host 2022-03-08 07:33:57 must be some Wayland thing with spice-vdagent only working with X 2022-03-08 07:34:07 since host is on X 2022-03-08 07:37:14 yeah, https://www.reddit.com/r/swaywm/comments/pg0rqi/clipboard_sharing_using_spicevdagent_not_working/ 2022-03-08 07:37:15 [REDDIT] Clipboard sharing using spice-vdagent not working (self.swaywm) | 3 points (03100.0%) | 2 comments | Posted by Barangtastic | Created at 2021-09-01 - 19:10:34UTC 2022-03-08 07:37:52 anyway, that's a decent enough workaround for now I guess 2022-03-08 10:52:38 Hello, this is my first time that I use IRC. Please forgive me, if I make mistakes. 2022-03-08 10:52:58 How do I disable ipv6 on alpine linux 3.15? 2022-03-08 10:55:13 if using syslinux, add "ipv6.disable=1" to the "append" line in /boot/extlinux.conf. at least, that kernel option used to be a way of doing it 2022-03-08 10:55:16 sysctl -w net.ipv6.conf.all.disable=1 2022-03-08 10:58:58 if I try sysctl -w net.ipv6.conf.all.disable=1, I get an error. sysctl: error: 'net.ipv6.conf.all.disable' is an unknown key 2022-03-08 11:02:00 looks like it's disable_ipv6 at the end 2022-03-08 11:02:15 see /proc/sys/net/ipv6/conf/all/ 2022-03-08 11:03:50 ok got it, thanks for your help. this did work: sysctl -w net.ipv6.conf.default.disable_ipv6=1 2022-03-08 11:07:33 you'll want to add "net.ipv6.conf.all.disable_ipv6 = 1" to /etc/sysctl.conf or /etc/sysctl.d/, if you want it to stick 2022-03-08 11:18:07 thanks lopid, I will do that 2022-03-08 13:02:48 Greetings. I've been using qutebrowser on alpine for some time now and I keep getting random hiccups and render process crashes on a regular basis. My guess it's related to musl, so I was wondering if anyone can hint me the right command to debug that since using gdb -ex r --args $(readlink -f $(which python3)) -m qutebrowser --debug only logs python symbols. 2022-03-08 14:01:58 use command -v instead of which 2022-03-08 14:02:15 and you also need --qt-flags=single-process or somesuch 2022-03-08 14:56:26 Hello71: ty for suggestion, but with --qt-flag=single-process qutebrowser hangs almost immediately and without it I still get nothing more than ERROR: Renderer process crashed (status 11). Guess I'll try to tinker with KDE's falkon for now. 2022-03-08 15:07:01 hi 2022-03-08 15:07:09 dmitz: its a shit browser 2022-03-08 15:07:15 vulnerable to 0-days 2022-03-08 15:07:20 since it has small team of devs 2022-03-08 15:07:23 don't use 2022-03-08 15:09:51 hi , how can I avoid installing all the firmware packages ? . I want to install only specific firmware packages , linux-lts and zfs-lts . Thanks 2022-03-08 15:12:24 enlight: and you came to this conclusion based on them using exactly the same engine? 2022-03-08 15:12:46 taendeda: if you install some firmware packages then the default set will be uninstalled 2022-03-08 15:13:08 (if you want to install no firmware there is linux-firmware-none) 2022-03-08 15:15:00 Hello71: no, they have small team of devs, it applies to all 2022-03-08 15:15:15 the rate with which chromium, gecko and webkit fix 0days and apply patches 2022-03-08 15:15:26 most hackers don't even look at qute browser for bugs 2022-03-08 15:15:35 thus it remain unpatched and far more vulnerbael 2022-03-08 15:15:37 because it uses the chromium engine? 2022-03-08 15:15:57 by that logic alpine is insecure 2022-03-08 15:16:31 occasionally there are vulnerabilities in the distro-specific parts but 99(.9)% of vulnerabilities are in the shared parts 2022-03-08 15:17:26 Hello71: many thanks . It worked . 2022-03-08 15:24:15 Hello71: out of curiosity, what browser did you use to get the stacktrace mentioned here: https://gitlab.alpinelinux.org/alpine/aports/-/issues/13333#note_199521? 2022-03-08 15:24:43 qutebrowser 2022-03-08 15:25:29 i don't remember command exactly. possibly something like gdb --args python3 /usr/bin/qutebrowser --qt-flags=single-process 2022-03-08 15:30:04 and what about debug packages? Should musl-dbg and qt5-qtwebengine-dbg suffice? 2022-03-08 15:30:49 theoretically 2022-03-08 15:31:16 possibly need apk upgrade -a 2022-03-08 15:31:56 ok, thanks a lot 2022-03-08 15:55:27 oh, also i think qt5-qtwebengine-dbg is broken in 3.15 2022-03-08 16:09:11 Hello71: yeah, i saw your MR addressing the issue, i'm on edge so not a big deal for me 2022-03-09 02:58:43 Hey, network connections aren't working right on my server. When I ifdown/ifup udhcp gets a lease, but I can't connect to anything unless I manually add routes. What could the problem be? If I can't figure out the root cause, is there a way to add the routes every time the interface comes up? 2022-03-09 03:09:22 probably your dhcp server is not providing any gateway 2022-03-09 03:09:34 check tcpdump 2022-03-09 03:09:46 well, not sure if tcpdump can decode dhcp options 2022-03-09 03:12:23 the thing is, even when I have a static configuration it still won't connect without manually adding routes 2022-03-09 03:25:25 what do you mean "I have a static configuration" 2022-03-09 03:26:02 when i specify an IP and gateway and netmask in /etc/network/interfaces 2022-03-09 03:26:12 instead of having it use dhcp 2022-03-09 05:10:30 is the custom openrc service still needed to install alpine on btrfs as mentioned here: https://wiki.alpinelinux.org/wiki/Btrfs 2022-03-09 05:43:54 is wine broken on alpine 3.15? 2022-03-09 07:18:40 aron: What are you experiencing? 2022-03-09 07:19:16 I'm having my APKBUILD prompt for polkit auth during the packaging stage... What did I manage to do this time? 2022-03-09 07:19:22 ACTION sent a code block: https://matrix.org/_matrix/media/r0/download/matrix.org/COujAqtjdTdDEZAaNeZkboti 2022-03-09 07:20:16 (gotta say, I'm pretty amped I made it to packaging before needing help this time 🤣) 2022-03-09 07:22:46 Saijin_Naib: I can't start my win32 program 2022-03-09 07:22:47 it's installing to / instead of pkgdir 2022-03-09 07:23:02 win32 can't be started on alpine because it requires wine32 2022-03-09 07:23:05 Ah, awesome. I'm just tinkering with that now 2022-03-09 07:23:10 we don't have multilib, so it's wine64 only 2022-03-09 07:23:19 which admittedly, makes most windows programs not.. usable 2022-03-09 07:23:24 wine is quite not good on alpine :) 2022-03-09 07:23:37 i think in like 5 releases wine will support win32 without multilib, but it's not there yet 2022-03-09 07:23:42 Yeah, I can run a number of 64bit clean programs, but not many installers (which apparently are still 32bit even for 64bit programs) 2022-03-09 07:23:47 yep 2022-03-09 07:23:48 so, it's more likely useless 2022-03-09 07:23:56 for most windows things, yes 2022-03-09 07:24:09 Is what you're using available 64bit without an installer? 2022-03-09 07:24:14 Saijin_Naib: if it's using gtk2 then it will most likely not be added 2022-03-09 07:24:18 Or, can you install it somewhere and copy the dir over? 2022-03-09 07:24:51 psykose: Are we deprecating gtk2 programs? 2022-03-09 07:24:58 they have been for a while 2022-03-09 07:25:03 existing things are kept, but nothing new is added 2022-03-09 07:25:10 😫 2022-03-09 07:25:21 it is after all a dead framework in the maintenance sense sadly 2022-03-09 07:25:41 No, makes sense, I guess. Just a number of cool things apparently still build against it 2022-03-09 07:25:59 Drat, guess this one is dead in the water then 2022-03-09 07:26:45 it seems, I have to move all that stuff into a vm 2022-03-09 07:26:59 (but I liked wine, it was super handy) 2022-03-09 07:27:13 there is also flatpak wine i think 2022-03-09 07:27:21 that one would work fine given multilib and all via flatpak 2022-03-09 07:28:11 Saijin_Naib: heh, it seems grip is even still maintained 2022-03-09 07:28:18 surprised they ported the build to meson 2022-03-09 07:28:40 Yeah 🙂 Problem is the current maintainer is somewhat like me, can't code 2022-03-09 07:28:46 psykose: what is flatpak wine? 2022-03-09 07:28:53 wine installed via flatpak 2022-03-09 07:29:45 They have GTK3 port and other fixes in their ToDo as accepting patches since they can't implement it themselves 2022-03-09 07:29:51 makes sense 2022-03-09 07:30:04 is it something that similar to snap? 2022-03-09 07:30:09 somewhat 2022-03-09 07:30:15 eww 2022-03-09 07:30:22 VM will be fine 2022-03-09 07:30:34 If it makes you feel any better, I've had much better luck with flatpak than SNAP 2022-03-09 07:30:49 flatpak is better than snap and VM is a bit overkill 2022-03-09 07:31:12 I saw snap before, it's pure cancer 2022-03-09 07:32:05 VM is essential, I have to run 3D stuff, the VM has a dedicated graphics card 2022-03-09 07:32:32 flatpak has GPU access 2022-03-09 07:32:45 i play games just fine via flatpak steam :p 2022-03-09 07:32:46 It's pretty much native, aside from the install sizes 2022-03-09 07:32:50 not sure how much more 3d it gets than that 2022-03-09 07:32:56 Same for SteamLink flatpak 2022-03-09 07:33:40 nah, I pass 2022-03-09 07:36:23 psykose: Make the PR for this anyway, or waste of folks' time since GTK+2.0? 2022-03-09 07:37:05 up to you (though you should look at the abuild-meson usage in other places first) but i would probably not merge it :) 2022-03-09 07:38:02 just one more thing about wine: should work with win64 binaries? 2022-03-09 07:38:11 Hmmm... Have any suggestions for GUI CD rippers with CDParanoia support? 2022-03-09 07:38:39 it should 2022-03-09 07:38:40 Theres whipper which has CDParanoia and AccurateRip support, but is CLI only 2022-03-09 07:38:43 Maybe that's fine... 2022-03-09 07:38:53 sadly i don't rip many cds 2022-03-09 07:39:10 i remember using nero back in the day on windows 2022-03-09 07:39:11 haha 2022-03-09 07:39:18 but for making them 2022-03-09 07:39:33 :) 2022-03-09 07:40:02 I made a lot of (data) CD images too 2022-03-09 07:40:17 then mount them into the virtual CD drive 2022-03-09 07:40:47 Aron, I'm using Alpine wine to run 7zip and Tuniac daily, so it works fine for 64bit clean stuff 2022-03-09 07:41:28 Saijin_Naib: cool 2022-03-09 07:41:52 Ah, I rip a ton, hence my mp3val add yesterday, and goal of adding wxmp3gain, wxmp3val, and wxlame, and looking for a good secure CD ripper to add to Alpine 2022-03-09 07:42:15 unfortunately, my win32 programs are quite legacy, back from '00 2022-03-09 07:42:33 Ahh... Yeah. I've got that problem as well. No solution for that yet for me 2022-03-09 07:42:43 Though I go back to Win 3.1 stuff haha 2022-03-09 07:42:49 And OS/2 😀 2022-03-09 07:45:31 hehe 2022-03-09 07:46:22 I liked win 3.1's cardfile :P 2022-03-09 07:59:39 It's a comfy, but limited DE, yeah 2022-03-09 08:00:14 psykose: Is the gtk+2.0 deprecation why LXDE got dropped from Alpine (though they picked it back up for development alongside LXQT apparently)? 2022-03-09 08:00:23 no idea 2022-03-09 08:38:14 hah, I have an idea 2022-03-09 08:38:26 x86 has 32-bit wine 2022-03-09 08:39:12 so, I could create a 32-bit environment with i.e. qemu-i686, and run win32 stuff with wine the that environment 2022-03-09 08:41:33 *in that 2022-03-09 08:42:47 aron: isn't qemu-i686 (user mode emulation) going to be quite slow since it can't use virtualization? 2022-03-09 08:44:21 I think it can use virtualization, but I'll give it a try sometime 2022-03-09 08:48:44 do you have a suggestion to restart a service as non-root? we are creating a web application that allows restarting a service in our industrial application 2022-03-09 08:48:54 I could run the web application as root but... kinda dislike that 2022-03-09 09:05:09 add a doas.conf thing for a specific user to be able to restart something 2022-03-09 09:05:47 permit nopass myuser cmd rc-service args myservice restart 2022-03-09 09:05:48 yes that can be a good idea 2022-03-09 09:05:49 something like that 2022-03-09 09:05:56 let me try that, thanks 2022-03-09 09:09:09 sounds like a hack :) 2022-03-09 09:09:32 sounds like a limitation of openrc 2022-03-09 09:11:54 I could send a SIGINT to the process (since it's running as same user as the webapp kill() will work) but supervise-daemon may log that it has crashed (while it didn't) 2022-03-09 14:30:32 if the service is running as your user then just call rc-service restart 2022-03-09 14:31:04 the only problem is that it will inherit your web server environment 2022-03-09 14:31:42 you can mostly fix that with env -i rc-service restart and ignore rlimits etc 2022-03-09 15:06:39 hi all! i am looking to autoinstall alpine and have some clarifying questions. do i understand it right that there is no officially supported way to avoid manually logging into root and execing setup-alpine? 2022-03-09 15:24:17 you can probably use cloud-init or ansible 2022-03-09 15:26:20 well, that still requires logging into root manually, i presume they mean something like openstack disk images that have cloud-init preinstalled 2022-03-09 15:26:30 which alpine currently doesn't have 2022-03-09 15:29:12 some way of automating the install really 2022-03-09 15:29:30 isn't there some place in the ramdisk i can hook a script into? 2022-03-09 15:30:46 i'd gladly work with ansible, but that requires setting up the network connection and sshd 2022-03-09 15:36:34 Saijin_Naib: wrt cd-rippers, cdparanoia is already in the "main" repository. Would that work? 2022-03-09 15:38:20 Saijin_Naib: Never mind, reading the logs backwards. You want a GUI ripper. :) 2022-03-09 15:40:49 nangel: Yeah, I'm trying to make myself comfortable within Alpine so I can finally put Windows in a VM with Alpine as the host, but to do that, I need to find tools for all my personal workflows that are like, 20+ years old at this point so not terribly flexible 🤣 2022-03-09 15:41:31 I love ExactAudioCopy, so I wanted to package a GUI ripper with a similar secure copy mechanism 2022-03-09 15:41:45 Do I need it to be GUI? Grumble... No... 2022-03-09 15:42:18 Whipper is excellent and someone added it, but it's not the UX I want. I dunno. Being fussy I suppose 2022-03-09 15:46:10 Maybe if I put a bunch of crying face emoji in my MR psykose won't reject grip for still being gtk+2.0 based 🤔 2022-03-09 15:47:46 I loved ExactAudioCopy! :) 2022-03-09 15:48:45 i will do it for 4 headpats 2022-03-09 15:57:42 Steep price, but it must be paid 😔 2022-03-09 16:25:55 Hello someone able to help me solving a problem regarding updating my repository index? Using apk update results in ERROR http://dl-cdn.alpinelinux.org/alpine/v3.15/main: BAD archive. Fetching the file APKINDEX.tar.gz via wget works without problems. Architecture is armv7 running on an old rpi2 2022-03-09 16:30:38 maybe -v or -vv will help 2022-03-09 16:32:01 Sadly with -v or -vv there is the same output with fetch http://dl-cdn.alpinelinux.org/alpine/v3.15/main/armv7/APKINDEX.tar.gz 2022-03-09 16:32:04 ERROR: http://dl-cdn.alpinelinux.org/alpine/v3.15/main: BAD archive 2022-03-09 16:32:07 WARNING: Ignoring http://dl-cdn.alpinelinux.org/alpine/v3.15/main: BAD archive 2022-03-09 16:32:10 2 errors; 37 distinct packages available 2022-03-09 16:32:30 try switching to a different mirror. 2022-03-09 16:33:24 Tried with rwth-aachen, hs-esslingen and dl-cdn.alpinelinux.org so far. Do you have a mirror confirmed working? 2022-03-09 16:38:27 dl-cdn should always work 2022-03-09 16:38:47 if wget works then not sure why the regular update would fail 2022-03-09 16:41:24 I'm looking into it 2022-03-09 16:43:14 I cannot reproduce it 2022-03-09 16:43:15 psykose: My hunch is the file is there (wget works) but it is corrupt. ikke is working on it. :) 2022-03-09 16:43:26 oh, 3.15 2022-03-09 16:43:27 :) 2022-03-09 16:43:54 https://tpaste.us/DBk0 2022-03-09 16:43:56 same with 3.15 2022-03-09 16:44:17 Is there some proxy that is corrupting things? 2022-03-09 16:44:36 blechdose: ^ 2022-03-09 16:44:57 Nope only default HomeRouter with NAT 2022-03-09 16:44:59 inb4 the proxy returns an html document 2022-03-09 16:45:31 blechdose: what is the filesize and sha512sum of the apkindex file? 2022-03-09 16:45:31 does it still save the apkindex on failure 2022-03-09 16:46:09 blechdose: can you also try https? 2022-03-09 16:46:12 I see this is http 2022-03-09 16:46:35 curl -O https://dl-cdn.alpinelinux.org/alpine/v3.15/main/armv7/APKINDEX.tar.gz 2022-03-09 16:46:41 Should be in /var/cache/apk right? Dir is empty; Tried https first, same result but I try again 2022-03-09 16:48:41 Sadly I don't have curl, it's the default armv7 image. 2022-03-09 16:48:43 fc9e74818c1a309262bb7fa6fa6bfefc551fe187b04907d485a6b118fb4069ffeb6baa0a3d23f9a6c1c9620a2514f2fd3c24f661d00fd45f0404ed3e9bc16da1 APKINDEX.tar.gz 2022-03-09 16:48:50 wget is fine 2022-03-09 16:49:12 When I wget ^^; apk update does not load anything into /var/cache/apk 2022-03-09 16:49:20 The sha512sum on the builder is 670659f90a32872e17c043a2458354a0cb9072bd58752123eae194ba09b86965e0b132fd9d8b1633bbd0d779d6371d2937956512ed83995d57c8557 2022-03-09 16:49:51 sorry, I was on community for that one 2022-03-09 16:50:46 3.15/master: fc9e74818c1a309262bb7fa6fa6bfefc551fe187b04907d485a6b118fb4069ffeb6baa0a3d23f9a6c1c9620a2514f2fd3c24f661d00fd45f0404ed3e9bc16da1 2022-03-09 16:50:56 so that corresponds 2022-03-09 16:55:45 Maybe if it's helpful, it's a fresh installation. Logged in via root, entered the second parition of sd-card into /etc/fstab with defaults 0 0, did setup-alpine, set the cache dir to mmcblk0p2 did lbu commit -d rebooted, setup openntpd lbu commit -d again, tried to install doas, failed, started to investigate and noticed apk update is not running correctly 2022-03-09 16:58:24 is mmcblk0p2 mounted? 2022-03-09 16:59:00 ok, to American english, that sounds bad. Sorry. 2022-03-09 17:00:13 /dev/mmcblk0p2 on /media/mmcblk0p2 type ext4 (rw,relatime) 2022-03-09 17:00:22 so yes it's mounted 2022-03-09 17:00:28 and rw. 2022-03-09 17:01:19 oh... 2022-03-09 17:01:43 cache dir should include the /media/ prefix 2022-03-09 17:02:50 Link seems correct 2022-03-09 17:02:52 lrwxrwxrwx 1 root root 22 Mar 9 18:02 cache -> /media/mmcblk0p2/cache 2022-03-09 17:03:13 did you create a cache directory on mmblk0p2? 2022-03-09 17:03:36 No apk did I think 2022-03-09 17:03:45 But it's there 2022-03-09 17:05:03 And it's mode 755 and owned by root which looks right to me? 2022-03-09 17:05:12 yes 2022-03-09 17:08:38 You say /var/cache/apk dir is empty? 2022-03-09 17:09:09 Yes 2022-03-09 17:12:36 That is beyond weird. ok... I don't think this will work... but.... 2022-03-09 17:12:48 #cd /var/cache/apk 2022-03-09 17:13:04 #wget https://dl-cdn.alpinelinux.org/alpine/v3.15/main/armv7/APKINDEX.tar.gz 2022-03-09 17:13:12 And let's try again 2022-03-09 17:14:30 Did that earlier, apk does not seem to care about it... Ah damn... just noticed a bunch of EXT4-fs errors on mmcblk0p2. When I unmount the partition `apk update` runs successful 2022-03-09 17:15:17 I feel for you. Been there more times than I want to admit! 2022-03-09 17:16:21 I think the filesystem is corrupt or maybe the sdcard is broken. Thank you all for your help ^^ I'll check if reformatting or maybe a different sdcard will help 2022-03-09 17:18:19 Blechdose: TIL another debugging step. Thank you. 2022-03-09 17:19:16 nangel: will apk use a plain APKINDEX.tar.gz file in /var/cache/apk? 2022-03-09 17:19:24 Normally it has some kind of suffix 2022-03-09 17:20:05 ikke: no, it should not. Which is why I said "I don't think this will work." :) 2022-03-09 17:20:13 ah, sorry, missed that 2022-03-09 17:20:38 hi, is there an easy way to pull in a single kernel module (plus maybe deps?) from linux-lts package for a VM running linux-virt kernel? 2022-03-09 17:20:58 basically I wanted to use pppoe.ko in a VM, but that's only available in linux-lts, not linux-virt 2022-03-09 17:21:17 In American Football, it is called "The Hail Mary" play... you throw the ball as far as you can, and hope your team catches it. With very little sense of success. 2022-03-09 17:21:50 crono_log: not sure if you can use modules not comming from that kernel 2022-03-09 17:24:57 ikke: ok I think you're right, I asked too soon whilst still researching, I think vermagic check will prevent me doing that anyway 2022-03-09 17:25:25 guess I'll have to use more heavyweight linux-lts for the specific VMs that need pppoe module 2022-03-09 17:25:48 crono_log: you can add linux-firmware-none to prevent firmware being pulled in 2022-03-09 17:26:19 that sounds good, thanks 2022-03-09 18:27:27 How I can disable creation of swap partition during alpine-installation? 2022-03-09 18:29:10 -s 0 to setup-disk, or SWAP_SIZE=0 in env should work 2022-03-09 18:29:13 probably, maybe 2022-03-09 18:31:34 If I use setup-disk, I cannot use setup-alpine, right? 2022-03-09 18:34:03 you can ignore vermagic with insmod -f but it still won't work due to missing symbols 2022-03-09 18:34:54 Ermine: technically not, but the env var should work with either 2022-03-09 18:39:53 psykose: that worked, thanks! 2022-03-09 19:49:28 psykose: Do I need to manually bump pkgrel for geotagging since it depends upon ffmpeg4-libs now (and I think it should build just fine against ffmpeg5/ffmpeg)? 2022-03-09 19:49:45 geotagging? 2022-03-09 19:50:29 ah i see 2022-03-09 19:50:36 it doesn't depend on ffmpeg-libs at all 2022-03-09 19:50:49 Oh... what. 2022-03-09 19:51:11 I sudo apk del ffmpeg4-libs and geotagging is one of the ones holding that package back from being deleted 2022-03-09 19:51:14 It must be another dependency 2022-03-09 19:51:32 that doesn't show them correctly in my experience 2022-03-09 19:51:54 and it won't be deleted for a while 2022-03-09 19:52:09 most things depend on ffmpeg4-libs 2022-03-09 19:52:13 Yeah, I think I read you were rebuilding things against the newer ffmpeg5 2022-03-09 19:52:49 i skipped something completely broken, and chromium (which works with 5, but will just wait for next release) 2022-03-09 19:52:56 everything else needs patches or just doesn't support it 2022-03-09 20:03:30 Normal sized mess then? haha 2022-03-09 20:04:04 not really a mess 2022-03-09 20:04:06 :) 2022-03-09 22:40:03 Just got to my hotel room 2022-03-09 22:40:46 disregard previous message sent by mistake :/ 2022-03-09 23:12:57 Hello. How do I bypass any and all partitioning and boot-loader installation in setup-disk? Essentially it should just copy over files and set appropriate permissions. 2022-03-09 23:18:21 setup-disk /mnt/mydisk 2022-03-09 23:18:29 if it's a mountpoint it doesn't do partitioning 2022-03-09 23:18:39 (note the absolute path, relative won't work) 2022-03-09 23:20:37 Thanks for clarifying on partitioning. How does one bypassing bootloader installation? 2022-03-09 23:20:51 s/ing/ 2022-03-09 23:20:51 XFaCE2 meant to say: Thanks for clarify on partitioning. How does one bypassing bootloader installation? 2022-03-09 23:21:11 That is not what I meant but OK. 2022-03-09 23:23:11 From reading the documentation, the only obvious way to do this is not installing any of the bootloader tools thus failing those parts of the installation... seems very crude. 2022-03-10 00:54:27 ACTION goes to burgle zcrayfish's house 2022-03-10 00:55:06 i guess that assumes zcrayfish lives alone 2022-03-10 00:55:29 haha 2022-03-10 01:55:18 anyone running alpine on btrfs? 2022-03-10 01:57:54 probably someone is 2022-03-10 01:58:39 if you have a problem or question, just ask it straight away 2022-03-10 01:58:39 I swear I had it working at some time in the past but I've been having issues installing it on an old macbook this go around 2022-03-10 02:01:10 my question is as stated above, I can do xfs + luks no problem, but btrfs + luks fails, even following the wiki + guides etc. installing to plain btrfs fails as well which is why I ask if anyone is running it. 2022-03-10 06:12:21 ACTION how to compile with out reverse engineering , first you should encrypt your source code , than encrypt it again and encrypt the whole config again so is 3 time encryption , than encrypt again the config file or script , modify your compiler to be the same encryption without decrypting it 2022-03-10 07:27:14 ok 2022-03-10 08:49:53 atka: "fails" is extremely vague, which is probably why you're not getting much of response. Are there some error messages or symptoms you can show us? 2022-03-10 10:44:34 Another day another qemu sway day. 2022-03-10 10:44:44 I can't seem to fire up sway on intel 2022-03-10 10:46:21 I've install intel grapgic card, tried all 3 options from: https://wiki.alpinelinux.org/wiki/Sway#Prerequisites seatd, elogin... but when I start it like this: DG_RUNTIME_DIR=~/.run dbus-launch sway I get: Unable to receive IPC response 2022-03-10 10:51:12 This is debug log: https://termbin.com/6qiv 2022-03-10 11:35:22 anyone has libreoffice working on alpine edge? I get terminate called after throwing an instance of 'com::sun::star::uno::RuntimeException' 2022-03-10 11:36:18 frojnd: maybe is related wth 00:00:00.071 [ERROR] [wlr] [render/egl.c:281] Software rendering detected, please use the WLR_RENDERER_ALLOW_SOFTWARE environment variable to proceed 2022-03-10 12:04:58 wow, I'm just testing gnumeric. Seems nice and faster than libreoffie-calc 2022-03-10 12:17:31 donoban: even if it's related to that env, not sure what I should set it to 2022-03-10 12:19:34 try to set to 1 2022-03-10 12:20:29 https://www.reddit.com/r/swaywm/comments/my67f6/sway_16_and_headless/ 2022-03-10 12:20:29 [REDDIT] Sway 1.6 and headless (self.swaywm) | 9 points (03100.0%) | 6 comments | Posted by power461 | Created at 2021-04-25 - 11:13:06UTC 2022-03-10 12:21:49 it doesn't seem same problem but he tries with WLR_RENDERER_ALLOW_SOFTWARE=1 2022-03-10 12:29:00 donoban: well can't believe it 2022-03-10 12:29:05 setting to 1 started sway 2022-03-10 12:29:20 Thank you 2022-03-10 12:29:22 Interesting 2022-03-10 12:34:16 it seems that they forced it to fail because many people complained about it being slow 2022-03-10 12:42:05 Hm now I don't seem I can do mauch with qemu and sway 2022-03-10 12:42:49 Like mod+enter opens, my host's terminal not guets 2022-03-10 12:50:52 hm, I have no idea 2022-03-10 13:03:55 I'm having issues booting the lx2k clearfog board with the alpine 3.15 iso 2022-03-10 13:04:04 and my install I upgraded from 3.14 to 3.15 also no longer boots 2022-03-10 13:04:13 there's just nothing after grub 2022-03-10 13:05:53 3.14 works, 3.15 does not? 2022-03-10 13:05:56 yes 2022-03-10 13:06:10 if I remove quiet from the kernel cmdline I get a lot of these errors where it hangs: https://paste.sr.ht/~martijnbraam/7d304ab88913e6e164810a5f4830b6dbf7f7fef6 2022-03-10 13:06:21 I have similar with VIA x64 motherboards. Something changed in the kernel 2022-03-10 13:06:59 I thought I just did something weird in my install, but the same behavior on the official isos so I guess something is really broken 2022-03-10 13:08:39 I know Ariadne has worked with this hardware, hopefully I can get some brilliant insights? :D 2022-03-10 13:08:50 agreed. For me on VIA x86_64 chips, its a kernel panic. No time to dig into it (need to pay for food and rent!) but yes, something changed in 3.15 2022-03-10 13:09:15 martijnbraam: iommu=passthrough arm-smmu.disable_bypass=1 2022-03-10 13:09:56 martijnbraam: its the price of being able to use the NIC unfortunately :( 2022-03-10 13:12:23 hmm still getting the same error 2022-03-10 13:13:15 hmm 2022-03-10 13:13:19 arm-smmu.disable_bypass=0 2022-03-10 13:13:26 sorry, i have not had my coffee yet 2022-03-10 13:13:45 you need the iommu=passthrough too 2022-03-10 13:16:18 also not changing anything 2022-03-10 13:16:41 I guess I should try it on the installer kernel 2022-03-10 13:17:37 nope, also not working on the iso 2022-03-10 13:22:51 martijnbraam: iommu.passthrough=1 2022-03-10 13:23:03 martijnbraam: with the arm-smmu.disable_bypass=0 2022-03-10 13:37:57 \o/ it boots 2022-03-10 13:38:03 thanks Ariadne 2022-03-10 15:24:36 Hey all. I hope everyone is doing well. I'm in need of a bit of help. I just booted up an old PC with a standard alpine linux CD-ROM into it, but once the loading messages get to "loading hardware drivers" the screen goes blank. I've tried with a different monitor. Any ideas what could be going on and what I could do to troubleshoot? Thanks in advance for any help. 2022-03-10 15:36:17 What gpu do you have? 2022-03-10 16:03:08 it's part of the motherboard. It's a matrox G200eW 16MB DDR2 2022-03-10 16:43:00 I ran pip install -U yt-dlp and got a bunch of error messages after Running setup.py install for brotli ... 2022-03-10 16:43:08 anyone else have this problem 2022-03-10 16:43:41 bleb: any reason you did not install yt-dlp from the repository? 2022-03-10 16:44:08 https://pkgs.alpinelinux.org/package/edge/community/x86_64/yt-dlp 2022-03-10 16:44:24 I thought it wouldn't be as up to date 2022-03-10 16:45:38 fishsemxpinha: try adding nomodeset to kernel command line 2022-03-10 16:48:24 it said it was falling back on the legacy setup.py because wheel is not installed 2022-03-10 16:48:39 so I did pip install -U wheel 2022-03-10 16:49:00 but then I get a bunch of errors after Building wheel for brotli (setup.py) ... 2022-03-10 16:49:36 hmm python/_brotli.cc:2:10: fatal error: Python.h: No such file or directory 2022-03-10 16:51:32 Hello71: This is before installation. I'd have to create a custom bootable version? 2022-03-10 17:02:16 ok apk add python3-dev fixed it 2022-03-10 17:12:20 Hello71: I didn't know there was a boot manager coming up. Your suggestion worked. As the "boot:" line is displayed I typed in "lts nomodeset" and it worked. Where is the best place to document this? 2022-03-10 17:24:32 fishsemxpinha: afaik you should be able to press esc when bootloader shows 'boot:' message to edit boot parameters 2022-03-10 17:34:54 Yeah, and then you can use tab so it shows you the name of kernel, and then you enter the kernel name and 'nomodeset' parameter 2022-03-10 18:13:54 awesome Ermine. Many thanks! 2022-03-10 18:16:13 I'm trying to launch Alpine 3.15 in secure boot environment. I have my own secure boot key infrastructure, and I signed bootloader with my key. But I cannot boot alpine: grub says 'shim_lock protocol not found'. What's wrong? 2022-03-10 18:17:22 Should I sign the kernel as well? 2022-03-10 18:19:20 From what I know, secure boot is for booting only signed efi applications, such as bootloader, and then secure boot's role ends 2022-03-10 18:20:15 The kernel doesn't have to be signed 2022-03-10 18:53:53 Hmm 🤔 I too am interested in what you find out. I'd like to keep SecureBoot on for my installs 2022-03-10 19:41:40 I put rustup in makedepends, but then I get rustup: not found when I try to use it in the apkbuild 2022-03-10 19:42:14 j`ey: Do you have a community repo in /etc/apk/repositories? 2022-03-10 19:42:26 And are you on either x86_64 or aarch64? 2022-03-10 19:42:37 I assume so, since I get (1/12) Installing rustup (1.24.3-r1) 2022-03-10 19:42:41 x86_64 2022-03-10 19:43:06 the binary is called rustup-init 2022-03-10 19:43:08 https://pkgs.alpinelinux.org/contents?branch=edge&name=rustup&arch=x86_64&repo=community 2022-03-10 19:43:38 ohh! 2022-03-10 19:44:23 thanks 2022-03-10 19:44:32 np 2022-03-10 19:47:00 using rustup is probably not The Right way of doing this 2022-03-10 22:05:35 Hello! Has anybody been able to get GLX running inside a virtual machine (I'm using VirtManager) running Alpine Linux? 2022-03-10 22:10:51 Thanks for clarifying on partitioning. How does one bypass bootloader installation? 2022-03-10 22:11:10 Apparently modifying the setup scripts. 2022-03-10 22:14:23 I will see if the YOLO approach works. 2022-03-10 22:25:05 technically GLX works anywhere, as long as you count software rendering 2022-03-10 22:26:26 Hello71: ah, thanks. I didn't know that - interesting I could use it with virtio selected in the video tab 2022-03-10 22:26:46 thanks and goodbye! :) 2022-03-11 05:55:42 Anyone have the energy to help me debug a broken APKBUILD again? 2022-03-11 05:55:59 ok... 2022-03-11 05:55:59 Project is GPSBabel: https://github.com/GPSBabel/gpsbabel 2022-03-11 05:56:23 APKBUILD... (full message at https://matrix.org/_matrix/media/r0/download/matrix.org/JsxLtSuNcHpIvHFspDUUcwfP) 2022-03-11 05:56:51 Error:... (full message at https://matrix.org/_matrix/media/r0/download/matrix.org/EjaVJtTYVSTVfcaAPCfokGyQ) 2022-03-11 05:57:13 Is this a pkgconf issue, or did I do something wrong? libusb-dev is in makedepends, and I checked that it provides the needed libusb.h 2022-03-11 05:58:51 Is it because I didn't set the qmake options right (or at all)? 2022-03-11 05:58:52 https://github.com/GPSBabel/gpsbabel/blob/master/INSTALL#L11 2022-03-11 06:00:09 Saijin_Naib: It says that "included" is "default, macOS only", so this either means that there's no default on other systems or that is the default but it breaks on other systems 2022-03-11 06:00:23 I suggest setting it to "pkgconfig" 2022-03-11 06:00:48 What would that look like for qmake? 2022-03-11 06:00:51 Does that go into prepare? 2022-03-11 06:03:43 Saijin_Naib: I think adding "WITH_LIBUSB=pkgconfig" after the "qmake" command in build (in between "qmake" and "-makefile") should work 2022-03-11 06:04:01 Oooh, thank you. Let me see what that brings 2022-03-11 06:17:18 More pain haha... It can't find libusb.h no matter what it seems. So odd 2022-03-11 10:25:49 so much spam 2022-03-11 10:26:24 someone may put a temporary +b on endigma? 2022-03-11 10:43:24 yes that'd be lovely... 2022-03-11 11:51:42 I have an old mini computer with NVIDIA Corporation C79 [ION] (rev b1) I've installed mesa-dri-nouveau My sway keeps on "hanging" while Firefox is opened. What else should I install to use firefox for browsing? (only one tab though -- didn't even start browsing yet) It might be also alacritty has some issues 2022-03-11 11:52:46 you can try mesa-dri-gallium instead if the gpu supports it 2022-03-11 11:53:23 which is nouveau instead of nouveau_vieux 2022-03-11 11:54:10 Ok let me remove mesa-dri-nouveau 2022-03-11 11:54:50 aside from that, no idea 2022-03-11 12:03:34 Hm I thought, firefox was wayland 2022-03-11 12:03:52 Looking at sway log I see these: 00:03:11.405 [DEBUG] [wlr] [xwayland/xwm.c:796] unhandled X11 property 338 (_NET_WM_OPAQUE_REGION) for window 6291477 2022-03-11 12:03:58 While focus on Firefox 2022-03-11 12:04:14 you need to enable wayland explicitly for firefox 2022-03-11 12:04:59 00:04:17.758 [ERROR] [wlr] [libinput] event0 - Logitech USB Keyboard: client bug: event processing lagging behind by 18ms, your system is too slow :D 2022-03-11 12:05:23 Ermine: ok let me try that if it's gonna be better 2022-03-11 12:05:32 MOZ_ENABLE_WAYLAND=1 2022-03-11 12:05:48 usually the system firefox launcher checks for wayland display and sets it itself 2022-03-11 12:06:05 but on alpine this is uh 2022-03-11 12:06:07 not a script 2022-03-11 12:06:21 gotta do it yourself :) 2022-03-11 12:06:42 I added it to ~/.profile 2022-03-11 12:06:47 yep 2022-03-11 12:06:48 But how do I really know if it's on wayland? 2022-03-11 12:06:52 After launching ff 2022-03-11 12:06:55 swaymsg -t get_tree 2022-03-11 12:07:01 it will show up with an app_id in the tree 2022-03-11 12:07:08 xwayland things don't have one 2022-03-11 12:07:11 and will show xwayland too 2022-03-11 12:07:13 or whatever 2022-03-11 12:07:18 Ok thank you 2022-03-11 12:07:25 there's probably easier ways without 10 screens of text, but i always use that 2022-03-11 12:10:15 Hm 2022-03-11 12:10:24 I see Mozilla restore there but app_id: null 2022-03-11 12:10:45 strange 2022-03-11 12:10:52 but if you see no xwayland it's fine 2022-03-11 12:11:04 under shell: 2022-03-11 12:11:08 and they get no appid at all 2022-03-11 12:11:09 not even null 2022-03-11 12:12:15 I do see it :/ 2022-03-11 12:12:20 Under shell: xwayland 2022-03-11 12:12:26 So It didn't read my profile correctly 2022-03-11 12:12:27 are you sure it's in the env 2022-03-11 12:12:37 Because I didn't source ~/.profile before starting ff ;/ 2022-03-11 12:12:40 .profile doesn't get reread until you relogin 2022-03-11 12:12:41 haha 2022-03-11 12:12:49 Or source it 2022-03-11 12:12:58 I thought source would do it 2022-03-11 12:13:09 not outside the shell running it 2022-03-11 12:13:40 source ~/.profile 2022-03-11 12:18:19 Well apperantly even after logout/login env wasn't applied 2022-03-11 12:18:56 But I had another ssh session... maybe that was the reason ;/ 2022-03-11 12:19:09 Much better with wayland enabled 2022-03-11 12:19:14 Now I can use one tab! 2022-03-11 12:24:00 Well mozilla ignores .profile and & .bashrc 2022-03-11 12:24:07 Only if I start within terminal 2022-03-11 12:24:17 I made sure env is exported which is weird 2022-03-11 12:27:11 every application on your system ignores .profile, your login shell sources it 2022-03-11 12:27:44 if the env variable is defined at the point of starting sway, then it will be visible to everything 2022-03-11 12:28:05 if not, then sway won't see it, and so will nothing started outside whatever shell you source things in manually 2022-03-11 12:28:27 You can add env variable to firefox's .desktop file 2022-03-11 12:28:34 that also works 2022-03-11 12:29:19 I have problem with my profile. 2022-03-11 12:29:37 Before starting sway, only XDG_RUNTIME_DIR=/tmp works 2022-03-11 12:29:51 WHen I echo it, but MOZ_ENABLE_WAYLAND=1 won't echo 1 2022-03-11 12:32:40 Ok finally I made .profile work 2022-03-11 13:03:21 I guess I must use netsurf 2022-03-11 13:03:47 Firefox clogs system even when running wayland 2022-03-11 13:05:16 it is a very old gpu indeed 2022-03-11 13:07:03 Hm. I installed numix-themes-gtk3 package but not sure how to "enable" it? 2022-03-11 13:14:48 usually an env variable 2022-03-11 13:15:10 GTK_THEME was gtk2 only i think, not sure 2022-03-11 13:15:35 gsettings set org.gnome.desktop.interface gtk-theme '' 2022-03-11 13:15:38 should work for gtk3 2022-03-11 13:17:17 with the theme name in the quotes 2022-03-11 13:19:45 psykose: you sad as an env, so this should be: GTK_THEME="numix-themes-gtk3" in my ~/.profile ? 2022-03-11 13:20:08 the name of the theme is just 'Numix' pretty sure 2022-03-11 13:20:20 and sure, dunno if that works for gtk3 2022-03-11 13:20:34 the second thing does, but i think you need to run that after you have a graphical session or something 2022-03-11 13:20:51 there is also ~/.config/gtk3 or ~/.gtk3 that can have a config i think 2022-03-11 13:20:52 been a while 2022-03-11 13:21:23 I just have ~/.config/gtk-3.0/ 2022-03-11 13:21:46 sure 2022-03-11 13:24:20 GTK_THEME is gtk3 or later only, but you shouldn't normally set it 2022-03-11 13:24:33 use gsettings if you have it, or ~/.gtkrc-2.0 + ~/.config/gtk-3.0/settings.ini otherwise 2022-03-11 13:24:38 But I should install some gtk3 theme and it should be applied by default? 2022-03-11 13:24:44 I don't have it 2022-03-11 13:24:54 Ok settings.ini then 2022-03-11 13:24:55 you can create files, you know 2022-03-11 13:25:39 in X11, xsettingsd could also be used to "publish" the theme name to both gtk2 and gtk3 (just like gnome-settings-daemon would in full GNOME), but there's no wayland equivalent 2022-03-11 14:22:33 I would like to have this app on Alpine: https://github.com/iamtalhaasghar/yewtube When trying to install it with pip install yt I get: https://termbin.com/nqtq I see gcc inthere 2022-03-11 14:23:04 So I think it won't isntall using pip because it uses some make files to install it but agains glibc, which is not supported under alpine 2022-03-11 14:29:01 error: command 'gcc' failed: No such file or directory <- this is a good indicator that app is build a gains gcc and not musl? 2022-03-11 14:31:36 don't confuse gcc (compiler) with glibc (runtime library) 2022-03-11 14:35:45 you just need to apk add gcc 2022-03-11 14:35:49 and probably python3-dev 2022-03-11 14:43:43 Adding build-base did the job. Thank you 2022-03-12 05:43:36 I keep getting error[E0463]: can't find crate for `core` 2022-03-12 05:43:51 and error[E0463]: can't find crate for `std` 2022-03-12 05:44:01 while compiling a rust package. 2022-03-12 05:44:15 any idea which alpine package has those crates> 2022-03-12 05:51:18 what rust package 2022-03-12 05:51:42 the answer is none of them, but this depends on the target 2022-03-12 05:52:44 psykose: https://github.com/waycrate/swhkd/blob/main/Makefile#L47 2022-03-12 05:53:10 don't call make setup 2022-03-12 05:53:20 ya, I didnt 2022-03-12 05:53:28 this makefile will never work actually 2022-03-12 05:53:33 Im not familiar with rust but when I read what rustup does.... 2022-03-12 05:53:35 lmfao 2022-03-12 05:53:39 alpines triple is arch-alpine-linux-musl 2022-03-12 05:54:51 cargo build in the root should work 2022-03-12 05:55:18 psykose: if I just run cargo build in the root, I get those two errors 2022-03-12 05:55:29 `type -a cargo` 2022-03-12 05:56:02 why is there a cargo in my HOME 2022-03-12 05:56:07 rustup put it there 2022-03-12 05:56:14 remove that from your path and it will work 2022-03-12 05:56:45 in this case though i would just say it's alpine that is wrong, and we should not have a custom triple for rust really, but idk 2022-03-12 05:57:04 it breaks random things sometimes that do triple checking 2022-03-12 05:57:21 but yeah, it builds fine with regular alpine packaged rust 2022-03-12 05:58:15 psykose: error[E0463]: can't find crate for `core` 2022-03-12 05:58:27 and same for crate for `std` 2022-03-12 05:58:37 are you sure it's the system rust 2022-03-12 06:00:40 if i make a container and just apk add cargo alpine-sdk and then git clone the repo it works 2022-03-12 06:01:07 so you probably did not delete ~/.cargo/bin 2022-03-12 06:01:26 psykose: I dont have .cargo/bin 2022-03-12 06:01:56 i rm -rf'ed the directory that returned when I typed in type -a cargo 2022-03-12 06:02:34 probably something else still hanging around somewhere 2022-03-12 06:04:27 alright lets see what the CI has to say 2022-03-12 06:04:29 thanks psykose 2022-03-12 06:04:34 you're always a great help 2022-03-12 06:04:53 Im not the most familiar with the toolchain 2022-03-12 06:05:53 this makefile is never going to work 2022-03-12 06:06:13 if that's what you were running then that is why it can't find the crate 2022-03-12 06:06:16 the triple doesn't exist 2022-03-12 06:06:21 you can only use cargo build 2022-03-12 06:06:36 ohhhhhhh 2022-03-12 06:06:41 I cant make build eh 2022-03-12 06:06:47 yes 2022-03-12 06:06:52 it passes *-unknown-* 2022-03-12 06:07:03 but if you look at the files of rust-stdlib, they're all *-alpine-* 2022-03-12 06:07:46 it also hardcodes x86_64 2022-03-12 06:08:04 i would just call cargo build in the usual fashion and look at what files to install 2022-03-12 06:08:09 ya cargo build works..... 2022-03-12 06:08:23 should I send a patch upstream to fix their makefile? 2022-03-12 06:08:36 shrug 2022-03-12 06:08:57 all they would have to do is remove --target entirely 2022-03-12 06:09:16 you can patch that in the meantime if you want to use it 2022-03-12 06:09:35 you also need !s390x !riscv64 for rust 2022-03-12 06:09:53 and to add a `cargo fetch --locked` in prepare() 2022-03-12 06:10:03 and then --frozen instead in build instead of locked 2022-03-12 06:11:06 ok Ill do that after I fix the checking 2022-03-12 06:11:19 and no -dev/-doc subpackage 2022-03-12 06:11:36 psykose: do you know how I can get cargo fmt? 2022-03-12 06:11:40 what package provides that? 2022-03-12 06:11:41 rustfmt 2022-03-12 06:11:56 (edge) 2022-03-12 06:13:05 crap 2022-03-12 06:13:14 Im on stablke 2022-03-12 06:13:23 i only added it after 3.15 sadly 2022-03-12 06:14:08 ok 2022-03-12 06:14:13 Ill upload what I have so far 2022-03-12 06:14:16 I gotta go to sleep 2022-03-12 06:15:00 what's the point of running fmt as a check step 2022-03-12 06:15:09 it just formats the source files and doesn't do anything else 2022-03-12 06:15:16 oh true 2022-03-12 06:15:22 I was blindly following the makefile 2022-03-12 06:15:24 nice catch 2022-03-12 06:15:32 you also missed --release on the build 2022-03-12 06:15:39 and --frozen on the check 2022-03-12 06:17:01 also their check is wrong 2022-03-12 06:17:04 cargo test runs the tests 2022-03-12 06:17:07 the check doesn't do anything 2022-03-12 06:17:25 that's just a fast compile check when developing :) 2022-03-12 06:17:37 gah 2022-03-12 06:17:42 I spent so much time getting it working 2022-03-12 06:17:46 haha 2022-03-12 06:18:10 well, atleast my package is future proof 2022-03-12 06:18:11 also the install won't work 2022-03-12 06:18:13 for when they add tests 2022-03-12 06:18:18 because it relies on it being in ./bin/binary 2022-03-12 06:18:19 ya, I just saw it fail lmfao 2022-03-12 06:18:33 manually write the installs??? 2022-03-12 06:18:34 =) 2022-03-12 06:18:38 target/release/binaryname 2022-03-12 06:18:41 is the default spot 2022-03-12 06:19:20 they only had it changed because if you add --target it adds that target name in the middle too 2022-03-12 06:19:55 should I install the polkit rule too? 2022-03-12 06:19:59 yeah 2022-03-12 06:20:25 docs/swhkdrc has a sample config to put into /etc too 2022-03-12 06:20:28 that is not normally in the install 2022-03-12 06:20:33 or something 2022-03-12 06:20:46 though i guess it is pre-filled, so maybe could also be left out 2022-03-12 06:20:47 up to you 2022-03-12 06:37:47 nice, I got it working. Thanks again psykose 2022-03-12 06:38:02 now for bed. Ill make any changes if needed after bed 2022-03-12 06:38:13 :) 2022-03-12 06:38:28 it should be cargo test not check 2022-03-12 06:39:29 also the license is BSD-2-Clause 2022-03-12 20:56:28 I finally managed to start Sway 1.6 successfully under qemu with my intel integrated graphic card. Now I have the issue of keyboard presses not being captured by guest alpine sway... when I do mod+2 it actually switches host's workspace. Any ideas? 2022-03-12 20:59:34 Mouse seems to be captured by Alpine's Sway in Qemu but not keyboard 2022-03-12 21:10:03 Ah grab input 2022-03-12 21:59:18 I'm having trouble getting alpine to use ethernet on a raspberry pi after ipxe booting. 2022-03-12 21:59:44 Closest thing online I could find about it was https://bugzilla.redhat.com/show_bug.cgi?id=1943983 2022-03-12 22:01:16 And there doesn't seem to be a /lib/modules/.../mdio_bcm_unimac.ko when booting alpine's arm64 netboot. 2022-03-12 23:39:23 Hello everyone. 2022-03-13 01:13:20 Does the setup program install grub on uefi systems? 2022-03-13 01:22:37 yes, it's the preferred bootloader 2022-03-13 01:24:56 ptrc: Oh alright, thank you. 2022-03-13 11:42:37 Hello guys. Can you see my messages? I haven't used irc before. 2022-03-13 11:43:01 boombim: hello, I can see 2022-03-13 11:44:51 Cool. Do you guys know is it worth to use udhcpd instead of full dhcpd? I can't connect to my iPhone's hotspot and it dramatically increase boot time 2022-03-13 11:45:19 Dhcpd fixed this issues but I haven't tried to configure udhcp 2022-03-13 11:48:09 what's the reason for considering something else? 2022-03-13 11:48:29 you can configure your inferface via /etc/network/interfaces to use dhcp, and then udhcpd would be launched automatically 2022-03-13 11:49:11 Also you can try to configure your interface statically to save boot time 2022-03-13 12:00:04 lopid: just to use busybox utils and not install external tool 2022-03-13 13:44:17 hm: ERROR: busybox-1.35.0-r5: failed to commit bin/busybox: Permission denied 2022-03-13 13:44:39 trying to use apk with --root 2022-03-13 13:46:47 not as root? 2022-03-13 13:47:05 nope 2022-03-13 13:47:23 thankfully, that message looks a bit scary! 2022-03-13 14:04:00 how can I properly untar/retar the minimal rootfs image, to keep the permissions/uids etc 2022-03-13 14:41:45 tar -x -p --numeric-owner doesnt seem to do it 2022-03-13 14:53:58 scripts/mkimage.sh --profile minirootfs --arch aarch64 --hostkeys --repository 2022-03-13 14:54:11 http://dl-cdn.alpinelinux.org/alpine/edge/main --outdir tmproot, gives me an error about untrusted repos. 2022-03-13 15:04:05 ok not passing arch works, so its an issue with cross-compilation 2022-03-13 15:14:13 ok apparently you can't do a cross mkimage, making the --arch param confusing https://lists.alpinelinux.org/~alpine/devel/%3CCABErt4ekLFSqQSwUDzBrSmYPJ6aq2f1GaCoO-3cWynayKNNNVw%40mail.gmail.com%3E 2022-03-13 17:14:50 Hi! The wiki tells me this channel is okay for quick support questions... Can I ask a quick support question? 2022-03-13 17:16:59 In case anyone sees it: Wondering how to troubleshoot getting the Discord web app to recognize my microphone in Firefox. I can record and playback audio through my only recording device (built-in laptop mic) using `arecord`, but Discord web-app doesn't pick up any input from the mic. Audio output in the browser (including Discord) is fine. Not even sure where to start and it's 2022-03-13 17:17:01 been surprisingly tough to Google. 2022-03-13 17:17:36 how are you invoking arecord 2022-03-13 17:17:54 and also have you tried any other webrtc instances 2022-03-13 17:18:24 1. `arecord -f S16_LE -d 10 -c 2 --device="hw:0,0" ./test-mic.wav` 2022-03-13 17:18:52 2. I haven't -- is there another I could easily try? 2022-03-13 17:19:33 Ah I guess zoom uses WebRTC too I think 2022-03-13 17:48:04 What sound server are you using? Once I got pipewire and mediaplumber setup (with alsa, pulse, and jack support) things have (thankfully) Just Worked (TM) thus far 2022-03-13 19:32:19 why are you passing so many arecord arguments 2022-03-13 19:32:57 arecord -fdat -d10 test.wav 2022-03-13 19:33:41 most applications will only use the default device 2022-03-13 19:33:51 almost no applications will ever use raw hardware devices (hw:x,y) 2022-03-13 19:54:02 Ah, nice. I haven't set up Pipewire yet... I suppose I should try that. 2022-03-13 19:54:13 Anyway, need to log off for awhile. Thanks for your responses :) 2022-03-13 21:06:33 Thanks for clarifying on partitioning. How does one bypass bootloader installation? 2022-03-13 21:06:34 I will see if the YOLO approach works. 2022-03-13 21:06:34 Apparently modifying the setup scripts. 2022-03-13 21:07:02 Nope, time to continue the arduous task of mutilating this poor script 2022-03-14 07:13:55 hello, I want to package metasploit-framework. It has numerous ruby dependencies https://github.com/rapid7/metasploit-framework/blob/master/Gemfile.lock 2022-03-14 07:14:35 I found an example here on wiki https://wiki.alpinelinux.org/wiki/APKBUILD_examples:Ruby but it seems that it's quite old 2022-03-14 07:15:55 What is our policy on including ruby dependencies? Do we add all ruby dependencies under rube-* namespace, or do we just use language's dependency manager while building the package? 2022-03-14 07:17:18 i.e Do I have to package said ruby dependencies as separate alpine package each? 2022-03-14 07:18:46 given the number of hits on 'gem install --local' in the package() function of any current package I'd guess it's safe to just go ahead and install everything in the final package 2022-03-14 07:19:21 (it does have --ignore-dependencies though) 2022-03-14 07:20:38 (so I once again talked too fast and things look like properly packaged into individual packages, and will just go back to lurking) 2022-03-14 07:22:37 Yes, I did the same, and am confused as well. 2022-03-14 08:21:51 jw.org 2022-03-14 08:22:12 sorry, wrong window 2022-03-14 08:57:52 hello ppl 2022-03-14 08:58:21 I have an hardware issue with a Ryzen server 2022-03-14 08:58:34 it takes forever toboot with Alpine 3.15 2022-03-14 08:59:00 I get the following errors (dmesg reports) 2022-03-14 08:59:04 [ 65.190144] udevd[1835]: worker [1892] /devices/LNXSYSTM:00/LNXSYBUS:00/PNP0A08:00/device:34/IPI0001:00 is taking a long time [ 65.190164] udevd[1835]: worker [1898] /devices/pci0000:00/0000:00:14.3/IPI0001:00 is taking a long time [ 114.556784] ipmi_si IPI0001:00: There appears to be no BMC at this location [ 114.558086] ipmi_ssif: IPMI SSIF Interface driver 2022-03-14 08:59:46 should I blacklist ipmi drivers ? Or there's a way to tell udevd to ignore ipmi completely? 2022-03-14 09:58:48 I'm trying to build visurf: https://sr.ht/~sircmpwn/visurf/ I've installed build-base and bison package, yet I get following error: src/filter-parser.y:15:10: fatal error: filter-lexer.h: No such file or directory https://termbin.com/chuj Any dieas what I'm doing wrong? 2022-03-14 10:06:20 Not sure what should provide that file 2022-03-14 14:05:02 your link does not show any error? 2022-03-14 15:05:48 why might `su` not work when in sway, but it's fine at the console? 2022-03-14 15:05:54 just says 'incorrect password' 2022-03-14 16:59:59 hello 2022-03-14 17:01:54 I'm rather fresh and new to alpine linux, advanced user here and I think I'll stay with you for a while (or much longer) :) 2022-03-14 17:02:31 satmd: welcome 2022-03-14 17:30:08 Which package does contain 'w' binary? 2022-03-14 17:31:15 https://pkgs.alpinelinux.org/contents?file=w&path=&name=&branch=edge 2022-03-14 18:15:15 ikke: thank you 2022-03-14 18:28:16 j`ey: securetty 2022-03-14 18:28:59 Hello71: what should it have in it that would make that work under wayland? 2022-03-14 18:29:55 i think deleting it turns it off 2022-03-14 18:30:09 why don't you use doas instead 2022-03-14 18:30:51 I probably should 2022-03-14 18:31:27 hm, it seems that securetty is about /bin/login, not su? 2022-03-14 18:34:28 maybe installing shadow-login- 2022-03-14 18:35:00 but it will install linux-pam 2022-03-14 18:39:35 it seems that you can also add "pts/[0-9]*" to /etc/securetty 2022-03-14 18:40:06 I will look at the su source later, not sure if it's util-linux or busybox, we'll see 2022-03-14 18:40:16 but I feel that doas is a simpler way 2022-03-14 19:59:21 Can mdio_bcm_unimac.ko be added to the aarch64 kernel? Otherwise Raspberry Pi 4 can't access ethernet on boot. 2022-03-14 20:00:08 FallenSnow: best way to achieve that is by creating an issue on gitlab.alpinelinux.org/alpine/aports 2022-03-14 20:07:52 @ikke, thank you. 2022-03-14 20:07:53 https://gitlab.alpinelinux.org/alpine/aports/-/issues/13594 2022-03-14 20:37:15 ikke: do you know of a quick workaround? I've already tried just adding the module to the initramfs but it fails to boot. 2022-03-14 20:38:25 Where did you get the module from? 2022-03-14 20:38:54 The modloop-lts. 2022-03-14 20:46:19 Doesn't that mean the module is already enabled? 2022-03-14 20:46:30 Is it just an issue it's not part of the initramfs? 2022-03-14 20:46:47 Both yes. 2022-03-14 20:47:00 It's marked as a module in the config. 2022-03-14 20:47:05 ok 2022-03-14 20:47:11 How did you try to add it yourself? 2022-03-14 20:49:54 `gunzip modloop-lts` then use an archive editor to add the module in the right place. Then just boot that but it freezes in IPXE. 2022-03-14 20:50:00 Haven't tried using cpio. 2022-03-14 20:50:15 FallenSnow: take a look at /etc/mkinitfs/ 2022-03-14 21:02:32 Well I downloaded aarch64 standard onto a USB and booted that on the Pi4. It works with ethernet and shows mdio_bcm_unimac is loaded. 2022-03-14 21:02:39 Let me look at /etc/mkinitfs/ 2022-03-15 16:18:20 hello, I just tried to launch qpdf on Alpine and I have the following error: Error relocating **** symbol not found 2022-03-15 16:20:45 vlp: what version of alpine? 2022-03-15 16:20:58 3.15 2022-03-15 16:21:17 Seems to work for me on 3.15 2022-03-15 16:22:26 :/ 2022-03-15 16:22:37 Does it happen when you open a pdf? 2022-03-15 16:22:40 vlp: make sure you're not mixing edge and 3.15 2022-03-15 16:23:36 ahhh yes probably 2022-03-15 16:26:24 it also helps if you don't randomly censor the error message 2022-03-15 16:31:13 here is the uncencored error :) Error relocating /usr/bin/qpdf: _ZNK7QPDFJob11getExitCodeEv: symbol not found 2022-03-15 16:33:13 Hi again folks! anyone here managed to get a vm up and running with kvm using an alpine image using virt-install? as alpine boots i get a 'Mounting boot media: failed.' following a "Attached SCSI disk" message. the virt-install command I'm using is: doas virt-install --name alpine6 --ram 512 --disk path=/var/lib/libvirt/images/a6.qcow2,size=8 --vcpus 1 --os-t 2022-03-15 16:33:18 ype linux --os-variant generic --nographics --location /mnt/cdr,kernel=/mnt/cdr/boot/vmlinuz-virt,initrd=/mnt/cdr/boot/ 2022-03-15 16:33:47 oh the copy paste went wrong. sorry about that 2022-03-15 16:34:26 command being: 'doas virt-install --name alpine6 --ram 512 --disk path=/var/lib/libvirt/images/a6.qcow2,size=8 --vcpus 1 --os-type linux --os-variant generic --nographics --location /mnt/cdr,kernel=/mnt/cdr/boot/vmlinuz-virt,initrd=/mnt/cdr/boot/initramfs-virt --accelerate --extra-args="console=ttyS0"' 2022-03-15 16:34:37 many thanks for any reponse in advance 2022-03-15 16:35:03 vlp: that looks like mixing edge with stable 2022-03-15 16:35:28 i swear every time i upgrade something in edge that was not upgraded since gcc11 someone shows up to complain it broke in stable :p 2022-03-15 16:36:59 psykose: best option to fix that ? disable stable and full upgrade to edge ? 2022-03-15 16:37:08 or be on stable, yes 2022-03-15 16:37:14 or build the upgrades you want locally on stable 2022-03-15 16:40:11 +1 for building packages if you want a newer version. It's probably as easy as `git clone https://gitlab.alpinelinux.org/alpine/aports.git; cd aports/community/qpdf; abuild -r` 2022-03-15 16:41:00 only if you have abuild set up 2022-03-15 16:41:14 but that is just abuild-keygen and doas and making a distfiles folder 2022-03-15 17:58:30 Why does lighttpd want to start after sshd (according to its init script) 2022-03-15 20:51:47 \quit 2022-03-15 20:53:23 tks for the help, I have switch to edge and it works (no more mix stable/edge) ! 2022-03-15 21:16:41 Is there a way to install pkgs with wilcard? E.g.: apk add influxdb* 2022-03-15 21:16:52 Or I have to install them one by one? 2022-03-15 21:20:40 frojnd: afaik, there is no wildcard support for add 2022-03-15 21:20:46 (there is for del) 2022-03-15 21:20:57 Ok 2022-03-15 22:22:37 sry, i'm tired and not a chat-person. anyway, problem: i run alpine on my Lenovo ideapad 330-15IKB / Intel Core i5-8250U / since my upgrade to edge, wpa_supplicant doesnt start on boot: ip: ioctl 0x8913 failed: No such device (wlan0). it's working (sometimes) when i do a reboot and i got a connection the day after the lase linux-lts update. for 1 day. today, again no connection during boot. this did not happend an stable 2 week 2022-03-15 22:23:33 and sry for my horrible en 2022-03-15 22:35:17 https://pastebin.com/raw/81Ry5By2 2022-03-15 22:38:08 ps: i added hwdrivers to rl sysinit (https://gitlab.alpinelinux.org/alpine/aports/-/issues/9653 ) a few minutes ago.could this be helpful? 2022-03-15 22:53:36 i think it is possibly a race condition 2022-03-15 22:53:44 what does dmesg say in the failure condition? 2022-03-15 22:54:28 lets take a look one moment plz 2022-03-15 23:00:39 ..of course, i'm online now -- after a reboot (where its working) therefore no error to see. var/logs showed me "/sbin/wpa_supplicant, pid xxx, exited with return code 255". next time, i have to take a look at dmesg, sorry 2022-03-15 23:01:38 jessas, my english... 2022-03-15 23:38:08 .. Hello71, i started the syste 2 times, with and without hwdrivers, no error. therefore, no dmesg message to find out whats happening. it's like yesterday- wlan worked the whole day, today, with 3 systemstarts, the posted error. until now, for now i guess. it's more or less random. at least, it looks like that, but isn't of course. 2022-03-15 23:41:11 before i write more nonsens, i go to sleep. thx anyway,@everyone for alpine/hard work, every day. good night 2022-03-16 00:06:06 ..and it happened again. dmesg: wlan0: authenticated/associated - i cant see an error there. restart wpa_supplicant: no connection. restart networking: " 2022-03-16 00:06:44 wlan0 ... cat: can't open '/var/run/udhcpc.wlan0.pid': No such file or directory sh: you need to specify whom to kill / [ ok ] * Starting networking ... * lo ... / [ ok ] * wlan0 ... udhcpc: started, v1.35.0 2022-03-16 00:52:02 update: i should sleep, but i tried 2 system starts with hwdrivers added @sysinit. it seems to work now.. 2022-03-16 00:57:34 it's loading hardware drivers, i can see it in the script, but is there any documentation about this service? 2022-03-16 09:19:44 I would like to open port 3000 tcp used by grafana on localhost:3000 I setup awall. Now as I read docs I should add this by creating /etc/awall/grafana.json and this is my entry: https://termbin.com/nqwv is this correct approach? 2022-03-16 09:21:02 However when I do awall list, there is no grafana listed 2022-03-16 09:25:29 I think I'm missing more stuff 2022-03-16 10:09:51 hi 2022-03-16 10:10:25 cava in testing repo is built without alsa/pulseaudio support, wondering if that can be changed 2022-03-16 10:11:43 also noticed an issue with i3wm-gaps in testing, it doesn't come with i3-save-tree. seems like i3-save-tree is locked as a dependency only to i3wm 2022-03-16 10:12:08 had to install i3wm then copy i3-save-tree then install i3wm-gaps to get i3-save-tree 2022-03-16 10:12:18 hoping that can be changed as well 2022-03-16 10:13:33 i3wm-gaps has save-tree 2022-03-16 10:13:45 it's not split into a subpackage 2022-03-16 10:13:54 https://pkgs.alpinelinux.org/contents?file=&path=&name=i3wm-gaps&branch=edge 2022-03-16 10:14:17 /usr/bin/i3-save-tree 2022-03-16 10:19:13 oh that's weird, not sure what was going on that day then 2022-03-16 10:20:26 bewildering me atm, but w.e. 2022-03-16 10:20:49 just tried to uninstall/reinstall i3wm-gaps, and yea, i3-save-tree is there 2022-03-16 10:30:49 anothe thing, i keep running into an issue with apk and "package mentioned in index not found" 2022-03-16 10:30:52 is that a remote thing? 2022-03-16 10:31:23 tsuki: what mirror? 2022-03-16 10:31:29 everytime this happens it goes away when i update later on 2022-03-16 10:31:31 means the mirror is in the middle of rsync i guess, the index was synced but not the packages 2022-03-16 10:31:42 waterloo csclub 2022-03-16 10:31:59 ;; This buffer is for text that is not saved, and for Lisp evaluation. 2022-03-16 10:32:02 ;; To create a file, visit it with C-x C-f and enter text in its buffer. 2022-03-16 10:32:04 y 2022-03-16 10:32:06 oops 2022-03-16 10:32:17 C-x C-f 2022-03-16 10:33:15 mirror.csclub.uwaterloo.ca 2022-03-16 10:33:28 was copying from urxvt to weechat 2022-03-16 10:33:55 that makes sense about the rsync 2022-03-16 10:37:54 urxvt wouldn't let me highlight a portion of the link and i went to paste lol 2022-03-16 18:39:23 Waiting 5 seconds for no one to get the joke. 2022-03-16 19:26:36 I'm trying to debug why grafana won't run under different http address. I Would also like to see some logs from grafana but I'm unfamiliar with Alpine. I started service like this: service grafana start. All good. When I do ps aux |grep grafana I see this: root 0:00 supervise-daemon grafana --start --user grafana grafana /usr/sbin/grafana-server -- -config /etc/grafana.ini 2022-03-16 19:26:38 -homepath /usr/share/grafana cfg:paths.data=/var/lib/grafana/data cfg:paths.plugins=/var/lib/grafana/plugins cfg:paths.provisioning=/var/lib/grafana/provisioning cfg:server.http_addr=127.0.0.1 cfg:log.mode=syslog 2022-03-16 19:26:48 Note the last string cfg:log.mode=syslog whatever that means 2022-03-16 19:26:56 So it's logging somewhere 2022-03-16 19:27:19 syslog is /var/log/messages 2022-03-16 19:27:26 Ok 2022-03-16 19:28:49 No errors no nothing 2022-03-16 19:29:15 It should change http_addr if I change it within /etc/grafana.ini but it won't it only changes port 2022-03-16 19:29:45 you should put it in /etc/conf.d/grafana 2022-03-16 19:30:21 there is already a cfg:server line for it there 2022-03-16 19:30:27 it probably takes cli as precedence 2022-03-16 19:30:33 so whatever you put in the ini is overriden by that 2022-03-16 19:30:56 cfg:server.http_addr=127.0.0.1 i assume it's this that you want 2022-03-16 19:31:15 That's exactly what I want! 2022-03-16 19:31:24 psykose: how did you know about that file? 2022-03-16 19:31:29 i run grafana 2022-03-16 19:31:37 but uh there is usually a conf.d for a service 2022-03-16 19:31:48 But is there a way apk should tell about that conf file like pacman -Ql packages does? 2022-03-16 19:31:49 i just check there first and then other stuff 2022-03-16 19:32:05 apk info -L grafana-openrc 2022-03-16 19:32:06 Anywasys /etc/conf.d is good to check 2022-03-16 19:32:31 Thank you so much! 2022-03-16 19:32:39 Do you run it vanilla or in a docker? 2022-03-16 19:32:56 3.15 vanilla with that default localhost 2022-03-16 19:33:00 behind nginx 2022-03-16 19:34:36 I will first setup averyting inside virtual env before putting to real hardware 2022-03-16 19:35:04 i basically just wanted a status graph, so i went and set up prometheus+grafana+node_exporter 2022-03-16 19:35:09 it took me like 40 minutes for everything 2022-03-16 19:37:19 Well I have another problem. I own raspbeerry pi B 512MB of ram one of the first ones and I was running on it: grafana + influxdb + nodered And also ran on it some sensors and then I was feeding grafana with all different data. Problem was that it got clogged after 2 or 3 weeks... so now I got some old mini pc which have 2GB of ram and 2 intel cores which should be more tha 2022-03-16 19:37:21 nenough 2022-03-16 19:37:48 i run everything on a single rpi4b with 8gb 2022-03-16 19:37:55 but it never goes past like 2022-03-16 19:37:56 idk 2022-03-16 19:37:57 psykose: I can't order any rpi where I live 2022-03-16 19:38:13 no peak above 500mb in grafana logs 2022-03-16 19:38:32 rpi4 has also fasater ram afaik 2022-03-16 19:38:48 i think it's 3200mhz lpddr4x or something 2022-03-16 19:38:54 but overall it's slow as fuck anyway, so 2022-03-16 19:38:58 Do you run any firewalls? 2022-03-16 19:39:01 that old mini pc probably competes 2022-03-16 19:39:03 Like awall 2022-03-16 19:39:04 nope 2022-03-16 19:39:19 Everyting local I see 2022-03-16 19:39:38 it's publically accessible via https, but adding a firewall i would still be allowing the same ports i use 2022-03-16 19:39:46 the only point of it would be to disallow outbound connections 2022-03-16 19:40:03 or do specific packet level filtering 2022-03-16 19:40:05 too much effort 2022-03-16 19:40:13 Heh 2022-03-16 19:40:29 Since I'm gonna put some more sensors I would not love my data to get out :P 2022-03-16 19:40:40 yep 2022-03-16 19:40:42 With CO2 sensors you can pretty much tell when and how many ppl are in apartment :P 2022-03-16 19:58:53 Maybe I don't need nodered for broker, only mosquitto 2022-03-16 19:59:01 Even better 2022-03-16 20:41:44 Congrats on release! 2022-03-17 01:06:35 hi 2022-03-17 01:06:51 just noticed flatpak broke after after the last update 2022-03-17 01:06:57 not sure what's up 2022-03-17 01:07:14 error: While opening repository /var/lib/flatpak/repo: opening repo: opendir(/var/lib/flatpak/repo): No such file or directory 2022-03-17 01:08:46 maybe bubblewrap update ? can you run bwrap --version ? 2022-03-17 01:09:41 bubblewrap 0.6.1 2022-03-17 01:11:51 tsuki: Does /var/lib/flatpak/repo exist? Also what command did you run that gave this error? 2022-03-17 01:12:09 the /var/lib/flatpak directory is empty 2022-03-17 01:12:20 i was running flatpak upgrade 2022-03-17 01:12:47 flatpak search is the same result 2022-03-17 01:13:17 my .local/share/flatpak stuff is still there, but looks like nothing will start 2022-03-17 01:13:41 just using this for steam on alpine, i have no experience with flatpak 2022-03-17 01:14:07 i tried to readd the flathub repo, but that did not do anything 2022-03-17 01:14:34 re-add* 2022-03-17 01:17:15 tsuki: What do you mean by "last update"? Do you know what version of flatpak you were on previously? 2022-03-17 01:17:30 i update multiple times a day 2022-03-17 01:17:35 so w.e. was the previous version 2022-03-17 01:18:42 Are you on edge? 2022-03-17 01:18:45 yes 2022-03-17 01:19:01 Does "flatpak list" work? 2022-03-17 01:20:14 no 2022-03-17 01:27:56 tried to uninstall flatpak and delete the /var/lib/flatpak directory, and reinstall, but that directory doesn't get setup at all through install 2022-03-17 01:30:04 i figured it out 2022-03-17 01:30:31 running `flatpak remote-add --if-not-exists flathub https://flathub.org/repo/flathub.flatpakrepo` sets up the repo dir in the flatpak directory 2022-03-17 01:30:54 no idea why that's necessary since i have flathub setup with --user flag 2022-03-17 01:31:14 no idea why the last update would have broken that either :\ 2022-03-17 01:34:17 i am able to start steam fine as well 2022-03-17 01:34:29 i had to run the flatpak remote-add as root 2022-03-17 01:35:01 alright, thank you 2022-03-17 01:35:32 Hmm, that's weird. I've never used flatpak install/update/etc as non-root though 2022-03-17 01:36:07 And tsuki is gone... 2022-03-17 13:04:17 is there an offical way to make a copy from the entire (i.e.) 3.15 package repository? 2022-03-17 13:04:32 rsync it :) 2022-03-17 13:05:54 okay 2022-03-17 13:06:06 (an implicit question is answered: it is permitted :P) 2022-03-17 13:06:18 can you give me an estimation about its size? 2022-03-17 13:06:45 aron: that's what mirrors do :) 2022-03-17 13:06:57 ah, yes, makes sense ;) 2022-03-17 13:07:05 https://wiki.alpinelinux.org/wiki/How_to_setup_a_Alpine_Linux_mirror 2022-03-17 13:07:25 nice, tyvm 2022-03-17 13:07:57 it's not small, tho :P 2022-03-17 13:08:16 nope 2022-03-17 13:11:22 at least I can set a bw limit -- back then in my arch linux times I used to keep a local mirror, it was waaay faster, and I always wanted to setup an alpine mirror too 2022-03-17 13:11:36 I have more installations than bandwidth 2022-03-17 13:13:04 (ofc, the first rsync will be painful AF) 2022-03-17 16:04:38 I saw APKv3 mentioned here: https://gitlab.alpinelinux.org/alpine/tsc/-/blob/master/minutes/2022-03-08.md#apkv3, is there a document that describes the motivation, differences, etc for building the new version? can't find any relevant documentation online 2022-03-17 16:19:31 handlerug: some of it is here: https://lists.alpinelinux.org/~alpine/apk-tools 2022-03-17 16:24:08 ikke: thanks! does one compile it from the source tree to install it, or there's a package? https://lists.alpinelinux.org/~alpine/apk-tools/%3C43397dfce7287c7ea51d04ba0d47d81c%40dereferenced.org%3E mentions apk-tools@testing but it's gone it seems 2022-03-17 16:28:23 you can compile it yourself, it's fairly easy 2022-03-17 16:28:45 note that you cannot replace the alpine one with it to just use apk3, it's for personal use and testing (making your own package stuff, etc) 2022-03-17 16:29:15 gotcha, thanks 2022-03-17 16:29:26 the package format changed after all :) 2022-03-17 16:30:13 iirc, apk3 should be able to use apkv2 packages / repositories 2022-03-17 17:54:25 sry.again. since my upgrade to edge, the system cannot conntect on boot to wlan0: https://pastebin.com/raw/81Ry5By2 . hello71 meant in the last chat, it is maybe a race condition. dmesg does not give me an error. today i tried this advice: https://gitlab.alpinelinux.org/alpine/aports/-/issues/9079 (rc_need=udev-settle) and it worked. but why (i never had to do this last year)? and has this "tweak" side effects? whats the actu 2022-03-17 17:55:03 oleovit: it got truncated at "whats the actu" 2022-03-17 17:55:33 whats the actual problem? ;) 2022-03-17 17:58:35 ps: hwdrivers (pastebin) is not started (udev-trigger doas this work). was just a test. 2022-03-17 17:59:03 maybe something in the kernel or firmware changes, not sure 2022-03-17 17:59:44 yep, i guass some firmware update.. 2022-03-17 17:59:52 "guess" 2022-03-17 18:02:27 so, what are the side effects of adding rc_need=udev-settle at /etc/conf.d/networking (if any).. + i can't be the only one if it's a bug 2022-03-17 18:18:38 Hi there! I want to modify an existing APKBUILD file for use within my docker-container but I can not figure out how this can be done - is this possible? 2022-03-17 18:20:46 XHFHX: where do you want to build that modified APKBUILD? 2022-03-17 18:21:38 within and for my docker container - i need a newer version of a package and I would just need to bump up the version number to make it work 2022-03-17 18:21:40 I think that the simple way should be building it on other place, then add the repository with the modified package and install it on 'docker build' 2022-03-17 18:22:11 what package is? 2022-03-17 18:22:44 https://git.alpinelinux.org/aports/tree/community/gst-plugins-good/APKBUILD - i need 1.20.1 b/c they introduced something new that I need 2022-03-17 18:23:19 i would need to bump up the var pkgver but I dont know how to do this 2022-03-17 18:23:54 have you used abuild before? 2022-03-17 18:24:40 no, actually i used most of the time debian but recently found alpine more attractive b/c the packages are newer 2022-03-17 18:25:31 alpine isn't that much fresher than debian sid 2022-03-17 18:25:54 well, you should understand a little the process without the docker layer 2022-03-17 18:26:28 better try to just build the package and then add it to the container image 2022-03-17 18:26:44 you can't bump that to 1.20.1 without bumping main/gstreamer and main/gstreamer-plugins-base too 2022-03-17 18:26:59 and there is an mr for that to 1.20.1 :) should be merged soon enough 2022-03-17 18:27:10 mh but then i would need to add a binary to my git repo and also i only have docker available as i am on macosx 2022-03-17 18:27:55 oh nice news with the MR :) is it possible to switch the branch to the one with the MR included? 2022-03-17 18:28:06 not until it's merged 2022-03-17 18:28:11 and then it's just edge 2022-03-17 18:28:13 it probably have some problem if it's not merged 2022-03-17 18:28:30 yes the problem is i cannot press merge :) 2022-03-17 18:28:40 that's huge problem 2022-03-17 18:29:19 but it would be really nice to know if there is a way to modify the build process of an apkbuild - i only use gstreamer b/c ffmpeg is build w/o jack-dev bindings but then the gstreamer version is too low 2022-03-17 18:29:44 you can modify even the source code if you want 2022-03-17 18:30:51 but how? i mean i can build it manually but i would much prefer to just bump up the version varibale and let it build for me 2022-03-17 18:31:13 using abuild 2022-03-17 18:31:16 you can rebuild ffmpeg with jack if you want 2022-03-17 18:31:20 just add jack-dev to makedeps 2022-03-17 18:31:54 how do you build it 'manually'? 2022-03-17 18:32:10 yes but how do i run the build of such an APKBUILD? i tried apk add --repository /foo/bar/ gstreamer-custom 2022-03-17 18:32:33 where in /foo/bar/gstreamer-custom is the APKBUILD file i modified 2022-03-17 18:32:41 apk doesn't build anything 2022-03-17 18:32:47 you use abuild to build apkbuilds 2022-03-17 18:33:52 https://wiki.alpinelinux.org/wiki/Creating_an_Alpine_package 2022-03-17 18:34:21 i tried this https://wiki.alpinelinux.org/wiki/Abuild_and_Helpers but it seems I am required to use a non-root user and setup a keyfile - both steps are not so nice within a docker environment 2022-03-17 18:35:09 what? 2022-03-17 18:36:32 well you can create users inside the container, probably better&simpler than skip the root check 2022-03-17 18:38:07 yeah, but in my experince it really gets troublesome if deviating from root as it may install only for the user and not systemwide 2022-03-17 18:38:59 installing things with apk is always systemwide 2022-03-17 18:39:02 abuild has some suid binaries for handling it properly 2022-03-17 18:39:29 my dockerfile is already 150+ lines and takes 15 minutes to build, i would love to keep it more simple :D something like abuild build --install /root/MYAPK would be great 2022-03-17 18:39:30 it will build and package everything as a systemwide perspective 2022-03-17 18:40:11 I don't think that you should add abuild related things to your dockerfile 2022-03-17 18:40:39 well, I feel that they should be separated things 2022-03-17 18:40:41 i can't comment on what the best approach to build a single package is without seeing your dockerfile 2022-03-17 18:40:53 what would you recommend @donoban? 2022-03-17 18:41:02 why do i get the impression that you also don't know about layers 2022-03-17 18:41:27 build the package, and then pass it to your Dockerfile 2022-03-17 18:41:56 you can add a repository to a /etc/apk/repositories 2022-03-17 18:42:24 I'm not sure if you can just pass the package and install it 2022-03-17 18:42:37 @donoban: but then i would need to store the binary somewhere and i really try to avoid storing binaries in git repos 2022-03-17 18:43:56 well I don't know how is your infraestructure 2022-03-17 18:44:12 @Hello71: I know what layers are but the big problem is that many dependencies need to be compiled customly and b/c of network reasons it needs 3 apps in one container. it is a really nasty dockerfile :/ 2022-03-17 18:44:17 can't you have a custom apk repository available for docker-build? 2022-03-17 18:44:34 in the worseee case 2022-03-17 18:44:45 I will do a Docker file just for your modified APKBUILD version 2022-03-17 18:45:00 and the your other Dockerfile uses that image 2022-03-17 18:45:04 instead rebuilding all everytime 2022-03-17 18:45:47 yeah but i need a combination of 3 different softwares that need to be build in specific order to work 2022-03-17 18:48:09 my infastructure is not huge, it is only a github repo and a vps 2022-03-17 18:49:36 in your git repo you should have a aports clone with a branch for your modified packages 2022-03-17 18:49:57 or better in alpine gitlab :P 2022-03-17 18:50:16 you don't need a whole aports clone to build 3 things 2022-03-17 18:50:20 so you can rebuild your packages everytime you want without having binaries on a repo 2022-03-17 18:51:06 yeah I suppose, just for having a track of alpine mainline 2022-03-17 18:51:13 yeah but you only rebuild if you invalidate your cache or change the Dockerfile 2022-03-17 18:52:05 you only need to rebuild your alpine packages if you change the APKBUILD 2022-03-17 18:53:25 why do you want to rebuild them each time you rebuild your docker image? 2022-03-17 18:54:36 well to get support a stateless setup with few dependencies as possible - if i host the binary somewhere i need to be sure that it stays there which is difficult 2022-03-17 18:57:05 I see 2022-03-17 18:59:38 is there a way to build with abuild w/o generating a key? 2022-03-17 19:00:25 I think that not 2022-03-17 19:00:55 it doesn't sound like it would be very useful to use the apkbuild at all in your case 2022-03-17 19:01:10 you can just extract the prepare/build steps and makedeps and just fetch the source and build it manually 2022-03-17 19:01:36 or find some docker container that has what you want in it already 2022-03-17 19:04:14 psykose: well, yeah, but i need to build something with the dev option enabled and then only seeing default_dev and looking what it does is like: it looks like a massive source of errors to do it by myself 2022-03-17 19:05:12 no, default_dev only splits some files into a subpackage 2022-03-17 19:05:18 if you call i.e. `make install` they will be there already 2022-03-17 19:05:54 it doesn't install any extra files or anything, it's just moving things around for our packaging side 2022-03-17 19:07:12 oh, okay - i thought it also moves some files around. i still struggle to get behind the building steps, especially as every project is using something else: meson, make, bazel... 2022-03-17 19:07:33 Is gettext-dev being replaced by gettext-tiny-dev? 2022-03-17 19:07:47 well, then i will simply stick to the instructions on the website and build it from hand :) 2022-03-17 19:09:52 you can probably just copypaste the lines from build 2022-03-17 19:10:01 Saijin_Naib: maybe, we'll see 2022-03-17 19:10:13 ACTION sent a code block: https://matrix.org/_matrix/media/r0/download/matrix.org/dfhogjaRaboztvDKLvPaGHCi 2022-03-17 19:10:38 yeah, its great to have the build script available, really good resource :) 2022-03-17 19:10:39 gettext-dev being pulled in and a bunch of the warned files are in gettext-tiny-dev only presently, but that conflicts with like, everything apparently 2022-03-17 19:10:59 this warning doesn't mean anything 2022-03-17 19:11:50 Oh 2022-03-17 19:11:59 Scary looking then 2022-03-17 19:12:17 almost every autotools project spits out 1k lines of warnings like that for some reason 2022-03-17 19:12:29 i'm not sure why those m4's are only in -tiny, but i don't think they matter 2022-03-17 19:12:58 Yeah, it's failing, but I wasn't sure if it was that or not 2022-03-17 19:13:01 Thanks for the assist, as ever 2022-03-17 19:13:09 you might need autoreconf -fi 2022-03-17 19:13:15 or something 2022-03-17 19:13:41 Hopefully nothing too esoteric. I'd like this one to be a surprise haha 2022-03-17 19:13:42 given the guess stuff you could do update_config_guess and _sub in prepare() instead too 2022-03-17 19:13:57 but i haven't touched it 2022-03-17 19:22:54 I think I need help... I think this has vendored (?) dependencies that I don't know how to make and have available for the ./autogen.sh prepare step so that functionality gets enabled 2022-03-17 19:23:11 it's in a directory below the main source with its own automake file 2022-03-17 19:30:07 what does ./autogen.sh fail on 2022-03-17 19:34:44 currently, fails after no update needed for ./config.guess 2022-03-17 19:42:36 Alright, passing! Still can't figure out the vendored depdendency, though 2022-03-17 19:43:08 Is that technically a prepare step since it needs to be in place for the autoconf to detect and enable it? 2022-03-17 19:43:19 not quite 2022-03-17 19:43:21 Or is it something I need to split out and submit another APKBUILD for somehow 2022-03-17 19:43:25 nah 2022-03-17 19:43:33 i mean i have no idea what it is 2022-03-17 19:43:36 Is this type of thing common? 2022-03-17 19:43:43 maybe 2022-03-17 19:43:43 This practice, I mean 2022-03-17 19:44:04 First I'm seeing of it outside of GPSBabel and FreeFileSync, which I also can't figure out 2022-03-17 19:44:07 ACTION sent a code block: https://matrix.org/_matrix/media/r0/download/matrix.org/MMxmHmzDvgHoYNaaYajCJHyB 2022-03-17 19:44:33 Tadah! First Orage release in like, 6 years haha 2022-03-17 19:44:34 GTK+3.0 ported even 2022-03-17 19:44:43 does that work or 2022-03-17 19:44:55 Yeh 🙂 2022-03-17 19:45:01 * LIBICAL support: yes, and not building tz_convert 2022-03-17 19:45:19 That's the thing that I'm stuck on. tz_convert is right in the directory with its own automake file 2022-03-17 19:45:29 this looks identical to every other autotools project then 2022-03-17 19:45:35 I know from coordinating meetings with folks around the world that, yeah, that's important functionality for a calendar/alarm 2022-03-17 19:45:52 i' 2022-03-17 19:45:55 Yeah, so far it hasn't thrown much my way except the tz_local source being inside it 2022-03-17 19:45:58 i'm pretty sure what it means is libical does it 2022-03-17 19:46:04 Ooooh 2022-03-17 19:46:09 and it links to it fine 2022-03-17 19:46:15 I am not good at reading these log messages 2022-03-17 19:46:21 you could run it and see if timezones work fine 2022-03-17 20:00:59 ACTION uploaded an image: (71KiB) < https://matrix.org/_matrix/media/r0/download/matrix.org/auviYBsOQMNjaUpCxYYUhDQI/image.png > 2022-03-18 03:43:49 Works a treat 2022-03-18 06:12:23 Is it possible to add multiple repositories to alpine_repo? 2022-03-18 06:13:56 what do you mean with alpine_repo? 2022-03-18 06:21:05 https://wiki.alpinelinux.org/wiki/PXE_boot 2022-03-18 06:21:17 The alpine_repo cmdline option. 2022-03-18 06:23:45 it writes it directly to /etc/apo/repositories, so you'd need to find a way to embed newlines in it 2022-03-18 06:26:20 I tried repo1\nrepo2 but that didn't work. 2022-03-18 06:26:31 So I just rewrote part of /init. 2022-03-18 10:34:05 I would like to know how to open rar file, I don't see any unrar in repo 2022-03-18 10:34:53 bsdtar should be able to 2022-03-18 11:51:18 7z also should be able to do that 2022-03-18 15:24:54 Hi! I'm trying to understand the installation procedure to put alpine onto a raspberry pi. I'm essentially following this guide: https://wiki.alpinelinux.org/wiki/Raspberry_Pi_-_Headless_Installation and I can boot into it fine and run alpine-setup. Now I'm confused with step 3 there. It says " Assuming your SD card device is called mmcblk0 choose that device. It will ask you if you want to erase the disk, then configure it a 2022-03-18 15:24:56 anyone knows a simple method for auto-umount fuse mounts if they are unused for some while? 2022-03-18 15:25:22 Does that mean I should erase it? Because when I do I'm essentially destroying my installation as far as I'm concerned 2022-03-18 15:28:02 chmanie: it seems that it prepares the sd for booting diskless with ssh running, but then you reinstalls alpine on the sd in a sys-install 2022-03-18 15:28:28 I think I would like to run it in data mode 2022-03-18 15:29:12 Ah, sorry, So when I do erase the disk it just stays erased 2022-03-18 15:29:19 you will need to tweak it for be able to boot 2022-03-18 15:29:27 There's no installation following up, the setup-tool just exits 2022-03-18 15:29:39 it seems that there is some error 2022-03-18 15:29:43 is there anything on the output? 2022-03-18 15:30:07 let me do that again, then I can copy the output for you 2022-03-18 15:30:19 ok 2022-03-18 15:30:26 So saying yes to erasing the disk would be the right choice in my case? 2022-03-18 15:30:46 for doing a data install? 2022-03-18 15:31:07 Yes. Sorry maybe I'm confused about the modes 2022-03-18 15:31:30 if you want to do a data install you shouldn't erase the partitions where you put alpine 2022-03-18 15:32:06 if you use sys mode it will reinstall all root, boot partition etc.. so it should work properly 2022-03-18 15:32:08 Ah. Okay, so the setup-alpine basically reconfigures the current disk that I created in the steps before 2022-03-18 15:32:21 yes, since it's running on ram you can erase it 2022-03-18 15:32:32 but you need to setup something bootable for rebooting later 2022-03-18 15:33:02 I don't necessarily need to erase it, I think it's fine like that 2022-03-18 15:33:21 you can also setup lbu and continue using it 2022-03-18 15:33:31 just remember to lbu commit your changes before reboot 2022-03-18 15:34:15 So in data mode the ramdisk is created from the bootable media but certain folders on the bootable media are also mounted to /var (I think?) 2022-03-18 15:34:41 I'm not really sure, I don't use data mode since some yaers ago 2022-03-18 15:34:45 Does setup-alpine do the lbu commit for me or do I have to do it afterwards? 2022-03-18 15:34:57 but I suppose that you shouldn't overrwite your boot media with it 2022-03-18 15:35:10 Yes, that's what's got me confused 2022-03-18 15:35:12 yes 2022-03-18 15:35:27 Thank you, I'll try again! 2022-03-18 15:36:03 when setup-alpine asks for using some disks, ignore it 2022-03-18 15:36:13 and then it will ask for setting lbu 2022-03-18 15:36:48 right, okay! 2022-03-18 15:38:13 uhm 2022-03-18 15:39:22 re-reading the wiki page 2022-03-18 15:39:56 it seems that all is on /dev/mmcblk1p1 2022-03-18 15:40:05 wops 2022-03-18 15:40:24 /dev/mmcblk1p2 could be used for sys-data 2022-03-18 16:04:05 So after running alpine-setup and rebooting, none of the configurations have been saved 2022-03-18 16:05:37 Only that alpine-setup is gone apparently 2022-03-18 16:05:52 Argh scratch that 2022-03-18 16:08:28 ey chmanie I think that I missinterpreted the wiki page 2022-03-18 16:08:39 it seems that all is on /dev/mmcblk1p1 2022-03-18 16:08:44 /dev/mmcblk1p2 could be used for sys-data 2022-03-18 16:09:11 after doing finish setup-alpine, did you lbu_commit? 2022-03-18 16:12:25 I did not do lbu commit. Sorry I must have misunderstood earlier 2022-03-18 16:12:48 I did do that now, and it seems to have taken some settings 2022-03-18 16:13:02 Only wifi is the one that I used in the `wifi.txt` 2022-03-18 16:14:12 if you didn't commit lbu, when you reboot it will load again the /mnt/sd1/headless.apkovl.tar.gz 2022-03-18 16:14:40 i'm not sure if you end with two apkvol files what it will load 2022-03-18 16:16:24 I _think_ I'm almost good now, thanks for your help and sorry for the confusion 2022-03-18 16:17:22 great!, no worries I was pretty confused too 2022-03-18 18:21:28 is iwd meant to write logs somewhere? do I need to somehow make it depend on my device becoming available? 2022-03-18 18:21:37 it's crashing for me on startup 2022-03-18 18:24:15 start it yourself and see why it crashes 2022-03-18 18:24:37 problem is.. it works the second time 2022-03-18 18:24:47 which is making me think its starting too early 2022-03-18 18:25:20 putting it in default should be enough 2022-03-18 18:25:31 maybe 2022-03-18 18:25:32 i dunno 2022-03-18 18:26:39 https://git.alpinelinux.org/aports/tree/community/iwd/iwd.initd I'll look at tweaking this to be 100ms, not 50 or something 2022-03-18 18:27:16 you could also make it supervisor=supervise-daemon and it might just restart itself repeatedly up to 5 times and work 2022-03-18 18:27:22 it's already foreground default 2022-03-18 18:27:44 I'll give that a try too, thanks 2022-03-18 18:29:37 i do remember using that specifically for a long time before so it should be okay 2022-03-18 18:29:54 annoyingly it was working fine before :/ so some kinda timing issue I guess 2022-03-18 18:30:58 most likely 2022-03-18 18:31:17 such is painful life with openrc.. 2022-03-18 18:57:17 And that's why we need s6, right? Actual supervision 2022-03-18 18:58:53 the answer is Complicated 2022-03-18 18:59:17 yes but not because of this 2022-03-18 18:59:49 i think you only need to start the device first if you use usedefaultinterface=false? 2022-03-18 19:12:59 Hi, I want to use the watchdog on rpi4. I created a the “/etc/watchdog.conf” and added the watchdog service “rc-update add watchdog”.But starting the service “rc-service watchdog start” just reports “WATCHDOG_DEV is not set, ERROR: watchdog failed to start”. Any idea what I miss? 2022-03-18 19:14:55 probably has to be set in /etc/conf.d/watchdog or something 2022-03-18 19:42:22 psykose Thanks, found it and got kicked out right away :-) Do I get it right and this is not the standard linux watchdog service https://linux.die.net/man/8/watchdog https://linux.die.net/man/5/watchdog.conf ? 2022-03-18 19:42:40 i have no idea what it even is 2022-03-18 19:43:02 from the description it's the busybox watchdog 2022-03-18 19:43:31 so, not that one no 2022-03-18 19:44:36 with the busybox one you need to pass the device yourself it seems 2022-03-18 19:44:53 namely just adding /dev/watchdog to that WATCHDOG_DEV line, probably 2022-03-18 19:44:55 i've never used it 2022-03-18 19:45:12 I thought so, and it seems there is no alpine package watchdog. It seems to be limited in functionality… 2022-03-18 19:45:20 indeed not packaged 2022-03-18 19:46:42 I did add /dev/watchdog and the service started and rebooted the pi nearly right away. I need to dig a bit deeper it seems. 2022-03-18 19:47:04 from the description it just writes to that periodically 2022-03-18 19:47:13 i don't know why it would reboot itself on that write 2022-03-18 19:47:22 maybe you need to set some watchdog params kernel side too 2022-03-18 19:51:17 Thanks, I will try to find some more information. Mayby I just write a simple "software watchdog script" on my own and put it in /etc/periodic. That will do it probably too. 2022-03-18 20:00:35 never use linux.die.net 2022-03-18 20:01:34 it is extremely outdated, the formatting is broken, and it doesn't cite any sources 2022-03-18 20:10:54 I nowadays use man.archlinux.org 2022-03-18 20:16:28 Thanks, noted. 2022-03-18 20:40:17 man.voidlinux.org is probably closer to alpine 2022-03-18 20:55:21 will 'need dbus' in a depend cause dbus to start, or cause a service to wait? 2022-03-18 20:55:59 wait for start 2022-03-18 20:56:23 this doesn't necessarily mean anything in openrc because 'started' does not mean 'ready' in any way whatsoever 2022-03-18 20:56:37 just race conditions piled on race conditions 2022-03-18 20:56:59 heh 2022-03-18 20:57:07 but it does guarantee it will consider the needed thing 'started' before attempting to start 2022-03-18 20:57:28 just tried to run iwd manually and it mentioned dbus, but I dont have dbus in my default runlevel 2022-03-18 20:57:47 it does need dbus, and need dbus would pull it in 2022-03-18 20:57:54 eiwd does not need dbus as an alternative 2022-03-18 21:00:06 I dont mind dbus, just thought it might explain iwd issues, but apparently not 2022-03-19 08:28:01 Hi, is there any way to use setup-disk with a mountpoint and not insatll any bootloader? The bootloader selection logic in setup-disk seems to not allow specifying an empty bootloader on an EFI system since it will detect EFI and install grub-efi. 2022-03-19 08:33:14 ktprograms: I don't see an option to skip installing a bootloader 2022-03-19 08:34:58 ikke: I suppose the easiest solution will be to manually edit the setup-disk script? I want to experiment with dual booting by sharing the ESP and using efibootmgr. 2022-03-19 08:35:21 I suppose so as well 2022-03-19 08:35:23 https://gitlab.alpinelinux.org/alpine/alpine-conf/-/blob/master/setup-disk.in#L626 2022-03-19 08:38:18 And also remove the call to setup_bootloader_pkg, right? 2022-03-19 08:38:52 yeah, or set BOOTLOADER=" 2022-03-19 08:39:58 The problem is that setup_bootloader_pkg returns grub-efi if USE_EFI is set, which happens automatically if EFI is detected. 2022-03-19 08:41:37 Oh I can't read, the function is called select_bootloader_pkg 2022-03-19 08:43:18 ktprograms: i think `setup-disk /mnt` is what you're looking for 2022-03-19 08:43:46 i'm pretty sure it skips installing bootloader when it's an absolute path 2022-03-19 08:44:17 ptrc: It seems like it still installs a bootloader (looking at the script source) 2022-03-19 08:44:18 ptrc: Yes, I always install Alpine like so and always manually install a bootloader after that 2022-03-19 08:47:01 hm, yeah, apparently it still does that; don't know why i remembered otherwise 2022-03-19 08:48:47 begs: On a UEFI system even if you set BOOTLOADER="" it will still install the grub-efi package (not really a big deal but I'd rather it didn't) 2022-03-19 08:54:50 "Maxim Karasev: On a UEFI..." <- Oh, I should read the script then to get better understanding 2022-03-19 08:55:55 begs: It's at https://gitlab.alpinelinux.org/alpine/alpine-conf/-/blob/master/setup-disk.in#L643, that function is called regardless of if you specify a device or mountpoint. 2022-03-19 08:58:34 ktprograms: Yes, now I see. I thought the talk was about installing the bootloader to disk, not just it's package 2022-03-19 08:59:16 begs: Ideally I wouldn't want it to install the bootloader package, but yes that's less bad than it doing the bootloader setup 2022-03-19 09:03:49 ktprograms: it shouldn't be very hard to modify the script for desired behavior and send a merge request 2022-03-19 09:27:00 Why is acct installed by setup-disk? 2022-03-19 11:54:39 one ast time before i give up: https://pastebin.com/raw/dvr4Xbwn 2022-03-19 11:57:39 So you are just asking why you need to enable udev-settle? 2022-03-19 11:58:36 THe 'drawback' is that it delays to boot until everything has been processed 2022-03-19 12:00:35 if that's the only drawback. udev-settle starts anyway if i do not add it to a runlevel. something needs/want it. but it seems to be "too late". 2022-03-19 12:01:01 so, why? 2022-03-19 12:03:28 I would not know the answer myself 2022-03-19 12:04:25 thx anyway ikke. but it's frustrating. 2022-03-19 12:06:50 i will add udev-settle to sysinit in my public script too. dont care about "boot time". it should _work_ 2022-03-19 12:06:58 Yes, understood 2022-03-19 12:15:50 anyone using wireguard with ifupdown? I get: Line unrecognized: `Address=192.168.20.2/24' 2022-03-19 12:17:54 it seems output of: wg setconf wg0 /etc/wireguard/wg0.conf 2022-03-19 12:19:54 it seems that wg/wg-quick treat config file differently 2022-03-19 12:22:01 I just tweaked ifupdown for use wg-quick, http://ix.io/3SQS 2022-03-19 12:24:09 Address is a wg-quick thing and not wireguard 2022-03-19 12:24:44 you can add `address xyz/32` or whatnot as part of the iface definition in /etc/network/interfaces 2022-03-19 12:25:05 https://img.ayaya.dev/qvGj2GB1YgfY looks like this, like what you did 2022-03-19 12:25:09 yeah I was reading the wireguard quickstart 2022-03-19 12:25:33 I suppose that 'ideally' would be handle all in the same place 2022-03-19 12:27:33 I jus thave: http://ix.io/3SQU 2022-03-19 12:27:51 and thw ifupdown tweaked for call wg-quick 2022-03-19 12:29:16 maybe could be an additional , use wireguard_quick 2022-03-19 12:38:22 donoban: there is a wireguard executor 2022-03-19 12:38:24 https://github.com/ifupdown-ng/ifupdown-ng/blob/main/executor-scripts/linux/wireguard 2022-03-19 12:38:26 use wireguard 2022-03-19 12:39:27 what was the command to update grub config on alpine 2022-03-19 12:46:04 grub-mkconfig -o /boot/grub.cfg ? 2022-03-19 12:46:35 I've already cheated and used mkinitfs trigger from apk 2022-03-19 12:47:46 even better 2022-03-19 12:51:41 is it not possible to use PARTUUID on Alpine? 2022-03-19 12:52:30 I'm using it, without an initramfs 2022-03-19 12:56:42 yeah ikke , problem is that use wireguard relies on 'wg setconf' which seems incompatible with wg-quick options 2022-03-19 12:57:05 so maybe could be useful to have two differnets, 'use wireguard' and 'use wireguard-quick' 2022-03-19 13:04:47 donoban: I'd say, open a feature request / pull request there 2022-03-19 13:26:34 0 2022-03-19 13:33:47 -1 2022-03-19 13:40:01 NaN 2022-03-19 13:40:18 ACTION divides by 0 2022-03-19 13:41:25 sqrt(-2) 2022-03-19 13:42:38 ACTION kills SIGFPE to everybody involved 2022-03-19 13:46:17 ACTION coredumps 2022-03-19 13:49:28 ` 2022-03-19 15:16:03 i just do soft floats :p 2022-03-19 20:36:05 ikke: ok 2022-03-19 21:14:33 what package is "ss" in ? 2022-03-19 21:15:24 https://pkgs.alpinelinux.org/contents?file=ss&path=%2Fsbin&name=&branch=edge 2022-03-19 21:15:37 ahh edge only got it , thanks 2022-03-19 21:15:48 wait no 2022-03-19 21:15:53 iproute2-ss 2022-03-19 21:15:55 thanks 2022-03-20 12:52:09 I'm trying to install Alpine in "data" mode on a Rasbperry Pi 4 and the wiki isn't really helping. The `setup-disk` command demands an entire disk and doesn't accept a single partition to setup `/var`. I guess I can just create `/var` manually and make relevant entries in `/etc/fstab`. However, the `setup-lbu` commands seems to want a dir starting with `/media`. Can I change this to something like `/var/lbu`? 2022-03-20 12:52:45 There is `LBU_BACKUPDIR` but I don't understand its purpose. 2022-03-20 12:53:23 Also, the lbu script itself seems to want `/media` in some places such as https://github.com/alpinelinux/alpine-conf/blob/master/lbu.in#L593 2022-03-20 12:54:18 Unlike what's written here https://github.com/alpinelinux/alpine-conf/blob/master/lbu.in#L567 2022-03-20 13:38:12 Hello 2022-03-20 13:39:20 I think that one of grub config file has a shebang calling for bash, so updating grub fails when bash isn't installed 2022-03-20 13:40:17 ayushnix: if you want to take a look in the process of looking the ovl see https://gitlab.alpinelinux.org/alpine/mkinitfs/-/blob/7095d01d069a62541d18a4b0aa31617efd88b815/nlplug-findfs/nlplug-findfs.c 2022-03-20 13:40:57 /etc/grub.d/35_fwupd:#! /bin/bash 2022-03-20 13:40:57 nuc [~]# grep bash /etc/grub.d/* 2022-03-20 13:41:12 I think that it should found it if you save it on top of your data partition regardless fstab entries but I'm not reallly sure 2022-03-20 13:42:03 So my upgrade to 3.15.1 failed until I installed bash 2022-03-20 13:42:12 probably the hardcoded /media could be considered a bug 2022-03-20 13:42:18 raspbeguy: can you open an issue for this? 2022-03-20 13:42:28 Ok 2022-03-20 13:43:20 ikke: on aports repo ? 2022-03-20 13:43:27 raspbeguy: yes please 2022-03-20 13:43:42 donoban: unfortunately, I don't really understand/write C so I don't understand that file 2022-03-20 13:44:35 donoban: I'll try setting LBU_BACKUPDIR to `/var/lbu` and make `/var` mount at boot using an entry in `/etc/fstab` 2022-03-20 13:45:27 but looking at the lbu script, it seems that it assumes in a few places that `/media` is being used as the dir to store overlay 2022-03-20 13:46:10 the difference witn LBU_MEDIA and LBU_BACKUPDIR seems that LBU_BACKUP dir assumes just a directory on some already mounted device 2022-03-20 13:46:25 read /etc/lbu/lbu.conf 2022-03-20 13:47:06 I suppose that if something fails with your setup it should be simple to fix 2022-03-20 13:47:14 your plan here is to put lbu on /var, which is mounted by /etc/fstab, which is in lbu? 2022-03-20 13:49:44 Hello71: oops, didn't realize this 2022-03-20 13:49:51 not sure if it'll work 2022-03-20 13:49:55 maybe not 2022-03-20 13:50:53 so if it doesn't work, this means that using "data" mode on alpine needs two storage devices? one for `/var` and another for lbu? 2022-03-20 13:53:28 yes, typically you have a usb storage device to boot from, and a disk for additional data 2022-03-20 13:55:18 ikke: https://gitlab.alpinelinux.org/alpine/aports/-/issues/13616 2022-03-20 13:58:21 ikke, yeah, using `/var` for lbu didn't, I'll use two storage devices. 2022-03-20 13:58:26 *didn't work 2022-03-20 13:58:30 thanks 2022-03-21 03:19:27 hi 2022-03-21 03:19:43 looks like the latest edge kernel broke alsa/pulse 2022-03-21 03:20:22 weirdly enough audio works fine when connected to bluetooth tho 2022-03-21 03:20:28 no idea what's going on 2022-03-21 03:24:18 well kernel audio driver problem should not affect bluetooth as it is a "software" sound card vs. kernel drivers for specific hardware.. 2022-03-21 03:26:27 tsuki: what audio hardware do you have? 2022-03-21 03:26:53 nice, thats good to know 2022-03-21 03:27:03 it's realtek audio 2022-03-21 03:27:34 had a lot of issues with pipewire, couldn't get it working at all 2022-03-21 03:27:46 it's an old laptop i got for free 2022-03-21 03:28:49 you can trying installing the previous `linux-edge` version 2022-03-21 03:31:57 i might try that tomorrow 2022-03-21 03:32:08 very tired atm 2022-03-21 03:32:20 i am also concerned this issu will persist 2022-03-21 03:34:10 this laptop does not like building stuff, otherwise i would just maintain my own kernel 2022-03-21 03:41:06 linux-lts is another option 2022-03-21 03:44:47 yea, might switch back to lts 2022-03-21 03:44:55 what are the dahdi kernels? 2022-03-21 03:46:30 nvm, i guess it's for specific hardware 2022-03-21 03:47:14 do you know what's going on with edge kernel and if the issue will persist? 2022-03-21 04:16:06 alright, ty for the assistance 2022-03-21 04:16:11 gn 2022-03-21 09:29:50 it's one of the openrc init script that mounts /run as tmpfs? 2022-03-21 09:30:30 our appliance has lots of services that use a shared /run/ourservice directory and I was guessing how can I create it at one place (to avoid adding some start_pre functions all over the places) 2022-03-21 09:38:26 maybe bootmisc 2022-03-21 09:58:55 Seems like acpid silently dies on s390x 2022-03-21 11:19:05 Ermine: yeah, noticed that on our s390x hosts as well 2022-03-21 17:08:15 I've installed Alpine Linux on my Raspberry Pi 4B in "data" mode. However, the packages which I've added aren't being installed automatically at boot. The `/etc/apk/cache` is a symlink to `/var/cache/apk` and `/var` is mounted at boot. I also added `/var/cache/apk` to `/etc/apk/repositories` but that doesn't help either. Do I have to make any other changes for the packages to get installed at boot? 2022-03-21 17:08:34 you need to do some lbu stuff 2022-03-21 17:08:48 lbu_include lbu_commit 2022-03-21 17:08:51 i think 2022-03-21 17:11:29 lbu is already setup by default to include `/etc/`. `/var` is mounted as a rw partition in "data" mode. 2022-03-21 17:12:01 I also did lbu commit before rebooting and `/etc/apk/world` has the list of packages that I installed 2022-03-21 17:12:43 The `/etc/apk/world` survives after a reboot so lbu is working fine but the packages aren't being installed automatically, psykose 2022-03-21 17:14:04 I guess that the overlay is loaded without /var mounted 2022-03-21 17:14:13 is there any error about trying to install packages? 2022-03-21 17:15:11 donoban: nope, when I do `apk -U upgrade` after a reboot, my packages get installed normally 2022-03-21 17:15:19 but they don't get installed automatically 2022-03-21 17:16:37 oh, I did add `/var/cache/apk` in `/etc/apk/repositores` and when I do `apk -U upgrade`, I get a warning saying that `ignoring /var/cache/apk: no such file or directory` 2022-03-21 17:16:48 but `/var/cache/apk` does exist 2022-03-21 17:16:54 that is not right 2022-03-21 17:16:58 remove it from repositories 2022-03-21 17:17:07 you should have a symlink on /etc/apk/cache 2022-03-21 17:18:15 donoban: ok, I removed `/var/cache/apk` from `/etc/apk/repositories`. `/etc/apk/cache` is a symlink to `/var/cache/apk`. I've done a reboot. 2022-03-21 17:18:26 after doing `lbu commit` 2022-03-21 17:18:39 maybe you could tweak the fstab on the initramfs 2022-03-21 17:18:53 but my packages still aren't installed automatically, donoban 2022-03-21 17:18:58 but I don't know if apkovl will be handled before mounting it 2022-03-21 17:20:22 donoban: this is how my `/etc/fstab` looks like 2022-03-21 17:20:23 https://0x0.st/oNW-.txt 2022-03-21 17:21:17 I'm not using an initramfs I think 2022-03-21 17:21:52 unless the rootfs extracted inside `/media/mmcblk0p1` counts 2022-03-21 17:22:34 ah sorry, you are on modloop 2022-03-21 17:23:55 yeah, I think 2022-03-21 17:24:19 does that affect automatic installation of packages at boot? 2022-03-21 17:24:30 I feel that /var is not moutend yet when the lbu is loaded 2022-03-21 17:24:42 yeah, that's what I was thinking 2022-03-21 17:24:49 but I'm not sure how to debug the boot process 2022-03-21 17:24:58 `/var` is a USB flash drive and lbu is in a SD card 2022-03-21 17:25:23 after booting, apk fix installs everything properly¿ 2022-03-21 17:26:26 donoban: it does, but I'd like to solve this issue 2022-03-21 17:26:41 and not execute `apk fix` or `apk update` after boot 2022-03-21 17:26:42 yeah, just for confirm it 2022-03-21 17:45:02 does the apk cache need to be on the storage device where the lbu data is stored? right now, my apk cache is in `/var/apk/cache`, which is a different storage device 2022-03-21 17:45:44 probably is the most common but I don't think it's a enforced requeriment 2022-03-21 17:46:54 do you have enough space for move the cache on the other device? 2022-03-21 17:46:55 the `setup-apkcache` script chooses `/var/cache/apk` and `setup-lbu` chooses `/media/mmcblk0p1` by default, different storage devices essentially 2022-03-21 17:47:13 donoban: yeah, both storage devices are 64G 2022-03-21 17:48:02 hmm, so try to put together 2022-03-21 17:48:25 maybe I'm spoiled but I must say, the alpine wiki is extremely bad and sometimes reads like someone's personal journal rather than a wiki 2022-03-21 17:48:36 it basically is a personal journal 2022-03-21 17:49:06 my personal journal was move to sys-install :P 2022-03-21 17:49:29 nah, at that point, might as well install openbsd 2022-03-21 17:49:56 i was considering installing openbsd but switched to alpine because of the "officially" supported diskless and data mode setups 2022-03-21 17:50:54 i am not sure what openbsd has in common 2022-03-21 17:51:31 ayushnix: it's probably easy to fix 2022-03-21 17:52:36 psykose: I'm installing alpine on a rpi for a specific purpose, pi hole/adguard. I think both openbsd and alpine are suitable for that purpose but I preferred alpine because of the ram disk installation seemed relatively straightforward. 2022-03-21 17:53:08 donoban: yeah, i'll try moving the apk cache to the lbu storage device and report back in a few minutes after I've had dinner 2022-03-21 17:54:03 I meant with the cache on the other device, but if you have enough space 2022-03-21 18:20:47 donoban and psykose, yeah, moving the apk cache to the same storage device where lbu backup is kept worked 2022-03-21 18:21:32 should've been documented on the wiki 2022-03-21 18:21:52 the handbook looks decent but it doesn't cover diskless and data installation modes 2022-03-21 19:56:13 Not sure if this will be helpful but after going through the wiki and installing Alpine on a Raspberry Pi 4, I decided to write a short document so that I won't be confused the next time I do this 2022-03-21 19:56:15 https://gist.github.com/ayushnix/9641b6b843411f2b49a5571083274802 2022-03-21 20:00:16 you should put that on the wiki 2022-03-21 20:00:17 :p 2022-03-21 20:00:45 It does include personal opinions 2022-03-21 20:00:52 exactly like a journal 2022-03-21 20:00:54 perfect fit 2022-03-21 20:06:23 yeah, this gist isn't really meant to be included in a wiki and as ikke said, it does contain opinions 2022-03-21 20:07:08 but it does clarify some doubts I had about the `setup-disk`, `setup-lbu`, and `setup-apkcache` scripts and that isn't mentioned anywhere on the wiki or the handbook, as far as I know 2022-03-21 20:08:10 There is some (very slow) work being done on adding proper documentation on docs.alpinelinux.rg 2022-03-21 20:08:16 docs.alpinelinux.org 2022-03-21 20:08:26 But it's kind of stalled atm 2022-03-21 20:10:57 writing docs is certainly difficult 2022-03-21 20:11:38 We need someone pushing the effort 2022-03-21 20:12:09 But it's not always rewarding 2022-03-21 20:12:15 is anything 2022-03-21 20:12:32 Some things more than others 2022-03-21 20:13:12 psykose: mood 2022-03-21 20:21:01 ikke: does the handbook not intend to document the diskless and data mode? 2022-03-21 20:21:38 ayushnix: It should 2022-03-21 20:21:41 but does not atm 2022-03-21 20:22:34 This text in this section makes it sound that it doesn't 2022-03-21 20:22:34 https://docs.alpinelinux.org/user-handbook/0.1a/index.html#_what_is_it 2022-03-21 23:00:18 heya 2022-03-21 23:01:05 I've installed Alpine unto a nvme drive on a new Dell laptop and used the crypt option from the setup-alpine command 2022-03-21 23:02:01 and I'm being dropped on a busybox console with a "UUID=12345 could no such file or directory" 2022-03-21 23:02:52 "mounting UUID=12345 on /sysroot failed: No such file or directory" 2022-03-21 23:05:51 i remember there was some nvme crypt issue but i forgot what it was 2022-03-21 23:12:06 you should verify /etc/mkinitfs.conf after manually getting back into the drive, and that it has nvme in it (and all the other needed options) 2022-03-21 23:12:12 then `apk fix mkinitfs` will regenerate it 2022-03-21 23:18:23 ok, will try 2022-03-21 23:35:59 yeah, added nvme, keymap, apk fix mkinitfs 2022-03-21 23:36:30 it rebuilded the image, although didn't display anything on which hooks were being processed 2022-03-21 23:36:41 still the same problem :( 2022-03-21 23:40:35 is the uuid the uuid of the crypt partition or the inner partition 2022-03-21 23:56:20 psykose: does the apk fix replace the wiki step of mkinitfs -c /etc/mkinitfs/mkinitfs.conf -b / $(ls /lib/modules/) ? 2022-03-21 23:56:43 it reruns the trigger 2022-03-21 23:56:46 which does that latter step 2022-03-21 23:56:53 i just prefer it because it's much shorter and much easier to remember 2022-03-21 23:56:56 ok.. just wanted to make sure... 2022-03-21 23:57:32 I run nvme (not on alpine) and I know its a initram mess 2022-03-22 02:06:42 ok, so minimal has helped me to untangle the problem and it is actually related to nlplug-findfs not being able to detect the nvme drive (timeouting) 2022-03-22 02:07:06 when being dropped on a recovering shell and using debug_init on grub 2022-03-22 05:34:40 eletrotupi: sounds like a known issue 2022-03-22 05:35:22 #12325 2022-03-22 13:57:59 hi, I'm installing OpenZFS on my Alpine Linux box which is running edge. there doesn't seem to be a zfs-edge package but there is zfs-lts, am I supposed to install the latter? I don't think so, but there's no package for edge 2022-03-22 13:59:54 it's based on kernel not the alpine version 2022-03-22 14:00:03 zfs-lts works with linux-lts 2022-03-22 14:00:21 and there is no zfs-edge for linux-edge, but these are distinct from the actual alpine release 2022-03-22 14:04:48 yeah, I understand that but I don't understand what package I'm meant to install then 2022-03-22 14:09:05 zfs-lts 2022-03-22 14:09:33 for the linux-lts kernel 2022-03-22 14:09:40 or zfs-virt for the linux-virt kerenel 2022-03-22 14:10:03 I'm running linux-edge 2022-03-22 14:10:45 then there is no zfs module for it 2022-03-22 14:11:14 I'm supposed to compile it from scratch then, right? 2022-03-22 14:13:00 something like that 2022-03-22 14:13:47 thanks! 2022-03-22 14:17:56 zfs-src 2022-03-22 14:19:50 (together with akms) 2022-03-22 14:20:20 but zfs-src doesn't exist 2022-03-22 14:20:27 and yeah I do have akms installed for a few other modules 2022-03-22 14:20:54 hm, my mistake 2022-03-22 14:28:47 someone should make it 2022-03-22 15:01:46 I'm trying to find out why mosquitto service would crash: doas service mosquitto status gives me: * status: crashe I've checked /var/log/messages but couldn't find any logs for mosquitto. I can run mosquitto as normal user not as service and I can see stdout. But how can I debug why would mosquitto service crashed? 2022-03-22 15:03:25 frojnd: anything in dmesg? 2022-03-22 15:04:03 No nothing 2022-03-22 15:04:21 Last log: [39960.382599] tsc: Marking TSC unstable due to clocksource watchdog 2022-03-22 15:04:29 And it's the same log when I do doas service mosquitto restart 2022-03-22 15:16:00 unfortunately openrc doesn't track these things 2022-03-22 16:13:01 Is @alpine_linux_english official group in telegram? 2022-03-22 16:13:49 We have no official group on telegram, so everything is from the community 2022-03-22 16:14:04 Ok 2022-03-22 16:33:25 no, it's ran by someone that was banned from half of alpine 2022-03-22 16:34:08 can't recommend, only get headaches from that place 2022-03-22 17:01:04 it seems like alpine isn't using my 'console=' param, seems to be using tty1 instead of what I put there 2022-03-22 17:01:56 any ideas? 2022-03-22 17:02:35 j`ey: where are you specifying that? 2022-03-22 17:02:43 on the kernel command line 2022-03-22 17:03:07 Directly in the bootloader? 2022-03-22 17:03:13 oh wait, I think I get the issue 2022-03-22 17:03:16 It has always worked 2022-03-22 17:03:24 the console is fine, but it's the login screen that's wrong 2022-03-22 17:03:48 I probably need a *getty on /dev/console maybe? 2022-03-22 17:04:10 for me, it always has been enough to specify console=.. 2022-03-22 17:04:33 Did you change /etc/inittab? 2022-03-22 17:04:36 nope 2022-03-22 17:04:54 What do you provide as console? 2022-03-22 17:05:13 ttySAC0 2022-03-22 17:05:24 (a serial port) 2022-03-22 17:05:29 You can try to add that to /etc/inittab 2022-03-22 17:05:45 Oh 2022-03-22 17:05:49 Maybe just to /etc/securetty 2022-03-22 17:06:12 I think I added it to securetty already, will double check, and also tr inittab 2022-03-22 17:09:09 yeah inittab was the issue, login[627]: root login on 'ttySAC0' 2022-03-22 18:59:20 is this a known thing? https://twitter.com/kennwhite/status/1506334472765157380 2022-03-22 19:00:19 the dotnet part or the libc part 2022-03-22 19:01:09 I assume dalias is aware that trying to executing a binary with a non-existing linker would fail with file not found 2022-03-22 19:02:31 I assume many people know about "file not found" being usually different libc issue but it's such a misleading error message that I'd rather double check :) 2022-03-22 19:03:00 dalias being the author of musl, I'm sure he is aware 2022-03-22 19:03:25 but also, is that in context of dotnet package in alpine or just dotnet on alpine 2022-03-22 19:04:06 well aiui this isn't exec but dlopen 2022-03-22 19:04:29 hmm 2022-03-22 19:04:39 isn't dlopen some userspace thing? 2022-03-22 19:04:43 but i guess there's no easy way to programmatically assign a cause to dlopen failure since you don't get error codes only a string :/ 2022-03-22 19:05:03 so .NET probably just assigned "file not found" as the catch-all 2022-03-22 19:05:21 .net on alpine is quite recent 2022-03-22 19:05:35 So it does not supprise me people start running into these things 2022-03-22 19:06:31 https://gitlab.alpinelinux.org/alpine/aports/-/merge_requests/31451#note_222140 2022-03-22 19:07:18 although this seems to be some random non-dotnet dp 2022-03-22 19:07:21 s/dp/dep 2022-03-22 19:07:21 panekj meant to say: although this seems to be some random non-dotnet dep 2022-03-22 19:07:48 but I did run into issue mentioned in tweet myself when trying to package powershell 2022-03-22 19:10:01 ACTION realized why bash says 'no such file or directory' when trying to execute 32-bit binary on 64-bit architecture (not related to Alpine/musl) 2022-03-22 19:10:47 epiphany 2022-03-22 19:11:58 I heard fish in few years will emit useful error message instead of `no such file or directory` 2022-03-22 19:11:58 sadly i know not of a way around that one 2022-03-22 19:12:26 the people making the build systems fetch random glibc things from the internet and there is just no good reporting 2022-03-22 19:12:49 with that excluded, the dotnet toolchains seem to work perfectly ok 2022-03-22 19:12:53 I only know of dotnet being such 2022-03-22 19:13:07 i have ran into maybe 5 packages in aports that do similar things at build time 2022-03-22 19:13:22 completely disregarding everyone and everything, do it our way or get ****** 2022-03-22 19:13:37 for c-like stuff at least you get some fairly obvious linking errors 2022-03-22 19:13:56 ah right, isn't cmake capable of fetching random stuff from urls 2022-03-22 19:14:02 every build system is 2022-03-22 19:14:07 why.... 2022-03-22 19:14:16 well all of them support some shell() equivalent 2022-03-22 19:14:19 and the shell has wget.. 2022-03-22 19:14:34 and then people actually use it, for some reason 2022-03-22 19:14:35 hah 2022-03-22 19:14:53 i meant native capability 2022-03-22 19:15:00 ah 2022-03-22 19:15:01 not sure then 2022-03-22 19:15:13 but the cmake projects i saw doing were doing the shell wget thing 2022-03-22 19:15:35 cmake apparently has FetchContent? 2022-03-22 19:16:08 https://cmake.org/cmake/help/latest/module/FetchContent.html 2022-03-22 19:16:50 yessss 2022-03-22 19:16:53 Secooreety 2022-03-22 19:17:48 meson does not appear to have anything built-in 2022-03-22 19:17:55 at least nothing that a cursory search returns 2022-03-22 19:18:29 meson's dependency() relies on cmake or pkg-config afaik 2022-03-22 19:18:30 Oh, it has 2022-03-22 19:18:33 meson has something called wrapdb 2022-03-22 19:18:37 yeah 2022-03-22 19:18:41 https://mesonbuild.com/Wrap-dependency-system-manual.html 2022-03-22 19:20:02 Also people who publish glibc binaries think that glibc is the only linux libc, I suppose 2022-03-22 19:20:51 meson also has --wrap-mode=nodownload to completely disable that and tell you 2022-03-22 19:21:12 which is by far the friendliest version of such a thing i've seen 2022-03-22 19:21:26 should be default imo 2022-03-22 19:21:35 it is set with abuild-meson 2022-03-22 19:21:43 nice :) 2022-03-22 19:24:47 insane idea: abuild could disable network during prepare() and build() stages 2022-03-22 19:24:58 rootbld already does that 2022-03-22 19:26:18 we still need net in prepare() until we get 2-staged prepare() for rust/go 2022-03-22 19:26:32 panekj: we need it especially in prepare then 2022-03-22 19:27:05 unless we would get dedicated fetch stages for go and rust, but not sure if we want to hardcode these things in abuild 2022-03-22 19:28:02 sorry, right, 2 staged fetch 2022-03-22 19:28:34 can't think when hungry 2022-03-22 19:29:32 The 2nd fetch stage would need to run after unpack 2022-03-22 19:29:43 and after deps 2022-03-22 19:37:30 Is there a source for abuild-keygen and abuild-sign that I can use on a not-alpine linux os? 2022-03-22 19:39:59 The source is here: https://gitlab.alpinelinux.org/alpine/abuild/-/blob/master/abuild-sign.in 2022-03-22 19:40:09 https://gitlab.alpinelinux.org/alpine/abuild/-/blob/master/abuild-keygen.in 2022-03-22 19:40:39 Sounds good. Are there any deps, or can I just build & run? 2022-03-22 19:40:47 they are shell scripts 2022-03-22 19:41:15 It has internal dependencies 2022-03-22 19:41:25 But I don't see why you cannot build abuild on other platforms 2022-03-22 19:41:40 Sounds good. Thanks! 2022-03-22 21:16:16 I successfully built the abuild repo, and it seems to work - however, the index file is still untrusted 2022-03-22 21:16:57 I'm using it in a network boot context, so maybe there's something else going on? 2022-03-22 21:17:31 I assume you added the key to /etc/apk/repositories? 2022-03-22 21:17:49 also make sure the key names correspond 2022-03-22 21:17:56 no. Guess I should add that to the apkovl 2022-03-22 21:18:00 Thanks 2022-03-22 21:18:08 sorry 2022-03-22 21:18:10 /etc/apk/ekys 2022-03-22 21:18:12 /etc/apk/keys 2022-03-22 21:18:15 not repositories 2022-03-22 21:18:18 lol thanks 2022-03-22 21:34:12 Looks like the apkovl adds my repo key to the /sysroot, but inital package download & installation happens before the system re-roots to /sysroot 2022-03-22 21:34:44 The process works if I just use the apkindex downloaded from the main alpine repo, so I know it's an issue with my custom index 2022-03-22 21:35:52 unless I'm missing something, it looks like I'd have to modify the initramfs directly, or just stick with the official apkindex 2022-03-22 21:36:21 Is there anything I'm missing? 2022-03-22 21:41:26 Hi all :-) 2022-03-22 21:42:45 I'm having permission denied error on a build in CI. https://gitlab.alpinelinux.org/kmmndr/aports/-/jobs/669245 2022-03-22 21:42:51 Any idea ? 2022-03-22 21:43:52 yeah, it should not try to modify any system files.. 2022-03-22 21:44:11 Not sure why it would try to: 2022-03-22 21:44:13 rm: cannot remove './/usr/lib/ruby/3.1.0/rubygems/ssl_certs/rubygems.org/GlobalSignRootCA.pem': Permission denied 2022-03-22 21:44:29 because it's in package() ? 2022-03-22 21:44:49 # rm -f ./$_rubydir/rubygems/ssl_certs/*/*.pem 2022-03-22 21:44:49 There is a "cd "$pkgdir" just before that command 2022-03-22 21:45:07 don't use ./ 2022-03-22 21:45:17 don't cd, you can just rm "$pkgdir"/... 2022-03-22 21:45:28 And the APKBUILD file is very close to main/ruby package, which seems to build 2022-03-22 21:45:42 you rm from $_rubydir 2022-03-22 21:45:48 which is an absolute path 2022-03-22 21:46:03 _rubydir="/usr/lib/ruby/$_abiver" 2022-03-22 21:46:08 you need to prefix it with $pkgdir 2022-03-22 21:46:48 As there is a cd before, it leads to `rm .//usr/...` 2022-03-22 21:47:17 In the current working directory of $pkgdir 2022-03-22 21:48:30 But I'll follow your tips and re-organize these lines, thank you :-) 2022-03-22 21:49:15 it seems like main/ruby APKBUILD is quite old-ish and should be corrected as well 2022-03-22 21:49:50 but since it works, probably noone bothered to do so 2022-03-22 21:52:32 I do not understand why it worked for main/ruby as the lines do comes from that APKBUILD 2022-03-22 21:55:29 maybe something changed between 3.0.3 and 3.1.1 version 2022-03-22 22:03:55 or maybe it just broke? 2022-03-22 22:04:48 main/ruby fails for me on 'cannot create dir ... pkg/ruby/usr/share/licences: permission denied' 2022-03-22 22:12:19 panekj, thank you very much for you help ! :-) I'll keep on modernizing APKBUILD 2022-03-22 22:14:40 kmmndr: https://0x0.st/oNYr.txt cleaned up a bit APKBUILD 2022-03-22 22:15:39 I presume it cannot delete files because the owner of them is root 2022-03-22 23:23:45 I have a diskless install with 460MB of ram, and 230.0M of tempfs usage. When I try to install a package I get "ERROR: Failed to create some/path/: No space left on device". I can create a 100 MB file on the tempfs with no errors. is there a way to resolve this or do I have to use a sys mode install? 2022-03-22 23:39:06 goblins: try running `df -i` 2022-03-22 23:40:53 just 2% on /. 2022-03-22 23:46:04 hu, apk is using a ton of memory when installing the package. 2022-03-23 01:42:59 hi are packages supposed to be reinstalled after rebooting when using a data disk mode? i have my cache directory setup, and i'm committing changes before i reboot 2022-03-23 01:46:08 i have lbu.conf configured to point to the mapped data drive as well. tried enabling the lbu commit limit options too since lbu list-backup doesn't see apkovl files that don't include a timestamp 2022-03-23 01:47:24 maybe i'm confused about what data disk mode is intended to do 2022-03-23 01:57:45 bbiab 2022-03-23 05:26:33 Hopefully s6 will; Hello71 | unfortunately openrc doesn't track these things 2022-03-23 05:43:15 Hm 2022-03-23 05:43:38 I've purged mosqutto files and reinstall it so I can have a clean config file but service crashes 2022-03-23 05:43:58 Clean as in default config file 2022-03-23 05:45:54 frojnd: how soon does it crash 2022-03-23 05:45:55 Ah it was because pid was still there 2022-03-23 05:46:01 I did restart service and now status 2022-03-23 05:46:08 And it says started 2022-03-23 05:46:13 So it must have been my config file 2022-03-23 05:47:21 I can work from here by backuping config file before I modify it to my needs 2022-03-23 06:05:57 I heave found the culprint 2022-03-23 06:06:25 If I change log file dest from sys to something like: /var/log/mosquitto/mosquitto.log it crashes 2022-03-23 06:06:54 Not sure if this is openrc thing or not that everyting goes into /var/log/messages? 2022-03-23 06:07:39 everything that sends messages to syslog ends up in /var/log/messages 2022-03-23 06:07:54 does the user the service runs as have access to /var/log/mosquitto? 2022-03-23 06:09:42 I have to read it with doas 2022-03-23 06:10:08 By use you mean user who executes service mosquitto start ? 2022-03-23 06:10:16 s/use/user 2022-03-23 06:10:16 frojnd meant to say: By user you mean user who executes service mosquitto start ? 2022-03-23 06:11:51 frojnd: it runs as the mosquitto user 2022-03-23 06:12:04 so you need to make sure that user can write to /var/log/mosquitto/ 2022-03-23 06:12:33 Ok thank you 2022-03-23 06:29:27 doas chown -R mosquitto:mosquitto /var/log/mosquitto 2022-03-23 06:29:51 ls -l /var/log gives me: drwxr-xr-x 2 mosquitt mosquitt 2 Mar 23 07:24 mosquitto 2022-03-23 06:29:55 mosquitt ? 2022-03-23 06:30:10 ACTION scratches head 2022-03-23 06:46:47 busybox ls(1) trims user/group names 2022-03-23 06:47:09 try stat(1) 2022-03-23 07:00:54 Ah ok all good 2022-03-23 08:12:49 I have a teltonika router that runs a DNS (dnsmasq) server which works fine but even though I specify that address into my alpine machine's /etc/resolv.conf any attempt to resolv a domain (using curl/wget/ping) ends in 'bad address' error 2022-03-23 08:13:01 using nc on port 53 I can connect as well 2022-03-23 08:13:58 what does dig say? Try to ssh to router and resolve with nslookup from there 2022-03-23 08:18:21 https://bpa.st/NKFA 2022-03-23 08:19:25 so it looks like the connection works but the router's dnsmasq refuse any request 2022-03-23 08:19:48 but I have another router at work that works fine, so I'll have to find out over the ten thousand billion web admin pages ot the router why it fails 2022-03-23 08:23:25 markand: is dnsmasq listening on 0.0.0.0? 2022-03-23 08:24:51 yes 2022-03-23 08:24:51 hmm the only option I could find that makes dnsmasq refuse queries is --connmark-allowlist 2022-03-23 08:25:00 does the router use that maybe 2022-03-23 08:26:01 https://bpa.st/KAPA 2022-03-23 08:26:05 this is the dnsmasq configuration 2022-03-23 08:36:04 okay I think I got it, they have an option "disable DNS rebind protection" 2022-03-23 09:09:50 I'm trying to upgrade my aarch64 vm, and I get: "ERROR: https://mirror.yandex.ru/mirrors/alpine/v3.15/main: UNTRUSTED signature" 2022-03-23 09:12:29 Ermine: what about a different mirror? 2022-03-23 09:13:19 dl-cdn.alpinelinux.org gives the same error 2022-03-23 09:13:34 (Current version there is 3.14) 2022-03-23 09:16:25 Oh 2022-03-23 09:20:03 I've upgaded and fetched new ca-certificates bundle, and now everything works 2022-03-23 09:20:54 Ermine: did it pull in a new apk-keys? 2022-03-23 09:21:30 Apparently yes 2022-03-23 09:24:17 for 3.15, we started using new keys 2022-03-23 09:35:36 Ermine: for reference: https://wiki.alpinelinux.org/wiki/Release_Notes_for_Alpine_3.15.0#New_package_signing_keys 2022-03-23 09:38:41 Hi all :-) 2022-03-23 09:39:03 Is there something wrong with Alpine CI runners ? 2022-03-23 09:39:22 I've lots of Permission denied errors 2022-03-23 09:40:54 https://gitlab.alpinelinux.org/kmmndr/aports/-/jobs/669457 2022-03-23 09:49:58 It fails at https://git.alpinelinux.org/abuild/tree/abuild.in#n1913 which seems independent from APKBUILD file 2022-03-23 14:20:23 Congrats on new release! 2022-03-23 14:20:28 ACTION celebrates 2022-03-23 15:12:15 hello 2022-03-23 15:12:29 Hi 2022-03-23 15:12:51 what is the git repository for this distro? 2022-03-23 15:14:00 https://gitlab.alpinelinux.org/alpine 2022-03-23 15:15:02 aports are where the packages are 2022-03-23 15:16:18 ok 2022-03-23 15:16:38 https://gitlab.alpinelinux.org/alpine/apk-tools 2022-03-23 15:17:02 is apk written entirely in C? 2022-03-23 15:17:09 yes 2022-03-23 15:17:33 is there any tutorial on how to compile it locally? 2022-03-23 15:18:53 daniel_tirzuman: you can see here how it's compiled: https://gitlab.alpinelinux.org/alpine/aports/-/blob/master/main/apk-tools/APKBUILD 2022-03-23 15:21:45 hmm, so I want to checkout the apk repo then build it from the source code 2022-03-23 15:22:08 so should I use the Makefile via the make command? 2022-03-23 15:22:11 the APKBUILD lists the dependencies 2022-03-23 15:22:32 daniel_tirzuman: in the master branch (for APKv3), they switched to meson 2022-03-23 15:23:36 oh, I see 2022-03-23 15:24:26 so basically I need lines 13,14,15 on my host os 2022-03-23 15:24:32 then I can do a meson build 2022-03-23 15:24:47 and I should have the generated apk binary 2022-03-23 15:25:09 *12,13,14 2022-03-23 16:07:06 is there something wrong with this openrc init script? The service fails to start saying that 'unknown option' followed by the command_args list but if I start it manually, it works as expected. 2022-03-23 16:07:09 https://0x0.st/oNIe.txt 2022-03-23 16:09:48 /var/adguard/adguardhome seems like a strange place to put an executable 2022-03-23 16:10:32 aside from that i would guess it's the checkconfig that fails 2022-03-23 16:10:33 some of that could be uplifted to `/etc/conf.d` 2022-03-23 16:11:40 psykose: the adguard binary complains that the arguments passed are unknown 2022-03-23 16:12:03 don't quote "$command_args" 2022-03-23 16:12:17 perhaps that 2022-03-23 16:14:12 psykose, thanks, that worked 2022-03-23 16:14:20 well, that was weird 2022-03-23 16:14:33 why would the initscript fail because of the quotes? 2022-03-23 16:14:45 because it's passed as one argument if you quote it like that to the command 2022-03-23 16:14:58 so the "-lots -of -args" is indeed not a valid argument 2022-03-23 16:15:06 ah, right 2022-03-23 16:17:14 psykose: should I place the adguard binary in /usr/bin if I'm using data mode on an rpi? Add /usr/bin to lbu to preserve the binary? 2022-03-23 16:17:31 maybe, if you want 2022-03-23 16:18:19 hmm, /var seems fine for now because that's on a flash drive and the lbu backups are on a sdcard 2022-03-23 16:19:01 sure 2022-03-23 17:21:15 is blender known to be broken? 2022-03-23 17:21:16 i get 2022-03-23 17:21:20 : CommandLine Error: Option 'use-dbg-addr' registered more than once! 2022-03-23 17:21:21 LLVM ERROR: inconsistency in registered CommandLine options 2022-03-23 17:21:36 which i think means one of the .so's it loaded has a copy of LLVM.so static linked into it 2022-03-23 17:21:46 in addition to a copy being dynamically loaded 2022-03-23 17:22:19 No issue open for it 2022-03-23 17:26:38 is it possibly specific to my system? :( 2022-03-23 17:26:57 maybe some lib that should be upgraded but wasn't, or only certain graphics driver? 2022-03-23 17:37:25 mm 2022-03-23 17:37:33 i've seen that before on something else 2022-03-23 17:37:51 strange, it opens fine for me 2022-03-23 17:38:12 the other thing was mpv when i was doing some local llvm13 rebuilds, which doesn't statically link it at all 2022-03-23 17:38:27 it was magically fixed when everything was llvm13 in edge 2022-03-23 17:42:37 yeah, 3.1.0-r1 seems to work fine on edge 2022-03-23 17:42:38 strange 2022-03-23 18:36:18 I'm digging around in the alpine xen install iso, and I'm noticing multiple variations of xen.gz: xen.gz, xen-4,gz, xen-4.15.gz, and xen-4.15.1.gz 2022-03-23 18:36:37 What's the difference between these? 2022-03-23 18:41:19 the filename 2022-03-23 18:45:44 heh. that makes it easy. thanks. 2022-03-23 20:06:32 ok something using LLVM-12 is being dlopened... 2022-03-23 20:16:12 that's weird 2022-03-23 20:16:23 just pulling in blender doesn't pull in llvm12-libs at all 2022-03-23 20:17:15 maybe i missed a rebuild regardless 2022-03-23 21:32:36 does iptables -m state not work on alpine by default? 2022-03-23 21:32:45 i see there's a conntrack-tools in the repo 2022-03-23 21:41:38 well that package doesn't seem to help 2022-03-23 21:43:26 nm found the problem 2022-03-23 22:05:19 you didn't reboot after updating 2022-03-24 00:31:48 i did actually. my related,established line was missing the -j ACCEPT at the end... staring at iptables for too long i guess 2022-03-24 02:10:06 Hi all, can anyone here tip me off as to how to make crond email me reports? I found this twelve-year-old inquiry from ncopa that suggests it must be possible, somehow: https://busybox.busybox.narkive.com/461ZB3oi/cron-does-not-work-with-sendmail 2022-03-24 02:10:43 (I have postfix installed and operational.) 2022-03-24 03:06:27 birdcart: when you say reports are you refering to stdout and/or stderr of a cron job? 2022-03-24 03:08:27 for that you can add MAIL='user/address' to the top of a crontab entry to enable mailing the stderr/stdout to the specified user and or address. This is sent by crond using the `sendmail` command 2022-03-24 03:10:00 smcavoy: Yep, that's exactly what I'm refering to. Looks like you answered my question. Thank you! 2022-03-24 03:16:27 np 2022-03-24 03:17:15 smcavoy: Doesn't seem to work. I added the line to the top of /etc/crontabs/root. Did you mean a different entry? 2022-03-24 03:18:00 Or… maybe it works but postfix isn't confomant to the sendmail api in some crucial way. 2022-03-24 03:18:13 Any idea how to debug this? 2022-03-24 03:21:12 Just realizing I have a sendmail installed, and it's not postfix, which I hadn't realized until now. I guess alpine comes with sendmail stock. 2022-03-24 03:22:06 I take it back. It's installed by postfix. So this should work. I think you must have meant a different file. 2022-03-24 03:26:32 Confirmed that I can send via sendmail manually. 2022-03-24 03:30:21 you should edit crontabs with `crontab -e -u [user]` 2022-03-24 03:33:30 crond expects to use the command + arg `sendmail -ti` see https://git.busybox.net/busybox/tree/miscutils/crond.c#n98 2022-03-24 03:37:14 postfix installs its own sendmail at /usr/sbin/sendmail 2022-03-24 03:37:18 Yes, `crontab -e` safely writes to that file. I verified that `sendmail -ti` works. I suspect I have mis-sepcified that MAIL= setting. 2022-03-24 03:37:42 And yes, that postfix pseudo-sendmail is what I have. 2022-03-24 03:37:45 my bad 2022-03-24 03:38:31 MAILTO='user/address' not MAIL 2022-03-24 03:39:06 Yep, I just saw that in the source 2022-03-24 03:46:05 smcavoy: it worked! Thank you so much. 2022-03-24 03:47:58 np 2022-03-24 07:32:08 hello 2022-03-24 07:32:27 I have a question about VLC in the latest Alpine (3.15.2) 2022-03-24 07:32:50 I am trying to setup a "kiosk" machine with xf86+vlc 2022-03-24 07:33:25 everything works as expected but ti seems the vlc shipped with Alpine is missing srt support 2022-03-24 07:34:37 the version of VLC is OK (3.0.13-8) but if I do "vlc --list" I don't have the srt module 2022-03-24 07:35:03 and I need it 2022-03-24 07:35:11 should I recompile it ? 2022-03-24 07:36:36 or is in some apk package I am missing ? 2022-03-24 07:54:19 wyk72: i remember making a vlc "kiosk" with alpine a decade ago or so :) 2022-03-24 07:55:50 wyk72: do you know what teh compile option for srt support is? I think we should enable it in the alpine vlc package if it isnt 2022-03-24 07:57:40 I do not have the sources of vlc handy at the moment 2022-03-24 07:57:54 srt is quite a new protocol for streaming 2022-03-24 07:58:25 but I worked around it using ffmpeg, basically I connect to the stream with ffmpeg and stream it to an udp local socket 2022-03-24 07:59:15 thing is: I can't get any output from vlc 2022-03-24 07:59:29 checking for SRT... no 2022-03-24 08:01:26 something like "ffmpeg -re -i srt://192.168.111.142:4400 -c:v copy -c:a copy -f mpegts udp://localhost:65000" 2022-03-24 08:01:59 but if I do "vlc udp://@:65000" I get no output 2022-03-24 08:02:04 it usually works 2022-03-24 08:02:31 i can confirm that the vlc package in alpine is not built with srt support. I'm gonna fix it 2022-03-24 08:02:50 ncopa: thanks a million 2022-03-24 08:03:07 im gonna have breakfast first. can you please ping me in an hour or so? 2022-03-24 08:03:54 ncopa: btw, I built a video encoder box with alpine-edge, and works exceptionally well, ffmpeg w hw acceleration vaapi etc 2022-03-24 08:04:15 ncopa: with an ancient Ivy Bridge i3 machine 2022-03-24 08:04:28 ncopa: and booting from an usb key :) 2022-03-24 08:14:49 cool 2022-03-24 08:15:22 i pushed vlc update right now to edge. should be available from the mirrors within 15 mins or so 2022-03-24 08:21:29 btw, the "workaround" works, using ffmpeg, but strangely I have to use the ip address of the machine instad of "localhost" 2022-03-24 08:21:44 "ffmpeg -re -i srt://192.168.111.142:4400 -c:v copy -c:a copy -f mpegts udp://localhost:65000" - not working 2022-03-24 08:22:01 "ffmpeg -re -i srt://192.168.111.142:4400 -c:v copy -c:a copy -f mpegts udp://:65000" -> working 2022-03-24 08:22:07 weird 2022-03-24 08:22:52 maybe ffmpeg tried to resolve "localhost" (weird) 2022-03-24 08:30:35 I've also discovered that with clever use of the ffmpeg "tee" output you can avoid all the L2/multicasting mess, and use normal/dumb switches to duplicate the stream to a bunch of other machines 2022-03-24 08:30:56 did not think it was even possible 2022-03-24 09:53:10 If I want to package soundfont file for fluidsynth, https://github.com/FluidSynth/fluidsynth/blob/master/sf2/COPYRIGHT.txt, how should I do it? It's just one file to be placed at /usr/share/soundfonts/ and a copyright file. 2022-03-24 10:56:32 dhruvin: I would try to do as a fluidsynth subpackage, or just add to it if it's not too big 2022-03-24 10:57:05 hM, is there an incompatible license? 2022-03-24 10:57:49 The soundfont doesn't have a source, it's from another author, as mentioned in the copyright. 2022-03-24 10:58:54 I think I should create a separate package `soundfont-vintage-dreams-waves` from its original source https://analoguesque.x10host.com/SoundFonts/ 2022-03-24 10:59:22 ah great 2022-03-24 10:59:23 this way people can install the fonts without its association with fluidsynth 2022-03-24 11:01:44 I intend to package all of the author's fonts (I use some). Do we have a policy on how to name a soundfont package, like soundfont-* (like we have with font-* packages). Also, is this path /usr/share/soundfonts/.sf2 okay? or should we go with debian way of putting the fonts at /usr/share/sounds/sf2/.sf2? 2022-03-24 11:02:36 I don't think that there is some policy 2022-03-24 11:05:28 okay, thanks for your inputs :) 2022-03-24 11:08:26 I tried to search for /usr/share/sounds but it seems that something fails on https://pkgs.alpinelinux.org/contents 2022-03-24 11:09:03 hmm, it work snow 2022-03-24 11:09:08 works now* 2022-03-24 11:10:25 https://pkgs.alpinelinux.org/contents?path=%2Fusr%2Fshare%2Fsound%2A&page=4&arch=x86_64&branch=edge I think that there aren't any soundfonts or similar 2022-03-24 11:11:14 https://pkgs.alpinelinux.org/contents?file=*.sf2&path=&name=&branch=edge&arch=x86_64 2022-03-24 11:11:55 there are two packages with .sf2 files, both put them on their own /usr/share/XXXX 2022-03-24 11:37:13 fluidsynth developers recommend the former path for external soundfont files https://github.com/FluidSynth/fluidsynth/issues/1005 2022-03-24 11:37:21 thanks donoban for investigating 2022-03-24 11:41:55 I think we should package soundfonts as a separate package (due to different licenses) 2022-03-24 11:42:48 I don't think GeneralUser-v1.47.sf2 from KDE/minuet is licensed under the same license either. 2022-03-24 12:08:27 Will musl-native applications be affected by gcompat if I install it? 2022-03-24 12:27:27 "Will musl-native applications be..." <- No. gcompat just provides a stub ld.so with appropriate name and a tiny binary compatibility layer, not breaking anything. 2022-03-24 13:31:39 psykose, ikke: it was mesa-dri-gallium being outdated :/ working now! 2022-03-24 13:47:43 begs: ok, thank you 2022-03-24 14:56:58 dalias: hah, that would do it :) well glad you found it 2022-03-24 21:03:11 psykose, debugging tip: to find it I straced and looked for the open call before the open to LLVM-12 to see what was pulling it in :) 2022-03-24 22:38:04 Hello, I'm trying to create a new Proxmox VM using Alpine 3.15. I run through setup-alpine and it seems to set up my disk but when I reboot the VM it doesn't boot from the disk. I have no trouble setting up Ubuntu on the same system. Any ideas? 2022-03-24 22:40:38 I have tried both the "sys" and "lvm-sys" disk options 2022-03-24 23:06:57 looks like my disk wasn't configured as a boot options in the VM settings...weird. 2022-03-25 15:05:39 o/ 2022-03-25 15:05:49 I have an RTC chipset that store the date but with -1h offset 2022-03-25 15:06:00 (I'm in Europe/Paris timezone) 2022-03-25 15:06:24 what should I do? if I force hwclock --systohc and callback hwclock -r the date is identical to the system 2022-03-25 15:06:33 but once I reboot hwclock -r gives the date with -1 hour 2022-03-25 15:06:45 so the system date is out of date until ntp kicks in 2022-03-25 15:15:09 currently trying local instead of UTC in /etc/conf.d/hwclock but not really sure what's going on 2022-03-25 18:02:48 Hey, does anyone get constant Chromium crashes? 2022-03-25 18:03:10 Tab crashes* 2022-03-25 18:08:10 on the latest version from yesterday? 2022-03-25 18:08:10 Nulo: I've seen an issue in aports repo about Chromium crashes. Also it was discussed here or on -devel channel iirc 2022-03-25 18:08:15 or before 2022-03-25 18:08:21 because with sandbox it kinda did all the time 2022-03-25 18:08:25 No, it has been happening since ~forever 2022-03-25 18:08:31 it should be patched a bit since yesterday 2022-03-25 18:08:42 aside from that, yeah, it's been an issue forever 2022-03-25 18:08:46 but i think it should be much better now 2022-03-25 18:09:14 it was a sandbox thing, https://gitlab.alpinelinux.org/alpine/aports/-/merge_requests/32356 2022-03-25 18:09:30 if it /still/ crashes (and then stops crashing with --no-sandbox), you could report it in the issue some more 2022-03-25 18:09:51 Okay, I'll try that, thanks psykose and eragon 2022-03-25 18:09:57 *Ermine :/ 2022-03-25 18:10:03 :3 2022-03-25 18:38:52 Yeah, it hasn't crashed for a bit now. Will keep testing but much better! 2022-03-25 18:39:01 I also installed chromium-angle as mentioned in one of the issues 2022-03-25 18:55:04 There is a boot service called 'networking' . Which package does it belong to? Can I use iwd backen with it? 2022-03-25 18:58:46 https://pkgs.alpinelinux.org/contents?file=networking&path=&name=&branch=edge 2022-03-25 18:59:21 Nulo: chromium-angle doesn't exist anymore actually :) 2022-03-25 18:59:29 it's now just part of chromium and there for backward compat 2022-03-25 18:59:31 you can't not have it 2022-03-25 18:59:43 since it's been the default for like 10 versions 2022-03-25 18:59:53 but we forgot to unsplit it/depend on it 2022-03-25 19:00:17 and thankfully fixed by boomanaiden 2022-03-25 19:31:48 Oh 2022-03-25 20:06:35 What is the best way to use testing packages on a stable version? I have a few packages I need from testing (particularly GUI apps). Currently, I copy the APKBUILD and patch files and use abuild to build the packages myself similar to how I did stuff for the AUR. I was wondering if adding the edge repos and pinning those packages to testing was the preferred option. 2022-03-25 20:07:12 the way you already do it is optimal (aside from moving to edge entirely) 2022-03-25 20:08:36 Do people generally use edge for the desktop use case? 2022-03-25 20:09:10 i guess, i do 2022-03-25 21:08:08 sirfredrick: Adding the edge repos will work but will break when libraries differ (been there done thatt) 2022-03-26 12:43:31 Hello, I'm looking for some advice on the best way to implement a udev style rule for Alpine (i.e. when a device is connected I want to run a command). I notice that udev isn't installed by default so is there an Alpine equivalent for this or would it just be a case of installing a udev package? 2022-03-26 12:44:03 beetlebum: by default mdev (eudev) is used 2022-03-26 12:45:02 Ah, perfect. I'll have a look at the documentation for mdev. Thanks! 2022-03-26 12:46:09 ignore the (eudev) 2022-03-26 12:46:16 eudev is the udev implementation we use 2022-03-26 12:46:25 https://github.com/slashbeast/mdev-like-a-boss 2022-03-26 13:58:39 Hey there! Small question, does the alpine-virt kernel support a Btrfs root? 2022-03-26 14:02:53 xerz: it looks like it does 2022-03-26 14:03:43 Oh does it? I wonder how I broke my new server then lol, I've modified both the boot entry and /etc/modules and the init complains about /sbin/btrfs not existing 2022-03-26 14:04:12 you may need to rebuild the initramfs 2022-03-26 14:04:28 guess I'll try chrooting hm 2022-03-26 14:05:15 is the btrfs-progs package installed? 2022-03-26 14:05:21 Ye 2022-03-26 14:24:45 aaaaaaaaaaaand that worked, it was that simple 2022-03-26 14:24:54 raven523: thanks lots! 👋 2022-03-26 14:29:59 before I google, anyone know how to autologin? typing my username is getting annoying since Im rebooting every minute or so while I develop something 2022-03-26 14:35:06 my guess would be autologin package and inittab edits. ymmv. 2022-03-26 14:40:52 thanks I will look into that 2022-03-26 19:07:14 hi, long time passed when I was here last time 2022-03-26 19:08:22 I created draft guide and script to install alpine on apple silicon M1 here https://arvanta.net/alpine/install-alpine-m1/ 2022-03-26 19:10:12 would be appreciated if some have comments and/or improvements 2022-03-26 19:11:35 interesting that disk is /dev/sda, but i assume you tested this 2022-03-26 19:12:14 Hello71: usb disk is sda in my case 2022-03-26 19:12:21 ah, i see 2022-03-26 19:12:56 internal nvme is nvme0n1 plus partitions 2022-03-26 19:13:07 mmhmm 2022-03-26 19:13:26 nvme0n1p4 is usually ESP and nvme0n1p5 is linux 2022-03-26 19:14:31 but guide is draft for now, probably will be changed to be better in few days. some people already installed alpine following it 2022-03-26 19:32:39 mps: Let me know if you get it to a point where os x can be shrunk and it can be dual booted lol 2022-03-26 19:32:49 thats what the default is 2022-03-26 19:33:23 j`ey: on an M1? 2022-03-26 19:33:33 yes 2022-03-26 19:33:49 interesting 2022-03-26 19:34:11 I have a lot of drive space left over to test with 2022-03-26 19:34:48 you can use the asahi-installer to resize and set everything up 2022-03-26 19:35:14 and got 32gb ram.. on this guy too. been using parallels already to build arm images 2022-03-26 19:35:22 would be nice to do it fully native 2022-03-26 19:36:28 Tenkawa: I put url https://asahilinux.org/2022/03/asahi-linux-alpha-release in guide where one can read about asahi installer 2022-03-26 19:36:40 thanks. 2022-03-26 19:36:44 I'll take a read 2022-03-26 19:37:42 Tenkawa: keep in mind 'my' guide is draft, you can join #asahi on OFTC and you will get more detailed info about installer 2022-03-26 19:38:33 I mostly just want to research at this point.. my long term goal would still probably be Debian 2022-03-26 19:38:49 Tenkawa: I'm running alpine more than 4 months on M1, probably first one who use M1 as daily driver ;) 2022-03-26 19:39:07 I like Alpine too but I don't know it well 2022-03-26 19:39:15 Tenkawa: there is debian ready to install also 2022-03-26 19:39:26 I know Debian in my sleep 2022-03-26 19:39:49 heh, I was debian user for about 20 years 2022-03-26 19:40:15 1993+ here 2022-03-26 19:40:15 the official image is Arch based 2022-03-26 19:40:41 debian is later than 1993 2022-03-26 19:41:02 nope 2022-03-26 19:41:02 in these years I used slackware mostly 2022-03-26 19:41:06 1993 2022-03-26 19:41:16 Slackware is a bit older 2022-03-26 19:41:23 I started on slack first 2022-03-26 19:41:37 Slack was a great distro in those days 2022-03-26 19:41:39 ohm, so I'm becoming senile :D 2022-03-26 19:42:23 they started within a few months of each other 2022-03-26 19:43:00 ACTION really got his start on old Unix.. long long ago in the 70's 2022-03-26 19:43:29 I still miss HP-UX 2022-03-26 19:43:54 mps: any luck/eta on bluetooth btw? 2022-03-26 19:44:21 ACTION really is going to be in bad shape without his bt mouse 2022-03-26 19:44:47 is it a kernel issue? 2022-03-26 19:45:09 I don't miss any of those pre-90s OSes 2022-03-26 19:45:18 what I do miss is FreeBSD 4.x 2022-03-26 19:45:28 bt is not a high priority for now Tenkawa 2022-03-26 19:45:43 Tenkawa: BT is not on horizon yet, except usb dongle 2022-03-26 19:46:05 well darn... guess that rules it out for me for now 2022-03-26 19:46:30 (dexterity related) 2022-03-26 23:09:31 hi all, i 2022-03-26 23:09:44 :/ 2022-03-26 23:10:14 hi, you 2022-03-26 23:10:23 :) 2022-03-26 23:10:26 i'm looking to run kvm and have the VMs behind a virtual switch, so that the VMs are in the same L2 network as the host 2022-03-26 23:10:30 (stupid enter key) 2022-03-26 23:10:37 i've done this on openbsd and freebsd, but with alpine i 2022-03-26 23:10:43 FFFFFFFF enter key 2022-03-26 23:11:02 i can't see any indication on how to achieve this 2022-03-26 23:11:20 some pointers and/or assistance would be greatly appreciated 2022-03-26 23:13:47 nvm, it's late and i'm nto reading right 2022-03-26 23:13:52 sorry to bother you 2022-03-26 23:22:00 opv: be sure to stick around, though, maybe someone can help you out later 2022-03-26 23:22:23 kode54: thanks, but it's literally already documented :D i wasn't looking in the right place 2022-03-26 23:22:29 oh, good 2022-03-26 23:22:33 so you got it, nice 2022-03-26 23:22:42 yeah 2022-03-26 23:22:51 is the gateway a required parameter for a static eth definition? 2022-03-26 23:23:20 yeah, gateway is required for static, unless you're not peering with anything 2022-03-26 23:23:58 the other, less common method for linking a couple of /32 ips is to define a peer as the remote IP you're connecting to over the link, but this is mostly useful for things like WireGuard tunnels 2022-03-26 23:24:18 like ip addr add /32 peer /32 2022-03-26 23:24:24 er, dev 2022-03-26 23:25:08 I dunno, I mostly follow this distribution indirectly, and hang out with the devs for other reasons than just using the distro 2022-03-26 23:25:16 it's a nice project though 2022-03-26 23:25:40 I can help with Linux issues, but mostly don't know my way around purely Alpine just yet 2022-03-26 23:25:49 kode54: very nice project, you don 2022-03-26 23:25:52 FFFFFFF enter key 2022-03-26 23:25:56 it's the apostrophe 2022-03-26 23:25:59 meh, I hate that issue 2022-03-26 23:26:06 you don't know the half of it... check out github.com/yviel.de 2022-03-26 23:26:08 I've encountered my hands doing that on new keyboards 2022-03-26 23:26:14 github/yviel-de/fsa 2022-03-26 23:26:41 i'm working on kvm on alpine so i can set up unit tests with molecule and majorly refactor the code for v0.2.0 2022-03-26 23:27:37 working with the existing config keys, which were designed with openbsd-simplicity, is a challenge 2022-03-26 23:29:29 fancy 2022-03-26 23:30:22 yes 2022-03-26 23:30:36 i'm primarily building it for myself, and figured while i'm at it why not go the extra mile 2022-03-26 23:30:49 and free the code ofc 2022-03-27 12:24:48 anyone using xfce on your alpinelinux? how do Allow shutdown restart for non-root users ?? 2022-03-27 12:59:14 oooAooo: have you tried what is written in wiki? https://wiki.alpinelinux.org/wiki/Xfce#Allowing_shut_down_and_reboot 2022-03-27 12:59:57 panekj, it work for root. what i did is: apk add consolekit2 polkit-elogind 2022-03-27 13:00:05 i did not install 'elogind' 2022-03-27 13:00:42 shutdown,rebootbutton show up on 'logout' for root user 2022-03-27 13:04:08 try to install elogind 2022-03-27 13:09:09 i want to create a L2 bridge but can't find any documentation towards it 2022-03-27 13:09:19 could you point me to the right direction 2022-03-27 13:09:26 Ermine , i have added 'elogind' 2022-03-27 13:10:43 Ermine, non-root user still doesn't get to shutdown, reboot 2022-03-27 13:11:17 try to start elogind with 'rc-service elogind start' 2022-03-27 13:11:26 and restart your xfce session 2022-03-27 13:15:04 Ermine , did a system reboot 2022-03-27 13:15:54 non-root user still unable to shutdown, reboot 2022-03-27 13:19:07 on xfce 4.16 2022-03-27 13:22:12 do i need to 'rc-service polkit start' and 'rc-update add polkit' ? 2022-03-27 13:23:10 i did that. non-root user still unable to shutdown, reboot when click on the 'logout' 2022-03-27 13:26:20 you rebooted and elogind didn't start again 2022-03-27 13:29:35 elogind did boot up fine 2022-03-27 13:30:39 delete consolekit2. just in case 2022-03-27 13:32:21 consolekit2 had been removed. did a reboot as well. 2022-03-27 13:32:30 non-root user still unable to shutdown, reboot when click on the 'logout' 2022-03-27 13:33:07 what happens when you click on 'logout'? 2022-03-27 13:33:35 only 'logout' button . the other buttons like restart, shutdown are greyed out 2022-03-27 13:36:21 Hmmm... 2022-03-27 13:37:11 Which version of Alpine are you using? 2022-03-27 13:38:02 3.13 2022-03-27 13:38:59 3.13.0 or what? 2022-03-27 13:39:16 3.13.0 2022-03-27 13:39:26 alpine-standard-3.13.6-x86_64 2022-03-27 13:41:45 ACTION spins up this release 2022-03-27 13:44:32 why are you intentionally using an old one 2022-03-27 13:44:48 regardless xfce is in community and so only supported on 3.15 2022-03-27 13:48:19 I've did everything as in wiki and shutdown works 2022-03-27 13:48:37 ermine , on 3.13 ? 2022-03-27 13:48:42 yes 2022-03-27 13:48:45 hmm.. 2022-03-27 13:48:58 it work as 'non-root' user ? 2022-03-27 13:49:03 yes 2022-03-27 13:49:38 i am puzzled now.. 2022-03-27 13:50:10 did you install xorg with 'setup-xorg-base' script? 2022-03-27 13:50:22 ya 2022-03-27 13:50:48 xfce came up because i have X running 2022-03-27 13:52:14 do you start xfce with display manager or via startx? 2022-03-27 13:53:59 ermine, i have this 'slim' install which seem to help start up the xfce for me.. 2022-03-27 13:54:18 most likely it is the fault of 'slim' graphical login causing problem ? 2022-03-27 13:55:20 ermine, are you using lightdm as the graphical login ? 2022-03-27 13:55:26 yes 2022-03-27 13:56:24 What do you mean by 'slim' install? 2022-03-27 13:57:41 ermine, bingo.. 2022-03-27 13:57:53 'slim' is an alternative graphical login display manager 2022-03-27 13:58:21 i delete 'slim' and use lightdm-gtk-greeter. all the buttons i want show up for the non-root user.. 2022-03-27 13:58:45 This is really strange. Maybe some bug 2022-03-27 13:59:16 Do you have reason to stay on 3.13? Otherwise I recommend you to update 2022-03-27 14:00:04 hello people, I'm having a weird error when trying to configure flatpak on alpine... what I'm trying to do is add key for flathub... when I do run: flatpak --user remote-modify --gpg-import=flathub.gpg flathub 2022-03-27 14:00:12 I get: error: GPG: fchown: Operation not permitted 2022-03-27 14:00:21 do you run that as root 2022-03-27 14:00:27 ah, --user 2022-03-27 14:00:30 let me see 2022-03-27 14:00:31 Ermine, i am fine with 3.15 .. 2022-03-27 14:00:45 3.13.6 isn't too far from the latest release 2022-03-27 14:01:01 is there a major change between 3.15 and 3.13.6? 2022-03-27 14:01:16 one is supported and one is not 2022-03-27 14:01:42 there is also quite a lot of version changes for many packages 2022-03-27 14:01:56 i will use 'https://dl-cdn.alpinelinux.org/alpine/v3.15/releases/x86_64/alpine-standard-3.15.2-x86_64.iso' 2022-03-27 14:01:56 libsys: could you send me that gpg key 2022-03-27 14:02:01 i hope i get this correct 2022-03-27 14:03:20 You can change version in /etc/apk/repositories, then apk update && apk upgrade --available 2022-03-27 14:04:13 you can replace apk update with apk upgrade -U 2022-03-27 14:04:17 psykose: https://flathub.org/repo/flathub.gpg 2022-03-27 14:04:56 seems to just work for me 2022-03-27 14:04:56 weird 2022-03-27 14:05:13 I literally can't find the error on internet xd 2022-03-27 14:05:22 there's not much to find 2022-03-27 14:05:23 it's a fresh postmarketos installation 2022-03-27 14:05:23 ls -lah flathub.gpg ? 2022-03-27 14:05:37 smells like excessive use of sudo 2022-03-27 14:06:01 rather none 2022-03-27 14:06:19 panekj: -rw-r--r-- 1 user user 2.8K Mar 27 04:00 flathub.gpg 2022-03-27 14:06:22 even with it being unowned by user it works 2022-03-27 14:06:29 the user is user, so it's fine 2022-03-27 14:07:02 even if not, it should be fine xd 2022-03-27 14:07:51 you could post an entire strace 2022-03-27 14:07:57 why would it chown at all though 2022-03-27 14:08:12 Ermine, i got my problem resolved. non-root user is able to shutdown, reboot now 2022-03-27 14:08:26 oooAooo: great! 2022-03-27 14:08:41 ermine, do you know how to get 'volume' bar to show up in xfce4? 2022-03-27 14:09:01 volume control from xfce4 gui 2022-03-27 14:12:07 if I do `flatpak --user remote-delete flathub` and add it again, I get the same error 2022-03-27 14:12:16 will try with strace 2022-03-27 14:12:33 oooAooo: install pulseaudio and xfce4-pulseaudio-plugin. Then add the plugin to your panel 2022-03-27 14:13:26 ermine, if i have pulseaudio, do i still alsa? 2022-03-27 14:16:46 yes (pulseaudio-alsa) 2022-03-27 14:16:54 also install pavucontrol 2022-03-27 14:16:57 psykose: here the strace https://freedombox.hacktivista.org/bepasty/vctjzVaN/+inline 2022-03-27 14:17:18 Ermine, okay. i have to leave now. 2022-03-27 14:18:55 heh, does your user have a uid of 10000 2022-03-27 14:19:12 yes 2022-03-27 14:19:54 ls -lah /home/user/.local/share/flatpak/repo 2022-03-27 14:20:29 all files are owned by user (10000) 2022-03-27 14:22:10 can you post the output 2022-03-27 14:22:39 https://freedombox.hacktivista.org/bepasty/NRcmeguZ/+inline 2022-03-27 14:23:01 weird 2022-03-27 14:24:09 it does a `openat(9, ".", O_WRONLY|O_LARGEFILE|O_CLOEXEC|O_TMPFILE, 0600) = 15` with 9 being that dir, then fchowns that 15 2022-03-27 14:24:57 that sounds like it should be correct, assuming 10000:1000 is the correct uid:gid 2022-03-27 14:24:58 no clue 2022-03-27 14:25:20 my gid is also 10000 2022-03-27 14:25:36 well it's calling 1000 for gid 2022-03-27 14:25:52 fchown(15, 10000, 1000) = -1 EPERM (Operation not permitted) 2022-03-27 14:25:59 sounds like a dirty bug 2022-03-27 14:26:02 xd 2022-03-27 14:27:00 the previous random getuid/getgid calls return 10000 2022-03-27 14:27:07 maybe the gpg codebase is doing something fancy :) 2022-03-27 14:27:53 well... I guess I could touch the file it's trying to edit? 2022-03-27 14:29:43 the file is '.' 2022-03-27 14:29:46 it opens the directory 2022-03-27 14:29:52 or something 2022-03-27 14:30:00 idk how the o_tmpfile interacts 2022-03-27 14:30:32 fuck, I bet everyone else is using flatpak as root 2022-03-27 14:30:35 xd 2022-03-27 14:31:24 life as root is so much easy and wild 2022-03-27 14:33:05 i use it as user only 2022-03-27 14:33:30 i just don't add the gpg key.. and even that works since i don't have a high user id 2022-03-27 14:33:39 idk what causes the actual issue :) maybe it is actually gpg 2022-03-27 14:34:14 i /could/ try and reproduce it myself and see if i can fix it, but its gpg.. so 2022-03-27 14:35:05 yeah, I referred to pmOS users.. which for some reason decided to use higher ids 2022-03-27 14:38:25 well... group id doesn't exist, so I could simply use id 1000 2022-03-27 14:39:14 doesn't exist? 2022-03-27 14:40:13 I mean, group ID 1000 2022-03-27 14:46:51 ok guys, thanks a lot 2022-03-27 14:47:16 I'll open a bug report later 2022-03-27 16:50:52 hello, if i want to run an alpine ARM vm on my pc, i guess i have to use the virtual image at /downloads/ 2022-03-27 16:51:19 but which parameters should i give to qemu-system-arm to make it run? 2022-03-27 16:57:43 nrb: I suppose that it depends or your needs (memory, vpcus, disk images, network, etc...) 2022-03-27 16:58:44 network etc. is something i will figure out later. i just need a plain command to start the .iso in qemu-system-arch with my -hda disk.img 2022-03-27 16:58:55 so i can install it inside, or is it preinstalled on an img? 2022-03-27 16:59:09 create your image with qemu-img 2022-03-27 16:59:13 and then try something like 2022-03-27 16:59:27 i tried -nographic which gives me no output to stdio, neither i see anything without -nographic in the qemu window 2022-03-27 16:59:59 qemu-system-x86_64 -m 512 -boot c -nographic -cdrom alpine-virt-3.15.0-x86_64.iso -hda sda.img -hdb sdb.img -machine type=q35,accel=kvm 2022-03-27 17:00:01 -netdev user,id=n1 -device virtio-net-pci,netdev=n1 -netdev user,id=n1 -device virtio-net-pci,netdev=n1 2022-03-27 17:00:10 meh 2022-03-27 17:00:26 copy/paste problems 2022-03-27 17:00:41 this is for x86_64, do i just have to change the names to ARM? 2022-03-27 17:00:48 yes yes 2022-03-27 17:01:43 i am trying this and not working: 2022-03-27 17:01:44 qemu-system-arm -machine virt -cpu cortex-a15 -m 512 -boot c -nographic -cdrom alpine-virt.iso -hda disk.img 2022-03-27 17:02:14 i expect kernel bootup/shell prompt but there is nothing 2022-03-27 17:02:27 uhM 2022-03-27 17:03:22 change 2022-03-27 17:03:34 '-boot c' to '-boot d' 2022-03-27 17:03:39 but i don't think that it's the problem 2022-03-27 17:03:59 it's not, still not working 2022-03-27 17:04:26 i know that if i would have a kernel and initrd, i could give them as parameters and then with -append i could set the console to ttyS0 or ttyAMA0, whatever 2022-03-27 17:04:49 but i thought an iso labeled virtual to run arm7 would just work out of the box 2022-03-27 17:06:20 probably the machine should be also adjusted 2022-03-27 17:08:48 try disabling --nographic and check what is on the vnc 2022-03-27 17:09:11 only the gtk window of qemu pops up and i can give it some input 2022-03-27 17:09:13 it is not the OS though 2022-03-27 17:11:39 I thought that it should be similar to x86_64, but it seems that I was wrong :| 2022-03-27 17:11:53 yeah i thought the same, i am used to x86_64 and i understand what you typed for x86_64 2022-03-27 17:12:07 anyways, im off, thanks for trying to help though :) 2022-03-27 17:32:40 hi folks, any idea why Ansible wouldn't be able to change the hostname on Alpine Linux? 2022-03-27 17:33:52 here's the error: https://paste.sr.ht/blob/01282d98adb1ab6bc8afaad3ae2c5806672ea62c 2022-03-27 17:40:31 could the hostname that you try to set have relation with TypeError: can only concatenate str (not \"list\") to str ? 2022-03-27 17:40:59 I suppose, but it works fine on other Linuxes 2022-03-27 17:42:42 does it works with something simple like 'localhost'? 2022-03-27 17:43:37 the variable is just `local_hostname: dragon` 2022-03-27 17:43:44 one sec I'll paste the code 2022-03-27 17:44:23 https://paste.sr.ht/~skyfaller/b322d4fa0071101f71b0458073f55f1b640ab790 2022-03-27 17:44:27 this is probably some python code that returns a slightly different type for whatever reason when running on alpine 2022-03-27 17:44:55 i would guess it's an alpine issue, no idea what it actually is 2022-03-27 17:45:00 that makes sense, but what bug do I file where? 2022-03-27 17:45:01 er, ansible* 2022-03-27 17:45:05 i dunno 2022-03-27 17:45:12 it's probably easier to debug with a backtrace 2022-03-27 17:45:45 how do I do that? 2022-03-27 17:45:53 no idea, i've never used ansible 2022-03-27 17:46:02 you can try add use: "alpine" next to the name: to force os detection 2022-03-27 17:46:08 in case it magically guesses something else 2022-03-27 17:46:10 but i doubt it 2022-03-27 17:49:05 skyfaller: can you test just 2022-03-27 17:49:16 hostname: "dragon" 2022-03-27 17:49:45 hm, give me a minute, I'm trying to get a more detailed error message for the previous version 2022-03-27 17:50:03 if it gets it as a literal string, maybe it's taking {{ xx }}, as a list on a list 2022-03-27 17:50:41 uHm 2022-03-27 17:50:43 look here 2022-03-27 17:50:47 https://stackoverflow.com/questions/58965459/how-to-set-hostname-using-inventory-file 2022-03-27 17:50:55 they user that {{ }} without quotes 2022-03-27 17:51:22 here's the previous version with more Vs: https://paste.sr.ht/blob/7e8bc82842104312eda63465d7cd2234d7e51c93 2022-03-27 17:51:54 didn't seem to help much 2022-03-27 17:52:14 I did force it to use alpine strategy tho 2022-03-27 17:53:07 donoban: it definitely needs quotes 2022-03-27 17:53:35 in the command 2022-03-27 17:54:11 I see 2022-03-27 17:55:07 still failing with just `name: "dragon"` and not pulling the variable from elsewhere 2022-03-27 17:56:28 oh wait something weird is happening, if I ssh into the server the hostname is now "dragon" as expected, instead of "localhost" as default 2022-03-27 17:56:45 so the command is succeeding, but returning a failture? 2022-03-27 17:57:03 *failure 2022-03-27 17:57:23 I would try to check that 'ansible/modules/hostname.py' code 2022-03-27 17:58:00 devel seems a little different, https://raw.githubusercontent.com/ansible/ansible/devel/lib/ansible/modules/hostname.py , I don't see a string concatenation on line 882 2022-03-27 17:58:38 I'm running ansible [core 2.12.2] 2022-03-27 17:58:40 there is like 2022-03-27 17:58:50 kw['diff'] = {'after': 'hostname = ' + name + '\n', 2022-03-27 17:59:09 after changing the hostname, so maybe yes, it crashes when reporting th eresult 2022-03-27 17:59:18 but try to check your exact version 2022-03-27 17:59:25 I have to go for a while 2022-03-27 17:59:49 so it would be https://github.com/ansible/ansible/blob/v2.12.2/lib/ansible/modules/hostname.py 2022-03-27 18:01:09 yeah, I guess if it's succeeding in changing the hostname, the error appears to be in reporting the result 2022-03-27 18:01:22 should I file this as a bug for Ansible? 2022-03-27 18:05:24 ok wtf, if I put `use: redhat` instead of alpine, it doesn't error? 2022-03-27 18:06:17 https://docs.ansible.com/ansible/latest/collections/ansible/builtin/hostname_module.html#parameters "Older distributions should use 'redhat'." 2022-03-27 18:06:54 so maybe whatever strategy they're trying to use for alpine is wrong 2022-03-27 18:08:55 it's failing only on the line that reports the success 2022-03-27 18:09:04 name_before is not a string but a list 2022-03-27 18:09:11 and name_before = current_hostname 2022-03-27 18:09:19 and current_hostname = hostname.get_current_hostname() 2022-03-27 18:09:22 so whatever that does 2022-03-27 18:09:26 returns.. not a string 2022-03-27 18:09:27 somehow 2022-03-27 18:09:43 it does set your new hostname then the old one is magically not a string 2022-03-27 18:09:44 magic 2022-03-27 18:13:47 can anyone please review !32410 2022-03-27 18:13:52 I just think it's weird that using the "redhat" strategy doesn't produce this error 2022-03-27 18:14:27 https://github.com/ansible/ansible/blob/6c75cf5c83da044d1fd69bc444ce4de50d728d09/lib/ansible/modules/hostname.py#L331= 2022-03-27 18:14:29 this one is alpine 2022-03-27 18:14:43 https://github.com/ansible/ansible/blob/6c75cf5c83da044d1fd69bc444ce4de50d728d09/lib/ansible/modules/hostname.py#L288= 2022-03-27 18:14:44 redhat 2022-03-27 18:15:49 sysconfig doesn't actually exist so i think it's something else 2022-03-27 18:19:19 WebKitGTK seems to work pretty badly, similarly to Chromium (tabs randomly crash, sometimes specific ones) 2022-03-27 18:19:36 Both epiphany and badwolf have the same issue 2022-03-27 18:25:19 ok, I filed a bug for Ansible: https://github.com/ansible/ansible/issues/77363 2022-03-27 18:25:44 if any of you can help with details, I'd really appreciate it, I'm a noob 2022-03-27 18:28:00 that is as detailed as it gets 2022-03-27 18:28:07 :) 2022-03-27 18:30:14 ok... I have other problem with flatpak... when trying to select a file (with foliate), I get the following error: Can't open portal file chooser: GDBus.Error:org.freedesktop.DBus.Error.UnknownMethod: No such interface “org.freedesktop.portal.FileChooser” on object at path /org/freedesktop/portal/desktop 2022-03-27 18:30:36 I've already installed xdg-desktop-portal-gtk 2022-03-27 18:32:24 a filechooser is probably something that lets you pick files 2022-03-27 18:32:27 like a file manager 2022-03-27 18:32:28 not sure 2022-03-27 18:32:43 but i think gtk has a default one, so it should exist 2022-03-27 18:32:45 no idea tbh 2022-03-27 18:33:55 it successfully opens the default gtk one for me 2022-03-27 18:34:00 i would guess some part of the stack is missing 2022-03-27 18:34:33 yup, but I have no idea what I should install 2022-03-27 18:34:37 if you're on a wlroots compositor you also want portal-wlr, aside from that it's probably some random gtk thing 2022-03-27 18:34:53 certainly something of gtk 2022-03-27 18:36:21 gtk4.0 maybe 2022-03-27 18:36:34 all the filechooser stuff is owned by that or portal 2022-03-27 18:36:55 thanks for your help everyone! 2022-03-27 18:39:10 found this file installed on my system: /usr/share/dbus-1/interfaces/org.freedesktop.portal.FileChooser.xml 2022-03-27 18:39:31 it's from portal 2022-03-27 18:39:56 seems like it's not being catched by some reason 2022-03-27 18:40:03 do you have gtk4.0 2022-03-27 18:40:08 yes 2022-03-27 18:40:15 hmm 2022-03-27 18:50:03 ): they said flatpak was for easy installations 2022-03-27 18:50:05 xd 2022-03-27 18:53:33 ok... I'm installing phosh to see if things work 2022-03-27 18:53:47 (I use SXMO/wayland) 2022-03-27 18:55:34 hopefully 2022-03-27 18:59:13 it doesn't T_T 2022-03-27 19:43:24 ok... I've tried with a fresh phosh installation and it works with no problems 2022-03-27 19:44:07 so... It's not all lost... will need to find the requiered package somehow 2022-03-27 20:34:50 Setting up a new alpine VM, and just put a script in /etc/profile.d with some aliases and evironment variables in it. However, it doesn't seem to be being applied when I logout and login again via SSH. 2022-03-27 20:34:52 Any suggestions? 2022-03-27 20:35:58 does it have the sh extension? 2022-03-27 20:37:54 No 2022-03-27 20:38:06 give it 2022-03-27 20:38:48 There's a file in there already called 'color_prompt' without an extension 2022-03-27 20:39:12 Ok, that's done it though, thanks lopid 2022-03-27 20:39:24 look at the last paragraph in /etc/profile 2022-03-27 20:39:43 Ah yes. Someone might want to fix 'color_prompt' then :) 2022-03-27 20:39:45 Thanks lopid 2022-03-27 20:40:03 hope they don't fix the README, too 2022-03-27 20:40:37 There's no README there for me. 2022-03-27 20:41:03 only /etc/profile.d/color_prompt.sh.disabled seems to be provided 2022-03-27 20:41:09 Just realised I install 3.13 too, will update to latest. 2022-03-27 20:41:19 part of alpine-baselayout 2022-03-27 20:42:06 Ok. Will update and check again. 2022-03-27 20:59:34 Ok lopid, using the latest version did the trick. Thanks. 2022-03-28 02:47:24 ok guys... I've achieved it, I needed to install libportal, which for some reason was not being installed with flatpak 2022-03-28 02:48:16 (nor xdg-desktop-portal) 2022-03-28 02:48:37 should I report a missing dependency? ... or write to the wiki? 2022-03-28 03:28:43 That does seem like an important missing dependency... Do you know if it gets roped in on another MUSL distro like Void MUSL? 2022-03-28 04:20:59 nope 2022-03-28 04:37:19 ok... this got even weirder... I'm getting the error again: Can't open portal file chooser: GDBus.Error:org.freedesktop.DBus.Error.UnknownMethod: No such interface “org.freedesktop.portal.FileChooser” on object at path /org/freedesktop/portal/desktop 2022-03-28 04:37:41 I don't think I've changed anything :/ 2022-03-28 05:17:18 ok... I need to manually execute /usr/libexec/xdg-desktop-portal-gtk and it works 2022-03-28 12:00:11 hey there, do we have user namespaces enabled on alpine kernels ? I'm having issues with flatpak steam games using Steam Linux Runtime and the cause may be that we dont have unprivileged user namespaces enabled 2022-03-28 12:00:50 We don't 2022-03-28 12:01:29 Is it something intended ? Or is it discussable/discussed ? 2022-03-28 12:04:23 https://gitlab.alpinelinux.org/alpine/aports/-/blob/master/main/linux-lts/lts.x86_64.config#L38 CONFIG_USER_NS is set 2022-03-28 12:05:22 I spotted this. So we got user namespace ? Do we have unprivileged user namespace ? / Is it something toggleable ? 2022-03-28 12:05:41 I dont see the common /proc/sys/kernel/unprivileged_userns_clone indicator 2022-03-28 12:05:46 Me neither 2022-03-28 12:05:48 it needs a patch 2022-03-28 12:06:13 https://github.com/archlinux/svntogit-packages/blob/packages/linux-lts/trunk/0001-ZEN-Add-sysctl-and-CONFIG-to-disallow-unprivileged-C.patch 2022-03-28 12:06:18 and no, we will never enable this 2022-03-28 12:07:59 Okay 2022-03-28 12:08:06 I'm curious, is there any article or literature out there which compares the security aspects of using suid binaries vs using unprivileged user namespaces? I'm not sure which to prefer. 2022-03-28 12:09:14 afaik flatpak dissalow the suid now and required the unpriviliged user way 2022-03-28 12:09:42 but then that would completly prevent flatpak steam games that need the Linux Runtime container 2022-03-28 12:09:47 the linux-hardened kernel in arch also disabled unprivileged user namespaces and this results in software like bubblewrap (from flatpak) and the chromium sandbox program to be installed as suid programs to work as intended 2022-03-28 12:10:32 i'll try this : https://github.com/flatpak/flatpak/wiki/User-namespace-requirements#setuid-bubblewrap 2022-03-28 12:13:42 mhh with suid steam just fails to start : bwrap: Creating newroot failed: Value too large for data type 2022-03-28 12:18:59 ^ nvm I just sucked at setting suid. Anyway it looks like the steam linux runtime really need the unprivileged way : "--share-pids not supported by host portal" "This feature requires Flatpak to be using a bubblewrap (bwrap) executablethat is not setuid" 2022-03-28 12:22:11 ayushnix: afaik both way got tradeoffs. The user unprivileged way seems to open more doors at this time 2022-03-28 12:22:36 cause by example Steam or chromium just support this way 2022-03-28 12:22:53 (open doors as being more supported by softwares) 2022-03-28 12:30:12 make bwrap suid 2022-03-28 12:31:15 Ermine: it doest works with flatpack chromium or steam 2022-03-28 12:32:22 ah bah moi je peux maintenant 2022-03-28 12:32:26 (oops sorry) 2022-03-28 12:34:57 in the end, this kernel patch would disable the parameter by default. It would only allow the admin to enable it if needed. Isnt it a good compromise ? 2022-03-28 12:41:27 is there a history of this chat stored anywhere? I might have lost an answer to my problem u.u 2022-03-28 12:42:41 https://irclogs.alpinelinux.org/ 2022-03-28 12:43:50 thanks :D 2022-03-28 12:46:42 ok... no answer... I'll leave my problem here to see if you can guide me 2022-03-28 12:47:38 I have a problem with flatpak and xdg-desktop-portal-gtk ... on pmOS+phosh things work fine though not on pmOS+sway, so I'm pretty sure it's something lacking on my initialization configuration. The thing is that if I do manually execute /usr/libexec/xdg-desktop-portal-gtk previous to running a GTK application through flatpak (https://johnfactotum.github.io/foliate/), everything 2022-03-28 12:47:39 works fine 2022-03-28 12:48:01 but if I don't, I get the following error: 2022-03-28 12:48:19 Can't open portal file chooser: GDBus.Error:org.freedesktop.DBus.Error.UnknownMethod: No such interface “org.freedesktop.portal.FileChooser” on object at path /org/freedesktop/portal/desktop 2022-03-28 12:48:40 clearly is that xdg-desktop-portal-gtk doesn't get automatically initialized, but IDK where or how to do it properly 2022-03-28 15:24:29 error message : failed to execute child process no such file 2022-03-28 15:25:00 this happen after i run a binary file from thunar . if i run './binary' , it say file cannot be found 2022-03-28 15:25:22 oooAooo: where did you get this binary from? 2022-03-28 15:25:34 Either it's the wrong architecture, or built against glibc, not musl 2022-03-28 15:26:11 seamonkey 2022-03-28 15:26:30 it is a 'suite' of program like browser, email etc 2022-03-28 15:26:37 most likely it is build in glibc 2022-03-28 15:26:44 which result in that error.. 2022-03-28 15:27:56 oooAooo: so you would need to make sure it's compiled against musl 2022-03-28 15:29:13 any 'shortcut' to get these glibc-program to be run? 2022-03-28 15:30:24 apk add gcompat 2022-03-28 15:35:16 ok 2022-03-28 15:35:49 it still give error .. i think best to the program as i foresee more dependency errors 2022-03-28 15:36:07 what is a good bluetoohh for xfce desktop? 2022-03-28 15:38:24 What about building seamonkey from source? 2022-03-28 15:41:10 Ermine, it is possible. got to get gtk along with seamonkey for compilation 2022-03-28 18:54:39 After running an 'apk upgrade' is there any way I can check if the system requires a reboot to complete the upgrade? Looking to automate it using Ansible so would like to be able to check for the existence of a file or similar. 2022-03-28 18:58:26 check if /lib/modules/$(uname -r) exists? 2022-03-28 18:58:27 programs which were updated need to be restarted 2022-03-28 18:59:17 restarting the kernel requires restarting all programs, but the reverse is not true 2022-03-28 18:59:35 ikke, that doesn't cover updated libs 2022-03-28 18:59:49 Habbie: He asked about reboots 2022-03-28 18:59:59 ikke, which are the best way to reload libraries! 2022-03-28 19:00:17 Then assume you need to reboot after upgrading 2022-03-28 19:00:39 yes, i like this 2022-03-28 19:13:04 ACTION is used to use lsof -n and check for open files in use 2022-03-28 19:13:17 ... open *deleted* files in use 2022-03-28 19:19:35 wouldn't ldconfig be a good way for specific programs too? 2022-03-28 19:19:48 as long as they were't staticly compiled 2022-03-28 19:21:56 err ldd not ldconfig 2022-03-28 19:22:48 that would give you the list of linked libraries minimally 2022-03-28 19:23:04 for a program 2022-03-28 19:41:40 ikke: might want to mention to someone that wikipedia's page is a fair bit behind on 3.15 releases now 2022-03-28 19:41:51 its still on 3.15.0 2022-03-28 19:45:00 Can 2022-03-28 19:45:09 Oh.. Congrats on new release! 2022-03-28 19:45:54 Tenkawa: something something primary sources 2022-03-28 19:46:26 I take it thats a grumble? 2022-03-28 19:46:52 and .. 2022-03-28 19:47:07 ACTION agrees with Ermine  2022-03-28 19:47:24 nice to see a new release 2022-03-28 19:48:56 might be time to finally rebuild one of my arm boxes to run it 2022-03-28 19:51:20 umm... are you all still working on the readme/announcement/pages/etc? 2022-03-28 19:51:38 Current Alpine Version 3.15.2 (Released Mar 23, 2022) v3.15 2021-11-24 3.15-stable | 3.15.2 | 3.15.1 | 3.15.0 2022-03-28 19:52:08 no sign of 3.15.3 on the website here 2022-03-28 19:53:16 Tenkawa: yes, will land in a bit 2022-03-28 19:53:18 I've updated now, and only package which got updated is alpine-base. But this is the second upgrade today 2022-03-28 19:53:40 ok.. just making sure I'm not losing it (more than I already am lol) 2022-03-28 19:53:53 Ermine: this minor releases are mostly for boot media / docker images 2022-03-28 19:54:03 If you regularly upgrade, you already get all the updates 2022-03-28 19:54:22 s/minor/patch 2022-03-28 19:54:22 ikke meant to say: Ermine: this patch releases are mostly for boot media / docker images 2022-03-28 19:54:40 Ok 2022-03-28 19:55:07 Tenkawa: https://www.alpinelinux.org/posts/Alpine-3.12.11-3.13.9-3.14.5-3.15.3-released.html 2022-03-28 19:55:16 thanks 2022-03-28 19:55:50 Ermine: I'm about to do some "creative" building with it so I have to make sure it's all in place or it gets very messy 2022-03-28 19:56:22 minimal: laughing yet? 2022-03-28 19:57:37 ikke: Habbie: Hello71: Was just wondering if there was a file created like Debian does. As you say, a reboot is probably no big deal, was just trying to streamline the process. 2022-03-28 19:57:46 adhawkins: no 2022-03-28 19:58:17 Ok, understood ikke. Thanks. Will stick with a reboot then. 2022-03-28 19:59:29 https://askubuntu.com/questions/1337713/what-creates-the-file-var-run-reboot-required says that reboot-required is used on a case-by-case basis 2022-03-28 19:59:41 presumably it does not encompass all cases where a reboot is or might be necessary 2022-03-28 20:01:28 for example, suppose the service file is updated to use supervise-daemon. is a restart required? yes, in some sense, but it is dumb to restart the server just to add supervise-daemon if it very rarely crashes in your environment 2022-03-28 20:06:03 Hello71: In general, I think if a package that manages a service updates something, that package will restart the service (in Debian based distros). 2022-03-28 20:06:27 that sounds terrible 2022-03-28 20:07:02 and debian only does that in some cases and only for services that don't generally have long-running connections or other work 2022-03-28 20:15:00 I could easily be wrong Hello71. I often am :) 2022-03-28 20:15:26 Ok, next question, is there a file I can add a command to that gets run at boot? I want to start up the webhook daemon, and the Alpine package doesn't provide a service file. 2022-03-28 20:15:49 adhawkins: enable the local service 2022-03-28 20:16:02 and then /etc/local.d/* 2022-03-28 20:16:35 Perfect, thanks ikke 2022-03-28 20:16:40 adhawkins: it needs the .start suffix 2022-03-28 20:17:00 Just checking the README in that directory :) 2022-03-28 20:27:23 fyi, you _can_ provide your own service file as well 2022-03-28 20:29:05 Hey folks; no PGP signatures for 3.15.3? The links are still in the download pages. 2022-03-28 20:40:32 syscomet: I've pinged ncopa, but I suspect it will be tomorrow 2022-03-28 20:40:47 ikke: thank you! 2022-03-28 20:43:11 sha256's are on the server at least 2022-03-28 21:06:55 im signing 3.15.3 now. had to have something to eat 2022-03-28 21:07:27 ikke: Yeah, I realise I could, but it seems overkill for something as simple as webhook. 2022-03-28 21:07:58 Next question, can I put symlinks in the directories under /etc/periodic or do they need to be actual files? 2022-03-28 21:08:00 the init files can be really simple as well 2022-03-28 21:08:17 Afaik symlink should work 2022-03-28 21:08:41 Great, will try a symlink ikke 2022-03-28 21:21:36 Yep, symlinks work fine ikke. 2022-03-28 21:22:34 Anyone recommend a 'lightweight' web server to just serve some static files and images? nginx? 2022-03-28 21:22:40 lighttpd? 2022-03-28 21:23:09 Darkhttpd 2022-03-28 21:24:05 Never heard of that one Ermine, let me google it. 2022-03-28 21:24:23 python 2022-03-28 21:25:51 Oooh, Google seems to be broken for me. Can't get a search result... 2022-03-28 21:26:07 lightweight, simple, web-server: caddy. 2022-03-28 21:26:31 https://downforeveryoneorjustme.com/google seems to confirm it. Ooops. 2022-03-28 21:26:39 php -S 0:80 2022-03-28 21:27:07 I'd hardly call PHP 'lightweight'! 2022-03-28 21:27:19 adhawkins, google is up 2022-03-28 21:27:24 it's the binary 2022-03-28 21:27:26 +1 for darkhttpd 2022-03-28 21:27:28 hg serve 2022-03-28 21:27:45 heat: downforeveryone seems to disagree with you... 2022-03-28 21:27:49 hell, netcat 2022-03-28 21:28:13 adhawkins, your link tells me it's up 2022-03-28 21:28:16 netcat would need some scripting or something. Ideally I just want a binary I can run and say 'serve up the files in this directory' 2022-03-28 21:28:17 :) 2022-03-28 21:28:31 heat: Not here! Odd... 2022-03-28 21:28:48 maybe it's down on one location 2022-03-28 21:28:54 Yeah, possibly heat. 2022-03-28 21:29:09 google is fine here 2022-03-28 21:29:28 Ok, they just hate me then :) 2022-03-28 21:29:41 what about other sites? 2022-03-28 21:29:42 bing works (off to wash my keyboard now). Just wanted a URL for darkhttpd... 2022-03-28 21:29:51 Everything else on the web is just fine. 2022-03-28 21:29:58 (Well, the sites I've tried) 2022-03-28 21:30:09 python -m http.server 2022-03-28 21:34:55 darkhttpd path/to/root 2022-03-28 21:35:18 "serve up files in this directory": as long as you don't care about https and it's temporary/transient, I agree with lopid: python. I use a local script which wraps that for a bunch of extra features, but at its heart, it's still that. 2022-03-28 21:36:17 darkhttpd is 40kb 2022-03-28 21:36:35 python is installed in almost every system 2022-03-28 21:36:44 not on alpine by default 2022-03-28 21:37:07 yes but it's quite possible you'll need it and install it 2022-03-28 21:37:15 why? 2022-03-28 21:37:40 because you'll need to, erm, run python scripts :P 2022-03-28 21:38:29 There's also busybox httpd, but afaik it is not built on alpine 2022-03-28 21:39:59 busybox-extras 2022-03-28 21:40:22 https://git.alpinelinux.org/aports/tree/main/busybox/busyboxconfig-extras#n855 2022-03-28 21:44:29 does anyone know if alpine is bootable on a rpi zero 2 w 2022-03-28 21:44:42 i really want to run alpine on it but last I checked, it didn't work 2022-03-28 21:45:57 Thanks for the 3.15.3 PGP sigs, they all validate now. :) 2022-03-28 21:55:02 I've gone for darkhttpd. Seems to do the job. I do have python as it's needed for Ansible. 2022-03-28 21:55:14 (I'm configuring this VM using Ansible to make it easily reproducible) 2022-03-29 13:31:19 heat: "does anyone know if alpine is bootable on a rpi zero 2 w" I think mainline support for it was just merged into Linux 5.18 a few days ago. There's no release yet, but when 5.18 is out you could try linux-edge. 2022-03-29 15:07:56 How I can add packages to initramfs 2022-03-29 15:12:01 panekj: look at /etc/mkinitfs/ 2022-03-29 15:12:12 there's a features.d folder and a mkinitfs.conf file 2022-03-29 15:15:21 ok, just noticed it's not just *.modules in /etc/mkinitfs/features.d 2022-03-29 15:15:25 thanks ikke 2022-03-29 15:15:48 Does linux-headers package work for linux-virt kernel in 3.15? 2022-03-29 15:22:13 anyone had noticed slurp being executed two times when called as shell subcommand like 'grim -g "$(slurp)"' 2022-03-29 15:22:29 I discovered that it happens with zsh but not with bash 2022-03-29 21:06:35 hey - i'm looking to build custom alpine linux iso. i have successfully been building iso's until today when I started getting: 2022-03-29 21:06:39 >>> mkimage-x86_64: --> syslinux aa8fb55a3119ed871fb636d48a150c3b0d824e7b 2022-03-29 21:06:39 ERROR: syslinux-6.04_pre1-r9: temporary error (try again later) 2022-03-29 21:07:07 am i been rate limited or something similar? 2022-03-29 21:08:14 (fwiw, i have tried basing off of edge, latest-stable, and v3.15) 2022-03-29 21:17:20 sl33nyc: temporary error is some network issue, usually dns 2022-03-29 21:20:38 thanks @ikke! i found an errant block of code that was overriding /etc/resolv.conf 2022-03-29 21:26:58 another question: from https://wiki.alpinelinux.org/wiki/How_to_make_a_custom_ISO_image_with_mkimage, it's not clear to me how to install an init.d openrc script to autostart a process that doesn't come with its own init.d script. how do i do so? 2022-03-29 21:32:26 sl33nyc: you can look at apkovl files 2022-03-29 21:33:31 Though, afaik they are always separate from the iso itself 2022-03-30 03:07:49 Hey! I was running alpine aarch64 in a minimal chroot 2022-03-30 03:08:23 I tried to use apk and... 2022-03-30 03:08:23 https://paste.debian.net/1236089/ 2022-03-30 03:08:50 HThis is the command I used to get into the chroot 2022-03-30 03:08:51 doas chroot alpine-minirootfs-3.15.3-aarch64 /bin/qemu-aarch64-static /bin/sh 2022-03-30 03:23:25 that command doesn't seem like it could possibly work unless for some reason you preinstalled qemu-user into the chroot 2022-03-30 03:25:07 Hello71: I installed qemu-user-static into the chroot 2022-03-30 03:26:10 not sure qemu actually traps execve 2022-03-30 03:26:19 how exactly did you set up the chroot 2022-03-30 03:30:05 I got qemu-user-static from the aur and put it in the chroot 2022-03-30 03:31:42 copy/paste 2022-03-30 03:31:45 in /bin 2022-03-30 04:12:07 Hello71: Sorry, did you reply? I disconnected 2022-03-30 04:37:05 Wally: that error message sounds like dns is not working or something 2022-03-30 04:37:08 or some other network issue 2022-03-30 04:37:29 ah 2022-03-30 04:37:49 Its a minichroot, could that be the issue> 2022-03-30 04:38:08 Wally: can you check if there's an /etc/resolv.conf file? 2022-03-30 04:38:24 there isnt iirc 2022-03-30 04:38:56 try to make one 2022-03-30 04:40:14 did nill in terms of errors 2022-03-30 04:40:20 want a log? 2022-03-30 04:40:21 hmm 2022-03-30 04:40:23 pretty much the same 2022-03-30 04:43:08 I guess something like strace could help 2022-03-30 04:43:29 derp 2022-03-30 04:43:40 A /bin/sh: strace: not found 2022-03-30 04:44:03 you'd have to download the package manually outside the chroot 2022-03-30 04:44:58 https://dl-cdn.alpinelinux.org/alpine/v3.15/main/aarch64/strace-5.14-r0.apk 2022-03-30 04:45:45 thanks you 2022-03-30 04:46:04 I was having a hard time finding that :emabarrased: 2022-03-30 04:48:54 Sorry how do I install the .apk? 2022-03-30 04:49:53 apk add --allow-untrusted /tmp/strace-5.14-r0.apk spits out errors 2022-03-30 04:50:53 ah 2022-03-30 04:51:28 https://paste.debian.net/1236092/ 2022-03-30 04:51:34 more dependencies? 2022-03-30 04:53:36 apparently, https://pkgs.alpinelinux.org/package/v3.13/main/aarch64/strace does not list that, strange 2022-03-30 04:53:50 libelf provides that, but that has more dependencies 2022-03-30 04:54:16 ok, helps if I check 3.15 :-) 2022-03-30 04:54:51 yeah...just tried installing libelf as I noticed that...3 other libs came in its place 2022-03-30 05:03:28 https://pastebin.com/V5Jqaqai 2022-03-30 05:03:47 sorry for pbin, dpaste banned me for spam 2022-03-30 05:04:34 hmm, I suppose that's because of qemu-static 2022-03-30 05:04:58 derp 2022-03-30 05:05:02 We also have https://tpaste.us 2022-03-30 05:05:23 oh! thx 2022-03-30 05:05:54 Always just used dpaste since someone on irc got mad at me for putting 3 line error messages in irc 2022-03-30 05:06:31 yeah, it's customary to use a paste service in that case 2022-03-30 05:07:51 but not sure about the issue then myself 2022-03-30 05:08:23 Are there any alternatives to chrooting into a aarch64 chroot on amd64? 2022-03-30 05:08:47 qemu-user 2022-03-30 05:09:04 ? 2022-03-30 05:09:41 ah, without static? 2022-03-30 05:10:20 with binfmt 2022-03-30 05:10:38 binfmt misc? 2022-03-30 05:10:46 the abi mapping thingy? 2022-03-30 05:11:04 k 2022-03-30 05:11:51 yes 2022-03-30 05:13:50 https://github.com/jirutka/qemu-openrc/blob/master/qemu-binfmt.initd 2022-03-30 05:14:09 (this is written for alpine, but you can at least check what it does) 2022-03-30 05:15:47 nope 2022-03-30 05:16:06 same error verbatim 2022-03-30 05:16:24 errors* 2022-03-30 05:19:37 saw this on a stackoverflow thread "The key thing missing was that guest kernel didn't have virtio modules. With that in place, there are a few ways to achieve this networking. Here is one:" 2022-03-30 05:19:44 etc. etc. 2022-03-30 05:20:13 but I doubt that as im running a chroot 2022-03-30 05:30:19 gonna set up a vm and create a chroot on top of that 2022-03-30 05:30:36 vm ~50mb ill be fine 2022-03-30 06:45:48 you cannot strace inside qemu-user 2022-03-30 06:46:20 yea, the error made that clear :) 2022-03-30 06:46:29 but Wally left 2022-03-30 07:08:50 i think you can use gdb, nmeum did it before. 2022-03-30 07:09:45 yes, you can instruct qemu-user to setup a gdb stub to which you can then connect using `target remote :` from gdb-multiarch 2022-03-30 07:10:28 but i think wally doesnt completely understand qemu-user :) 2022-03-30 07:18:49 s/strace/walk/ and s/qemu-user/mordor/ 2022-03-30 11:46:20 anyone run AppImage app on aplinelinux successfully ? 2022-03-30 11:48:38 it doesn't work currently: https://github.com/AppImage/AppImageKit/issues/1015 2022-03-30 11:49:53 what do i need that lsusb shows what kind of device it is... on alpine 2022-03-30 11:50:10 3.15 2022-03-30 11:50:17 libusb is installed 2022-03-30 11:50:17 panekj , ok. 2022-03-30 11:50:55 panekj, do you know any 'palemoon' web browser available via repository for aplinelinux ? 2022-03-30 11:51:39 usbutils was missing 2022-03-30 11:53:06 palemoon is not available in alpine repos, but it's source is available so possibly it can be built on alpine(?) 2022-03-30 12:07:03 panekj ok 2022-03-30 15:15:46 Hi everyone 2022-03-30 22:32:06 Hey, with the newest updates, I cant launch firefox 2022-03-30 22:32:13 I get segmentation fault 2022-03-30 22:32:20 Im on release 3.15.3 2022-03-30 22:32:59 firefox-esr? 2022-03-30 22:37:50 psykose: yes 2022-03-30 22:37:58 could you post a backtrace 2022-03-31 00:10:39 Hi, I've a fresh install of alpine 3.15 and both firefox and chromium chrashes for me. Firefox gets a segmentation fault in nssSlot_AddRef() in /usr/lib/libnss3.so 2022-03-31 00:10:47 I've no idea on how to move forward here 2022-03-31 00:22:32 hmm 2022-03-31 00:27:00 chromium stops on ssl_client_socket_impl.cc handshake failed. So it seems related and is probably some ssl error 2022-03-31 00:27:25 the former might be caused by an nss patch from today 2022-03-31 00:29:29 hold on, i'll break my whole system by downgrading to 3.15 and see if i can reproduce it :) 2022-03-31 00:30:22 :) 2022-03-31 00:31:24 indeed 2022-03-31 00:31:26 i can get the same error 2022-03-31 00:32:52 nice, thanks for testing 2022-03-31 00:33:05 so what's the solution here? A recompile of firefox and chromium? 2022-03-31 00:33:19 looking 2022-03-31 00:39:12 latest version seems to work 2022-03-31 00:39:17 i will guess the patch by itself breaks nss 2022-03-31 00:41:42 latest version? I have the latest version 2022-03-31 00:41:42 the version you have is 3.72.1 2022-03-31 00:41:43 i bumped it to 3.76.1 2022-03-31 00:41:43 upgrade in like.. 10-15 minutes 2022-03-31 00:41:43 for firefox 2022-03-31 00:41:43 then it should work 2022-03-31 00:41:43 for chromium, i have no idea 2022-03-31 00:41:46 ah, I see. thanks! 2022-03-31 00:41:50 open an issue and i will ping the chromium guy 2022-03-31 00:42:18 will do, once I get firefox up and running. Thanks for a quick turnaround here! 2022-03-31 00:47:27 psykose: looks like it actually solved the chromium problem as well 2022-03-31 00:47:31 oh neat 2022-03-31 00:47:44 ah, it does depend on nss 2022-03-31 00:47:49 yeah, it was the nss patch from today 2022-03-31 00:47:57 probably the same for another person in here too 2022-03-31 00:47:59 should be fixed now 2022-03-31 00:48:23 anjan: an upgrade should fix it for you too :) 2022-03-31 03:47:51 psykose: ah thank you. It did fix it on my pinebook pro 2022-03-31 03:48:01 psykose: do you not use alpine on your desktop? 2022-03-31 03:48:08 i do 2022-03-31 03:48:13 surprised no one else ran into it 2022-03-31 03:48:19 that person right there did 2022-03-31 03:48:28 i'm just on edge 2022-03-31 03:48:31 ah, interesting lol 2022-03-31 12:37:32 I wonder why consolefont does not work if you use console=/dev/null as parameter (would like a very clean boot) 2022-03-31 12:37:44 but after than, login in the tty and restarting consolefont applies the font 2022-03-31 12:51:22 Hi there 2022-03-31 12:51:44 What would be the best way to limit my download speeds? To set download speed on a linux? 2022-03-31 13:40:51 I'm having difficulty getting PCI devices VFIO isolated via the kernel parameters. When the system boots, I can manually unload relevant device drivers, detach the PCI devices and attach them to virts without issue, but the idea is to have this done very early on during boot time so I don't have to write unreliable scripts. dmesg: https://controlc.com/28f911d7 2022-03-31 13:45:43 I intend to update a few pages on the wiki if I manage to get a conversation going because there were quite a few headaches around grub or syslinux relying on update-extlinux. I kinda feel that my lack of understanding of Alpine's boot process is the problem here, so I'm hoping to contribute 2022-03-31 14:09:09 wow, mounting XFS on raspberry is so SLOW 2022-03-31 14:12:37 markand: really? 2022-03-31 14:12:49 thats all I use (not on alpine though) 2022-03-31 14:13:22 xfs has been the fastest filesystem in all of my testing 2022-03-31 14:14:05 (granted I don't use any microsd so I can't speak for those) 2022-03-31 14:41:36 Is there any way to add fonts to st terminal (local alpine package)? 2022-03-31 14:48:44 Tenkawa, it takes somewhat like 3-4 to mount / 2022-03-31 14:49:11 well thats odd 2022-03-31 14:49:21 its instantaneous here 2022-03-31 14:49:21 GenericNick2: build it yourself with custom patches 2022-03-31 14:54:00 Thanks. One more question, is there any way to use Alpine without mesa? 2022-03-31 14:54:31 s/alpine/xorg 2022-03-31 14:54:31 GenericNick2 meant to say: Is there any way to add fonts to st terminal (local xorg package)? 2022-03-31 14:55:33 why did meetbot use the first message instead of the latest one 2022-03-31 14:55:50 case sensitivity is such a gift 2022-03-31 14:56:00 GenericNick2: alpine != Alpine 2022-03-31 14:56:03 ah 2022-03-31 14:56:54 anyway, all the xf86 packages seem to depend on mesa-dri drivers 2022-03-31 15:04:36 Me again - my issue has been resolved thanks to minimal! My issue was that I didn't know that I was using an outdated mkinit "features" format. The listed drivers need to be ".ko.*" and not ".ko" as the document showed. I'll gather notes and update the doco later. Thanks again minimal! 2022-03-31 18:45:45 trying to netboot (ipxe) a physical machine. it has 2 nics. Only the first is connected. I set "ip=dhcp" on kernel cmdline. udhcpc configures the first, but just keeps trying the second. Is there anyway to disable the second or have it bail after n tries?