2019-02-01 03:12:35 I have noticed dovecot 2.3.3 does not have ARGON2 hashing support on Alpine (it's standard since 2.3.0 on other platforms). Is there a reason for it's non support of this? 2019-02-01 03:13:53 it uses libsodium which is available but I have been unable to compile the dovecot plugin probably because of libc musl issues 2019-02-01 03:14:40 using v3.9 btw 2019-02-01 03:20:08 Nataneal Copa is listed as package maintainer but he doesn't have a dovecot repository on github. Are there any other email or dovecot maintainers on irc who might have an idea? Many thanks for any attention. 2019-02-01 05:04:48 clandmeter: around? to conitnue my lxc journey: http://tpaste.us/Q1d4 2019-02-01 05:07:04 clandmeter: this is what my mounts look like now: http://tpaste.us/Kgog 2019-02-01 07:59:53 oculux: edge now has sodium support 2019-02-01 08:01:17 otaku42: is that on boot? 2019-02-01 08:06:56 steinex: are you sure you have enough entropy? 2019-02-01 08:08:12 clandmeter: yes, that was the messages from boottime. i had to plug the box to a serial adapter, since its headless. and by the way, the reason i could not reach it after reboot yesterday was because sshd forgot about the custom port, and 22/tcp is intentionally blocked at the firewall. 2019-02-01 08:27:07 otaku42: something is preventing cgroups to mount correctly. there must be some changes in your system config that prevents it. 2019-02-01 08:28:11 maybe grep etc directory for cgroups or similar. and make sure you system and configs are all uptodate. 2019-02-01 08:34:52 clandmeter: i did the config-update thing yesterday, and there were no changes that appeared to me to be related to cgroups. will grep for cgroups in /etc next and let you know what i find. 2019-02-01 08:35:38 otaku42: you could also try to ask in lxc channel (if they have one) 2019-02-01 08:38:20 clandmeter: yeah, they have. 2019-02-01 08:41:20 otaku42: or maybe even openrc. cgroups is managed by openrc. 2019-02-01 08:41:35 do you have any changes in rc.conf? 2019-02-01 09:01:45 clandmeter: i made no changes to rc.conf, and the only non-commented-out line i see in there is: rc_tty_number=12 2019-02-01 09:02:12 what hardware is this and which kernel? 2019-02-01 09:03:22 clandmeter: pc engines apu2, cpu is an AMD GX-412TC (x86) 2019-02-01 09:03:33 clandmeter: Linux lxc 4.19.18-0-vanilla #1-Alpine SMP Sat Jan 26 21:59:33 UTC 2019 x86_64 Linux 2019-02-01 09:04:23 heh 2019-02-01 09:06:29 clandmeter: ? :) 2019-02-01 09:06:48 i have a few of them 2019-02-01 09:06:58 but non with lxc 2019-02-01 09:07:27 use them only for networking 2019-02-01 09:07:51 i guess your containers are on ssd? 2019-02-01 09:08:08 clandmeter: i use that box as my personal web-and-everything-server at my home internet pipe. lxc comes in handy for my purposes, for separation of different server things 2019-02-01 09:08:34 i wonder if you would boot of usb with vanilla install if you could start lxc 2019-02-01 09:09:00 you could just dd the iso to usb and boot. 2019-02-01 09:09:08 clandmeter: yes, it's a 250GB sata ssd that i managed to fit into the standard pcengines case 2019-02-01 09:09:58 clandmeter: i don't think that the ssd is the cause of this issue. 2019-02-01 09:10:17 im not talking about hw here. 2019-02-01 09:10:27 vanilla alpine with just lxc installed. 2019-02-01 09:11:11 i would never boot from ssd anyways in that setup :) 2019-02-01 09:13:37 tbh, i was thinking whether it would be worthwile to move the base system onto a usb stick (or an internal msata 16gb ssd) to boot from that, and mount the 250gb ssd under /var/lib/lxc to run the containers off of that 2019-02-01 09:13:57 yep 2019-02-01 09:14:00 or sdcard 2019-02-01 09:14:09 it would boot into memory 2019-02-01 09:15:11 partition the mstata/usb stick into two, to be able to have a spare partition as a fallback in case the next upgrade fails like this one :) 2019-02-01 09:15:48 if you use tmpfs install you have your config in an overlay. 2019-02-01 09:16:03 why would you use usb when it has an sdcard slot? 2019-02-01 09:16:50 clandmeter: well, i have heard of the apus being picky on sdcards. not sure if that is true, and didn't try by now 2019-02-01 09:17:16 ok, that i dont know. i used it mostly on alix 2019-02-01 09:17:41 now i just buy an msata as its dirt cheap 2019-02-01 09:18:01 but in your case sdcard could be better due to heat 2019-02-01 09:18:30 clandmeter: usb has the advantage of being accessible from outside, without opening the case. that a plus for me, since when i have to open the case i will have to fool around again with the 2,5" ssd to get the case closed again :) 2019-02-01 09:18:49 true 2019-02-01 09:18:57 i tent to never open them again :) 2019-02-01 09:19:04 but thats not always possible... 2019-02-01 09:20:22 i would hope to not have to, too, but... upgrades seem to fail :) 2019-02-01 09:22:52 Hi, I need some hints, Is it possible to make diskless image of alpine started from netboot and resident on ramfs only on the node? I know there is a raspberry π version, but I like to do something like that on x86, and with netboot. 2019-02-01 09:23:13 i did that once 2019-02-01 09:25:32 clandmeter: hmm. grepping /etc for cgroup possibly revealed something: 2019-02-01 09:26:33 clandmeter: in update-extlinux.conf i have: default_kernel_opts="... cgroup_enable=memory ..." 2019-02-01 09:26:53 aha 2019-02-01 09:27:08 can you remove it and try again 2019-02-01 09:27:46 clandmeter: not sure if that is default, was suggested by the lxc recipe in the wiki, or has been introduced by the custom install image that i got from someone here to have a serial console during installation 2019-02-01 09:28:57 im not sure, but lots has changed in v3 so all that information is probably incorrect. 2019-02-01 09:29:20 if that fixes your issue please add it to the wiki. 2019-02-01 09:31:28 clandmeter: what is required to make the change "applied" after i've altered the update-extlinux.conf file? i guess i have to run update-extlinux? 2019-02-01 09:31:44 yes 2019-02-01 09:45:11 clandmeter: according to the kernel-parameters.txt of 4.19.9, the cgroup_enable does no longer exist and thus should have no effect. from what i grasp, it seems that in the past cgroups had to be enabled; now they are enabled by default, and one may disable controllers by using the cgroup_disable parameter. 2019-02-01 09:45:19 well, i'll try it anyway. 2019-02-01 09:52:58 clandmeter: aha. interesting: cgroups service now starts up without error messages, but still "lxc failed to start". looking into what is going on now. 2019-02-01 09:56:11 clandmeter: many more mounts unter /sys/fs/cgroup now: http://tpaste.us/jX0n 2019-02-01 09:56:33 yes ofc :) 2019-02-01 09:57:06 i dont think it will be the same error now. 2019-02-01 09:57:42 clandmeter: 4 out of 6 containers are running. the c1-alpine test container i've created yesterday does not (because of autorun=0), and the container named "mail" that i was testing with yesterday is not - the latter certainly being the reason of "lxc failed to start" 2019-02-01 09:57:46 clandmeter: getting closer now :) 2019-02-01 09:58:07 i think you can solve the rest yourself. 2019-02-01 10:03:04 clandmeter: aye. after i removed the one container from the onboot group (it was the only one in that group, and i added it during testing yesterday) it now comes up fine during boot 2019-02-01 10:03:34 clandmeter: so, lxc issue resolved now. thanks a ton for your help and sparring. i owe you a beer or two! 2019-02-01 10:05:13 and now i have to see what broke within the containers due to the apline upgrade. the one with wordpress complains about not being able to get a database connection, for example *sigh* 2019-02-01 10:43:26 how do i add opencl in alpine 2019-02-01 10:56:03 for python or C? check apk search opencl 2019-02-01 10:56:22 C 2019-02-01 10:56:26 pki 2019-02-01 10:56:37 # apk search opencl 2019-02-01 10:56:37 libreoffice-common-5.0.5.2-r0 2019-02-01 10:57:31 helpp!!! 2019-02-01 10:59:26 did you enable the community repos? 2019-02-01 10:59:40 how ? 2019-02-01 10:59:51 btw other things are added 2019-02-01 11:00:01 but opencl no repo found 2019-02-01 11:00:16 https://www.google.com/search?hl=en&q=alpine%20linux%20enable%20community%20repo 2019-02-01 11:01:49 uhm, instead of using edge/community in the url, you might need to use v3.8/community if you use Alpine v3.8 2019-02-01 11:02:04 same with other stable alpine versions 2019-02-01 11:02:31 multiarch/alpine:x86_64-v3.3 2019-02-01 11:02:41 am using this one 2019-02-01 11:02:51 on docker. 2019-02-01 11:07:13 v3.3 is unsupported for over a year now 2019-02-01 11:07:34 ok letme install 3.8 2019-02-01 11:08:04 Welcome to Alpine Linux 3.8 2019-02-01 11:08:04 Kernel \r on an \m (\l) 2019-02-01 11:08:07 done 2019-02-01 11:08:16 letme try to add opencl 2019-02-01 11:08:44 we dont ship opencl 2019-02-01 11:09:04 so i can't cmake opencl stuff 2019-02-01 11:09:42 clandmeter: but? afaik its a framework 2019-02-01 11:10:01 i dont know, i just mention its not in aports :) 2019-02-01 11:10:24 so apk add opencl doesnt work 2019-02-01 11:10:39 any alternative :( 2019-02-01 11:10:46 p3rL: no need to msg me in private. 2019-02-01 11:11:36 https://pkgs.alpinelinux.org/packages?name=*opencl*&branch=v3.8 2019-02-01 11:12:07 even better: https://pkgs.alpinelinux.org/packages?name=*opencl*&branch=v3.8&arch=x86_64 2019-02-01 11:12:55 # apk add opencl-icd-loader 2019-02-01 11:12:55 (1/1) Installing opencl-icd-loader (2.2.12-r0) 2019-02-01 11:12:55 OK: 231 MiB in 66 packages 2019-02-01 11:13:02 letme try to compile. nw 2019-02-01 11:18:05 p3rL: please dont copy and paste data into channel. use a pastbin service if you want to share. 2019-02-01 11:27:40 p3rL: pastbin => https://bpaste.net/ 2019-02-01 11:30:38 p3rL: https://wiki.alpinelinux.org/wiki/Main_Page you can find everything about alpine here. you can ask here if you face a real problem 2019-02-01 11:30:54 am facing really 2019-02-01 11:30:59 letme paste 2019-02-01 11:31:11 https://bpaste.net/show/02a0d210ada1 2019-02-01 11:31:15 clandmeter: no, i'm not. 2019-02-01 11:32:05 p3rL: have you installed opencl ? 2019-02-01 11:32:16 https://bpaste.net/show/9b52f71d0a4f 2019-02-01 11:32:27 clandmeter: should i use something like haveged or what do you recommend? 2019-02-01 11:32:31 -- Found OpenCL: /usr/lib/libOpenCL.so (found version "2.1") 2019-02-01 11:32:53 apk add opencl-icd-loader-dev opencl-headers opencl-icd-loader 2019-02-01 11:32:58 i used those 2019-02-01 11:33:04 steinex: what hardware you have 2019-02-01 11:34:04 if it is low end box without much sources for entropy then yes, it helps a lot 2019-02-01 11:34:05 hm 2019-02-01 11:34:16 what changed that is causing low entropy? 2019-02-01 11:34:26 there must be something 2019-02-01 11:34:40 disabled RNRND in kernel 2019-02-01 11:34:52 between 3.8 and 3.9? 2019-02-01 11:35:00 Ai9zO5AP did you saw ? 2019-02-01 11:35:08 kernel 4.19.x to be precise 2019-02-01 11:35:13 steinex: yes you could try haveged 2019-02-01 11:35:19 mps: ahhh 2019-02-01 11:35:22 see if it makes a difference. 2019-02-01 11:35:25 well thats silly 2019-02-01 11:35:31 but maybe this warrants a post or something 2019-02-01 11:35:40 so it doesn't need to be explained repeatedly 2019-02-01 11:35:52 jwh: I don't agree that it is silly 2019-02-01 11:36:02 p3rL: Could you tell me what do you want to compile ? 2019-02-01 11:36:07 it's essentially a regression at this point though, mps 2019-02-01 11:36:26 since previously working just fine machines now don't work properly 2019-02-01 11:36:34 this morning I thought to write about that, but didn't have time 2019-02-01 11:36:46 jwh is right, we should add it to the relnotes 2019-02-01 11:36:58 it's up to the end user to decide on policy, if they can trust the rng or not is up to their local policy, the default should "just work" 2019-02-01 11:37:01 heh 2019-02-01 11:37:05 but i was waiting for ncopa, but it seems he is enjoying his earned holiday 2019-02-01 11:38:02 Ai9zO5AP can i msg you in private ? 2019-02-01 11:38:21 i am sorry, no 2019-02-01 11:38:41 Ai9zO5AP i am trying to compile https://github.com/bogdanadnan/ariominer "ariominer" 2019-02-01 11:38:48 https://bpaste.net/show/9b52f71d0a4f 2019-02-01 11:38:51 it should be explained in short note how to enable it in boot menu, and how to add it in extlinux.conf 2019-02-01 11:39:09 Ai9zO5AP am getting error in 95% /usr/lib/gcc/x86_64-alpine-linux-musl/8.2.0/../../../../x86_64-alpine-linux-musl/bin/ld: cannot find -lOpenCL 2019-02-01 11:39:35 p3rL: you should install opencl 2019-02-01 11:39:52 Ai9zO5AP already installed those > apk add opencl-icd-loader-dev opencl-headers opencl-icd-loader 2019-02-01 11:39:59 Ai9zO5AP -- Found OpenCL: /usr/lib/libOpenCL.so (found version "2.1") 2019-02-01 11:41:40 p3rL: what is the package that you want to compile ? 2019-02-01 11:41:54 https://github.com/bogdanadnan/ariominer 2019-02-01 11:42:11 installed haveged, /proc/sys/kernel/random/entropy_avail now shows 2681, still the same issue. I'm not sure if the value is high enough though. My workstation has something around ~4000 and is fine 2019-02-01 11:42:39 steinex: ok, it was just a guess as it comes up a lot recently. 2019-02-01 11:42:48 and it could be related to ssl 2019-02-01 11:43:01 Another issue is that openntpd seems to completely stall if the first constraint check fails, so this will be an issue nonetheless if not enough entropy is available when it first starts i think. 2019-02-01 11:43:28 Ai9zO5AP am trying to compile ArioMiner. 2019-02-01 11:43:50 clandmeter: openssl s_client to www.google.com works fine. it's strange. 2019-02-01 11:44:15 steinex: your workstation also uses smtpd? 2019-02-01 11:44:38 you mean openntpd? yes. thats a Void Linux though. 2019-02-01 11:44:46 ok 2019-02-01 11:45:00 this issue happend after you upgraded? 2019-02-01 11:45:17 this is a fresh install of 3.9. I never used Alpine before. :-) 2019-02-01 11:45:38 ah, thats good advertisement :) 2019-02-01 11:46:02 i have no experience with osmptd 2019-02-01 11:46:17 SpaceToast uses it though. 2019-02-01 11:46:45 theres an issue re openntpd on github for this 2019-02-01 11:46:58 or maybe bugs.a.o 2019-02-01 11:47:43 jwh: yes, i linked to it yesterday. you mean https://bugs.alpinelinux.org/issues/9635 i guess 2019-02-01 11:47:52 ye 2019-02-01 11:47:53 It seems like the same issue indeed 2019-02-01 11:48:15 jwh: on github? 2019-02-01 11:48:26 ah ok 2019-02-01 11:48:59 if its not entropy its probably the switch to openssl 2019-02-01 11:49:27 p3rL: you should ask the developer of ArioMiner 2019-02-01 11:50:02 Ai9zO5AP its work fine in simple installation on ubuntu. 2019-02-01 11:50:12 am trying to build static using alpine 2019-02-01 11:50:38 steinex: i think its a openssl compat issue 2019-02-01 11:51:34 ok. i can run openntpd without the constraint check for now. it's not an urgent issue :) 2019-02-01 11:52:14 yes itr looks more openssl 2019-02-01 11:52:45 https://github.com/OpenSMTPD/OpenSMTPD/issues/738 2019-02-01 11:52:52 look at the last msg 2019-02-01 11:53:38 clandmeter: i'm talking about OpenNTPD btw. But yea, OpenSMTPD is from the same camp :-) 2019-02-01 11:54:22 lol 2019-02-01 11:54:29 ok my bad 2019-02-01 11:55:28 "therefore we no longer want to cope with the extra-work for us of accommodating OpenSSL." 2019-02-01 11:55:34 luckily better MTAs exist :) 2019-02-01 12:04:52 I have another question about apk. After installing the system and doing "apk list -IO" basically all installed packages are orphans. Isn't there a "base-system" package or such that pulls in these packages so not every single package is orphaned? 2019-02-01 12:05:36 this would help a lot later keeping the system clean and being able to have a glance at what packages i really manually installed 2019-02-01 12:06:54 im surprised alpine has orphaned packages 2019-02-01 12:07:20 there is alpine-base 2019-02-01 12:08:55 found that too. it seems that every package after installation is marked as orphan though. 2019-02-01 12:10:14 im wondering how a package can even be orphaned? either its depended by world or some other pkg, or it is removed 2019-02-01 12:11:51 it could just mean "no other package depends on it" and it would be a good indicator for packages that were manually installed. 2019-02-01 12:11:52 ohh 2019-02-01 12:11:56 It least thats what i hope :D 2019-02-01 12:12:04 no, than they would be dependet by world 2019-02-01 12:12:10 i think i found it 2019-02-01 12:12:20 they become orphaned when they dont have a upstream source anymore 2019-02-01 12:12:47 check with `apk policy` 2019-02-01 12:13:08 I'm still confused by the term "world" 2019-02-01 12:14:04 /etc/apk/world 2019-02-01 12:14:15 essentially dependencies the user has on packages 2019-02-01 12:14:49 ah-ha! this is exactly what i was looking for... 2019-02-01 12:15:28 thanks :-) 2019-02-01 12:15:49 some users just edit that file and then run apk fix 2019-02-01 12:19:45 neat. 2019-02-01 12:35:15 steinex: indeed, many packages seem incorrectly marked as orphaned 2019-02-01 13:24:32 this /etc/apk/world approach is awesome 2019-02-01 13:24:37 i'm in love with apk fix already 2019-02-01 13:27:27 do you know how the installation preseed mechanism for alpine works? 2019-02-01 13:27:53 the apkovl is given to the installer, its essentially an tarball with stuff like /etc/apk/world inside 2019-02-01 13:30:44 i hadn't have a closer look yet. just using alpine since yesterday. but i will definitely dig deeper as i go :-) 2019-02-01 13:35:43 oh boy. the upgrade from 3.8 to 3.9 seems to have silently changed the mariadb default configuration, resulting in mariadb to no longer listen on network 2019-02-01 13:36:29 did cost me another ~ two hours or so of troubleshooting 2019-02-01 13:36:41 i'm getting to old for this sh*t 2019-02-01 13:37:46 too old, even 2019-02-01 14:24:20 I am buiilding pyarrow inside an alpine docker container. It builds fine and the pyarrow tests pass but pyarrow isn't accessible globally. I tried to set the pythonpath without any success. Is there another way I should be adding it to the path to make it accessible? 2019-02-01 14:28:12 what happens if you invoke the binary by absolute path, shade34321 ? 2019-02-01 14:29:39 appears to work, If I open a python shell inside the dir where the binary is located I can import it and what I want but other shells being spawned within the python process, i.e. running pytest in another dir, don't work 2019-02-01 14:33:51 oh, a python module, not a binary 2019-02-01 14:34:29 yeah, sorry if I was unclear 2019-02-01 14:34:49 When I run a sequence of commands in a Docker container, it correctly installs a program then uninstalls *-dev packages. When I run the same sequence creating a Docker image (docker build), I get an error like "The command '/bin/sh -c apk del --no-cache hdf5-dev@testing' returned a non-zero code: 1". Same for stable packages. What may be wrong? How do I get more information out of Docker? 2019-02-01 14:35:10 It would actually be nice if I could just build it then pip install it 2019-02-01 14:54:42 tkzv: is the dockerfile aailable anywhere? can sometimes get more info out by instantiating the intermediate containres created during the build and poking around inside them 2019-02-01 15:00:55 mort___: What do you mean? I have the file (I'm editing it). I replaced "apk del --verbose --no-cache hdf5-dev@testing" with "apk del --verbose --no-cache -X testing hdf5-dev", it gave a warning "WARNING: Ignoring testing/x86_64/APKINDEX.tar.gz: No such file or directory", but worked without errors. 2019-02-01 15:01:45 what was teh error code ("echo $?") after you ran that command? 2019-02-01 15:02:29 Docker stopped reporting 1. 2019-02-01 15:04:29 What's the difference between "-X testing hdf5-dev" and "hdf5-dev@testing" and why is it only manifest for docker build? 2019-02-01 15:08:12 I thing -X testing tries to pull the package and all its dependencies from testing, $PACKAGE@testing tries to pull only that package from testing and tries to satisfy the rest from other repos iirc 2019-02-01 15:08:20 or maybe the other way around, not sure now 2019-02-01 15:10:02 What's the difference for apk del? 2019-02-01 15:13:35 no idea 2019-02-01 15:13:49 maybe just to keep the interface the same 2019-02-01 15:51:32 Hi folks! I was hoping to find anyone here who could enlighten me on the state of DMVPN in alpine linux? 2019-02-01 15:51:58 Most up-to-date info I can find links to Alpine - nhrpd + strongswan with patch 2019-02-01 15:52:30 same as the state on any other linux distro I'd imagine since it's largely kernel relevant and I think only one maybe two software packages support it 2019-02-01 15:52:55 these days you're limited to frr really as not much else is actively maintained 2019-02-01 15:53:16 but, I expect that may be an absolute nightmare to build on musl 2019-02-01 15:53:24 Well, the kernel specifics seems to be properly mainlined, the devil is in the details. Quagga has an extension which requires some patches on strongswan, which noone but Alpine has 2019-02-01 15:54:43 theres still many kernel bugs, still 2019-02-01 15:54:51 ugh, s/, still// 2019-02-01 15:55:45 stopped tracking it long ago as it was a total mess 2019-02-01 15:55:58 Ok, so the short answer is "don't go there" 2019-02-01 15:56:00 that's good enough for me 2019-02-01 15:56:22 tbf, it's probably usable but I'd limit it to just nhrpd, sans ipsec 2019-02-01 15:56:53 Ok, thank you very much 2019-02-01 15:57:08 however, look at the frr notes: https://github.com/FRRouting/frr/blob/master/nhrpd/README.nhrpd 2019-02-01 15:58:00 So, if i understand correctly there once was a standalone nhrpd, which has been integrated in both ffr and quagga 2019-02-01 15:58:05 basically forked from there 2019-02-01 15:58:23 looks like it, that readme suggests the alpine patches also 2019-02-01 15:59:20 details on the wiki seem complete enough though, maybe it's stable enough these days... 2019-02-01 15:59:42 it all smells a bit stale, and I wonder how many people really use this 2019-02-01 15:59:47 hard to tell 2019-02-01 16:00:23 do you actually need dmvpn? 2019-02-01 16:00:31 *I* don't 2019-02-01 16:00:47 but we have potential customers who are absolute cisco lovers 2019-02-01 16:00:53 ah gotcha 2019-02-01 16:01:13 so I spent the day investigating the state of affairs and understanding what it actually is 2019-02-01 16:01:28 (answer: a big mess of stacked protocols which is likely a nightmare to debug) 2019-02-01 16:02:28 heh 2019-02-01 16:02:54 it's fine if you stick to cisco (having used it a long time ago), but I wouldn't even use it on cisco these days either 2019-02-01 16:03:11 theres much better (albeit maybe more complex to setup) solutions, or better... non-cisco ones 2019-02-01 16:03:37 I guess. It just seems very convoluted to me, but there are probably Good Reasons for that 2019-02-01 16:03:47 I just tend to install Tinc and stop caring 2019-02-01 16:03:58 it's a reasonable idea, and it was the only sensible choice at the time 2019-02-01 16:04:29 these days theres a plethora of solutions and software/hardware that can do similar but with less horribleness 2019-02-01 16:04:41 True. But legacy 2019-02-01 16:05:12 problem obviously if your clients are cisco centric, you may have a hard time finding a cross section of what you want to/can support and what their devices support 2019-02-01 16:05:21 always like that. 2019-02-01 16:05:33 Well, thanks again, I'll report back and go do Fun things instead 2019-02-01 16:05:36 have a nice day! 2019-02-01 16:05:52 lol, you too 2019-02-01 16:22:01 how would i best re-read /etc/inittab? 2019-02-01 16:26:20 init -q seems to do the job. 2019-02-01 17:29:07 am I being a potato, what do I need to enable to read /etc/fstab at boot time? 2019-02-01 18:54:02 uhm yeah, that /etc/fstab question is interesting. i wonder this myself? 2019-02-01 18:55:00 hm maybe I'm missing something 2019-02-01 18:55:26 i have changed mount option in /etc/fstab but they are not applied aswell 2019-02-01 19:02:48 hm, the remount with updated mount options doesn't seem to work for / 2019-02-01 19:15:54 to be more precise: i changed rw,relatime to rw,noatime for / in /etc/fstab but that change isn't reflected on reboot. i tried to mkinitfs but that doesn't change it. any ideas? 2019-02-01 19:16:31 yeah 2019-02-01 19:16:48 I think you need localboot enabled 2019-02-01 19:16:51 ugh, localmount 2019-02-01 19:17:59 WARNING: localmount has already been started. If i restart is just for the sake of testing it, it's still relatime 2019-02-01 19:18:11 hm 2019-02-01 19:18:25 UUID=a43faf59-1bd9-4611-8ffc-92ea404c6032 / ext4 rw,noatime,data=ordered,discard 0 1 2019-02-01 19:18:30 those get applied on my test one 2019-02-01 19:18:36 /dev/sdd3 on / type ext4 (rw,noatime,discard) 2019-02-01 19:20:17 not for me. weird. 2019-02-01 19:21:58 how did you enable localmount? i think it's default on my installation. 2019-02-01 21:41:18 ah, I generated a mangled image, so probably wasn't enabled by default on mine 2019-02-01 21:43:52 for me, / still does not get (re)mounted with my options from /etc/fstab. localmount is enabled. ¯\_(ツ)_/¯ 2019-02-01 21:44:06 localmount is for non-/ 2019-02-01 21:44:24 but i lieu of it recognising fstab options, you can put them on kernel cmdline 2019-02-01 21:45:07 i now use the local service and have a one-liner in /etc/local.d which remounts / like i want. that works for me for now. 2019-02-01 22:34:41 not really ideal but yeah 2019-02-02 03:48:39 err 2019-02-02 03:48:41 http://dl-cdn.alpinelinux.org/alpine/v3.9/releases/armhf/alpine-rpi-3.9.0-armhf.tar.gz.asc 2019-02-02 03:48:45 why does that not exist? 2019-02-02 03:49:18 the pgp signatures exist for RC1 2019-02-02 03:50:31 http://dl-cdn.alpinelinux.org/alpine/v3.9/releases/x86_64/ even for x86_64 2019-02-02 03:56:49 i am surprised nothing is mentioned on the mailing lists about that 2019-02-02 03:58:50 i guess this is a question for alpine-devel 2019-02-02 03:59:39 because i'm not upgrading until it's rectified 2019-02-02 03:59:51 particularly as those isos also come over http:// 2019-02-02 04:10:11 yeah, people should verify signatures 2019-02-02 04:10:14 ncopa: any clue? ^ 2019-02-02 04:10:18 this is not good 2019-02-02 04:10:31 i have posted to alpine-devel 2019-02-02 04:10:52 so continuing discussion about it should be public (not so much irc) 2019-02-02 04:12:07 i just saw the email, and agreed 2019-02-02 04:12:18 it just seems... weird... 2019-02-02 04:12:30 particularly as there's nothing mentioned in any release notes about it 2019-02-02 04:12:39 i mean i want to think it's just a mistake, but these days..... 2019-02-02 04:12:41 some bug might have occurred somewhere in the CI/CD pipeline 2019-02-02 04:12:48 i suspect it's a mistake 2019-02-02 04:12:51 yeah i'm guessing that's the case 2019-02-02 04:13:03 but i am also surprised nobody noticed or said anything for 3 days 2019-02-02 04:13:22 or any of the other releases from RC2 to RC5 2019-02-02 04:13:33 that's the weirdest part 2019-02-02 04:15:01 something isn't quite right for sure 2019-02-02 04:15:06 that's the thing often things like this do look like innocent mistakes 2019-02-02 04:15:59 https://freedom-to-tinker.com/2013/10/09/the-linux-backdoor-attempt-of-2003/ :P 2019-02-02 04:36:36 danieli: should i open a bug? 2019-02-02 04:37:01 there doesn't seem to be any 2019-02-02 04:38:02 also affected versions on the bug tracker, 3.8.3 is the latest 2019-02-02 04:38:38 danieli: what would you set the priority to? 2019-02-02 04:38:47 i'm guessing high/urgent 2019-02-02 04:42:02 oh wait no 3.9 is there. 2019-02-02 04:42:54 weirdly it goes 3.8.0, 3.8.1, 3.8.2, 3.9.0, 3.10.0, 3.5.4, 3.6.7, 3.8.3 :P 2019-02-02 04:44:00 https://bugs.alpinelinux.org/issues/9952 2019-02-02 10:05:29 hiya 2019-02-02 10:10:52 hi 2019-02-02 10:34:16 darn, no obconf in apk 2019-02-02 18:08:38 yay! 3.9! 2019-02-02 18:12:57 <_ikke_> \o/ 2019-02-02 18:31:26 still some packages on my machine that doesn't upgrade since 3.8, except for when I `apk fix -u $pkgname`... does this indicate a bigger problem, or is it safe? 2019-02-02 18:46:12 <_ikke_> trfl: did you use --availe when running apk upgrade? 2019-02-02 18:46:19 <_ikke_> --available 2019-02-02 18:47:10 yes, I always add -a just in case 2019-02-02 18:47:31 looking at the options, maybe -l would be a good idea? 2019-02-02 20:21:12 hi, i think i have founded a recently bug with "apk" 2019-02-02 20:24:45 here : https://privatebin.net/?36095a3d8f7dd380#hPXjmNMzGMhuetVpElHtKcDYNh7XUWeYUrPRJ+8tdyc= 2019-02-02 20:25:52 virtual package is not increment with recent version of apk 2019-02-02 20:26:31 apk add --virtual .test pkg1 and later... apk add --virtual .test pkg2 << is not added and not installed in my virtual package 2019-02-02 20:29:07 If this is a bug, and that's what i think, i can open an issue _ikke_ ? 2019-02-02 20:30:02 This broke some images because my Dockerfile use virtual package 2019-02-02 20:30:08 no problem before 2019-02-02 20:30:22 <_ikke_> Tetsumaki: If that same sequence used to work, that it probably is a bug. Opening an issue would probably be waranted 2019-02-02 20:33:44 _ikke_: OK, i will try to determine from which version exactly (2.10.0 or 1 or 2 or other) this does not work and i open a issue 2019-02-02 20:34:07 I just did a git bisect 2019-02-02 20:34:11 b06e3b991dbe2cb5f9ca65ed50a443a5ecaaf383 introduced it 2019-02-02 20:34:26 <_ikke_> pardis: +1 2019-02-02 20:35:13 https://git.alpinelinux.org/apk-tools/commit/?id=b06e3b991dbe2cb5f9ca65ed50a443a5ecaaf383 2019-02-02 20:37:22 ho thanks pardis 2019-02-02 20:38:31 i open a issue 2019-02-02 20:38:53 with my fantastique english 2019-02-02 20:40:59 https://bugs.alpinelinux.org/issues/9651 2019-02-02 20:41:16 3 month bug, change priority ? 2019-02-02 20:42:16 <_ikke_> I would not change the priority. You could add a comment, including what pardis found 2019-02-02 20:43:23 OK _ikke_ 2019-02-02 20:46:30 comment added 2019-02-02 20:55:48 Pending correction, I use apk-tools-2.10.1-r0.apk 2019-02-03 00:14:47 clandmeter: thank you for the info on sodium on edge. My mailserver needs a bit more stability than edge I think. Maybe I'll spin up a container and try it out. 2019-02-03 02:35:25 > FreeBDSM ╡ [13:13:41] yay! 3.9! 2019-02-03 02:35:37 giggity 2019-02-03 02:35:39 yeah i'm excited to deploy this new server but, i am a bit concerned 2019-02-03 02:35:49 https://bugs.alpinelinux.org/issues/9952 2019-02-03 02:36:49 also i'm waiting on the disclosure https://lists.alpinelinux.org/alpine-user/0619.html 2019-02-03 02:40:22 https://www.kb.cert.org/vuls/ doesn't look like they've published yet. 2019-02-03 02:41:46 i'm thinking it might not be long until a 3.9.1 but who knows. 2019-02-03 02:43:24 what i am kind of surprised and shocked at is that between RC2 and the final release none of the community has mentioned the lack of signatures on the mailing lists or on the bug tracker 2019-02-03 02:43:47 shame on you (users) if you download alpine linux and don't even verify it. 2019-02-03 14:17:32 are network installations possible? I am looking to replicate the functionality provided by OpenBSDs ramdisk installer -- boot a small kernel, feed answers to the installer, installer installs on local hard drive, reboot to fresh new system 2019-02-03 14:17:49 in reading the wiki, I'm not entirely sure if this is possible 2019-02-04 07:31:09 mm ncopa must be on a holiday 2019-02-04 07:31:23 i guess a deserved one after releasing 3.9.0 2019-02-04 07:32:05 unless of course aliens kidnapped him to run their next-generation space ship 2019-02-04 07:32:22 (because they think alpine linux is a good choice) 2019-02-04 07:50:12 tya99: :D 2019-02-04 07:50:49 you made me smiling this morning 2019-02-04 08:06:45 mps: good choice until evil aliens decide to muck with their computers! 2019-02-04 08:06:52 https://bugs.alpinelinux.org/issues/9952 2019-02-04 08:07:27 this is assuming they don't have huge quantum inter-dimensional computers capable of breaking RSA ;) 2019-02-04 08:07:42 they might be shitty evil aliens! 2019-02-04 08:10:22 tya99: we are aware but but the person who can sign is current not available. 2019-02-04 08:10:41 tya99: meh, bug 9952 will be solved. till then you can use previous release :/ 2019-02-04 08:10:43 yeah i figured 2019-02-04 08:10:52 clandmeter: how does alpine do signatures 2019-02-04 08:10:55 mps: am doing 2019-02-04 08:11:17 maldridge: what do you mean? 2019-02-04 08:11:33 do you just have individuals with keys, or do you make use of a central keying system 2019-02-04 08:11:51 currently only ncopa signs the releases. 2019-02-04 08:11:52 tya99: I also use previous on one box where I should upgrade 2019-02-04 08:12:08 ah 2019-02-04 08:13:01 ncopa is enjoying his holiday, if he can find time im sure he will fix it. 2019-02-04 08:13:16 if you look at history of #alpine-devel just after release you can see why it is not signed 2019-02-04 08:13:19 I was primarily curious if you guys had anything better than how we do it with void 2019-02-04 08:13:55 not yet, it has not been discussed yet. 2019-02-04 08:13:58 maldridge: you are void dev? 2019-02-04 08:14:03 mps: 2019-02-04 08:14:05 mps: indeed 2019-02-04 08:14:15 (hit enter too soon) 2019-02-04 08:14:25 the idea is to redo the build infra and setup releng team. 2019-02-04 08:14:29 clandmeter: I'd be very interested in anything that comes of that discussion 2019-02-04 08:14:42 ah, then you remember last year problem with keys ;) 2019-02-04 08:14:53 I'm currently considering a combination of consul/vault and redoctober to do this sort of thing 2019-02-04 08:15:10 or, two years, I forgot exact time 2019-02-04 08:21:34 clandmeter: it's nice to know that ncopa is healthy and not in the brig of an alien spaceship! 2019-02-04 08:48:31 moin. anyone else had problems with postfix not starting up anymore after updating to alpine v3.9? 2019-02-04 08:50:04 according to /var/log/messages, /usr/lib/postfix/postfix-script can not be found. it has been moved from /usr/lib/postfix (3.8) to /usr/libexec/postfix (3.9), and something apparently does not yet took notice of that change. 2019-02-04 08:51:34 otaku42: postfix starts fine here on vanilla install 2019-02-04 08:52:01 clandmeter: fresh install, or updated from v3.8? 2019-02-04 08:52:11 fresh 2019-02-04 08:53:12 clandmeter: so it's most likely an issue that the upgrade caused. hmmkay. 2019-02-04 08:53:38 can you pastein the error? 2019-02-04 08:54:43 otaku42: i think you should compare the config files new and current. 2019-02-04 08:55:08 apk should have added a main.cf.apk-new (if you ever modified it) 2019-02-04 08:55:48 clandmeter: http://tpaste.us/Q1dz 2019-02-04 08:56:00 Guys, can you recommend any budget armv8 vps provider except Scaleway? 2019-02-04 08:56:52 otaku42: do you have daemon_directory = setting? 2019-02-04 08:57:19 clandmeter: thanks, that's it. i thought i had gone through the update-conf thing, but obviously didn't 2019-02-04 08:57:39 clandmeter: the daemon_directory directive has changed. 2019-02-04 08:59:19 clandmeter: yep, that was it. after adjusting daemon_directory postfix starts fine now. /me feels dumb now :) 2019-02-04 09:12:00 i noticed that firefox was removed for arm, is there anywhere official i can find where mozilla say they don't support arm? 2019-02-04 09:12:08 i had a bit of a look around. 2019-02-04 09:12:23 seems like a bit of a shitty limitation to moving to rust if it only works on x86_64 2019-02-04 09:12:35 or is that something that is a WIP 2019-02-04 09:12:55 running firefox on arm seems like something people might want to do 2019-02-04 09:13:48 https://www.theverge.com/2018/12/6/18129456/qualcomm-snapdragon-pc-firefox-web-browser-64-bit-native-mozilla interesting 2019-02-04 09:15:31 https://bugzilla.mozilla.org/show_bug.cgi?id=1450742 wow maybe that's true 2019-02-04 09:18:20 although it seems that the esr compiles on everything https://packages.debian.org/sid/firefox-esr except powerpc and hppa 2019-02-04 09:19:50 yeah see it seems to support armhf there in debian. 2019-02-04 09:21:09 and it certainly seemed to in 3.8 https://pkgs.alpinelinux.org/package/v3.8/community/armhf/firefox-esr 2019-02-04 09:21:51 so i'd be curious if i can get more information about "Firefox is only available on x86_64 due to Rust." 2019-02-04 09:23:24 https://lists.alpinelinux.org/alpine-user/0317.html 2019-02-04 09:23:25 also 2019-02-04 09:23:27 in regards to that 2019-02-04 09:23:31 > PS: no seamonkey for x86? it is my ever-favourite :) Pale Moon could 2019-02-04 09:23:33 > also be interesting 2019-02-04 09:23:36 NO. 2019-02-04 09:23:58 should you need to know why https://github.com/privacytoolsIO/privacytools.io/issues/375 2019-02-04 09:26:15 does hypermail let you get the messageid? 2019-02-04 09:26:31 if you want to reply to something and you're not on the list you could do that with piper 2019-02-04 09:35:53 it does seem the message id is at the top of the page, commented out 2019-02-04 09:35:59 if you view source 2019-02-04 09:36:08 i wonder if i can input a specific message id in thunderbird 2019-02-04 09:38:12 yesterday I compıled pale moon on alpine 2019-02-04 09:39:13 was running quite smooth even without stack-overflow patch which generally applied firefox derived browsers on musl 2019-02-04 09:40:31 firefox has much better (low) cpu usage on video playback 2019-02-04 09:43:31 for anyone wanting to read backlogs in browser, https://github.com/insteps/scripts/tree/master/alpine 2019-02-04 09:44:46 are network installations possible? I am looking to replicate the functionality provided by OpenBSDs ramdisk installer -- boot a small kernel, feed answers to the installer, installer installs on local hard drive, reboot to fresh new system 2019-02-04 09:44:58 in reading the wiki, I'm not entirely sure if this is possible 2019-02-04 09:45:21 I understand apkovl is *probably* what I want -- but is there some way to script that instead of doing a full install and creating some kind of overlay to package up? 2019-02-04 09:46:14 I'm trying to get as close as one file to answer installation questions, and another set of scripts or what-have-you to install packages and mangle/place config files not in base 2019-02-04 09:46:47 vkrishn: line 28: can't open '../../../bash/inc/color.inc': No such file or directory 2019-02-04 09:46:58 fedora/redhat has that pyvpx 2019-02-04 09:47:07 have you looked at anaconda/kickstart? 2019-02-04 09:47:29 i wonder how hard it would be to adapt it to install alpine instead of fedora 2019-02-04 09:49:43 well, I'm not going to run anything but alpine 2019-02-04 09:50:05 terra just clone the repo 2019-02-04 09:50:16 not very big :) 2019-02-04 09:50:20 as much as I'd love to keep using openbsd for this, I can't. and I have 50 instances on 12 different VM providers (all KVM save for two Xen) 2019-02-04 09:50:56 pyvpx: if you need that for a anumber of clients, you can use pxe boot over nfs 2019-02-04 09:51:03 tya99: https://lwn.net/Articles/777963/ look near at the end of comments 2019-02-04 09:51:23 pyvpx: but I'm not sure alpine initramfs capable of mount nfs root 2019-02-04 09:51:44 terra: these are on VPS providers. there's nothing to NFS and all but one (vultr) dont support PXE boot 2019-02-04 09:52:13 I currently boot an openbsd ramdisk preseeded with installer answers, that then pulls site specific configurations from an HTTP server 2019-02-04 09:52:23 trying to replicate similar functionality anyway I can 2019-02-04 09:52:28 what're trying to do? 2019-02-04 09:52:37 what're you* 2019-02-04 09:52:55 what about cloud init? 2019-02-04 09:53:12 automatic/unattended installation & configuration of alpine linux virtual machines on a variety of "VPS" providers 2019-02-04 09:53:40 my understanding of cloud init is you need a metadata server that is adjacent. if whatever I boot can setup an ipsec tunnel so link-local addressing works...sure? :p 2019-02-04 09:53:42 hm, seems like a job for cloud-init or post-install ansible or similar 2019-02-04 09:53:45 but l2tp seems a big involved. 2019-02-04 09:53:59 well you can seed enough of a config from say vultr cloud-init to get it into a state you can use 2019-02-04 09:54:11 but theres nothing really generic, of course 2019-02-04 09:54:16 yes, that's vultr though. I have 11 other providers who give no api and have no smarts 2019-02-04 09:54:33 sensible providers *should* have cloud-init 2019-02-04 09:54:41 but a lot just seem to use off the shelf crap 2019-02-04 09:54:53 unfortunately the majority of the providers these VMs are on are not sensible 2019-02-04 09:55:05 (personally I don't particularly like cloud-init but I can see why its useful) 2019-02-04 09:55:11 indeed 2019-02-04 09:55:37 ansible is probably the better option for this tbf, or something similar 2019-02-04 09:55:47 at least if you can get it installed (snapshots, iso, or something) 2019-02-04 09:56:14 you'll likely have to knock something up yourself either way as the starting point might be an iso, or another distro et 2019-02-04 09:56:18 etc* 2019-02-04 09:56:35 well, okay then -- step one, how do I configure networking unattended (so ansible would even be functional)? 2019-02-04 09:56:47 I'd hope dhcp works 2019-02-04 09:56:54 at least for the most part 2019-02-04 09:56:56 many of these are statically configured 2019-02-04 09:57:08 do they have dhcp though? 2019-02-04 09:57:10 aka no dhcp 2019-02-04 09:57:11 nope 2019-02-04 09:57:13 as in, does somerthing answer 2019-02-04 09:57:16 no 2019-02-04 09:57:18 unfortunately not 2019-02-04 09:57:33 yeah, without an API of some kind you've got no chance :P 2019-02-04 09:57:36 but I have all the relevant network/disk/etc details in a database I can pull from 2019-02-04 09:57:52 unless they can deploy templates or snapshots and reconfigure the network when they're deployed 2019-02-04 09:58:05 odd. I have a well working pipeline with openbsd and it's all shell scripts. I'm a bit surprised I can't replicate it with alpine 2019-02-04 09:58:12 oh 2019-02-04 09:58:17 well if you know it in advance, yeah 2019-02-04 09:58:22 no they just offer the usual redhat/ubuntu/debian basic installs 2019-02-04 09:58:54 yes I have everything I need in advance. 2019-02-04 09:58:57 sorry I wasn't clear 2019-02-04 09:59:06 I am looking for an unattended installation method 2019-02-04 09:59:19 where I can "preseed" (I know, debian term) the installer questions 2019-02-04 09:59:32 tbh the only one I get on with is ansible and it has some limited notion of data stores/inventory 2019-02-04 09:59:39 but pick whatever works for you 2019-02-04 10:02:33 https://wiki.alpinelinux.org/wiki/Bootstrapping_Alpine_Linux 2019-02-04 10:02:41 unfortunate. thats kind of exactly what I was looking for, heh 2019-02-04 10:04:07 if you need to install alpine over an existing distro thats gonna be more tricky 2019-02-04 10:09:30 is there something that grub can boot, that I can pre-configure with network details? 2019-02-04 10:09:49 hm 2019-02-04 10:09:59 you'd have to build an image for each thing 2019-02-04 10:10:13 dhcp would make this *much* easier, couldn't you just limit your selection of terrible providers? 2019-02-04 10:10:19 configure bootable file w/ network -> configure grub to boot it -> reboot -> boots w/ network access -> run alpine install on to hard drive 2019-02-04 10:10:31 no, these are all specifically picked for location and the fact they provide a full BGP table 2019-02-04 10:10:39 o 2019-02-04 10:10:41 so, I have to work around the annoyances 2019-02-04 10:10:48 wanna share your list? :P 2019-02-04 10:10:57 https://bgp.services/ 2019-02-04 10:11:06 oh 2019-02-04 10:11:08 the main ones are vultr and m247 2019-02-04 10:11:11 I went through them, bleh 2019-02-04 10:11:22 thought you might have some secrets 2019-02-04 10:11:24 it's regularly updated ;) 2019-02-04 10:11:27 nah 2019-02-04 10:11:31 melbicom is great 2019-02-04 10:11:35 misaka is hit or miss 2019-02-04 10:11:37 tbh vultr caters for most locations 2019-02-04 10:11:53 indeed. my usecase here is as many full table feeds in as many palces as possible 2019-02-04 10:12:01 and is convenient because dhcp and same details everywhere 2019-02-04 10:12:03 heh 2019-02-04 10:12:22 yes vultr is a dream compared to every other provider I deal with 2019-02-04 10:12:39 designed properly I guess 2019-02-04 10:13:03 ok so do you typically only have one VM per provider/location? 2019-02-04 10:14:32 yes 2019-02-04 10:14:48 yeah so it's not even worth making something for each location heh 2019-02-04 10:14:57 but once they're configured, thats it right... 2019-02-04 10:15:25 so in openbsd, I create a bootable ramdisk (bsd.rd) that includes an answerfile (auto_install.conf) and point grub to boot the bsd.rd 2019-02-04 10:15:48 that loads up, the answer file configures networking, downloads the install files, installs them, then grabs a site specific file 2019-02-04 10:16:00 hm 2019-02-04 10:16:12 I dunno enough about apkvol, guessing you've tried that? 2019-02-04 10:16:20 the answerfile is site specific, of course, and I generate that from a basic shell script (since I have all the disk/network specifics) 2019-02-04 10:16:34 yes, perhaps I'm not fully appreciating/understanding apkvol 2019-02-04 10:16:47 but I understood it to be "do a full install, configure what you need in an overlay, tar up the overlay" 2019-02-04 10:16:58 so I'd have to do the last two steps for every site? 2019-02-04 10:17:01 yeah thats my undertanding too 2019-02-04 10:17:34 what about just throwing in initramfs (built into the kernel) - almost the same as bsd.r 2019-02-04 10:17:37 bsd.rd 2019-02-04 10:18:23 you'd only need to add the ip/routes to the initramfs, if you plot up a webserver somewhere you could just grab configs or generate them on the fly 2019-02-04 10:18:59 or... actually 2019-02-04 10:19:06 just pass the ip info from grub 2019-02-04 10:28:06 jwh: that sounds like what I'm looking for :D 2019-02-04 10:28:27 where do I read more on this? sorry, I've used OpenBSD (or some other BSD) exclusively for the past 15 years I know next to nothing about linux except installing OpenWRT a few times :p 2019-02-04 10:28:58 for convenience I was going to suggest buildroot to make the initramfs (what openwrt spawned from) 2019-02-04 10:44:15 mps: ah so it's to do with not using glibc then that makes more sense 2019-02-04 10:45:08 i figured it was a WIP and it seemed like too much of a failure for upstream 2019-02-04 10:53:45 mps: also i solved my other question https://superuser.com/questions/1177870/how-to-manually-set-the-in-reply-to-header-in-thunderbird 2019-02-04 11:00:00 tya99: no, it is to bootstrap rust lang on Alpine arch's, for now at aarch64 and armv7 2019-02-04 11:00:25 oh yeah :P that as well thanks for the link 2019-02-04 11:00:27 there are no issues with firefox 2019-02-04 11:02:01 I managed to build firefox-esr 60.4.0-r1 for aarch64 and armv7, and it works quite fine, and is faster than the previous version 2019-02-04 11:02:39 but, the problem is to get to 'native rust' on these arch's 2019-02-04 11:04:05 and, in the 'play' arrives llvm with it's quirks and loopholes 2019-02-04 11:05:36 upstream of llvm and rust mostly 'ignore' musl, i.e. musl is low priority in their work, that my impression 2019-02-04 12:59:31 just use a browser that doesn't require rust? :P 2019-02-04 13:05:27 hiro: i'd love it if there was a umatrix for qutebrowser but yeah. 2019-02-04 13:05:37 it's a toss up of firefox or chromium. 2019-02-04 13:05:57 and at this point firefox is better for privacy than joining the google borg 2019-02-04 13:06:25 it's not a major issue as mps said, and i bet it will be ironed out :) 2019-02-04 13:08:45 tya99: you can try if PM build for you (it does for x86 and x86_64) https://github.com/tanertas/aports/tree/palemoon/testing/palemoon 2019-02-04 13:09:10 terra: hahaha no you'll never get me to try palemoon 2019-02-04 13:09:59 terra: perhaps you should read from there on https://github.com/privacytoolsIO/privacytools.io/issues/375#issuecomment-454704963 2019-02-04 13:10:24 i think I made a pretty good point while palemoon shouldn't be an option for anyone at this point 2019-02-04 13:10:30 point(s) 2019-02-04 13:10:47 [12:23] > PS: no seamonkey for x86? it is my ever-favourite :) Pale Moon could [12:23] > also be interesting 2019-02-04 13:10:56 and one of the nooby palemoon devs tried to argue with me about it 2019-02-04 13:11:16 apparently every point i made was "fake news" 2019-02-04 13:11:32 tya99: agreed, i also like to recommend umatrix, though it's all *way* too much work normally :) 2019-02-04 13:11:32 you want some popcorn if you want to read that drama 2019-02-04 13:11:41 wasn't it you that talking about PM ? 2019-02-04 13:11:56 no, someone else was, and i just dropped a link to that ticket. 2019-02-04 13:12:03 *issue on another project i work on 2019-02-04 13:12:06 tya99: and firefox is part of the google borg 2019-02-04 13:12:18 no, that's chromium 2019-02-04 13:12:26 but i guess you'd be okay with ungoogled chrome or brave 2019-02-04 13:12:29 tya99: no, mozilla is mostly funded by google 2019-02-04 13:12:43 well a good portion of it yes, but i guess that's the world we live in 2019-02-04 13:12:48 tya99: and it's not like mozilla is inputting or having any power over the development of the web standards 2019-02-04 13:13:05 like i said, though palemoon isn't the solution 2019-02-04 13:13:24 tya99: and since most security problems come from shitty basic web standard design issues, even that doesn't matter. 2019-02-04 13:13:34 (in that issue) in fact it cannot even be considered a reasonable contender to anything 2019-02-04 13:13:48 if you want privacy right now your best bet is not to use a browser 2019-02-04 13:13:57 :) 2019-02-04 13:14:00 we could all browse the internet stallman style 2019-02-04 13:14:19 we should at least try to get as close to his way as possible 2019-02-04 13:14:33 but i wonder whether richard still does that these days 2019-02-04 13:14:40 or at least try to make sure people that *do* use the internet like him aren't sabotaged by some shitty standards 2019-02-04 13:14:49 https://stallman.org/stallman-computing.html 2019-02-04 13:14:56 tya99: i'm aware :) 2019-02-04 13:15:11 i would probably use icecat if i wasn't using firefox tbh 2019-02-04 13:15:12 i would love to be able to pull it off like stallman or tannenbaum 2019-02-04 13:15:45 i think the only way it is possible to do what they suggest is if you visit a specific set of sites that, that kind of usage works 2019-02-04 13:16:06 for a large part of the internet, you won't have so much luck. 2019-02-04 13:18:38 tannenbaum just has somebody print out stuff he might want to read 2019-02-04 13:19:07 sounds a bit like donald knuth 2019-02-04 13:19:15 actually 2019-02-04 13:19:21 i might be conflating knuth and tannenbaum 2019-02-04 13:19:25 :) 2019-02-04 13:19:36 it was one of them at least 2019-02-04 13:20:00 but besides the point. the problem is most of us cannot afford a secretary that will unfuck the web for us 2019-02-04 13:20:11 true :( 2019-02-04 13:20:27 but i like to imagine that, cause it's easy to do. 2019-02-04 13:20:30 terra: the most amusing part is PM's website has google adverts lol 2019-02-04 13:20:51 and then i can compare my actual experience with what i imagine to be knuth's truly awesome life 2019-02-04 13:20:59 and see what's most lacking 2019-02-04 13:21:25 puts things into the right perspective 2019-02-04 13:21:32 and that dev (mattatobin) said he would respond "until the heat death of the universe", but when i mentioned the bit about google adverts he stopped responding :( 2019-02-04 13:22:19 and not because he had better things to do, because if he did have them, he wouldn't have made me repeat myself in 10 figillion ways 2019-02-04 13:22:25 also since you brought up the borg: you know they would try to make mozilla one of their's like they did with the captain 2019-02-04 13:22:33 s/dev/cheersquad member/ 2019-02-04 13:22:35 i don't know if that was clear earlier ;) 2019-02-04 13:24:05 they don't only get money from google though. 2019-02-04 13:24:10 https://en.wikipedia.org/wiki/Mozilla_Foundation#Donations 2019-02-04 13:24:16 https://techcrunch.com/2017/11/14/mozilla-terminates-its-deal-with-yahoo-and-makes-google-the-default-in-firefox-again/ 2019-02-04 13:24:55 https://en.wikipedia.org/wiki/Mozilla_Foundation#Financing 2019-02-04 13:26:17 https://old.reddit.com/r/firefox/comments/a98gmi/donations_to_mozilla_foundation_are_not_used_for/ 2019-02-04 13:26:43 but yeah i'd probably donate to the tor project, because mozilla will match the donation! 2019-02-04 13:28:43 i don't trust any of this shit! :) 2019-02-04 13:28:51 tor was always too near to US government 2019-02-04 13:29:35 and just when US government got even less even remotely trustable they claim to now diversify their donors so that they aren't mainly government funded 2019-02-04 13:29:59 but who guarantees us they aren't still involved 2019-02-04 13:30:21 does alpine have a correspondig package to base-devel ? 2019-02-04 13:30:31 i mean even if you ignore the money, they will be heavily invested in the technology 2019-02-04 13:30:40 Piraty: alpine-sdk ? 2019-02-04 13:30:40 a meta package to install common build tools? 2019-02-04 13:30:44 so they are up to date, i cannot track TOR development, but i'm sure they do... 2019-02-04 13:30:56 mps: thanks 2019-02-04 13:31:19 tya99: to me mozilla+tor are too much connected 2019-02-04 13:31:27 in a lxc container, based on alpine-3.8, gcc is super old, why is that? 2019-02-04 13:31:42 Piraty: look at wiki on development section 2019-02-04 13:31:44 well i meant in terms of support of a project which would help contribute upstream 2019-02-04 13:31:44 gcc (Alpine 6.4.0) 6.4.0 2019-02-04 13:31:52 mps: will do, thanks 2019-02-04 13:32:10 now gcc is 8.2 2019-02-04 13:32:19 hiro: tor project is independently run, sure the US government seeded it with money but so what, that doesn't make it bad 2019-02-04 13:32:35 the spec is open, the source is open, and it gets a lot of research. 2019-02-04 13:32:54 it's quite likely the US government uses it, yes, (that's almost certain) but yeah. 2019-02-04 13:33:57 tya99: i remember that were were onion urls with the US .gov ssh motd 2019-02-04 13:33:58 tya99: i'm not saying it's bad, i'm just saying i don't trust it because for long times us controlled too much of the used infrastructure, and they control it to some level through money and they definitely have a lot of people watching it 2019-02-04 13:34:16 > long times us controlled 2019-02-04 13:34:16 no software system is flawless 2019-02-04 13:34:20 no it's not and never has been 2019-02-04 13:34:24 huh? 2019-02-04 13:34:39 i thought that was widely accepted knowledge 2019-02-04 13:34:51 https://en.wikipedia.org/wiki/Tor_%28anonymity_network%29#History 2019-02-04 13:35:09 that is not the same thing as "controlling it" 2019-02-04 13:35:29 by that logic you could say the whole internet is "controlled by the US because DARPA" 2019-02-04 13:36:40 you know because TCP/IP https://en.wikipedia.org/wiki/History_of_the_Internet#Merging_the_networks_and_creating_the_Internet_(1973%E2%80%9395) 2019-02-04 13:37:06 i'd say DARPA just shows their intention 2019-02-04 13:37:44 and if the snowden leaks are anything to go by they haven't got a magic backdoor ;) 2019-02-04 13:37:53 being decentralized to a big degree and even if overlayed by TOR, i still see the US trying to follow very closesly all the details of the technology 2019-02-04 13:38:20 which means that since software is flawless, they will be one of the big entities that will be able to exploit small flaws earlier than the average guy, or even programmer 2019-02-04 13:38:25 if they fund it and use it they would want to keep an eye on things. 2019-02-04 13:38:31 of course 2019-02-04 13:38:48 but i think you'll find state programs just look for 0days in everything 2019-02-04 13:38:55 not specific projects. 2019-02-04 13:39:02 they would be looking at all the commits for all the things 2019-02-04 13:39:18 and it would not just be them, but other projects as well, and all the governments (that have money) 2019-02-04 13:39:35 you know they call it "amassing that cyber arsenal etc" 2019-02-04 13:40:13 i have no proof for anything sensational that people would get really anxious about if put into mainstream news. but i feel like there are enough pointers that show the US might be able to master the technology in a way that it's used against TOR's official goals. 2019-02-04 13:41:13 so if you have nothing to hide (TM) from big and powerful US intelligence, i think it's totally fine to use TOR 2019-02-04 13:41:21 i think you'll find though they have the same ways as everything else 2019-02-04 13:41:45 and many people (that don't live in the US) truly don't think their main threat vector is the US government 2019-02-04 13:43:07 there is (understandably) slightly more attention given to TOR bec. of it's history of being (partly successfully) used by people trying to hide from the US government 2019-02-04 13:43:11 and i think you'll find FOXACID being one of those ways https://www.schneier.com/blog/archives/2013/10/how_the_nsa_att.html 2019-02-04 13:43:30 and since a lot of stuff goes out through limited number of tor exit nodes, it's very easy to attack this quite centrally 2019-02-04 13:43:32 hiro: one of the major ways that the NSA did things was downgrade attacks 2019-02-04 13:43:44 because people did things shitty and allowed shittier things for compatibility 2019-02-04 13:43:48 sure, and all this shows is that TOR has their attention 2019-02-04 13:44:01 we can be fortunate with things like TLS 1.3 that people are getting much more strict when writing new standards 2019-02-04 13:44:33 but being distributed to only a *very* limited amount (if you don't use tor-internal onion urls), i'd say it's practically too centralized around tor exit nodes that might be getting extra screening scrutiny 2019-02-04 13:44:48 like if i was US government i would record everything that ever leaves a tor exit node 2019-02-04 13:44:54 but that's a known fact 2019-02-04 13:45:01 just cause it's very likely that people who have something to hide will traverse through them 2019-02-04 13:45:04 fortunately most of the internet is over https these days 2019-02-04 13:45:16 and the standards have improved significantly, OCSP, HSTS etc. 2019-02-04 13:45:20 tya99: yeah, but we're talking about adding layers of imperfect security on top of each other 2019-02-04 13:45:33 web security is a complete joke by design 2019-02-04 13:45:34 well nothing is perfect 2019-02-04 13:45:38 exactly 2019-02-04 13:45:53 but that doesn't mean we all give up and go home 2019-02-04 13:46:13 now if nothing is perfect i don't think it helps to give people product recommendations like a snake-oil seller, at least not without the correct disclaimers 2019-02-04 13:46:27 and as i said my disclaimer is: don't use it to hide from US government 2019-02-04 13:47:07 possibly try not to use it to hide from *any* big government 2019-02-04 13:47:18 i think that is making the US government sound like a bunch of magicians which they are not 2019-02-04 13:47:53 overall though if something is to be learned by the snowden leaks, Tor works pretty damn well 2019-02-04 13:48:17 and sure there might be certain things like FOXACID and QUANTUMINSERT that work in *some* cases, it's certainly not something they can just apply anywhere 2019-02-04 13:48:33 come on, don't bring in snowden 2019-02-04 13:48:38 anyway this is getting offtopic. 2019-02-04 13:48:40 we all knew what they were doing for decades 2019-02-04 13:48:51 well exactly, and i think people are more proactive now 2019-02-04 13:48:59 i'm just trying to tell you to add a disclaimer next time you talk about TOR 2019-02-04 13:49:00 hi. is there nice way to check which (possibly not installed) package provides a certain binary? 2019-02-04 13:49:02 in terms of thinking "how might this be broken" 2019-02-04 13:49:21 because people are stupid and every problem for them has a product as a solution 2019-02-04 13:49:27 hiro: and i think you're being paranoid ;) 2019-02-04 13:49:28 but sometimes it's not like that. 2019-02-04 13:49:33 tya99: not at all 2019-02-04 13:49:56 tya99: i tell those people to go low-tech and don't waste their time sweating about bullshit digital security questions 2019-02-04 13:49:56 > because people are stupid and every problem for them has a product as a solution 2019-02-04 13:50:05 well a threat model is obviously something you need to consider 2019-02-04 13:50:11 i just tell them they shouldn't pretend to themselves that any of their shit is secure 2019-02-04 13:50:19 and that they shouldn't hope technology will solve that for them 2019-02-04 13:50:24 Piraty: you can look for file at https://pkgs.alpinelinux.org/contents 2019-02-04 13:50:24 but it doesn't mean that you should just use nothing and send carrier pigeons either. 2019-02-04 13:50:31 tya99 and hiro: consider moving to #alpine-offtopic 2019-02-04 13:50:36 bec. this way they'll just keep actually private shit to themselves 2019-02-04 13:50:42 and not on their computer 2019-02-04 13:50:46 oh sorry. i'll stop now :) 2019-02-04 13:50:53 no worries 2019-02-04 13:50:54 i'd participate in the conversation too, but i'm a little preoccupied 2019-02-04 13:52:03 mps: thanks. no cli tool? apk magix? 2019-02-04 13:53:04 huh, I think there is or were http to json interface but couldn't remember now. You should ask on #alpine-devel channel 2019-02-04 13:53:57 iirc there is an apk command for it (without any API stuff), but i don't remember it off the top of my head 2019-02-04 13:54:39 yeah, apk-file package 2019-02-04 13:55:30 it is in testing 2019-02-04 13:57:13 Piraty: you can use: apk search cmd:cmdname 2019-02-04 13:57:29 but it will only work for binaries in path 2019-02-04 13:58:29 apk-file works fine, mostly 2019-02-04 13:58:31 where would I follow progress of porting cargo/rustc to alpine x86? 2019-02-04 13:59:58 heh i didnt know apk-file :) 2019-02-04 14:00:17 but like i mention, if you just need things from path the cmd:... will work fine. 2019-02-04 14:01:09 OP wanted something to search content of packages which are not installed 2019-02-04 14:01:43 right, yeah 2019-02-04 14:01:48 pkgs.alpinelinux.org/contents 2019-02-04 14:01:57 he asked for binary 2019-02-04 14:02:00 not just file 2019-02-04 14:02:13 a binary file is a file 2019-02-04 14:02:18 everything is a file ;) 2019-02-04 14:02:38 binaries are mostly placed in path, so cmd:... would work. 2019-02-04 14:03:03 executables i mean 2019-02-04 14:03:05 but missing support for armv7 arch in apk-file 2019-02-04 14:03:35 and searching for cmd:... will be much faster :) 2019-02-04 14:04:07 does it works for not installed packages? 2019-02-04 14:04:15 yes 2019-02-04 14:04:26 didn't know, thanks 2019-02-04 14:04:33 cmd are in the index 2019-02-04 14:05:00 similar as you could do: apk add cmd:dmesg 2019-02-04 14:05:26 busybox applets are not added i think 2019-02-04 14:05:37 so it would replace the applet with the app providing it. 2019-02-04 14:06:42 you have to be careful because multiple apps can provide the same thing. the one with the highest version will win. 2019-02-04 14:08:13 'apk search cmd:ash' show noting, so busybox applet are 'skipped' 2019-02-04 14:09:38 they are generated post install 2019-02-04 14:10:59 understand, just tried to see how cmd: works 2019-02-04 14:12:16 even shorted names works: 'apk search cmd:busy'. nice 2019-02-04 15:45:54 Hi, does apk keep logs of upgrades/installed packages? 2019-02-04 15:46:30 Or is there a way to check if and when an upgrade of an installation took place? 2019-02-04 15:46:43 sebix: i dont think so 2019-02-04 16:38:55 pyvpx: hey just out of curiosity, did you find any in central europe that weren't germany (cz, at etc) 2019-02-04 16:38:58 ? 2019-02-04 16:46:59 doesn't seem to be much between germany an asia 2019-02-04 16:47:01 and* 2019-02-04 17:15:35 jwh: yeah I have instances in prague, vienna, sofia 2019-02-04 17:35:57 oh, who do you get those off? 2019-02-04 17:43:45 melbicom and m247 2019-02-04 17:43:50 vps247.com 2019-02-04 17:47:06 so can I build a initramfs that includes network configuration, then that pulls the alpine installer with an answerfile? is that a feasiable route to investigate further? 2019-02-04 17:47:41 you could set network config from the cmdline to make it easier (kernel can do that on its own) 2019-02-04 17:48:10 then you wouldn't need multiple initramds 2019-02-04 17:48:13 initramfs 2019-02-04 17:49:29 hey uh, mind sharing gateway or something for vps247 in prg? 2019-02-04 17:50:18 wanna see how far away it is from home 2019-02-04 17:51:56 sent 2019-02-04 17:52:07 ugh, sec I have +g 2019-02-04 17:52:12 can you resend pls 2019-02-04 17:52:13 ah, sorry 2019-02-04 17:52:18 resent :) 2019-02-04 17:52:46 telia, but at least it stays in the city 2019-02-04 17:52:47 heh 2019-02-04 19:33:58 I'm trying to get Xorg to work on my desktop computer with an AMD GPU. Currently, when I issue "startx", the shell doesn't return? 2019-02-04 19:34:22 Currently, it just outputs "xauth: timeout in locking authority file ..." 2019-02-04 19:36:13 azarus: which Grafical env did you install ? 2019-02-04 19:36:19 Right, got it to flicker again. My Xorg log: https://ptpb.pw/GKfj 2019-02-04 19:36:23 Ai9zO5AP: none, just dwm 2019-02-04 19:36:41 (via .xinitrc) 2019-02-04 19:38:29 azarus: could you paste bin your .xinitrc ? 2019-02-04 19:38:39 Ai9zO5AP: it's two lines: 2019-02-04 19:38:42 exec dwm & 2019-02-04 19:38:42 st 2019-02-04 19:38:48 2019-02-04 19:39:58 azarus: did you run setup-xorg-base ? 2019-02-04 19:40:02 yes 2019-02-04 19:40:11 (and many other things to enable kernel modesetting) 2019-02-04 19:41:06 azarus: could pastebin the xorg log 2019-02-04 19:41:13 Got it working now! 2019-02-04 19:41:13 :) 2019-02-04 19:41:18 please do not copy paste here 2019-02-04 19:41:29 i have, already ;) 2019-02-04 19:41:37 azarus: what was the problem ? 2019-02-04 19:41:41 azarus: how? 2019-02-04 19:41:54 AinNero: kernel parameters contained "nomodeset" 2019-02-04 19:41:58 which is a strange default 2019-02-04 19:43:56 azarus: did you mean earlyprintk ? 2019-02-04 19:44:22 Ai9zO5AP: nope, never heard of that parameter before 2019-02-04 19:44:36 azarus: are you using VM ? 2019-02-04 19:44:47 Ai9zO5AP: no, else how would I have an AMD GPU? 2019-02-04 19:45:12 sorry I forgot, but strange 2019-02-04 19:46:15 If i am right nomodeset must be add for example with qemu/kvm or even xen as a boot option when booting up 2019-02-04 21:14:48 Quick question: Has anybody ran Tiled ( https://mapeditor.org ) on Alpine Linux? The AppImage they provide can't be run, even with gcompat :/ 2019-02-04 21:17:49 wasn't appimage the one that does open /proc/self/exe to retrieve a squashfs? 2019-02-04 21:17:59 because that breaks with gcompat 2019-02-04 21:19:00 AinNero: https://ptpb.pw/qJkb 2019-02-04 21:19:04 if that helps 2019-02-04 21:19:55 yea wont run on alpine 2019-02-04 21:20:24 hm, the other option would be to create an aport, right 2019-02-04 21:22:31 i guess 2019-02-04 21:22:58 It doesn't look like a pig to port; it does depend on qt5 however 2019-02-04 23:06:09 Hi, is there a simple method to include an arbitrary file (in my case a firmware.rom binary) in an ISO generated by mkimage? 2019-02-05 09:07:51 hi, trying a live upgrade from 3.8 to 3.9 -> alpine:~$ sudo apk upgrade 2019-02-05 09:07:52 1 error; 2098 MiB in 487 packages 2019-02-05 09:08:13 how can I find where the '1 error' comes from? 2019-02-05 09:09:55 <_ikke_> apk fix 2019-02-05 09:10:22 tru_tru: try 'apk fix -v' 2019-02-05 09:10:54 _ikke_: eh, you are fast 2019-02-05 09:11:20 tru_tru: what is the error you got when you upgraded? 2019-02-05 09:11:43 should be in the list of upgraded packages. 2019-02-05 09:20:54 alpine:~$ sudo apk fix -v 2019-02-05 09:20:55 (1/1) Reinstalling xorgproto (2018.4-r0) 2019-02-05 09:20:55 OK: 487 packages, 3656 dirs, 37018 files, 2098 MiB 2019-02-05 09:20:57 thx :D 2019-02-05 09:21:39 the initial error was lost in the scroll back lines... when I ran 'sudo apk update && sudo apk upgrade' 2019-02-05 09:22:14 0 error -> much better, rebooting 2019-02-05 09:22:29 famous last words 2019-02-05 09:23:43 rebooted, and still alive ;P 2019-02-05 09:28:38 cg 2019-02-05 11:50:25 pyvpx: so I'm intrigued, what are you building? :D 2019-02-05 11:50:46 I recognise name/domain but I can't place you heh 2019-02-05 12:01:04 hm, is there an env I can pass to setup-alpine to skip swap? 2019-02-05 12:01:20 or better yet, tell it to use a file not a partition? 2019-02-05 12:54:14 jwh: dfz.watch :-) 2019-02-05 12:54:49 o 2019-02-05 12:55:05 are you planning a party for 1024 day? 2019-02-05 12:55:13 768k day you mean? 2019-02-05 12:55:18 that is definitely the next party 2019-02-05 12:55:20 might be a bit late for that, it's close 2019-02-05 12:55:34 but no, now that bgpmon is shutting down, I'm trying to whip up a replacement service 2019-02-05 12:55:44 ah 2019-02-05 12:55:57 I started building similar, I just never got round to finishing it 2019-02-05 12:56:03 as57335 was more of a "comically large anycast instance" that I meant to blog about and turn into a marketing thing. but now it has a real use so... 2019-02-05 12:56:19 ah 2019-02-05 12:56:36 I had one but I let it all expire because I've been way too busy with real networks heh 2019-02-05 12:56:42 unfortunately running openbsd on these VPS providers is...painful 2019-02-05 12:56:46 yeah 2019-02-05 12:57:01 and I need a bit more monitoring. so if I'm going to stop using things in base openbsd, might as well switch to a linux 2019-02-05 12:57:10 and alpine is the only linux I'll consider running :) 2019-02-05 12:57:31 I used to use openbsd prettty extensively (in production too) but it's just not fast enough or has the features I need 2019-02-05 12:57:49 the only thing it has over linux atm is it has a working VPLS/PW data plane 2019-02-05 12:58:21 but patches exist for linux, when I get round to fixing them up 2019-02-05 12:59:04 I'd really like pure v6 VPLS (or even l2tpv3 pw, whatever really), but alas the bits aren't there yet 2019-02-05 12:59:17 no EVPN either *sigh* 2019-02-05 12:59:31 I feel you 2019-02-05 12:59:44 claudio@ is doing great foundational work on bgpd 2019-02-05 12:59:53 so hopefully in 2020 we'll see some more features like EVPN and what not 2019-02-05 13:00:29 yeah 2019-02-05 13:00:49 unfortunately it's not really bgpd that stops me 2019-02-05 13:01:06 platform support, any management (like being able to see optic info), horrendous performance 2019-02-05 13:01:09 :( 2019-02-05 13:01:32 curious what kind of perf you're experiencing, specifically. but probably not on topic for #alpine ;) 2019-02-05 13:01:37 it's a shame, but I had to ditch it as the network grew as it wouldn't scale with simple design 2019-02-05 13:01:49 #alpine-offtopic please :) 2019-02-05 13:01:49 and yeah, I've been meaning to port freebsds driver for DDM 2019-02-05 13:01:58 PM is probably better 2019-02-05 13:02:08 sorry 2019-02-05 13:02:17 I mean, I'm interested, I'm reading along with half an eye 2019-02-05 13:02:31 one of my pet projects is related to large-scale routing and monitoring 2019-02-05 13:02:35 oh, if you wanna join in we can certainly discuss it in offtopic 2019-02-05 13:02:36 are* 2019-02-05 17:00:56 Hi, executing 'apk upgrade' on 3.9 gives me this currently: "ERROR: tini-static-0.18.0-r0: BAD signature" How can I fix this? 2019-02-05 17:37:24 turns out alpine is the first distro to start using xfce 4.13, which is p. cool - but kinda curious why since it hasn't even been released yet? 2019-02-05 17:38:09 noticed that almost every theme broke except for greybird and the default one, which really isn't that bad since greybird is the best :p 2019-02-05 22:10:25 I'm not sure but I think it's an Alpine problem and one that has persisted for years 2019-02-05 22:10:48 at least the entire time I've used Alpine (3.6 or so) it's been that way 2019-02-06 02:54:40 anybody alive 2019-02-06 02:55:28 I have a quick question running Ubuntu 18.04.1 as a DomU? 2019-02-06 02:56:50 It appears that netplan is breaking Xen network driver. 2019-02-06 02:57:47 Has anybody being able to run Ubuntu 18.04 on 3.8.2? 2019-02-06 04:03:27 Has anybody being able to run Ubuntu 18.04.01 on 3.9.0? 2019-02-06 04:03:38 I just upgrade to 3.9 release 2019-02-06 04:28:30 exut 2019-02-06 04:28:32 exit 2019-02-06 04:28:33 quit 2019-02-06 11:42:34 hello 2019-02-06 11:42:52 https://pkgs.alpinelinux.org/package/edge/main/x86_64/libc-dev what this packages? musl-libc or glibc? 2019-02-06 11:44:38 FreeBDSM: meta package 2019-02-06 11:44:48 AinNero: meaning? 2019-02-06 11:45:18 To pull it correct libc, which is musl at the momebnt 2019-02-06 11:45:22 *moment 2019-02-06 11:45:37 should be in the description field 2019-02-06 11:47:19 AinNero: so it will install https://pkgs.alpinelinux.org/package/edge/main/x86_64/musl-dev right? 2019-02-06 11:47:30 yes 2019-02-06 11:47:41 thank you 2019-02-06 11:47:52 look at the right part of the page, under depends 2019-02-06 11:48:29 yeah, seen it, just wanted to make sure I got it right 2019-02-06 11:48:43 FreeBDSM: what do you think where we look at? 2019-02-06 11:53:14 do I get it right that there's no glibc available for alpine at all? 2019-02-06 11:53:24 it isn't meant to run glibc 2019-02-06 11:53:35 i believe there is a glibc package, let me check 2019-02-06 11:53:44 no not in aports 2019-02-06 11:53:46 by 'it' you mean alpine itself? 2019-02-06 11:53:58 you cannot just swithc libc 2019-02-06 11:54:01 yeah, no glibc in aports 2019-02-06 11:54:04 I mean https://pkgs.alpinelinux.org/packages?name=*glibc* returns 0 2019-02-06 11:54:07 all the software on alpine is built against glibc either way 2019-02-06 11:54:09 yes, it isn't in aports 2019-02-06 11:54:17 alpine has no multiarch, so having glibc somethere except in cross toolchains doesnt make sense 2019-02-06 11:54:23 +1 2019-02-06 11:54:27 https://wiki.alpinelinux.org/wiki/Running_glibc_programs 2019-02-06 11:54:31 see this 2019-02-06 11:54:48 danieli: it has a big WIP-Warning 2019-02-06 11:55:01 I don't have a particular need to run glibc programs 2019-02-06 11:55:04 right, yeah, that's a given - disclaimer: *big WIP label* 2019-02-06 11:55:12 alpine just isn't meant to run glibc 2019-02-06 11:55:19 it doesn't support systemd either 2019-02-06 11:55:36 you mean, systemd rejects patches to support us, lol 2019-02-06 11:55:52 it goes both ways, let's not argue over that 2019-02-06 11:56:22 ("it doesn't support systemd either" isn't meant to push guilt around) 2019-02-06 11:56:47 what was the reason why alpine at start was decided in favor of being based on musl rather than gnu libc? 2019-02-06 11:57:00 being different, and musl being more sane 2019-02-06 11:57:06 just overall more sane, in a lot of ways 2019-02-06 11:57:17 it has its quirks, but eh, not a big tradeoff 2019-02-06 11:57:25 flatpak should help a lot to run glibc-only stuff, not? 2019-02-06 11:57:30 *coughs* LDSO paths *coughs* 2019-02-06 11:58:42 steinex: i'd usually shoo away people using container engines to get away with not having to update openssl 2019-02-06 11:59:27 i try keeping personal opinions out of it most of the time 2019-02-06 11:59:38 this was more a technical remark, not a ideological one :) 2019-02-06 12:03:01 from the technical point, it surely works 2019-02-06 12:57:25 hi, can you guys please help me how can i easily get existing APKBUILD to modify it and rebuild 2019-02-06 12:57:38 <_ikke_> devachandra: clone aports 2019-02-06 12:57:51 thx 2019-02-06 15:12:24 hello! is it possible to install some package from edge on stable? 2019-02-06 15:13:15 I see that openconnect may be available in edge, but I would like to remain on the stable channel. 2019-02-06 15:14:17 you cag tag repositories in the repo list 2019-02-06 15:14:27 '@edge http://....' 2019-02-06 15:14:33 and then apk add foobar@edge 2019-02-06 15:14:48 but i would not recommend it for production 2019-02-06 15:15:45 will that make all packages available in that repository available and installable when I do apk update && apk upgrade? 2019-02-06 15:17:02 no 2019-02-06 15:17:07 yeah; if you have a package installed by "apk add pkg@edge" then upgrades to that package come from edge as well 2019-02-06 15:17:26 ah, I see! 2019-02-06 15:17:28 neat 2019-02-06 15:17:30 if you want to install a pkg from that repo, you need to append '@edge' to it 2019-02-06 15:17:52 thank you, I'll try that 2019-02-06 15:17:56 so AinNero's answer is the same as mine although completely opposite :) 2019-02-06 15:18:07 yeah noticed that weirdly as well 2019-02-06 15:18:16 i blame the questions phrasing 2019-02-06 15:55:07 is there a way to list information about packages, like all installed files and what package a file belongs to? 2019-02-06 15:55:36 if a pkg is installed, you can use apk info -L to list its files 2019-02-06 15:58:20 I can see a "dnsd" service in init.d. how do I find out who installed that service and what it is good for? this is why I asked the second question. 2019-02-06 15:59:04 also, I started the ntpd service and it said "starting busybox ntpd". however, I also had ntpsec installed. 2019-02-06 15:59:28 dnsd is special since its i no package, its a busybox symlink 2019-02-06 15:59:30 File: '/usr/sbin/dnsd' -> '/bin/busybox-extras' 2019-02-06 16:00:17 I get that most default services are part of busybox 2019-02-06 20:24:23 Did anyone tried to run ajenti on alpine? 2019-02-07 05:10:56 Could someone explain what are the difference in different ''version'' there https://alpinelinux.org/downloads/ 2019-02-07 07:13:34 Any recommended Dockerfile's to look at for supporting running some PHP scripts on a http? No SSL required since there would be a load balancer doing that. 2019-02-07 07:17:02 hendry: I can't speak to dockerfiles, but it should just be a matter of installing and launching php-fpm 2019-02-07 09:38:21 Is possible to override job name based on label ? I am using generic job for scraping k8s endpoints, but I need to rename job name in prometheus for node-exporter and kube-state-metrics. 2019-02-07 09:39:30 ups, wrong channel 2019-02-07 10:03:46 hi folks... i just realized that there is no working graphical webbrowser on alpine-x86 2019-02-07 10:03:56 or did i miss something? 2019-02-07 10:04:09 midori only gives me a white screen 2019-02-07 10:04:51 ah, thats sad, but sadly a results of ongoing trends 2019-02-07 10:05:07 midori should be fixed tho 2019-02-07 10:06:04 oh, didn't test in the last days, but it was a gtkwebkit bug. other browsers, which uses the library had the same bug. will test later today 2019-02-07 10:06:24 (surf for example, if i remember correctly) 2019-02-07 10:06:35 rust on 32bit musl is still broken, rihgt? 2019-02-07 10:09:07 romanis: most upstreams are not 'friendly' to musl 2019-02-07 10:10:36 :-/ 2019-02-07 10:10:50 especially not for 32-bit musl, i guess. 2019-02-07 10:11:13 but if midori is fixed, it's good enough for me. 2019-02-07 10:12:04 imo, 'big players' related to some corporation/big money are those who ignores musl 2019-02-07 10:12:45 and doesn't matter 32bit or 64bit 2019-02-07 10:14:39 for example, upstream rust is not ready for arm 64bit 2019-02-07 10:22:28 ah, k. that's a shame. firefox was nicely portable before the switch to rust. 2019-02-07 10:29:58 romanis: I'm working on bootstraping rust on aarch64 and armv7 for Alpine. Actually i did that and firefox works on arm64 and arm32 on Alpine v3.9 but for now only in my testing boxes 2019-02-07 10:31:38 x86 and ppc looks harder but maybe someone with deep knowledge will port it also on these arch'es 2019-02-07 10:34:11 actually, I have patches for x86 and ppc to build rust on musl, but don't have cross-build system 2019-02-07 10:34:42 mps: that's great news. 2019-02-07 10:36:19 thanks for your work! 2019-02-07 10:42:49 you are welcome, I work that mostly for selfish reasons, I use it and I like Alpine to be better ;) 2019-02-07 11:26:32 the entire Linux world never considered the possibility of having multiple libc's, resulting in such heavy dependency in glibc that using everything else takes a lot of extra effort 2019-02-07 11:30:54 We should make poetterings "non-useful libc" a banner here 2019-02-07 11:46:52 first versions of linux played with bsd libc, iirc 2019-02-07 11:49:13 I forgot exact versions and story (that was nearly thirty years ago, few years less) but for sure there were idea and even some actual work to use libc, but after some time switch to glibc is done 2019-02-07 11:58:15 btw, openwrt was built on uclibc and nowadays on musl, so it is not big problem for linux kernel, some big apps are problematic 2019-02-07 11:59:20 I currently don't use musl on my desktop since stuff I want to use segfaults and I don't have the capacity to diagnose/fix it :/ 2019-02-07 12:05:44 yeh, the kernel works fine with musl, it's the apps that are coded using non-standard glibc extensions 2019-02-07 12:17:32 i switched to alpine completely a few months ago. my base x64-based laptop works nearly flawlessly and extremely fast. 2019-02-07 12:17:58 the very few applications that won't run out of the box are dockerized. 2019-02-07 12:18:14 (mostly closed source, so dockerize them is a nice security gain, too. 2019-02-07 12:18:15 ) 2019-02-07 12:18:24 romanis: so if I need to setup Alpine on desktop/laptop, I know who is the master to ask. Thanks 2019-02-07 12:19:04 last time I tried with vbox/qemu+kvm, still strugled with xfce ... 2019-02-07 12:19:11 i use mate 2019-02-07 12:19:19 okay, good to know 2019-02-07 12:19:35 almost as lightweight. 2019-02-07 12:19:42 i don't remember window managers being any trouble on Alpine 2019-02-07 12:20:00 circa 3.5, 3.6 or something ... 2019-02-07 12:20:08 well, it's not the window-manager, but the plugins. 2019-02-07 12:20:28 like sound control, battery indicator, power-management etc... 2019-02-07 12:27:56 XFCE had a problem with theming at some point, don't know if it's fixed now that some components of XFCE are at 4.13 2019-02-07 12:31:36 right! i remember icons missing and stuff 2019-02-07 12:39:33 tmhoang: my wife runs Alpine with xfce for about one year 2019-02-07 12:40:21 mps: Idk if it's wierd to ask for her contact *joking* 2019-02-07 12:41:03 tmhoang: just look for the nerd girls, drop them a iso image and they'll figure it out themselves 2019-02-07 12:43:44 tmhoang: nothing special, firefox, gimp, geeqie, claws-mail, mpv and maybe something else 2019-02-07 12:45:29 but I run exclusively on alpine all my work, only have one lxc with minimal debian for kicad (electonic design software) and sometimes cross building some kernels 2019-02-07 12:46:38 the icon problem was easily fixed by installing an icon theme, I used Faenza ... but gtk apps lost their themes completely if you ever switched themes via the XFCE control panel 2019-02-07 12:49:05 TBB: I use awesome and installed some random theme from net in ~/.themes configured it in .gtkrc-2.0 and .config/gtk-3.0/settings.ini. don't run settings daemon, all works fine 2019-02-07 12:50:24 although I build apk for xsettingsd and used it some time but stopped because it becomes useless if manual config are right 2019-02-07 12:51:41 I'm an i3wm guy myself 2019-02-07 12:51:54 I thought to post patch to aports for xsettingsd but forgot and now not sure if it would be useful. 2019-02-07 12:54:54 I'm surprised when people ask (every few days or weeks) does anyone runs Alpine on desktop. It is not hard and complicated and (I) don't understand why people have such perception 2019-02-07 12:55:36 mostly because a lot of things don't "just work" for desktop use 2019-02-07 12:55:54 they do work but require a bit of finagling 2019-02-07 12:56:18 I think that could be fixed by a couple of extra packages with the required settings changes 2019-02-07 12:56:56 I use awesome, and fonts don't work out of the box, for example 2019-02-07 12:57:14 you have to figure out what to install and configure 2019-02-07 12:58:47 also, firefox was pretty broken for a long time, it works now but it wasn't a good look for desktop users 2019-02-07 12:58:55 qman__: I think that the people have to understand tools they use and not blindly install everything they see in distro or on the net 2019-02-07 13:00:23 last drop of water which forced me to switch from debian to alpine after twenty years was firefox, and I don't regret it is better 2019-02-07 13:00:24 fonts are a pretty basic part of a GUI and should be a dependency of those packages, in my opinion 2019-02-07 13:01:25 not saying all fonts, just at least one standard set of mono, sans, and serif 2019-02-07 13:01:41 so that apps expecting them work correctly 2019-02-07 13:02:09 'dependency hell' was one of thing which droves me from debian, I prefer to decide what I will have on my system and not give decision to package manager 2019-02-07 13:03:04 on the other side, on alpine there are users complaining that some things dont depend on other things they need quite often 2019-02-07 13:03:06 at the very least, it should be in the documentation 2019-02-07 13:03:11 it isn't today 2019-02-07 13:04:00 or at least wasn't when I last ran through an install a month or so ago 2019-02-07 13:04:46 Alpine is 'Small' so it couldn't (and shouldn't) be next debian/ubuntu or similar distros where everything is ready made 2019-02-07 13:05:04 this is the firefox problem I'm referring to, by the way https://bugs.alpinelinux.org/issues/4248 2019-02-07 13:06:54 don't misunderstand, I like alpine a lot, and I'm actively using it as a desktop, but if you want to know why people think you can't, these are the reasons 2019-02-07 13:08:07 keep in mind that upstream providers of big software doesn't care much for musl and other minimalistic system, so you can't expect everything to work 'out-of-box' as on glibc distros 2019-02-07 13:08:17 qman__: for people like me, they are the reason why im using alpine at all 2019-02-07 13:08:34 mps: you can't expect that, either 2019-02-07 13:08:53 glibc distros still need alot of patches as well, just not as much as for musl.. 2019-02-07 13:09:01 yep 2019-02-07 13:09:33 maybe we need a measurement for brokenness, like 'LoC patches per downstream' 2019-02-07 13:09:35 AinNero: personally no, but I presume that i know their resons and not happy with my (mis)understanding 2019-02-07 13:09:37 you can't expect any given software to just work from upstream, but the people building the distro make it work, so the end user doesn't have to hack it together 2019-02-07 13:09:50 firefox being unusable for two years isn't a good look 2019-02-07 13:12:10 obviously, they had other priorities, and that's fine, but that's a bad desktop experience 2019-02-07 13:12:14 qman__: then we have different understanding of term "usable", it is quite usable to me about last one and a half of year on Alpine (maybe just two, but not sure, I don't count time seriosly( 2019-02-07 13:12:42 s/(/)/ 2019-02-07 13:14:13 none of the context menus working is pretty unusable to me 2019-02-07 13:14:39 which was the case for a long time 2019-02-07 14:19:43 On the subject of usability for desktop -- I've been working with Alpine on openbox for the last week on one of my older laptops and it's been great, but I am running into issues with icons missing in thunar even after installing faenza 2019-02-07 14:21:22 afaik the icons need to be selected using something like lxappearance 2019-02-07 14:21:45 That makes sense 2019-02-07 14:22:15 It would be cool if obconf was available, need to create an issue for that 2019-02-07 14:38:26 it would be cool if cowsay would be available xD 2019-02-07 14:56:48 Lol :( 2019-02-07 15:34:43 the same script on alpine3.8 gcc6 links statically and on 3.9-gcc8 it links everything except musl itself... ? 2019-02-07 15:36:24 https://pastebin.com/raw/tNX5hc9d 2019-02-07 18:03:25 i guess it was the cmake version switch that changed some defaults, since it fixed it adding search start and search end static 2019-02-07 18:16:28 untoreh, can you properly report what the broken link command is? 2019-02-07 18:16:36 i suspect it's a regression in cmake package that needs to be fixed 2019-02-07 18:17:14 probably someone upstream at cmake thought "you can't possibly actually want to static link everything, let's just static link individual libs and let system stuff stay dynamic" 2019-02-07 18:17:23 like what libtool does 2019-02-07 20:56:17 blaaah 2019-02-07 21:00:08 so, what is the correct way of creating a bridge? 2019-02-07 21:00:25 it just errors with unknown interface br0 or whatever if I do what the wiki says 2019-02-07 21:01:10 what does the wiki say? 2019-02-07 21:01:22 https://wiki.alpinelinux.org/wiki/Bridge#Bridging_for_KVM 2019-02-07 21:02:31 jwh: if you "lsmod | grep bridge" 2019-02-07 21:02:34 do you get anything? 2019-02-07 21:02:49 yup, it's loaded 2019-02-07 21:03:03 just use the raw brctl commands 2019-02-07 21:03:09 to create a bridge 2019-02-07 21:03:13 and add some interfaces to it 2019-02-07 21:03:17 how does that help me at boot? :D 2019-02-07 21:03:18 see if that works at least 2019-02-07 21:03:40 it helps you determine if it even works in the first place for one 2019-02-07 21:03:50 well obviously that works, thats all fine 2019-02-07 21:03:55 unless you wanna try and debug this by running an init script over and over 2019-02-07 21:04:03 the network scripts aren't doing what they're supposed to 2019-02-07 21:04:12 without actually understanding whether or not vbridging actually _works_ 2019-02-07 21:04:36 it isn't my first rodeo, if it was another reason I'd have noticed heh 2019-02-07 21:05:12 well, I'm not particularly interested in digging into alpine's busted initscripts right now 2019-02-07 21:05:18 lols 2019-02-07 21:05:19 so I wish you luck 2019-02-07 21:05:33 as much as I love the simplicity, sometimes I wish for better network config 2019-02-07 21:05:50 netifrc from gentoo seems like it's been pretty solid for me 2019-02-07 21:06:02 it's been ported to an alpine derivative that I use 2019-02-07 21:06:22 oh hm 2019-02-07 21:06:32 unfortunately it doesn't seem to be in alpine's repos 2019-02-07 21:06:34 someone should just port netifd and be done with it heh 2019-02-07 21:07:01 but fwiw, network configuration sucks in general on linux 2019-02-07 21:07:09 I haven't seen a good solution yet 2019-02-07 21:07:11 oh, yeah, it absolutely does 2019-02-07 21:07:11 to do it 2019-02-07 21:07:22 been spoilt by systemd-networkd heh 2019-02-07 21:07:27 there are things that "work" 2019-02-07 21:07:40 but that's about the extent of it 2019-02-07 21:08:21 hm 2019-02-07 21:08:25 wonder what else is in the repo 2019-02-07 21:09:01 not much it would appear :( 2019-02-07 21:09:44 local.d it is until I have time to port something over 2019-02-07 21:21:15 jwh: you did 'apk add bridge' 2019-02-07 21:22:30 pretty sure 2019-02-07 21:22:55 although that just raises new questions 2019-02-07 21:23:25 I hadn't even considered that its kinda absurd to have to install something 2019-02-07 21:24:48 it worked for me after adding bridge apk 2019-02-07 21:25:05 huh, maybe it broke, coz its there in my history 2019-02-07 21:25:34 nm 2019-02-07 21:25:58 interfaces is still pretty hideous 2019-02-07 22:34:19 hi. Is there any chance to add nginx vts module as a package? 2019-02-07 22:35:12 I didn't mind networkmanager once I went through the #@$¡ effort to configure it 2019-02-07 23:09:07 FreeBDSM: if you commit it and everything is in order, I don't see why not 2019-02-08 00:07:30 Could anyone explain the difference between the multiple ''versions'' of alpine linux on the official download pages ? 2019-02-08 00:08:56 musl makes me sad sometimes, cockroach is still a PITFA 2019-02-08 00:16:22 ? 2019-02-08 00:40:22 What's the difference between standar and extended 2019-02-08 01:50:13 KH405_TV: check aports/scripts/mkimg.standard.sh 2019-02-08 01:54:51 Evening all looking for a starting point on how to cross-compile aports of my desktop if possible? 2019-02-08 03:12:19 danieli: Weld I would have like to know without downloading them all ... 2019-02-08 10:21:41 KH405_TV: you can view it at git.alpinelinux.org or github 2019-02-08 15:46:45 danieli: not sure if you remember me from a while back (100% cpu with dockerd running under vmm/vmd on OpenBSD), but I was able to "solve" my issue 2019-02-08 15:47:25 upgrading from alpine 3.8.2 -> 3.9.0 fixed it for me. I suspect something in the kernel/userland was fixed since I appeard to be running the same version of docker in each 2019-02-08 15:47:35 hmm. no updates to vmm/vmd on the host? 2019-02-08 15:48:21 negative - no relevant patches to vmm (and I was able to reproduce on the latest snapshot of OpenBSD) 2019-02-08 15:48:29 hm, that's odd 2019-02-08 15:48:40 maybe it was a timekeeping issue or something - no idea, really 2019-02-08 15:48:42 strace in the guest show anything interesting on the offending process? 2019-02-08 15:48:52 i don't remember what troubleshooting you did last time 2019-02-08 15:49:39 I honestly never really got that far. The strange thing was that a Lenovo T420 ran perfectly fine with docker/alpine as a client container. The X1 Carbon 5th gen I had was the offender 2019-02-08 15:49:58 I assumed it was a vmm problem due to it being so new (it still could have been, but haven't heard back from mlarkin@) 2019-02-08 15:52:40 but anyway, thanks for alpine. I've really gotten to enjoy it over the past few months. It seems like the other distros are actively pushing me away 2019-02-08 17:13:11 esdererI fins alpine a good Linux dist 2019-02-08 17:13:22 I Find* 2019-02-08 17:23:24 <_ikke_> thanks 2019-02-08 18:06:22 can I somehow force uninstall a package, even if some other package is still depending on it? 2019-02-08 18:07:02 apk del -r 2019-02-08 18:09:00 although it will uninstall (delete) all dependant packages 2019-02-08 18:10:49 well I don't want to delete the packages depending on it. this is just for testing so stuff may break because of it 2019-02-08 18:11:28 seems `apk del --force-broken-world` does the trick 2019-02-08 18:11:34 actually no ignore that 2019-02-08 18:15:19 hi short question hiw can i add mióre temmolates to lxc 2019-02-08 18:15:38 i have only busybox template 2019-02-08 18:16:35 PureTryOut[m]: right, but you then have 'broken world' 2019-02-08 18:17:42 installing/deinstallig packages on Alpine is fast, anyway :) 2019-02-08 18:18:35 hm, out of curiosity I tried out that command on the packages that refuse to upgrade on my system -- apk went into an infinite loop lol 2019-02-08 18:19:29 yeah but I just don't want to deinstall the packages depending on it. it's out of the question for what I'm trying to do 2019-02-08 18:19:56 the package depending on it is a meta package which would uninstall a lot of other stuff too 2019-02-08 18:20:27 hm, how about creating a new meta package with everything you want to keep that doesn't depend on the package first? 2019-02-08 18:21:04 since most likely there's some packages in there that would break without that dependency being installd 2019-02-08 18:21:35 I guess, but that seems like an ugly workaround for something that a simple `--force` flag could do in a package manager 2019-02-08 18:22:25 such a --force flag isn't easy to provide or work with 2019-02-08 18:22:34 it would interfere with all future package activity 2019-02-08 18:23:05 the right way to implement such a --force is probably to install a fake empty package that "provides" the package being removed 2019-02-08 18:23:57 I'd think that future package activity would make that package be installed again, which would've been fine for my use-case 2019-02-08 18:24:17 I'll just stick with the "copy of meta package but not just" approach then... 2019-02-08 19:44:19 so i got fixed it 2019-02-08 19:44:50 in lxc package is not included the download script 2019-02-08 19:45:31 i have copied manuelly to it and fix the parametes and it's worked :-D 2019-02-08 19:47:06 Hi. How might i drop root privileges/drop to a user in a sh script? 2019-02-08 19:47:33 sudo/su and re-execute the script 2019-02-08 19:48:19 [ "$id" = 0 ] && su -l username $0 "$@" 2019-02-08 19:48:37 (i just wrote this from my head, dont quote me on that) 2019-02-08 19:50:01 AinNero: thanks .. simply "su " seems to not work (https://pastebin.com/1JE5NP2m) ..It will drop me to a shell under the user, but when i type 'exit' it will drop to the tmux session 2019-02-08 19:50:59 PureTryOut[m]: its not so ugly as suggested. you can do that with --virtual 2019-02-08 19:51:38 ola_norsk: for your usecase, you might try su -c 'tmux ....' 2019-02-08 19:52:04 take care because this makes correct quoting more difficult 2019-02-08 19:52:46 AinNero: tried that, and yeah, i came across problems getting the tmux command encompassed by "'s 2019-02-08 19:54:08 another idea: you could try sh -l username < with heredoc syntax and pipe it into the shell started by su 2019-02-08 19:54:30 oh, i meant su -l username, not sh 2019-02-08 19:56:24 AinNero: [ "$id" = 0 ] && su -l mysticbbs $0 "$@" 2019-02-08 19:56:48 AinNero: will still start the following tmux as root user 2019-02-08 19:57:50 ah, it doesnt exit there 2019-02-08 19:58:08 'exec su' instead of 'su' 2019-02-08 19:58:18 so the root-owned process will stop there 2019-02-08 20:01:36 i'm guessing it'll be less clutter to move the tmux stuff to a seperate script then, and use 'exec su mysticbbs -c