#archlinux32 | Logs for 2020-01-31

[01:24:14] <buildmaster> i686/cargo-tree is broken (says eurobuild6-3): https://archlinux32.org
[01:25:26] <buildmaster> pentium4/cargo-tree is broken (says eurobuild6-6): https://archlinux32.org
[01:27:26] -!- buildmaster has quit [Remote host closed the connection]
[01:27:38] -!- buildmaster has joined #archlinux32
[01:27:38] <buildmaster> !rq buildmaster
[01:27:39] <phrik> buildmaster: <buildmaster> I might be insane, but never confused ... ;-)
[03:19:27] <buildmaster> i686/cbindgen is broken (says eurobuild6-5): https://archlinux32.org
[03:21:19] <buildmaster> pentium4/cbindgen is broken (says eurobuild6-6): https://archlinux32.org
[03:23:19] -!- buildmaster has quit [Remote host closed the connection]
[03:23:30] -!- buildmaster has joined #archlinux32
[03:23:30] <buildmaster> !rq buildmaster
[03:23:31] <phrik> buildmaster: <buildmaster> I might be insane, but never confused ... ;-)
[07:15:24] -!- abaumann has joined #archlinux32
[07:15:24] <buildmaster> Hi abaumann!
[07:15:24] <buildmaster> !rq abaumann
[07:15:25] <phrik> buildmaster: <abaumann> mmh.. my build slaves are on strike? Maybe I should pay them? ;-)
[07:27:10] <abaumann> java8-openjdk was a local checksum patching of mine which I removed now..
[07:27:28] <abaumann> ..so, I should maybe not do this normally or check periodically for such checksum patches
[07:34:51] <abaumann> This rsync lacks old-style --compress due to its external zlib. Try -zz.
[07:34:51] <abaumann> Continuing without compression.
[07:34:53] <abaumann> hehe. :-)
[07:35:43] <abaumann> ==> ERROR: Failure while downloading rsync://rsync.osuosl.org
[07:36:17] <abaumann> mmh. but now it works when building mariadb manually.. mmh.
[07:36:26] <abaumann> let's check what rsync version we have..
[07:56:30] <abaumann> mmh works.. so, rescheduling mariadb
[08:13:47] <abaumann> zstd: error 11 : Allocation error : not enough memory
[08:13:54] <abaumann> yeah. coolMem: 64302 7881 1776 30 54644 55682
[08:14:04] <abaumann> and I cannot create a mariadb package
[08:14:46] <abaumann> so, the next compression algorithm in pacman will require 1 TB of RAM? ;-)
[08:17:12] -!- abaumann has quit [Quit: leaving]
[08:23:12] -!- AndrevS has joined #archlinux32
[08:25:00] <girls> !grab abaumann
[08:25:01] <phrik> girls: ๐ŸŽ‰
[08:25:25] <girls> I turned off zst compression on my raspis, too - I now do the compression externally on an x86_64 machine
[08:25:46] <buildmaster> i686/d-mir-core is broken (says rechenknecht): https://archlinux32.org
[08:26:22] <buildmaster> pentium4/d-mir-core is broken (says rechenknecht): https://archlinux32.org
[08:29:58] -!- T0MuX has joined #archlinux32
[08:31:00] <T0MuX> girls: Hi, I meant, why the package "wine" (and not "wine-staging") aren't in community repo ?
[08:31:09] <girls> ah, ok
[08:31:16] <girls> it was being built
[08:33:29] <girls> plus: our policy is, that packages must traverse [testing]/[community-testing] first
[08:33:35] <girls> and ideally be tested meanwhile
[08:33:44] <girls> and only *then* enter the stable repositories
[08:33:56] <girls> I force-moved wine to [community] now
[08:37:57] <T0MuX> Yes, I understand that
[08:38:15] <T0MuX> and this is also why you choose wine-staging instead of wine ?
[08:38:22] <girls> no
[08:38:33] <T0MuX> (because wine-staging is also a testing version itself)
[08:38:34] <girls> I guess, some script ran amok and deleted wine
[08:38:41] <T0MuX> oh crap
[08:38:47] <T0MuX> okay
[08:39:09] <T0MuX> so last question ๐Ÿ™‚ wine will comes in community repo a day ?
[08:39:10] <girls> we had wine in the past: https://archive.archlinux32.org
[08:39:12] <phrik> Title: Index of /packages/w/wine/ (at archive.archlinux32.org)
[08:39:18] <girls> it should be there already
[08:39:21] <girls> (since ~1/4 h)
[08:39:27] <T0MuX> okay I understand
[08:39:33] <T0MuX> oh ?
[08:39:36] <T0MuX> you mean right now
[08:39:40] <girls> yeah :-)
[08:39:41] <T0MuX> I check..
[08:40:00] <girls> I moved the i686 wine a few minutes ago, the pentium4 wine ~1/2 day ago
[08:40:01] <T0MuX> and that's a win :D
[08:40:05] <girls> :-D
[08:40:09] <T0MuX> thank you !
[08:40:13] <girls> np
[08:41:15] <T0MuX> oh, wine seams to have a lot more deps than wine-staging
[08:41:32] <T0MuX> pacman just installed 50 more packages
[08:42:06] <buildmaster> i686/linux is broken (says eurobuild6-6) - already flagged out-of-date upstream on 2020-01-27: https://archlinux32.org
[08:43:06] <buildmaster> pentium4/linux is broken (says eurobuild6-4) - already flagged out-of-date upstream on 2020-01-27: https://archlinux32.org
[08:43:23] <T0MuX> so, I'm facing a keyboard layout issue : I have KEYMAP=fr in /etc/vconsole.conf and this is not working
[08:43:24] <T0MuX> still qwerty
[08:44:55] <girls> I think, there was an issue, that you needed to uncompress the layout file manually, because gzip failed in early boot
[08:45:29] <girls> abaumann: we should really fix this ^ (or was it unfixable?)
[08:48:35] <buildmaster> i686/dmapi is broken (says rechenknecht): https://archlinux32.org
[08:49:24] <buildmaster> i686/libpagemaker is broken (says rechenknecht): https://archlinux32.org
[08:49:56] <buildmaster> pentium4/libpagemaker is broken (says rechenknecht): https://archlinux32.org
[08:52:08] <buildmaster> i486/mariadb is broken (says eurobuild6-7-i486): https://archlinux32.org
[08:54:21] <buildmaster> i686/linux-zen is broken (says rechenknecht) - already flagged out-of-date upstream on 2020-01-27: https://archlinux32.org
[08:59:23] -!- DepositePirate has quit [Ping timeout: 240 seconds]
[09:00:01] <buildmaster> pentium4/linux-zen is broken (says rechenknecht) - already flagged out-of-date upstream on 2020-01-27: https://archlinux32.org
[09:00:42] -!- DepositePirate has joined #archlinux32
[09:15:21] <buildmaster> i686/mariadb is broken (says eurobuild6-1): https://archlinux32.org
[09:15:52] -!- deep42thought has joined #archlinux32
[09:15:52] <buildmaster> Hi deep42thought!
[09:15:52] <buildmaster> !rq deep42thought
[09:15:54] <phrik> buildmaster: <deep42thought> golang-asvnp-acwduac-sefc-evfsx-vb4fs-sefvcy is broken (says me)
[09:16:45] <deep42thought> abaumann: as a workaround for the compression issues, some other machine could handle the compression
[09:24:23] -!- abaumann has joined #archlinux32
[09:24:24] <buildmaster> Hi abaumann!
[09:24:24] <buildmaster> !rq abaumann
[09:24:25] <phrik> buildmaster: <abaumann> they do all kind of weird stuff, nobody else does, because they are known not to work. :-)
[09:24:29] <deep42thought> Hi abaumann!
[09:24:32] <abaumann> morning deep42thought
[09:25:10] <abaumann> the gzip-charset-issue was some problems in early boot with gunzip, something libsecomp related, IIRC
[09:25:15] <abaumann> not that trivial to fix
[09:25:28] <deep42thought> ok
[09:25:40] <abaumann> mariadb-compression problems only happend on my slaves, maybe they had some 4GB limit set in nspawn or so, must check
[09:25:47] <deep42thought> maybe we could display that information somewhere more prominently
[09:25:55] <deep42thought> (I could not find the related thread in the forum)
[09:26:04] <abaumann> yes, as soon the forums are joined :-)
[09:26:10] <deep42thought> ah, right :-D
[09:26:24] <deep42thought> I also tried to find it in the archives, though - but I couldn't locate it
[09:26:29] <abaumann> I'm currently testing some merger plugins for fluxbb
[09:26:45] <deep42thought> :-)
[09:26:57] <abaumann> I could add a read-only version of the old fluxbb on buildmaster.archlinux32.org
[09:27:11] <deep42thought> old-bbs.archlinux32.org?
[09:27:14] <abaumann> but I'm not sure whether you can browse and search if you don't have a user.
[09:27:22] <abaumann> that's something I have to check first
[09:27:23] <deep42thought> na, never mind
[09:27:42] -!- marc2377 has joined #archlinux32
[09:27:45] <abaumann> I also though joining the forums is an efford better spent.
[09:27:52] <deep42thought> exactly
[09:28:19] <abaumann> by git.archlinux32.org fiddling yesterday may have sent some packages onto the broken build list
[09:28:51] <abaumann> zabbix, #warning "The <sys/sysctl.h> header is deprecated and will be removed." and -Werror :-)
[09:29:00] <deep42thought> hehe
[09:29:12] <deep42thought> I think, breaking git will not bring packages on the broken-build-list
[09:29:26] <deep42thought> either the build slaves do have the requested commit and everything is fine
[09:29:30] <abaumann> it only breaks the slaves, or rather doesn't update them anymore
[09:29:35] <deep42thought> or they don't and will refuse to build the package entirely
[09:29:41] <abaumann> yeah. My slave scripts have a git pull as PreRun
[09:29:43] <deep42thought> yes, exactly
[09:29:46] <abaumann> that's why they complained
[09:30:33] <deep42thought> line 339 upto 347
[09:30:43] <abaumann> ah?
[09:31:07] <deep42thought> https://git.archlinux32.org
[09:31:08] <phrik> Title: build-packages ร‚ลค bin - builder - Archlinux32 build system (at git.archlinux32.org)
[09:31:10] <abaumann> oh. I see
[09:31:24] <abaumann> in build-packages, so the system is auto-updating itself already.
[09:31:24] <deep42thought> it has a `|| true` - so failure is non-fatal
[09:31:27] <buildmaster> i686/dotnet-core-2.2 is broken (says eurobuild6-6): https://archlinux32.org
[09:31:30] <deep42thought> yes
[09:31:34] <abaumann> I remove my pre-execution then..
[09:31:40] <deep42thought> :-)
[09:32:13] <deep42thought> it's *supposed* to be a no-hands-on script
[09:33:05] <abaumann> I should have read the code :-)
[09:33:20] <buildmaster> pentium4/dotnet-core-2.2 is broken (says eurobuild6-4): https://archlinux32.org
[09:33:20] <deep42thought> maybe it's such a pain to read my code?
[09:33:25] <abaumann> old saying: writing code is easier that reading code..
[09:33:32] <deep42thought> !grab abaumann
[09:33:32] <abaumann> ..well.. shellish..
[09:33:33] <phrik> deep42thought: Tada!
[09:33:56] <abaumann> quite ok for shell
[09:34:01] <deep42thought> lol
[09:34:37] <abaumann> ah.. my archlinux32-devel server has disappeared.. most likely it's on a broken VM somewhere..
[09:34:50] <buildmaster> pentium4/mariadb is broken (says eurobuild6-5): https://archlinux32.org
[09:35:01] <abaumann> I have a huch about mariadb.. :->
[09:35:18] <deep42thought> you're right
[09:35:45] <abaumann> so, all packages with a certain size cannot be compressed.
[09:35:52] <abaumann> I wonder what upstream thinks about this..
[09:36:05] <deep42thought> I noticed this on my raspi with bareos
[09:36:49] <abaumann> yeah, the internet is full of complains about zstd (mips, arm, etc. out-of-memory)
[09:36:55] <deep42thought> the biggest one is 1MB - so not *that* big ...
[09:37:18] <deep42thought> I meant 1MB uncompressed
[09:37:19] <abaumann> I would hate to deviate in the compression algorithm for Archlinux32
[09:37:33] <abaumann> but if I cannot update my Alix or build packages for it..
[09:37:36] <abaumann> ah. :-)
[09:37:53] <deep42thought> I would guess, that decompression is much more memory efficient
[09:38:03] <abaumann> yes.
[09:38:05] <T0MuX> girls abaumann but, I think the issue is not related to the "unzip", because I can load my layout with loadkeys fr
[09:38:08] <T0MuX> and it works
[09:38:11] <abaumann> and usually this doesn't need a lot of memory
[09:38:22] <deep42thought> T0MuX: yes, that's right
[09:38:28] <deep42thought> but in early boot, gunzip fails
[09:38:36] <abaumann> T0MuX: the problem is that it works normally, but not when booting in the initramdisk
[09:39:07] * abaumann digs through old forum pages, maybe he finds the related forum..
[09:40:20] <abaumann> https://old.archlinux32.org Trouble setting the propper keymap layout _ Installation _ Arch Linux 32 Forums.html
[09:41:06] <abaumann> There you have my whole investigation into the problem..
[09:41:09] <buildmaster> pentium4/dmapi is broken (says eurobuild6-3): https://archlinux32.org
[09:41:37] <T0MuX> this is weird I cant open this link
[09:41:55] <abaumann> yeah. go over the old.archlinux32.org webpage
[09:42:10] <abaumann> it has too many brackets and spaces probably :-)
[09:42:15] <abaumann> https://bbs.archlinux32.org
[09:42:16] <abaumann> mmh
[09:42:20] <abaumann> about keymap hooks
[09:42:37] <abaumann> hence it is so important to get the original ids right again.
[09:42:59] <abaumann> yeah, again. I rather try to join the forums now.
[09:49:57] <T0MuX> haha when I search "keymap" in the forum, I find nothing
[09:49:59] <T0MuX> https://bbs.archlinux32.org
[09:50:00] <phrik> Title: Info / Archlinux32 Forum (at bbs.archlinux32.org)
[09:51:06] <buildmaster> i486/dmapi is broken (says eurobuild6-7-i486): https://archlinux32.org
[09:51:45] <abaumann> yes, that's because the now forum doesn't contain the old pages.
[09:51:52] <buildmaster> i486/libpagemaker is broken (says eurobuild6-7-i486): https://archlinux32.org
[09:52:03] <abaumann> we had a..well.. hickup some time ago :-)
[09:54:20] <abaumann> compress_as in /usr/share/makepkg/util/compress.sh uses zstd -c -z -q - as default which can be overloaded with a COMPRESSZST env variable
[09:54:54] <abaumann> but only see a compression -# option for setting something else than default:3
[09:54:57] <deep42thought> you intent to send the package somewhere else for compression?
[09:55:10] <abaumann> no --please-do-not-use-enormous-amounts-of-memory-while-compressing ;-)
[09:55:17] <deep42thought> !grab abaumann
[09:55:18] <phrik> deep42thought: Bazinga!
[09:55:35] <abaumann> I was thinking to fiddle with the zstd options so that maybe all slaves have a chance to compress the package
[09:55:52] <abaumann> we can always go back to xz
[09:56:25] <abaumann> or a fallback? use zstd, unless it fails, them use xz?
[09:56:31] <abaumann> *then
[09:57:13] <deep42thought> the compression is (currently) determined by what is in the databse
[09:57:26] <deep42thought> so return-assignment expects to get *.zst packages
[09:57:49] <abaumann> yes, furthermore we should not deviate from upstream, if not necessary
[09:58:36] <deep42thought> yes
[09:58:44] <deep42thought> the question is: is it necessary ;-)
[09:59:07] <abaumann> mmh. https://github.com
[09:59:08] <phrik> Title: Using too much memory ยท Issue #1328 ยท facebook/zstd ยท GitHub (at github.com)
[10:00:29] <abaumann> I'm always sceptical when big tech companies are "donating" code to the public. :->
[10:00:37] <deep42thought> !grab abaumann
[10:00:38] <phrik> deep42thought: Tada!
[10:00:39] <deep42thought> :-D
[10:02:09] <T0MuX> wait, fb is open-source ??
[10:02:24] <T0MuX> abaumann: yes me too
[10:02:38] <T0MuX> I always think there is an hidden purpose...
[10:02:48] <T0MuX> something bad
[10:03:15] <abaumann> it's not about that.
[10:03:16] <T0MuX> just like the "chown" (lol) of github to microsoft
[10:03:33] <T0MuX> ah?
[10:03:33] <abaumann> the software is intended to run on big clusters of machines
[10:03:40] <abaumann> they serve their purpose well
[10:03:52] <abaumann> but they might not be suitable for other applications.
[10:04:07] <abaumann> it's all about tradeoffs: zstd is faster because it uses more memory
[10:04:15] <abaumann> how much "more" is "too much" depends on the application.
[10:04:45] <abaumann> what's more the problem is people gladly accepting software falling down the 'Silicon Valley Sky' :-)
[10:06:52] <abaumann> yeah, I start to understand the benefits of an architecture like OpenBuild has it: dedicated signers, compressors, etc. So you can optimize the specialized workers for their task.
[10:07:40] <abaumann> the problem is a little bit that compression is a task _inside_ makepkg, so we would have to build the package with no compression first, then send it somewhere for compression and signing.
[10:08:00] <abaumann> having dedicated signers also limits the risk of exposing sign keys as an additional benefit
[10:08:08] <abaumann> something for the todo list :-)
[10:08:15] <deep42thought> or we could hook up script inside there, with the env variable you wrote
[10:08:43] <abaumann> I would actually try first to use different compression parameters (quality)
[10:08:44] <deep42thought> though, sending uncompressed packages around might not be the best idea :-)
[10:08:57] <deep42thought> yeah, your approach is much easier
[10:09:08] <abaumann> if the better we compress means more memory usage, then we can maybe just fiddle a little bit the parameters
[10:09:25] <abaumann> besides: we should maybe first see how many packages are affected by that.
[10:09:39] <abaumann> we could add a error class to the buildmaster (like verify, source, etc.)
[10:09:57] <abaumann> it's in the database IIRC?
[10:10:02] <deep42thought> right
[10:10:15] <abaumann> the old method: monitor first, act if needed :-)
[10:10:31] <abaumann> who needs a new mariadb now anyways? ;-)
[10:10:46] <deep42thought> :-D
[10:13:58] <abaumann> must leave fast, be back soon..
[10:13:59] -!- abaumann has quit [Quit: leaving]
[10:17:58] <buildmaster> i486/vulkan-tools are broken (says nlopc46-i486bs1): https://archlinux32.org
[10:19:21] <buildmaster> i686/vulkan-tools are broken (says rechenknecht): https://archlinux32.org
[10:21:21] -!- buildmaster has quit [Remote host closed the connection]
[10:21:33] -!- buildmaster has joined #archlinux32
[10:21:33] <buildmaster> !rq buildmaster
[10:21:34] <phrik> buildmaster: <buildmaster> I might be insane, but never confused ... ;-)
[10:21:54] <T0MuX> Okay
[10:22:17] <T0MuX> I come back to my vconsole issue
[10:22:19] <T0MuX> I tried to add keymap hook in mkinitcpio.conf but nothing changed
[10:22:21] <T0MuX> this workaround is not working
[10:23:12] <buildmaster> pentium4/vulkan-tools are broken (says eurobuild6-5): https://archlinux32.org
[10:23:15] <deep42thought> that's what I remember, too
[10:23:27] <deep42thought> but IIRC, gunzipping the keymap file works
[10:23:41] <deep42thought> ... maybe we should just ship uncompressed keymap files >:-)
[10:25:59] <deep42thought> it would increase the package size by 1.6MB
[10:26:08] <deep42thought> (from 1.1MB to 2.7MB or so)
[10:26:10] <T0MuX> I also added consolefont in mkinitcpio (and removed keymap because useless) : I see the font changing just after the init is loaded, and when I come to the TTY login, everything disappear !! and get back to the "ugly oem font"
[10:26:15] <T0MuX> what does this mean
[10:26:30] * deep42thought has no idea
[10:27:51] <T0MuX> hilarious : I type "sudo systemctl start systemd-vconsole-setup" and tadam ! my font is loaded and also my keymap
[10:28:06] <T0MuX> this means THIS service is not correctly loaded at boot ?
[10:29:00] <deep42thought> what does `sudo systemctl status systemd-vconsole-setup` tell you about this (before starting it)?
[10:29:05] <deep42thought> maybe it needs some extra dependency?
[10:29:11] <deep42thought> also check journalctl for that service
[10:30:24] <T0MuX> I check
[10:30:33] <T0MuX> (have to reboot lol)
[10:31:05] <T0MuX> ah so smooth and fast ๐Ÿ™‚ already rebooted :D
[10:32:16] <T0MuX> Loaded: loaded (/usr/lib/systemd/system/systemd-vconsole-setup.service; static; vendor preset: disabled)
[10:32:18] <T0MuX> Active: inactive (dead)
[10:32:31] <deep42thought> maybe you just need to enable it?
[10:36:13] <T0MuX> it fails :/
[10:36:15] <T0MuX> I tried !
[10:38:12] <T0MuX> "The unit file have no installation config (WantedBy=, RequiredBy=, Also=, Alias= settings in the [Install] section, and DefaultInstance= for template units). This means they are not meant to be enabled using systemctl."
[10:38:26] <deep42thought> ah, so it gets activated by something else
[10:38:29] <deep42thought> hmmm
[10:39:25] <T0MuX> (I think this issue is still happening because lot of english people using qwerty, so they probably never facing this)
[10:39:44] <deep42thought> and because it only happens on 32 bit
[10:39:51] <T0MuX> (i'm french, so... :) )
[10:39:57] <T0MuX> yes also
[10:40:14] <T0MuX> but, I already had, in the past, this issue on x64
[10:40:18] <deep42thought> maybe, we should make the default keyboard layout klingon and then those kind of issues would be solved much quicker ;-)
[10:40:34] <T0MuX> mmmmm wait.... maybe it was on the old x86 archlinux... i'm not sure
[10:40:51] <T0MuX> xD
[10:41:18] <T0MuX> I had wrote a warkaround on my personnal wiki
[10:42:06] <deep42thought> you can add a systemd unit which *depends* on systemd-vconsole-setup and which is enabled (symlinked in multi-user.target)
[10:42:14] <deep42thought> this might solve the issue, too
[10:42:17] <T0MuX> my workaround is simple : create a service that starts on boot and execute localctl
[10:42:29] <deep42thought> yeah
[10:42:34] <T0MuX> but I can replace localectl by 'systemctl start systemd-vconsole-setup ^^
[10:42:34] <deep42thought> similarly good
[10:42:39] <deep42thought> :-D
[10:42:58] <T0MuX> I have this line in it : WantedBy=getty.target
[10:43:10] <T0MuX> so it starts just before the TTY login prompt shows up
[10:43:21] <deep42thought> yeah, also good
[10:43:33] <T0MuX> girls: check this out ^
[10:44:11] <T0MuX> deep42thought: ah yes, I didnt read your suggestion which is quite similar to mine
[10:44:12] <deep42thought> she's the same person (my bouncer
[10:44:13] <T0MuX> ^^
[10:44:14] <deep42thought> )
[10:44:21] <T0MuX> xD
[10:44:22] <T0MuX> okay
[10:44:36] <T0MuX> the door ? okay.
[10:47:34] <T0MuX> I reboot, to test !
[10:48:41] <T0MuX> okay great. my "manual service" starts well manually, but not on boot.
[10:48:50] <T0MuX> Oh, I said nothing !
[10:48:53] <T0MuX> it started !
[10:48:57] <T0MuX> my keyboard is FR, but my consolefont is not loaded
[10:50:01] <T0MuX> to get my font, I have to START AGAIN systemd-vconsole-setup
[10:50:15] <deep42thought> that sounds wrong
[10:50:19] <T0MuX> yes
[10:51:26] <T0MuX> but, what I dont understand is : systemd-vconsole-setup loads /etc/vconsole.conf. In vconsole.conf I have : a keymap (layout), and a font (ter-218b). So, it sounds logical that systemctl start systemd-vconsole-setup loads the layout AND the font, right ?
[10:51:40] <T0MuX> so, why at boot it only load the layout and not the font ?? xD
[10:51:40] <deep42thought> yes
[10:52:16] <deep42thought> abaumann: db-update now also uses the intentions-queue
[10:52:23] <T0MuX> (just to remind it : I have the consolefont hook in mkinitcpio, but dont know if its related)
[10:53:02] <deep42thought> starting the service afterwards should load both
[10:53:11] <deep42thought> no matter what it did during boot
[10:53:23] <deep42thought> ... or is it already running?
[10:53:32] <deep42thought> (systemd won't start the service a second time)
[10:53:50] <deep42thought> maybe, you need to change `systemctl start ...` to `systemctl restart ...` in your manual unit
[10:53:50] <T0MuX> yes because it is a oneshot service
[10:54:01] <T0MuX> ah good idea
[10:54:02] <T0MuX> have to try
[10:54:03] <deep42thought> and not RemainAfterExit=true, right?
[10:54:21] <deep42thought> at least on my box it's not
[10:54:45] <T0MuX> I have RemainAfterExit=yes
[10:54:46] <T0MuX> it shouldnt ?
[10:55:03] <deep42thought> your unit or systemd-vconsole-setup?
[10:55:51] <T0MuX> my unit
[10:56:09] <deep42thought> that's not a problem, then
[10:58:09] <T0MuX> okay, just to look at it, I opened /lib/systemd/system/systemd-vconsole-setup.service
[10:58:28] <T0MuX> I see : ExecStart=/usr/lib/systemd/systemd-vconsole-setup
[10:58:36] <T0MuX> maybe I should use the exact same line in my own service unit
[10:58:50] <deep42thought> you can extract all the info which systemd sees, by issuing `systemctl cat $unit-name`
[10:58:52] <T0MuX> to avoid usind systemctl which will use this exec command
[10:59:05] <deep42thought> (just FYI)
[10:59:29] <T0MuX> (okay thanks)
[11:01:14] <T0MuX> ok, reboot, and test .. (using same exec command than vconsole service in my own unit)
[11:01:40] <T0MuX> (also, I removed consolefont from mkinitcpio.... finally useless in my case)
[11:02:06] <T0MuX> okay SAME weird crap issue : layout loaded, not the font.
[11:02:08] <T0MuX> pfff
[11:02:55] <T0MuX> well, it's not so bad. the most important is the layout, the consolefont is "bonus"
[11:03:11] <deep42thought> yes, but it's still very strange
[11:03:27] <deep42thought> maybe the unit gets activated too early?
[11:03:40] <deep42thought> can you look for some dependency which might get it scheduled later?
[11:04:01] <deep42thought> or even better: have a look at journalctl first, if it spits out any errors about loading the font
[11:05:03] <T0MuX> you mean in the stock vconsole unit ?
[11:05:16] <deep42thought> or yours
[11:05:29] <deep42thought> especially yours
[11:05:56] <deep42thought> I would guess, the stock vconsole unit experiences similar issues like the gunzip-issue during early boot
[11:06:02] <deep42thought> but yours should really work
[11:06:58] <T0MuX> mhmm interesting : /usr/bin/setfont failed with exit status 65
[11:07:45] <T0MuX> so this is an error with setfont, which occures only on boot
[11:07:54] <T0MuX> not after boot
[11:08:58] <deep42thought> can you run setfont directly in your unit and add `-v` to see what it does?
[11:11:42] <T0MuX> oooh wait https://bbs.archlinux.org
[11:11:43] <T0MuX> have to try this
[11:11:43] <phrik> Title: [SOLVED] Can't set KEYMAP and FONT in /etc/vconsole.conf / Applications & Desktop Environments / Arch Linux Forums (at bbs.archlinux.org)
[11:18:00] <T0MuX> regen mkinit... reboot....
[11:26:16] <buildmaster> pentium4/wireguard-arch is broken (says rechenknecht): https://archlinux32.org
[11:28:22] <T0MuX> okay so this solution didnt solved in my case.
[11:29:04] <T0MuX> but it is interesting to see systemd loaded by the init : when it's over, the tty login shows up immediatly
[11:29:34] <T0MuX> so, it is longer at the initrd loading step, but way shorter after (I think there is not change on the total time)
[11:30:45] <T0MuX> when /etc/profile is loaded ?
[11:30:48] <T0MuX> after login ?
[11:31:30] <T0MuX> and /etc/environment before login ? right ?
[11:34:09] <T0MuX> this issue is really pissing me off.
[11:34:26] <deep42thought> I think, the unit is run too early
[11:34:35] <deep42thought> it does work if you start it manually, right?
[11:37:05] <T0MuX> yes
[11:37:12] <T0MuX> can we add delay to it ?
[11:37:30] <T0MuX> (I mean, to the stock unit)
[11:37:36] <deep42thought> no, not really
[11:37:49] <deep42thought> it would just delay the whole boot process
[11:38:03] <deep42thought> because it has `Before=initrd-switch-root.target`
[11:40:23] <T0MuX> how can I start my own unit (named "kbfr") just before the tty login prompt ?
[11:40:31] <T0MuX> I think there something to do with getty
[11:40:33] <T0MuX> but donno how
[11:40:45] <deep42thought> Before=tty.target
[11:40:47] <deep42thought> or something
[11:40:56] <deep42thought> (I don't know the real name of the tty target)
[11:42:06] <T0MuX> this have to be in [Install] section ?
[11:42:11] <T0MuX> or [Service]
[11:42:20] <deep42thought> [Unit], I think
[11:42:29] <deep42thought> yes, [Unit]
[11:47:31] <deep42thought> gotta go, bbl
[11:47:42] -!- deep42thought has quit [Quit: Leaving.]
[12:09:47] <T0MuX> ssssssssssssooooooooooooooooooooooooooooolveeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeed
[12:09:49] <T0MuX> :D
[12:13:05] <T0MuX> girls: check my unit on my wiki ๐Ÿ™‚ https://wiki.mux.re
[12:13:06] <phrik> Title: archlinux:cmd [WikiMuX] (at wiki.mux.re)
[12:13:12] <T0MuX> it is french, but you can read the unit content
[12:13:37] <T`aZ> year 2020, still linux keyboard issue, check
[12:28:06] <girls> T0MuX: nice :-) However, I'd recommend putting manual units into /etc/systemd/system/ - if you (plan to) package this unit, then /usr/lib/systemd/system is fine. (But that's mainly an aesthetical issue)
[12:28:30] -!- deep42thought has joined #archlinux32
[12:28:30] <buildmaster> Hi deep42thought!
[12:28:30] <buildmaster> !rq deep42thought
[12:28:31] <phrik> buildmaster: <deep42thought> I know this feeling: doing hundrets of laser shots (manually) just to discover, that your spectrometer was on auto-trigger, but not on auto-save ...
[13:23:07] -!- drathir has quit [Ping timeout: 265 seconds]
[14:23:19] -!- T`aZ has quit [Remote host closed the connection]
[14:40:36] -!- drathir has joined #archlinux32
[14:59:32] -!- T`aZ has joined #archlinux32
[15:05:03] -!- DepositePirate has quit [Ping timeout: 240 seconds]
[15:06:13] -!- DepositePirate has joined #archlinux32
[15:22:02] <nit-picker> key A0B250C0FC9FC079EC04ADB7A50C0F20AEC3AF00 (from Polichronucci (Arch Linux 32 Master Key) <polichronucci@archlinux.gr>) in package archlinux32-keyring-20200113-1.0-any.pkg.tar.zst expires on 2020-05-05 (in 94 < 100 days).
[15:36:34] -!- abaumann has joined #archlinux32
[15:36:35] <buildmaster> Hi abaumann!
[15:36:35] <buildmaster> !rq abaumann
[15:36:36] <phrik> buildmaster: <abaumann> I never used zen at all :-)
[15:36:38] <deep42thought> Hi abaumann!
[15:36:42] <abaumann> hi deep42thought
[15:37:13] <abaumann> I installed a Pentium III machine lately and also had distorted fonts.. I thought it's because the Intel chip there is way old.. but maybe this is also a vconsole issue.
[15:37:14] <deep42thought> sry, gotta go
[15:37:17] <abaumann> np
[15:37:20] <abaumann> cu :-)
[15:37:27] <deep42thought> cu
[15:37:28] -!- deep42thought has quit [Quit: Leaving.]
[16:03:32] <buildmaster> i486/vulkan-validation-layers are broken (says nlopc46-i486bs1): https://archlinux32.org
[16:16:58] <buildmaster> i486/ukui-media is broken (says eurobuild6-7-i486): https://archlinux32.org
[16:23:23] -!- abaumann has quit [Quit: leaving]
[16:34:25] <T0MuX> I am a genius.
[16:34:56] <T0MuX> https://cloud.mux.re
[16:34:58] <phrik> Title: MuX Cloud (at cloud.mux.re)
[16:35:29] <T0MuX> guys, this is why I spend all this time to make it works.... Let me introduce you my CS 1.6 LAN USB KEY ๐Ÿ˜
[16:35:56] <T0MuX> powered by archlinux32
[16:36:50] <buildmaster> i486/yubikey-manager-qt is broken (says nlopc46-i486bs1): https://archlinux32.org
[16:37:37] <T`aZ> nice
[16:46:22] <buildmaster> i486/dotnet-core-2.2 is broken (says nlopc46-i486bs0): https://archlinux32.org
[17:38:57] <buildmaster> pentium4/intel-graphics-compiler is broken (says eurobuild6-5): https://archlinux32.org
[17:43:18] <buildmaster> i686/intel-graphics-compiler is broken (says rechenknecht): https://archlinux32.org
[18:12:47] <nit-picker> key A0B250C0FC9FC079EC04ADB7A50C0F20AEC3AF00 (from Polichronucci (Arch Linux 32 Master Key) <polichronucci@archlinux.gr>) in package archlinux32-keyring-transition-20200113-1.0-any.pkg.tar.zst expires on 2020-05-05 (in 94 < 100 days).
[19:15:58] <buildmaster> pentium4/dino is broken (says eurobuild6-3): https://archlinux32.org
[20:40:07] -!- crns has quit [Quit: bye]
[22:02:41] <buildmaster> i486/fscrypt is broken (says nlopc46-i486bs0): https://archlinux32.org
[23:59:19] <buildmaster> i486/psutils are broken (says nlopc46-i486bs1): https://archlinux32.org