#archlinux32 | Logs for 2025-11-29

Back
[00:00:59] <socksinspace> alpine linux's "main" is a lot more achievable of a target than the bajillion packages of arch linux
[00:02:21] <KitsuWhooa> yeah
[00:02:58] <KitsuWhooa> I also remember Qt being a pain in the ass
[00:03:32] <KitsuWhooa> if you encounter any build issues, feel free to look at the patches in the a32 packages repo
[00:03:49] <KitsuWhooa> I think I yoinked a few alpine ones :p
[00:05:07] -!- abouvier has joined #archlinux32
[07:24:20] -!- abaumann has joined #archlinux32
[07:24:46] <abaumann> KitsuWhooa: there is librsvg-og for circumventing rustification, and old thin-provisioning-tools-og.
[07:25:02] <KitsuWhooa> I would rather not run those on my system, honestly
[07:25:03] <abaumann> but the trend to an inmature toochain like rust will kill our efforts.
[07:25:34] <abaumann> oh. sorry. my history didn't catch up. I'm in the middle of a discussion now :-)
[07:26:16] <abaumann> Qt6 never compiled for me on 32-bit
[07:26:41] <abaumann> And after the n-th time trying to figure out how to bootstrap Python to get working python packages I also gave up.
[07:26:47] <KitsuWhooa> I did get Qt6 to compile
[07:26:53] <abaumann> Ah, that's good news :-)
[07:27:32] <KitsuWhooa> https://archlinux32.org
[07:27:32] <abaumann> I have to be frank: the goal in Archlinux32 was since a year or so just to keep some machines alive by adding binary shim packages.
[07:27:33] <phrik> Title: Arch Linux 32 - qt6-base 6.9.1-2.0 (pentium4) (at archlinux32.org)
[07:27:44] <abaumann> The effort for the rolling release just got too big.
[07:27:47] <KitsuWhooa> it still compiles apparently
[07:27:51] <KitsuWhooa> yeah
[07:27:58] <KitsuWhooa> that's why I'm thinking of jumping ship to alpine
[07:28:16] <KitsuWhooa> it is really unfortunate
[07:28:24] <abaumann> That's an idea. I started to look into t2sde and I started a i486tcclinux.
[07:28:45] <KitsuWhooa> once upon a time I was going to run arch32 on my router, but I had 0 time so after more than a year of "I'll get to it" I went crawling back to openwrt
[07:29:18] <abaumann> That's where I still keep arch32 alive. I also keep an old ARMv6 backup system alive, a NAS, etc.
[07:29:43] <abaumann> But there is a big difference between a working rolling release and a patched system just working fine for the job at hand.
[07:30:03] <KitsuWhooa> yeeeeah
[07:30:24] <abaumann> The BSDs are still full i686-ish, so that's also an option.
[07:30:31] <abaumann> OpenBSD as a firewall appliance is just nice. :-)
[07:30:43] <KitsuWhooa> I am not a fan of the BSDs personally
[07:31:12] <abaumann> For a desktop you will always miss just that one application you miss, and the you have to run in in Linux emulation mode.
[07:31:24] <KitsuWhooa> I just don't see the point personally
[07:31:38] <KitsuWhooa> plus I don't like the license
[07:32:06] <abaumann> They are much nicer to port, because they don't have things like coreutils in c and new coreutils in rust. The userspace is part of the project.
[07:32:28] <abaumann> Linux always suffered from the "we do just kernel", "we just do a new version of userland tool X"
[07:32:43] <KitsuWhooa> that's true, but I'm not sure how much of an issue it actually is
[07:33:03] <abaumann> urgh.. my PGP keys have a SHA-1 error, nice. and limited trust because of the 3-people sign rule.
[07:33:23] <KitsuWhooa> :)
[07:33:30] <abaumann> There is ONE firewall on BSDs, not iftables, ebtables, berkley filters, etc.
[07:33:59] <abaumann> man pages are much more usable on BSDs
[07:34:12] * KitsuWhooa shrugs
[07:34:14] <abaumann> configuration parsing is standardized (at least on OpenBSD)
[07:34:26] <KitsuWhooa> you say "miss that one app" but in reality I just miss GNU
[07:34:26] <abaumann> BSDs are more like Unix was. :-)
[07:34:54] <abaumann> Well, you can use the GNU ports of coreutils and install them in /usr/local or so.
[07:35:20] <KitsuWhooa> yeah, that same argument goes for actually using alpine
[07:35:28] <KitsuWhooa> "you can use full blown utilities" instead of busybox
[07:35:37] <KitsuWhooa> but by the time you install everything, you're better off not using it
[07:35:39] <KitsuWhooa> :p
[07:36:07] <abaumann> isn't alpine musl-based?
[07:36:09] <KitsuWhooa> yeah
[07:36:15] <KitsuWhooa> which I suspect will also make life easier
[07:36:32] <KitsuWhooa> openwrt also uses musl
[07:36:47] <abaumann> Don't get me wrong, I like musl, I use it for i486tcclinux too, mainly because it's easy to be conviced to compile with something not gcc/clang, but tcc.
[07:37:12] <KitsuWhooa> the issue with musl is that it's not glibc :p
[07:37:25] <abaumann> which in most ways is a good thing :->
[07:37:35] <KitsuWhooa> until you try to run a blob
[07:38:14] <abaumann> does snap, appimage, etc. count as blob? do they bundle the glibc, I doubt.
[07:38:22] <KitsuWhooa> I do believe they bundle glibc :p
[07:38:28] <KitsuWhooa> well, flatpak
[07:38:32] <KitsuWhooa> I assume snap does the same
[07:38:33] <abaumann> I know go can also do it.
[07:38:38] <KitsuWhooa> I don't think appimage does
[07:38:39] <abaumann> 'go=the language go'
[07:38:46] <KitsuWhooa> I hate go
[07:38:49] <KitsuWhooa> not because of the language
[07:38:55] <KitsuWhooa> but because no one ever packages go applications
[07:39:02] <KitsuWhooa> so I have to manually update every single one of them
[07:39:14] <KitsuWhooa> "just download the binary and run it, it's so easy to use"
[07:39:22] <KitsuWhooa> and then you end up with a 2 year old out of date web service that gets pwned
[07:39:38] <abaumann> it was designed for easy binary shippment for micro-services. so.
[07:39:47] <KitsuWhooa> yes but something needs to keep that up to date...
[07:40:08] <abaumann> That's a problem with all softwware. Software you don't share that is.
[07:40:09] <KitsuWhooa> I guess "just throw out the docker container" is a solution
[07:40:12] <abaumann> It is a tradeoff.
[07:40:42] <KitsuWhooa> you can still make a deb out of a single go binary :p
[07:40:48] <KitsuWhooa> and shove it in an apt repo
[07:41:13] <abaumann> That's what arch packges are doing, it just increased build time, as for rust stuff.
[07:41:23] <KitsuWhooa> rustc-bin my beloved
[07:41:29] <abaumann> The time of the shared library is over.
[07:41:36] <KitsuWhooa> yes
[07:41:40] <abaumann> I'm not a fan of the situation, but, oh well.
[07:42:33] <KitsuWhooa> at least you are not a Debian/Ubuntu maintainer :)
[07:42:55] <abaumann> mmh. don't remember. I might have done some packages in the past.
[07:42:55] <KitsuWhooa> they have to keep bumping rustc on ancient LTSes
[07:43:04] <abaumann> oh.
[07:43:50] <abaumann> In theory you build the first rust with mrust and then you just keep all packages like rustXYY around and make it depend on rustXYY-1 or so
[07:44:13] <abaumann> But it's a maintainance mess.
[07:44:16] <KitsuWhooa> :)
[07:47:46] <abaumann> I did a small test with some C programs running through as many C compilers I could find working on Linux (gcc, clang, tcc, pcc, cproc, cparser, ack, nwcc, smlrcc, gcc-4.7, 8cc, chibicc, lacc), also then you find out what exacly is implemented where. And you find interesting corner cases. :-)
[07:48:15] <abaumann> But there are standards, and those don't change on every release.
[07:49:14] <abaumann> mmh. have to find I way to get those PGP keys signed..
[07:50:15] <buildmaster> Hi abaumann!
[07:50:16] <buildmaster> !rq abaumann
[07:50:17] <phrik> buildmaster: <abaumann> deep42thought: you cannot break cmake, it's already broken.. ;-)
[07:50:46] <abaumann> oh. high "buildmaster" :-)
[07:50:51] <abaumann> *hi
[07:51:10] <KitsuWhooa> lol that was delayed
[07:51:35] <KitsuWhooa> How long has arch32 been around for, anyway?
[07:51:43] <KitsuWhooa> guess one could say it's had a good run
[07:51:43] <abaumann> Since 2017
[07:51:59] <abaumann> yeah, would say, longer than expected.
[07:52:24] <KitsuWhooa> not bad
[07:53:09] <KitsuWhooa> I remember switching from Arch to Arch32
[07:53:15] <KitsuWhooa> it really has been that long
[07:53:36] <abaumann> time flies in a pandemic ;-)
[07:53:55] <KitsuWhooa> s/ in.*//
[07:53:58] <KitsuWhooa> :p
[07:56:46] <abaumann> [GNUPG:] NEWSIG
[07:56:47] <abaumann> [GNUPG:] KEYEXPIRED 1763112304
[07:56:47] <abaumann> [GNUPG:] KEYEXPIRED 1763112304
[07:56:47] <abaumann> [GNUPG:] KEYEXPIRED 1763112304
[07:56:47] <abaumann> [GNUPG:] KEYEXPIRED 1763112304
[07:56:49] <abaumann> [GNUPG:] KEY_CONSIDERED 16194A82231E9EF823562181C8E8F5A0AF9BA7E7 0
[07:56:51] <abaumann> [GNUPG:] KEYEXPIRED 1763112304
[07:56:54] <abaumann> [GNUPG:] SIG_ID Jidi9kKoW79tuwelMpQxgF0DUyE 2025-11-29 1764402967
[07:56:56] <abaumann> [GNUPG:] EXPKEYSIG C8E8F5A0AF9BA7E7 Andreas Baumann (sign) <mail@andreasbaumann.cc>
[07:56:59] <abaumann> [GNUPG:] VALIDSIG 16194A82231E9EF823562181C8E8F5A0AF9BA7E7 2025-11-29 1764402967 0 4 0 1 10 00 16194A82231E9EF823562181C8E8F5A0AF9BA7E7
[07:57:02] <abaumann> 'return-assignment' reports a signature error.
[07:57:05] <abaumann> oh sweet
[07:57:07] <KitsuWhooa> rip
[07:57:24] <KitsuWhooa> want me to fire up a builder?
[07:57:45] <abaumann> worth a try, your key still looks alive and kicking.
[07:57:50] <KitsuWhooa> I set it to never expire :)
[07:57:50] <abaumann> though my master key expired.
[07:57:53] <KitsuWhooa> oh
[07:58:00] <KitsuWhooa> okay maybe it won't work
[07:58:05] <abaumann> well, worth a try
[07:58:47] <KitsuWhooa> > ==> Making package: newsboat-og 2.13.2-3.0 (Sat Nov 29 07:58:41 2025)
[07:59:02] <KitsuWhooa> you'll need to do some magic to force the package on my builder
[07:59:14] <abaumann> yes, hang on, what's the name of the builder?
[07:59:35] <abaumann> build-tasossah-c2d
[07:59:37] <KitsuWhooa> yup
[08:01:51] <abaumann> ok, forced
[08:02:11] <KitsuWhooa> 2025-11-29 08:02:08: building package "newsboat-og" (revisions 0000000000000000000000000000000000000000 4293a7084fab9d8f60ef0cd27f1576a476790a8c, repository extra, straw :mirrored_source_by_hash:) for pentium4 ...
[08:02:15] <KitsuWhooa> guess we have to wait for this to finish
[08:02:20] <abaumann> yes.
[08:02:36] <abaumann> I'll fiddle with my signing key in the meantime
[08:03:13] <KitsuWhooa> This is a low end Core 2 Duo machine connected to the internet via ADSL over WiFi
[08:03:19] <KitsuWhooa> so don't expect it to be fast :p
[08:03:32] <abaumann> newsboat-og is C++, not rust :->
[08:03:43] <abaumann> so it will be reasonably fast - if it builds..
[08:03:51] <KitsuWhooa> the storage is a RAID0 of 3 slow laptop HDDs
[08:03:54] <KitsuWhooa> oh it just finished
[08:04:01] <abaumann> see :-)
[08:04:04] <KitsuWhooa> get-assignment told me:
[08:04:06] <KitsuWhooa> 3: come back after the next run of get-package-updates or when
[08:04:08] <KitsuWhooa> some currently building packages are returned - currently
[08:04:14] <KitsuWhooa> there are no pending buildable packages
[08:04:20] <abaumann> *sigh*
[08:04:26] <KitsuWhooa> I, too, sighed
[08:04:49] <abaumann> my job is stuck in post-build..
[08:05:54] <abaumann> uid [ultimate] Andreas Baumann (sign) <mail@andreasbaumann.cc>
[08:05:54] <abaumann> sub rsa2048 2017-11-02 [E] [expires: 2026-11-27]
[08:06:00] <abaumann> on the master user in the pgp keyring.
[08:06:08] <abaumann> let's hope, the buildmaster takes that one.
[08:06:42] <KitsuWhooa> mine still says the same
[08:06:47] <abaumann> ah. the buildmaster might do some package updates, that might run a while, it didn't for some time.
[08:07:24] <KitsuWhooa> I'll leave it running
[08:08:16] <KitsuWhooa> 2025-11-29 08:07:57: building package "jupyterlab-rise" (revisions 7b38b6759bdf69bf13c7f175c25a176d6c9cebe3 0000000000000000000000000000000000000000, repository extra, straw :mirrored_source_by_hash:) for pentium4 ... failed.
[08:08:18] <KitsuWhooa> nope
[08:09:21] <abaumann> mmh
[08:09:44] <abaumann> I used the ~/force script in the master HOME on the buildmaster
[08:10:11] <abaumann> again.
[08:10:25] <abaumann> maybe the build list is still the same..
[08:11:39] <KitsuWhooa> come back after the next run of get-package-updates - currently there are no pending packages
[08:11:59] <abaumann> yeah, let's give the buildmaster some time to do the package update.
[08:12:08] <KitsuWhooa> archlinux32-keyring is still assigned to euronuc
[08:12:12] <abaumann> ah.
[08:12:24] <KitsuWhooa> can you let it run so that it fails?
[08:12:33] <KitsuWhooa> or does it not move on
[08:12:35] <abaumann> maybe because it fails due to failed signatures it gets resacheduled to the same build slave over and over again?
[08:12:38] <abaumann> could be.
[08:12:47] <KitsuWhooa> ugh
[08:12:55] <KitsuWhooa> maybe you can edit the database
[08:12:56] <abaumann> it's a table somewhere..
[08:12:57] <KitsuWhooa> I forgot how to do it
[08:12:58] <abaumann> in mysql
[08:12:59] <KitsuWhooa> hah
[08:13:01] <abaumann> me too
[08:13:35] <KitsuWhooa> I saw streusel-{1..3} in the build list and now I want something sweet
[08:13:35] <abaumann> HOWTO.RemoveBuildJobForSlave
[08:13:37] <abaumann> aha
[08:13:43] <KitsuWhooa> nice
[08:13:52] <KitsuWhooa> *builder list
[08:16:03] -!- morriset has quit [Remote host closed the connection]
[08:17:19] -!- morriset has joined #archlinux32
[08:18:12] <abaumann> aha. I see something building.
[08:18:16] <KitsuWhooa> 2025-11-29 08:18:06: building package "archlinux32-keyring" (revisions 0000000000000000000000000000000000000000 11d90d4350108d3d3a8bd4db32c0a6e808333a3b, repository core, straw :mirrored_source_by_hash:) for pentium4 ...
[08:18:35] <KitsuWhooa> ...it failed to build
[08:18:37] <KitsuWhooa> what
[08:19:01] <KitsuWhooa> maybe nuking the chroot will fix it
[08:23:43] <KitsuWhooa> nope, it failed to build
[08:23:45] <KitsuWhooa> ugh
[08:24:11] <abaumann> ok, never mind. I'm close to fixing my key situation on the buildmaster :-)
[08:28:36] <abaumann> gpg: Note: third-party key signatures using the SHA1 algorithm are rejected
[08:28:36] <abaumann> gpg: (use option "--allow-weak-key-signatures" to override)
[08:28:39] <abaumann> hrmph
[08:29:06] <abaumann> it takes the keys from pacman as root on the buildmaster, not the local gpg keyring of the master user.
[08:29:12] <abaumann> now I was able to build :-)
[08:29:19] <abaumann> with weak signatures though.
[08:29:44] <abaumann> I would have expected that gpg gives me an option to change to safer hash algorithms, if it is deprecating old ones.
[08:29:53] <abaumann> I don't want to make new keys..
[08:30:17] <abaumann> ok, I can build a keyring, now, where does the keyring take the keys for the keyring from..
[08:30:56] <abaumann> https://sources.archlinux32.org
[08:30:59] <abaumann> mmh. ok.
[08:31:21] <abaumann> -rw-r--r-- 1 sources sources 39593 Jun 30 21:08 archlinux32-keyring-v20250630.tar.gz
[08:31:35] <abaumann> *investigations ongoing
[08:32:01] <abaumann> the package is in git and has a nice Makefile IIRC
[08:32:32] <abaumann> update-keys looks promising
[08:32:35] <abaumann> key is not fully trusted: 5FDCA472AB93292BC678FD59255A76DB9A12601A deep42thought
[08:32:38] <abaumann> key is not fully trusted: 80EC18799E8BCD375C6E64ABE4D41569196B1160 KitsuWhooa
[08:32:43] <abaumann> hrmph
[08:33:16] <abaumann> make dist
[08:33:18] <abaumann> gpg: skipped "release@archlinux32.org": Unusable secret key
[08:33:18] <abaumann> gpg: signing failed: Unusable secret key
[08:33:28] <abaumann> oha. the master keys of all keys..
[08:33:45] <abaumann> sec rsa4096 2020-01-20 [SC] [expired: 2025-02-04] 33CA3597B0D161AAE4173F65C17F1214114574A4
[08:33:48] <abaumann> uid [ expired] Archlinux 32 Release Key <release@archlinux32.org>
[08:34:38] <abaumann> ok, I have to contact deep42thought for that.
[08:35:35] <abaumann> 5773 Feb 2 2023 release-archlinux32.key
[08:35:47] <abaumann> that is the last one I have
[09:03:08] <socksinspace> re: glibc binaries on alpine: there is gcompat, which works for some things, but as far as i can tell the common advice is: just use flatpak
[09:08:56] <abaumann> ah. thanks. interesting.
[09:21:50] -!- abaumann has quit [Quit: leaving]
[11:36:49] <girls> abaumann: if you have the secret key, you can just extend the release key - but I sent you an updated key anyways. Would be nice, if we could get keys.php working again, though.
[11:57:04] -!- h|weechat has quit [Ping timeout: 260 seconds]
[12:07:40] -!- abaumann has joined #archlinux32
[12:07:40] <buildmaster> Hi abaumann!
[12:07:40] <buildmaster> !rq abaumann
[12:07:41] <phrik> buildmaster: <abaumann> ipv4 addresses are not that many..easy to remember.. ;-)
[12:13:56] <abaumann> hi girls.
[12:14:14] <abaumann> sorry, I nuked the .gnupg folder in /srv/http while migrating www.archlinux32.org
[12:14:26] <abaumann> and I also added an Anubis challenge to it, which didn't help :-)
[12:28:56] <abaumann> I uploaded your release key and update archlinux32-keyring (also with my master/sign keys). Trying to build it and see what I get..
[12:39:22] -!- abaumann has quit [Quit: leaving]