#archlinux32 | Logs for 2019-03-08

[00:05:37] -!- AndrevS has quit [Read error: Connection reset by peer]
[00:12:14] <buildmaster> i686/libreoffice-fresh is broken (says rechenknecht).
[00:19:16] -!- yans has quit [Quit: chaos is the only true answer]
[01:07:27] <buildmaster> i686/jupyter-notebook is broken (says rechenknecht).
[01:29:23] <buildmaster> i686/haskell-keys are broken (says rechenknecht) - I rescheduled: haskell-free.
[01:39:13] <buildmaster> i686/haskell-lens-aeson is broken (says nlopc46) - I rescheduled: haskell-adjunctions, haskell-invariant, haskell-kan-extensions, haskell-lens.
[01:41:13] <buildmaster> i686/haskell-pointed is broken (says rechenknecht).
[01:43:13] -!- buildmaster has quit [Remote host closed the connection]
[01:43:25] -!- buildmaster has joined #archlinux32
[01:49:19] <buildmaster> i686/haskell-cryptohash-conduit is broken (says rechenknecht) - I rescheduled: haskell-conduit-extra.
[01:52:42] <buildmaster> i686/haskell-kan-extensions are broken (says nlopc46) - I rescheduled: haskell-invariant.
[01:57:33] -!- jason0597 has quit [Ping timeout: 245 seconds]
[02:03:47] <buildmaster> i686/haskell-deferred-folds are broken (says nlopc46) - I rescheduled: haskell-rebase.
[02:04:47] <buildmaster> i686/haskell-persistent-sqlite is broken (says buildknecht2).
[02:15:38] <buildmaster> i686/hedgewars are broken (says nlopc46).
[02:21:43] <buildmaster> i686/haskell-html-conduit is broken (says rechenknecht) - I rescheduled: haskell-xml-conduit.
[02:37:34] -!- Dimtree has joined #archlinux32
[02:57:22] <buildmaster> i686/haskell-warp is broken (says nlopc46) - I rescheduled: haskell-http2.
[03:01:38] <buildmaster> i686/haskell-ipynb is broken (says rechenknecht) - I rescheduled: haskell-microlens-aeson.
[03:10:56] <buildmaster> i686/haskell-persistent-qq is broken (says nlopc46) - I rescheduled: haskell-monad-logger, haskell-persistent.
[03:10:59] <buildmaster> i686/haskell-esqueleto is broken (says rechenknecht).
[03:21:39] <buildmaster> i686/hledger-ui is broken (says buildknecht2).
[03:28:43] <buildmaster> i686/idris are broken (says nlopc46).
[03:34:51] <buildmaster> i686/haskell-hpack is broken (says rechenknecht) - I rescheduled: haskell-http-client-tls, haskell-interpolate.
[03:53:37] <buildmaster> i686/reflector is broken (says nlopc46).
[07:11:49] -!- ofara_ has quit [Ping timeout: 268 seconds]
[07:19:09] <KitsuWhooa> deep42thought: for when you're around, I assume librsvg has been built, however it still crashes with SIGILL :(
[07:19:13] <KitsuWhooa> I hope i installed the right package, anyway
[07:19:38] <KitsuWhooa> either way, thought I'd mention that in the future you'll probably need to edit the regex again https://gitlab.gnome.org
[07:19:39] <phrik> Title: Makefile.am · master · GNOME / librsvg · GitLab (at gitlab.gnome.org)
[07:47:34] -!- ichernev has joined #archlinux32
[07:50:47] <ichernev> I booted arch32 (from usb made with dd) on a amd x3216 (32bit) cpu with 8gb ram, but free only shows 3.5gb ram. CPU supports PAE, am I missing something?
[07:58:16] -!- abaumann has joined #archlinux32
[07:58:16] <buildmaster> Hi abaumann!
[07:58:16] <buildmaster> !rq abaumann
[07:58:17] <phrik> buildmaster: <abaumann> who would have thought digital books burn too. ;-)
[07:59:15] <abaumann> KitsuWhooa: the librsvg package might be stuck in 'testing' because of dependencies..
[08:00:00] <KitsuWhooa> abaumann: I manually downloaded the package from testing and installed it with pacman -U
[08:00:07] <abaumann> and still SSE2 in it?
[08:00:13] <KitsuWhooa> yup
[08:00:28] <abaumann> ok. maybe rust changed the 'pentium3' target to include SSE2?
[08:01:04] <ichernev> I think I figured it out -- after I install it, I'd need to use linux-pae kernel, which has pae enabled
[08:01:46] <abaumann> huh?
[08:02:11] <abaumann> ah.
[08:02:18] <abaumann> second thread of communication, sorry. :-)
[08:02:29] <KitsuWhooa> abaumann: not sure. I hope not
[08:02:34] <abaumann> yes, pae is your friend here.
[08:02:40] <KitsuWhooa> I do wonder if the patch is being applied properly though
[08:02:46] <abaumann> note, that a user process can still not use more than 4 GB address space..
[08:03:01] <abaumann> ..but now you can split your 8 GB to two user processes ith 4 GB address space each..
[08:03:06] -!- ofara has joined #archlinux32
[08:03:22] <abaumann> KitsuWhooa: that's what I'm checking right now..
[08:03:33] <KitsuWhooa> How do the pkgbuilds get modified from upstream?
[08:03:40] <KitsuWhooa> It doesn't look like a diff
[08:04:00] <abaumann> it's a snippet of shell script which gets appended to the upstream PKGBUILD
[08:04:17] <KitsuWhooa> Ah, okay
[08:04:40] <abaumann> usually some variables or functions are then patched using seds.
[08:05:27] <KitsuWhooa> as a sidenote, I'm not sure if it's known
[08:05:30] <KitsuWhooa> rustc also uses SSE2
[08:05:42] <ichernev> abaumann, can the kernel use more than 4g for caches and stuff? This machine will be used for a NAS, mainly
[08:06:04] <abaumann> not sure about that, but I heavily doubt it.
[08:06:21] <abaumann> the kernel address space is hardly bigger than 4 GB..
[08:07:38] <abaumann> for a NAS you have nfsd and smb user processes, they hardly benefit from more memory.
[08:08:38] <ichernev> what about disk caches -- that will only use 4g, correct?
[08:09:59] <abaumann> I'm not sure whether PAE is used for creating user address spaces only or if PAE is also used to allocate pages for caching..
[08:10:26] <KitsuWhooa> You can probably test it though
[08:10:38] <abaumann> also check /proc/cpuinfo: address sizes
[08:10:39] <abaumann> address sizes : 32 bits physical, 32 bits virtual
[08:10:51] <abaumann> for instance doesn't help, it should say >32 bits physical there
[08:14:15] <ichernev> I haven't run linux-pae kernel yet
[08:15:03] <ichernev> 48 bits physical, 48 bits virtual
[08:15:18] <abaumann> and this is a 32-bit cpu?
[08:15:50] <ichernev> amd opteron x3216 ... 64 bit ubuntu wouldn't even boot, so after a bit of reading I figured its 32bit ... and 32bit arch booted
[08:16:26] <ichernev> can I tell from the /proc/cpuinfo flags?
[08:16:41] <abaumann> not really, but check flags. do you see a 'pae' there?
[08:16:51] <ichernev> yes, pae is there
[08:17:57] -!- deep42thought has joined #archlinux32
[08:17:57] <buildmaster> Hi deep42thought!
[08:17:57] <buildmaster> !rq deep42thought
[08:17:58] <phrik> buildmaster: <deep42thought> so you're a garbage collector?
[08:18:14] <deep42thought> good morning, everyone
[08:18:17] <abaumann> hi deep42thought
[08:18:25] <ichernev> according to wikipedia "most linux kernels ship with pae enabled", but if there is linux-pae, then arch32 kernel is not pae enabled, correct?
[08:18:57] <deep42thought> iirc, it fails to boot on ancient hardware, so we disabled it
[08:20:26] <ichernev> deep42thought: that makes sense. I still can't understand why a brand new machine will ship with 32bit cpu ...
[08:21:05] <deep42thought> yeah, this sounds ... interesting
[08:21:18] <abaumann> it's amazing how bad the product information for this CPU is..
[08:21:31] <abaumann> but this should be a 64-bit CPU
[08:21:39] <deep42thought> (we were looking for 32bit machines lately for testing purposes and found no new ones)
[08:21:55] <ichernev> abaumann: where did you read that
[08:22:06] <abaumann> just guessing.
[08:22:07] <abaumann> :-)
[08:22:23] <abaumann> I thought nobody is shipping new machines anymore with 32-bit cpus
[08:22:40] <abaumann> ok. HP could be the exception..
[08:22:46] <deep42thought> LOL
[08:23:22] <abaumann> https://www.virten.net
[08:23:23] <phrik> Title: Homeserver – ESXi on HPE ProLiant MicroServer Gen10 | Virten.net (at www.virten.net)
[08:24:07] <ichernev> abaumann: exactly, that is what I bought. From ebay, for $360, brand new
[08:24:41] <abaumann> getconf LONG_BIT
[08:24:59] <abaumann> this should tell whether it's 32 or 64 bits
[08:25:22] <ichernev> 32
[08:25:29] <abaumann> interesting.
[08:25:47] * deep42thought sees abaumann buying this very same box just for testing purposes
[08:25:52] <abaumann> ok. so HP shipped 32-bit CPUs in 2017.. that seems to be a world record.. :-)
[08:25:53] <ichernev> I'd be happy to donate one of these for testing packages
[08:25:57] <KitsuWhooa> wouldn't that need a 64 bit kernel?
[08:25:59] <abaumann> * abaumann: right :-)
[08:26:18] <ichernev> what "needs" 64 bit kernel?
[08:26:30] -!- jason0597 has joined #archlinux32
[08:27:00] <abaumann> technically, nothing 'needs' 64-bit. there could be some problems with EFI-only machines booting only in 64-bit mode though..
[08:27:04] <ichernev> what is more, the machine supports 32 gb of ram, and comes preinstalled with 8gb
[08:27:14] <abaumann> and is a 32-bit cpu.
[08:27:22] <abaumann> as I said.. HP.. :->
[08:27:28] <ichernev> this I just found out :)
[08:28:22] <abaumann> deep42thought: /usr/bin/meinproc5: error while loading shared libraries: libdouble-conversion.so.3.0.0: cannot open shar
[08:28:38] <KitsuWhooa> ichernev: when you say the 64 bit kernel wouldn't boot, what was the error?
[08:28:51] <ichernev> the ubuntu animation just hung
[08:28:55] <abaumann> libdouble-conversion was somehow published before kdoctools had a chance to get rebuilt against libdouble-conversion.so.3.1.2
[08:29:13] <KitsuWhooa> you wouldn't get that far if 64 bit wasn't supported
[08:29:16] <deep42thought> abaumann: ah, yes, I've seen this
[08:29:16] <ichernev> KitsuWhooa: is is possible to run 64 bit kernel on 32 bit cpu??
[08:29:26] <deep42thought> ichernev: no
[08:29:51] <KitsuWhooa> in order for plymouth to start, so does the kernel
[08:30:00] <KitsuWhooa> and I'm sure a 64 bit kernel wouldn't start if the CPU didn't support it
[08:30:13] <abaumann> good point: plymouth is a user process..
[08:30:17] <ichernev> maybe they splashed the first frame somehow and then it hung
[08:30:19] <KitsuWhooa> so chances are the hang was a different issue
[08:30:21] <deep42thought> yes, it will fail pretty early (in the bootloader)
[08:30:25] <abaumann> ..it would hardly get started if the kernel was not up.
[08:30:33] <ichernev> I guess I can start with plymouth disabled
[08:30:41] <ichernev> via grub
[08:30:42] <KitsuWhooa> just remove the silent splash args from the cmdline
[08:30:43] <abaumann> you can try to boot in verbose mode and see the kernel messages instead of the nice splash screen
[08:30:51] <KitsuWhooa> er
[08:30:53] <KitsuWhooa> quiet
[08:31:01] <ichernev> let me try
[08:32:11] <KitsuWhooa> Also, I don't think LONG_BIT indicates what the cpu supports
[08:32:40] <abaumann> could be it reports only what the kernel tells it to report.
[08:32:58] <KitsuWhooa> Also, I really doubt there are x86 only CPUs with DDR4 out there
[08:33:40] <abaumann> there were cpus, which are technically 64-bit once, but the instruction set was crippled to support only 32-bit opcodes (IIRC)
[08:33:47] <abaumann> but this was Intel..
[08:34:50] <KitsuWhooa> fwiw, my FX8350 also reports address sizes : 48 bits physical, 48 bits virtual
[08:35:28] <abaumann> just tried rsvg-convert twitter.svg on testing with no sse2
[08:35:32] <abaumann> seems to work
[08:35:40] <KitsuWhooa> hm
[08:35:55] <abaumann> could also be some qt library using sse2
[08:36:08] <KitsuWhooa> let me run gdb
[08:36:13] <abaumann> good idea.
[08:36:22] <abaumann> librsvg 2:2.44.13-1.3
[08:36:27] <abaumann> from testing
[08:37:14] <KitsuWhooa> it'd help if I had X running :p
[08:37:40] <abaumann> what window manager are you using along with X
[08:37:46] <KitsuWhooa> awesome
[08:38:00] <abaumann> ah , ok. that one is light on libraries :-)
[08:38:07] <KitsuWhooa> the desktop starts up fine
[08:38:16] <KitsuWhooa> it's certain applications like audacious and other gtk ones that fail
[08:38:26] <KitsuWhooa> Thread 1 "audacious" received signal SIGILL, Illegal instruction.
[08:38:27] <KitsuWhooa> 0xb3b893eb in ?? () from /usr/lib/librsvg-2.so.2
[08:38:49] <abaumann> ah. ok. let me try on my test machine..
[08:38:51] <KitsuWhooa> (gdb) x/i $pc
[08:38:53] <KitsuWhooa> => 0xb3b893eb: movsd 0x10(%esi),%xmm0
[08:38:58] <abaumann> yep.
[08:38:59] <KitsuWhooa> let me also make sure I have the correct version installed
[08:39:08] <ichernev> ok, now in legacy boot, ubuntu booted without splash, even displayed the cursor, but I can't enter tty or gui...
[08:39:35] <KitsuWhooa> ichernev: yeah your cpu definitely supports 64 bit in that case
[08:40:02] <KitsuWhooa> wild guess, try booting with nomodeset
[08:41:00] <abaumann> if you're unlucky, your graphic card is a AMD embedded one and X doesn't support it anymore, so no GUI (but how needs an installer with graphics anyway to install a NAS) :->
[08:41:29] <KitsuWhooa> I'd suspect it's a bug
[08:41:30] -!- oaken-source has joined #archlinux32
[08:41:34] <KitsuWhooa> nomodeset would cause it to not load the amd driver, so you'd at least get a desktop
[08:41:54] <abaumann> I hope, Ubuntu still has a text-only installation mode..
[08:41:58] <KitsuWhooa> even installing and then updating the system might fix it
[08:42:25] <KitsuWhooa> or try a different version, or a different distro
[08:43:03] <KitsuWhooa> abaumann: tatokis@nightopia:~$ pacman -Q librsvg
[08:43:04] <KitsuWhooa> librsvg 2:2.44.13-1.3
[08:44:08] <abaumann> yep. the same. what I don't get: I have a virtual machine running without SSE2 enabled, and everything works.
[08:44:20] <abaumann> I fear, this is again sloppy emulation in qemu/kvm..
[08:44:24] <KitsuWhooa> it's possible that sse is just not advertised
[08:44:34] <KitsuWhooa> wouldn't kvm let it run if the host cpu supported it?
[08:44:42] <KitsuWhooa> *sse2
[08:44:57] <abaumann> there are flags, but they don't have an effect for some opcodes..
[08:45:01] <KitsuWhooa> ah
[08:45:09] <abaumann> ..only if you truly emulate the CPU, then it fails.
[08:45:11] <ichernev> KitsuWhooa: it booted, it sees 7.5gb ram now. getconf LONG_BIT reports 64
[08:45:15] <abaumann> but that's unbearably slow..
[08:45:20] <KitsuWhooa> abaumann: I can imagine
[08:45:24] <KitsuWhooa> ichernev: nice
[08:45:39] <KitsuWhooa> I assume the .5 is reserved for graphics
[08:45:57] <abaumann> or BIOS, or I/O ports for cards or other stuff..
[08:46:14] <KitsuWhooa> yeah
[08:46:48] <abaumann> *abaumann starts a really old machine with a Pentium 3 without SSE2
[08:46:50] <KitsuWhooa> abaumann: do the build VMs also run with the same qemu flags?
[08:46:53] <ichernev> btw, on 2 machines now, arch from dd usb won't boot in uefi mode, ubuntu boots fine
[08:47:08] <KitsuWhooa> it would explain how those VMs can compile librsvg in the first place
[08:47:16] <abaumann> no, they usually run 64-bitish and build in a 32-bit chroot..
[08:47:20] <KitsuWhooa> ah
[08:47:54] <KitsuWhooa> Hm. Wouldn't that mean virtualbox VMs would also work?
[08:47:58] <abaumann> which makes them vulnerable to setarch mismatches and some other quirks people come up when probing for the cpu in /proc/cpuinfo directly. :->
[08:48:15] <deep42thought> we could replace all SSE2 instructions by ones that fail even in the vms
[08:48:17] <abaumann> I never saw virtualbox supporting fiddling with instruction sets..
[08:48:55] <KitsuWhooa> Me neither, but if the build VMs are 64 bit, I could possibly spin up a vbox one
[08:49:20] <deep42thought> KitsuWhooa: there's always need for more build vms :-)
[08:49:30] <abaumann> rsvg-convert twitter.svg > twitter.png
[08:49:30] <abaumann> Illegal instruction (core dumped)
[08:49:37] <abaumann> aha. on the real machine it SIGILLs
[08:49:44] <abaumann> I start to hate virtualization..
[08:49:45] <KitsuWhooa> yup, just like mine
[08:50:16] <abaumann> so, the idea of testing releases of Archlinux32 on a virtualized system is just not feasible..
[08:50:53] <KitsuWhooa> is it even possible to get rustc to work without sse2?
[08:51:09] <abaumann> it should respect the target-cpu flags.
[08:51:23] <abaumann> it had hard times to support i486, but everything else should work.. :-)
[08:51:30] <KitsuWhooa> ...well
[08:51:47] <KitsuWhooa> tatokis@nightopia:~$ gdb rustc
[08:51:49] <KitsuWhooa> Program received signal SIGILL, Illegal instruction.
[08:51:51] <KitsuWhooa> :p
[08:52:22] <abaumann> librsvg-2:2.44.13-1.3, the same. ok confirmed.
[08:52:38] <abaumann> the pentium3 flag obviously doesn't disable SSE2 anymore..
[08:52:48] <KitsuWhooa> that would explain it, yeah
[08:53:06] <abaumann> *abaumann refrains to start his usual rant about the librsvg/rust disaster..
[08:56:45] <KitsuWhooa> maybe we can try pentium2? :p
[08:57:03] * KitsuWhooa digs for a changelog that would explain this
[08:57:06] <abaumann> I'm reading throut target-cpu and target-feature now in rust..
[08:59:04] <abaumann> rustc -C target-cpu=help
[08:59:04] <abaumann> Target CPU help is not supported by this LLVM version.
[08:59:11] <abaumann> sometimes I get a headache..
[08:59:20] <KitsuWhooa> oh dear
[09:00:31] -!- ichernev has quit [Ping timeout: 256 seconds]
[09:04:19] <abaumann> there used to be a '-C target-features=-sse2'
[09:04:29] <abaumann> now it says error: unknown codegen option: `target-features`
[09:04:31] <abaumann> cool
[09:04:46] <abaumann> I like a running target as a programming language for a central parsing library for SVG
[09:05:13] <abaumann> nowhere there is an explanation what 'pentium3' actually enables exactly..
[09:06:41] <KitsuWhooa> <abaumann> now it says error: unknown codegen option: `target-features` <-- isn't it `target-feature` ?
[09:06:51] <abaumann> oh. :-)
[09:07:02] <abaumann> thanks :-)
[09:07:24] <KitsuWhooa> np
[09:07:38] <abaumann> I start to get allergic shocks just by having to deal with rust.. :-)
[09:07:46] <deep42thought> !grab abaumann
[09:07:47] <phrik> deep42thought: Tada!
[09:08:40] <KitsuWhooa> also, maybe try rustc --print target-cpus
[09:08:58] <abaumann> Target CPU help is not supported by this LLVM version.
[09:09:28] <KitsuWhooa> ...oops
[09:09:29] <KitsuWhooa> that looks like a bug :p
[09:10:08] <KitsuWhooa> it does work on my desktop
[09:10:37] <abaumann> I'm on testing with llvm 7.
[09:10:42] <abaumann> that might be the problem..
[09:10:58] <KitsuWhooa> my desktop runs llvm 8
[09:11:11] <abaumann> interesting..
[09:12:14] <abaumann> info: latest update on 2019-02-28, rust version 1.33.0 (2aa4c46cf 2019-02-28)
[09:12:17] <abaumann> info: downloading component 'rustc'
[09:12:27] <KitsuWhooa> ah
[09:12:28] <abaumann> jippie: a compiler downloading itself when compiling librsvg
[09:12:37] <abaumann> really cool..
[09:13:05] <abaumann> there comes the day I will rewrite librsvg using a decent language like C or C++..
[09:13:06] <deep42thought> you got no net connection? well, then you can't compile this hello world program.
[09:13:25] <abaumann> "always connected" is the anti-pattern here.
[09:13:30] <deep42thought> yeah
[09:13:42] <deep42thought> I know how it feels to be *not* always connected
[09:13:47] <deep42thought> :-(
[09:14:35] <abaumann> yeah, I sympathise with you..
[09:14:52] <KitsuWhooa> Hm, I see i386/i486/i586/i686 target processors in the rustc output
[09:15:14] * deep42thought has currently a downlink and uplink of 64GB/day, with a latency of 1 day per request :-/
[09:15:24] <KitsuWhooa> I know i686 definitely uses SSE2
[09:15:28] <abaumann> KitsuWhooa: yea. try run one of those.. :->
[09:15:41] <abaumann> ok. I try now to build with -C target-feature=-sse2
[09:15:48] <abaumann> let's see what happens :-)
[09:15:57] <KitsuWhooa> well, I can't really cross compile, and the compiler won't run on my laptop :p
[09:16:17] <abaumann> no worries.. :-)
[09:18:09] <KitsuWhooa> not sure how relevant this is, but https://github.com
[09:18:11] <phrik> Title: llvm/X86.td at master · llvm-mirror/llvm · GitHub (at github.com)
[09:18:42] <abaumann> this looks good and correct.
[09:18:52] <abaumann> but otoh: https://github.com
[09:18:54] <phrik> Title: SSE2 instructions in i686 binaries · Issue #1196 · rust-lang/rustup.rs · GitHub (at github.com)
[09:18:56] <KitsuWhooa> yeah, that's the confusing part
[09:18:59] <KitsuWhooa> hm
[09:19:13] <abaumann> so there were issues with SSE2 creeping into rust in the past..
[09:19:59] <KitsuWhooa> their prebuilt i686 binaries definitely need SSE2
[09:20:00] <KitsuWhooa> they sigill instantly
[09:20:32] <abaumann> BTW: if you know a nice tutorial on how to bootstrap rustc for a new architecture..
[09:20:42] <abaumann> ..I still want to do this with i486.
[09:20:57] <KitsuWhooa> I actually know nothing about rust
[09:21:15] <abaumann> But using mrust seemed to be the better approach than to go with the rustc/cargo craziness..
[09:21:40] <deep42thought> I gotta go, bbl
[09:21:43] <KitsuWhooa> see ya
[09:21:46] <abaumann> ok. cu
[09:21:47] -!- deep42thought has quit [Quit: Leaving.]
[09:22:01] <KitsuWhooa> sadly I don't think I can help, but it _should_ be possible
[09:23:09] <abaumann> np, you helped already by adressing the issue :-)
[09:23:21] <KitsuWhooa> I assume the simple i486-unknown-gnu while cross compiling doesn't work :p
[09:23:54] <KitsuWhooa> https://github.com hmm
[09:23:56] <phrik> Title: Bootstrap cargo crashes on i486 machine · Issue #54740 · rust-lang/rust · GitHub (at github.com)
[09:24:19] <KitsuWhooa> *i486-unknown-linux-gnu
[09:25:59] <abaumann> yeah. they will have a nice discussion with the Fedora people, they still require i586 as base for their builds
[09:26:30] <KitsuWhooa> at least they support some form of i686
[09:26:37] <KitsuWhooa> could've been way worse :p
[09:26:50] <abaumann> yes. that's right. they could go 64-bit only.
[09:27:13] <abaumann> we have to acknoledge, that using 32-bit cpus older than 10 years is a corner usage case. :-)
[09:27:33] <KitsuWhooa> yeah
[09:27:59] <KitsuWhooa> that said, a lot of schools here run old computers with linux
[09:28:11] <KitsuWhooa> like P3 old
[09:35:29] <KitsuWhooa> Unrelated. Anyone had any luck getting speedstep working? I have CONFIG_X86_SPEEDSTEP_SMI=y but nothing is happening
[09:36:41] <abaumann> I'm using p4-clockmod and cpupower..
[09:37:09] <abaumann> my eeepc steps nicely from 225 Mhz to 900 Mhz
[09:37:25] <KitsuWhooa> I'm not sure that will work on my cpu
[09:38:34] <KitsuWhooa> It's on a i440BX
[09:43:00] -!- heradon has joined #archlinux32
[09:43:03] <heradon> Moin
[09:43:08] <abaumann> hi
[09:43:14] <heradon> all fine?
[09:43:30] <abaumann> yeah. business as usual. :-)
[09:43:47] <heradon> another day another sh*t ^.^
[09:56:49] <abaumann> I have a mug saying "same sh*t, different day!" :-)
[09:57:07] <heradon> lol cool
[09:57:24] <heradon> i have a mug with Aperture Laboratories (u know?)
[09:57:56] <abaumann> :-)
[10:01:06] <KitsuWhooa> Is there a guide to set up a VM for arch32 builds? What sort of specs are needed? What ports/software?
[10:01:58] <heradon> i dont know, i host only 2 vms for the admin :>
[10:03:05] <abaumann> https://bbs.archlinux32.org
[10:03:07] <phrik> Title: How to set up a build slave / Building / Arch Linux 32 Forums (at bbs.archlinux32.org)
[10:04:22] <abaumann> that's the slave. as for the buildmaster and such..
[10:04:35] <abaumann> you have to ask deep42thought for details..
[10:04:43] <KitsuWhooa> I want to see if I can contribute a build VM, that's all :p
[10:04:52] <abaumann> ah.
[10:04:56] <heradon> you have 2 options:
[10:04:57] <abaumann> then a slave vm is enough.
[10:05:14] <heradon> install a vm and give the login data to deep2throut
[10:05:36] <heradon> or install a vm and install a build slave and ask deep2throut for upload your packages :D
[10:07:22] <KitsuWhooa> at some point it might be worth hosting a clean VM base image that people can just download, run, and change a few variables
[10:07:47] <abaumann> [root@eurobuild2 ~]# rsvg-convert twitter.svg > twitter.png
[10:07:47] <abaumann> Illegal instruction (core dumped)
[10:07:49] <abaumann> damn, again.
[10:07:51] <KitsuWhooa> :(
[10:08:30] <abaumann> pentium2 approach :-)
[10:09:06] <KitsuWhooa> there's also pentium3m
[10:09:44] <KitsuWhooa> hm, pentium2 doesn't have SSE, maybe you can add it on top?
[10:09:59] <abaumann> -C target-cpu=pentium2 -C target-feature=+sse
[10:10:02] <abaumann> yep. next try
[10:10:14] <abaumann> though I start to think those flags are just decoration.. :->
[10:10:26] <KitsuWhooa> I hope not...
[10:11:18] <heradon> i do illegal things :>
[10:11:32] <heradon> i am on work, i cannot chat here with irc.
[10:11:43] <KitsuWhooa> mmm
[10:11:56] <KitsuWhooa> abaumann: is there any way to check if the flags are actually getting passed through?
[10:12:00] <abaumann> you know that those irc chats are recorded on our webpage. :->
[10:12:03] <heradon> But i am not dumb - installed a gentoo VM and connect to it with x2go (over ssh) and i have a complete desktop *hrhr
[10:12:10] <abaumann> they are.
[10:12:24] <KitsuWhooa> ah, fair
[10:12:37] -!- deep42thought has joined #archlinux32
[10:12:37] <buildmaster> Hi deep42thought!
[10:12:37] <buildmaster> !rq deep42thought
[10:12:39] <phrik> buildmaster: <deep42thought> so you're a garbage collector?
[10:12:43] <abaumann> but if there are hard-coded optimizations or when half of the internet is downloaded via cargo with SSE2 in it.. there is not much you can do.
[10:12:45] <heradon> hey dude
[10:12:52] <deep42thought> Hi vollzornbrot!
[10:13:00] heradon is now known as Vollzornbrot
[10:13:07] <Vollzornbrot> yes vollzornbrot *lol*
[10:13:25] <deep42thought> I really like that name :-D
[10:13:27] <KitsuWhooa> abaumann: true, although I thought cargo didn't download binaries
[10:13:40] <KitsuWhooa> crates, anyway
[10:13:45] <Vollzornbrot> okay, lets talk about war....
[10:13:57] <Vollzornbrot> i need the sync.php -.-
[10:14:16] <deep42thought> umm, it was on _your_ host ...
[10:14:36] <Vollzornbrot> yes but deleted :(
[10:16:38] <deep42thought> Vollzornbrot: you still have the cronjob (or some residual thereof)?
[10:16:54] <Vollzornbrot> hmm?
[10:17:11] <deep42thought> I think, you were running some sort of cronjob to rsync the mirror
[10:17:23] <deep42thought> the question is if there is any trace of that left
[10:17:26] <Vollzornbrot> yes i can run a cron
[10:19:06] <deep42thought> abaumann: what parameters do you use to sync your mirror?
[10:20:22] <abaumann> http://archlinux32.andreasbaumann.cc
[10:20:48] <deep42thought> Vollzornbrot: ^
[10:21:31] <Vollzornbrot> thanks
[10:21:42] <Vollzornbrot> this is for cron right?
[10:21:47] <abaumann> yes.
[10:22:02] <Vollzornbrot> okay give me 1 or 2 min
[10:22:03] <abaumann> */5 * * * * /data/arch32/scripts/sync_mirror
[10:38:05] <Vollzornbrot> syncing
[10:38:09] <deep42thought> \o/
[10:43:08] <deep42thought> Vollzornbrot: you really want to have the sync log and lock available on your mirror?
[10:44:32] <Vollzornbrot> why not?
[10:44:44] <Vollzornbrot> ok no
[10:45:28] <Vollzornbrot> removed
[10:47:05] <deep42thought> and you changed the url, too?
[10:47:43] <Vollzornbrot> no archlinux32.vollzornbrot.de why?
[10:47:53] <deep42thought> archlinux32.vollzornbrot.de/mirror/
[10:48:03] <Vollzornbrot> argh *FUUUUUUUUUUUUUUUUUUUUU
[10:48:37] <deep42thought> gotcha
[10:48:40] <Vollzornbrot> fixed
[10:49:03] <deep42thought> yay, now even the pedantic deep42thought has nothing to complain about ;-P
[10:49:12] <abaumann> :-)
[10:49:15] <Vollzornbrot> not yet :D
[10:49:19] <deep42thought> lol
[10:49:43] <deep42thought> "no one likes pedantic people" - "that's not true: _almost_ nobody likes them"
[10:49:55] <abaumann> other pedantic people do. :-)
[10:50:06] <deep42thought> I'm sure, they don't really do
[10:50:21] <abaumann> well.. there is some common ground.. :-)
[10:50:47] <deep42thought> yes, but imagine having a perfect copy of yourself - I wouldn't last an hour ...
[10:51:03] <deep42thought> so much for "common ground"
[10:51:35] <deep42thought> Vollzornbrot: is this mirror running systemd?
[10:51:46] <Vollzornbrot> debian system
[10:51:49] <Vollzornbrot> debian 9
[10:51:57] <deep42thought> then you could add the sync script as a systemd service + timer
[10:52:04] <deep42thought> and activate the service from the php script
[10:52:09] <Vollzornbrot> deep42thought: i have 0 plan of systemd
[10:52:19] <deep42thought> np
[10:52:21] <Vollzornbrot> now the cron works ^
[10:52:29] <deep42thought> good point
[10:52:30] <deep42thought> :-D
[10:52:47] <deep42thought> sync.php is still todo
[11:01:50] <Vollzornbrot> sync is finish
[11:03:10] <deep42thought> regarding systemd: put https://ptpb.pw into a file /etc/systemd/system/sync-mirror.service and https://ptpb.pw into /etc/systemd/systemd/sync-mirror.timer
[11:03:20] <deep42thought> then you can 'systemctl enable sync-mirror.timer'
[11:03:29] <deep42thought> and 'systemctl start sync-mirror.timer'
[11:03:49] <deep42thought> and from the php scrip you can do 'systemctl start sync-mirror.service'
[11:25:34] <abaumann> rsvg-convert twitter.svg > twitter.png
[11:25:34] <abaumann> Illegal instruction (core dumped)
[11:25:42] <abaumann> again, with pentium2 and +sse flags.
[11:25:49] <abaumann> this start to get annoying..
[11:25:53] -!- oaken-source has quit [Ping timeout: 245 seconds]
[11:34:05] <KitsuWhooa> abaumann: would it be possible to build an older version of librsvg to test?
[11:34:14] <KitsuWhooa> to test if it's a rust change or a librsvg one
[11:34:30] <KitsuWhooa> if I recall correctly, the 42 one works fine
[11:35:33] <abaumann> good point. I personally doubt it's in librsvg, as they don't play with SSE2.
[11:36:15] <KitsuWhooa> true
[11:42:36] <Vollzornbrot> true == right?
[11:43:33] <abaumann> true == !false ;-)
[11:43:37] <deep42thought> true == right, but ! true === right
[11:43:43] <abaumann> LOL
[11:43:51] <Vollzornbrot> wtf
[11:44:53] <abaumann> just celebrating nerddom.. ;-)
[11:44:55] <Vollzornbrot> What do you call a deer with no eyes?
[11:45:16] <Vollzornbrot> No eye deer
[11:45:33] <Vollzornbrot> What you you call a deer with no eyes or legs?
[11:45:33] <deep42thought> obviously
[11:45:48] <Vollzornbrot> Still no eye deer
[11:46:17] <Vollzornbrot> what do you call a deer with no eyes, no legs and no genitalia?
[11:46:48] <abaumann> *the censorship police steps into the room now..
[11:46:49] <Vollzornbrot> still no fu*king eye deer
[11:46:57] <abaumann> lol
[11:47:00] <deep42thought> lol
[11:47:06] <Vollzornbrot> ^.^
[11:48:49] <Vollzornbrot> all work and no play makes deep42thought a dull boy!
[11:49:28] <deep42thought> I am?
[11:49:38] <Vollzornbrot> yes
[11:49:41] <deep42thought> :'-(
[11:49:47] <abaumann> redrum
[11:50:06] <deep42thought> !rq deep42thought
[11:50:06] <phrik> deep42thought: <deep42thought> "und wieder ist ein Tag vollbracht und wieder ist nur Murks gemacht"
[11:50:21] <deep42thought> well, that was one of the lesser appropriate quotes in this context
[11:50:26] <Vollzornbrot> exactly
[11:50:26] <deep42thought> THANKS phrik!!
[11:50:28] <abaumann> lunch time. AFK
[11:50:32] <deep42thought> cu
[11:50:45] <deep42thought> I have fun ocassionally, too
[11:50:50] <deep42thought> but it's less obvious ;-P
[11:51:00] <deep42thought> buildmaster: say something new!
[11:51:00] <buildmaster> deep42thought: Frifd
[11:51:09] <Vollzornbrot> !rq heradon
[11:51:10] <phrik> Vollzornbrot: <Heradon> Nein!
[11:51:14] <Vollzornbrot> lol
[11:51:27] <deep42thought> !rq deep42thought
[11:51:27] <phrik> deep42thought: <deep42thought> ahm, what is actually in the package "vulkan-headers" on i486? I thought, we disabled all vulcans?
[11:51:43] <Vollzornbrot> !rq heradon
[11:51:43] <phrik> Vollzornbrot: <Heradon> Nein!
[11:51:50] <Vollzornbrot> !rq Vollzornbrot
[11:51:50] <phrik> Vollzornbrot: Error: Couldn't get a random quote for that nick.
[11:51:58] <Vollzornbrot> *drop*
[11:52:01] <deep42thought> is that your only quote?
[11:52:43] <Vollzornbrot> i think yes
[11:58:56] <Vollzornbrot> the most audio cable comes from auxburg
[12:05:01] -!- oaken-source has joined #archlinux32
[13:17:05] <deep42thought> Vollzornbrot: I see your download spike https://jeti100.ioq.uni-jena.de hehe
[13:17:44] <buildmaster> i686/calibre is broken (says buildknecht2).
[13:40:36] <Vollzornbrot> lol
[13:56:31] <deep42thought> Vollzornbrot: your certificate on the mirror is invalid, too
[13:56:44] <deep42thought> (sry for bringing up all that suff)
[14:14:00] -!- alad has joined #archlinux32
[14:14:19] <alad> hi, I was wondering if you guys have ever seen arch-specific (x86_64 or otherwise) in a PKGBUILD's prepare() function?
[14:15:18] <deep42thought> I haven't
[14:17:57] <deep42thought> let me grep for it ...
[14:22:31] <deep42thought> there are some
[14:22:35] <deep42thought> e.g. community/cmucl
[14:27:58] <alad> thanks, interesting
[14:28:11] <alad> another question, does the same hold for pkgver() too? at least that one sounds unlikely
[14:32:58] <deep42thought> nope, none found
[14:33:46] <deep42thought> (note, that I was not looking through aur packages)
[14:41:58] <alad> thanks
[15:30:03] -!- ofara_ has joined #archlinux32
[15:41:06] -!- T`aZ has quit [Ping timeout: 252 seconds]
[15:42:52] -!- T`aZ has joined #archlinux32
[16:01:57] -!- abaumann has quit [Quit: leaving]
[16:39:37] -!- deep42thought has quit [Quit: Leaving.]
[16:52:27] -!- oaken-source has quit [Ping timeout: 240 seconds]
[16:59:51] <buildmaster> i686/js60 is broken (says buildknecht).
[17:12:13] -!- yans has joined #archlinux32
[17:16:23] <buildmaster> i686/kresus are broken (says buildknecht).
[17:28:58] <buildmaster> i686/qca is broken (says nlopc46).
[18:37:41] -!- jason0597 has quit [Remote host closed the connection]
[18:42:19] -!- jason0597 has joined #archlinux32
[19:35:52] -!- MrBIOS has joined #archlinux32
[19:42:05] -!- samantaz has joined #archlinux32
[20:21:06] -!- alad has parted #archlinux32
[20:27:06] -!- MrBIOS has quit [Quit: MrBIOS]
[21:05:34] -!- jason0597 has quit [Ping timeout: 255 seconds]
[21:11:10] -!- oaken-source has joined #archlinux32
[21:52:25] -!- yans has quit [Quit: chaos is the only true answer]
[22:01:45] <buildmaster> i686/vtk is broken (says buildknecht2).
[22:22:32] <buildmaster> i686/kde-cli-tools are broken (says buildknecht2).
[23:04:47] <buildmaster> i686/kamoso is broken (says rechenknecht).
[23:12:28] -!- samantaz_ has joined #archlinux32
[23:13:48] -!- oaken-source has quit [Ping timeout: 245 seconds]
[23:14:43] -!- samantaz has quit [Ping timeout: 255 seconds]