#archlinux32 | Logs for 2021-05-25

Back
[00:59:46] -!- jonathon has joined #archlinux32
[01:05:15] -!- doskoi has joined #archlinux32
[01:16:01] -!- doskoi has quit [Quit: kthxbye]
[01:16:15] -!- doskoi has joined #archlinux32
[01:17:16] jonathon is now known as fez
[01:52:37] -!- bill-auger has joined #archlinux32
[03:05:37] -!- JerryXiao has joined #archlinux32
[03:26:26] fez is now known as jonathon
[03:28:53] -!- Foxboron has quit [Ping timeout: 264 seconds]
[04:01:14] -!- Namarrgon has joined #archlinux32
[04:01:45] <Namarrgon> do you want phrik to rejoin here?
[04:55:48] jonathon is now known as nohtanoj
[07:38:05] -!- ytg has joined #archlinux32
[08:01:17] -!- abaumann has joined #archlinux32
[08:01:18] <076AAAJ5E> Hi abaumann!
[08:01:18] <076AAAJ5E> !rq abaumann
[08:01:29] <abaumann> mmh, that's a weird name :-)
[08:01:58] <abaumann> deep42thought: I had bugwrangler as cloak, I'm also fine with dev, is shorter. :-)
[08:56:16] <girls> Namarrgon: yes, that would be great! Thanks
[08:56:23] <girls> buildmaster: what's up?
[08:56:24] <076AAAJ5E> up? I'm up for 1 day, 8 hours, 18 minutes, load average: 2.57, 2.82, 3.51
[08:56:35] <girls> it's indeed him, ... strange
[08:56:53] <girls> that's his *true* name, I guess
[08:57:19] 076AAAJ5E is now known as buildmaster
[08:58:42] <girls> 1621885786 076AAAJ5E Nick collision, forcing nick change to your unique ID
[08:58:42] <girls> 1621885786 -!- buildmaster changed nick to 076AAAJ5E
[08:59:11] <girls> so: not our fault :-)
[11:15:25] <abaumann> ah, ok :-)
[12:22:51] -!- CPurpurea has joined #archlinux32
[12:22:55] -!- CPurpurea has parted #archlinux32
[12:44:17] -!- phrik has joined #archlinux32
[12:44:22] <Namarrgon> !ping
[12:44:22] <phrik> pong
[14:44:20] -!- elibrokeit has joined #archlinux32
[15:11:26] <girls> !rq deep42thought
[15:11:26] <phrik> girls: <deep42thought> otoh I have the feeling, archlinux32 dies anyway if I get overrun by a bus
[15:13:47] -!- deep42thought has joined #archlinux32
[15:13:48] <buildmaster> Hi deep42thought!
[15:13:48] <buildmaster> !rq deep42thought
[15:13:49] <phrik> buildmaster: <deep42thought> otoh I have the feeling, archlinux32 dies anyway if I get overrun by a bus
[15:14:06] <deep42thought> abaumann, what's your opinion on pushing gcc11 to testing?
[16:23:16] <abaumann> deep42thought: it build and run the test suits? so it might hang for some stupid reason..
[16:23:47] <deep42thought> it's waiting for stupid dependencies :-)
[16:23:58] <abaumann> I wonder which ones.
[16:24:01] <deep42thought> gcc-libs :-)
[16:24:06] <abaumann> oh.
[16:24:10] <deep42thought> well, stuff depends on gcc-libs
[16:24:15] <abaumann> yeah, and almost half of the world depends on gcc-libs
[16:24:29] <abaumann> in theory gcc-libs is ABI-backwards-compatible
[16:24:35] <deep42thought> that's how I noticed: I rebuilt bareos and cannot install, because I built against gcc-libs11 from staging
[16:24:56] <deep42thought> oh it is?
[16:25:02] <abaumann> I thought so.
[16:25:15] <deep42thought> there's an easy way to find out >:-)
[16:25:18] <abaumann> but to make sure I usually push binutils, gcc, glibc in one package.
[16:25:35] <abaumann> well. pushing onces from staging to testing is not such a problem.
[16:25:45] <abaumann> we just have to fix it afterwards (eventually)
[16:26:33] <abaumann> let me fast do an update test
[16:27:16] <abaumann> just updating gcc and gcc-libs in a stable vm from staging.
[16:27:53] <deep42thought> you're lucky: I already started db-update, but I used the wrong arguments, so nothing was forced :-D
[16:28:50] <abaumann> mmh. looks.
[16:28:52] <abaumann> good.
[16:29:08] <abaumann> I can build a cmake-based C project against gcc 11 when pushed to stable
[16:29:24] <deep42thought> I'll move gcc, glibc, binutils then?
[16:29:57] <abaumann> yep.
[16:30:08] <abaumann> from now it's _our_ fault. :-)
[16:30:20] <deep42thought> well, it's alway *our* fault :-)
[16:30:32] <abaumann> :-)
[16:36:52] <deep42thought> ok, done
[16:38:33] <deep42thought> now, I can finally update my machines again :-D
[16:43:13] <deep42thought> if I'm suddenly offline, then the i486 branch is broken :-D
[16:49:31] <abaumann> :-)
[18:36:45] <deep42thought> what's the easiest way to generate some sse2 instruction to see, whether my cpu can actually execute it?
[18:53:35] <abaumann> http://archlinux32.andreasbaumann.cc
[18:53:35] <phrik> Title: Index of /SSE2 (at archlinux32.andreasbaumann.cc)
[18:53:42] <abaumann> I played once with some code..
[18:54:38] -!- abaumann has quit [Quit: leaving]
[19:31:34] <deep42thought> thanks! Your test.c snippet correctly(?) detects sse2
[19:31:56] <deep42thought> on the box, where __builtin_cpu_supports("sse2") returns 0
[19:32:51] <deep42thought> but, this snippet only checks for some bits in some flags register, right?
[19:33:22] <deep42thought> I'd really like some sse2-only instruction, which makes my program dump core, if it's run on the wrong hardware :-)
[20:11:11] -!- phrik has quit [Remote host closed the connection]
[20:11:25] -!- phrik has joined #archlinux32
[20:12:59] -!- abaumann has joined #archlinux32
[20:12:59] <buildmaster> Hi abaumann!
[20:12:59] <buildmaster> !rq abaumann
[20:13:00] <phrik> buildmaster: <abaumann> All the times I was debugging some network settings, just to find out, that I tripped a cable. :-)
[20:13:27] <abaumann> deep42thought: one idea would be to execute inline assembly which uses SSE2 and catch the SIGILL signal in a terminal handling
[20:14:35] <deep42thought> yes, exactly
[20:14:39] <deep42thought> something like that
[20:15:04] <deep42thought> my problem is, that I a) do not know how to write inline-assembly and b) don't know any sse2 instructions off the top of my head :-D
[20:19:32] <abaumann> yoy could call a sse2 compiler intrinsic like _mm_cvtsi128_si32 or so
[20:19:48] <abaumann> I'm checking what /proc/cpuinfo is doing exactly..
[20:19:55] <deep42thought> thx :-)
[20:29:42] <abaumann> classical approach for a quorum system: have N guessers and if N/2+1 answer SSE2, it must be SSE2
[20:30:01] <abaumann> it seem the linux kernel in arch/x86 does some probing with cpuid
[20:31:32] <deep42thought> you're on your alix, right now, right?
[20:31:38] <deep42thought> can you compile and test your test.c?
[20:31:41] <abaumann> aeh. not really.
[20:31:43] <deep42thought> (I don't have gcco n mine)
[20:31:50] <deep42thought> ah, ok
[20:31:54] <abaumann> this is one problem. I don't have a real non-SSE2 machine running.
[20:31:58] <deep42thought> then it's the same effort for both of us :-)
[20:32:13] <deep42thought> I compiled test.c and ran it on the alix - it crashed
[20:32:18] <abaumann> I have an AMD something around, let me check, but maybe that one.
[20:32:23] <abaumann> mmh.
[20:32:24] <deep42thought> is this intended on a non-sse2 system?
[20:32:42] <abaumann> no, crashing is not the idea..
[20:32:49] <abaumann> ..though this can be trapped too. ;-)
[20:32:51] <deep42thought> ok, then the crash is something else
[20:35:40] <deep42thought> ah, actually it's test2.c, which I tried
[20:35:47] <deep42thought> (test.c does the exact same thing as pacman)
[20:38:36] <abaumann> hah. an AMD Athlon with sse and no sse2. good. :-)
[20:42:16] <abaumann> all tests end in "No SSE2" and no crashes
[20:51:14] <deep42thought> I'm currently installing gcc on the alix and testing test2.c there
[20:51:20] <deep42thought> => "No SSE2"
[20:51:22] <deep42thought> good
[20:51:50] <deep42thought> though, I think, I'll report the bug regarding __builtin_cpu_supports() to gcc
[21:13:01] nohtanoj is now known as jonathon
[21:36:10] -!- abaumann has quit [Quit: leaving]
[21:40:11] -!- deep42thought has quit [Quit: Leaving.]
[22:17:48] -!- abaumann has joined #archlinux32
[22:17:48] <buildmaster> Hi abaumann!
[22:17:48] <buildmaster> !rq abaumann
[22:17:49] <phrik> buildmaster: <abaumann> it usually automatically fails. :-)
[22:18:51] <abaumann> I have a test4.c with SSE2 intrinsincs..
[22:19:01] <abaumann> In the same location..
[22:19:38] <abaumann> has to be compiled with -msse2 on a i486
[22:23:03] -!- abaumann has quit [Quit: leaving]
[22:33:03] -!- ytg has quit [Quit: bye]
[22:45:02] -!- jonathon has parted #archlinux32
[22:51:16] -!- jonathon has joined #archlinux32
[22:54:32] -!- phrik has quit [Remote host closed the connection]
[23:03:59] -!- phrik has joined #archlinux32
[23:11:19] -!- phrik has quit [Remote host closed the connection]
[23:11:48] -!- phrik has joined #archlinux32
[23:43:50] -!- phrik has quit [Read error: Connection reset by peer]
[23:44:04] -!- phrik has joined #archlinux32
[23:52:20] -!- phrik has quit [Remote host closed the connection]
[23:52:36] -!- phrik has joined #archlinux32