#archlinux32 | Logs for 2018-05-21
Back
[01:47:08] -!- NoobAlice has joined #archlinux32
[03:02:00] -!- bill-auger has quit [Remote host closed the connection]
[03:05:20] -!- bill-auger has joined #archlinux32
[05:50:59] -!- isacdaavid has quit [Ping timeout: 260 seconds]
[06:20:50] -!- NoobAlice has quit [Quit: Leaving.]
[08:46:57] -!- eduardoeae_ has quit [Ping timeout: 248 seconds]
[11:42:58] -!- eduardoeae_ has joined #archlinux32
[15:53:01] -!- isacdaavid has joined #archlinux32
[17:40:25] -!- isacdaavid has quit [Read error: Connection reset by peer]
[18:20:16] -!- buildmaster has joined #archlinux32
[18:20:34] -!- deep42thought has joined #archlinux32
[18:20:35] <buildmaster> Hi deep42thought!
[18:20:42] <deep42thought> Hi buildmaster, what's up?
[18:21:11] <buildmaster> up? I'm up for 1 minute, load average: 0.81, 0.40, 0.15 ... and I'm insane :-D
[18:22:05] <tyzoid> wb
[18:32:07] * buildmaster resumes sanity.
[18:32:11] <deep42thought> \o/
[18:40:04] -!- abaumann has joined #archlinux32
[18:40:05] <buildmaster> Hi abaumann!
[18:44:50] <abaumann> hi there. :-)
[18:45:23] <abaumann> The glibc/binutils issue is fixed, but I'm hesitating to backport things into a 2.30 binutils, maybe it's safer to wait for a new binutils release..
[18:45:40] <abaumann> firefox breaks because a too old rust version is in build-support, it seems.
[18:52:37] <tyzoid> deep42thought / abaumann: Any thoughts as to why uname would show i586 on archiso32?
[18:53:01] <tyzoid> https://mirror.archlinux32.org
[18:53:03] <phrik> Title:#archlinux32 | Logs for 2018-05-18 (at mirror.archlinux32.org)
[18:54:58] <abaumann> mmh. Because the CPU is not a full i686
[18:55:01] <abaumann> maybe.
[18:55:18] <abaumann> In my i486-version of the ISO I get i586 on an AMD-K6.
[18:55:32] <tyzoid> abaumann: The cpu in that machine is one of these: https://en.wikipedia.org
[18:55:34] <phrik> Title:Transmeta Crusoe - Wikipedia (at en.wikipedia.org)
[18:55:45] <tyzoid> which is awesome
[18:55:55] <tyzoid> but also that I have no idea if we can support it
[18:56:11] <tyzoid> is this something that we'll just need to wait for i486?
[18:57:21] <abaumann> i486 would run for sure.
[18:57:28] <abaumann> Seems to be a PIII or so.
[18:57:46] <abaumann> without a cat /proc/cpuinfo seeing the CPU flags it's hard to tell.
[18:57:48] <tyzoid> Pentium 3 should be P6
[18:58:20] <tyzoid> but yeah
[18:58:30] <tyzoid> If they come back, I'll ask
[18:59:00] <abaumann> They are not fully i686 compliant. Some small things are missing.
[18:59:14] <abaumann> if they have CMOV, the ISO should boot.
[18:59:45] <abaumann> It's similar to the AMD Geode or a VIA thingy. They may lack MMX, SSE and stuff..
[19:00:41] <tyzoid> IIRC you should be able to boot without SSE
[19:00:49] <tyzoid> just be limited in what you can run
[19:00:57] <abaumann> exactly.
[19:01:21] <tyzoid> CMOV yeah... gonna be hard to work without that
[19:01:56] <abaumann> That's for instance the only reason, why my AMD K6 machine is not running.. :-)
[19:02:06] <abaumann> ..and needs the i486 version.
[19:09:50] <tyzoid> abaumann: speaking if 486, does your vm need more resources?
[19:10:08] <tyzoid> and/or you going to test ipv6 any time soon?
[19:10:28] <abaumann> ah. thanks, no, resources are fine.
[19:10:33] <tyzoid> it's still pegging 480mb of 512 mb ram
[19:10:35] <tyzoid> and I assume swapping
[19:10:40] <abaumann> yes.
[19:10:51] <abaumann> My target archicture is 64M, :-)
[19:10:56] <abaumann> not for building, of course.
[19:11:01] <tyzoid> lol
[19:11:06] <tyzoid> want a smaller vm?
[19:11:09] <abaumann> only few packages need more than 1/2 GB RAM to build.
[19:11:10] <abaumann> :-)
[19:11:12] <tyzoid> I can spin you up another
[19:11:26] <abaumann> My problem is at the moment time..
[19:11:38] <tyzoid> Ain't that the truth
[19:13:57] <deep42thought> buildmaster-fail-reason: "The power cable wasn't inserted all the way and the drive came unplugged."
[19:14:11] <abaumann> lol, classic :-)
[19:14:34] <deep42thought> that's why you use servers with redundant power supplies ;-)
[19:15:18] <deep42thought> plus: you can plug them in (different) switchable power sockets ;-)
[19:15:37] <deep42thought> that's what's done in our lab - except the redundancy :-/
[19:16:40] -!- belanthor has joined #archlinux32
[19:23:28] <abaumann> why, tensorflow depends on cuda and cudnn, cuda is on the blacklist, so tensorflow should not get onto the build list?
[19:23:56] <abaumann> ah. it's a makedepend.
[19:29:42] <abaumann> tensorflow needs bazel to build, which is in the blacklist, so I'm blacklisting tensorflow.
[19:31:25] <deep42thought> makepdepends shoud suffice to blacklist a package
[19:31:30] <deep42thought> checkdepends should not
[19:31:48] <abaumann> mmh.
[20:18:21] -!- abaumann has quit [Quit: leaving]
[20:19:01] <buildmaster> sbcl is broken (says rechenknecht).
[20:29:34] <deep42thought> abaumann: I found the bug - the query said "the dependency must be relevant for building and relevant for the binary package" to be a valid blacklist reason
[20:29:56] <deep42thought> I changed that to "or", now tensorflow is automatically blacklisted
[20:48:24] <buildmaster> gitlab-runner is broken (says buildknecht3).
[21:07:08] -!- deep42thought has quit [Quit: Leaving.]
[22:28:32] <buildmaster> imagemagick is broken (says buildknecht3).
[23:41:09] -!- belanthor has quit [Quit: Leaving]