#archlinux32 | Logs for 2018-08-09
Back
[00:55:47] -!- davor_ has joined #archlinux32
[01:01:21] -!- woshty has quit [Ping timeout: 248 seconds]
[01:02:47] -!- davor has quit [*.net *.split]
[01:02:48] davor_ is now known as davor
[01:05:19] -!- Alina-malina has quit [Ping timeout: 260 seconds]
[01:25:12] -!- Alina-malina has joined #archlinux32
[02:13:55] -!- phrik has quit [Remote host closed the connection]
[02:14:03] -!- phrik has joined #archlinux32
[02:14:04] <buildmaster> Hi phrik!
[02:17:46] -!- elibrokeit has quit [Ping timeout: 256 seconds]
[02:19:19] -!- phrik has quit [Ping timeout: 265 seconds]
[02:30:05] -!- phrik has joined #archlinux32
[02:30:05] <buildmaster> Hi phrik!
[02:32:21] -!- elibrokeit has joined #archlinux32
[02:50:36] Guest12145 is now known as eschwartz[m]
[04:17:07] <buildmaster> deepin-dock is broken (says buildknecht).
[04:43:20] -!- bill-auger has quit [Remote host closed the connection]
[04:43:56] -!- bill-auger has joined #archlinux32
[07:45:04] -!- deep42thought has joined #archlinux32
[07:45:05] <buildmaster> Hi deep42thought!
[07:45:12] <deep42thought> Hi buildmaster, what's up?
[07:45:13] <buildmaster> up? I'm up for 1 day, 6 hours, 3 minutes, load average: 0.66, 0.74, 0.78
[07:59:56] -!- abaumann has joined #archlinux32
[07:59:57] <buildmaster> Hi abaumann!
[08:00:03] <abaumann> hi deep42thought
[08:02:28] <deep42thought> Hi abaumann!
[08:02:47] <deep42thought> did you have a look at python, yet?
[08:05:39] <abaumann> no. it's building on my machine. :-)
[08:05:44] <deep42thought> oh, hmm
[08:05:52] <abaumann> https://packages.archlinux32.org
[08:06:02] <abaumann> What I see is glibc binutils and gcc running in parallel..
[08:06:07] <abaumann> Is this good?
[08:06:15] <deep42thought> umm
[08:06:21] <deep42thought> this is definitely not intended
[08:06:37] <abaumann> maybe older versions and then the build schedule catches up later?
[08:07:14] <abaumann> I'll do a complete rebuild of i486, as I can do it more or less automatic now. I'm curious if some toolchain issues disappear..
[08:07:17] <deep42thought> I implemented some logic to not remove some toolchain stuff from the build list if they are built the first time
[08:07:36] <abaumann> ok. let's see if it works. :-)
[08:07:51] <deep42thought> probably, it won't
[08:08:04] <deep42thought> I mean, it already failed in not handing out the stuff in parallel :-D
[08:09:04] <abaumann> Do we have enough build slaves to handle this big rebuild?
[08:09:22] <deep42thought> good question
[08:09:30] <deep42thought> I threw in my two spare machines
[08:09:47] <deep42thought> so we now have 5 of mine and one of you and tyzoid each
[08:10:06] <abaumann> tyzoid-srv0-bs0 and buildknecht3 are not running, it seems.
[08:10:14] <deep42thought> buildknecht3 was shut down
[08:10:22] <abaumann> ah
[08:10:34] <deep42thought> the buildknechte are Vollzornbrot's
[08:10:41] <abaumann> ah right.
[08:10:41] <deep42thought> and he needed some cpu time for himself :-D
[08:10:50] <abaumann> very understandable. :-)
[08:11:15] <deep42thought> but first, we need to pull the cork
[08:11:29] <deep42thought> the rest is then only a matter of how long it takes to compile the pile
[08:11:37] <abaumann> right.
[08:11:50] <deep42thought> python failed in check(), it seems
[08:11:57] <deep42thought> did it pass check on your box?
[08:12:16] * deep42thought recalls having python check() issues in the past
[08:12:17] <abaumann> still running in it's first attempt
[08:12:38] <abaumann> testCongestion (test.test_socket.RDSTest) ... /usr/bin/xvfb-run: line 181: 18052 Killed DISPLAY=:$SERVERNUM XAUTHORITY=$AUTHFILE "$@" 2>&1
[08:12:53] <abaumann> not X again..
[08:12:57] <abaumann> maybe the patch failed..
[08:13:20] <deep42thought> :-(
[08:13:31] <deep42thought> I think, that issue is still open in the python bug tracker
[08:13:43] <deep42thought> the other issue was to add -fvwrap (not sure about the exact letters) to the flags
[08:14:10] <abaumann> python started to generate a postscript file with 'gs' and open it, when python crashes.
[08:14:14] <abaumann> what a dump idea.
[08:14:22] <abaumann> *dumb
[08:14:48] <deep42thought> why should that crash?
[08:15:11] <abaumann> you have a python script using a c-interface and this one segfaults.
[08:15:30] <abaumann> so the error gets written to a postscript file. neat. but not really nice if you are on a server without graphic output.
[08:15:49] <abaumann> very soon linux will only output to a mesa-optimized output frame buffer or so.. :->
[08:16:16] <deep42thought> systemd will, I'm sure
[08:16:50] <abaumann> actually, all snark on systemd asside, it runs almost without any modifications on a i486. I give them credit for that..
[08:16:55] <deep42thought> isn't there a "--fancy" option to the system log which generates a pdf with clickable links and everything?
[08:17:09] <deep42thought> nice :-)
[08:17:12] <abaumann> yeah. but it's not the default and I have an option to switch it off.
[08:17:22] <abaumann> ..or on.
[08:17:25] <abaumann> this is ok.
[08:17:26] * deep42thought was joking about --fancy
[08:17:51] <abaumann> Aeh. LOL. Got me there. :-)
[08:18:02] <abaumann> there is no python directory in the packages?
[08:18:09] <abaumann> packages32..
[08:18:16] <deep42thought> huh?
[08:18:27] <deep42thought> python2
[08:18:30] <deep42thought> ah, right
[08:18:32] <abaumann> yeah.
[08:18:41] <abaumann> python2 is about the broken tests..
[08:18:57] <deep42thought> ah, right
[08:19:02] <deep42thought> so this is "different"
[08:19:03] <deep42thought> ...
[08:20:14] <abaumann> mmh. there is no patching away of xvfb-run in python2
[08:20:26] <abaumann> now I recall. I did that in the i486 branch, because I have no X (yet).
[08:22:51] <abaumann> testCongestion (test.test_socket.RDSTest) ... /usr/bin/xvfb-run: line 181: 18052 Killed DISPLAY=:$SERVERNUM XAUTHORITY=$AUTHFILE "$@" 2>&1
[08:22:59] <abaumann> mmh. this could also be the test being killed.
[08:23:11] <abaumann> It does not necessarily mean that something is wrong with xvfb-run
[08:23:49] <abaumann> Though I really don't see the poing to do all low-level tests in python with a running framebuffer X-server.
[08:23:51] <deep42thought> I think, something is wrong with the test when executed with xvfb-run
[08:24:05] <deep42thought> but it's only a guess
[08:24:09] <abaumann> hang on..
[08:24:15] <abaumann> I built a python, which one was it..
[08:24:39] <abaumann> python-3.6.6
[08:24:42] <abaumann> in i486
[08:25:01] <abaumann> so 2.7.15 would also fail in my branch most likely.
[08:25:10] <abaumann> * 3.7.0
[08:26:50] <abaumann> yeah. so. i'll move all my current i486 packages to stable and then I start a complete rebuild.
[08:28:11] <abaumann> :: installing perl (5.28.0-1.0) breaks dependency 'perl<5.27' required by net-snmp
[08:28:14] <abaumann> :: installing perl (5.28.0-1.0) breaks dependency 'perl<5.27' required by perl-locale-gettext
[08:28:17] <abaumann> :: installing perl (5.28.0-1.0) breaks dependency 'perl<5.27' required by perl-xml-parser
[08:28:20] <abaumann> :: installing perl (5.28.0-1.0) breaks dependency 'perl<5.27' required by xmms2
[08:28:23] <abaumann> was there a big Perl rebuild too?
[08:28:33] <deep42thought> I saw some perl packages
[08:28:40] <deep42thought> but I'm not sure it was a big rebuild ..
[08:28:56] <abaumann> There should be one, because I saw perl version errors also on 64-bit.
[08:29:26] <deep42thought> perl-5.26.2-2.0-i686.pkg.tar.xz perl-5.28.0-1.0-i686.pkg.tar.xz
[08:29:34] <abaumann> :: installing perl (5.28.0-1) breaks dependency 'perl<5.27' required by hivex
[08:29:37] <abaumann> :: installing perl (5.28.0-1) breaks dependency 'perl<5.27' required by perl-sys-virt
[08:30:40] <deep42thought> umm
[08:30:51] <deep42thought> there are like >100 perl packages on the build list
[08:31:48] <abaumann> ok. let's wait then.
[08:32:18] <abaumann> binutils: /usr/share/locale/pt/LC_MESSAGES/bfd.mo exists in filesystem
[08:32:21] <abaumann> yes. this is fun.
[08:32:56] <abaumann> deleting it should be fine.
[08:35:56] <abaumann> pacman -Qqo '/usr/lib/perl5/5.26'
[08:35:56] <abaumann> gdal
[08:35:56] <abaumann> graphicsmagick
[08:35:56] <abaumann> graphviz
[08:35:56] <abaumann> imagemagick
[08:35:59] <abaumann> libproxy
[08:36:01] <abaumann> perl-error
[08:36:04] <abaumann> perl-mailtools
[08:36:06] <abaumann> perl-timedate
[08:36:09] <abaumann> perl-uri
[08:36:11] <abaumann> subversion
[08:36:14] <abaumann> xfconf
[08:36:16] <abaumann> oj.
[08:36:30] <deep42thought> well, the rebuild is still in progress
[08:36:33] <deep42thought> or rather on hold
[08:38:11] <abaumann> python: configure: WARNING: unrecognized options: --with-threads
[08:38:16] <abaumann> mmh.
[08:38:30] <abaumann> threads are highly overrated :-)
[08:38:45] <deep42thought> who uses them, anyways?
[08:39:06] <abaumann> me. in python and even in perl. :-)
[08:40:18] <abaumann> yep. the warning is there also on 64-bit.
[08:42:00] <abaumann> mmh. I guess the option is obsolete and threads are now simply required by python 3.7
[09:04:01] <abaumann> mmh. my build slave just tanked.. do python tests now use 32GB of RAM?
[09:04:16] <deep42thought> most probably not
[09:04:31] <abaumann> * sigh
[09:05:01] <abaumann> * abaumann walks to his cabinet and hard-resets the machine..
[09:32:17] <elibrokeit> we had pkgbuild.com crash with oom a whole bunch of times when rebuilding python 3.7
[09:32:51] <deep42thought> sounds like fun
[09:34:57] <elibrokeit> it was kernel memory so the whole server locked up and rebooted :p
[09:34:58] <elibrokeit> 306 (test_socket)
[09:35:06] <elibrokeit> ^^ the troublesome python test
[09:36:48] <abaumann> oh. :-)
[09:38:01] <abaumann> there was once the idea of 'make test' for normal unit tests and 'make torture' to do the real resource consuming and stress tests. :-)
[09:38:35] <deep42thought> I like that name and distinction
[09:38:38] <abaumann> 1:00:41 load avg: 2.01 [306/415] test_socket
[09:38:47] <abaumann> thats the last message I see in the logs. :-)
[09:41:23] <abaumann> Ironically disabled swap gives a more stable build slave, then the kernel at least kills the offending test. After enabling swap the machine went into trashing.. :-)
[10:06:57] -!- davor has quit [Ping timeout: 248 seconds]
[10:07:42] -!- davor has joined #archlinux32
[10:14:48] -!- eduardoeae has quit [Ping timeout: 260 seconds]
[10:18:03] -!- oaken-source has joined #archlinux32
[10:47:32] -!- abaumann has quit [Quit: leaving]
[11:06:37] -!- abaumann has joined #archlinux32
[11:06:38] <buildmaster> Hi abaumann!
[11:06:42] <deep42thought> Hi abaumann!
[11:06:48] <abaumann> I get tons of core dumps when running python tests.
[11:06:53] <abaumann> seems completly borked.
[11:06:58] <abaumann> for 32-bit :-)
[11:07:04] <deep42thought> errrm
[11:07:05] <deep42thought> hmmm
[11:07:12] <deep42thought> proposal:
[11:07:17] <deep42thought> skip tests for now
[11:07:25] <deep42thought> and bug upstream with it :-)
[11:07:34] <abaumann> if python is broken for 32-bit then this is pretty much the end..
[11:07:47] <deep42thought> maybe only the tests?
[11:07:48] <deep42thought> :-/
[11:07:54] <abaumann> I doubt.
[11:08:21] <abaumann> do they claim to support 32-bit?
[11:08:33] <deep42thought> I doubt they don't
[11:08:53] <abaumann> macos and windows 32-bit has downloadable installers.
[11:09:22] <deep42thought> maybe they will drop support for building _on_ 32 bit at some point
[11:09:33] <abaumann> or at this very point :->
[11:09:34] <deep42thought> but I doubt, they'll drop 32 bit support anytime soon
[11:09:55] <abaumann> what could also be: toolchain issues..
[11:10:05] <deep42thought> well, but then the build should fail (oom? like firefox?), not the test suite - right?
[11:10:06] <abaumann> ..manifest in python.
[11:10:40] <abaumann> not necessarily: using a rarely used programming feature maybe which then breaks in the python binary itself.
[11:28:34] -!- abaumann has quit [Quit: leaving]
[11:38:04] -!- woshty has joined #archlinux32
[12:10:02] -!- eduardoeae has joined #archlinux32
[13:13:36] -!- abaumann has joined #archlinux32
[13:13:37] <buildmaster> Hi abaumann!
[13:13:56] <abaumann> deep42thought: I have a comletely different idea, why things are crashing in Python..
[13:13:59] <abaumann> Run tests in parallel using 3 child processes
[13:14:09] <abaumann> The default is to spawn many tests in parallel..
[13:14:31] <abaumann> the compiled python seems sort of ok, at least when started manually in the build srcdir
[13:15:07] <deep42thought> hmm
[13:15:14] <abaumann> args.extend(['-j', '0']) # Use all CPU cores
[13:15:15] <deep42thought> why should tests fail if run massively in parallel?
[13:15:26] <abaumann> because of resource depletion.
[13:15:34] <deep42thought> yeah, ok
[13:15:37] <abaumann> not to mention, that my build server was killed two more times now.
[13:15:46] <abaumann> I'll try with -j 1 :-)
[13:15:54] <deep42thought> :-)
[13:16:04] <deep42thought> on i686 or i486?
[13:16:13] <abaumann> i686
[13:16:21] -!- oaken-source has quit [Ping timeout: 240 seconds]
[13:16:26] <abaumann> I never build something on i486 which doesn't work on i686 just fine.
[13:16:32] <deep42thought> :-D
[13:16:35] <deep42thought> good point :-D
[13:16:47] <abaumann> I did in the past.. and got punished. :-)
[13:18:59] <abaumann> == CPU count: 1
[13:19:02] <abaumann> Run tests in parallel using 3 child processes
[13:19:09] <deep42thought> O.o
[13:19:36] <abaumann> make EXTRATESTOPS='-j 1' test
[13:19:39] <abaumann> no effect
[13:19:52] <deep42thought> is this a typo
[13:19:53] <deep42thought> ?
[13:19:57] <deep42thought> or is it really "OPS"
[13:20:05] <abaumann> oh. thanks :-)
[13:20:20] <abaumann> Run tests in parallel using 1 child processes
[13:20:21] <abaumann> :-)
[13:20:26] <abaumann> let's see..
[13:21:14] <abaumann> The next step will be to distribute code to neighbour servers or use the Amazon cluster in the build script to run tests..
[13:21:27] <deep42thought> !grab abaumann
[13:21:28] <phrik> deep42thought: Tada!
[13:21:35] <abaumann> * abaumann's zynical remark of the day :->
[13:37:51] -!- woshty has quit [Ping timeout: 240 seconds]
[13:56:46] -!- N0rootz has joined #archlinux32
[14:07:03] -!- woshty has joined #archlinux32
[14:22:43] -!- N0rootz has quit [Ping timeout: 260 seconds]
[14:33:30] -!- abaumann has quit [Quit: leaving]
[14:40:01] -!- woshty has quit [Ping timeout: 248 seconds]
[14:45:15] -!- woshty has joined #archlinux32
[15:01:05] -!- abaumann has joined #archlinux32
[15:01:06] <buildmaster> Hi abaumann!
[15:01:26] <abaumann> hah. I;m in test 314 now (passed 306)
[15:01:33] <deep42thought> \o/
[15:01:42] <abaumann> so, I'll make a patch in python with EXTRATESTFLAGS='-j 1'
[15:01:49] <deep42thought> thank you
[15:02:08] <abaumann> 23 tests are failing, also some trivial ones.. this doesn't make me too happy.. but as you sad.. we ignore them for now.
[15:02:12] <deep42thought> "EXTRATESTOPTS", right?
[15:02:17] <abaumann> *said
[15:02:19] <abaumann> damn..
[15:02:22] <abaumann> again :-)
[15:02:24] <abaumann> yes. :-)
[15:02:41] <deep42thought> :-D
[15:02:48] <deep42thought> good to have a pedantic around ...
[15:03:02] <abaumann> I just want to let it finish so I can save the workspace for inspecting the failing tests later.
[15:03:13] <deep42thought> ok
[15:03:18] <abaumann> ..pedantism pays off. sometimes. :-)
[15:03:35] <abaumann> in coding.. not in real life.
[15:05:03] <deep42thought> you'd be surprised ;-)
[15:09:46] <buildmaster> deepin-session-ui is broken (says buildknecht2).
[15:11:11] -!- oaken-source has joined #archlinux32
[16:16:11] <abaumann> testCongestion (test.test_socket.RDSTest) ... make: *** [Makefile:1066: test] Killed
[16:16:17] <abaumann> aha. so that's the one.
[16:16:36] <abaumann> oom_reaper
[16:17:13] <deep42thought> is this a real thing?
[16:17:22] <abaumann> yes.
[16:17:45] <abaumann> some kernel daemon IIRC
[16:20:36] <deep42thought> maybe, -j0.5 would be a good option to have ...
[16:28:54] <abaumann> ..yes, especially because the concurrent test is still starting 3 threads..
[16:29:30] <abaumann> * abaumann packs his outdoor office and flees inside.. heavy thunderstorms ahead..
[16:29:32] -!- abaumann has quit [Quit: leaving]
[17:25:09] -!- deep42thought has quit [Quit: Leaving.]
[17:33:29] <buildmaster> deepin-session-ui is broken (says tyzoid-srv0-bs0).
[17:36:42] -!- bill-auger has quit [Quit: No Ping reply in 60 seconds.]
[17:46:59] -!- bill-auger has joined #archlinux32
[17:48:17] -!- bill-auger has quit [Client Quit]
[17:54:26] -!- bill-auger has joined #archlinux32
[19:28:10] -!- bill-auger has quit [Quit: No Ping reply in 60 seconds.]
[19:28:24] -!- bill-auger has joined #archlinux32
[20:23:24] -!- abaumann has joined #archlinux32
[20:23:25] <buildmaster> Hi abaumann!
[20:23:39] <abaumann> mmh.. gcc, binutils, glibc are all uploading for some time now..
[20:27:35] -!- deep42thought has joined #archlinux32
[20:27:36] <buildmaster> Hi deep42thought!
[20:27:44] <deep42thought> abaumann: probably I broke something
[20:27:47] <deep42thought> let me check ...
[20:29:09] <deep42thought> "I cannot determine, where this package (or some part of it) should be published"
[20:29:14] <deep42thought> yep :-D
[20:34:29] <deep42thought> it built the package with the same sub_pkgrel
[20:34:33] <deep42thought> which leads to errors
[20:38:23] <deep42thought> ah, no
[20:45:24] <deep42thought> hmm, git.archlinux32.org seems to take forever
[20:45:33] <deep42thought> might be ipv6 only, though
[20:46:31] <abaumann> oh..
[20:47:59] <deep42thought> I circumvented the ipv6 timeout by going over my ipv4-only machine :-s
[20:48:37] <deep42thought> I think I solved the return-assignment problem
[20:49:24] <deep42thought> it did not notice, that some binary_packages of the returned build_assignment were already built (as this was not intended in the past, but is used for the toolchain)
[20:49:39] <deep42thought> so it expected to receive gcc-8.2.0-2.0 and gcc-8.2.0-2.1
[20:49:47] <deep42thought> but it obviously only got the latter :-)
[20:51:29] <abaumann> good spotting it now :-)
[20:51:45] <deep42thought> well, it was unavoidable :-D
[20:55:05] <abaumann> buildknecht python trials 3..
[20:55:14] <abaumann> .. my patching seems not to be successful
[20:55:15] <deep42thought> not good
[21:26:26] <deep42thought> hmmm, how often is the buildmaster going to hand out gcc, glibc and binutils O.o
[21:35:45] -!- abaumann has quit [Quit: leaving]
[21:44:44] -!- deep42thought has quit [Quit: Leaving.]