IRC logs for #baserock for Monday, 2015-06-22

*** edcragg has quit IRC00:37
*** rdale has joined #baserock02:47
*** petefoth has quit IRC05:08
*** mike has joined #baserock05:41
*** mike is now known as Guest2360905:41
*** petefoth has joined #baserock06:24
*** perryl_ is now known as perryl07:27
*** edcragg has joined #baserock07:35
*** petefoth_ has joined #baserock07:52
*** petefoth has quit IRC07:53
*** petefoth_ is now known as petefoth07:53
*** bashrc_ has joined #baserock08:05
*** edcragg has quit IRC08:10
*** tiagogomes_ has joined #baserock08:12
*** petefoth has quit IRC08:15
*** gary_perkins has joined #baserock08:16
*** petefoth has joined #baserock08:18
*** jonathanmaw has joined #baserock08:32
*** tiagogomes_ has quit IRC08:39
*** tiagogomes_ has joined #baserock08:39
*** pacon has joined #baserock08:54
*** ssam2 has joined #baserock09:07
*** ChanServ sets mode: +v ssam209:07
* pedroalvarez upgrades masons morph version to the one used in the latest release09:10
pedroalvarezso that they can understand definitions version 509:11
*** CTtpollard has joined #baserock09:17
*** lachlanmackenzie has joined #baserock09:20
*** mariaderidder has joined #baserock09:21
ssam2the YBD continuous builder I set up has now done 18 identical builds of some chunks...09:32
ssam2it's also done 21 non identical builds of other chunks, but it's a start09:32
*** pacon has quit IRC10:12
pedroalvarezhi! I had problems to locate this page
pedroalvarezshould we make it more visible? or is just me.. :)10:13
petefothThat page shuold probably be linked (conspicuously) from the front page10:13
pedroalvarezlooks like it is10:14
ssam2let's make it more visible10:14
pedroalvarezI was looking for a link in the "contribute" page10:14
*** pacon has joined #baserock10:14
ZaraI'd link it where we link to the 'developer experience' bit, since I think the content of 'doing stuff with baserock' is a lot more interesting to developers.10:16
*** persia has quit IRC10:43
*** persia has joined #baserock10:43
*** persia has joined #baserock10:43
*** rdale has quit IRC10:57
*** rdale has joined #baserock10:58
*** zoli__ has joined #baserock11:15
*** zoli___ has joined #baserock11:16
*** zoli__ has quit IRC11:19
*** paulw has quit IRC11:21
*** paulw has joined #baserock11:22
pedroalvarezpaulsher1ood:  re your problem11:26
pedroalvarez<paulsher1ood> pedroalvarez: if i have lots of tmp files around, they need to be mounted readonly during sandboxing. the issue is that there's a hard limit of 5011:26
pedroalvarezLooks like it was a linux-user-chroot limitation11:26
pedroalvarezAlso, it looks like there wasn't any release of linux-user-chroot with that fix in11:27
paulsher1oodpedroalvarez: ah, ok. good to know11:28
pedroalvarezSo I assume that Ubuntu's linux-user-chroot doesn't have the limitiation change11:28
pedroalvarezthis change happened on 2013, though11:29
*** zoli___ has quit IRC11:31
*** zoli__ has joined #baserock11:32
ssam2I believe we sent the patch11:34
ssam2to increase the mount limit in linux-user-chroot11:34
*** jonathanmaw has quit IRC11:48
*** jonathanmaw has joined #baserock11:52
richard_mawthe patch was applied upstream IIRC11:57
pedroalvarezyup, the issue here is that ubuntu doesn't have it11:59
*** Guest23609 has quit IRC12:30
pdar_has any one who can merge things a few moments to merge this please?
*** pdar_ is now known as pdar12:33
* SotK looks at it12:34
pdarta SotK12:34
SotKpdar: done12:35
*** Guest23609 has joined #baserock12:44
tiagogomes_did anyone looked at the possibility of using llvm in build essential? It needs python to build...12:47
richard_mawtiagogomes_: not to my knowledge12:48
* pedroalvarez builds from a definitions (git) checkout 13:00
pedroalvarezthis is great13:00
*** sambishop has quit IRC13:05
*** sambishop has joined #baserock13:06
ssam2del, ctrl-left and ctrl-right finally work in Baserock 15.25 !13:30
ssam2thanks jjardon !13:30
*** edcragg has joined #baserock13:39
*** pacon has quit IRC13:42
*** jonathanmaw has quit IRC14:15
nowsterleft right left right up up down down morph build?14:25
*** jonathanmaw has joined #baserock14:28
paulsher1oodtiagogomes_: you mean instead of gcc, or earlier in stage1/stage2?14:59
*** jonathanmaw has quit IRC14:59
paulsher1ood2015-06-22 03:55:45 [build-system-armv8l64] Now cached as build-system-armv8l64.3e7aa37c542cd707cb1dbe221b93d6e0bbdc08a996d113cb6863ab278149ca5c14:59
paulsher1ood(built on Ubuntu, on Moonshot)15:00
tiagogomes_paulsher1ood isn't instead of gcc the same as earlier in stage1/stage2?15:03
tiagogomes_anyway building from scratch using llvm is not feasible without adding a bunch of chunks to build-essential15:04
paulsher1oodtiagogomes_: there is stage1-gcc, stage2-gcc, gcc. i think trying to have llvm in stage1/stage2 would be overkill but i may be wrong15:05
tiagogomes_paulsher1ood it is. llvm needs python to build...15:06
*** jonathanmaw has joined #baserock15:06
paulsher1oodif it's just that, i wouldn't see it as a barrier15:07
paulsher1oodbut again i may be wrong15:07
pedroalvarezdoes it really need to go in build-essential>?15:08
tiagogomes_paulsher1ood you would have to move a few chunks to build essential, which is very bad15:08
tiagogomes_pedroalvarez depends. If you want all stage 3 chunks built with llvm from scratch... yes15:10
tiagogomes_anyway, last time that I heard (FOSDEM) Linux was still not building with llvm15:13
pedroalvarezI see15:13
paulsher1oodi wonder if we could end up with build-essential-gcc, and build-essential-llvm etc?15:18
paulsher1oodand separate stage1 and stage2 in to bootstrap-essential15:18
tiagogomes_I thought that was the plan, having different build essential for each combination of compiler and libc15:19
* paulsher1ood wonders where the 'plan' is :)15:19
rdalei've currently got an openwrt and openwrt-musl build essentials15:20
tiagogomes_I am getting this error when I try to deploy: `ImportError: No module named writeexts`. Solutions?15:21
pedroalvareztiagogomes_: so, it needs python to build. but once built, does it need python to run?15:21
paulsher1oodrdale:but are they branches of build-essential, or separately named?15:21
paulsher1oodtiagogomes_: using latest morph? or ybd?15:21
rdalecurrently they are both called build-essential in two different branches, but i'm thinking how to improve that15:22
tiagogomes_Using latest morph15:22
paulsher1oodSotK: ^^?15:22
SotKtiagogomes_: can you paste the traceback that is probably in morph.log?15:22
SotKtiagogomes_: also, what deployment extension are you using?15:22
tiagogomes_SotK ok, I was using and old cluster that wasn't updated by prepending 'extensions' on 'type'15:27
tiagogomes_but know I've another error:
paulsher1oodrdale: i'd suggest we go with build-essential-variant (where variant for you would be musl)?15:28
SotKtiagogomes_: that is definitely my bad :/15:29
rdalepaulsher1ood: yes, that's what i was thinking. then we'll need a little script to fix up the actual build-essential name in the strata and system before building15:29
* SotK is sure he tested kvm deployment :s15:30
paulsher1oodrdale: erk.... that needs a *bit* more thought :)15:30
paulsher1oodtime to try to get to the bottom of this semantics and versioning pit, i think15:31
tiagogomes_SotK let me double check that I've the cluster right15:31
SotKtiagogomes_: it's because the line raising the error should say `except subprocess.CalledProcessError` I'd imagine, I'll send a patch15:32
tiagogomes_SotK, ok15:34
SotKtiagogomes_: on looking closer, it should be a writeexts.ExtensionError15:35
SotKI must have tested kvm before I reworked the ssh_runcmd stuff and not after :/15:35
paulsher1oodthat's interesting... how do we test deployments in general?15:36
paulsher1ooddid/do morph's tests cover it?15:36
SotKmorph's tests still use the extensions which are in morphlib (which are now deprecated)15:37
paulsher1oodok so this exposes an interesting conundrum... testing of definitions. and if definitions include deployments, then testing of those in definitions15:38
SotKI hope that the Zuul CI will eventually test these kinds of situations15:38
paulsher1oodSotK: irrespective of which ci, how did/do deployments in morph actually get tested?15:39
SotKpaulsher1ood: for this, I did a load of test deployments15:40
SotKpaulsher1ood: in general, there are scenario tests in morph's test suite for deployment, but they deploy using the extensions in morphlib rather than definitions and so need updating somehow15:40
ssam2bit sticky, actually15:40
ssam2previously we've always tried to keep Morph's tests independent from the reference definitions15:41
paulsher1oodit seems a bit hard to do, for me. i know how to test builds.15:41
ssam2the tests only test quite a subset of deployment, in anycase. I think just rawdisk and tar15:41
paulsher1oodmorph's tests are unit tests afaict15:41
SotKpaulsher1ood: the yarns are scenario tests15:42
paulsher1oodoh, ok. i haven't fully understood the difference, yet15:42
SotKssam2: that sounds about right15:43
ssam2I think the solution is to remove most of the deployment tests from Morph's test suite, and just keep one trivial .write extension in morph.git for test purposes (like tar.write) to provide a quick way of checking the `morph deploy` code path15:43
ssam2the deployment extensions now belong to the reference systems, so they should be tested in whatever way we test the reference systems15:44
ssam2Mason deploys to disk images, so in fact we don't really lose anything in terms of test coverage15:44
*** cosm has joined #baserock15:45
paulsher1oodSotK: +115:45
tiagogomes_thanks SotK15:48
tiagogomes_I got another error now: NameError: global name 'subprocess' is not defined15:48
tiagogomes_you missed an import subprocess on kvm.write15:49
SotKargh, sorry :(15:50
paulsher1oodSotK: not to worry - easy to diagnose, fast to fix at least15:51
* SotK updates the change15:54
persiaOn supporting multiple libcs: I think that we should have multiple strata/systems that depend differently on different libs (and not every strata needs to support every libc), so build-essential-variant seems right to me.15:56
tiagogomes_I still can't deploy15:56
persiaI would expect the reference development system to always use glibc, but other systems might be better with musl, uclibc, etc.15:57
SotKtiagogomes_: what is the error now?15:57
rdaleanother reason to have different build-essentials, apart from libc's, is when you need different busybox config options to the default ones15:58
SotKtiagogomes_: looks like I really broke something when changing that ssh code, I'll investigate16:00
ssam2'bufsize must be an integer' from subprocess.Popen is usually to do with passing wrong arguments16:03
ssam2bufsize is the first keyword arg16:03
SotKaha, I call check_call wrongly in kvm.write at one part16:05
* SotK updates the change again16:16
tiagogomes_SotK I'll tell you what the next error is :)16:17
*** jonathanmaw has quit IRC16:17
SotKoops, forgot to commit :316:17
tiagogomes_that was the next error16:17
SotKOK, updated16:18
*** gary_perkins has quit IRC16:18
ssam2seems the Debian reproducible builds project is getting some funding from the Linux Foundation!
tiagogomes_2015-06-22 16:19:56 [systems/build-system-x86_64.morph][release]Creating btrfs filesystem16:20
tiagogomes_ERROR: device scan failed '/src/tmp/deployments/tmp3a9596/tmpokM3ph/tmpzTmGb_' - Block device required16:20
tiagogomes_is still deploying though16:20
tiagogomes_sh: root: not found16:26
tiagogomes_clang works16:27
SotKtiagogomes_: did the deployment work correctly then?16:30
tiagogomes_SotK yes, it succeed16:31
tiagogomes_but that "device scan failed" makes me apprehensive16:31
tiagogomes_can some review ? I submitted that patch assuming that there is interest in having clang in devtools16:34
SotKtiagogomes_: I think that has started showing up (along with some other output elsewhere) because we no longer swallow the output, but I don't know if that error is a problem or not16:37
perrylpaulsher1ood: just sent a pull request for YBD fixing up some issues i had missed earlier, sorry i didn't pick them up sooner!16:37
paulsher1oodperryl: merged! thanks!16:55
*** ssam2 has quit IRC16:59
*** bashrc_ has quit IRC17:01
*** Guest23609 has quit IRC17:02
*** mariaderidder has quit IRC17:04
*** lachlanmackenzie has quit IRC18:53
*** zoli__ has quit IRC19:58
*** cosm_ has joined #baserock20:19
*** franred has joined #baserock20:23
*** cosm has quit IRC20:23
*** franred has quit IRC20:37
*** cosm_ has quit IRC23:32

Generated by 2.15.3 by Marius Gedminas - find it at!