IRC logs for #baserock for Monday, 2016-11-28

*** fay has joined #baserock08:07
*** fay is now known as Guest576208:08
*** ctbruce has joined #baserock08:29
*** franred has joined #baserock08:31
*** paulwaters_ has joined #baserock08:37
*** gtristan has quit IRC08:39
*** toscalix has joined #baserock08:57
*** locallycompact has joined #baserock09:18
* locallycompact wonders if today might be the day https://gitlab.com/baserock/ybd/merge_requests/26309:22
*** Guest5762 is now known as faybrocklebank09:23
*** anahuelamo has joined #baserock09:25
locallycompactsome lucky person could be the next one to click yes on this historic merge request https://gitlab.com/baserock/ybd/merge_requests/26309:50
*** rdale has joined #baserock09:52
paulsherwoodlocallycompact: the real purpose of this is for your proposed V10 changes, iiuc?09:55
locallycompactpaulsherwood, purpose of python 3 is improving comprehension so that fixes have a shorter turnaround / are intelligible09:57
locallycompactthe recursive submodules fix is waiting for this09:57
locallycompactand V1009:57
paulsherwoodfor future reference, smaller patches (and smaller series) are easier to review09:58
locallycompact3.5 only was a lot smaller, 3.4 didn't like circular imports09:59
locallycompactwhich is most of it09:59
paulsherwoodeven so... some of the component patches could have been trivially merged10:01
paulsherwood(without switching to 3 for example)10:01
paulsherwoodmerged. please make future changes as easy to grok as possible :)10:02
locallycompactactually they couldn't have been because python2 wouldn't have complained at the right things10:03
locallycompactpaulsherwood, thanks!10:04
locallycompacthere's another one https://gitlab.com/baserock/ybd/merge_requests/26510:04
pedroalvarezoh, that's the one i liked10:05
locallycompactlib has docstrings now10:06
* paulsherwood has no idea what it does, so declares himself unable to review10:06
paulsherwoodi like that it removes code, though10:06
pedroalvarezit's moving some logic to another library10:07
*** gtristan has joined #baserock10:08
paulsherwoodoh... so just *hiding* code then :)10:10
paulsherwoodlocallycompact: so net result of that is no-one has reviewed your 15 gitmachine commits, i expect?10:11
locallycompactis a lot more linear now10:11
paulsherwoodwhat does 'linear' mean?10:11
locallycompactnot recursing through three different _checkout/checkout/checkout_submodules functions10:12
locallycompactlibrary has independent test/ci for nested modules10:13
locallycompactand docs10:13
paulsherwoodack10:14
paulsherwoodwould you mind adding pep8 to its ci?10:14
pedroalvarezheh, we have also hidden the sandboxing code then10:15
locallycompactalright10:16
paulsherwoodlocallycompact: https://gitlab.com/baserock/ybd/pipelines/5162760 !10:18
locallycompactfatal: unable to connect to git.baserock.org:10:19
locallycompactgit.baserock.org[0: 185.43.218.183]: errno=No route to host10:19
* paulsherwood retries10:21
pedroalvarezhas that been around  10:11:48 ?10:22
*** CTtpollard has joined #baserock10:22
locallycompactyup10:22
pedroalvarezi think that our gitlab workers and the machine that has also seen the network failure are in the same network, so I can't tell where the problem is10:23
pedroalvarezbut it's gone10:24
*** tiagogomes has joined #baserock10:24
*** jude_ has quit IRC10:33
*** jude_ has joined #baserock10:48
*** brlogger has joined #baserock10:49
*** anahuelamo has quit IRC10:51
locallycompactgetting frequent g.b.o contact errors building locally10:56
*** franred has quit IRC11:00
*** franred has joined #baserock11:03
pedroalvarezI can see that too11:08
pedroalvarezvms that are in t he same network as g.b.o are not seeing any failure, so i expect these are network problems11:09
pedroalvarezyup, pinging the server i can see some packet loss11:10
pedroalvarezI wonder if gary_perkins knows anything about it11:11
pedroalvarezsame when pinging irclogs.baserock.org11:12
pedroalvarezI don't think this is g.b.o11:12
*** anahuelamo has joined #baserock11:16
*** brlogger has joined #baserock11:17
*** brlogger has joined #baserock11:20
*** anahuelamo has quit IRC11:20
pedroalvarezmail from DC, they are trying to fix these problems11:24
*** anahuelamo has joined #baserock11:34
pedroalvarezthey claim it has been fixed11:56
pedroalvarezlocallycompact: let us know if you see more problems like this11:56
locallycompacthttps://gitlab.com/baserock/definitions/merge_requests/1611:59
paulsherwoodlocallycompact: build failed12:05
paulsherwoodgtristan: do i understand correctly that you are not in favour of locallycompact's V10 changes?12:05
gtristanoh this12:24
*** jude_ has quit IRC12:25
gtristanpaulsherwood, I very much wanted include-mode and am using something similar; the concept of assemblages however present a difficult migration path; since they turn strata into something which is not by itself buildable (one needs to now tell a strata what to build on top of)12:26
locallycompactjust make a thing with build-essential and that strata in12:28
gtristanthis change makes depending on foreign pipelines difficult since one would need to tell that pipeline how to order its builds, and just breaks down when building more than 2 pipelines recursively - so instead we've done "variants" and allow depending on any "stack" (or element) from a foreign pipeline, or a given variant of that12:28
gtristanwe could migrate from an already plotted pipeline12:29
gtristanbut then if baserock grows definitions which stack strata in different ways, migrating a given set of definitions wholesale is tricky12:29
*** jude_ has joined #baserock12:40
paulsherwoodlocallycompact: i may be mistaken, the above looks to me like a -1 for V1012:53
locallycompactgtristan, sorry, downstream needs to be able to pull in strata from upstream and be able to swap out sysroots and we need this pretty much now. I will migrate any data you want.12:58
*** locallycompact has quit IRC13:23
*** locallycompact has joined #baserock13:28
*** leeming has quit IRC13:33
*** locallycompact has quit IRC13:37
*** locallycompact has joined #baserock13:43
*** locallycompact has quit IRC13:52
*** locallycompact has joined #baserock13:57
*** brlogger has joined #baserock13:58
*** toscalix has quit IRC14:05
*** toscalix has joined #baserock14:14
*** franred has quit IRC14:37
locallycompactpedroalvarez, still getting connection problems14:38
pedroalvarezYeah, me too14:40
*** tiagogomes has quit IRC14:41
pedroalvarezthey just said they are looking at it14:42
radiofreelocallycompact: do you have any idea why https://gitlab.com/baserock/definitions/merge_requests/16/pipelines is failing?14:51
radiofreeseems like a perfectly fine change14:51
radiofreegbo problems?14:51
locallycompactsome of it was typos some of it was gbo14:51
radiofreeah14:52
locallycompactoh no14:52
locallycompacthttps://gitlab.com/baserock/definitions/builds/673613814:52
locallycompactnot sure14:52
*** tiagogomes has joined #baserock14:56
*** jude_ has quit IRC15:00
radiofreeERROR: Failed to install split components  :'(15:09
locallycompactIt's some unicode thing on python 315:09
locallycompacto such file or directory: b'/root/ybd/tmp/tmp2yf60da1/usr/share/ca-certificates/mozilla/T\xc3\x9cB\xc4\xb0TAK_UEKAE_K\xc3\xb6k_Sertifika_Hizmet_Sa\xc4\x9flay\xc4\x15:09
locallycompactit's like that in the metafile15:15
locallycompactfor ca-certificates15:15
pedroalvarezUGH15:20
pedroalvarezugh*15:20
* paulsherwood wonders if he'll live to regret approving the move to p315:21
locallycompactHow do you get around this15:55
*** tardyp has quit IRC16:01
*** tardyp has joined #baserock16:08
*** leeming has joined #baserock16:08
locallycompactIt says No such file or directory /root/ybd/tmp/tmpdo2wp38a/usr/share/ca-certificates/mozilla/TÜBİTAK_UEKAE_Kök_Sertifika_Hizmet_Sağlayıcısı_-_Sürüm_3.crt16:16
locallycompactBut the one in the artifact got truncated at TÜBİTAK_UEKAE_Kök_Sertifika_Hizmet_Sağlayıcısı_-_Sürüm_16:17
locallycompactthe meta file gives the full name but the file missing a 3.crt16:17
* paulsherwood doesn't understand the problem properly. previously leeming helped with unicode16:32
rjekIs Python still fundamentally broken when it comes to paths?16:33
rjek(ie, it assumes all valid paths can be encoded to valid Unicode)16:33
paulsherwoodrjek: are you having one of those days where everything is impossible again?16:33
rjekI'm asking if it's still the case16:34
rjekPython /used/ to be fundamentally broken in this respect16:34
* paulsherwood doesn't know... leeming ?16:34
*** leeming has quit IRC16:35
rjekPerfect timing :D16:35
paulsherwood?16:35
locallycompactOk so16:35
persiarjek: python is so fundamentally broken, but there are a number of known workarounds.16:35
persiaAs  simple check, what locale applies when the filename is truncated?16:35
locallycompactin the sandbox dn['install'] it has the correct file name TÜBİTAK_UEKAE_Kök_Sertifika_Hizmet_Sağlayıcısı_-_Sürüm_3.crt16:36
locallycompactthen it calls utils.make_deterministic_gztar_archive(cachefile, dn['install'])16:36
locallycompactand in the tarball it has a truncated filename16:36
paulsherwoodcould it be path truncation?16:37
paulsherwoodie the actual length, not unicode?16:37
locallycompactthere are longer filenames in the same folder16:37
persiaMore bytes, or more codepoints?16:37
locallycompactah maybe yeah16:38
locallycompactstill though why16:39
locallycompactit's definitely tar16:48
locallycompactf_tar.add(name=name, arcname=arcname, recursive=False)16:48
locallycompactarcname at that point is fine16:49
*** jonathanmaw has joined #baserock16:56
locallycompactmakes no sense17:19
*** ctbruce has quit IRC17:26
*** toscalix has quit IRC17:47
*** gtristan has quit IRC18:06
*** jonathanmaw has quit IRC18:20
*** anahuelamo has quit IRC18:33
*** jude_ has joined #baserock20:35
*** locallycompact has quit IRC21:25
*** jude_ has quit IRC23:03
*** gtristan has joined #baserock23:23

Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!