IRC logs for #baserock for Wednesday, 2016-06-01

*** gtristan has quit IRC03:04
*** gtristan has joined #baserock03:31
*** gtristan has quit IRC05:25
*** gtristan has joined #baserock05:51
*** fay has joined #baserock07:09
*** fay is now known as Guest9173307:10
rjekI like your new name, fay07:11
*** rdale has joined #baserock07:25
Guest91733snappy isn't it :)07:35
*** Guest91733 is now known as faybrocklebank07:35
*** CTtpollard has joined #baserock07:44
*** edcragg has joined #baserock07:53
*** rdale has quit IRC08:26
*** jonathanmaw has joined #baserock08:30
*** rdale has joined #baserock08:37
*** edcragg has quit IRC08:40
*** toscalix has joined #baserock08:59
*** edcragg has joined #baserock09:01
*** toscalix has quit IRC09:08
*** toscalix has joined #baserock09:09
*** locallycompact has joined #baserock09:16
*** toscalix has quit IRC09:19
*** toscalix has joined #baserock09:21
*** franred has joined #baserock09:33
*** franred_ has joined #baserock11:08
*** franred has quit IRC11:11
*** franred has joined #baserock11:28
*** franred_ has quit IRC11:31
*** toscalix has quit IRC11:57
*** toscalix has joined #baserock12:34
*** franred_ has joined #baserock12:48
*** franred has quit IRC12:52
*** gtristan has quit IRC13:17
locallycompacthow does symlinking things in the target work in baserock? there's lines like13:25
locallycompactln -sf libncurses.so "$DESTDIR/${PREFIX}/lib/libcurses.so"13:25
locallycompactThat's nonsense isn't it?13:25
pedroalvarezwhy?13:27
richard_mawI remember this from linux-from-scratch, libncurses is ABI compatible with libcurses, and some stuff links against libcurses13:27
locallycompactI don't understand what this is doing. Isn't libncurses.so in the build directory? and so doesn't end up in the buildroot?13:28
locallycompactThen what is it symlinked to?13:28
locallycompactHow do I symlink say /usr/include/json-c to /usr/include/json, which is what I'm trying to do13:28
locallycompactin the target13:29
richard_mawln -sf json-c "$DESTDIR/${PREFIX}/include/json"13:29
richard_mawwait13:29
richard_mawmight have that the wrong way around, is the name it gets installed to by default json-c or json?13:29
locallycompactjson-c currently13:30
locallycompacti want json13:30
locallycompactto exist13:30
richard_mawok, so it is the other way around then: `ln -sf json "$DESTDIR/${PREFIX}/include/json-c"`13:30
pedroalvarezrichard_maw: I think you were right before13:30
richard_mawso I was13:31
richard_mawaaaa13:31
locallycompactno path on json-c, why?13:31
locallycompactwait13:31
* locallycompact confuses13:31
richard_mawln -s LINK_TARGET LINK_LOCATION13:31
locallycompactmy link target is $DESTDIR/usr/include/json-c surely?13:32
richard_mawif LINK_TARGET is a relative path, it is interpreted relative to LINK_LOCATION when you look up paths13:32
locallycompactohh13:32
locallycompactok13:32
* locallycompact tries13:32
pedroalvarezif you put "DESTDIR" in link_target, it's going to end up really broken13:33
leemingSuggestions for why a .meta file would throw the WARNING? #149 https://github.com/devcurmudgeon/ybd/blob/master/ybd/splitting.py14:04
leemingcauses AttributeError to be thrown on #23814:06
leemingsimple fix, will log it14:06
leemingsorry, *is there a* simple fix?14:06
pedroalvarezleeming: you could try to remove that 'except' and leave python to show you the error, and figure out why14:06
pedroalvarezmy guess is that yaml.safe_load(f) in #145 is failing14:07
leemingmalformed morph file? what generates the .meta ?14:07
edcraggthe build tool14:07
pedroalvarezleeming: a quick workaround would be to put a:14:09
pedroalvarezif not metadata:14:09
pedroalvarez    continue14:09
pedroalvarezafter #23714:09
SotKor return an empty dict instead of None on #150 I expect14:09
pedroalvarezSotK: empty dict will cause other errors, for example in #24914:10
leemingthis is on gitlab btw :\14:10
*** gtristan has joined #baserock14:11
SotKpedroalvarez, leeming: ah yeah, ignore me :)14:11
pedroalvarezleeming: the proper fix would be to check why is not valid yaml, and make sure the tool (ybd) generates it correctly14:11
SotKindeed14:11
leemingis there a stand alone yaml validator?14:12
persialeeming: There are several, but the better ones need to be fed a schema for the specific yaml to be validated.14:12
* pedroalvarez always uses webpages for that14:13
pedroalvarezlike http://www.yamllint.com/14:13
rjekDoes that llint yams?14:13
pedroalvarezNo.14:14
leemingok so error origin is incorrect parsing of ymal format?14:18
* pedroalvarez fails to parse that ^14:21
* pedroalvarez didn't understand what you meant by "this is on gitlab" either14:22
leemingsorry, builds are remote. I don't have direct access to the runners to modify the source for ybd for further investigation14:22
*** toscalix_ has joined #baserock14:23
pedroalvarezleeming: ah, I see14:23
leemingso there is clearly a bug here in the exception handling in YBD, but there is something that is causing it to happen, i.e. the origin14:23
pedroalvarezthen, you could try to reproduce the error locally14:23
pedroalvareznot sure that is the origin (thanks for clarifying), given that ybd is hiding the error and just dropping a warning14:24
*** toscalix has quit IRC14:24
leemingso the warning it throws is... "WARNING: problem loading metadata {'build-depends': [ .... ], .... }"14:25
pedroalvarezyes, but that warning happens because something in the lines 144-147 is raising an error14:27
pedroalvarezwhat error? we don't know14:27
pedroalvarezthat warning should be an error, and it should "raise" the exception instead of returning None14:28
SotKleeming: I guess since you don't have direct access you also can't get to the metadata file that is causing the error?14:28
locallycompactwhat's a metadata file and where does it live?14:29
SotKwhatever https://github.com/devcurmudgeon/ybd/blob/master/ybd/splitting.py#L153 would return for a given component14:29
pedroalvarezyou can also use Morph :)14:31
pedroalvarezalso, report the issue14:33
*** toscalix_ is now known as toscalix15:00
paulsherwoodleeming: i expect you could reproduce the problem locally, provided your local machine a) had ybd installed and b) was using the same cache-server and definitions15:21
locallycompactthe problem was a mismatched name in the stratum chunk description and the chunk file itself15:25
pedroalvarezfunky15:26
paulsherwoodlocallycompact: can you raise an issue with an example, so i can fix it?15:27
locallycompactpaulsherwood, leeming is on it15:28
paulsherwoodok great15:28
paulsherwoodeven better, a patch :)15:29
paulsherwoodleeming: ^^ :)15:29
leemingpaulsherwood, submitted an issue15:46
pedroalvareznice error15:48
*** toscalix has quit IRC16:11
*** CTtpollard has quit IRC16:16
*** toscalix has joined #baserock16:19
*** faybrocklebank has quit IRC16:28
*** toscalix has quit IRC16:29
*** jonathanmaw has quit IRC16:37
*** franred_ has quit IRC17:35
*** locallycompact has quit IRC17:36
*** rdale has quit IRC17:39
*** edcragg has quit IRC17:42
paulsherwoodnot so nice... there are quite a few examples in master definitions http://paste.baserock.org/nupejodego17:49
paulsherwoodleeming: master will now warn for that. if you want it to exit, set check-definitions: exit17:58
*** edcragg has joined #baserock18:36
*** edcragg has quit IRC19:46

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