IRC logs for #baserock for Wednesday, 2016-03-09

*** gtristan has quit IRC01:27
*** edcragg has quit IRC01:27
*** gtristan has joined #baserock01:40
*** gtristan has quit IRC02:19
*** gtristan has joined #baserock02:35
*** vgrade has quit IRC04:51
*** vgrade has joined #baserock04:52
*** gtristan has quit IRC07:16
*** franred has joined #baserock08:09
*** edcragg has joined #baserock08:17
*** gtristan has joined #baserock08:36
*** gtristan has quit IRC08:39
*** gtristan has joined #baserock08:39
*** ctbruce has joined #baserock08:45
*** jonathanmaw has joined #baserock09:04
*** bashrc has joined #baserock09:04
*** wdutch has left #baserock09:05
gtristanWhat is the right way to clone a repo with it's submodules from g.b.o ?09:24
gtristangit clone <module> && cd <module> && git submodule <smth-I-dont-know-about> ?09:25
rjekIsn't there a recursive flag?09:26
CTtpollardgtristan: --recursive should checkout and each module09:26
CTtpollards/and/09:26
rjek--recurse-submodules or something?09:26
gtristanmmmmyeah09:27
franredor if you have already clone the repo you could do "git submodule update"09:27
gtristangit clone --recursive works... for some definition of works... it clones the submodules from some arbitrary version of the .gitmodules file... possibly master09:28
franredummm although I think you need "git submodule init" beforehand09:28
gtristangit submodule update didnt do anything09:28
* gtristan tries git submodule init...09:28
franredgtristan, what about git clone && cd && git submodule init && git submodule update09:28
gtristanok so is "upstream:<name>" a concept proper to baserock ?09:29
* gtristan was under the impression that was a git feature09:29
franredgtristan, yep, it is baserock related09:30
gtristanI see09:31
gtristanok... modifying the .gitmodules, git submodule init && git submodule update... does it09:32
CTtpollardinit then update works, but --recursive clone is a cleaner method imo09:32
pedroalvarezIME "it clones the submodules from some arbitrary version of the .gitmodules file" is not true09:33
franredCTtpollard, the problem that gtristan is having is that git doesn't understand "upstream:" in .gitmodules so it won't clone recursively09:33
franredI think^^09:33
gtristannod09:33
gtristanworked around09:34
gtristanok this is odd...10:05
gtristanpedroalvarez, 2708be3df0ecafc70046212c627ede87bd3e6487 <-- fix migration for version 710:07
gtristanthat's been a while actually10:07
gtristannow... I'm just raising this because... it's odd :-/10:08
pedroalvarezhey! :)10:08
pedroalvarezwhat exactly is odd?10:08
gtristangood morning :D10:08
gtristanSo what I am seeing... I am building libtool from core.morph... not libtool-tarball10:09
gtristanand libtool does a pre-configure-commands ./bootstrap10:10
gtristanthen... ybd does ./bootstrap again10:10
gtristanand it fails10:10
gtristanbut... if I do the whole thing in the checkout dir, with just ./bootstrap and then ./configure --prefix=/usr... it doesnt fail10:10
gtristansomething is fishy10:11
*** bashrc has quit IRC10:11
*** edcragg has quit IRC10:11
gtristanlet me try again after a git clean -xdf10:11
gtristanpedroalvarez, possibly a ybd bug in it's interpretation of what is in DEFAULTS ?10:12
gtristanlike... the suspicious looking '>-' ?10:13
gtristanunder autotools: configure-commands: ?10:13
*** bashrc has joined #baserock10:13
*** ssam2 has joined #baserock10:16
*** ChanServ sets mode: +v ssam210:16
gtristanIn DEFAULTS... what is the meaning of '>-' ?10:18
richard_mawit's a yaml thing10:18
richard_mawit changes how newlines are interpreted I think10:18
richard_mawwell, I know it does10:19
richard_mawI just can't remember which particular mode >- means10:19
richard_mawI think it's fold newlines and chomp off the trailing newline10:19
gtristanhmmm10:21
*** ssam2 has quit IRC10:25
pedroalvarezsorry, distracted by other things10:26
*** ssam2 has joined #baserock10:27
*** ChanServ sets mode: +v ssam210:27
gtristanpedroalvarez, yeah well... it looks like with that change we should at least be able to remove libtool.morph from core10:28
gtristanthat at least is a given, it otherwise runs bootstrap twice10:28
ssam2mmm, killing libtool10:28
gtristanyeah, wouldnt that be nice10:28
gtristanman10:28
gtristanthis is very strange10:28
gtristanso if I run the build as ybd does it... running ./bootstrap first10:29
gtristanand then running:10:29
gtristanexport NOCONFIGURE=1; if [ -e autogen ]; then ./autogen; elif [ -e autogen.sh ]; then ./autogen.sh; elif [ -e bootstrap ]; then ./bootstrap; elif [ -e bootstrap.sh ]; then ./bootstrap.sh; elif [ ! -e ./configure ]; then autoreconf -ivf; fi10:29
gtristanthe first time bootstrap gets it right10:29
gtristanthe second time is destructive10:29
gtristanand errors with:10:29
*** edcragg has joined #baserock10:29
gtristanconfigure.ac:152: error: required file 'build-aux/ltmain.sh' not found10:29
gtristanautoreconf: automake failed with exit status: 110:29
gtristanBut... seeing that pedroalvarez's commit is from december, I have to imagine that libtool was built with ybd dozens of times since then10:30
gtristanso I have to wonder... why does it break ?10:30
ssam2it's not the tarball cache is it?10:31
ssam2chunk hardlink cache thingy?10:32
ssam2not sure why it would be, but that's something that can really cause confusing errors10:32
ssam2if the write-protection over the cache of unpacked chunk files isn't working for some reason, maybe the first build modifies the actual unpacked files, and then the second build goes haywire because of that10:32
gtristanssam2, it could be cached10:32
gtristanssam2, one way it could be cached, is iirc ybd used to sport it's own DEFAULTS file10:33
rdaleit still has its own defaults10:33
gtristanso maybe... version 7 happened... and ybd happily built libtool and cached the artifact ignoring definitions DEFAULTS10:34
gtristanuntil one day ybd was fixed10:34
gtristannot effecting the cache key of the cached libtool artifact10:34
gtristanrdale, you have a running ybd right ?10:36
gtristanrdale, care to rebuild libtool with a recent version and see if it works there ?10:37
rdaleyes - i haven't looked at the defaults code for a few days - i don't know if it has changed10:37
gtristanI have rebased ybd as of... last week anyway10:37
rdaleok, i'll try building libtool - but that is part of build-essential i assume and i've never had problems with it before10:38
gtristanit's part of core10:38
gtristannot build-essential10:39
rdaleah ok10:39
rdalei think ybd first loads its defaults, and then it load DEFAULTS on top of that - i'm not sure what the end result actually is. it doesn't only load its own defaults if it can't find DEFAULT10:39
pedroalvarezgtristan: morph build log for libtool: http://paste.baserock.org/rilabagafo10:41
gtristanpedroalvarez, I see :-/10:42
gtristanmkay10:42
gtristanso it succeeds where I fail10:42
pedroalvarezand I agree, those preconfiure-commands are redundant10:42
gtristanalbeit, it does so at the cost of double bootstrap, but it does "GEN      build-aux/ltmain.sh" both times around10:43
gtristanmine does not10:43
*** bashrc has quit IRC10:48
*** bashrc has joined #baserock10:53
rdalethis is a libtool log from one i've just built with ybd: http://paste.baserock.org/ebafuwilez11:02
rdaleref: 5e11a17c39a2a976a09c33a568431a9f7ad00d4811:04
gtristanrdale, thanks11:06
gtristanso it would seem that doing the double ./bootstrap on the virtfs mount gets confused... possibly due to symlinks11:07
*** bashrc has quit IRC11:45
*** gtristan has quit IRC11:45
*** bashrc has joined #baserock11:49
*** gtristan has joined #baserock12:03
*** ssam2_ has joined #baserock12:26
*** ssam2 has quit IRC12:27
jjardonHi, Im getting this error in morph when adding a new command in DEFAULTS: "ERROR: Invalid DEFAULTS file: Additional properties are not allowed ('bootstrap-commands' was unexpected)" ybd works fine; should i report a bug?12:54
ssam2_when you say "ybd works fine", what do you mean?12:58
pedroalvarezmorph drops that error when validating with the schema13:03
*** ssam2_ has quit IRC13:06
pedroalvarezI kind of hope that a simple change in 'morphlib/schemas/defaults.json-schema' will fix that error13:08
pedroalvarezalthough...13:08
*** ssam2 has joined #baserock13:08
*** ChanServ sets mode: +v ssam213:08
pedroalvarezwell yes, it will fix that error, but you need to do more changes in morph to support those new bootstrap-commands13:09
jjardonssam2: seems to start building the chunks with no problems13:09
ssam2have you given it any bootstrap-commands  ? does it execute them at the time you expect them to?13:10
ssam2I think the bug is that YBD doesn't validate DEFAULTS according to the json-schema for it13:10
* pedroalvarez nods13:11
jjardonoh, seems so; ybd simply ignore the new commands13:13
jjardonso, what would be the steps to add a new command step in DEFAULTS?13:13
ssam2propose it as a change to the spec13:14
ssam2I think we agreed 'bootstrap-commands' was a confusing name though because it has lots of different meanings13:14
ssam2call it prepare-commands or something, i forget13:14
pedroalvarezyeah, I've been trying to find irc logs for of that discussion, but I couldn't find any13:15
ssam2might have taken place in Koffee Pot or something ;-)13:15
jjardonhaha13:15
ssam2I guess there should be a detailed guide on how to propose changes to the spec, but the idea is make a patch to spec.git, basically13:21
ssam2updating spec.md and any json-schemas in in schemas/ dir should be enough13:21
edcragglooks like a useful document14:04
*** toscalix has joined #baserock14:35
*** CTtpollard has quit IRC15:53
*** jonathanmaw_ has joined #baserock15:54
*** jonathanmaw_ has quit IRC15:54
*** jonathanmaw_ has joined #baserock15:55
*** CTtpollard has joined #baserock15:56
*** jonathanmaw has quit IRC15:59
*** CTtpollard has quit IRC17:02
*** jonathanmaw_ has quit IRC17:02
*** ctbruce has quit IRC17:19
*** toscalix has quit IRC17:39
*** ssam2 has quit IRC17:45
*** bashrc has quit IRC18:00
*** franred has quit IRC18:00
*** edcragg has quit IRC18:28
*** toscalix has joined #baserock18:49
*** rdale_ct has joined #baserock19:18
*** rdale has quit IRC19:21
*** rdale has joined #baserock19:24
*** rdale_ct has quit IRC19:27
*** rdale_ct has joined #baserock19:36
*** rdale_ct_ has joined #baserock19:37
*** rdale has quit IRC19:38
*** rdale_ct has quit IRC19:40
*** rdale has joined #baserock19:46
*** edcragg has joined #baserock19:46
*** rdale_ct_ has quit IRC19:47
*** rdale_ct has joined #baserock20:07
*** rdale has quit IRC20:08
*** rdale has joined #baserock20:08
*** rdale_ct has quit IRC20:09
*** rdale_ct has joined #baserock20:13
*** rdale has quit IRC20:13
*** rdale has joined #baserock20:21
*** rdale_ct has quit IRC20:21
*** rdale_ct has joined #baserock20:23
*** rdale_ct_ has joined #baserock20:25
*** rdale has quit IRC20:26
*** rdale has joined #baserock20:28
*** rdale_ct has quit IRC20:29
*** rdale_ct has joined #baserock20:31
*** rdale has quit IRC20:31
*** rdale_ct_ has quit IRC20:31
*** rdale has joined #baserock20:33
*** rdale_ct has quit IRC20:35
*** toscalix has quit IRC21:26
paulsherwoodplease could folks raise issues on ybd... otherwise i may miss them in discussions like those above21:31
*** cosm has quit IRC22:39
*** cosm has joined #baserock22:56
*** edcragg has quit IRC23:08
*** edcragg has joined #baserock23:09
*** edcragg has quit IRC23:16
*** cosm has quit IRC23:16
*** edcragg has joined #baserock23:17
*** cosm has joined #baserock23:31

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