IRC logs for #baserock for Thursday, 2018-05-31

*** rdale has quit IRC01:30
*** toscalix has joined #baserock07:54
*** rdale has joined #baserock08:13
ironfoot:( upgrading bst broke the build?08:20
*** jonathanmaw has joined #baserock08:20
jjardonironfoot: trying to use newer bst seems to break the build yes; seems a regression between 1.1.2 and 1.1.308:30
* paulsherwood wonders how bst does its testgin08:30
jjardonI have opened to track this08:32
jjardonpaulsherwood: bst have tests they run with
*** jonathanmaw has quit IRC09:08
*** jonathanmaw has joined #baserock09:21
juergbiI consider this a Baserock issue, see #414, but we can discuss this further09:28
ironfootjuergbi: thanks for the explanation. Yes, it definitely looks like a definitions issue09:36
noisecellironfoot, I've been looking at the STAGE2_SYSROOT in definitions history, and we were setting it to $(dirname $(pwd)) very early on. jjardon, do we have a way to check that change in YBD (if we are currently still maintaining definitions for this build tool) as well as buildstream10:02
ironfootmaster of definitions shouldn't support any other than buildstream10:03
ironfootIIRC there are no .morph files in there10:04
ironfootWhat I mean is that we shouldn't need to test anything else when fixing this issue for buildstream10:05
ironfootfor the new version of buildstream*10:06
noisecellironfoot, ok. is buildstream backwards compatible? e.g. if we change this in definitions 1.0.1 should also work?10:07
ironfootA fix for this should work in 1.0 and in 1.1.3 from my point of view10:07
ironfootif not, then maybe is not a good fix10:07
noisecellif jjardon is not faster than me I might create a patch for it and see if it works for current version, and from them we can check also in jjardon's branches10:12
noisecellironfoot, ^^10:12
jjardonthere is a branch that point to the latest definitions based on .morphs; the branch name is "morphs"10:14
jjardonIt's what we use to check the conversion script keeps working10:14
noisecelljjardon, ok, I can branch out from that branch10:16
ironfootThis is part of the stack might not be part of the converted morphs, but probably we need to fix something in that branch too10:17
noisecellironfoot, you might be right, I cans see bsts in that branch for the gnu-toolchain, etc which probably need to be changed too10:25
noisecellin any case I will do in that branch and submit a MR using master?10:26
noisecellheh, I have more issues, when building the minimal system -- when running the commands in the initramfs-gz.bst. we set init and shutdown with executable permissions, but in my build for some reason init is under /usr/bin and not /sbin and shutdown doesn't seems to be installed in the chroot :/11:09
noisecellNOTE: this is a separated issue from the previous question11:09
noisecellIm building with bst 1.0.1 and in a Fedora-28 VM11:13
*** radiofree has joined #baserock12:21
*** toscalix has quit IRC14:24
*** toscalix has joined #baserock14:25
noisecellI can confirm this fails also in a debian VM for the same reason14:47
noisecellok, I think I found what is going on. -- the init and shutdown mention here: should be installed using: which is called:
* noisecell reads logs15:00
ironfootgah, I was going to test this, and then forgot about it15:00
ironfootmany things on my plate today15:00
paulsherwoodnoisecell: can you fix it?15:00
noisecellpaulsherwood, I am looking into it15:01
*** jonathanmaw has quit IRC16:50
*** toscalix has quit IRC17:09
*** rdale has quit IRC20:42

Generated by 2.15.3 by Marius Gedminas - find it at!