IRC logs for #baserock for Friday, 2017-09-29

*** gtristan has joined #baserock06:14
*** ctbruce has joined #baserock07:30
*** toscalix has joined #baserock07:39
*** locallycompact has joined #baserock07:56
*** rdale has joined #baserock08:12
*** jonathanmaw has joined #baserock08:52
*** ssam2 has joined #baserock08:57
*** ChanServ sets mode: +v ssam208:57
*** locallycompact has quit IRC09:10
ssam2jonathanmaw, so actually the way I set up the access control for ostree.baserock.org is that there is a specific set of people who are allowed to push over ssh09:39
ssam2https://gitlab.com/baserock/infrastructure/blob/master/baserock_ostree/ostree-access-config.yml09:39
ssam2probably best if you can send me an ssh public key for you and i'll add that to the list09:39
ssam2then you can push your armv8l64 artifacts09:39
jonathanmawssam2: is sending it via PM ok?09:40
ssam2it would be nicer to avoid giving people access and gate everything through the CI, but we clearly don't have time to set up a gitlab-ci runner on any of our arm servers right now09:40
*** rdale has quit IRC09:40
ssam2PM is fine yeah09:40
*** rdale has joined #baserock09:45
gitlab-br-botinfrastructure: merge request (sam/jonathanmaw-ostree-access->master: Give Jonathan Maw push access to ostree.baserock.org) #7 changed state ("opened"): https://gitlab.com/baserock/infrastructure/merge_requests/709:46
ssam2ironfoot, gary_perkins ... ok with ^^ ?09:47
ironfootfor one second i thought i was looking at bitbucket09:47
ssam2the ui change in gitlab? :-)09:48
gitlab-br-botinfrastructure: merge request (sam/jonathanmaw-ostree-access->master: Give Jonathan Maw push access to ostree.baserock.org) #7 changed state ("merged"): https://gitlab.com/baserock/infrastructure/merge_requests/709:48
paulsherwoodyup... takes a bit of getting used to09:48
jonathanmawssam2: ok, how do I ostree push? iirc it's an external tool, rather than an ostree builtin09:52
ssam2one sec while i deploy the change09:53
gtristanjonathanmaw, ostree-push is not *really* a thing09:53
gtristanah09:54
ssam2we'll set up buildstream.conf so that it can push09:54
* gtristan reads backlog and suspects ssam2 has setup ostree-push somewhere09:54
ssam2jonathanmaw, put this in your ~/.config/buildstream.conf : http://paste.baserock.org/osokowuhuj09:55
ssam2that will cause buildstream to push everything it builds to the cache09:55
ssam2or fail, if your SSH key isn't available09:55
ssam2(i've been using ssh agent forwarding to make the key available, and have sometimes left builds running overnight only for them to fail as soon as the first element succeeds because my laptop is no longer sharing the key :-)09:56
ssam2i don't think there's anything you can do wrong to break the cache, but make sure to comment out the 'push-url' and 'push-port' lines when you're not doing clean builds (like when modifying the guts of buildstream itself)09:57
ssam2you can also `bst push` stuff that you already built09:57
jonathanmawok, so I'll be wanting to do `bst push --deps all $TARGET`09:59
ssam2yeah10:00
jonathanmawaha, I should wait until I have a finalised toolchain added to "base", otherwise the uploaded artifacts won't ever be found.10:00
ssam2right yeah10:01
ssam2so first build and push bootstrap/stage3-sysroot.bst10:01
jonathanmawssam2: pushing the sysroot in a diffferent way than pushing it to the cache, I assume10:01
ssam2no, just push it to the cache10:02
ssam2this is an interim solution; i'm still figuring out the "correct way"10:02
ssam2as per https://mail.gnome.org/archives/buildstream-list/2017-September/msg00038.html10:03
jonathanmawssam2: ok, and it'll be addressed in a way other than by divining the cache key, iiuc10:03
ssam2right now you just have to divine the correct ref, indeed10:03
*** locallycompact has joined #baserock10:03
*** locallycompact has joined #baserock10:04
ssam2i already did this for ppc64b and x86_64: https://gitlab.com/baserock/definitions/commit/16a183b01f20f737189ab36e45d67c7527bdedfa10:04
ssam2it'll work fine until we need to prune the cache :-)10:04
ssam2but the buildstream conversion is still work-in-progress, so whatever10:05
ssam2by the way, you can put the ostree ref name into the "track" field, then run "bst track gnu-toolchain/base.bst" and it will figure out the "ref" field for you :-)10:06
ssam2and the log file generated by `bst push` should give you at least a clue at the ref name10:06
jonathanmawssam2: looking at the logs, it looks like I have two refs, I assume one is the strong ref, and one is the weak ref. Does it matter which one I choose? I can't really tell them apart.10:17
ssam2i prefer the strong ref10:19
ssam2i think in practice it doesn't matter10:19
ssam2i think the ref you see abbreviated in the bst console output is the strong one10:19
gitlab-br-botdefinitions: merge request (sam/bst-artifacts-url->master: project.conf: Set the artifact cache URL here) #57 changed state ("opened"): https://gitlab.com/baserock/definitions/merge_requests/5710:20
gitlab-br-botdefinitions: merge request (sam/bst-artifacts-url->master: project.conf: Set the artifact cache URL here) #57 changed state ("merged"): https://gitlab.com/baserock/definitions/merge_requests/5710:22
ironfootjonathanmaw: I've left a comment in https://gitlab.com/baserock/definitions/commit/e17e4b36c493b5ca50b27f888674c835adaf7bb710:23
ironfootlong time since I've looked at baserock bootsrap for different architecture10:24
ironfoots10:24
ironfootmeh, english, you know what I meant10:24
jonathanmawssam2: new base defined, started a full build10:32
jonathanmawperhaps I should push the commit as well, since it'll be pushing the artifacts to the cache as it builds.10:32
ssam2sure10:32
gitlab-br-botdefinitions: merge request (sam/remove-morph-add-buildstream->master: build-system: Remove Morph and add BuildStream) #60 changed state ("opened"): https://gitlab.com/baserock/definitions/merge_requests/6010:53
gitlab-br-botdefinitions: merge request (sam/armv8-ppc64->master: Various fixes for armv8 and ppc64) #59 changed state ("opened"): https://gitlab.com/baserock/definitions/merge_requests/5911:16
gitlab-br-botdefinitions: merge request (sam/armv8-ppc64->master: Various fixes for armv8 and ppc64) #59 changed state ("opened"): https://gitlab.com/baserock/definitions/merge_requests/5911:17
ssam2seems we've triggered a buildstream bug... https://gitlab.com/baserock/definitions/-/jobs/3447800411:26
* ssam2 will try and fix11:27
toscalixif the gitlab wiki is not being used, I recommend to go to Settings -> Integrations and in the Project Services section, click in the external wiki option and add the wiki URL. Then when you click in the wiki option, it will redirect you to the "actual wiki"11:42
ssam2good idea11:44
* ssam2 tries in vain to find "Settings" in the new UI11:45
ssam2oh, it's per-repo11:45
toscalixper group / per project/repo11:45
ssam2I don't see "Integrations" in there though11:45
ssam2oh, LHS again11:46
toscalixyou might not have admin rights in that specific repo then11:46
ssam2aha, it works11:46
*** adds68 has quit IRC12:14
*** adds68 has joined #baserock12:19
*** gtristan has quit IRC12:40
gitlab-br-botdefinitions: merge request (jonathan/buildstream-bootstrap->sam/buildstream-bootstrap: Add aarch64 base toolchain) #61 changed state ("opened"): https://gitlab.com/baserock/definitions/merge_requests/6113:10
*** toscalix_ has joined #baserock13:58
*** toscalix has quit IRC14:00
*** ctbruce has quit IRC14:57
*** ssam2 has quit IRC15:55
*** toscalix_ has quit IRC16:46
*** locallycompact has quit IRC17:02
*** jonathanmaw has quit IRC17:20
*** gtristan has joined #baserock20:27
*** gtristan has quit IRC21:32
*** rdale has quit IRC22:25

Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!