IRC logs for #baserock for Sunday, 2015-10-11

*** edcragg has quit IRC00:49
*** gtristan has joined #baserock02:25
*** lachlanmackenzie has quit IRC02:33
*** lachlanmackenzie has joined #baserock04:04
gtristanjjardon, http://picpaste.com/baserock-gnome-shell-D7puKghz.png !08:12
SotKgtristan: \o/08:22
gtristanThat... was really not as easy as I expected :)08:23
gtristanand, it's not got gdm properly configured... pam modules also need proper attention, no applications08:23
gtristanbut... at least it's a milestone08:24
*** lachlanmackenzie has quit IRC08:36
gtristanfwiw, I'll require one additional lorry: https://gerrit.baserock.org/#/c/1165/08:39
gtristanfor the above to build08:39
*** edcragg has joined #baserock09:09
*** lachlanmackenzie has joined #baserock09:24
gtristanhrm... git review seems buggy09:40
gtristanSo, this happens in a fresh checkout, no matter how many times I download `commit-msg' into .git/hooks: http://paste.baserock.org/okevadeziv09:47
gtristanAnd, in another checkout, I get this error: http://paste.baserock.org/fezihigeza09:48
gtristaneven though git show 815fb0c09f6a97909dc388b15d8c054e73107941... shows a commit with the correctly set codething.co.uk author09:48
paulsherwoodgtristan: http://git.baserock.org/cgi-bin/cgit.cgi/delta/telepathy-logger.git/09:50
gtristanthe above is using: git review -R -t make-gnome-shell-startup (to upload a patchset containing close to 20 commits, which *really* should not be squashed into a single commit)09:50
* paulsherwood can't help with the gerrit magic, sadly09:50
gtristan:)09:50
paulsherwoodin other news, i seem to have a working ybd+vagrant setup09:51
paulsherwoodhttp://wiki.baserock.org/ybd/?updated#index1h209:52
* gtristan learns what vagrant is at: https://www.vagrantup.com/09:54
* gtristan -> supper09:56
*** gtristan has quit IRC09:56
paulsherwoodit's basically some ruby+go to simplify spinning up virtual machines for developers09:56
paulsherwoodvery popular with app folks i think09:56
pedroalvarezI've been using it for testing Ansible playbooks :)10:05
pedroalvarezyou can configre Ansible as a provider for Vagrant, making the cycle "create fresh VM" -> "run Ansible playbooks against it" really easy10:06
pedroalvarezpaulsherwood: you can create a Vagrantfile that also installs the dependencies (and even ybd), so that when you do `vagrant up` everything is set up10:10
paulsherwoodpedroalvarez: oh, ok. i need to read more then :)10:11
* paulsherwood is only skating on the surface of the pond10:11
pedroalvarezAnd... you can even put this file in the ybd repo, so the instructions to get started would be: clone ybd, `vagrant up`10:12
* rjek tends to test ansible playbooks against a VM with a snapshot, so he can roll-back to the before state10:15
pedroalvarezrjek: my usecase is to make sure that they work for a fresh VM :)10:18
paulsherwoodpedroalvarez: noted. am working on it :)10:20
* pedroalvarez continues playing with the raspberry 210:21
*** zoli_ has joined #baserock10:21
pedroalvarezIf u-boot upstream works, I can stop using another binary blob :)10:22
paulsherwoodcool :-)10:22
*** gtristan has joined #baserock10:27
* paulsherwood wonders why we're now using linux-stable as well as linux10:28
pedroalvarezpaulsherwood: different versions of linux for different BSPs10:31
pedroalvarezthere were some discussion about that here, some people didn't like the idea, others didn't disagree10:32
pedroalvarezThe patch ended up being merged10:32
rjekpedroalvarez: My fresh VMs are made from clones of existing VMs with snapshots.10:33
rjekIt works well10:33
pedroalvarezyay, this u-boot works :)10:34
pedroalvarezNext task: get extlinux.conf working, so that we can upgrade them :)10:34
paulsherwoodpedroalvarez: it adds quite a few minutes to my 'startup experience'10:37
paulsherwoodpedroalvarez: you're on fire :)10:37
pedroalvarezI am, everything works for me this weekend10:38
pedroalvarezpaulsherwood: We could try to get the same kernel for all bsps we care about this week10:38
paulsherwood+110:38
pedroalvarezBTW: I think for raspberry pi 2 I can't use upstream kernel tree10:39
pedroalvarezIt's on my todo list of things to investigate10:39
*** zoli_ has quit IRC11:13
*** zoli_ has joined #baserock11:13
*** zoli_ has quit IRC11:19
gtristanAha I see now11:50
gtristanthe commit-msg hook doesn't run for patches applied with git am11:50
gtristanbut with interactive rebase, doing 'reword', you can force it11:50
* gtristan thinks he is able to work around this gerrit monster and forcefeed it some patches :)11:51
gtristanyay !11:52
gtristandoes this look about right: https://gerrit.baserock.org/#/q/status:open ?11:54
gtristanseems to have created one gerrit entry for each patch, as opposed to one entry with a list of patches (but they are all at least on the same 'startup-gnome-shell' branch)11:55
* gtristan thinks he will avoid git-review in the future and would rather be exposed to the gory details - otherwise when it fails, you just have no clue why11:56
gtristanalso, I think there is some kind of bug, really11:56
gtristanas I usually configure my email address only locally to the repository in question11:56
gtristanbut it seems that when pushing to gerrit, it does something and only pulls my globally configured email address11:57
gtristan(causing that message which actually makes no sense)11:57
*** edcragg has quit IRC13:09
*** edcragg has joined #baserock13:17
* gtristan files https://code.google.com/p/gerrit/issues/detail?id=3597 ... and hits the hey14:17
persiagtristan: each patch as a separate review entry in the same branch is normal.  Not ideal, but one of the less frustrating warts gerrit has.  Unfortunately, nothing else has the base features, making dependence on gerrit something we all work around with distaste.14:42
gtristan:)14:45
gtristanyou can at least view all the patches of a topic branch together14:46
persiaYes, and with tools like gertty, the experience is almost sane.14:52
* paulsherwood is failing dismally to get ybd working on shared filesystems :/15:21
*** gtristan has quit IRC15:23
rjekpaulsherwood: What are you running into?15:33
paulsherwoodybd uses mknod to create /dev/null in the sandboxes. that doesn't seem to work for vboxfs or nfs15:36
rjekAlmost certainly not, and for good security reasons.15:36
rjekDevice nodes are only normally creatable locally, and by root.15:39
rjekI would be surprised if vboxfs supports device nodes at all15:39
rjekpaulsherwood: Perhaps use a local tmpfs for /dev ?15:40
paulsherwoodrjek: that was edmund's suggestion. my brain is so far failing to join the dots, though16:19
*** edcragg has quit IRC17:19
*** edcragg has joined #baserock19:21
*** edcragg has quit IRC20:40
*** edcragg has joined #baserock20:46
pedroalvareznope, extlinux.conf doesn't work straightaway21:06
pedroalvarezand I don't have any clue of why :)21:07
*** edcragg has quit IRC22:31

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