| jjardon | pedroalvarez: patch series to fix the issue sent to the list | 01:51 |
|---|---|---|
| *** zoli__ has joined #baserock | 07:43 | |
| *** grahamfinney_ has joined #baserock | 08:03 | |
| *** mike has joined #baserock | 08:06 | |
| *** mike is now known as Guest19217 | 08:07 | |
| *** rdale has joined #baserock | 08:27 | |
| *** zoli__ has quit IRC | 08:31 | |
| *** mariaderidder has joined #baserock | 08:33 | |
| *** zoli__ has joined #baserock | 08:33 | |
| *** zoli___ has joined #baserock | 08:38 | |
| *** zoli__ has quit IRC | 08:40 | |
| *** jonathanmaw has joined #baserock | 08:49 | |
| pedroalvarez | jjardon: wow, that was fast, thank you | 08:58 |
| pedroalvarez | jjardon: I hope you didn't think I was "angry" at you :/ I mean, the fix could have waited a bit | 09:02 |
| *** bashrc has joined #baserock | 09:04 | |
| radiofree | It's actually my fault | 09:05 |
| radiofree | This is exactly the error I said we wouldn't have to worry about when I gave it +1 | 09:05 |
| jjardon | Not at all, I wanted it fixed as well; that's the good think about the mason instance! | 09:06 |
| pedroalvarez | I'm now worried about genivi-weston, is it now the same? | 09:07 |
| radiofree | Genivi Weston is the one I tested | 09:09 |
| radiofree | It's old enough to not enable the libinput backend by default | 09:10 |
| pedroalvarez | :) | 09:10 |
| *** tiagogomes_ has joined #baserock | 09:13 | |
| pedroalvarez | And my last worry (due my lack of graphic stack experience) is: Will this old weston-genivi work with the wayland upgrade? | 09:13 |
| *** mdunford has joined #baserock | 09:14 | |
| radiofree | pedroalvarez: yeah, I use master Wayland with genivi Weston on my laptop. But I'll double check when I get in | 09:17 |
| pedroalvarez | radiofree: thank you very much :) | 09:17 |
| *** gary_perkins has joined #baserock | 09:33 | |
| jjardon | pedroalvarez: is the genivi system in mason? If not and the project cares about it I think it should be added | 09:57 |
| *** ssam2 has joined #baserock | 09:57 | |
| *** ChanServ sets mode: +v ssam2 | 09:57 | |
| perryl | hi, i'm currently running baserock in a chroot, and when i run ./check to test any changes i've made work correctly i get a couple of btrfs issues. is this an issue i will need to work around? | 09:59 |
| pedroalvarez | perryl: IIRC you need to enable btrfs support in your host OS | 10:00 |
| ssam2 | could you paste the errors somewhere ? | 10:01 |
| ssam2 | enter-baserock should run `modprobe btrfs` so btrfs should be enabled in the host kernel already | 10:01 |
| perryl | http://paste.baserock.org/kotibecare | 10:02 |
| ssam2 | 2 of the errors are due to missing extlinux/syslinux in the chroot | 10:03 |
| ssam2 | weirdly enough I do have extlinux in my chroot | 10:03 |
| ssam2 | ok, I see the problem | 10:04 |
| ssam2 | devel-system-x86_64-chroot contains the 'bsp-x86_both-tools' stratum which contains syslinux | 10:04 |
| ssam2 | but build-system-x86_64-chroot doesn't | 10:04 |
| ssam2 | so morph's test suite doesn't actually work in the reference systems we release as images... | 10:04 |
| perryl | ahh | 10:05 |
| persia | Yet another reason we shouldn't trust humans to run tests. | 10:05 |
| persia | Can we cause Mason to add running morph's test suite to the test sequence somehow? | 10:05 |
| persia | (for certain classes of system) | 10:05 |
| ssam2 | persia: yes, but I don't think that's at all related to this | 10:05 |
| ssam2 | as a developer, I need to be able to run the test suite manually | 10:06 |
| ssam2 | also, I think it's a mistake that the 'bsp-x86_both-tools' isn't in the 'build' system | 10:06 |
| ssam2 | since it prevents anyone using a 'build' system from deploying an x86 rawdisk system | 10:06 |
| persia | Yes, but I don't trust your developer-customised system to be a clean environment, or the process to create it to be documented. | 10:06 |
| persia | So this *particular* problem should be fixed (unrelated to Mason) | 10:06 |
| ssam2 | I do think adding Morph's test suite into Mason is a good idea | 10:07 |
| pedroalvarez | ssam2: then an arm system can't deploy a x86 rawdisk system? | 10:07 |
| persia | But *also* it would be good to have Mason run morph's test suite in a defined environment to protect against developers making mistakes in the future. | 10:07 |
| ssam2 | I see your point now. | 10:07 |
| ssam2 | pedroalvarez: yes | 10:07 |
| ssam2 | so maybe we should, perversely, add 'bsp-x86_both-tools' to all chroots? | 10:08 |
| pedroalvarez | ssam2: and to all devel systems (just in case only devel x86 systems have it) | 10:09 |
| ssam2 | perryl: how do you feel about doing that, building a new chroot, and testing that it fixes morph's ./check ? | 10:10 |
| perryl | ssam2: if there are instructions available i believe that should be fine | 10:10 |
| ssam2 | cool. adding a stratum to a system .morph is just a matter of an extra entry in the 'strata': list | 10:11 |
| ssam2 | deploying the new chroot can be done with `morph deploy clusters/release.morph build-system-x86-64-chroot` (will output a .tar file in your current directory) | 10:13 |
| ssam2 | you can then move the .tar into your host OS and run `manage-baserock add chroot-new ./build-system-x86_64-chroot.tar` then `enter-baserock chroot-new` to enter the new chroot | 10:14 |
| ssam2 | chroot-new is just a label, you can give it a better name if you want :) | 10:14 |
| perryl | i'll give it a go :) | 10:15 |
| *** zoli__ has joined #baserock | 10:56 | |
| *** zoli___ has quit IRC | 10:56 | |
| *** CTtpollard has quit IRC | 11:03 | |
| pedroalvarez | Here we go, a Basic PoC of Openstack running on Baserock! https://vimeo.com/119221370 | 11:10 |
| ssam2 | nice! | 11:12 |
| ridgerunner | radiofree: I've been through and cloned the repos you indicated. I ran ./autogen.sh in cbootimage/ and tegrarcm/ | 11:12 |
| pedroalvarez | The morphologies needed to build it are in the branch baserock/franred/openstack. Is a WIP so don't expect it to work perfectly | 11:12 |
| ridgerunner | radiofree: tegra-uboot-flasher/build-tools still fails. should the other repos be in a subdirectory of tegra-uboot-flasher? | 11:14 |
| radiofree | ridgerunner: build them manually and ignore ./build-tools | 11:15 |
| radiofree | README.developer + .user should tell you what to do next | 11:15 |
| perryl | are there any steps i need to take between building and deploying a system? | 11:16 |
| ridgerunner | ok. Any idea how long itwill take? | 11:16 |
| persia | perryl: It's often a good idea to double-check your cluster definition | 11:17 |
| radiofree | ridgerunner: not long, you basically just have to build u-boot and flash it (which ./build will do for you) | 11:18 |
| radiofree | don't forget to export CROSS_COMPILE | 11:18 |
| *** CTtpollard has joined #baserock | 11:18 | |
| perryl | persia: what for? | 11:18 |
| perryl | by which i mean is there anything in particular i should check in the cluster definition | 11:20 |
| persia | perryl: Make sure you are deploying to the intended target, with the right parameters, etc. | 11:20 |
| persia | I've become distracted building systems a few times, and ended up changing something such that I needed to alter my deployment parameters to avoid being disappointed. | 11:20 |
| ridgerunner | radiofree: my CROSS_COMPILE is currently set to the version I was using when doing kernel patches a few months ago. Any ideas as to what it should be set to? | 11:21 |
| bashrc | what architecture are you compiling to? | 11:21 |
| ridgerunner | TK1 BE. | 11:21 |
| bashrc | export CROSS_COMPILE=arm-linux-gnueabihf- | 11:22 |
| perryl | persia: i'm following the instructions on the wiki; morph build systems/build-system-x86_64-chroot.morph, then the deploy command ssam2 mentioned above | 11:22 |
| radiofree | Whatever you use for the kernel will be fine | 11:22 |
| persia | perryl: If you're trying to follow instructions from the wiki, then just look at it, so you know what it will do. Some of the reference clusters have placeholder variables that need to be adjusted. | 11:23 |
| straycat | thoughts on http://docs.openstack.org/developer/ceilometer/install/dbreco.html ? | 11:38 |
| pedroalvarez | straycat: we are using PostgreSQL | 11:38 |
| pedroalvarez | We will have to move to mysql if we want to move to the Openstack Kilo release | 11:39 |
| pedroalvarez | although maybe ceilometer will still support postgresql | 11:40 |
| persia | It won't be tested | 11:41 |
| straycat | I meant more about the recommendation to have the database on a dedicated host | 11:41 |
| persia | Or rather, it won't be part of the gate tests | 11:41 |
| persia | That's a reliability thing: if usage is to be high, then it needs large resources. Sharing resources with other things usually means lower availability. | 11:41 |
| persia | large-scale IaaS environments may be able to provide the level of reliability required in a VM. | 11:42 |
| ridgerunner | bashrc: Thanks, just recompiling the kernel using that setting, seems to be going ok. Will move on to U-Boot if the kernel completes correctly. | 11:43 |
| straycat | I guess if it needs to be dedicated that could be done at deployment | 11:44 |
| persia | I think that's the right time for the choice to be made. It really depends on the deployment needs. | 11:45 |
| persia | For test purposes, sticking on the same host as the services makes sense, whereas for large production environments, one needs to be able to point it at some HA DB cluster somewhere. | 11:45 |
| franred | persia, I think at the moment all the services in openstack have their dedicated databases - one for each one | 11:47 |
| franred | straycat, ^^ | 11:48 |
| persia | franred: Their own dedicated databases, or their own dedicated database servers? | 11:48 |
| franred | own databases - same postgresql server, I fear | 11:50 |
| *** mdunford has quit IRC | 11:50 | |
| persia | No need to fear: I think that's the right answer. | 11:51 |
| persia | But I think that it needs to be a deploy-time decision, as there are reasons why one might want to have the DB server separate from the services server. | 11:52 |
| persia | Or even to use different DB servers for different services (based on differing reliability requirements) | 11:52 |
| straycat | "The recommended Ceilometer storage backend is MongoDB." | 11:54 |
| persia | Interesting. That's different than most. | 11:55 |
| perryl | ssam2: deployed the new chroot, still fails in the same manner :( | 11:56 |
| ssam2 | perryl: what happens if you do `which extlinux` ? | 11:57 |
| perryl | no output | 11:58 |
| ssam2 | right, so somehow syslinux didn't end up in your chroot | 11:58 |
| franred | straycat, urgh! and fun | 11:58 |
| ssam2 | i'll come have a look | 11:58 |
| radiofree | why do you need extlinux in a chroot? | 11:58 |
| pedroalvarez | franred: no no no, no fun, he can still use postrgresql | 11:58 |
| radiofree | if you use BOOTLOADER_INSTALL: "none" morph won't bother trying to run extlinux | 11:59 |
| straycat | I'm not sure what we want to do yet, but it might not be wise to ignore their recommendation. I'm going to follow their devstack install guide to get a better idea of how things fit together | 12:00 |
| radiofree | unless you're trying to build a rawdisk image in a chroot, in which case you probably do want to run extlinux... | 12:00 |
| ssam2 | radiofree: deploying to a rawdisk image, and also running morph's test suite (which tests deploying to a rawdisk image) | 12:02 |
| *** zoli__ has quit IRC | 12:02 | |
| ssam2 | both things that the images we provide on download.baserock.org should probably support :) | 12:02 |
| radiofree | oh yeah you'll want a bootloader there | 12:03 |
| *** mdunford has joined #baserock | 12:13 | |
| persia | Is there a working write extension to manage flash storage? Something that manages UBI containers, or does squashfs with loopback, or similar? | 12:17 |
| ridgerunner | ./cbootimage-configs/README.txt | 12:18 |
| grahamfinney_ | do we have a top-level guide for baserock/morph? I don't mean a how-to, I mean a doc detailing what all the different parts do? | 12:18 |
| radiofree | ridgerunner: i meant README.* in uboot-flasher-scripts | 12:18 |
| radiofree | read developer first, then user | 12:19 |
| persia | grahamfinney_: There's some docs in the source code, both textual and as comments. I don't know of anything else. | 12:19 |
| *** mdunford has quit IRC | 12:20 | |
| grahamfinney_ | thanks persia | 12:20 |
| ridgerunner | Am reading developer. As far as I can see, the repos you said to clone already contain the tools etc that it talks about, am I right? | 12:20 |
| ridgerunner | radiofree: ^^^ | 12:20 |
| ssam2 | grahamfinney_: http://wiki.baserock.org/moving-parts/ and http://wiki.baserock.org/glossary/ may be of use | 12:21 |
| grahamfinney_ | Nothing describing what the various components do? | 12:22 |
| radiofree | ridgerunner: yes, you've done that right? just read about how to use the flashing tools now | 12:24 |
| ridgerunner | I've cloned those repos, yes. | 12:24 |
| straycat | persia, I don't think there's a write ext for flash storage, no | 12:24 |
| radiofree | you need to build and install things | 12:25 |
| radiofree | if you've cloned everything into the same folder, cd into uboot-flash scripts and run ./build-tools build | 12:26 |
| ssam2 | grahamfinney_: the intention of 'moving-parts' is to describe what the various components do. In what way do you find it lacking? | 12:26 |
| *** zoli__ has joined #baserock | 12:26 | |
| ssam2 | i'm sure it is lacking (in detail, if nothing else) but it'd be useful to get feedback from someone fresh to the project | 12:27 |
| rdale | i'm getting 'No space left on device' errors in my baserock image, but when i do a 'df -h' command there is plenty of space | 12:28 |
| straycat | what's giving the 'No space left' error? | 12:28 |
| ssam2 | rdale: are you trying to deploy a system with a too small DISK_SIZE ? | 12:29 |
| *** zoli__ has quit IRC | 12:29 | |
| rdale | http://paste.baserock.org/aqoyowuvit | 12:30 |
| rdale | i'm just trying to do 'morph build' | 12:31 |
| ssam2 | are you using btrfs? | 12:31 |
| ssam2 | for /src ? | 12:31 |
| rdale | yes | 12:32 |
| rdale | UUID=c14ddad4-3a11-417e-a567-7b35605876bd /home btrfs subvol=/state/home,defaults,rw,noatime 0 2 | 12:32 |
| rdale | LABEL=src /src btrfs defaults 0 2 | 12:32 |
| ssam2 | what does 'btrfs filesystem df /src' say ? | 12:33 |
| ssam2 | btrfs breaks 'df' | 12:33 |
| *** zoli__ has joined #baserock | 12:33 | |
| ssam2 | it also handles running out of disk space very, very badly | 12:33 |
| ssam2 | I recommend starting again with ext4 for your /src partition, if possible | 12:33 |
| rdale | http://paste.baserock.org/caxuhisoro | 12:34 |
| rdale | i'll copy it off onto ext4 then | 12:35 |
| ssam2 | I don't know how to interpret that paste. It seems like it should be fine, but clearly it isn't :) | 12:36 |
| rdale | what about btrfsck - does that exist? | 12:36 |
| ssam2 | it does, it's apparently not that useful, but maybe it improved since I last tried it | 12:37 |
| rdale | hmm | 12:38 |
| perryl | woohoo, extlinux is present, adding bsp-x86_both-tools to the strata fixes the btrfs issue | 12:39 |
| ssam2 | does ./check pass for you now ? | 12:40 |
| perryl | yup :) | 12:40 |
| ridgerunner | radiofree: ./build build fails: No such file or directory: '/home/robjones/develop/uboot/u-boot' | 12:41 |
| perryl | well, except for bad copyright years in morphlib/exts/virtualbox-ssh.write but that's an easy enough fix :) | 12:41 |
| radiofree | ridgerunner: clone u-boot | 12:44 |
| ridgerunner | sure, where from? | 12:46 |
| ridgerunner | denx.de? | 12:47 |
| radiofree | sure, upstream should work | 12:48 |
| tiagogomes_ | hey jjardon, are you planning to update the openssh as well? I have updated it to 6.5 in a branch | 12:54 |
| jjardon | tiagogomes_: maybe, why you didnt submit that already? ;) | 12:55 |
| tiagogomes_ | jjardon, because is still wip, it was need for aarch64 | 12:57 |
| persia | tiagogomes_: Please submit small separate patches, as anything else causes duplicate work (largely because people don't know what you are doing), or unacceptable delays (because people can't wait). | 12:58 |
| jjardon | tiagogomes_: you should submit what is ready as soon as possible instead sending a mega patch when everything is ready (my opinnion) | 12:58 |
| SotK | straycat: do you have time to glance over my v2 of "Speeding up construction of the build graph" to check that my changes are acceptable? | 12:58 |
| persia | The mega patches take forever to review as well, and there was already a post by a reviewer this month refusing to review a patch series because it was too big. | 12:58 |
| jjardon | tiagogomes_: 6.5 is already 1 year old anyway, current version is 6.7 | 13:01 |
| jjardon | persia: maybe you have some free time to reply to my question in the mailing list? ;) | 13:02 |
| persia | The one about CI? | 13:05 |
| jjardon | green in http://mason-x86-64.baserock.org/ :) | 13:05 |
| jjardon | persia: the one about using morph offline | 13:06 |
| persia | Oh, ugh. I meant to do that last week :( I'll dig that up and try to reply to it today. | 13:06 |
| jjardon | can I have another +1 for the GTK+ patches? Its not being used for any "official" system yet and Ive tested here without problems for x86 | 13:07 |
| jjardon | persia: np, thanks! | 13:08 |
| persia | Found the mail: the quick answer is: "Use file:///${PATH} entries for repo:" | 13:09 |
| * persia will send a longer response later | 13:09 | |
| ridgerunner | ok, I've downloaded the u-boot source from denx.de but can find nothing to tell me how to build it for the TK1. Can anybody help? | 13:21 |
| ridgerunner | Is this sort of stuff really "expected knowledge and expertise"? | 13:21 |
| persia | TK1 BE is largely unexplored. It has been done, but the wizardy has yet to be removed. | 13:29 |
| *** fay__ has joined #baserock | 13:29 | |
| *** fay_ has quit IRC | 13:31 | |
| CTtpollard | if I wanted to get access the setup file of the ikiwiki hosting w.b.o , who is the person to ask? | 13:35 |
| *** fay__ has quit IRC | 13:37 | |
| ssam2 | anyone on the ops team, in general. what do you need to do? | 13:40 |
| persia | I thought that was hosted by branchable: does ops team have access? | 13:41 |
| *** fay__ has joined #baserock | 13:42 | |
| ssam2 | persia: yes, all of us can get at the setup interface. As can Kinnison and liw-orc | 13:44 |
| ssam2 | and richard_maw, I think | 13:44 |
| persia | Ah, cool. | 13:44 |
| straycat | SotK, *nod* yes I think so :] | 13:54 |
| SotK | straycat: thanks :) | 13:55 |
| CTtpollard | it's in relation to internal codethink infrastructure, so I'm not sure how much I can discuss on here | 14:10 |
| pedroalvarez | CTtpollard: if it's codethink related, then ask to some people there :) | 14:12 |
| CTtpollard | I got pointed here :P | 14:12 |
| *** rdale_ has joined #baserock | 14:32 | |
| *** rdale has quit IRC | 14:35 | |
| *** grahamfinney_ has quit IRC | 14:40 | |
| pedroalvarez | tiagogomes_: heheh you have sent an email to Riku Voipio <riku.voipio@linaro.org> :) | 14:43 |
| Zara | today I discovered the 'use ls twice in a row, just to check' technique: http://paste.baserock.org/iletahelec | 14:45 |
| tiagogomes_ | pedroalvarez ops, I tend to use suppress-cc=self, I should had used 'all' in this case. I am sure that he will not mind the noise and he'll be glad that his commit is being used :) | 14:46 |
| persia | Zara: Indeed: never trust the backing store to be stable | 14:46 |
| SotK | Zara: did they actually produce different output or did your terminal cut the bottom line off the output from the first command when you re-ran it? (this happens to me sometimes) | 14:48 |
| Zara | SotK: I'm not sure, I ran it again absentmindedly because I couldn't find the openstack-cluster.morph, and forgot that I'd just tried running 'ls', which makes me think it produced different output, but I guess we'll never know | 14:50 |
| persia | For me, it usually provides different output on two successive runs, unless I really didn't do anything at all in the meantime. | 14:51 |
| CTtpollard | ty pedroalvarez , ssam2 | 14:53 |
| ssam2 | perryl: seems that I gave you bad advise with the extlinux patch. could you do a second version that only adds that stratum it to the x86 'build' systems ? | 14:55 |
| franred | does someone knows which is the latest release in qemu? it looks like it is 2.1.3 but the previous version was 2.2.0?? http://wiki.qemu.org/Main_Page | 14:55 |
| perryl | ssam2: yup, no problem | 14:55 |
| persia | franred: 2.1.3 is the most recent, 2.2.0 is the most advanced. | 14:56 |
| persia | From my reading, 2.1.3 includes backport bugfixes to 2.1.2, whereas 2.2.0 has new features over 2.1.2 | 14:56 |
| franred | persia, thanks, that clarifies it a little bit more | 14:57 |
| *** grahamfinney_ has joined #baserock | 14:58 | |
| pedroalvarez | franred: I'd try with 2.1.3 for now | 14:58 |
| *** grahamfinney_ has quit IRC | 14:59 | |
| straycat | pedroalvarez, have you tried baserock on the rasberry pi 2? | 14:59 |
| *** grahamfinney has joined #baserock | 15:00 | |
| franred | pedroalvarez, ok, I will give it a go | 15:00 |
| pedroalvarez | straycat: I almost by one to try that | 15:00 |
| *** grahamfinney has joined #baserock | 15:00 | |
| straycat | pedroalvarez, shame a friend of mine seems to be going to a lot of trouble to get a minimal system for his pi 2, apparently minibian isn't small enough. I imagine one of the armv7 images could work on the pi 2, but I don't know what they need to boot etc | 15:04 |
| persia | None of the Baserock reference images are super-small either. | 15:05 |
| straycat | I thought the minimal system was? | 15:05 |
| jmacs | Why does it need to be small? | 15:06 |
| straycat | No idea, I assume he has good reasons | 15:06 |
| straycat | oh, minibian apparently hasn't updated for the pi 2 yet | 15:12 |
| rjek | Just a matter of the kernel, I'd have thought? | 15:14 |
| rjek | Stock Debian armhf port should work quite nicely on it | 15:14 |
| rjek | (As it now has an ARMv7 CPU) | 15:14 |
| *** mdunford has joined #baserock | 15:19 | |
| persia | I think our minimal is smaller than debootstrap's minimal, but it isn't that important, as one is more likely to choose Baserock vs. debootstrap, live-build, etc. based on familiarity with tooling and the need for reproducibility rather than the output of some random sample system. | 15:20 |
| *** Krin has quit IRC | 15:21 | |
| rjek | qemu-arm-static is some kind of witchcraft | 15:35 |
| rjek | debootstrap --arch=armhf --foreign wheezy my-arm-chroot && chroot my-arm-chroot -> shell on an ARM system | 15:36 |
| rjek | (I have not mentioned the second stage debootstrap as that's not related to my impression of qemu's magic.) | 15:36 |
| * persia prefers mk-schroot, but no longer believes that toolchain is useful due to reproducibility issues | 15:43 | |
| *** rdale_ has quit IRC | 15:44 | |
| *** rdale has joined #baserock | 15:45 | |
| *** flatmush has quit IRC | 15:45 | |
| *** flatmush has joined #baserock | 15:46 | |
| *** flatmush has quit IRC | 15:48 | |
| *** flatmush has joined #baserock | 15:52 | |
| *** ssam2 has quit IRC | 15:54 | |
| pedroalvarez | oh, yeah. readline again | 16:14 |
| pedroalvarez | I feel like we need to think about this problem | 16:14 |
| nowster | what is the tool which updates refs in definitions based on the unpetrify-ref ? | 16:15 |
| *** ssam2 has joined #baserock | 16:15 | |
| *** ChanServ sets mode: +v ssam2 | 16:15 | |
| pedroalvarez | nowster: there is no tool to do that I believe | 16:16 |
| nowster | so what's the significance of unpetrify-ref? | 16:16 |
| persia | historical baggage. | 16:16 |
| franred | can someone check this for a fast review http://paste.baserock.org/imigutumot ? | 16:16 |
| SotK | we used to have morph petrify and morph unpetrify to do that | 16:16 |
| persia | Some people use it as a comment field, because morph doesn't support comments. | 16:16 |
| nowster | ah | 16:16 |
| pedroalvarez | franred: for what exactly do you need this? | 16:17 |
| nowster | sounds like it could be (or have been) useful | 16:17 |
| franred | pedroalvarez "Add qemu submodule openhackware" to update to v2.1.3 - because our current version of qemu does not compile with the latest changes made in gcc (I guess) | 16:18 |
| pedroalvarez | franred: I see. I guess you have verified that v2.1.3 compiles? | 16:19 |
| persia | Why don't we want 2.2.0 again? | 16:19 |
| franred | persia, I valuated more bugfixing/fresher version over features - but I can be wrong | 16:20 |
| pedroalvarez | franred: openhackware will be needed anyway, so +1 | 16:21 |
| persia | I don't think 2.1.3 is fresher. | 16:21 |
| persia | I think 2.2.0 is fresher. | 16:21 |
| persia | I think 2.1.3 is an attempt to backport the bugfixes already in 2.2.0 to 2.1.2 | 16:21 |
| persia | But I suspect there were bugs they didn't know how to backport, because of other changes. | 16:21 |
| franred | pedroalvarez, I verified that compiles, I can hold the patch until I have the image tested | 16:22 |
| pedroalvarez | that is a good argument to go to 2.2.0 instead | 16:22 |
| franred | persia, 2.1.3 is from 22nd of January and 2.2.0 is from december 2014 | 16:22 |
| persia | franred: Right, so it took three weeks to determine which bugs could be backported. | 16:23 |
| persia | I still think 2.2.0 has fewer old bugs than 2.1.3 (although it may have more new ones) | 16:23 |
| persia | If you say "I want 2.1.3 because it is more stable", I'll stop promoting 2.2.0. I just think 2.1.3 is the wrong choice if we want something "fresh" or "new" | 16:24 |
| franred | persia, Im not against 2.2.0 - I always prefer latest tags - I will give a try and if 2.2.0 does not have any problems I will go with it | 16:25 |
| persia | 2.2.0 is what most folk would call "latest" for that. 2.1.3 would be a "stable branch update" | 16:25 |
| persia | On an unrelated note, is the readline in the build systems still the GPLv2 readline? | 16:27 |
| pedroalvarez | yes | 16:29 |
| persia | And texinfo? | 16:29 |
| radiofree | hmm texinfo in core is gplv3 | 16:32 |
| radiofree | but it gets stripped afterwards | 16:32 |
| SotK | I think it gets stripped from genivi | 16:33 |
| persia | My worry is wheter artifacs constructed by GPLv3 tools might end up carrying any obligations. I hope that my fear is unfounded. | 16:33 |
| *** sambishop has joined #baserock | 16:35 | |
| *** Krin has joined #baserock | 16:42 | |
| jjardon | persia: gcc is GPLv3 | 16:43 |
| jjardon | so I do not think its a problem | 16:43 |
| rjek | Depends on licence. | 16:43 |
| rjek | gcc has a specific waiver in its licence for the shared code that every produced binary includes | 16:43 |
| radiofree | so can we not just strip readline afterwards, or is that going to break python + other things? | 16:44 |
| * persia doesn't know | 16:44 | |
| rjek | radiofree: By "strip", do you mean "remove the shared object form the system" ? | 16:44 |
| rjek | if so, then yeah; every binary linked with it will stop working | 16:44 |
| radiofree | i think that's what the strip extension does | 16:44 |
| radiofree | and maybe it strips things built against it as well? | 16:44 |
| rjek | Then you rapidly end up with a system with no binaries in it :) | 16:45 |
| persia | I think it's better to just not install things than install and then try to remove. Being sure one got everything is tricky. | 16:45 |
| jjardon | thats why I explained in the review that the strip way will not work | 16:45 |
| rjek | I wrote something ABI-compatible with readline that is mostly just stubs that do anything apart from the actual readline() call that just proxies to fgets(), called fakeline | 16:46 |
| rjek | But there was little interest in it. | 16:46 |
| * rjek wishes people used libeditline instead, which a) is BSD licenced, and b) has an API that was actually designed | 16:46 | |
| jjardon | the problem is more general that expecifically readline | 16:46 |
| rjek | Indeed, but readline always comes up | 16:47 |
| persia | rjek: Is there a libeditline wrapper that provides a readline-compatible API? | 16:48 |
| tiagogomes_ | I am planning to redo the readline update, but keep the texinfo one | 16:48 |
| rjek | persia: You could probably write one, but readline's API is a typical GNU suppurating sprawling horror which is underdocumented and overabused | 16:48 |
| rjek | http://www.rjek.com/readline.h <-- readline API surface | 16:49 |
| persia | rjek: Understood, but such a wrapper solves my problem today, and I don't have enough cricket bats to cause people to stop using readline. | 16:49 |
| *** ssam2 has quit IRC | 16:49 | |
| jjardon | tiagogomes_: before you spend time rewriteing the patches, I think we should think in a general solution for how integrate GPLv3 libraries | 16:50 |
| rjek | persia: If you just wanted line editing but none of the tab completion, macros, history, etc, making fakeline call out to libeditline is probably a 10 minute job. | 16:50 |
| * persia happily waits 10 minutes in hopes of receiving repo and ref data for the preparation of a patch | 16:51 | |
| rjek | start with bzr+http://bzr.rjek.com/public/fakeline ;-) | 16:52 |
| persia | And the ref including the libeditline callouts? | 16:53 |
| rjek | Exists only in one of our possible futures. | 16:55 |
| persia | Ah, pity. | 16:56 |
| *** grahamfinney has quit IRC | 16:57 | |
| rjek | Actually, the situation's better than that; it exists in many of our possible futures. | 16:57 |
| rjek | Probably at least half. | 16:57 |
| *** grahamfinney has joined #baserock | 16:59 | |
| *** grahamfinney is now known as 5EXAAGQLD | 16:59 | |
| *** 21WABANV9 has joined #baserock | 16:59 | |
| *** ssam2 has joined #baserock | 17:02 | |
| *** ChanServ sets mode: +v ssam2 | 17:02 | |
| ssam2 | could anyone give a second review to my 'Add time-zone-database to foundation' patch? | 17:03 |
| pedroalvarez | ssam2: +1, will send an email | 17:04 |
| pedroalvarez | I'm still wondering what problems does it fix | 17:04 |
| ssam2 | not being able to set a timezone | 17:05 |
| ssam2 | it was bundled in EGLIBC 2.15, but for 2.20 it's not bundled | 17:05 |
| ssam2 | GLIBC 2.20 that is | 17:05 |
| *** zoli__ has quit IRC | 17:23 | |
| *** mariaderidder has quit IRC | 17:29 | |
| pedroalvarez | hm.. I'm trying to figure out from where can I get "/usr/sbin/dhcp_release". | 17:31 |
| pedroalvarez | looks like it's included in dnsmasq-utils package in some distros | 17:31 |
| *** bashrc has quit IRC | 17:31 | |
| *** gary_perkins has quit IRC | 17:32 | |
| * pedroalvarez still can't find source code for it | 17:35 | |
| persia | https://sources.debian.net/src/dnsmasq/2.72-2/contrib/wrt/dhcp_release.c/ | 17:38 |
| pedroalvarez | aha, so it is in dnsmasq repo but is not being built (in baserock) | 17:40 |
| persia | Right. It's in the contrib/wrt directory, which complicates building. Probably needs some magic hints during the build. | 17:42 |
| pedroalvarez | persia: indeed, but looks like fedora is already doing these magic hints :) | 17:46 |
| persia | https://sources.debian.net/src/dnsmasq/2.72-2/debian/rules/ also | 17:46 |
| pedroalvarez | same magic | 17:47 |
| persia | Debian also strips that binary in differently than normal ways. | 17:47 |
| jjardon | pedroalvarez: can you ask why you need that for? AFAIK in baserock dhcp is currently done by networkd | 17:57 |
| pedroalvarez | jjardon: openstack tries at some points to run the following command, and looks like it needs dhcp_release in the system | 17:58 |
| pedroalvarez | ip netns exec qdhcp-63399d35-f830-4972-86b5-fafaf4eca392 dhcp_release tapadbb6630-f5 192.168.1.4 fa:16:3e:5c:86:14 | 17:58 |
| *** sambishop has quit IRC | 18:00 | |
| pedroalvarez | jjardon: also, this is only going to be present in systems with dnsmasq, which is in virtualization.morph | 18:00 |
| persia | Might be interesting to ask the neutron folk *why* they run that command. | 18:00 |
| * straycat grumbles about python requirements that don't specify versions | 18:00 | |
| * persia suspects straycat would get along swimmingly with the current PEP curators | 18:01 | |
| *** jonathanmaw has quit IRC | 18:03 | |
| *** gary_perkins has joined #baserock | 18:05 | |
| *** mdunford has quit IRC | 18:19 | |
| *** 21WABANV9 has quit IRC | 18:19 | |
| *** 5EXAAGQLD has quit IRC | 18:19 | |
| *** Guest19217 has quit IRC | 18:31 | |
| *** ssam2 has quit IRC | 18:42 | |
| *** gary_perkins has quit IRC | 19:18 | |
| *** edcragg has quit IRC | 19:25 | |
| *** rdale_ has joined #baserock | 19:36 | |
| *** rdale__ has joined #baserock | 19:38 | |
| *** rdale_ has quit IRC | 19:38 | |
| *** rdale_ has joined #baserock | 19:39 | |
| *** rdale has quit IRC | 19:39 | |
| *** rdale has joined #baserock | 19:41 | |
| *** rdale__ has quit IRC | 19:43 | |
| *** rdale_ has quit IRC | 19:44 | |
| *** rdale has quit IRC | 19:53 | |
| *** rdale has joined #baserock | 19:58 | |
| *** rdale has quit IRC | 20:07 | |
| *** rdale has joined #baserock | 20:08 | |
| *** rdale_ has joined #baserock | 20:17 | |
| *** rdale has quit IRC | 20:17 | |
| *** rdale has joined #baserock | 20:46 | |
| *** rdale__ has joined #baserock | 20:48 | |
| *** rdale_ has quit IRC | 20:49 | |
| *** rdale has quit IRC | 20:51 | |
| *** rdale has joined #baserock | 20:54 | |
| *** rdale__ has quit IRC | 20:57 | |
| *** rdale has joined #baserock | 21:00 | |
| *** rdale has quit IRC | 21:11 | |
| *** rdale has joined #baserock | 21:13 | |
| *** rdale_ has joined #baserock | 21:15 | |
| *** rdale has quit IRC | 21:15 | |
| *** rdale has joined #baserock | 21:17 | |
| *** rdale__ has joined #baserock | 21:19 | |
| *** rdale_ has quit IRC | 21:20 | |
| *** rdale_ has joined #baserock | 21:21 | |
| *** rdale__ has quit IRC | 21:23 | |
| *** rdale has quit IRC | 21:23 | |
| *** rdale_ has quit IRC | 21:23 | |
| *** rdale has joined #baserock | 21:23 | |
| *** rdale_ has joined #baserock | 21:37 | |
| *** rdale has quit IRC | 21:40 | |
| *** rdale_ has quit IRC | 21:59 | |
| *** zoli__ has joined #baserock | 23:25 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!