| *** radiofree has quit IRC | 02:36 | |
| *** radiofree has joined #baserock | 02:38 | |
| *** zoli__ has joined #baserock | 06:27 | |
| *** zoli___ has joined #baserock | 06:31 | |
| *** zoli__ has quit IRC | 06:33 | |
| *** petefoth has joined #baserock | 06:40 | |
| *** franred has joined #baserock | 07:39 | |
| *** rdale has joined #baserock | 07:59 | |
| *** a1exhughe5 has joined #baserock | 08:04 | |
| *** zoli___ has quit IRC | 08:04 | |
| *** zoli__ has joined #baserock | 08:07 | |
| *** gfinney has joined #baserock | 08:17 | |
| *** CTtpollard has joined #baserock | 08:19 | |
| *** sambishop has joined #baserock | 08:33 | |
| *** mdizzle has joined #baserock | 08:40 | |
| *** bashrc has joined #baserock | 09:06 | |
| *** jonathanmaw has joined #baserock | 09:21 | |
| *** tiagogomes_ has joined #baserock | 09:28 | |
| *** gary_perkins has joined #baserock | 09:31 | |
| bashrc | I guess that morph build both builds and installs | 09:32 |
|---|---|---|
| Kinnison | 'morph build' prepares artifacts | 09:33 |
| Kinnison | from the PoV of the actions taken with regards the commands defined in the chunk morphologies, that means the whole configure/build/install cycle | 09:33 |
| bashrc | ok | 09:34 |
| bashrc | suppose I wanted a particular system with particular software added or removed. Is there an equivalent to buildroot make menuconfig? | 09:36 |
| Kinnison | No | 09:36 |
| Kinnison | You get to edit the morphologies | 09:36 |
| Kinnison | there are no tools for manipulating them because it's exceedingly hard for computers to edit things in ways humans like | 09:36 |
| bashrc | how would I disable a chunk or stratum? | 09:37 |
| Kinnison | baserock systems are not meant to be 100% parameterised things, instead they are collections of software (strata) joined into coherent wholes (systems) | 09:37 |
| Kinnison | for a stratum, copy the system definition and in your copy, remove the stratum you no longer want | 09:37 |
| tiagogomes_ | how is system-version-manager deploy supposed to be used? How the orig snapshot is transferred to the client os? | 09:37 |
| Kinnison | for a particular chunk, you (a) need to be careful regarding build or runtime dependencies and (b) copy the system and the stratum in question and then make your edits in your copies | 09:37 |
| pedroalvarez | tiagogomes_: ssh-rsync write extension uses it | 09:38 |
| pedroalvarez | tiagogomes_: I don't think that the `deploy` command is useful without it | 09:38 |
| bashrc | could there be a menuconfig, or will selecting software always be down to manual editing? | 09:39 |
| Kinnison | In a future form of definitions, there may be computer-based selections, but for now, it's editing | 09:39 |
| tiagogomes_ | pedroalvarez, that makes fore sense, and it is useful to me that ssh-rsync runs 'deploy' :) | 09:40 |
| *** edcragg has joined #baserock | 09:46 | |
| pedroalvarez | cool | 09:47 |
| * pedroalvarez discovers "cmake -DCMAKE_INSTALL_PREFIX=/usr" in buildsystem.py | 09:47 | |
| pedroalvarez | shouldn't that be $PREFIX? | 09:48 |
| rdale | no, it is correct | 09:48 |
| Kinnison | pedroalvarez: It probably should be $PREFIX yes | 09:48 |
| pedroalvarez | everything else is $PREFIX in that file | 09:49 |
| pedroalvarez | rdale: any luck with `make` in a staging area? | 09:49 |
| Kinnison | pedroalvarez: I wonder if it's worth fixing that now, or leaving it until build systems are declarative? | 09:49 |
| pedroalvarez | Kinnison: that's true | 09:50 |
| Kinnison | pedroalvarez: I guess if you're in there for any other reason in the meantime, fix it, otherwise leave it? | 09:50 |
| rdale | pedroalvarez: no i haven't done anything since i found out it needed /proc in the chroot yesterday | 09:50 |
| bashrc | here's another general question: is there any reason for baserock not being officially mirrored on Github? You'll get more visibility/discoverability | 09:50 |
| Kinnison | Github is fundamentally a pain | 09:51 |
| Kinnison | Its model for contributions does not match ours | 09:51 |
| Kinnison | it's a more centralised VCS than Subversion could ever have hoped to be | 09:51 |
| Kinnison | and it's exceedingly hard to script it | 09:51 |
| Kinnison | (having tried) | 09:51 |
| bashrc | Github a a den of proprietaryness, but it has lots of eyeballs. People search for stuff there | 09:52 |
| Kinnison | Github's single-layer project/repository structure doesn't match what we need | 09:52 |
| Kinnison | Its rules are insufficient to express what we need | 09:52 |
| Kinnison | etc. | 09:52 |
| Kinnison | The eyeballs would be nice, but IMO not worth the hassle | 09:52 |
| rdale | i can't see any mention of $PREFIX in the default cmake rules in buildsystem.py, and there shouldn't be | 09:52 |
| Kinnison | rdale: why shouldn't there be? | 09:52 |
| radiofree | tiagogomes_: i sent a patch to use deploy for upgrading from rawdisk images | 09:52 |
| bashrc | apart from the technical stuff. I'm thinking of the sociology if you want to grow a live open source project | 09:52 |
| pedroalvarez | radiofree: oh yeah, was it merged? | 09:53 |
| radiofree | (deploy in system-version-manager) i tink you essentially tell it "there's a rootfs in this folder here... upgrade the current system using that" | 09:53 |
| radiofree | pedroalvarez: no, needs some comments adding to it, that i've not go round to yet | 09:53 |
| Kinnison | bashrc: Personally I'd rather find a way to grow the project without having to rely on github's eyes | 09:54 |
| Kinnison | bashrc: But I'm sure the ops guys have noted your points and interest :) | 09:54 |
| rdale | Kinnison: because the correct way to set up an install prefix in a cmake project is to use -DCMAKE_INSTALL_PREFIX=<prefix> | 09:54 |
| radiofree | tiagogomes_: "morph upgrade" with ssh-rsync calls it (system-version-manager deploy...) after rsyncing over the new system to a new subvoulme | 09:54 |
| Kinnison | rdale: right | 09:54 |
| *** ssam2 has joined #baserock | 09:54 | |
| *** ChanServ sets mode: +v ssam2 | 09:54 | |
| Kinnison | rdale: so why can't it be "cmake -DCMAKE_INSTALL_PREFIX=$PREFIX" ? | 09:54 |
| rdale | ah good point | 09:55 |
| tiagogomes_ | radiofree, yep, just saw, that, upgrades should work then even on a /boot partition | 09:55 |
| radiofree | tiagogomes_: if you set BOOT_PARITION and ROOT_PARITION, yeah | 09:55 |
| tiagogomes_ | without extra fiddling around | 09:55 |
| tiagogomes_ | radiofree, I look at the code, and I didn't find a reason to use ROOT_PARTITION | 09:56 |
| tiagogomes_ | ROOT_PARTITION seems to be only needed for creating /etc/fstab | 09:56 |
| radiofree | if you don't set ROOT_PARTITION i believe morph defaults to /dev/sda? | 09:57 |
| radiofree | http://git.baserock.org/cgi-bin/cgit.cgi/baserock/baserock/morph.git/commit/?id=de35a0f256999e107ee524a4647151aab6cbfc87 | 09:59 |
| radiofree | sorry i always get that wrong | 10:01 |
| radiofree | BOOT_DEVICE and ROOT_DEVICE | 10:01 |
| tiagogomes_ | radiofree, root device is being used for fstab and kernel args on extlinux. The latter doesn't matter, because I didn't get sysboot to work, so the kernel args are going to be read from boot.src | 10:01 |
| radiofree | the former matters though? | 10:01 |
| radiofree | actually it probably won't, because the upgrade uses the UUID anyway | 10:02 |
| radiofree | however.... i think it's nicer to just put ROOT_DEVICE and BOOT_DEVICE in your cluster so it's obvious these things are different there | 10:02 |
| tiagogomes_ | radiofree agreed | 10:02 |
| pedroalvarez | yeah, I tried to fix the fstab issue in this commit: http://git.baserock.org/cgi-bin/cgit.cgi/baserock/baserock/morph.git/commit/morphlib/exts/ssh-rsync.write?id=558b7dbfb6a62490e65bb4a4667ab8bf03495f68 | 10:03 |
| bashrc | is the root of definitions the best place for moonshot? Should it not be in some hardware subdiectory? | 10:09 |
| Kinnison | Very little should be in the root of definitions | 10:10 |
| Kinnison | what file in particular is there? | 10:10 |
| bashrc | there's a moonshot directory and moonshot-kernel.configure | 10:10 |
| Kinnison | Oh | 10:10 |
| * Kinnison isn't sure what our rules are about that kind of thing | 10:11 | |
| bashrc | maybe better to have something like hardware/moonshot | 10:11 |
| Kinnison | If it's there in master then I guess someone liked it enough for now | 10:11 |
| * SotK doesn't think we have a rule for configure extensions or stuff for install-files yet | 10:11 | |
| Kinnison | I believe it's temporary | 10:11 |
| Kinnison | Since I imagine it's only there until the M400's DTB is merged upstream into the kernel | 10:11 |
| *** wschaller has joined #baserock | 10:21 | |
| *** Krin has joined #baserock | 10:22 | |
| tiagogomes_ | radiofree http://paste.baserock.org/azugoxoqaf | 10:25 |
| radiofree | tiagogomes_: it can't read ** File not found /boot.scr ** | 10:25 |
| radiofree | is that file on / of your boot partition? | 10:25 |
| radiofree | heh, "Trying scsi 0 /boot.scr.bak..." | 10:26 |
| radiofree | how sensible of them :) | 10:26 |
| tiagogomes_ | ah right, I know what is going on | 10:26 |
| *** mdizzle has quit IRC | 10:35 | |
| *** wschaller has quit IRC | 10:38 | |
| *** tiagogomes_ has quit IRC | 10:39 | |
| bashrc | another question. When I see "upstream:node" in the nodejs definition, where is that upstream defined? | 10:43 |
| pedroalvarez | upstream: is an alias of "git://TROVE_HOST/delta/" | 10:45 |
| pedroalvarez | e.g. git://git.baserock.org/delta/node | 10:45 |
| pedroalvarez | http://git.baserock.org/cgi-bin/cgit.cgi/delta/node.git/ | 10:46 |
| pedroalvarez | (cgit url) | 10:46 |
| bashrc | ok | 10:46 |
| *** zoli__ has joined #baserock | 10:52 | |
| bashrc | suggestion: change unpetrify-ref to branch-ref | 11:02 |
| bashrc | incidentally, is there anywhere to log issues for baserock? Things tend to get lost on mailing lists | 11:05 |
| petefoth | bashrc: StoryBoard is / will be the official bug- and feature-tracker. People are using the instance at https://storyboard.baserock.org/#!/project/list | 11:07 |
| * petefoth is reminded to progress some of the tasks he has worked on recently | 11:09 | |
| pedroalvarez | right, patches needed for the genivi release sent | 11:14 |
| *** gfinney has quit IRC | 11:22 | |
| *** gfinney has joined #baserock | 11:29 | |
| *** petefoth has quit IRC | 11:44 | |
| *** petefoth has joined #baserock | 11:44 | |
| *** petefoth_ has joined #baserock | 11:55 | |
| *** petefoth has quit IRC | 11:56 | |
| *** petefoth_ is now known as petefoth | 11:56 | |
| pedroalvarez | If we are releasing today 15.10, I think it should have latest morph: http://paste.baserock.org/sucusalebi | 12:01 |
| petefoth | I want to add a story to StoryBoard realting to the the Import Tool [1]. Do we need a new Project in StoryBoard? Should I use the 'Infrastructure' project? ([1] A short story - adding a link in the README to the guides on the Wiki) | 12:06 |
| pedroalvarez | petefoth: isn't easier to add the link without creating the story? :p | 12:07 |
| pedroalvarez | pedroalvarez: I think you should create another project for that | 12:08 |
| straycat | pedroalvarez, +1 | 12:08 |
| pedroalvarez | it doesn't belong to baserock/infrastructure | 12:08 |
| petefoth | pedroalvarez: but then we'd have no 'story' and all changes should (IMHO) be trackable to 'stories; I'll do it anyway, but.... | 12:09 |
| pedroalvarez | petefoth: I agree, let's create a new project for it | 12:09 |
| petefoth | and I agree it isn't infrastructure. I think we should have a 'Miscellaneous' project to cover stuff that isn't morph, definitions, infrastructure or wiki | 12:10 |
| petefoth | pedroalvarez: ^^ not sure we need a whole project just for the import tool | 12:11 |
| pedroalvarez | I wouldn't mind the creation of one project in storyboard for every project we have in git | 12:11 |
| pedroalvarez | so people interested in the import tool can go to that project and see what needs to be done | 12:12 |
| CTtpollard | that would be the goal of stroyboard, when gerrit is integrated properly into it | 12:12 |
| pedroalvarez | petefoth: also, the story you want to create, wouldn't it belong to baserock/wiki? | 12:12 |
| petefoth | I think we should get a few more opinions on this before we do tto much. | 12:12 |
| petefoth | pedroalvarez: no. The 'wiki' story is 'Have alink in the Import Tool wiki page to the README. Tha story is implemented (though it's not in SB). This stoiry is 'Add links in the Import Tool README'. | 12:14 |
| pedroalvarez | ahh! right | 12:14 |
| petefoth | I'll make the change without a SB story, but we should agree how to deal with all this fairly soon | 12:15 |
| petefoth | I'll track it in my employer's task tracking system thereby satisfying my demand for traceability :) | 12:17 |
| *** jmalk has left #baserock | 12:26 | |
| * bashrc builds system | 12:32 | |
| *** gary_perkins has quit IRC | 12:45 | |
| *** fay_ has quit IRC | 12:45 | |
| *** edcragg has quit IRC | 12:45 | |
| *** jonathanmaw has quit IRC | 12:45 | |
| *** flatmush has quit IRC | 12:45 | |
| *** a1exhughe5 has quit IRC | 12:45 | |
| *** bashrc has quit IRC | 12:45 | |
| *** sambishop has quit IRC | 12:45 | |
| *** paulw has quit IRC | 12:45 | |
| *** nowster has quit IRC | 12:45 | |
| *** ssam2 has quit IRC | 12:45 | |
| *** CTtpollard has quit IRC | 12:45 | |
| *** mauricemoss_ has quit IRC | 12:45 | |
| *** Krin has quit IRC | 12:45 | |
| *** franred has quit IRC | 12:45 | |
| *** edcragg has joined #baserock | 12:45 | |
| *** jonathanmaw has joined #baserock | 12:45 | |
| *** nowster has joined #baserock | 12:45 | |
| *** ssam2 has joined #baserock | 12:45 | |
| *** ChanServ sets mode: +v ssam2 | 12:45 | |
| *** franred has joined #baserock | 12:45 | |
| *** Krin has joined #baserock | 12:45 | |
| *** sambishop has joined #baserock | 12:45 | |
| *** mauricemoss_ has joined #baserock | 12:45 | |
| *** paulw has joined #baserock | 12:45 | |
| *** a1exhughe5 has joined #baserock | 12:45 | |
| *** fay_ has joined #baserock | 12:45 | |
| *** bashrc has joined #baserock | 12:45 | |
| *** CTtpollard has joined #baserock | 12:45 | |
| *** flatmush has joined #baserock | 12:45 | |
| *** gary_perkins has joined #baserock | 12:46 | |
| *** edcragg has quit IRC | 12:49 | |
| *** edcragg has joined #baserock | 12:51 | |
| pedroalvarez | <pedroalvarez> If we are releasing today 15.10, I think it should have latest morph: http://paste.baserock.org/sucusalebi | 12:55 |
| straycat | 12:08 straycat$ pedroalvarez, +1 | 13:12 |
| pedroalvarez | he | 13:12 |
| pedroalvarez | I have the ability to miss things on IRC :/ | 13:12 |
| pedroalvarez | straycat: feeling confident to give it a +2? | 13:13 |
| SotK | pedroalvarez: +1 from me too | 13:13 |
| pedroalvarez | yay | 13:13 |
| pedroalvarez | the release begins! | 13:13 |
| CTtpollard | :) | 13:14 |
| CTtpollard | are baserock planing to keep upto date with storyboard upstream? | 13:16 |
| CTtpollard | or with storyboard.baserock to be left as is? | 13:17 |
| pedroalvarez | CTtpollard: storyboard.b.o is running in fedora for now. If we ever have it running in baserock then I guess we can stay as close to upstream as possible | 13:21 |
| CTtpollard | cool, it's just so I can structure my response to petefoth's email to baserock-dev better | 13:38 |
| *** locallycompact has joined #baserock | 13:54 | |
| locallycompact | Hello, is there any chance to have qt5-devel things on cache.baserock.org? | 13:58 |
| *** tiagogomes_ has joined #baserock | 13:59 | |
| bashrc | looks like gnuplot isn't on baserock yet | 14:01 |
| pedroalvarez | locallycompact: do you mean to include a qt system in the ci? or just build a qt system once to populate cache.baserock.org? | 14:01 |
| locallycompact | pedroalvarez: The ci I imagine, I like to keep up with upstream definitions which means qt5 wants rebuilds often. | 14:02 |
| locallycompact | and I don't have the power to do it before the next change to devel generally :P | 14:03 |
| pedroalvarez | locallycompact: right, in that case it depends on how long takes the system to build | 14:03 |
| locallycompact | looooong | 14:04 |
| locallycompact | way long | 14:04 |
| pedroalvarez | :) | 14:04 |
| locallycompact | but I don't know how long on a capable ci server | 14:04 |
| pedroalvarez | or ci server is not that fast really, is a distbuild of a single node | 14:04 |
| pedroalvarez | s/or/our/ | 14:05 |
| pedroalvarez | If you have the resources, you can try to deploy a mason instance to build qt, to populate your artifact-cache-server | 14:06 |
| locallycompact | That sounds like fun | 14:07 |
| pedroalvarez | heh :) | 14:09 |
| pedroalvarez | I have one currently building from a development branch, and it makes things easier for the whole team that is working in the same project | 14:10 |
| kejiahu | what's the mechanism does the morph distbuild use to connect to build controller? I got an error on fail to connect to controller, but I can ssh into the controller at same time. http://paste.baserock.org/lenowerasa | 14:11 |
| pedroalvarez | kejiahu: take a look at the services running in 192.168.0.119 | 14:13 |
| pedroalvarez | `systemctl` | 14:13 |
| pedroalvarez | morph-controller and morph-controller-helper should be running | 14:14 |
| pedroalvarez | instead of ssh, you should use `telnet 192.168.0.119 7878` to see if something is listening in that port | 14:14 |
| kejiahu | pedroalvarez, thanks | 14:16 |
| * SotK notices the distbuild test in morph's ./check passing even though the distbuild failed | 14:21 | |
| Kinnison | :( | 14:21 |
| *** mdizzle has joined #baserock | 14:24 | |
| *** a1exhughe5 has quit IRC | 14:29 | |
| *** CTtpollard has quit IRC | 14:30 | |
| tiagogomes_ | radiofree, booting from m.2 with new boot layout | 14:39 |
| tiagogomes_ | I'll test an upgrade | 14:40 |
| *** a1exhughe5 has joined #baserock | 14:43 | |
| *** CTtpollard has joined #baserock | 14:44 | |
| *** zoli__ has quit IRC | 14:44 | |
| radiofree | tiagogomes_: quick checklist (/boot is / of your boot partition) 1) you have /boot/extlinux.conf 2) you have /boot/systems/factory 3) /boot/systems/default is a symlink to /boot/systems/factory | 14:44 |
| *** zoli__ has joined #baserock | 14:45 | |
| *** wschaller has joined #baserock | 14:45 | |
| tiagogomes_ | I should have all of those | 14:45 |
| radiofree | 4) your upgrade cluster contains BOOT_DEVICE: "/dev/sda1" and ROOT_DEVICE: "/dev/sda2" | 14:45 |
| radiofree | if that's the layout | 14:45 |
| tiagogomes_ | it is | 14:45 |
| tiagogomes_ | mmm, what I don't have is network connectivity in the node | 14:46 |
| radiofree | there was some discussion here about that the other day? | 14:51 |
| radiofree | Kinnison: richard_maw? | 14:51 |
| radiofree | something systemd related | 14:51 |
| richard_maw | hm what? | 14:53 |
| * richard_maw has no context | 14:53 | |
| radiofree | [14:46:45] <tiagogomes_> mmm, what I don't have is network connectivity in the node | 14:54 |
| richard_maw | there was some stuff which was networkd not handling ipv6 not being enabled in the kernel | 14:55 |
| richard_maw | which resulted in chewing CPU if DHCP=yes was set, and not being able to bring links up, because it couldn't initialise ipv6 on it | 14:56 |
| tiagogomes_ | DHCP=v4 is not working for me either | 14:57 |
| kejiahu | tiagogomes_: add 'scripts/config -e IPV6' in bsp kernel works on at least distbuild nodes | 14:57 |
| richard_maw | tiagogomes_: that's because systemd doesn't know how to bring the link up properly when ipv6 is not enabled | 14:58 |
| kejiahu | tiagogomes_: I didn't test on other system build, but I think they should work the same way | 14:58 |
| *** persia_ has quit IRC | 15:01 | |
| *** persia has quit IRC | 15:01 | |
| *** jonathanmaw has quit IRC | 15:06 | |
| *** jonathanmaw has joined #baserock | 15:09 | |
| *** a1exhughe5 has quit IRC | 15:09 | |
| *** a1exhughe5 has joined #baserock | 15:13 | |
| *** persia has joined #baserock | 15:13 | |
| *** persia_ has joined #baserock | 15:13 | |
| *** persia_ has joined #baserock | 15:13 | |
| pedroalvarez | Status report of the Baserock 15.10 release: Everything is going smooth, the systems are being deployed right now in our cloud infrastructure. | 15:16 |
| paulsherwood | pedroalvarez: cool! :) | 15:17 |
| jjardon | paulsherwood: When you said Baserock 15.10 release, you mean baserock project is releasing a specific version of the reference systems or what exactly? | 15:27 |
| pedroalvarez | jjardon: was that addressed to me? | 15:28 |
| paulsherwood | i think it's a release to coincide with genivi, which has an approximately 6-weekly cycle | 15:29 |
| jjardon | pedroalvarez: yep, sorry | 15:29 |
| pedroalvarez | A baserock release means publishin the reference systems defined in clusters/release.morph at the moment the release is being done | 15:30 |
| pedroalvarez | This specific release is being done because a) Morph has new dependencies (pylru) and it would be nice to have it in the latest released system b) I have to do a genivi release and releasing everyhing vs only genivi is just a matter of wrinting some nice release notes | 15:32 |
| pedroalvarez | c) there is probably more reasons :) | 15:33 |
| pedroalvarez | s/is/are/ | 15:33 |
| jjardon | pedroalvarez: rigth, thanks for the explanation. Maybe it would be better to call it Baserock Reference Systems 15.10 ? | 15:50 |
| perryl | when running check in morph, i'm getting an error that there is no pylru module. whereabouts does check look for this so i know where to put it? | 15:51 |
| richard_maw | either in your PYTHONPATH, or /usr/lib/python2.7/site-packages | 15:52 |
| * richard_maw just installed it on top with `pip install pylru` | 15:52 | |
| jjardon | perryl: this is how I run it in my baserock system: TMPDIR=/src/tmp PYTHONPATH="/src/morph:/src/pylru" ./check | 15:52 |
| richard_maw | as I wasn't in a position to update my system at that point | 15:52 |
| pedroalvarez | when i hit that issue I decided to install pylru as well | 15:52 |
| perryl | jjardon: that works for me, ta! | 15:53 |
| perryl | richard_maw: i see to recall i had an issue with pip install pylru not finding the package, though i'll try again so i'm not having to constantly set the path via cmdline | 15:54 |
| CTtpollard | that's odd perryl | 15:55 |
| * richard_maw maintains a .bashrc in his VM for that kind of thing | 15:55 | |
| perryl | CTtpollard: it's probably me doing something wrong when trying, wouldn't be the first time :) | 15:55 |
| pedroalvarez | python setup.py install worked for me :) | 15:56 |
| tiagogomes_ | radiofree, for the jetson's, boot device and root device are 2 partitions on the same disk? | 16:01 |
| *** franred has quit IRC | 16:03 | |
| radiofree | tiagogomes_: yes, i use BOOT_DEVICE: "/dev/mmcblk0p1" ROOT_DEVICE: "/dev/mmcblk0p2" | 16:04 |
| ssam2 | Draft 15.10 release notes! Please review: http://wiki.baserock.org/releases/baserock-15.10/ | 16:14 |
| ssam2 | lots of stuff changed, everyone has been working hard | 16:14 |
| ssam2 | calling it 'Baserock reference systems 15.10' might indeed be better, although it's a bit of a mouthful | 16:15 |
| CTtpollard | stoked that Apache is included | 16:16 |
| tiagogomes_ | It may be worth to say something about the improved build speeds | 16:16 |
| ssam2 | are they improved? build graph calculation is faster, but stuff will build at the same speed as ever I think | 16:17 |
| tiagogomes_ | ssam2 gcc usually took 1h to build and now builds in 10m, is that right pedroalvarez | 16:18 |
| pedroalvarez | maybe it wasn't that slow, but yeah, I've noticed some build speed improvements | 16:19 |
| straycat | pretty cool release | 16:21 |
| straycat | i'm really enjoying the --local-changes=ignore option | 16:22 |
| straycat | though it's forced me to rearrange my workspaces so that they correspond to the refs i'm working on which makes for a very long path, so my ps1 is now just '% ' | 16:23 |
| straycat | force isn't the right word | 16:23 |
| pedroalvarez | ssam2: s/QtCreater/QtCreator/ | 16:23 |
| straycat | 'encouraged' | 16:24 |
| pedroalvarez | ssam2: also, the "how do I get in contact?" section looks over indented | 16:24 |
| straycat | should we add virtualenv to the list of added components or does no one care about that? :) | 16:25 |
| CTtpollard | I'd say venv is a pretty big deal | 16:26 |
| straycat | s/updates/additions/ | 16:26 |
| tiagogomes_ | ssam2, the change that build-depends is not no longer necessary isn't on this release? | 16:26 |
| jjardon | Very nice release, congrats everyone! | 16:26 |
| pedroalvarez | indeed! well done everyone :) | 16:27 |
| richard_maw | do we need to bump the definitions format version for removing build-depends? | 16:27 |
| jjardon | ssam2: I would split the changes in sections "Reference systems:", "Morph:" ... but we can do it for the next one | 16:27 |
| pedroalvarez | richard_maw: IIRC that change was included in the patch | 16:28 |
| richard_maw | pedroalvarez: cool | 16:28 |
| pedroalvarez | but it hasn't been merged yet | 16:30 |
| richard_maw | oh | 16:30 |
| jjardon | reviewers welcomed! ;) | 16:31 |
| ssam2 | I don't want to add something about reduced GCC build time without knowing why it happened ... | 16:32 |
| ssam2 | maybe it was richard_maw's recent branch ? | 16:32 |
| ssam2 | no, that just touched split rules.. | 16:32 |
| ssam2 | I've made the following changes so far, thanks for all the comments: http://source.baserock.branchable.com/?p=source.git;a=blobdiff;f=releases/baserock-15.10.mdwn;h=97b7865696cef17050b8f0a7ecc7f2be7037986e;hp=07209b413cdb9a5f0097e1f603f70f165a122a3a;hb=20d6be383d33e7f50d8d35898e700e7b3355f073;hpb=bebdfa5580eb52cc7f393aa49f4bb6a781458641 | 16:33 |
| radiofree | i don't think anything in morph would have caused the build time improvements | 16:35 |
| radiofree | maybe gcc 4.9 is just quicker to build? tiagogomes_ enabled some optimisations in stage1/2 (i believe) which could also help things | 16:35 |
| pedroalvarez | tiago said that it was because he enabled some optimization flags in stage2-gcc | 16:35 |
| jjardon | ssam2: maybe is a good oportunity to announce the storyboard instance? or its still in "testing"? | 16:39 |
| tiagogomes_ | pedroalvarez heh, I said it could be | 16:41 |
| radiofree | maybe gcc 4.9 is faster at building gcc 4.9... | 16:42 |
| radiofree | could be quite a few things really | 16:42 |
| ssam2 | jjardon: not yet, I think | 16:42 |
| *** CTtpollard has quit IRC | 16:53 | |
| *** zoli__ has quit IRC | 17:06 | |
| *** zoli__ has joined #baserock | 17:06 | |
| *** edcragg has quit IRC | 17:09 | |
| *** edcragg has joined #baserock | 17:11 | |
| radiofree | tiagogomes_: i think i know why your u-boot doesn't sysboot | 17:14 |
| radiofree | there's no "do_ext2load", only "do_ext4load", however cmd_pxe doesn't know that.... | 17:16 |
| radiofree | looks like /boot would have to be a fat partition for it to sysboot | 17:16 |
| pedroalvarez | Genivi release notes draft: http://paste.baserock.org/esevosogeh | 17:18 |
| paulsherwood | pedroalvarez: just checking, in the release are we at version 0, or 1? | 17:19 |
| pedroalvarez | 0 | 17:20 |
| paulsherwood | pedroalvarez: could include a link there to the relevant meetup notes on | 17:20 |
| tiagogomes_ | radiofree fat? you mean FAT? | 17:20 |
| radiofree | tiagogomes_: yep | 17:20 |
| radiofree | tiagogomes_: for your board CONFIG_MMC is the one that sets CONFIG_CMD_EXT2 | 17:20 |
| radiofree | however it also does #undef CONFIG_MMC.... | 17:20 |
| radiofree | and only enables SATA, which only enables CONFIG_CMD_EXT4 | 17:21 |
| paulsherwood | pedroalvarez: otherwise all looks good to me +1 | 17:21 |
| paulsherwood | pedroalvarez: shall i do the spreadsheet thing again? | 17:21 |
| pedroalvarez | paulsherwood: actually, the release has been done from VERSION 0 but maybe morph can understand VERSION 1 if that patch was merged | 17:21 |
| radiofree | it would be a simple fix if you put your own version of u-boot on there, but you can't :( | 17:21 |
| tiagogomes_ | there's a firmware upgrade coming | 17:22 |
| radiofree | can you contact your vendor and ask them to enable CONFIG_CMD_EXT2 as well as CONFIG_CMD_EXT4? | 17:22 |
| pedroalvarez | paulsherwood: I probably should do it and stop bothering you with it | 17:22 |
| radiofree | tiagogomes_: sadly even upstream u-boot makes the check for CONFIG_CMD_EXT2, so if it's disabled in the config for your board in the same way it's never going to work | 17:23 |
| paulsherwood | pedroalvarez: i'm happy to do it. i just frab ybd and run it on definitions | 17:23 |
| paulsherwood | i should send a patch to morph to make it output better upstream version info in its log | 17:23 |
| radiofree | tiagogomes_: however I believe it would work if it was a FAT32 partition... but you can't do symlinks on that (right?) | 17:23 |
| radiofree | so.. not an option | 17:23 |
| straycat | you can't do symlinks on fat32? | 17:24 |
| radiofree | i don't know? | 17:24 |
| straycat | i can imagine that maybe you can't do hard links | 17:24 |
| pedroalvarez | paulsherwood: just wondering, were you suggesting adding a link to the relevant meetup notes in the genivi release notes or in the main release notes? | 17:25 |
| *** a1exhughe5 has quit IRC | 17:25 | |
| radiofree | tiagogomes_: you might be lucky enough to get a more modern version in the firmware update, in which case "sysboot scsi 0 any ...." | 17:26 |
| radiofree | if you have a good relationship with vendor you might be able to get to enable CONFIG_CMD_EXT2 for the board | 17:27 |
| straycat | radiofree, looks like you can't, i didn't know that | 17:27 |
| radiofree | straycat: i think you can on ntfs | 17:27 |
| straycat | symlinks are basically just a file with a path in them, so i'm a bit surprised | 17:28 |
| SotK | radiofree: I'm pretty sure that is the case | 17:28 |
| radiofree | tiagogomes_: if you're feeling particularly adventurous you could try building your own u-boot and chain-loading it.... but you'd probably want to have physical access to the machine in question if you did that | 17:29 |
| tiagogomes_ | radiofree not in a friday | 17:30 |
| *** mdizzle has quit IRC | 17:30 | |
| radiofree | tiagogomes_: :) | 17:30 |
| radiofree | maybe something like http://fpaste.org/194402/14256630/ | 17:30 |
| tiagogomes_ | system-version-manger worked fine with a /boot partition, so I'm good | 17:31 |
| radiofree | :D | 17:31 |
| paulsherwood | pedroalvarez: main release notes (the versioning bit) | 17:31 |
| jjardon | pedroalvarez: is http://mason-x86-64.baserock.org/ running latest morph? If not it will break | 17:31 |
| tiagogomes_ | radiofree, what's u-boot.bin? | 17:31 |
| paulsherwood | jjardon: what will break it? | 17:31 |
| radiofree | tiagogomes_: the version of u-boot you build from source | 17:31 |
| paulsherwood | oh, the build-depends :) | 17:31 |
| tiagogomes_ | ah, you load u-boot with u-boot? nasty | 17:32 |
| ssam2 | pedroalvarez: ah, we need to update the masons before we merge the definitions patch to remove build-depends: []... good point | 17:33 |
| ssam2 | jjardon: maybe best to revert for now, i don't think any of us are going to want to spend friday night upgrading masons | 17:34 |
| jjardon | ssam2: sure, will do | 17:36 |
| ssam2 | thanks :) | 17:36 |
| ssam2 | what we need is masons that upgrade to master after every build, but that's a bit of a way off | 17:36 |
| jjardon | ssam2: is that on storboard? If not I will add an entry | 17:37 |
| ssam2 | please do | 17:37 |
| pedroalvarez | the upgrade of mason should be easy, I can take care of it if we don't want to revert the patch | 17:38 |
| jjardon | pedroalvarez: I let you decide if revert or upgrade mason then | 17:39 |
| *** jonathanmaw has quit IRC | 17:58 | |
| *** ssam2 has quit IRC | 17:59 | |
| pedroalvarez | Done! Baserock 15.10 is released! | 17:59 |
| *** bashrc has quit IRC | 18:00 | |
| pedroalvarez | jjardon: has it been merged then? | 18:00 |
| pedroalvarez | I don't see mason complaining | 18:00 |
| jjardon | pedroalvarez: because I think its in the middle of a masive build | 18:01 |
| pedroalvarez | jjardon: why massive? | 18:01 |
| jjardon | (busybox has been upgrade) | 18:01 |
| pedroalvarez | oooo | 18:01 |
| jjardon | pedroalvarez: so I guess you have plenty of time to upgrade mason if you want to follow that route ;) | 18:02 |
| pedroalvarez | upgrade mason == 10 minutes(ish) | 18:02 |
| pedroalvarez | I'll do that over the weekend, now they are busy building stuff | 18:04 |
| *** nowster has quit IRC | 18:05 | |
| tiagogomes_ | 2015-03-06 18:07:08 Starting build 278d07d9cd1c403db509065d6dcd8b58 | 18:07 |
| tiagogomes_ | 2015-03-06 18:07:08 Collecting morphologies involved in building systems/devel-system-armv8l64.morph from baserock/tiago-ed/m2-moonshot-deployment | 18:07 |
| tiagogomes_ | ERROR: Stratum NetworkManager-common has no build dependencies for chunk libndp in strata/NetworkManager-common.morph | 18:07 |
| pedroalvarez | tiagogomes_: use latest morph | 18:07 |
| tiagogomes_ | right | 18:07 |
| pedroalvarez | jjardon: you havent upgraded morph sha1 when merging the patch for morph :/ | 18:07 |
| pedroalvarez | so current definitions can't build themselves | 18:08 |
| tiagogomes_ | :( | 18:08 |
| pedroalvarez | Now I think that reverting the patch is the best option | 18:09 |
| pedroalvarez | then upgrade morph sha1 in definitions.git | 18:10 |
| pedroalvarez | and then merge the patch again | 18:10 |
| *** gary_perkins has quit IRC | 18:13 | |
| pedroalvarez | sha256 checksums updated, I think I'm done now | 18:13 |
| jjardon | pedroalvarez: sorry, didnt realize about that and nobody pointed out that in the review | 18:14 |
| pedroalvarez | no worries. People normally upgrade the sha1 in definitions when doing big changes in morph.git | 18:15 |
| jjardon | pedroalvarez: Do you want me to revert or are you going to do it? Wouldnt be better to upgrade the sha for morph and we are done? | 18:16 |
| *** edcragg has quit IRC | 18:17 | |
| pedroalvarez | jjardon: then.. how can I build a system that can build version 1? | 18:18 |
| jjardon | pedroalvarez: you can manually upgrade morph in that system? Maybe not ideal though. | 18:24 |
| jjardon | pedroalvarez: but yeah, I can revert and reapply the patches now if you think is better | 18:25 |
| *** tiagogomes_ has quit IRC | 18:26 | |
| *** rdale has quit IRC | 18:31 | |
| pedroalvarez | yeah, I know is possible, but is not nice though | 18:32 |
| pedroalvarez | not to many of us around to make a decission | 18:32 |
| pedroalvarez | I'd say revert definitions.git patch and then upgrade morph sha1 | 18:33 |
| pedroalvarez | whoever upgrades the sha1 has my +2 | 18:33 |
| *** zoli__ has quit IRC | 18:33 | |
| jjardon | pedroalvarez: will do it before going to the pub ;) | 18:36 |
| straycat | not sure it's worth reverting, we've already got a point in history where definitions won't be buildable, reverting won't fix that? | 18:39 |
| * straycat needs to go to the vet | 18:39 | |
| *** wschaller has quit IRC | 18:41 | |
| jjardon | I think its still worth reverting to remember to someone thats looking the git history how the thinks should be done properly | 18:47 |
| jjardon | Fix pushed | 18:48 |
| straycat | haha, we're getting steam in baserock? | 19:53 |
| *** radiofree has quit IRC | 20:00 | |
| *** gfinney has quit IRC | 20:04 | |
| * SotK looks forward to it | 20:09 | |
| straycat | SotK, are you still around? | 20:58 |
| *** persia is now known as persia__ | 21:32 | |
| *** persia_ is now known as persia | 21:32 | |
| *** persia__ is now known as persia_ | 21:32 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!