**** BEGIN LOGGING AT Fri May 29 02:59:59 2015 May 29 03:39:00 Crofton|work: did you ever ack? May 29 06:16:46 morning all May 29 07:12:17 My custom kernel is getting an IP by DHCP and uses NFS for the rootfs, during kernel boot I can see the IP config with nameservers and the system boots nicely. After it is booted I can ping IPs but no hostnames as it doesn't resolve them using the nameservers. resolv.conf is empty and when I place my nameservers in it everything works. Is there any way of auto configuring the resolv.conf for the nameservers my kernel fetches during boot? May 29 07:24:53 RagBal: perhaps u can take a look at the resolvconf recipe and go from there May 29 07:26:57 challinan, I have tried resolvconf but it doesn't seem to add the nameservers to resolv.conf May 29 07:27:16 challinan, sorry meant chankit May 29 07:35:01 RagBal: I think you can take a look the recipe and check which file inside the temp dir gets copied to /etc/resolv.conf. Then, u know where you can add the name servers May 29 07:35:29 not the most elegant one though. let's see if there's more expert here :-) May 29 07:39:00 RagBal: perhaps this page http://www.yoctoproject.org/docs/1.4.2/dev-manual/dev-manual.html#usingpoky-extend-addpkg-postinstalls can provide better clues May 29 07:50:32 parrot, thanks, will have a look at it May 29 08:17:29 is there a kind of metapackage that just installs all kernel modules without having to list time explicitly? May 29 08:20:08 ah, probably just "kernel-modules" :) May 29 09:06:42 Hello guys. I'm trying to write a recipe for the software tinyproxy 1.8.4. There is a `autogen.sh` and there are some Makefile.am in some folders. I tried with just `inherit autotools` but I got this error: https://gist.github.com/943c5c1b6e5dd0cdbb86 May 29 09:06:42 It seems that the `autogen.sh` does all the necessary work here, but I couldn't figure out if autotools-class also uses this and I couldn't find another recipe where there is also a `autogen.sh` involved. May 29 09:08:24 zwerch: did u define any methods in your recipe? May 29 09:10:16 parrot: no, i thought autotools would do all the work for me May 29 09:10:25 otavio: btw, thanks for the status update, very glad to hear that a fix is in the queue! May 29 09:11:52 zwerch: did u check log.do_compile to make sure that tinyproxy is indeed compiled? May 29 09:14:52 parrot: just did, you were right: May 29 09:14:52 NOTE: nothing to compile May 29 09:14:52 so, how do I approach this? execute the autogen.sh in the do_configure and then what? May 29 09:15:24 zwerch: so you might want to go one further and check log.do_configure? May 29 09:16:29 parrot: May 29 09:16:29 DEBUG: Executing shell function do_configure May 29 09:16:29 NOTE: nothing to configure May 29 09:16:50 parrot: so, autotools does not help at all here? May 29 09:18:29 zwerch: might have something to do with source directory thingy....I'm actually surprised do_configure and do_compile didnt complain at all. May 29 09:18:39 do you use git or http downloads to get the source? May 29 09:18:57 parrot: git May 29 09:19:33 zwerch: add S="${WORKDIR}/git" to your recipe and see if it's any better? May 29 09:20:29 hi all May 29 09:22:12 I need some help: I've got a kernel panic error -8 (NOEXEC). please can anyone give some hint May 29 09:22:50 parrot: i'm gonna try, mom May 29 09:26:59 zwerch: gonna leave now so good luck :-) bluelightning would be able to help if you are still stuck May 29 09:28:10 bluelightning: first, please say thanks to parrot, he really helped, but now I'm stuck with this: https://gist.github.com/0dc34f9890927b15ff3d May 29 09:32:06 bluelightning: it seems that this is a problem with tinyproxy itself, found an old version that did not have this problem, gonna look for a fix now, thanks May 29 09:42:10 I'm currently experiencing a weird issue where my bash 4.3.30(1) can not handle subshell IO redirection... is that already known? May 29 09:42:13 # echo foo &> >(logger -s) May 29 09:42:14 -sh: syntax error near unexpected token `>' May 29 10:20:13 bluelightning: How can I build a native package without fetching it ? May 29 10:21:47 E.g. I need to build a native version of postgresql. But adding BBCLASSEXTEND="native" gives fetch error. As is tries to fetch source for postgresql-native.tar.gz May 29 10:25:16 the SRC_URI line for the postgres recipe presumably contains ${PN} May 29 10:25:24 change that to ${BPN} May 29 10:29:09 ok..let me try May 29 10:34:14 joshuagl: Yes, it was ${P} actually. So changing to ${BP} worked. Thanks! May 29 10:35:00 np May 29 10:41:44 nerdboy, busy with a work conference May 29 10:54:56 Can I tell autotools to add some parameters to `configure`? May 29 10:55:12 I would like to stage a file only for the native package build using a common recipe. How should I do it? May 29 10:58:08 zwerch: yes, you need to define EXTRA_OECONF in your recipe with the extra options you want to pass May 29 10:58:53 joshuagl: thanks! i've seen this name quite often, but i never thought what it was for :D May 29 10:59:06 milan: you can define a do_install_append_virtclass-native() May 29 10:59:28 ok May 29 11:36:21 hi May 29 11:36:44 bluelightning: hmm, PATH=/sbin:$PATH did not quite work out as it got replaced by Yocto's development environment PATH May 29 11:36:53 bluelightning: is there a way to echo literally? May 29 11:37:45 ah, backslash! May 29 12:56:20 bluelightning: when I su foo, my path does not get set because I think /home/foo/.profile is not sourced May 29 12:56:26 but I do not have anything like su - either. May 29 13:04:12 oh, I just had to use busybox.suid su - nic due to the separation (again). May 29 13:15:27 otavio: hey, we are affected by another Vivante driver Wayland bug, this time https://bugzilla.yoctoproject.org/show_bug.cgi?id=7825. Could you please forward that to the GPU guys again? May 29 13:15:28 Bug 7825: normal, Undecided, ---, otavio, NEW , Vivante: Spurious frame sent to compositor after hiding Wayland window May 29 13:15:41 As before, contains a testcase and a detailed description of the problem. May 29 13:21:43 tmcguire: great bug report May 29 13:22:39 thanks. I have been on the receiving end of bug reports for a long time, so I guess writing them is easy for me :) May 29 13:23:22 :) May 29 13:24:23 I wish Vivante had a bug tracker, reporting to yocto.org feels kinda wrong. May 29 13:26:26 yoctoproject.org :) May 29 13:35:45 anyone know what the correct way to include locale data etc. in images is? I can't find anything in the docs May 29 13:36:07 currently I'm just including LIBC_LOCALE_DEPENDENCIES in IMAGE_INSTALL, but there must be a cleaner way? May 29 13:38:19 tmcguire: yes May 29 13:39:03 joshuagl: IMAGE_LINGUAS is not what you are after either? May 29 13:41:36 lpapp: I thought it might be, but setting it doesn't add any extra stuff to my image May 29 13:42:07 joshuagl: is there extra stuff created though? May 29 13:44:35 the locale stuff is all created, it's just not included in my images May 29 13:44:49 I guess they end up in core-image-sato through some dependency chain... May 29 13:46:16 * joshuagl decides to spend time with the output of bitbake -g May 29 13:47:52 bitbake -g -u depexp is also useful from time to time May 29 13:48:46 kergoth: aye, very familiar with that one May 29 13:48:52 sadly running on a headless system May 29 13:54:33 that is not sadly, headless systems are cool! :-) May 29 14:00:59 depexp is cooler :-p May 29 15:54:49 when I want to add a service on my device I type "systemctl enable x.service" into the console, this creates symlinks in the /etc/systemd/system/y.wants dir May 29 15:55:14 If I want this to be a part of my image, do I manually create those symlinks or is there some more automatic way to do it? May 29 15:55:22 i.e. I want to enable a service by default May 29 15:56:58 gardarh: your recipe should inherit systemd and set the appropriate variables May 29 15:57:11 (the recipe that installs the service file, that is) May 29 15:57:35 gardarh: there are some pointers here: http://www.yoctoproject.org/docs/current/ref-manual/ref-manual.html#ref-classes-systemd May 29 16:06:16 bluelightning: Thanks, was actually already trying to inherit systemd.bbclass but stumbled upon that section in the docs, it's quite good May 29 16:06:24 guess I should look at the docs before googling :) May 29 16:07:01 no problem, unfortunately the docs (or at least the latest docs) don't always show up in google May 29 16:07:21 they really should, but we don't have full control over that as you might guess May 29 16:07:39 right :) May 29 16:42:15 I've got a recipe that clones another git repo during the compiling. when trying to build the recipe with bitbake, the clone fails. May 29 16:42:21 any guesses why? May 29 16:42:39 fatal: unable to access 'https://github.com/USCiLab/cereal.git/': Failed to connect to github.com port 443: Connection timed out May 29 16:43:23 not offhand, but that's not ideal. It means the recipe doesn't obey BB_NO_NETWORK and the like. ideally, *any* contact to external sources should be done by our fetching May 29 16:43:42 i'd recommend adding that repo to SRC_URI with subdir= set appropriately May 29 16:49:01 erm May 29 16:49:13 is BB_NO_NETWORK material in this case? May 29 16:49:31 it sounds like a simple connectivity issue May 29 16:50:35 if you are behind some form of proxy it may be that parts of the environment that are needed to enable a proxy connection aren't exported through to the environment in which do_compile runs May 29 16:52:09 yeah, i'm sure it is, but the recipe should still be fixed :) May 29 16:53:51 right, I agree it shouldn't be doing any fetching outside of do_fetch May 29 16:53:58 npm is a massive offender in that regard :/ May 29 16:55:40 kergoth: what do you mean by "our fetching"? May 29 16:56:49 I've been exporting http_proxy before running bitbake... but it's unclear if that environment is carrying through May 29 16:56:58 (and https_proxy...) May 29 16:57:20 tripzero: within do_fetch May 29 16:57:41 (i.e. that's what he meant) May 29 16:58:12 ahh. May 29 16:59:51 * rink__ bestelt bij new-line.nl es wat spul May 29 16:59:55 oops, wrong window, sorry May 29 17:02:48 bluelightning: npm being nodejs? May 29 17:02:58 tripzero: right, yes May 29 17:03:13 root May 29 17:03:38 npm is not a separate package from nodejs I think May 29 17:03:46 but its origins are nodejs May 29 17:04:11 yeah, I guess I really meant to say nodejs rather than npm May 29 17:04:52 I've been more on the periphery of building and packaging it rather than actually using it May 29 17:05:50 BB_NO_NETWORK is an environment variable? May 29 17:05:59 yeah, I think I have done a recipe for it in recent past May 29 17:06:13 tripzero: its a bitbake variable May 29 17:06:26 you set it in config metadata somewhere May 29 17:06:36 local.conf is a prime location May 29 17:06:52 BB_NO_NETWORK isn't going to actually help resolve this situation, in case that's what you're hoping... May 29 17:07:29 all it does is tell bitbake's fetching code to blow up if it actually gets to the point of trying to fetch something from the network May 29 17:07:41 ahh May 29 17:07:47 (for cases where you don't want anything to be fetched) May 29 17:08:12 in this case bitbake's fetching code isn't involved, so it wouldn't be affected by BB_NO_NETWORK May 29 17:08:20 that makes sense May 29 17:08:36 so it's some other cause in my case. May 29 17:08:48 like proxy or general connectivity issues May 29 17:12:21 tripzero: your network is blockinh 443 and git protocol i think May 29 17:12:50 Can you try a normal git clone https://github.com/USCiLab/cereal.git May 29 17:12:54 does that work on this box May 29 17:13:25 and share the assignment of SRC_URI in your recipe as well May 29 17:13:37 for more details May 29 17:13:45 yes May 29 17:14:13 i can clone it on the box outside of bitbake May 29 17:14:18 ok May 29 17:14:27 then show recipe plz May 29 17:15:25 SRC_URI = "git://gerrit.iotivity.org/gerrit/iotivity;protocol=https;name=iotivity \ May 29 17:15:51 here's a pastebin of the entire recipe: http://paste.ubuntu.com/11436104/ May 29 17:33:14 khem`: anything obviously wrong with the recipe? May 29 17:33:35 didnt look let me see May 29 17:34:19 hmmm May 29 17:34:25 you showed me a github link May 29 17:34:33 this recipe has some gerrit link May 29 17:34:54 gerrit uses ssh based authentication May 29 17:35:01 so you need to set it up May 29 17:35:22 in your ~/.netrc you should set the username/password for gerrit access May 29 17:35:54 machine gerrit.iotivity.org May 29 17:36:52 hmmm May 29 17:36:54 git clone https://gerrit.iotivity.org/gerrit/iotivity May 29 17:36:56 works for me May 29 17:47:07 * kergoth ponders May 29 18:36:54 Hi, I use dizzy and I'm trying to build an image with systemd but somehow sysvinit is built and it RDEPENDS on "initscripts-functions" which can't be installed. Can anyone help me with that? May 29 18:48:14 kido, how did you select systemd May 29 18:48:20 RP: around ? May 29 18:49:10 http://www.thelins.se/johan/blog/2014/06/yocto-part-i-baseline-boot-time/ see the first four lines in this one May 29 18:50:00 which add systemd to the features, set systemd to be the init_manager and then remove sysvinit from considerations + make sure initscripts don't get installed May 29 18:50:58 khem`: http://pastebin.com/Dk7mMMgE like that May 29 18:51:30 and I added manualy systemd to DISTRO_FEATURES some lines before May 29 18:51:44 so basicaly like your instructions JEEBsv May 29 18:51:45 you need DISTRO_FEATURES_append = " systemd" May 29 18:51:46 I think it's not necessary to override the udev choice since the addition of systemd to features should have it provide that package May 29 18:52:20 at least I didn't have to do it, but I started with 1.7 so~ May 29 18:52:21 khem`: even if I added systemd manualy to DISTRO_FEATURES? May 29 18:52:52 that's the same effect methinks :P if it's in DISTRO_FEATURES one way or another it's there May 29 18:53:04 indeed :/ May 29 18:53:35 DISTRO_FEATURES goes across whole metadata May 29 18:55:01 DISTRO_FEATURES_append = " systemd" May 29 18:55:01 DISTRO_FEATURES_BACKFILL_CONSIDERED += "sysvinit" May 29 18:55:01 VIRTUAL-RUNTIME_init_manager = "systemd" May 29 18:55:01 VIRTUAL-RUNTIME_initscripts = "systemd-compat-units" May 29 18:55:03 would do it May 29 18:56:07 alright, I will try. thank you :) May 29 19:49:50 hi all! anyone can help me plz? May 29 20:26:01 halstead: how long would it take to throw up a placeholder for the compliance page on yp.org? May 29 20:28:21 darknighte, Do we have content already for the placeholder? Or can it be blank? May 29 20:29:33 No content, but AFAIAC, it could be blank, or "coming soon" May 29 20:31:03 halstead: ^^ May 29 20:40:01 darknighte, Should be easy. Do you just need the URL or should it be linked into the menus? May 29 20:41:07 halstead: not sure what the difference will mean to the end user May 29 20:41:26 halstead: page in question is already live: https://www.yoctoproject.org/ecosystem/yocto-project-compliance-program May 29 20:41:48 * darknighte considers May 29 20:43:16 yeah, ideally, there would be no reference to a "compliance program" at all, in the menus or otherwise May 29 20:54:10 darknighte, I'm sorry. I think I'm missing context. What's the goal? May 29 21:24:28 bluelightning: my current thinking is to create a `recipetool appendsrcfile` subcommand, which works much like appendfile, but instead of operating against target paths / target files, it arranges to overwrite files in WORKDIR or S before the build starts May 29 21:24:33 thoughts? May 29 21:26:06 kergoth: that sounds OK to me; the alternative I had thought of was to add an option to appendfile but then it's probably got enough options already and it's a different use case, so a separate command might be better as you suggest May 29 21:29:16 Hmm, I'll prototype it this way, and if it has a lot of overlap and feels like it might be better merged, we can do that, but probably better to have small focused commands than larger ones, and just shift common code into common modules May 29 23:42:06 hello, hello, i'm trying to build a image with read-only-rootfs it works if the rootfs is ext3 or 4 but not if its squashfs. it boot but the systemd fail to start networkd (at least) and there is absolutly no logs (journalctl is empty) May 29 23:42:22 any clues ? **** ENDING LOGGING AT Sat May 30 02:59:59 2015