**** BEGIN LOGGING AT Tue Nov 02 02:59:57 2010 Nov 02 06:26:43 gm Nov 02 06:55:39 ka6sox-work: ping Nov 02 06:57:13 or ka6sox3 :-) Nov 02 07:02:55 hmm.. Nov 02 07:03:00 eFfeM_work, pong Nov 02 07:03:03 morning Nov 02 07:03:12 morning Jay7 Nov 02 07:03:20 ka6sox-work: you administer melo ? Nov 02 07:03:21 gm all Nov 02 07:03:22 * Jay7 have failed all 3 test builds on libfontenc.. Nov 02 07:03:30 eFfeM_work, I try :D Nov 02 07:03:33 likewise: are you up early or up late ? Nov 02 07:03:37 early :-) Nov 02 07:03:47 late here. Nov 02 07:04:07 eFfeM_work, wazzup? Nov 02 07:04:15 ka6sox-work: I was wondering triggered by an email on the ML whether there is anything at all in http://sources.openembedded.org/ Nov 02 07:04:20 ping tells me it is melo Nov 02 07:04:27 or is this just an empty placeholder Nov 02 07:05:03 can't browse it, figured you could peek in the config files and dir Nov 02 07:05:26 I've not heard of that subdomain. Nov 02 07:05:31 but I will look at the configs Nov 02 07:05:32 jas Nov 02 07:05:47 tnx Nov 02 07:05:49 it exists: Nov 02 07:05:50 frans@frans-desktop:~/workspace/openembedded/conf$ ping sources.openembedded.org Nov 02 07:05:50 PING sources.openembedded.org (140.211.169.165) 56(84) bytes of data. Nov 02 07:05:50 64 bytes from melo.openembedded.org (140.211.169.165): icmp_seq=1 ttl=51 time=188 ms Nov 02 07:06:32 yes, I can believe that its a CNAME...I just need to see what the webserver thinks. Nov 02 07:06:44 oh it resovles Nov 02 07:06:54 this is the content Nov 02 07:06:55 Openembedded main mirror for upstream source code Nov 02 07:07:16 so there is actually a server behind it Nov 02 07:07:20 hmmm...okay I don't know who would populate that. Nov 02 07:08:01 apparently rolf did something there, but I have no idea what is below it (if anything at all) Nov 02 07:08:10 iirc, Crofton was switched mirrors site to melo yesteray Nov 02 07:08:22 ah..okay its populated. Nov 02 07:08:26 let me check permissions Nov 02 07:09:04 ya, looks like permissions aren't quite right Nov 02 07:09:08 are you seeing anything? Nov 02 07:09:19 only this: Nov 02 07:09:19 Openembedded main mirror for upstream source code Nov 02 07:09:45 which I guess is index.html or index.php Nov 02 07:09:45 that appears as index.txt Nov 02 07:10:00 browsing http://sources.openembedded.org/ for that Nov 02 07:10:11 I see a number of sources including md5sums but the permissions look wrong. Nov 02 07:11:02 hmm Nov 02 07:11:08 is it by any chance recent? Nov 02 07:11:56 the sources are from 5yrs ago to last week. Nov 02 07:12:10 ah ok, so it seems autopopulated one way or another Nov 02 07:12:25 actually it looks like a tar extraction Nov 02 07:12:46 but the ownership seems busted. Nov 02 07:12:56 it is not the same as the angstrom src mirror, is it? Nov 02 07:12:57 group: rolf. Nov 02 07:13:02 no its not. Nov 02 07:13:27 well..the apache config doesn't support angstrom domain names. Nov 02 07:13:34 (at this time) Nov 02 07:15:05 hm, guess we need to find out how to get it working and how it gets its data (or ask rolf, but I have no idea where he hangs out nowadays, might still have some email address from him though) Nov 02 07:16:14 ah ok, I think he also mentioned its existence on the list Nov 02 07:16:37 right...but the group perms are wrong...let me fix that. Nov 02 07:50:29 good morning Nov 02 07:51:51 nn Nov 02 07:54:21 * Jay7 is looking on http://tinderbox.openembedded.net/public/logs/task/9477641.txt Nov 02 07:54:56 hm.. is font-util required package now? Nov 02 08:13:50 ah.. that was fixed already in master Nov 02 08:28:56 morning Nov 02 08:30:23 hi hrw Nov 02 08:30:54 hi eFfeM_work Nov 02 08:30:58 how was OEDEM? Nov 02 08:35:03 interesting discussions wrt yocto, but little got decided Nov 02 08:35:14 don't have my notes handy Nov 02 08:36:30 actually there was supposed to be an announcement yesterday, but didn't see it yet Nov 02 08:37:31 ~curse my 24" lcd Nov 02 08:37:32 May you be reincarnated as a Windows XP administrator, my 24" lcd ! Nov 02 08:37:44 backlight died... need to use 20" :( Nov 02 08:45:05 hi there Nov 02 08:46:40 03Koen Kooi  07org.openembedded.dev * r30e362c22a 10openembedded.git/recipes/vlc/ (vlc.inc vlc_1.1.4.1.bb): Nov 02 08:46:40 vlc: update to 1.1.4.1 Nov 02 08:46:40 Signed-off-by: Koen Kooi Nov 02 08:46:45 eglibc-initial-2.12-r15.6+svnr11762 fails to compile for me (armv4t arch, rx1950 machine, jlime-2010.1 distro) Nov 02 08:47:03 actually, it fails to install Nov 02 08:47:19 here's tail of install log: http://dpaste.org/hnHf/ Nov 02 08:53:10 can anyone help me to fix this issue? Nov 02 09:25:49 mickeyl, cbrake can I be added to http://wiki.openembedded.org/index.php/OpenEmbedded_Developers ? I'd do that myself but I am not sure about my git_key_name Nov 02 09:49:07 good morning Nov 02 10:08:32 has the OEDEM minutes been posted somewhere? Nov 02 10:12:38 not sure, but likewise sent me a copy at least Nov 02 10:12:48 hi mickey|office Nov 02 10:13:10 esbenh: I'm editing the notes so they are readable for non-attenders. Nov 02 10:13:14 florian: ^ Nov 02 10:13:32 likewise: oh that's great! Nov 02 10:16:08 let me focus and finish that right away :-) Nov 02 10:16:26 likewise: great, I look forward to reading them :-) Nov 02 10:20:46 hi Nov 02 10:22:42 hi ericben Nov 02 10:26:18 hi florian Nov 02 10:48:00 likewise Nov 02 10:48:11 thanks, they are pretty incoheerent at times Nov 02 10:48:35 for non-attenddees, the minutes are collected collaboratively with gobby Nov 02 10:57:10 how do I make bitbake build 2.6.36rc kernel? bitbake linux-omap_2.6.36bb fails.. and it by default picks 2.6.29 Nov 02 10:57:13 Crofton|work: I didn't add much coherency, just added a bit of context here and there. Nov 02 11:00:57 heh Nov 02 11:01:32 thanks, we do need to spread the OE work load around Nov 02 11:02:30 nm got it Nov 02 11:05:56 * XorA spreads more of it USA direction Nov 02 11:06:41 we are all asleep Nov 02 11:06:59 * XorA whips the lazy merkins, get up yah lazy bums Nov 02 11:07:20 hehe Nov 02 11:08:27 Crofton|work: enjoy trip to UK? Nov 02 11:08:49 morning folks Nov 02 11:08:55 hey hey mr mickey|office Nov 02 11:09:38 yes Nov 02 11:09:43 too much work time though Nov 02 11:09:47 heh Nov 02 11:10:04 * XorA was kayaking round Loch Lomond while you guys were discussing OE Nov 02 11:10:11 I did stay with my family there the night before I left Nov 02 11:11:08 likewise: minutes did not land here yet Nov 02 11:11:54 also, no minutes here yet... Nov 02 11:12:01 I visited Kennedy's Space Center during oedem Nov 02 11:12:34 * XorA thinks all future OE meetings should be in mountains with lakes Nov 02 11:14:28 if you want to come around bordeaux, there is the ocean, no mountains but quite interesting castle to visit and taste ;-) Nov 02 11:15:42 I want a tropical island Nov 02 11:18:14 XorA: we know, how many in the kayak? :-) Nov 02 11:19:17 likewise: single sea kayaks Nov 02 11:19:39 ericben: for people from uk the dunes might count as mountains ;-) Nov 02 11:19:46 did you practice eskimo rolls? :-) Nov 02 11:20:05 florian: and at the bottom of the dune it's possible to practice sea kayak ;-) Nov 02 11:20:11 florian: for people from .nl every bump in the road counts as a mountain Nov 02 11:20:42 hehe Nov 02 11:20:44 florian: San Francisco would be ok for the next OEDEM (especially if LF hires a plane for us) Nov 02 11:21:14 or multiple flights (to reduce the risk of OpenEmbedded crashing) Nov 02 11:22:32 eFfeM_work, esbenh: I had to re-subscribe and re-send, apparently my account overflooded during my absence. Should be there now, if I may post... Nov 02 11:22:52 i'll check Nov 02 11:23:02 yup it is there Nov 02 11:23:12 likewise: no, I prefer the option of staying upright, but I need to spend winder doing my self recues so I can work on getting my trip leader qualifications next year Nov 02 11:24:39 * likewise votes for mountain with lake, has slight preference for Swedish gi^W^WSweden. Nov 02 11:24:56 likewise: Id agree with that :-) Nov 02 11:29:48 so openembedded stable in december and them sort of merge of oe & yocto ? Nov 02 11:29:58 hi mickey|office Nov 02 11:31:33 ok lets break busybox :-D Nov 02 11:31:39 03Eric Bénard  07org.openembedded.dev * r3be416325f 10openembedded.git/recipes/busybox/ (51 files in 17 dirs): (log message trimmed) Nov 02 11:31:39 busybox: remove unused version Nov 02 11:31:39 keep the following versions : Nov 02 11:31:39 1.2.1 (used by oplinux) Nov 02 11:31:39 1.7.2 (used by nylon) Nov 02 11:31:39 1.13.2 (used by angstrom, kaeilos and slugos) Nov 02 11:31:39 1.17.1 (used by angstrom) Nov 02 11:31:49 03Eric Bénard  07org.openembedded.dev * r0356d3dccf 10openembedded.git/recipes/busybox/busybox-static_1.17.3.bb: Nov 02 11:31:49 busybox-static: add the latest version Nov 02 11:31:49 tested using initramfs-image Nov 02 11:31:49 Signed-off-by: Eric Bénard Nov 02 11:31:49 Acked-by: Khem Raj Nov 02 11:31:49 Acked-by: Frans Meulenbroeks Nov 02 11:31:50 03Eric Bénard  07org.openembedded.dev * rcd5170857a 10openembedded.git/recipes/busybox/busybox_1.17.3.bb: Nov 02 11:31:51 busybox: 1.17.3 is stable and can be the default fallback version Nov 02 11:31:51 Signed-off-by: Eric Bénard Nov 02 11:31:51 Acked-by: Khem Raj Nov 02 11:31:52 Acked-by: Frans Meulenbroeks Nov 02 11:33:55 ericben afaik it was decided to have a discussion on whether to merge Nov 02 11:34:36 ti merge what ? Nov 02 11:34:41 ah yocto & oe ? Nov 02 11:36:09 yes Nov 02 11:37:50 that needs a lot of discussion Nov 02 11:38:04 the yocto thing was a bit of a surprise Nov 02 11:39:16 likewise: thanks for the minutes, I feel they are quite accurate (and yes, at some times it was chaotic) Nov 02 11:39:23 away for lunch Nov 02 11:39:51 Crofton|work: we cant complain about some paid engineers though :-) Nov 02 11:40:06 of course wwe can Nov 02 11:40:12 we can complain about anything :) Nov 02 11:40:31 * XorA goes to fetch his best baseball bat Nov 02 11:40:36 * Crofton|work grumbles about Germans on vacation :) Nov 02 11:40:49 heh Nov 02 11:41:20 Steffen has been doing some wiki work, but I remember an email saying he would be on vacation for a while Nov 02 11:41:27 I need cbrake to wake up :) Nov 02 11:41:38 From time to time I really wonder how much time Germans spent on vacations Nov 02 11:42:30 Crofton|work: I wonder if he is on vacation too... no reply to my mail from friday. Nov 02 11:42:56 * florian wants a donation button! Nov 02 11:43:14 yeah Nov 02 11:43:41 He did some work, but I clearly remember saying he would be away for a month Nov 02 11:45:05 "The Yocto Working Group shall <...> be open for participation only to LF members and to such individuals as may from time to time be invited to participate by Working Group personnel, based upon their skills and position in the community." Nov 02 11:46:13 likewise: ... which makes a merge a little bit unlikely Nov 02 11:48:24 florian: we can start by merging the sources both ways Nov 02 11:48:26 these are things we need to figure out Nov 02 11:49:20 right... the whole scenario is rather complex and not only a technical issue we are used to solve. Nov 02 11:49:25 heh Nov 02 11:49:27 yeah Nov 02 11:49:34 * hrw needs to move to better hosting Nov 02 11:49:39 it is a political problem with technical implications Nov 02 11:49:45 15K emails of oeml is problem for current one Nov 02 11:49:50 * Crofton|work needs to figure out u-boot scripting Nov 02 11:49:56 Crofton|work: or the other way around Nov 02 11:51:32 * florian fights v4l driver framework changes Nov 02 11:52:36 * hrw fights with ^^@#$@# ubuntu ways of installing distro on pandaboard Nov 02 11:52:54 the "everyone runs pandaboard from SD" stigmat Nov 02 11:53:04 * hrw has 320GB sata disk for it Nov 02 11:54:42 hrw: nice Nov 02 11:54:59 XorA: side effect of adding 60GB SSD to desktop Nov 02 11:55:19 decided that 4 sata disks in desktop is one too much Nov 02 12:12:00 likewise: thx for notes Nov 02 12:20:10 thx for the notes Nov 02 12:20:32 I must admit I am a bit surprised 'bout the merge discussion. Nov 02 12:20:56 well, the yocto project announcement was a bit of a surprise to many Nov 02 12:21:29 good idea to try to merge sources, but organization wise, I simply don't see the future even in the distance. Nov 02 12:21:30 yup Nov 02 12:21:59 LF run project does not seem very community-like IMHO. Nov 02 12:22:32 moreover if yocto is just the answer of intel to arm/linaro this is not clear how long the effort will last Nov 02 12:22:46 could hope they will see the light sometime in the future, but for now, I would not like to put my work into a LF project Nov 02 12:22:48 * hrw wears Linaro hat Nov 02 12:23:12 I do not see Yocto as answer to Linaro - different goals of projects Nov 02 12:23:28 Linaro is not distribution or build system Nov 02 12:24:48 hrw: ok I don't know enough both to speak but from a very far view, both seem to have similar goals : bring a sort of standard to embedded systems Nov 02 12:25:13 the "standard" linaro seems to want to bring is canonical Nov 02 12:25:43 linaro is all about arm, right? Nov 02 12:25:58 afaik yes Nov 02 12:26:05 ericben: at Linaro we work on getting toolchain faster on Cortex A8/9/15 only (and no harm to older) + work on simplifing arm kernel Nov 02 12:26:05 esbenh: yes that's the major difference Nov 02 12:26:19 yocto aims wider coverage Nov 02 12:26:19 yocto is all about embedded, which most definitely includes Intel CPU's ;-) Nov 02 12:26:22 at least for now Nov 02 12:26:26 there are also other groups which works on powermanagement + tools to measure it Nov 02 12:26:36 and other groups Nov 02 12:26:49 I lost track a bit as linaro is no more 20 people Nov 02 12:26:53 hrw: yes I've seen all that. Linaro has technical groups for several aspects which bring improvement to all the community Nov 02 12:27:02 that's a major difference you're right Nov 02 12:27:15 ericben: OE gcc-4.5 already uses Linaro work Nov 02 12:27:18 yes Nov 02 12:27:21 ~hail khem for it Nov 02 12:27:22 * ibot bows down to khem for it and chants, "I'M NOT WORTHY!!" Nov 02 12:27:38 * hrw has 666 unread mails now Nov 02 12:27:43 Recently I have wonder more and more about what the role of OE will be in the future. Not that I am a meego fan, or anything close to that, but package based embedded Linux systems will be a tough area to compete in. Nov 02 12:28:14 esbenh: meego is mostly marketing + cash. they do not even got to "lets upgrade from 1.0 to 1.1" phase Nov 02 12:28:30 esbenh: meego now is "reflash, not upgrade" 'distro' Nov 02 12:28:37 source based build systems for "deeply embedded systems" still misses the grand unified / common platform. Nov 02 12:29:12 * hrw slowly moves from really embedded systems to more desktoplike ones Nov 02 12:29:23 maybe that's because these systems are to specific to be able to get unified Nov 02 12:29:29 my latest board just installs gnome now Nov 02 12:29:39 hrw: lol, yes, and that is really hillarious. the strong point about their package based approach should be the upgrade path, not the reflash. Nov 02 12:30:59 ericben: I believe they are to specific to work with unified packages, but I _DO_ believe it is possible to have unified build recipes, but these must embrace the need to customzie to the extreme. Nov 02 12:31:33 esbenh: yes, take recipes from oe, put them in an overlay, tune them and you have your optimized setup ;-) Nov 02 12:31:47 yes, that is the current OE way Nov 02 12:31:49 optimized and stable setup Nov 02 12:32:20 03Dmitry Eremin-Solenikov  07master * r5ad749e24d 10openembedded.git/recipes/pimlico/ (tasks-0.13/libkoto-fix-compile.diff tasks_0.13.bb): Nov 02 12:32:20 tasks: fix compilation due to missing config.h include Nov 02 12:32:20 Signed-off-by: Dmitry Eremin-Solenikov Nov 02 12:32:31 03Dmitry Eremin-Solenikov  07master * r1775a986a1 10openembedded.git/recipes/mysql/mysql5-native_5.1.40.bb: Nov 02 12:32:31 mysql5-native: drop obsolete do_rm_work replacement Nov 02 12:32:31 Signed-off-by: Dmitry Eremin-Solenikov Nov 02 12:32:42 but I would like to see much more "handles" (trying hard to avoid saying the taboo U word here) to not have to copy recipes to an overlay to "tune" them. Nov 02 12:32:45 tasks 0.13... does it is 3 years old now? Nov 02 12:32:59 esbenh: U-word? Nov 02 12:33:09 ending in -flags ;-) Nov 02 12:33:24 like in G...oo? Nov 02 12:33:36 :p Nov 02 12:33:37 shhhhh Nov 02 12:33:59 as you can already overide any flag in OE easilly we have infinite userflags Nov 02 12:34:12 as much as I don't think it makes sense for a desktop distro, I even more think it makes sense for a build system for "deeply embedded systems" Nov 02 12:34:19 and you can add DISTRO/MACHINE features at will Nov 02 12:34:36 I know. Nov 02 12:34:52 XorA: the example of bluetooth comes to mind Nov 02 12:34:55 but it conflicts with the need for stable feeds. Nov 02 12:35:10 esbenh: not if they are dealt with correctly by the distro Nov 02 12:35:26 esbenh: this is why we are against USERFLAGS Nov 02 12:35:56 you can have global and per-package useflags, what we actually do is hacking the .configure to our needs, without a core-logic Nov 02 12:35:57 let's say I have a recipe with 50 different tunable "features", 5 of them being of string type and not plain boolean Nov 02 12:36:06 XorA: yes, and I understand and appreciate it. Nov 02 12:36:27 so we have some recipes pulling in BT and other not Nov 02 12:36:31 for a build system that tries to compete in the package/feed based distro game. Nov 02 12:37:29 if/when using OE for building embedded systems not using feeds, it really is a shame to not have a platform that is based on more tunable features. Nov 02 12:38:10 what's worst is that in some case the optional deps are found because are staged for some reason and the package is compiled with those. You won't notice-it. Nov 02 12:38:43 yes, that is another key issue for me. Nov 02 12:39:13 and why I will go really far to avoid a common staging directory Nov 02 12:39:20 I dont think there is anything wrong with a recipe that does EXTRA_OECONF = "${FEATUREONE} ${FEATURETWO}" as long as defaults for those are in the recipe, the openldap does exactly this Nov 02 12:39:36 each recipe should completely decide what goes into it's staging dir and what does not. Nov 02 12:42:25 could probably do with extending bitbake syntax to allow ${FEATURE}?-yes:-no type syntax Nov 02 12:42:29 XorA: finally that check is done in each Gentoo recipes too Nov 02 12:44:17 here are the two bits we are lacking atm (IMHO) Nov 02 12:44:17 http://www.gentoo.org/proj/en/devrel/handbook/handbook.xml?part=2&chap=3 Nov 02 12:45:23 ant_work: the IUSE var? Nov 02 12:45:42 (useflhgs and specific ver dependency) Nov 02 12:45:58 IUSE="" would be EXTRA_OECONF = "" Nov 02 12:46:21 well, you see the point Nov 02 12:46:33 no, IUSE is a generic thing, EXTRA_OECONF implies autotools Nov 02 12:46:35 you have to list the potential deps of the recipe Nov 02 12:47:12 version specific dependency would be nice Nov 02 12:47:58 heh Nov 02 12:48:09 did you discuss the OE2 / BB2 agenda topic on OEDEM? Nov 02 12:49:29 esbenh, that all depends on how we handle the yocto project Nov 02 12:49:37 with chainsaws :-) Nov 02 12:49:59 oh sorry, halloween passed :-( Nov 02 12:50:09 there will be one next year ;-) Nov 02 12:50:24 YAY!!!! Nov 02 12:50:34 * XorA thinks it should be halloween all the time Nov 02 12:51:28 the important thing to consider is where we want to be in 2-3 years and further Nov 02 12:52:12 esbenh, world domination Nov 02 12:52:13 personally I want a phone/PDA/router OS that can be as featureful as Fedora, but can be small and tiny Nov 02 12:52:28 continouing the success of OE might depend on setting the right goals, not trying to target everything at once. Nov 02 12:52:40 I want to be able to create an Active Directory server on my 5W arm hardware Nov 02 12:52:57 Crofton|work: why be so modest :-P Nov 02 12:53:42 XorA: if this is the goals, then we surely will have to compete with Android and Meego for some time. Nov 02 12:53:52 want to be able to get rid on x86 except for brute force Nov 02 12:54:02 not that it should stop anyone :-) Nov 02 12:54:37 XorA: you found >11" arm laptop/netbook? Nov 02 12:54:39 meego is dead, Android is the only real competition Nov 02 12:54:50 hrw: I want 10" max Nov 02 12:54:52 * hrw has 10.1" arm netbook on desk Nov 02 12:55:21 XorA: dead on arrival Nov 02 12:55:44 cant really take meego serious when next week nokia will decide to re-invent it for billionth time make all work to date worthless Nov 02 12:55:59 I am in a completely different world. I need a system for building software platforms for industrial systems / deeply embedded systems. Nov 02 12:56:21 esbenh: so you want OE indeed Nov 02 12:56:28 I don't need to compete with Android or Meego, and don't really have to care about package feeds and upgrading. Nov 02 12:56:30 esbenh: arm926 systems? Nov 02 12:56:51 any systems that customers want. Nov 02 12:57:12 currently mustly powerpc, and some arm (920t and A8) Nov 02 13:09:53 XorA, Crofton, florian: do we have our meeting now? Nov 02 13:10:00 mickey|office, yes Nov 02 13:10:04 hang on Nov 02 13:10:33 lets move to a special room Nov 02 13:10:50 esbenh: I'm in the same situation Nov 02 13:23:10 anyone know anything about the qt4-x11-free recipes? Nov 02 13:23:26 i was wondering why all the qt4-x11-free-gles recipes PROVIDE qt4-x11-free, but the qt4-x11-free recipes don't Nov 02 13:24:45 mrj10: recipes don't need to PROVIDE their own names, that is implicit. Nov 02 13:25:26 with hindsight this is possibly a misfeature, and it might have been better to just set PROVIDES = "${PN}" by default in bitbake.conf, but the current behaviour is what it is. Nov 02 13:26:02 gotcha. i am getting an error at the start of my build saying that both qt4-x11-free-gles and qt4-x11-free are being built, and that this usually happens when one recipe PROVIDES something that the other mistakenly does not Nov 02 13:27:33 i'm building angstrom 2008 for beagleboard Nov 02 13:29:22 the fact that both of them are being built causes problems for me because qt4-x11-free installs its libraries to sysroots/.../usr/lib, and sysroots/.../usr/lib is in front of work/qt4-x11-free-gles.../.../lib on the compiler's library search path, so one gets link errors on all the GLES-related stuff Nov 02 13:30:06 yeah, having both built is definitely not desirable Nov 02 13:30:09 it looks like qt4-x11-free-gles replaces qt4-x11-free, so maybe I should set a preferred provider to "gles" somewhere? i'm a bit new Nov 02 13:30:30 the easiest way to debug this is probably to mask out the one you don't want (qt4-x11-free), run the build again, and look to see if you get an unresolved dependency error. Nov 02 13:30:46 how do i mask it? Nov 02 13:30:56 also, yes, if you don't already have PREFERRED_PROVIDER_qt4-x11-free set, set it to qt4-x11-free-gles Nov 02 13:31:01 mrj10: see BBMASK in the manual Nov 02 13:31:05 ok, thanks Nov 02 13:34:04 yeah, it's strange that no machines or distros set that preferred provider, since no recipes explicitly call out -gles in their DEPENDS and it seems like you'd always want the -gles version if your chip runs opengl es Nov 02 13:36:54 03Dmitry Eremin-Solenikov  07master * r5efecae8b7 10openembedded.git/classes/base.bbclass: Nov 02 13:36:54 base.bbclass: fix build error if building with bitbake 1.8.x Nov 02 13:36:54 Signed-off-by: Dmitry Eremin-Solenikov Nov 02 13:42:54 so should the following work? Nov 02 13:42:54 bitbake -b sources/openembedded/recipes/qt4/qt4-x11-free_4.6.3.bb -c distclean; export BBMASK="/qt4-x11-free_4.6.3.bb"; bitbake qt4-x11-free; Nov 02 13:43:18 i'm not sure what command to use to get the installed libs from qt4-x11-free_4.6.3.bb out of sysroots Nov 02 14:08:21 Hi Nov 02 14:11:40 I would like to make some patches on oe tree..but i don't know how i'll do that,what directory i run diff,what i do...how i do... Have any tutorial on wiki ? Nov 02 14:12:02 yeah, there is, let me find it Nov 02 14:12:19 ok thank you Nov 02 14:12:25 google "how to submit a patch to openembedded" Nov 02 14:12:26 i found this but is in blank Nov 02 14:12:27 http://wiki.openembedded.org/index.php/OePatch Nov 02 14:12:29 ok Nov 02 14:12:40 cool Nov 02 14:12:41 thank you Nov 02 14:12:44 no prob Nov 02 14:27:18 Question: What's the best way to tune a recipe? I need to alter the default mysql config file (my.cnf). Should I copy/paste the entire mysql recipe dir into my overlay? Nov 02 14:28:11 i've found that i need to cp -r the whole directory to get it to work right for recipes in overlays Nov 02 14:28:38 ideally i'd prefer to just have the files i changed, but it doesn't seem like it works. maybe my tree just wasn't set up right when i tried that though Nov 02 14:29:17 ok. I didn't have luck in the past either. Temporarily, I've been working on a custom local git branch right in OE. Nov 02 14:29:17 mrj10: I'm making the tutorial,but i don't want submit a new recipe like that command "git add recipes/nodejs/" but i want edit a file(existing) and subimit,what is the command? Nov 02 14:30:31 mrj10: any idea? Nov 02 14:30:50 git add just adds a file to the list of files that will be committed next time you do a "git commit" Nov 02 14:31:01 hum... Nov 02 14:31:04 it's not just for adding new files to the repository like "svn add" Nov 02 14:31:11 ah Nov 02 14:31:12 so Nov 02 14:31:16 for example Nov 02 14:31:25 git add recipes/fluxbox/fluxbox_1.1.1.bb Nov 02 14:31:36 did you make a new branch? Nov 02 14:31:40 no Nov 02 14:31:44 i mean Nov 02 14:31:55 i do the git add after or before the edit? Nov 02 14:32:01 after Nov 02 14:32:24 ah...ok thank you again :) Nov 02 14:32:26 what i would do is make a new temporary branch just for your change, make the edit, git add the file you edited, git commit, then git format-patch Nov 02 14:32:49 no problem Nov 02 14:32:54 just learned this myself a few days ago Nov 02 14:33:20 you say to make a new temporary branch,local isn't? Nov 02 14:33:32 make sure that your patch has the "Signed Off By" line at the bottom, it's policy here and they'll make you resubmit your patch if it doesn't have it Nov 02 14:34:06 the way my git repo is set up anyway, i have local branches that track master and org.openembedded.dev or whatever it's called Nov 02 14:34:18 i made a third branch off of master called mychange Nov 02 14:34:25 switched to mychange Nov 02 14:34:35 the command to use "Signed Off By" is : git format-patch -1 -s Nov 02 14:34:36 made the edit, git add, git commit, git format-patch, git send-email Nov 02 14:34:44 yep Nov 02 14:34:47 ok Nov 02 14:34:52 i'll do another branch Nov 02 14:34:59 i think that'll be more easy Nov 02 14:35:19 it just makes it easier to switch back to master and be able to fast-forward your tree to get changes made on the remote master Nov 02 14:35:44 ah ok Nov 02 14:35:54 i'll try to remember the command to create a new branch Nov 02 14:39:21 try git checkout -b mybranch Nov 02 14:40:04 checkout switches between branches, and -b implicitly creates the branch if it doesn't exist already Nov 02 14:40:12 ah ok Nov 02 14:40:13 thanks Nov 02 14:40:38 np Nov 02 14:41:35 I'm mainteiner of distro metano.conf Nov 02 14:41:57 That distro only get on OE tree yesterday Nov 02 14:42:33 cool, congrats Nov 02 14:42:38 :) Nov 02 14:43:00 i need make patches before them remove,because it is at this time only a minimal image with some packages Nov 02 14:45:51 help...do you know how to remove blank lines on vi ? Nov 02 14:45:57 (im newbie on that) Nov 02 14:46:36 dd Nov 02 14:46:51 Oh thank you Nov 02 14:46:52 :) Nov 02 14:52:18 or dto delete somenumber+1 lines Nov 02 14:56:55 i make a issue here Nov 02 14:57:04 how to change the file that contents the changes Nov 02 14:57:56 can you explain another way? Nov 02 14:58:03 what is the name of the file you want to change Nov 02 14:58:08 see: Nov 02 14:58:22 i did git commit -s Nov 02 14:58:30 and i edited a file with nano Nov 02 14:58:38 but...how i re-edit that file? Nov 02 14:58:48 how do you change your commit message? Nov 02 14:58:57 is that what you're asking? Nov 02 14:59:00 yes Nov 02 14:59:04 sorry my bad english Nov 02 14:59:24 no problem, your english is fine, better than many native speakers :P Nov 02 14:59:29 just trying to understand Nov 02 14:59:56 so i'm not a git expert yet, but my understanding is that you need to do a git rebase Nov 02 15:00:15 ah, maybe not: http://stackoverflow.com/questions/179123/how-do-i-edit-an-incorrect-commit-message-in-git Nov 02 15:00:26 i'll see that Nov 02 15:00:42 git commit -s --amend Nov 02 15:01:03 if that's the last commit : git commit --amend Nov 02 15:01:11 yes Nov 02 15:01:12 worked Nov 02 15:01:16 thanks ! Nov 02 15:01:17 if not : git rebase --interactive ID of the commit before the one you want to change Nov 02 15:01:33 then change the 1st word of the line of the commit you want to edit to edit Nov 02 15:01:37 i had place the signed-off in first line Nov 02 15:01:37 then commit --amend Nov 02 15:01:41 the rebase --continue Nov 02 15:01:52 then, not the Nov 02 15:02:57 hi there Nov 02 15:03:01 hi Nov 02 15:03:04 hi Nov 02 15:03:11 i'm having some problems with my oe rig Nov 02 15:03:36 i havent used it in a while, and in the meantime i upgraded from kubuntu 10.04 to 10.10 Nov 02 15:03:56 i also updated bitbake to the latest version and cloned a fresh copy of oe data Nov 02 15:04:17 and now i get this;EOL while scanning string literal (FILESPATH, line 1) while parsing Nov 02 15:04:32 for every .bb file in openmebedded/recipes Nov 02 15:05:52 here's a full example of one error: Nov 02 15:05:54 ERROR: EOL while scanning string literal (FILESPATH, line 1) while parsing /home/dromede/zaurus/oe/openembedded/recipes/avahi/avahi_0.6.23.bb Nov 02 15:05:54 NOTE: :EOL while scanning string literal (FILESPATH, line 1) while evaluating: Nov 02 15:05:54 ${@':'.join(uniq(os.path.normpath(os.path.join(fp, p, o)) for fp in '/home/dromede/zaurus/oe/openembedded/recipes/avahi'.split(':') for p in 'avahi-0.6.23-r1:avahi-0.6.23:avahi:avahi-0.6.23:avahi:files:.'.split(':') for o in reversed([''] + filter(None, 'local:akita:${MACHINE_CLASS}:zos:linux-gnueabi:arm:build-linux:fail-fast:pn-avahi'.split(':')))))} Nov 02 15:07:56 dromede: have you looked at the .bb files to see if you can see anything wrong with the string literal in question? Nov 02 15:08:21 i cloned a fresh copy of oe data twice Nov 02 15:08:25 any weird trailing backslashes or anything like that? Nov 02 15:08:35 i get a error when i do git-email Nov 02 15:08:35 http://angelox_123.pastebin.com/T5fLMhAd Nov 02 15:09:14 angelox_123: are you running linux on your host machine? Nov 02 15:09:37 yes Nov 02 15:09:39 i am running ubuntu 10.04 and had to apt-get install git-email Nov 02 15:09:41 arch linux Nov 02 15:09:51 maybe there is a similar package for arch? Nov 02 15:10:02 hopefully that satisfies the perl ssl module dependence Nov 02 15:10:04 every error refers the same string, (FILESPATH, line 1) Nov 02 15:10:30 and no, there are no weird characters in bb files Nov 02 15:10:33 i think that git-mail is installed Nov 02 15:10:38 see the begin: Nov 02 15:10:39 http://angelox_123.pastebin.com/SQsEHtqb Nov 02 15:11:09 dromede: any weird characters in your BB-related environment variables? Nov 02 15:11:30 i.e. stuff in ~/.oe/environment Nov 02 15:11:43 maybe a path with a space in it (that has to be escaped with a backslash)? Nov 02 15:13:48 here's my local.conf Nov 02 15:13:50 http://pastie.org/1266816 Nov 02 15:14:40 dromede: i think the 1 line are wrong Nov 02 15:15:09 BBFILES? Nov 02 15:15:14 yes Nov 02 15:15:24 you have two recipes directory? Nov 02 15:15:29 yes Nov 02 15:15:34 strange Nov 02 15:15:46 that's ok Nov 02 15:16:11 i haven't made any changes to my local.conf before this started to happen Nov 02 15:17:18 now i edited my BBFILES to point to a single directory, no effect Nov 02 15:17:33 yeah BBFILES should have spaces Nov 02 15:17:52 environment (as in ones set in your shell) variables used by BB should not Nov 02 15:18:00 do you have a ~/.oe/environment script? Nov 02 15:18:26 or how do your BBPATH, OE_BASE, et. variables get set? Nov 02 15:18:29 i have here ~/.oe/environment for my distro,i can give to you and you change it Nov 02 15:18:38 oh, sorry, i meant dromede Nov 02 15:18:56 i set my BBPATH and PATH in konsole Nov 02 15:18:58 me too Nov 02 15:19:19 dromede: i can give to you my ~/.oe/environment to you test Nov 02 15:19:30 ok, let's try that Nov 02 15:19:30 you'll only need change the machine and distro Nov 02 15:20:06 http://angelox_123.pastebin.com/KyG7nTfs Nov 02 15:20:27 change the things Nov 02 15:21:18 angelox: your smtp server requires SSL right? Nov 02 15:21:26 if so, you need to install the SSL perl module Nov 02 15:21:28 cpan Net::SMTP::SSL Nov 02 15:21:54 mrj10: i had how tutorial,with gmail account Nov 02 15:21:56 it's weird that your git-email package doesn't include all the perl modules for all supported authentication modes Nov 02 15:22:02 i'd have to change pretty much every var in your example angelox Nov 02 15:22:36 but that'd help very Nov 02 15:22:48 and i don't think my local.conf or my env vars are wrong, they used to work before i switched to kubuntu 10.10 just fine Nov 02 15:23:12 dromede: i use here(works fine in any distro) angstrom-setup-scripts Nov 02 15:23:17 and i change the things Nov 02 15:23:25 i'm gonna try with a new, separate oe folder Nov 02 15:24:18 yeah, try that Nov 02 15:24:25 my environment script is here: http://pastebin.com/FSLxPwgT Nov 02 15:25:04 if you want to try modifying that Nov 02 15:25:14 but otherwise i dont think i'll be of much help, sorry Nov 02 15:25:59 thanks anyway Nov 02 15:27:22 arch is hard to me Nov 02 15:27:32 and i don't have any CD to burn fedora Nov 02 15:27:42 can you boot from usb? Nov 02 15:27:43 so..i'll try send patchs manually Nov 02 15:27:51 i don't have usb too :( Nov 02 15:27:57 bad bad bad Nov 02 15:28:06 so cpan didn't work for you? Nov 02 15:28:11 Greetings! It this still valid, or is there another procedure to change a kernel configuration for a certain machine/image? http://groups.google.com/group/beagleboard/browse_thread/thread/f9edc50ba03c94a6/e71e0f6b82017e4c?pli=1 Nov 02 15:29:11 (specifically linux kernel configuration for beagleboard, I want to the change the default beagleboard linux configuration rather than creating a new one from scratch) Nov 02 15:29:16 does your recipe directory have a config file in it? called something like defconfig? Nov 02 15:30:18 i have changed my config for beagleboard, but only for linux-omap-2.6.36rc, and that one had a defconfig file in it, so once i figured out how to generate the defconfig file i just put it there and it worked Nov 02 15:32:07 Ah ok, I think the problem then, was that I modified the .config in the linux directory but not the defconfig Nov 02 15:32:47 azariah4:you can always make a separate dir for your customised recipes Nov 02 15:33:01 you can use bitbake collections for that Nov 02 15:33:23 here's a local.conf example:http://docs.openembedded.org/usermanual/usermanual.html#special_bitbake_collections Nov 02 15:34:04 dromede: thanks! Nov 02 15:34:37 so you make another dir, include the dir in your BBFILES var Nov 02 15:34:59 that dir will always be searched first by bitbake when it parses recipes Nov 02 15:35:25 you then copy your linux recipe and defconfig into your custom dir and make changes Nov 02 15:35:58 How i can send manually patches ? Nov 02 15:36:09 people also call them overlays, in case that helps you google Nov 02 15:36:18 dromede: nice, will make things easier Nov 02 15:36:44 just send the contents of the patch to openembedded-devel@lists.openembedded.org Nov 02 15:36:56 with a subject "[PATCH] " Nov 02 15:37:05 <mrj10> you have to sign up for that list first Nov 02 15:37:16 <mrj10> takes a few hours to get approved usually (did for me, anyway) Nov 02 15:37:43 <mrj10> well, you need the commit message and the signed-off line too, above the patch, but basically whatever git send-email was going to send, just paste that into an email Nov 02 15:38:06 <angelox_123> ok Nov 02 15:38:11 <angelox_123> do you have any example to me? Nov 02 15:38:21 <mrj10> let me see Nov 02 15:38:47 <angelox_123> ok Nov 02 15:41:23 <mrj10> http://pastebin.com/ZsQ1YrVU Nov 02 15:41:42 <mrj10> the subject was "[PATCH] linux-omap-2.6.36rc -- Consistent MAC address for BeagleBoard xM" Nov 02 15:42:24 <angelox_123> thanks Nov 02 16:03:52 <ericben> Tartarus: for gdb autotools doesn't work directly Nov 02 16:04:02 <ericben> Tartarus: configure.ac:32: error: Please use exactly Autoconf 2.64 instead of 2.65. Nov 02 16:04:27 <ericben> Tartarus: hi btw :-) Nov 02 16:10:42 <ericben> is "sed -i 's/2.64/2.65/' config/override.m4" acceptable ? Nov 02 16:22:29 <mrc3> hello! regarding licenses, i wanted to point out how portage is handling that nowadays Nov 02 16:26:00 <mrc3> it uses tokens separated by spaces and some logical modifiers. for example: Nov 02 16:26:21 <mrc3> tagsoup-1.2.ebuild:LICENSE="Apache-2.0" Nov 02 16:26:29 <mrc3> tagsoup-1.1.ebuild:LICENSE="|| ( AFL-3.0 GPL-2 )" Nov 02 16:27:16 <mrc3> moreover, some special licenses agreed to can be defined in make.conf: Nov 02 16:27:20 <mrc3> ACCEPT_LICENSE = "Broadcom dlj-1.1 skype-eula googleearth AdobeFlash-10.1 cadsoft" Nov 02 16:34:17 <tharvey> what is needed in a recipe for a simple qt4 app to build? looked at a couple of recipes and figured it was simply 'inherit qmake' but the moc isn't getting run Nov 02 16:38:45 <ericben> tharvey: do you remember why http://git.openembedded.org/cgit.cgi/openembedded/commit/?id=90cdd0e3813a274a19a2625f664542ef7f93e8f7 was needed ? Nov 02 16:39:49 <tharvey> ericben, was failing do_configure with a note about autoconf version conflict - but I believe that patch did not fix the issue and there was a followup Nov 02 16:39:52 <tharvey> one sec, I'll find it Nov 02 16:40:57 <ericben> yes I've seend the message Nov 02 16:41:57 <tharvey> http://git.openembedded.org/cgit.cgi/openembedded/commit/?id=d823e741892288283a5aaffe91693fefd8b973ed Nov 02 16:42:28 <ericben> yes but why is it needed to havd 2 x oe_runconf ? Nov 02 16:43:05 <tharvey> I never had a good understanding of that myself - I'm not that up with autotools Nov 02 16:43:20 <ericben> I don't see what will change between both Nov 02 16:43:44 <tharvey> I think the point was to run oe_runconf twice Nov 02 16:43:55 <ericben> yes but why ? Nov 02 16:44:07 <tharvey> khem, can you comment? Nov 02 16:49:51 <ericben> tharvey: for you qt4 app : here I'm using inherit qt4e and that works Nov 02 16:53:10 <tharvey> thx I'll give that a try - there are no comments in qmake.bbclass and I see a qmake2.bbclass - its almost as if there is a sep class for diff vers of qt Nov 02 16:54:05 <ericben> tharvey: if using qt4 for X&& you may ant inherit qt4x11 instead of qt4e Nov 02 17:04:31 <tharvey> ericben, thx that worked - we need more comments in the bbclass's Nov 02 17:05:04 <tharvey> ericben, I'm kicking off a build of meta-toolchain with the 'extra' oe_runconf removed, I don't see why its needed. I'm guessing the failure I ran into was fixed somewhere else? Nov 02 17:06:54 <tharvey> I see there were patches to those recipes the same day I was having issues Nov 02 17:28:01 <tharvey> anyone familiar with the gstreamer recipes? I keep getting things like: NOTE: consider defining a PREFERRED_PROVIDER entry to match gst-plugin-tcp Nov 02 17:28:01 <tharvey> NOTE: multiple providers are available for runtime gst-plugin-avi (gst-plugins-bad, gst-plugins-good, gst-plugins-base, gst-plugins-ugly) yet I don't see how any of those recipes even specify what they provide - only one of those provides gst-plugin-tcp Nov 02 17:28:19 <tharvey> er gst-plugin-avi in this case Nov 02 17:59:37 <TAsn> Hey, I'm trying to compile shr (using oe) for the smartq v7, I changed the machine type in the config, and for some reason, although I set PREFERRED_VERSION_kernel = "2.6.36" it still tries to build kernel 2.6.28, where does it take this value from? thanks. Nov 02 18:04:52 <TAsn> ok, it seems that DEFAULT_PREFERENCE Nov 02 18:04:55 <TAsn> was the issue. Nov 02 18:05:01 <TAsn> thanks. Nov 02 18:07:14 <tharvey> anyone know why ifupdown_0.6.10.bb and ifupdown_ubuntu_0.6.8.bb don't inherit alternatives and why ifupdown_0.6.10.bb doesn't specifcy 'PROVIDES = ifupdown' so that you can choose between one and the other with a PREFERRED_PROVIDER? Nov 02 18:18:12 <tharvey> hmmm perhaps I don't understand the problem. I want to install ifupdown-ubuntu because ifupdown uses /etc/network/run for ifstate which is non-volatile and ubuntu verison uses /var/run/network - I'm using task-proper-tools which uses ifupdown - how can I make ifupdown/ifupdown-ubuntu alternate providers? Nov 02 18:18:48 <tharvey> I thought as long as both had PROVIDES = "ifupdown" I could then use PREFERRED_PROVIDER_ifupdown = "ifupdown-ubuntu" Nov 02 18:24:22 <kergoth_> tharvey: ifupdown always PROVIDES ifupdown. see the default PROVIDES in bitbake.conf. ${PN}. ${PN}-${PV}, and ${PN}-${PV}-${PR} are provided by default for any recipe Nov 02 18:24:37 <kergoth_> perhaps you meant that ifupdown_ubuntu should PROVIDES ifupdown, that it should, i expect :) Nov 02 18:25:34 <tharvey> ah ok, didn't catch the default. in that case both have PROVIDES = "ifupdown" - so how come PREFERRED_PROVIDER_ifupdown = "ifupdown-ubuntu" is not pulling in that instead I wonder? Nov 02 18:25:47 <tharvey> | * check_data_file_clashes: Package ifupdown-ubuntu wants to install file /home/tharvey/oe/commonOS-sync/tmp/rootfs/firstlook-ocu-image/etc/init.d/ifup Nov 02 18:26:07 <kergoth_> hmm, not sure Nov 02 18:26:19 <kergoth_> perhaps you hve a recipe pulling in the ubuntu one directly Nov 02 18:26:37 <tharvey> the ubuntu one is the one I want not the other Nov 02 18:26:49 <kergoth_> ah, right. hmm Nov 02 18:26:53 <tharvey> unless there is some std way of making the other use /var/run/network? Nov 02 18:27:16 <tharvey> seems strange to use /etc/run/network for ifstate - makes 2nd boot not configure network interfaces Nov 02 18:27:40 <kergoth_> that does seem strange Nov 02 18:28:06 <tharvey> or perhaps I need to clean something I'm not - since task-proper-tools is what includes ifupdown its that I suspec I should clean right? Nov 02 18:29:22 <tharvey> perhaps ifupdown-ubuntu recipe naming is the issue? '-' ? Nov 02 18:30:50 <tharvey> and I'm still unsure why both of those recipes don't inherit alternatives and configure ifup/ifdown as alternatives for busybox provided? Nov 02 18:53:20 <TAsn> hey guys, can having python pointing to python3 (instead of python2) on my host machine cause issues? Nov 02 18:53:26 <TAsn> I search and replaced Nov 02 18:53:32 <TAsn> bitbake Nov 02 18:53:37 <TAsn> to specifically use python2 Nov 02 18:53:42 <TAsn> but is there something else I'm missing? Nov 02 18:54:15 <mrj10> bitbake is itself a python program Nov 02 18:54:33 <TAsn> as I said Nov 02 18:54:36 <mrj10> so if it is not written to work on python3, you may be busted out of the gate Nov 02 18:54:39 <TAsn> <TAsn> I search and replaced Nov 02 18:54:40 <TAsn> <TAsn> bitbake Nov 02 18:54:40 <TAsn> <TAsn> to specifically use python2 Nov 02 18:54:43 <LexsZero> your host distro don't provide sane python versions switching method? Nov 02 18:54:59 <TAsn> LexsZero, no idea actually, I'm using arch Nov 02 18:55:07 <TAsn> I wonder why the hell they made the switch so prematurely Nov 02 18:55:10 <mrj10> TAsn: doh, my bad. second time i've misread a critical character in 5 minutes Nov 02 18:55:18 <TAsn> mrj10, :P Nov 02 18:55:47 <mrj10> have you tried a build any it gave you problems? or asking preemptively? Nov 02 18:55:59 <TAsn> LexsZero, asked there in their chan, waiting for a response Nov 02 18:56:04 <TAsn> anyhow, is there something else Nov 02 18:56:10 <TAsn> ? Nov 02 18:56:21 <mrj10> i suppose recipes may call python themselves Nov 02 18:56:24 <LexsZero> for example, in gentoo you can do smth like eselect python set blahblah Nov 02 18:56:33 <mrj10> haven't seen it but it could happen. seems like an easy grep job Nov 02 18:56:34 <TAsn> LexsZero, they said I should just change the symlink Nov 02 18:56:39 <TAsn> do you think that makes sense? Nov 02 18:56:45 <LexsZero> yes Nov 02 18:57:05 <mrmoku> khem: is armv6j supposed to work gcc/eglibc-wise? Nov 02 18:57:45 <mrj10> yeah, i think the arch guys might just be trying to bully people into writing python3 code going forward, but it shouldn't break anything to change your symlink back Nov 02 18:58:30 <TAsn> hehe they are fighting hard Nov 02 18:58:35 <TAsn> they are saying I should Nov 02 18:58:46 <TAsn> just search and replace all of oe Nov 02 18:58:56 <TAsn> lol Nov 02 18:58:59 <TAsn> ok, will change the symlink Nov 02 18:59:44 <TAsn> ok, retrying Nov 02 19:08:23 <TAsn> nope, changing python did no good Nov 02 19:08:26 <TAsn> still fails Nov 02 19:37:23 <Crofton|work> I'm building a recpe that unpacks a tarball and install's data files manually Nov 02 19:37:30 <Crofton|work> is tehre a good example recipe for this? Nov 02 19:44:23 <Tartarus> What do you mean? Nov 02 19:44:27 <Tartarus> Just an empty do_compile Nov 02 19:44:30 <Tartarus> and a manual do_install? Nov 02 19:50:04 <Crofton|work> yeah Nov 02 19:50:27 <Crofton|work> just want to make sure I copy the right do_install :) Nov 02 19:50:51 <Tartarus> best to just write one i think Nov 02 19:51:36 <woglinde> jo Nov 02 19:51:40 <Crofton> ok Nov 02 19:51:50 <woglinde> ping khem Nov 02 19:54:12 <Jay7> hm.. Nov 02 19:54:32 <Jay7> tinderbox thinking that my builds are still incomplete Nov 02 19:54:37 <Jay7> http://tinderbox.openembedded.net/builders/Jay7-tb/ Nov 02 19:55:20 <Jay7> top 3 builds was finished successfull about 4 hours ago Nov 02 19:55:40 <Jay7> btw, bitbake was returned 1 for all 3 builds Nov 02 19:55:50 <Jay7> but according to log all was ok Nov 02 19:59:12 <Jay7> sh*t, bitbake remembers previous tmp location.. Nov 02 19:59:40 <Jay7> so workflow 'do clean build then move it to do incremental build' isn't working... Nov 02 20:15:54 <Jay7> hm.. buildbot is interesting thing Nov 02 20:19:16 <Jay7> we can setup own buildbot (on tinderbox) and use current build-testers as build-slaves Nov 02 20:19:23 <Jay7> or we can use Yocto's Nov 02 20:19:42 <Jay7> as zenlinuxPDX pointed in his letter Nov 02 20:20:49 <zenlinuxPDX> btw, a (slightly generalized) buildbot config we're using can be found here: http://git.pokylinux.org/cgit/cgit.cgi/poky-autobuilder/ Nov 02 20:21:14 <zenlinuxPDX> that repo includes various build and support scripts we're using, too Nov 02 20:21:28 <Jay7> hm.. but buildbot have no integration with bugzilla seems Nov 02 20:21:35 <Jay7> our tinderbox have Nov 02 20:21:58 <zenlinuxPDX> Jay7, true - tinderbox is a mozilla project, I believe Nov 02 20:22:15 <zenlinuxPDX> but interestingly, the lead developer of buildbot just got hired by mozilla to work on the project full-time Nov 02 20:22:57 <Jay7> well.. let's other people to say something :) Nov 02 20:23:08 <woglinde_> something Nov 02 20:23:24 <Jay7> and woglinde_ wins this contest Nov 02 20:34:24 <woglinde_> jo jama Nov 02 20:35:27 <JaMa> jo woglinde_ Nov 02 20:41:54 <woglinde_> jo effem Nov 02 20:42:16 <eFfeM> hi woglinde_ , hi all Nov 02 21:04:04 <anarsoul> hi Nov 02 21:04:46 <anarsoul> guys, what about taking this patch: http://bugs.gentoo.org/attachment.cgi?id=248616 into eglibc recipe stuff Nov 02 21:05:10 <anarsoul> 'cos it fails to build with make-3.82 without this (or similar) patch Nov 02 21:05:39 <anarsoul> ? Nov 02 21:05:48 <anarsoul> khem: I know you're responsible for toolchain stuff Nov 02 21:10:10 <tharvey> ericben, tested the removal of the second oe_runconf and had no build issues of meta-toolchain - I think my issue was likely caused by some other breakage that was already fixed Nov 02 21:11:57 <CIA-1> 03Eric Bénard <eric@eukrea.com> 07org.openembedded.dev * r04386651d9 10openembedded.git/recipes/gdb/gdb-cross-sdk.inc: (log message trimmed) Nov 02 21:11:57 <CIA-1> gdb-cross-sdk: reconfigure before running oe_runmake configure-host Nov 02 21:11:57 <CIA-1> so we have : configure then make configure-host which generates the Nov 02 21:11:57 <CIA-1> Makefile that we modify using sed to statically link the libs Nov 02 21:11:57 <CIA-1> Without this patch, we don't statically link libncurses, libtinfo Nov 02 21:11:57 <CIA-1> and libz which leads to : Nov 02 21:11:58 <CIA-1> /usr/local/angstrom/arm/bin/arm-angstrom-linux-gnueabi-gdb: Nov 02 21:22:29 <cbrake> has anyone used valgrind on a ARM system? Nov 02 21:22:36 <cbrake> specifically a PXA270 class machine Nov 02 21:22:37 <woglinde> not yet Nov 02 21:23:14 <cbrake> 3.6 claims to support ARM/Linux Nov 02 21:23:23 <woglinde> I know Nov 02 21:40:53 <Jay7> cbrake: are you here? Nov 02 21:41:07 <cbrake> Jay7: yes Nov 02 21:41:31 <Jay7> cbrake: what is previous testing-next branch? 'testing' or 'testing-prev'? Nov 02 21:41:48 <cbrake> Jay7: testing Nov 02 21:41:56 <Jay7> 10x Nov 02 21:42:05 <cbrake> Jay7: we only move it forward if things are reasonably stable Nov 02 21:42:09 <cbrake> Jay7: so its a couple weeks old Nov 02 21:42:15 <Jay7> hm.. Nov 02 21:42:24 <Jay7> I'm about incremental build Nov 02 21:42:40 <cbrake> Jay7: you could get last weeks branch from the tested tag Nov 02 21:43:08 <Jay7> seems it will require manual operations Nov 02 21:44:05 <cbrake> Jay7: nod, so you are looking to automate Nov 02 21:44:30 <cbrake> Jay7: I could start creating a testing-last branch or some less confusing name if anyone has ideas Nov 02 21:46:21 <Jay7> cbrake: well.. it required only for first time Nov 02 21:46:29 <Jay7> or may be not required at all.. Nov 02 21:46:54 <cbrake> Jay7: ok, if you want something like that, I can add to the weekly procedure Nov 02 21:47:20 <Jay7> cbrake: ok, I'll rethink my testbuilder again Nov 02 21:47:27 <florian> re Nov 02 22:11:26 <Jay7> - automated tests and result reports on qemu targets/on remote hw board targets Nov 02 22:11:28 <Jay7> I like this Nov 02 22:19:04 <khem> all day was mess Nov 02 22:22:50 <khem> I hope no one was looking for me with a baseball bat in hand today Nov 02 22:41:45 <grg> khem, is that usually a problem? Nov 02 22:42:55 <khem> heh no unless its wielded Nov 02 22:46:08 <dj-death> khem: crappy day too ? :) Nov 02 22:47:18 <khem> there is so much papers you need to keep on getting stamps on to keep yourself legal in this country Nov 02 22:47:59 <dj-death> usa ? Nov 02 23:02:09 <TAsn> hey guys, I'm trying to build shr for smartq v7, taking a look at the build, it builds kernel 2.6.28, it does that because DEFAULT_PREFERENCE_smartqv7 = "1" is set. but I think 2.6.36 already has support in mainline so maybe I can just use it, what's the correct way to choose 2.6.36? thanks... Nov 02 23:03:29 <fidencio> TAsn: are you the same TAsn from edevelop? Nov 02 23:03:45 <TAsn> yep. Nov 02 23:03:55 <TAsn> and from many other chans on freenode :P Nov 02 23:04:03 <TAsn> (j/k) Nov 02 23:04:12 <TAsn> (but I am the same man) Nov 02 23:05:58 <fidencio> ehhehe. First time that I meet a "household" name from efl in this channel Nov 02 23:06:24 <TAsn> nah, I've been coming in and out this chan when I was active in shr Nov 02 23:07:07 <TAsn> anyhow, now I'm trying to run shr on my smartq :P Nov 02 23:07:12 <TAsn> already running efl on their basic ubuntu Nov 02 23:07:19 <TAsn> but /me wants more. Nov 02 23:08:04 <fidencio> TAsn: about your queston, is there a recipe to kernel 2.6.36 with the confs that you need? Nov 02 23:08:53 <TAsn> fidencio, from what I've seen Nov 02 23:08:59 <TAsn> 2.6.36 has no specific configs Nov 02 23:09:02 <TAsn> for any device. Nov 02 23:09:24 <mrj10> there is one for the beagleboard from the linux-omap tree :P Nov 02 23:09:26 <TAsn> I figured I'd have to config it for mine. Nov 02 23:09:33 <TAsn> i.c Nov 02 23:09:42 <TAsn> well, I'll have to stick with the .28 then Nov 02 23:09:47 <TAsn> or create a config Nov 02 23:09:52 <mrj10> you might be able to create your own recipe Nov 02 23:10:35 <fidencio> hmmm. and shr don't work with a specific config just keeping a lot of patches? (freescale works in this way) Nov 02 23:10:59 <fidencio> s/woek/have/ Nov 02 23:11:58 <TAsn> fidencio, no idea, but I'm building to a device different than the regular shr one. Nov 02 23:12:12 <TAsn> shr guys usually build for Nov 02 23:12:13 <TAsn> dream Nov 02 23:12:14 <TAsn> n900 Nov 02 23:12:17 <TAsn> om-gta0* Nov 02 23:12:21 <TAsn> and palmpre Nov 02 23:12:28 <TAsn> I'm buliding for the smartqv8 Nov 02 23:12:30 <TAsn> *7 Nov 02 23:13:37 <TAsn> fidencio, anyhow, seems I'll just have to get myself a better kernel after I first get this basic image working :) Nov 02 23:14:07 <fidencio> :-) Nov 02 23:14:48 <fidencio> TAsn: sigdinner, I'll leave that people with more knowledge can help you. see ya in #edev :-) Nov 02 23:14:52 <TAsn> I wonder if there's someone actively building OE for the smartq Nov 02 23:15:11 <TAsn> fidencio, ha, yeah, see you there :P Nov 02 23:15:46 <mrj10> can you check tinderbox to see if anyone has that as their MACHINE? Nov 02 23:16:04 <likewise> mrj10: yes Nov 02 23:16:20 <TAsn> mrj10, good idea. Nov 02 23:16:21 <likewise> mrj10: sorry, I read that question different, never mind Nov 02 23:17:21 <mrj10> likewise: if colloquial english were less messed up, it wouldn't be impolite to say "you should check tinderbox" :P Nov 02 23:17:47 <TAsn> mrj10, no one. :| Nov 02 23:17:57 <likewise> mrj10: I read it as "is it possible to check tinderbox to see..." Nov 02 23:18:37 <likewise> good nite all (as in proper intl english) Nov 02 23:20:19 <mrj10> bummer :/ Nov 02 23:23:38 <TAsn> yep. Nov 02 23:27:18 <Jay7> TAsn: set DEFAULT_PREFERENCE_smartqv7 = "1" in kernel recipe you wish to build Nov 02 23:27:37 <TAsn> yeah, that's obvious. Nov 02 23:27:40 <TAsn> I did that Nov 02 23:27:45 <TAsn> earlier today Nov 02 23:27:48 <Jay7> better to do it in your local branch Nov 02 23:27:55 <TAsn> but I'm wondering if there's a better way to do it Nov 02 23:28:21 <TAsn> sounds odd I'd need to edit a shared file Nov 02 23:28:27 <TAsn> to choose a kernel version Nov 02 23:28:49 <Jay7> well.. may be there is way to change via PREFFERED_VERSION_blablabla Nov 02 23:28:57 <Jay7> in local.conf Nov 02 23:29:13 <Jay7> but I'm unsure about right way Nov 02 23:29:30 <TAsn> thanks. Nov 02 23:29:34 <TAsn> anyhow, I'm going to sleep. Nov 02 23:29:35 <TAsn> night. Nov 02 23:29:38 <TAsn> tty tomorrow. Nov 02 23:57:07 <ka6sox-work> Crofton, you around? Nov 02 23:59:39 <Crofton> yeah Nov 02 23:59:51 <Crofton> ka6sox-work, ping Nov 03 00:01:22 <ka6sox-work> Sources.openembedded.org is ready for use? Nov 03 00:01:32 <Crofton> er Nov 03 00:01:43 <Crofton> did you ask me to do something? Nov 03 01:09:38 <Jay7> what directory in $TMPDIR can I check to be sure that build at least was started there? Nov 03 01:09:56 <Jay7> directory or file Nov 03 01:11:50 <mrj10> work? Nov 03 01:12:15 <Jay7> I think the same :) Nov 03 01:13:03 <mrj10> most commands of the form "bitbake x" will do a bunch of stuff in work/, then stage it into sysroots/ and or deploy/<your libc variant>/<images|ipk|sources> Nov 03 01:13:44 <Jay7> well.. work should be enough for my need Nov 03 01:14:52 <mrj10> yeah, each package will have its own directory in dir/x/<package> Nov 03 01:15:10 <Jay7> hm.. good idea to place $TMPDIR on separate partition or tmpfs on autobuilder Nov 03 01:15:16 <mrj10> where x depends on what sort of package it is (host, MACHINE-specific, architecture-specific, etc. Nov 03 01:15:25 <Jay7> then it can be cleaned by newfs/umount Nov 03 01:15:37 <mrj10> depends Nov 03 01:15:47 <mrj10> what is $TMPDIR for you? does it encompass the entire build? Nov 03 01:16:10 <mrj10> not a bad idea to be able to wipe it easily, the staging areas can easily become corrupted if you have problems halfway through a build Nov 03 01:16:16 <Jay7> I'm writing autobuilder script :) Nov 03 01:16:34 <mrj10> but you want to move the stuff in deploy/ somewhere safe before wiping, that's the final output :P Nov 03 01:16:48 <Jay7> no, I don't there :) Nov 03 01:17:17 <Jay7> clean builds should be started with empty $TMPDIR Nov 03 01:17:49 <Jay7> it's only for testing purposes Nov 03 01:18:13 <Jay7> mrj10: but.. you give me interesting idea.. Nov 03 01:19:03 <Jay7> results of clean build can be used to automated checking in qemu later.. Nov 03 01:19:08 <mrj10> yeah Nov 03 01:19:26 <Jay7> but anyway.. I'll have incremental builds too with saved tmpdir Nov 03 01:22:25 <kergoth_> hey guys Nov 03 01:22:36 <Jay7> hello kergoth_ Nov 03 01:23:01 <kergoth> Jay7: how's it going? Nov 03 01:23:09 <Jay7> kergoth: sleepy :) Nov 03 01:23:10 * kergoth peruses the OEDEM summary since he wasn't able to make it Nov 03 01:23:32 <Jay7> too late here but I wish start builder before going to bed Nov 03 01:23:40 <kergoth> ah Nov 03 01:24:17 <kergoth> i should really buy a dedicated build machine for personal builds one of these days Nov 03 01:24:25 <Jay7> I like idea of automated testing with qemu Nov 03 01:24:33 <Jay7> we should think about this Nov 03 01:24:48 <kergoth> agreed Nov 03 01:24:55 <kergoth> also chroots. for build stuff, they work well Nov 03 01:25:06 <kergoth> oh, you mean runtime testing, not build testing Nov 03 01:25:07 <Jay7> I'm building in LXC Nov 03 01:25:07 <kergoth> agreed Nov 03 01:25:11 <kergoth> cool Nov 03 01:25:16 * kergoth should play with that at some point, know nothing about it Nov 03 01:25:25 <Jay7> and I have almost ready chroot with debian lenny Nov 03 01:26:18 <kergoth> <3 debootstrap :) Nov 03 01:26:57 <Jay7> about runtime testing.. I think running image in qemu and parsing kernel output on serial should be easy thing Nov 03 01:27:10 <Jay7> may be connect to ssh after Nov 03 01:27:21 <kergoth> agreed Nov 03 01:27:40 <kergoth> could get ltp going Nov 03 01:27:54 <kergoth> also unit tests for the target gcc and glibc, perhaps Nov 03 01:40:09 <Jay7> well.. should work now Nov 03 01:54:47 <Jay7> builder started so I can go to bed Nov 03 01:54:55 <Jay7> good night Nov 03 01:55:08 <kergoth> night Nov 03 02:10:19 <angelox_123> anybody out me uses EZX devices as arch ? Nov 03 02:10:39 <angelox_123> like a1200,a780 Nov 03 02:47:14 <angelox_123> Hi back Nov 03 02:47:28 <mrj10> hey Nov 03 02:47:28 <angelox_123> i have here a patch to /etc/init.d/halt.... Nov 03 02:47:41 <angelox_123> but i don't know what bb file create that Nov 03 02:47:45 <angelox_123> hi mrj10 Nov 03 02:48:03 <angelox_123> that patch allows shutdown when booted by NFS Nov 03 02:48:55 <mrj10> oh cool Nov 03 02:49:04 <angelox_123> do you know what file create that other file? Nov 03 02:51:02 <angelox_123> do anybody knows ? Nov 03 02:51:05 <mrj10> hmm.. are you using util-linux-ng? Nov 03 02:51:16 <angelox_123> i don't know :) Nov 03 02:51:19 <angelox_123> i'll check Nov 03 02:51:35 <mrj10> actually, i think it's initscripts Nov 03 02:51:51 <mrj10> mrj10@mjlap:~/beagleboard/kernel/build-validation-image/angstrom-setup-scripts/sources/openembedded$ grep -r init\.d\/halt * Nov 03 02:51:51 <mrj10> recipes/shr/initscripts-shr_0.0.1.bb: ln -sf ../init.d/halt ${D}${sysconfdir}/rc0.d/S90halt Nov 03 02:51:51 <mrj10> recipes/initscripts/initscripts_1.0.bb: ln -sf ../init.d/halt ${D}${sysconfdir}/rc0.d/S90halt Nov 03 02:51:51 <mrj10> recipes/obsolete/initscripts/initscripts-openmoko_1.0.bb: ln -sf ../init.d/halt ${D}${sysconfdir}/rc0.d/S90halt Nov 03 02:52:03 <angelox_123> yes Nov 03 02:52:05 <angelox_123> i'll check Nov 03 02:52:51 <angelox_123> yes you found it Nov 03 02:52:52 <angelox_123> thanks Nov 03 02:52:54 <angelox_123> :) Nov 03 02:53:03 <angelox_123> that's will be my first OE patch Nov 03 02:54:14 <mrj10> nice, do you know the procedure for submitting a patch? Nov 03 02:54:19 <angelox_123> +/- Nov 03 02:54:20 <angelox_123> :) Nov 03 02:54:27 <mrj10> i have pointed a few people to the tutorial but i forget if you're one of them :P Nov 03 02:54:41 <angelox_123> np Nov 03 02:54:43 <mrj10> http://wiki.openembedded.org/index.php/How_to_submit_a_patch_to_OpenEmbedded Nov 03 02:55:05 <angelox_123> now i'm remembering Nov 03 02:55:09 <angelox_123> you pointed yes Nov 03 02:55:10 <angelox_123> :) Nov 03 02:55:17 <angelox_123> some hours ago Nov 03 02:55:22 <angelox_123> like 6/7 hours Nov 03 02:55:37 <angelox_123> i had the problem with a library Nov 03 02:55:39 <mrj10> oh, right, were you having a problem with a perl SSL module? Nov 03 02:55:44 <angelox_123> yes Nov 03 02:55:49 <mrj10> yeah, did you try installing it via cpan? Nov 03 02:55:53 <angelox_123> no Nov 03 02:56:22 <angelox_123> i never hear about it! :) to you see how i'm newbie Nov 03 02:56:46 <angelox_123> i'm trying it now Nov 03 02:58:31 <mrj10> "cpan Net::SMTP::SSL" as root Nov 03 02:58:55 <angelox_123> trying Nov 03 02:58:56 <angelox_123> thanks Nov 03 02:59:18 <angelox_123> hi grg Nov 03 02:59:26 <grg> hello angelox_123 **** ENDING LOGGING AT Wed Nov 03 02:59:57 2010