**** BEGIN LOGGING AT Fri Jan 18 02:59:59 2013 Jan 18 05:31:24 re Jan 18 07:38:00 hey everyone how do i mount a .img file system image Jan 18 07:38:00 ? Jan 18 07:38:10 i tried -o loop it says specify type Jan 18 07:38:31 and i tried multiple fs types but nothing worked. It kept spewing wrong fs Jan 18 07:44:44 spyzer: .img is generic extension Jan 18 07:44:49 how did you make the fle Jan 18 08:27:17 good morning Jan 18 08:34:06 khem, i donwloaded fedora 17 arm image for pandaboard Jan 18 08:35:40 gah Jan 18 08:35:49 ask fedora dudes Jan 18 08:35:55 you will have better luck Jan 18 08:36:58 morning all Jan 18 08:37:23 silvio: gm Jan 18 08:37:47 gm mckoan Jan 18 08:37:59 khem: okay so there is one thing i definitely need help in. i was using narcissus to make an image for pandaboard. But the tar.gz image doesn't have an MLO or u-boot.bin Jan 18 08:38:25 and if I am trying to compile it, i am getting certain issues. Is there a way for me to get out of this Jan 18 08:38:26 ? Jan 18 09:30:29 Hi Jan 18 09:31:35 I am new to the system, just followed the user manual steps to build helloworld.bb.. but getting error " Can not check RPATH, scanelf (part of pax-utils-native) not found" ... any idea ? Jan 18 09:37:42 hi Jan 18 09:43:16 morning all Jan 18 09:53:56 I am new to oe... I am just trying to the default hellloworld.bb in oe. getting error " Can not check RPATH, scanelf (part of pax-utils-native) not found" ...Please help Jan 18 09:56:56 debi: no clues, try eplaining what you are doing: machine, distro, image... Jan 18 09:59:32 i used the TARGET_ARCH = "arm" ... TARGET_OS = "linux".......DISTRO = "angstrom-2008.1" Jan 18 09:59:40 this is the configuration I am using Jan 18 10:00:08 i issued bitbake -b openembedded/recipes/helloworld/helloworld_1.0.0.bb Jan 18 10:01:46 can you suggest if there is any issue with the configuration that I am using Jan 18 10:03:34 I am trying to build it for Linux for a arm board Jan 18 10:07:13 debi: oe-classic ? which machine? Jan 18 10:11:42 I am trying to create individual package on an ubuntu machine Jan 18 10:12:09 debi: more than likely the use of -b is preventing the appropriate dependencies from being built Jan 18 10:12:31 debi: you should not use -b unless you know that the appropriate dependencies are already built Jan 18 10:13:03 so how should I build helloworld.bb Jan 18 10:13:15 I want an individual package Jan 18 10:13:30 debi: bitbake helloworld Jan 18 10:14:10 I tried this Jan 18 10:14:11 bitbake openembedded/recipes/helloworld/helloworld_1.0.0.bb Jan 18 10:14:24 ERROR: ExpansionError during parsing /home/user/Graphics/OEBASE/openembedded/recipes/wlan-ng/wlan-ng-modules_svn.bb Jan 18 10:14:59 Failure expanding variable SRCPV, expression was ${@bb.fetch.get_srcrev(d)} which triggered exception MalformedUrl: The URL: '#' is invalid and cannot be interpreted ERROR: Command execution failed: Exited with 1 Jan 18 10:15:25 debi: are yui following any document online fo that? Jan 18 10:15:46 would be better to pastebin your local.conf at least Jan 18 10:15:47 I am following the user manual for oe Jan 18 10:16:32 debi: which is your machine? Jan 18 10:16:42 http://pastebin.com/52zkMaku Jan 18 10:17:08 host: ubuntu x86 target_arch = arm Jan 18 10:17:47 debi: you have to set MACHINE Jan 18 10:18:00 which is your target board? Jan 18 10:19:26 I think I should use MACHINE = "qemuarm" Jan 18 10:19:37 target board is an ARM board Jan 18 10:20:15 debi: ok so set MACHINE = "qemuarm" then Jan 18 10:20:28 okj Jan 18 10:21:46 I am following the user manual for oe Jan 18 10:21:58 ERROR: ExpansionError during parsing /home/user/Graphics/OEBASE/openembedded/recipes/wlan-ng/wlan-ng-modules_svn.bb: Failure expanding variable SRCPV, expression was ${@bb.fetch.get_srcrev(d)} which triggered exception MalformedUrl: The URL: '#' is invalid and cannot be interpreted ERROR: Command execution failed: Exited with 1 Jan 18 10:22:11 I am getting the same old Expansionerror Jan 18 10:22:47 debi: please post the link of your "user manual for oe" Jan 18 10:24:08 AFAIK the OE manual is quite out of date. http://www.openembedded.org/wiki/Documentation Jan 18 10:24:22 http://docs.openembedded.org/usermanual/html/index.html Jan 18 10:24:47 oe-classic is stalled and replaced by oe-core and yocto Jan 18 10:24:58 debi: that document is out of date Jan 18 10:25:28 debi: angstrom-2008.1 is out of date Jan 18 10:25:35 sigh..again Jan 18 10:25:41 mckoan : oh ok .... I see ... Can you please suggest where can I get the updated details Jan 18 10:26:03 debi: I'd suggest you to give a try with this http://www.kaeilos.com/download Jan 18 10:27:11 You mean I must get he oe-core form the link .. right ? Jan 18 10:27:17 debi: but consider that I don't use it for a long time and that I am currently upgrading it to oe-core/yocto Jan 18 10:28:16 That is not at all an issue ...rather thanks for your interest :) Jan 18 10:28:20 bluelightning: I think it's now time for the wiki/Main_Page to redirect to wiki/OpenEmbedded-Core Jan 18 10:28:35 ant_work: I did update the main page already... Jan 18 10:28:49 go straight to oe-core ;) Jan 18 10:29:09 it already points to the OE-Core getting started page though Jan 18 10:29:43 it's not evident enough...we have a fresh victim ;) Jan 18 10:30:26 I agree, the links and their description are cryptic IMHO Jan 18 10:31:29 I got the Getting Started page..... Jan 18 10:31:50 this http://www.openembedded.org/wiki/Getting_started is getting start to a new seek Jan 18 10:31:58 there do you mean I should check the yocto poject Jan 18 10:32:15 not providing a real step-by-step doc Jan 18 10:32:26 As mckoan said angstrom is obsolete Jan 18 10:32:37 debi: Yocto is a good starting point Jan 18 10:32:51 mckoan: er, it does, it links to several step-by-step documents Jan 18 10:33:05 debi: no, angstrom-2008.1 is out of date, not Angstrom distro at all Jan 18 10:33:40 debi: you can quick start with Angstrom or Yocto Jan 18 10:33:45 bluelightning: the fact is, we should purge Google Jan 18 10:33:56 ^_^ Jan 18 10:34:27 I think I can start with yocto project.... Jan 18 10:34:49 ant_work: if you add a clear link to oe main page it would be enough Jan 18 10:35:29 ant_work: a link saying that you can't do anything with oe alone, you need a distro Angstrom or Yocto too Jan 18 10:35:53 it depends Jan 18 10:35:53 mckoan: but that's not correct, you can build a functioning linux system using OE-Core + bitbake Jan 18 10:36:09 my idea of 'core' is OE-Core Standalone Setup Jan 18 10:36:31 yes we need a new OE-Core Standalone Setup doc Jan 18 10:36:40 I think there must be some information about oe being upgraded to oe-core and that the manual must be removed as mckoan said its obsolete Jan 18 10:36:57 right -- we need a new OE-Core Standalone Setup doc Jan 18 10:37:04 mckoan: what's wrong with that? Jan 18 10:37:45 the fact that Distro's are hiding you thedetails won't help to understand Jan 18 10:38:09 ant_work: yes it's a misunderstanding Jan 18 10:38:20 mckoan: ah, I see...we could discuss about having bitbake under oe-core... Jan 18 10:38:58 ant_work: I wanted to say that we need a new step-by-step doc Jan 18 10:40:10 mckoan: what's wrong with http://www.openembedded.org/wiki/OE-Core_Standalone_Setup (linked from the new getting started page) ? Jan 18 10:40:19 ant_work: which is basically this http://www.openembedded.org/wiki/OE-Core_Standalone_Setup ;-) Jan 18 10:40:31 :-D Jan 18 10:40:42 I do understand that a lot of people will come directly to the manual from google, we probably do need to do something about that Jan 18 10:40:50 ok, we all know this page but looks like newbies can't find it Jan 18 10:41:01 :-D Jan 18 10:41:40 so debi you caould follow it as minimal starting point and then decide if you need a distro ;-) Jan 18 10:42:11 debi: out of interest, how did you get to the user manual? Jan 18 10:42:46 I came through it on google Jan 18 10:44:08 bluelightning: I'm prety sure kergoth was laughing lou about keeping bitbake under /oe-core but I think i'ts acceptable as Yocto is using it Jan 18 10:44:35 it = this layout Jan 18 11:04:42 ant_work: it shouldn't be too much of an issue, the bitbake/ directory is in .gitignore in the OE-Core repository Jan 18 11:07:06 still, you have to cd there before pulling bitbake updates. One could think that after git-pull of oe-core bitbake is updated as well Jan 18 11:09:57 what is the best way to see if a package is in your image so you can optionally add configure options for it? Jan 18 11:10:20 i.e. if (mysql) then --with-mysql Jan 18 11:12:53 it's the other way around, you RDEPENDS, RRECOMMENDS, RSUGGESTS runtime dependency on mysql in the packages using it Jan 18 11:13:19 jackmitchell: in oe-classic you got a list of the packages, but now I don't know Jan 18 11:14:39 essentially what I am trying to do is configure PHP so that if you have mysql in your image then it will configure php with mysql support, otherwise it won't Jan 18 11:14:52 at the moment php depends on mysql5 Jan 18 11:15:09 I could try putting it in RSUGGESTS Jan 18 11:15:30 and then assume that PHP is clever enough to pickup if it is available or not... Jan 18 11:15:55 but it might get built after PHP and then we would have a configured php (without mysql) but mysql in the image Jan 18 11:16:02 ant_work: if you weren't familiar with the way git works, I guess so Jan 18 11:16:07 jackmitchell: a dependent configuration... interesting but I don't know how to do that Jan 18 11:16:43 at the moment php depends on mysql5 Jan 18 11:16:57 so mysql should be built and staged before php Jan 18 11:17:03 jackmitchell: you can't really do things dynamically like that, you should be prescriptive up-front as to what the configuration should be Jan 18 11:17:21 indeed, but I don't want mysql5 in my imagem but I do want PHP Jan 18 11:17:34 at the moment I use an out of tree PHP recipe Jan 18 11:17:48 and just take out the mysql5 bits Jan 18 11:18:05 but I think it would be better to make it a more generic recipe Jan 18 11:18:08 jackmitchell: maybe do a custom php recipe ? Jan 18 11:18:09 can php not be built in a modular fashion so that the mysql-specific bits can be packaged separately? Jan 18 11:18:34 bluelightning: maybe, it's a mighty beast with minimal build documentation Jan 18 11:18:57 yes, I noticed when I was working on the modphp recipe :/ Jan 18 11:18:57 I think that might be the way to go Jan 18 11:19:32 I might investigate the ArchLinux package as that builds it all seperatly Jan 18 11:21:46 urgh, a big shared library config Jan 18 11:21:47 https://projects.archlinux.org/svntogit/packages.git/tree/trunk/PKGBUILD?h=packages/php Jan 18 11:23:08 I've looked at this before and decided I just don't have the drive for it - they really need to sort out some documentation for it Jan 18 11:23:31 it would be a matter of starting small and building up the recipe I guess Jan 18 11:25:31 it looks like it would have a main php recipe and then lots of little recipes which get built but not necessarily go into the image unless specified Jan 18 11:30:27 I wouldn't have thought it necessary to have multiple recipes, just separate packages Jan 18 12:27:22 when I start Hiawatha webserver I get the following errors: Jan 18 12:27:23 root@beaglebone:/etc/hiawatha# Warning: couldn't create logfile /var/log/hiawatha/system.log Jan 18 12:27:23 Warning: couldn't create logfile /var/log/hiawatha/garbage.log Jan 18 12:27:23 Warning: couldn't create logfile /var/log/hiawatha/exploit.log Jan 18 12:27:23 Warning: couldn't create logfile /var/log/hiawatha/access.log Jan 18 12:27:23 Warning: couldn't create logfile /var/log/hiawatha/error.log Jan 18 12:28:26 if I create the directory /var/log/hiawatha it can create the logs fine Jan 18 12:28:49 I have tried to create the direcory in the reciepe but it seems as though /var/log is remounted to somewhere else on boot Jan 18 12:29:05 and so the directory is non-existant Jan 18 12:36:23 JaMa: have you looked at php patch? Jan 18 12:39:10 hrw: yes it's in contrib/jansa/in-test Jan 18 12:39:16 and world build is running Jan 18 12:39:18 cool Jan 18 12:39:29 slowly.. because libffi change invalidated almost all sstate checksums :) Jan 18 12:41:11 and allowed me to make another 'git rm -r' Jan 18 12:42:42 ok, time to rebuild qt4-x11-free again Jan 18 12:49:57 Running task 9957 of 23185 Jan 18 12:50:29 who knows if 9,5G free space will be enough to finish this one Jan 18 12:51:21 if you got qt4 already than maybe Jan 18 12:51:39 libjson.so.0 - which package Jan 18 12:51:40 ? Jan 18 12:52:01 yes that's what I've built and removed from tmp/work after running out of space for first time Jan 18 12:52:15 the same for webkit-efl and qt4-embedded Jan 18 12:53:01 webkit-efl needs aarch64 treatment Jan 18 12:53:23 maybe I should upgrade it again now with harfbuzz in oe-core Jan 18 12:55:42 224M :) Jan 18 12:56:53 ok, looks like pulseaudio needs dependencies fixed Jan 18 12:57:42 yep. Jan 18 12:59:14 nope, solved Jan 18 13:08:52 jackmitchell: I think maybe it ought to be creating the directory on the fly, most daemons do AFAIK Jan 18 13:12:41 bluelightning: so Hiawatha should be creating the hiawatha directory in /var/log/ instead of it expecting it to already be created Jan 18 13:12:51 jackmitchell: I think so yes Jan 18 13:13:03 ok, I'll make some enquiries upstream Jan 18 13:28:25 looks like 18GB is enough to build qt4-x11-free Jan 18 13:31:50 WARNING: The free space of none is running low (0.075GB left) Jan 18 13:31:52 ;( Jan 18 13:34:54 I was building in much slower tmpfs, but you have to build it in steps (to force rm_work execution) Jan 18 13:35:31 the problem is that rm_work needs assistance anyway Jan 18 13:35:42 I killed also deploy Jan 18 13:36:05 it would be nice to kill S when D is created Jan 18 13:36:23 thogh about it Jan 18 13:36:30 https://bugzilla.yoctoproject.org/show_bug.cgi?id=3221 Jan 18 13:36:32 but what if do_install/package fails? Jan 18 13:37:06 you can replace do_compile with _setscene like when rm_work is executed before do_build Jan 18 13:38:50 and I think that most packages fail in do_compile or do_install, so if you clean it between do_install and do_package you're pretty safe Jan 18 13:41:11 JaMa: not if you work on fixing recipe and forget to disable rm_wokr Jan 18 13:52:45 has anyone used perf recently? Jan 18 13:52:57 I am trying to use it for my application as I have done in the past Jan 18 13:53:18 but the perf.data file that it spits out causes perf report to use 100% cpu and never actually report Jan 18 13:53:35 I have also tried moving the file to my i7 machine and using perf report there but the same problem occurs Jan 18 14:00:57 gm Jan 18 14:07:28 Collected errors: Jan 18 14:07:29 | * satisfy_dependencies_for: Cannot satisfy the following dependencies for busybox: Jan 18 14:07:31 | * update-alternatives-cworth * update-alternatives-cworth * Jan 18 14:07:32 | * opkg_install_cmd: Cannot install package busybox. Jan 18 14:08:03 was opkg built? Jan 18 14:08:19 JaMa: tnx, checking. Jan 18 14:09:25 JaMa: seems it was not, does this mean I have package support disabled somewhere (could be)? Jan 18 14:25:00 likewise_: JaMa: I think I know why that is happening... we add the dependency on VIRTUAL-RUNTIME_update-alternatives too late for the system to figure out it needs to build opkg to satisfy it Jan 18 14:25:24 see OE-Core commit 8a4c5765fb56d17c3d82ff6db47863e934c9fcbe Jan 18 14:26:54 workaround for now would be just to explicitly mention VIRTUAL-RUNTIME_update-alternatives somewhere e.g. in IMAGE_INSTALL Jan 18 14:29:44 bluelightning: just mentioning it as in IMAGE_INSTALL += "VIRTUAL-RUNTIME_update-alternatives" > Jan 18 14:29:49 > = ? Jan 18 14:30:52 hm, I'm trying to add a depend on the build kernel in my recipe by adding: Jan 18 14:30:53 do_install[depends] += "virtual/kernel:do_deploy" Jan 18 14:31:00 but this does not work Jan 18 14:31:21 the dependency is not added, anyone seeing an obvious problem ? Jan 18 14:33:21 qt build times are insane Jan 18 14:35:53 today is qt4 build day. Jan 18 14:38:26 that will take most of the day :) Jan 18 14:40:04 same here, lots of clock cycles on Qt Jan 18 14:44:20 first one was 'lets check how it fail for aarch64 target'. second is 'lets build it with my patch'. third will be 'I have patch for OE so probably will rebuild' Jan 18 15:59:50 khem: dec6f912d68f52748f645b6af411f189cf464f85 this breaks core-image-sato-sdk for me. Jan 18 16:02:08 (for TCLIBC="uclibc") Jan 18 16:57:35 Summary: There were 5268 ERROR messages shown, returning a non-zero exit code. Jan 18 16:57:37 L( Jan 18 18:51:35 morning all Jan 18 19:12:55 Hi pb_ longtime Jan 18 20:20:27 are there any write-ups for how people manage their OE builds in production? For example, they must keep local forks of all of the OE repos, but then how do they branch those for their own changes, etc. Jan 18 20:40:02 how can i force to use a special kernel recipe when building an image ? Jan 18 20:45:31 hi khemm Jan 18 20:45:36 s/mm/m/ Jan 18 20:54:45 DJWillis. ping **** ENDING LOGGING AT Sat Jan 19 02:59:59 2013