**** BEGIN LOGGING AT Thu Jul 19 02:59:59 2012 Jul 19 03:05:09 Prowell: heh, there are tons of scripts/setups to get going, there's no single best method. that page definitely makes it sound like that one is official, which isn't the case Jul 19 03:05:26 Prowell: e.g. my personal setup is https://github.com/kergoth/oe-core-template Jul 19 03:05:33 but angstrom and poky are the main ones, the most popular Jul 19 03:05:36 shr does their own thing too Jul 19 03:06:57 thanks. Looking at Angstrom now. I'm definitely new to embedded linux. Done RTOS before (non-linux), but this is first go at a proper embedded Linux system. Jul 19 03:07:21 np Jul 19 03:07:22 * kergoth nods Jul 19 03:07:50 I have two devices, one can actually run a standard distro like Red Hat. The other is definitely embedded or nothing (no cmov or PAE). Jul 19 03:08:00 angstrom and poky ahve slightly different focuses. angstrom is a great deal like a desktop distribution, in that it focuses on availability of package feeds, coherent and functional package upgrades and good on the fly installs from said feeds, etc Jul 19 03:09:01 I'm looking for something that can survive a read-only disk system. And a tiny one at that. Jul 19 03:09:42 * kergoth nods Jul 19 03:09:48 poky or even poky-tiny might be a good bet for you Jul 19 03:09:51 OE looked possible, but not very good start here docs. :( Jul 19 03:09:57 but i highly recommend exploring your various options Jul 19 03:10:13 angstrom is oe based, poky is yocto based, but both share the same underlying core tools and metadata Jul 19 03:10:18 content is fairly granular nowadays Jul 19 03:10:26 oe isn't just a single repository anymore Jul 19 03:10:30 oe is basically oe-core + meta-oe Jul 19 03:10:34 yocto is oe-core + meta-yocto Jul 19 03:10:38 both using bitbake, of course Jul 19 03:10:47 s/yocto/poky/ Jul 19 03:11:22 meta-oe harder to get going with? Jul 19 03:14:00 what i just described was about what layers are included, not how they're accessed or set up Jul 19 03:14:19 angstrom uses oe-core+meta-oe+other layers, and has its own setup scripts to obtain those things and do your builds Jul 19 03:14:30 poky uses oe-core+meta-yocto+others and uses the oe-core setup script Jul 19 03:34:46 thank you very much for your time kergoth. I have a lot of reading, and not much time. Jul 19 03:35:05 np Jul 19 06:11:31 hello all Jul 19 06:12:19 other than angstrom distribution... it is possible to build with debian distro in OE ... and how? Jul 19 06:30:00 snkt not sure what you want to achieve, you can create your own image/distro with oe Jul 19 06:30:22 building debian with oe seems a little bit odd as debian has its own cross build system iirc Jul 19 06:31:55 ok Jul 19 06:35:09 snkt: or do you want to build OE projects under debian (which should be quite possible) Jul 19 06:35:41 eFfeM_work, ya... but how? Jul 19 06:36:11 actually I want to see the filesize optimization with different toolchain Jul 19 06:36:12 snkt probably best follow the ubuntu instructions as ubuntu is debian based Jul 19 06:36:37 filesize for build env? or filesize of generated image ? Jul 19 07:02:08 eFfeM_work, filesize of generated image... Jul 19 07:02:50 snkt: filesize of generated img should be the same independent of your build host Jul 19 07:03:06 as oe will build its own gcc and use the same gcc version on both hosts Jul 19 07:03:48 you might want to try e.g. codesourcery instead of gcc though (and use code sourcery as prebuild toolchain) Jul 19 07:03:52 haven't tried that myself Jul 19 07:04:08 eFfeM_work, ok I got it... that means it uses the same gcc for all distro... Jul 19 07:04:24 yes (for all host distro) Jul 19 07:04:43 different target distro's could use different gcc Jul 19 08:51:03 hi Jul 19 08:51:34 have some issue regarding building bitbake image ! It would be great if any soul could help me Jul 19 08:51:35 ! Jul 19 08:55:01 guys i have installed oe, for crosscompiling projects to beagleboard. And i used to use it without error until some time ago... I think it began after i updated it. Now when i compile my project with "make all" it gives this error : ti/c6accel/ce/C6ACCEL.xdc", line 17: can't resolve name (ti.sdo.ce.universal.IUNIVERSAL Jul 19 08:57:12 aamir_: generally it helps if you explain the issue Jul 19 08:57:52 I am using BitBake Build Tool Core version 1.12.0, bitbake version 1.12.0 Jul 19 08:58:09 i am newbie Jul 19 08:58:16 working on imote2 Jul 19 08:58:39 i have c based code and I want to run that code on imote2 Jul 19 08:58:51 parana: no idea on this Jul 19 08:59:00 for now I am building image of bitbake Jul 19 08:59:11 using the following command Jul 19 08:59:21 $bitbake linux Jul 19 08:59:40 Can I serialize the do_fetch-function of two recipies? Background is that this two recipies depend on the same overloaded svn-server. Jul 19 08:59:45 eFfeM_work: what should i do ? google says nothing.. Jul 19 08:59:48 aamir_: i suggest to read the yocto doc and start with the tutorials Jul 19 09:00:20 parana: no idea what you did and what version of oe you are using. is this a std gcc build or something else Jul 19 09:00:42 it is the default one Jul 19 09:00:48 c6accel seems very ti specific Jul 19 09:00:54 how can i learn version? Jul 19 09:01:18 where ddi you get it from/what did you do Jul 19 09:01:34 http://code.google.com/p/opencv-dsp-acceleration/wiki/Instruction_For_Building_Examples Jul 19 09:01:52 aamir_: if you are a newbe I suggest to go for yocto (which is oe based) and use the latest release Jul 19 09:02:31 thanks eFfeM_work!! thanks for your help Jul 19 09:02:36 yw Jul 19 09:03:25 aamir_: http://www.yoctoproject.org/documentation Jul 19 09:03:32 see e./g. quick start guide Jul 19 09:03:39 okay! Jul 19 09:03:45 have fun! Jul 19 09:04:06 :) would sure do! I am really grateful ! Jul 19 09:04:17 parana: not sure what you want to do but the wiki page you gave is from aug 2010; this is probably highly outdated Jul 19 09:05:01 can you give me the recent tutorial.. Jul 19 09:05:14 so i can compile oe from the beggining Jul 19 09:05:22 parana: you might want to peek at http://www.angstrom-distribution.org/ and/or ask in #angstrom or #beagle Jul 19 09:06:01 parana: http://www.openembedded.org/wiki/Main_Page or www.yoctoproject.org might also be good starters Jul 19 09:06:38 ok thanks a lot :) Jul 19 09:07:38 yw Jul 19 09:15:45 parana building c6accel wasnt tested for a long time Jul 19 09:15:49 at least from me Jul 19 09:17:24 :/ i cant find the solution at #beagle also, they never replied.. probably i will do it from the beggining.. Jul 19 09:21:48 morning all Jul 19 09:24:01 hi bluelightning Jul 19 09:24:05 and gn ka6sox Jul 19 09:24:17 hi woglinde Jul 19 09:53:23 eFfeM_work: I was working on yocto project quick start guide Jul 19 09:53:37 ok, and ? Jul 19 09:53:56 in the guide i used the following command $ bitbake -k core-image-sato Jul 19 09:54:03 which gaves the following error :( Jul 19 09:54:13 $ bitbake -k core-image-sato Build Configuration: BB_VERSION = "1.12.0" METADATA_BRANCH = "master" METADATA_REVISION = "f5ea9a0" TARGET_ARCH = "“arm”" TARGET_OS = "linux" MACHINE = "imote2" DISTRO = "angstrom" DISTRO_VERSION = "v2012.07" ERROR: Nothing PROVIDES 'core-image-sato' NOTE: Resolving any missing task queue dependencies ERROR: Nothing RPROVIDES 'core-image-sato' NOTE: Runtime t Jul 19 09:54:38 NOTE: Runtime target 'core-image-sato' is unbuildable, removing... Missing or unbuildable dependency chain was: ['core-image-sato'] NOTE: Tasks Summary: Attempted 0 tasks of which 0 didn't need to be rerun and 0 failed. Jul 19 09:54:48 pastebin Jul 19 09:55:26 aamir_: for long texts please use pastebin Jul 19 09:55:47 I think you did not follow the instructions good enough Jul 19 09:56:56 have you installed bitbake from a repository? (e.g. apt-get install bitbake) if so, better uninstall it Jul 19 09:57:45 aamir_: and this cmd from the tutorial should have pu tyou in the build directorry Jul 19 09:57:46 source poky-denzil-7.0/oe-init-build-env poky-denzil-7.0-build Jul 19 09:58:00 ok! Jul 19 09:58:07 that is the place were you need to issue the bitbake cmd Jul 19 09:58:45 ok Jul 19 10:08:04 how can i make a conditional to a layer.conf file? The file is the one in meta-systemd in meta-openembedded. I only want to add the meta-systemd/meta-gnome etc. if I have already added meta-gnome itself Jul 19 10:13:11 mertsas hm one solution might be make a branch and delete all stuff besides meta-systemd/meta-gnome Jul 19 10:13:18 how can i solve qt4-x11-free and qt4-x11-free-glx provding the same ... as my image now builds both .. which i dont want Jul 19 10:16:15 woglinde: yeah, I already have that. But was thinking of making a conditional. So if meta-gnome is not added. meta-systemd/meta-gnome is not added. This way I could push it upstream instead of sitting on a local patch Jul 19 10:16:40 hm Jul 19 10:22:45 mertsas: rather that deleting just use BBMASK as indicated in the meta-systemd layer README Jul 19 10:24:52 I'm not sure you can really do a conditional on a layer being included Jul 19 10:28:24 oh right BBMASK Jul 19 10:28:25 hm Jul 19 10:38:16 Can I serialize do_fetch-functions from two recipies? Background is, this two recipies depend on the same overloaded server. Jul 19 10:50:13 aaah... that's what I wanted. Thanks bluelightning Jul 19 10:54:30 bluelightning: how do I that bbmask? Adding BBMASK += "${TOPDIR}/sources/meta-openembedded/meta-systemd/meta-gnome" in bblayers.conf? or do I have to do it in the meta-openembedded? Jul 19 10:54:52 mertsas: you would do it in local.conf Jul 19 10:55:30 bluelightning: so I have to do it in the openembedded layer then? Jul 19 10:56:08 mertsas: no, it's not in the layer, it's in your local configuration in /conf/local.conf Jul 19 10:56:27 cool, and it supports wildcards? Jul 19 10:56:34 mertsas: it's a regex IIRC Jul 19 10:58:17 ok, thanks Jul 19 10:58:18 mertsas: you don't need the full path, i.e. BBMASK = "meta-openembedded/meta-systemd/meta-gnome|meta-openembedded/meta-systemd/meta-efl" Jul 19 10:58:23 that should work Jul 19 10:58:54 cool, thanks Jul 19 11:31:21 is meta-oe broken atm due to the 8->4 ws conversion? Jul 19 11:32:37 yes Jul 19 11:33:31 oe-core itself is also being worked on atm Jul 19 11:35:38 is there a tool which helps in this? I noticed in my stuff that it is very difficulty to reindent python without the logic :( Jul 19 11:41:28 hi jama Jul 19 11:46:16 woglinde: hi Jul 19 11:46:51 ensc|w: ask RP.. I've used vim.. but for both python and shell tasks (in meta-smartphone) Jul 19 11:47:28 now waiting for decision about changing style guide to use 4 space indentation also in shell tasks Jul 19 12:02:44 hi all Jul 19 12:05:00 hi silvio__ :-) Jul 19 12:05:09 lol Jul 19 12:05:20 too many Silvio :-) Jul 19 12:05:26 hihi Jul 19 12:29:53 ji hrw Jul 19 12:30:24 ho woglinde Jul 19 13:43:53 someone use qt4-embedded-gles? Jul 19 13:48:53 didnt tried it sofar Jul 19 13:51:34 i have a strange error on gcc 4.4 Jul 19 13:52:04 uhm? Jul 19 13:52:06 classic? Jul 19 13:52:13 yes Jul 19 13:52:22 sorry dont support that Jul 19 13:52:38 :( Jul 19 13:52:50 ); Jul 19 18:56:48 woglinde: hey Jul 19 19:03:25 he khem Jul 19 19:04:06 khem I pushed some meta-java patches Jul 19 19:04:17 woglinde: I have some more Jul 19 19:04:34 woglinde: zerovm is default right ? Jul 19 19:04:45 for arm yes Jul 19 19:06:09 woglinde: how about x86-64 Jul 19 19:06:15 woglinde: I have an interesting fix Jul 19 19:06:32 that fixes a problem in the way we were using qemu user mode Jul 19 19:07:06 hm Jul 19 19:07:11 let me see Jul 19 19:07:14 and a powerpc compile fix Jul 19 19:07:42 woglinde: I will send updated request Jul 19 19:07:55 woglinde: btw. whats default vm for say x86_64 Jul 19 19:07:58 dont forget to pull Jul 19 19:08:16 hm the jdk one too Jul 19 19:08:31 no its still using zerovm Jul 19 19:09:10 woglinde: btw, I have a patch for ppc for openjdk Jul 19 19:09:18 but the patching is little confusing Jul 19 19:09:27 some patches get applied and some dont Jul 19 19:09:46 this patch should be applied to openjdk-src-dir Jul 19 19:09:46 sure Jul 19 19:09:50 whats the method Jul 19 19:09:53 hm? Jul 19 19:10:09 some patches are for icedtea buildsystem Jul 19 19:10:15 the other are patches for openjdk Jul 19 19:10:30 which will be applied thorugh icedtea build process Jul 19 19:10:41 ok so default is to apply to icedtea build system right ? Jul 19 19:10:48 depends Jul 19 19:10:53 what you need to patch Jul 19 19:11:00 then you have code which copied some patches into openjdk-src-dir Jul 19 19:11:17 ./openjdk-src-dir/hotspot/src/os_cpu/linux_zero/vm/os_linux_zero.hpp Jul 19 19:11:23 is the file I need to patch Jul 19 19:11:40 okay thats openjdk Jul 19 19:12:31 so how do we go about it Jul 19 19:12:50 so they need to go to DISTRIBUTION_PATCHES Jul 19 19:21:07 woglinde: unbuildable dependency chain was: ['openjdk-6-jre', 'llvm2.8'] Jul 19 19:21:14 after your stuff Jul 19 19:21:49 hm Jul 19 19:23:19 NOTE: Error expanding variable populate_packages | ETA: --:--:-- Jul 19 19:23:22 NOTE: Error during finalise of /home/kraj/work/yocto/poky/meta-java/recipes-core/llvm/llvm2.8_2.8.bb Jul 19 19:23:25 NOTE: Error expanding variable populate_packages Jul 19 19:23:27 NOTE: Error during finalise of /home/kraj/work/yocto/poky/meta-java/recipes-core/llvm/llvm2.9_2.9.bb Jul 19 19:23:30 ERROR: Unable to parse /home/kraj/work/yocto/poky/meta-java/recipes-core/llvm/llvm2.8_2.8.bb Jul 19 19:24:44 python populate_packages_prepend () { Jul 19 19:24:48 does that work ? Jul 19 19:30:13 woglinde i think tjats the issue Jul 19 19:33:26 khem here it works Jul 19 19:33:33 pull Jul 19 19:34:29 I am away now for some time Jul 19 19:35:53 will you ne herw later Jul 19 19:36:38 did u commit some thing to fix it Jul 19 20:10:22 re Jul 19 20:10:27 I fixed the tabs Jul 19 20:17:44 yes I fixed them locally too :) Jul 19 20:17:50 did you commit Jul 19 20:20:25 -> 21:04 < woglinde> khem I pushed some meta-java patches Jul 19 20:20:49 woglinde: oh there is bug Jul 19 20:20:55 hang on let me pastebin it Jul 19 20:20:57 for you Jul 19 20:21:14 o.O Jul 19 20:21:23 I switched to master branch Jul 19 20:21:32 woglinde: http://paste.ubuntu.com/1100782/ Jul 19 20:21:34 and it was parsed fine Jul 19 20:21:37 you need this patch Jul 19 20:21:48 the indentation is screwed this tha untabification you did Jul 19 20:22:20 I am on poky Jul 19 20:22:31 in any case more spaces wont hurt Jul 19 20:23:01 https://github.com/woglinde/meta-java/commit/8e11ccd8a46abcf46fe3854d40ead208b5605b53 Jul 19 20:23:09 xalan-j native has some weird issue too Jul 19 20:23:29 it fails fist time and then I do -cclean and then it rebuilds ok :/ Jul 19 20:23:56 khem does github lie or what? Jul 19 20:24:04 woglinde: it does not lie Jul 19 20:24:09 or did you not pull? Jul 19 20:24:14 woglinge I think its just number of spaces Jul 19 20:24:18 it should be 8 Jul 19 20:24:28 no Jul 19 20:24:29 4 Jul 19 20:24:31 with tab it was indented right Jul 19 20:24:40 policy is 4 Jul 19 20:24:56 did you look at the populate_packages code ? Jul 19 20:25:33 woglinde: here is xalan-j issue http://paste.ubuntu.com/1100789/ Jul 19 20:26:37 woglinde: anyway populate_packages in package.bbclass has tabs and they are aligned at 8 spaces Jul 19 20:27:29 so convert it to spaces by all means but keep the spacing 8 Jul 19 20:27:36 for best results Jul 19 20:27:53 until everywhere it becomes 4 spaces Jul 19 20:29:49 uhm Jul 19 20:29:54 no Jul 19 20:29:57 I will wait Jul 19 20:30:05 instead of changing it twice Jul 19 20:30:10 for nothing Jul 19 20:30:26 woglinde: whats your version of bitbake Jul 19 20:30:39 master ? Jul 19 20:30:43 if not try it Jul 19 20:30:45 yes Jul 19 20:31:00 OE-Core master with bb master ? Jul 19 20:31:12 hm I should scubscribe to bitbake ml Jul 19 20:31:16 to see the commits Jul 19 20:31:21 all master yes Jul 19 20:31:31 strange Jul 19 20:33:18 args Jul 19 20:33:29 woglinde: can you see anything wrong with my xalan-j issue Jul 19 20:33:32 for openjdk-7 branch I missed an update Jul 19 20:33:36 mom Jul 19 20:33:45 need to fix this one first Jul 19 20:34:02 stupid me used push for all branches Jul 19 20:34:28 hello Jul 19 20:35:07 bluelightning: so what to do for cleaning shell tasks having mixed tab/spaces? Jul 19 20:35:29 woglinde: ok I see that OE-Core has changed the indentation of package.bbclass Jul 19 20:35:51 commit 604d46c686d06d62d5a07b9c7f4fa170f99307d8 Jul 19 20:35:51 Author: Richard Purdie Jul 19 20:35:51 Date: Wed Jul 11 17:33:43 2012 +0000 Jul 19 20:35:51 Convert tab indentation in python functions into four-space Jul 19 20:35:51 Signed-off-by: Richard Purdie Jul 19 20:35:54 this is the commit Jul 19 20:35:56 which did it Jul 19 20:35:58 I dont have it Jul 19 20:36:12 so that bit it sorted Jul 19 20:49:52 ant__: well, officially it's tabs for shell scripts... I've never particularly liked that, but that's what our convention is Jul 19 20:50:26 ok, there is some work to do in meta-hh Jul 19 20:52:11 q Jul 19 20:52:51 ant__: JaMa was keen to change that though by the sounds of it Jul 19 20:55:09 good Jul 19 20:55:24 I'm momentanly lost configuring kernels Jul 19 20:55:49 hm so now let take a look at the xalan recipe Jul 19 20:55:57 I'm not sure it's worth worrying too much about outside of the core; the python ones are much more of an issue Jul 19 20:56:59 http://www.arm.linux.org.uk/mailinglists/faq.php#f3 Jul 19 20:57:11 I feel really confused about what's the difference with gcc soft-float support and the kernel NWFPE/FastFPE support? Jul 19 20:57:26 this appeared recently for pxa Jul 19 20:57:53 khem: ^? Jul 19 20:58:30 khem I see no problem offhand Jul 19 20:58:38 code isnt parallized Jul 19 20:58:42 its shell Jul 19 21:00:11 khem hm maybee we should set an explicit rm before creating the dir again Jul 19 21:00:26 if someones interrupts the compile Jul 19 21:00:53 for soft-float it is indifferent isn't? I'l let that disabled Jul 19 21:02:48 khem try http://paste.ubuntu.com/1100856/ Jul 19 21:02:50 and report Jul 19 21:03:51 woglinde in a meeting right now will try after thst Jul 19 22:43:03 Has anyone done a pull within the last 4-5 hours from oe-core? The recipe opkg-utils_git.bb is failing in the do_fetch step. Some reason the recipe's fetch command includes git remote add --mirror=fetch origin git://git.yoctoproject.org/opkg-utils which fails with the following error STDERR: error: option `mirror` takes no value Jul 19 22:45:18 it shouldn't be possible for that to happen unless you moved your build dir from one machine to another. it means it didn't build git-replacement-native Jul 19 22:56:42 kergoth: Building git-replacement-native fixed the issue. I didn't touch my build directory. I actually started from scratch 5-6 hours ago and everything worked fine. I just did a pull and tried to rebuild something again when I experienced that issue. I did make branches on various layers to fix certain warnings and kept switching back and forth between branches. **** ENDING LOGGING AT Fri Jul 20 02:59:59 2012