**** BEGIN LOGGING AT Tue Apr 20 03:00:04 2010 Apr 20 06:03:31 03Martin Jansa  07org.openembedded.dev * r6689cb4cc2 10openembedded.git/recipes/shr/frameworkd-config-shr_git.bb: frameworkd-config-shr: bump SRCREV Apr 20 06:10:33 hi everyone, i am trying to emulate opie-image by qemu. I have done a bitbake opie-image for machine-qemuarm and distro=angstrom Apr 20 06:10:58 but as soon as I try to emulate the .ext3 image generated qemu return an error as to unknown parameter Apr 20 06:11:04 please help me out here Apr 20 06:24:14 goodmorning Apr 20 06:30:16 03Koen Kooi  07org.openembedded.dev * rd37850fd54 10openembedded.git/recipes/zaurus-utils/zaurus-updater.bb: zaurus-update: add back PACKAGE_ARCH = "${MACHINE_ARCH}" otherwise you'll miss out on gnu-tar for multi-machine builds where spitz isn't first Apr 20 07:15:07 03Koen Kooi  07org.openembedded.dev * r5453d2a893 10openembedded.git/recipes/linux/multi-kernel.inc: multi-kernel: update depends after kernel.bbclass changes Apr 20 08:02:35 Hi, I am working on this problem, http://article.gmane.org/gmane.comp.handhelds.openembedded/32076. What could cause curl to choose the system linker and not the cross linker when the .o files are normally cross-compiled for ARM ? Does anybody else have the same problem ? Apr 20 08:04:03 more complete do_install log : http://pastebin.org/161245 Apr 20 08:38:09 good morning Apr 20 08:46:29 good morning Apr 20 08:47:53 hi Apr 20 09:02:14 hi, I'd like to ask something about /etc/device_tables it looks like it does not contain stuff specified in IMAGE_DEVICE_TABLES, is that right? Apr 20 09:06:10 IIUC, IMAGE_DEVICE_TABLES is used to populate /dev when building the rootfs, but on boot makedevs overrides that picking up stuff from /etc/device_table. The question is, shouldn't /etc/device_table modified following what's in IMAGE_DEVICE_TABLES too? Apr 20 09:40:03 ah there is an issue already about that: http://bugs.openembedded.org/show_bug.cgi?id=3593 Apr 20 10:02:01 Hej! When building initramfs-image.bb the directory /sys isn't created and therefore sysfs nmounted and udev not started. Where should /sys be created in this case? Apr 20 10:02:21 When building full-featured images /sys is created in base-files Apr 20 10:04:52 I've built x11-gpe-image for the h5000, I have it loaded, but X isn't starting. I can start it via the serial console and even load apps, but the touch screen doesn't seem to be functioning. Can anyone point me in the direction that I will need to look in to get the touch screen working? Apr 20 11:04:44 03Sergey Lapin  07org.openembedded.dev * rc2df527b55 10openembedded.git/recipes/sysklogd/ (files/sysklogd sysklogd.inc): sysklogd: installing init script Apr 20 11:07:14 hi, when i'm in {OETREE}/openembedded git pull is up to date, but git pull --rebase tells me linux 2.6.28.bb needs an update Apr 20 11:07:24 whats the command to update this? Apr 20 11:07:55 nlCortana: hi Apr 20 11:08:23 hi Apr 20 11:08:29 This means you changed the file and you would have to check it in and update it in Git. Apr 20 11:08:45 But you can handle local changes with Git too: Apr 20 11:08:56 "git stash" saves these changes Apr 20 11:09:07 aha, hmm it wasnt really me who changed it, but friend helped me for adding machine support and indeed overided it Apr 20 11:09:10 then you should be able to rebase Apr 20 11:09:40 after this you get your local changes back with "git stash apply" Apr 20 11:27:22 re Apr 20 11:28:06 MWelchUK: h5000? thats old... Apr 20 11:28:38 Yeah - got one lying around, got nowt else to do with it :-) Apr 20 11:29:18 Plus it has a screen, a touch screen at that, unlike just about any other target I've used. Apr 20 11:34:18 03Martin Jansa  07org.openembedded.dev * r356e71b1a7 10openembedded.git/recipes/shr/shr-launcher_svn.bb: shr-launcher: bump SRCREV Apr 20 11:38:57 03Koen Kooi  07org.openembedded.dev * r7bdb604776 10openembedded.git/classes/angstrom-mirrors.bbclass: angstrom mirrors: fix regression introduced by http://cgit.openembedded.org/cgit.cgi/bitbake/commit/?h=1.10&id=11375c3f74511ce59ad5a625bfda4121a8714345 Apr 20 11:55:03 hi Apr 20 11:57:45 i'm looking for some stable ARM target with proper X support that will work fine under qemu ? (I need something as an example for a seminar) Any suggestions are much appreciated :) I'm rather flexible as you can tell Apr 20 12:01:49 qemuarm Apr 20 12:27:43 hrw: qemuarm + X working? hmmm.... Apr 20 12:30:47 mckoan: I do not know current state but it was working when last time I used it Apr 20 12:33:35 is git.openembedded.org super slow or is it just me ? only get like 30kb/sec Apr 20 12:38:55 gilligan_: in EU is quite normal Apr 20 12:47:07 mckoan, x_X Apr 20 12:56:32 03Sebastian Spaeth  07org.openembedded.dev * r4321d61d1e 10openembedded.git/recipes/shr/ (libphone-ui-shr_git.bb libphone-ui_git.bb): Apr 20 12:56:32 libphone-ui*: bump to recent version Apr 20 12:56:32 to catch up with recent EFL changes Apr 20 12:56:32 Signed-off-by: Sebastian Spaeth Apr 20 13:36:51 hi, when trying to build linux 2.6.33 for x86, I met a problem at staging. I sent a fix : http://thread.gmane.org/gmane.comp.handhelds.openembedded/32081 which seems to fix the problem. May you please check this is the right thing to do ? Apr 20 14:29:37 huhu Apr 20 14:30:11 03Koen Kooi  07org.openembedded.dev * r717855f88d 10openembedded.git/recipes/libtiff/tiff_3.9.2.bb: tiff 3.9.2: fix QA error during configure Apr 20 14:43:59 I know it's very much off topic, if someone to flash-program something using the freescale usb tap thingie, please /msg me :> Apr 20 14:44:07 aem, succeeded that is Apr 20 14:51:02 hi, im trying to build something with oe, but is seems like bitbake searches for the frong toolchain. It is searching for arm-none-linux-gnueabi, but it needs arm-angstrom-linux-gnueabi. How to change this Apr 20 14:57:25 I'm building u-boot_2009.8 and I'm getting a ld error. It seems that the arm-oe-linux-gnueabi-ld linker does not like the -Wl option as if I removed it from LDFLAGS in the compile script u-boot links with no issues. How can I remove the -Wl perminantly from the u-boot LDFLAGS? Apr 20 14:58:37 when I say compile script I'm refering to temp/run.do_compile.xxxx Apr 20 15:01:53 is it safe just to remove it from conf/bitbake.conf LDFLAGS or will that break other packages? Apr 20 15:05:02 svolpe: it sounds like there's a problem with u-boot's Makefiles Apr 20 15:08:37 Tartarus: well even if I do arm-oe-linux-gnueabi-ld -Wl from the command line it tells me its not a valid option for LD Apr 20 15:14:09 svolpe: -Wl is a gcc argument, to pass arguments along to ld. the LDFLAGS from the bitbake metadata is expected to be given to gcc to link, not ld directly Apr 20 15:15:20 kergoth: thank you, so that is why Tartarus said its probably an issue with the u-boot makefile Apr 20 15:20:34 kergoth Tartarus thanks, I think I found the offending line in the main u-boot Makefile. I will patch it and give it a go. Apr 20 15:24:35 The xmodmaps seem to be being placed in /etc/X11, but /etc/X11/Xinit.d/12keymap expects them in /etc/X11/xmodmaps Apr 20 15:24:56 x11-image build for h5000 Apr 20 15:39:41 svolpe: indeed. good luck with it :) Apr 20 16:07:04 I have identified the offending code in the u-boot makefile but I'm not having any luck fixing it. Here is the offending code: http://pastebin.com/zvw82Bkh I would appriciate any ideas anyone may have. Apr 20 16:37:35 looking for "ant" Apr 20 16:38:07 I have succesfully built openembedded and generated several images for the beagleboard Apr 20 16:38:45 I am looking for a doc on how to install the images on an sd card to boot them.... Apr 20 16:38:52 any help is appreciated Apr 20 16:41:10 Try #beagle Apr 20 16:41:58 Tartarus: tried beagle no luck there... Apr 20 16:41:59 midnight1111_: http://elinux.org/BeagleBoard#MMC.2FSD_boot Apr 20 16:43:30 svolpe: thx been through that web page... the websites all have instructions for pre-built images... Apr 20 16:44:09 I am after how to take some images just after a build and install them on an sd card "copy to on the sd card" Apr 20 16:45:25 midnight1111_: do you mean copy the rootfs or a file system image? Apr 20 16:46:20 svolpe: so the list of files in my /stuff/build/tmp/deploy/glibc/images/om-gta01 directory include several .tar.gz rootfs and base-image files Apr 20 16:46:41 I want to copy these to the equivalent files on the sd card to boot the beagle board Apr 20 16:47:46 uImage-om-gta01-latest.bin probably goes to uImage.bin on the sd card Apr 20 16:48:34 then I still have modules-2.6.24-oe5+gitrfb42ce6724576fc173faf8abfb04aa2c36d213b7-r1-om-gta01.tgz, base-image-om-gta01.tar.gz, and Angstrom-base-image-glibc-ipk-2010.4-test-20100415-om-gta01-testlab Apr 20 16:49:08 and Angstrom-base-image-glibc-ipk-2010.4-test-20100415-om-gta01.rootfs.tar.gz which looks like the rootfs.bin file Apr 20 16:49:28 well then I recommend setting up an ext2/3 partition on the sd card, copy the rootfs over to that card and then pass the kernel the root=/dev/'whateveryoursdcardpartis' Apr 20 16:51:08 hi all, i did a bitbake opie-image for qemuarm MACHINE and DISTRO angstrom but when i am following the steps mentioned at http://linuxtogo.org/gowiki/Angstrom/qemu-system-arm I am getting this error Apr 20 16:51:15 qemu: unknown parameter '�' in 'Angstrom-opie-image-glibc-ipk-2010.4-test-20100420-qemuarm.rootfs.ext3 Apr 20 16:51:17 please help Apr 20 16:58:38 umm... anyone Apr 20 17:19:57 please help me out anyone Apr 20 17:25:02 Spyzer, maybe instead of copy-pasting the wiki,try to type the command Apr 20 17:35:52 ../qemu-system-arm -M versatilepb -usb -usbdevice wacom-tablet -show-cursor -m 64 -kernel zImage-2.6.25+2.6.26-rc4-r6-qemuarm.bin -drive Angstrom-opie-image-glibc-ipk-2010.4-test-20100419-qemuarm.rootfs.ext3 -append "root=/dev/sda rw" Apr 20 17:36:03 this is the command @GNUtoo Apr 20 17:36:43 Spyzer, does it work now? Apr 20 17:37:16 no it is not working it is always giving the same error i mentioned above/earlier Apr 20 17:37:20 Spyzer, or I've another solution Apr 20 17:37:30 yes please Apr 20 17:37:34 rename Angstrom-opie-image-glibc-ipk-2010.4-test-20100419-qemuarm.rootfs.ext3 to something simplier Apr 20 17:38:20 because '�' is an unknown character Apr 20 17:38:39 just to be shure we understand each other,in my IRC client '�' appears as '?' Apr 20 17:39:25 i did that earliere and now too, i renamed it to opie-angstrom.ext3 and now am getting the same error qemu: unknown parameter '?' in 'Angstrom-opie-image-glibc-ipk-2010.4-test-20100420-qemuarm.rootfs.ext3 Apr 20 17:39:38 oh sorry Apr 20 17:39:48 qemu: unknown parameter '�' in 'opie-angstrom.ext3' Apr 20 17:39:54 this is the new one after renaming Apr 20 17:40:05 mmm strange Apr 20 17:40:31 you have a strange character somewhere,you must track it Apr 20 17:41:04 do you think it has something to do with a possible machine and distro combination, earlier when it ran i actually gave DISTRO=minimal and machine=qemuarm Apr 20 17:41:19 now qemuarm MACHINE and DISTRO angstrom Apr 20 17:41:29 use -drive file= Apr 20 17:41:36 http://linuxtogo.org/gowiki/Angstrom/qemu-system-arm Apr 20 17:42:47 man i cannot believe that was the error Apr 20 17:43:01 :) Apr 20 17:43:14 thanks a lot Apr 20 17:43:17 :) Apr 20 17:43:38 no problem I use qemu a lot to test images Apr 20 17:44:07 so have u created any customized images of your own too??? Apr 20 17:44:32 kind of Apr 20 17:45:03 just have a look at images and tasks dirs Apr 20 17:45:38 the manual also explains it quite well Apr 20 17:46:00 ok well if u can review this... Apr 20 17:46:03 IMAGE_INSTALL = "qtehelloworld" Apr 20 17:46:05 RDEPENDS = "qte" Apr 20 17:46:05 require base-image.bb Apr 20 17:46:05 inherit image Apr 20 17:46:17 where qtehelloworld is a recipe created by me Apr 20 17:46:26 which bitbakes fine enough Apr 20 17:46:55 so do u think an image will be created which has support for qte and has the application and can run it too Apr 20 17:47:15 can anyone suggest me some arm machine that builds fine with angstrom/X and also works fine with qemu ? Apr 20 17:48:22 Spyzer: sorry I don't know it Apr 20 17:49:08 well if u can just tell that if using the rdepends variable, will that install all the qte runtime libraries in the created image?? Apr 20 17:51:24 oh.. there's qemuarm.conf .. that should do then ;-) Apr 20 17:52:27 03Koen Kooi  07org.openembedded.dev * r15188a35e2 10openembedded.git/conf/distro/include/angstrom-2008-preferred-versions.inc: angstrom: prefer xserver-common 1.33 Apr 20 17:52:38 03Koen Kooi  07org.openembedded.dev * rda8608ce9f 10openembedded.git/recipes/xserver-common/ (3 files in 2 dirs): xserver-common: add 1.33 with some extra fixes Apr 20 17:52:48 Spyzer: I don't know it. I allways put them also to the image install. someone else migth know the answer Apr 20 17:53:01 oh ok :) Apr 20 17:53:18 but it might be not a clean solution at all Apr 20 17:53:30 clean?? Apr 20 17:53:35 best would be to ask at the mailing list Apr 20 17:55:07 with clean I wanted to say the best way to do it Apr 20 18:12:43 * kergoth_ thinks about changing var expansion to expand nonexistant vars to the empty string and comparing bitbake -e output Apr 20 18:16:56 gm to all Apr 20 18:18:15 hey khem Apr 20 18:24:38 re Apr 20 18:27:17 kergoth_: hey Apr 20 18:27:33 been tired of flying for last 48 hours Apr 20 18:27:45 these volcano's are not helpful :) Apr 20 18:34:05 khem: oops Apr 20 18:34:14 jo Apr 20 18:39:57 brb Apr 20 18:48:42 florian: I was to fly into Munich from Dubai but they cancelled it and packed me on a 16 hour flight to toronto Apr 20 18:49:06 jo khem Apr 20 18:49:17 and then I had to wait there for another 10 hours to get on a flight to San fransisco Apr 20 18:49:21 khem where are you now? Apr 20 18:49:43 woglinde_: heh I have reached in valley no more transits :) Apr 20 18:50:04 khem: heh... oh well. Toronto instead of Munich... that's quite a difference :-) Apr 20 18:56:00 sheesh, just found 5 different places in bitbake itself where it was requested an expanded version of a non-string bitbake internal variable Apr 20 18:56:03 sad Apr 20 18:56:33 wait, no, more than 5 now Apr 20 18:57:11 * kergoth_ is experimenting with teaching expand to str(value) instead of choking on non-string Apr 20 18:57:15 kergoth still at cleanup? Apr 20 18:58:14 on occasion, its a good background task Apr 20 18:58:57 these angstrom bluez/fso-apm/libiconv messages can bite me. Apr 20 19:08:08 and here's a few expand()s on None.. hrmph Apr 20 19:20:49 hrmph.. MIRRORS/PREMIRRORS is nicely flexible, but what crappy usability.. i wonder how we could improve that interface for the user Apr 20 19:22:28 SRC_TARBALL_STASH was certainly easier to use than the new method, if less flexible Apr 20 20:08:02 Hi all. Have a question regarding how Bitbake creates packages. I'm specifically looking at the omniorb recipe. It was my understanding that the packages created are controlled by the PACKAGES variable, and that the files packaged are controlled by FILES_ However, the recipe I'm looking at doesn't have a PACKAGES variable but still creates -dev, -dbg, and -static packages. Where would this be specified? Apr 20 20:08:27 read conf/bitbake.conf Apr 20 20:08:38 global config metadata flows into the recipes, including things like FILES and PACKAGES Apr 20 20:08:54 also, bitbake -e is helpful Apr 20 20:09:03 bitbake -e omniorb|grep \^FILES_ Apr 20 20:09:07 bitbake -e omniorb|grep \^PACKAGES Apr 20 20:10:05 cool. thanks kergoth! Apr 20 20:10:09 np Apr 20 20:10:35 note that classes are free to manipulate those variables too. bitbake -e will show you the final values, as used in the build Apr 20 20:12:37 when I add items to my IMAGE_INSTALL variable, do I list packages there then? I was thinking it was recipe names, but then I only get the base package included with my image. Apr 20 20:12:58 yes, IMAGE_INSTALL controls what binary packages get installed into an image, and PACKAGES controls what binary packages are emitted. Apr 20 20:13:05 recipes aren't installed in an image Apr 20 20:13:07 packages are Apr 20 20:13:35 makes perfect sense! Apr 20 20:13:36 thx Apr 20 20:13:38 np Apr 20 20:34:21 hmm. Apr 20 21:27:09 hrm Apr 20 21:27:14 * kergoth_ grumbles Apr 20 21:30:29 grrr Apr 20 21:34:55 JaMa|Opera, I found a pastebin of your with the same problem i am experiencing: http://www.google.it/search?q=%22build/lib.linux-x86_64-2.6%22%20wrong%20ELF%20class%3A%20ELFCLASS32 did you solve it any way? Apr 20 21:38:11 hmmmmm Apr 20 21:55:55 ao2: seems like its failing to detect the host system type Apr 20 21:56:00 which package is it ? Apr 20 21:56:20 libffi ? Apr 20 22:06:03 ao2: that google result of mine returns "That is an invalid ID, or the post has expired." and from yours I don't recognize that problem :/ Apr 20 22:06:31 ao2: so maybe it wasn't pastebined by me.. or it's really old Apr 20 22:06:49 JaMa|Zzzz: http://ja.pastebin.ca/1869784 Apr 20 22:07:56 ah so it's the same as ao2's version Apr 20 22:09:30 then this is what was enough to resolve it back then http://git.openembedded.org/cgit.cgi/openembedded/commit/?id=a070023d0d3c1af845dd41703a8419af818d70f3 Apr 20 22:10:09 and it's still used in 2.6.4 as it was in 2.6.2 Apr 20 22:13:51 03Andrea Adami  07org.openembedded.dev * re9ec37937e 10openembedded.git/conf/machine/ (5 files in 2 dirs): Apr 20 22:13:51 zaurus-2.6.inc: merge the jffs2 bits from zaurus-clamshells.inc. Apr 20 22:13:51 * remove the legacy headers needed for vanilla Sharp images. Apr 20 22:13:51 * (the checksum was anyway broken) Apr 20 22:14:00 03Andrea Adami  07org.openembedded.dev * rde4a5bc6f2 10openembedded.git/conf/machine/include/ (zaurus-2.6.inc zaurus-kernel.inc): Apr 20 22:14:00 zaurus-2.6.inc: reorder common Zaurus settings. Apr 20 22:14:00 * reabsorb zaurus-kernel.inc Apr 20 22:28:54 JaMa|Zzzz, here's the full log: http://people.openezx.org/ao2/tmp/log.do_configure.31040 _ctypes is ntill not compiled for me... Apr 20 22:29:29 ah sorry: http://people.openezx.org/ao2/tmp/log.do_compile.4827 Apr 20 22:31:45 ao2: python for target is using libffi from build doesnt sound good to me Apr 20 22:32:13 ao2: it should use libffi compiled for target i.e. arm in your case Apr 20 22:32:21 khem, sure not, but I don't know how to fix that Apr 20 22:32:27 if you can figure it our why Apr 20 22:32:29 ah Apr 20 22:32:47 it should be easy go through the config.logs Apr 20 22:37:44 kergoth_: ERROR: no recipe files to build, check your BBPATH and BBFILES? Apr 20 22:37:56 after updating to latest on master Apr 20 22:38:15 nothing else is changed in my env I am still following same steps as I used to Apr 20 22:38:27 and BBPATH is set correctly. Apr 20 22:38:32 any hints ? Apr 20 22:44:59 khem: using collections i assume? Add require conf/collections.inc to your local.conf Apr 20 22:45:15 it was an oe change, not a bitbake one, collections.inc isn't included by default now Apr 20 22:45:23 due to interference with the bblayers stuff Apr 20 22:45:49 ah Apr 20 22:47:43 thx kergoth_ now its parsing 2 weeks is enough to get outdated in OE world :) Apr 20 22:47:49 :) Apr 20 22:47:51 np Apr 20 22:47:57 been playing around with setup using bblayers, its nice Apr 20 22:48:01 just cd into your area, bitbake Apr 20 22:48:04 no env vars necessary Apr 20 22:48:14 hmmm seems nice Apr 20 22:48:27 is bblayers a replacement for collections Apr 20 22:48:38 or a suppliment Apr 20 22:50:13 collections is easier to set up, slightly, but bblayers has other advantages. tradeoffs, atm Apr 20 22:50:20 definitely one or the other Apr 20 22:50:30 bblayers has more boilerplate necessary, but works nice once you have it set up Apr 20 22:50:35 not just one var like collections Apr 20 22:50:41 ah I see Apr 20 22:50:48 http://www.mail-archive.com/bitbake-dev@lists.berlios.de/msg00716.html Apr 20 22:50:51 I will read it Apr 20 22:51:00 03Andrea Adami  07org.openembedded.dev * r884ca29a3d 10openembedded.git/conf/machine/include/zaurus-2.6.inc: zaurus-2.6.inc: remove forgotten comments (#) and unbreak jffs2 Apr 20 22:51:58 khem: check out http://github.com/kergoth/OE-BBLayers Apr 20 22:52:02 khem: example setup Apr 20 22:53:15 ah nice Apr 20 22:56:21 kergoth_: How would I include my own overlay in your example Apr 20 22:56:33 is BBLAYERS a list ? Apr 20 22:56:36 add it to BBLAYERS and give it a conf/layer.conf like the oe one Apr 20 22:56:37 yep Apr 20 22:57:10 list separated by : Apr 20 22:57:12 you see what i mean about the setup.. if you want the bitbake priority handling for dealing with recipes in both overlays, you'll need to uncomment the BBFILE bits in the layer.confs Apr 20 22:57:15 space separated Apr 20 22:57:29 hmm ok Apr 20 22:59:46 kergoth_: now I dont have to export BBPATH Apr 20 22:59:53 yep Apr 20 22:59:57 just need to make sure bitbake is in your path Apr 20 23:00:02 i did that in my ~/.profile Apr 20 23:00:27 it searches up the tree to find conf/bblayers.conf, so as long as you're within your build area, you can bake away Apr 20 23:00:52 conf/bblayers.conf should be in build area ? Apr 20 23:01:03 actually I build in /scratch Apr 20 23:01:13 and metadata is in my /home/oe Apr 20 23:01:28 put it wherever you want, i think it makes the most sense in the root of your build, it sets TOPDIR relative to that Apr 20 23:01:37 (at least in my example it does, could do it differently) Apr 20 23:04:46 kergoth_: so now all customizations will be in overlay/conf/layer.conf right Apr 20 23:04:59 local.conf does not make sense anymore ? Apr 20 23:05:03 it makes sense. Apr 20 23:05:15 local.conf is bound to the build, layers are for a specific overlay Apr 20 23:05:19 not all your settings are overlay specific Apr 20 23:05:20 ok I see only BBFILES stuff Apr 20 23:05:29 right Apr 20 23:05:29 the example bblayers.conf i show sets BBPATH to include TOPDIR Apr 20 23:05:33 most of it is common Apr 20 23:05:37 so you can still use conf/local.conf relative to there Apr 20 23:08:15 * kergoth_ fixes some of the bitbake annoyances surrounding use of non-string objects in variables Apr 20 23:08:34 kergoth_: so it looks for layer.conf ? Apr 20 23:08:42 automatically ? Apr 20 23:08:44 it parses conf/layer.conf in each layer Apr 20 23:08:47 requires it, actually Apr 20 23:08:48 ah Apr 20 23:08:49 will error without it Apr 20 23:09:01 and how does it find bblayers.conf ? Apr 20 23:09:13 it searches up from your current location searching for 'conf/bblayers.conf' Apr 20 23:09:16 until it hits / Apr 20 23:09:23 ah Apr 20 23:09:56 so as long as I invoke bitbake in some subtree it will find it Apr 20 23:10:05 * khem tried bitbake -DDD Apr 20 23:10:05 though, i guess you could do all your build wide settings in bblayers.conf directly rather than local.conf. i have both still Apr 20 23:10:09 yeah, exactly Apr 20 23:10:31 the thing to remember with bblayers.conf and layer.conf is they're all parsed before bitbake.conf. Apr 20 23:10:43 obviously, since the BBPATH changes are needed to find bitbake.conf :) Apr 20 23:11:08 overlay priority should also be in layer.conf I guess Apr 20 23:11:27 * kergoth_ thinks it would've been less confusing to call this project.conf or something, people tend to find the notion of a "project" less confusing Apr 20 23:11:45 yeah, thats what i did in the example. BBFILE_COLLECTIONS += "whatever", BBFILE_PRIORITY_whatever .. Apr 20 23:12:23 not as elegant as the priority ordered COLLECTIONS, but its a one time thing at least Apr 20 23:12:56 ok cool Apr 20 23:13:26 * kergoth_ thinks about possibly adding variable typing directly into bitbake now that the str() calls are in the right places Apr 20 23:13:36 * khem needs to find who is pulling in gcc-cross-4.3.2+csl-arm-2008q3-39-r1 Apr 20 23:31:04 hmmmmm Apr 21 00:02:16 03Andrea Adami  07org.openembedded.dev * r3f528c6b5a 10openembedded.git/recipes/zaurus-utils/ (zaurus-installer.bb zaurus-legacy-tar.bb zaurus-updater.bb): Apr 21 00:02:16 zaurus-installer: deploy legacy gnu-tar here, not in zaurus-updater. Apr 21 00:02:16 * (only for spitz) Apr 21 01:42:01 man.. the problem with trying to add variable typing is *when* does the string become the type.. Apr 21 01:42:10 if its too early, things explode when _append/_prepend/etc are attempted Apr 21 01:42:13 hrm Apr 21 01:42:56 could do it really really lazily, construct on getVar(), but would get bitten at the same spot.. hmm Apr 21 01:44:05 i have a first pass version that does it at finalize() time, but before the anonymous python functions are exec'd.. which works, but bitbake can't leverage the typed values, nor can ConfigParsed event handlers Apr 21 01:45:28 * kergoth thinks about a possible revamp of _append/_prepend/override handling Apr 21 01:54:03 hmmmm Apr 21 01:55:20 * kergoth wonders if there are any cases where _append/_prepend are utilized unconditionally that could be converted easily to +=/=+/.=/=. Apr 21 01:55:21 yarr Apr 21 01:55:55 greetings fellow OE-eres Apr 21 01:56:08 the cool thing about this typing stuff is, it can let us ditch these really common constructs: (bb.data.getVar("SOMEVAR", True, d) or "").split() Apr 21 01:56:11 hey dit Apr 21 01:56:13 er, aditya_111 Apr 21 01:56:26 * kergoth is talking to himself some, clearly.. always dead around this time Apr 21 01:56:40 yeah I was just going to say you seem to be talking to yourself as usual Apr 21 01:56:49 abotu w/e changes you're making to recipes Apr 21 01:57:09 of course, even if people were here, I'm often talking to myself... there's a real shortage of people that delve into the OE/bitbake internals Apr 21 01:57:24 [5;5~heh Apr 21 01:57:47 that reminds me Apr 21 01:57:52 I should work on that GUI thing a bit more Apr 21 02:00:00 kergoth, well bitbake just works. So no need to look at it :P Apr 21 02:01:05 heh. well, thats the thing, the vast majority of the devs are using OE for a specific project or product, and getting it out the door is the priority, not improving the code or working on future changes Apr 21 02:01:14 which is okay, and understandable Apr 21 02:05:12 * kergoth thinks about trying to make 'inherit' work in a declarative way Apr 21 02:09:36 if montavista decides to hire in australia, i'd be happy to look at oe/bitbake guts Apr 21 02:15:18 okay, yow, moving the class inherit before the recipe makes all sorts of stuff explode Apr 21 02:15:19 hmm Apr 21 02:59:15 grr **** ENDING LOGGING AT Wed Apr 21 02:59:56 2010