**** BEGIN LOGGING AT Sun Jan 22 02:59:57 2012 Jan 22 03:01:36 even rpm needs opkg-utils. checkmate :.( goin to sleep Jan 22 05:54:20 hello folks! Jan 22 05:55:48 hi Jan 22 06:00:06 hi khem :) Jan 22 06:06:59 i finally managed to built my first core-image-minimal jffs2 image. and after that fiddled around a bit to get initramfs-kexecboot-image built. now i got a new problem. none of my jffs2 images are bootable. at least the crl open handhelds arm bootloader does not do much if i try to boot "flash" after erasing and succesfully writing the oe images to flash. i dont have a serialcradle so watching what the bootloader does isnt possible right Jan 22 06:22:41 comparing the first few bytes of the jffs images created via oe-core with the working ones from hh.org i can see some difference from byte 3 on... Jan 22 06:59:36 bye Jan 22 10:19:36 heelo folks! Jan 22 11:26:00 could it be that the jffs2 images i built for the h3600 dont include a kernel? i still havent gigured out how to dump the content of the jffs2 images, but looking into the cpio, or tar.gz ball right next to them there is no kernel included. Jan 22 11:27:09 on the othere hand, my kexecboot image is around 2MB in size, so this looks like there is a kernel inside :) Jan 22 11:30:33 anyon familiar with jffs2dump? Jan 22 11:36:08 ok, the jffs2dump problem is almost solved. havent had mtd support enabled on my dev machine. Jan 22 11:40:51 ipaqlor: Sunday is usually quiet here, good luck :) Jan 22 11:41:24 likewise: hi. well, as long as im talking, no problem ;) Jan 22 11:41:49 ipaqlor: haha. well, I cannot help you if you suddenly stop talking. Jan 22 11:41:53 i may get scared off though if i cant see my echo anymore :) Jan 22 11:56:05 is it right that do_rootfs always needs opkg-utils? I switched everything to rpm, but still some task like task-self-hosted.bb insist on installing opkg-utils. Jan 22 11:56:22 This is a total showstopper, not a single package can be built -.- Jan 22 11:56:30 into rootfs Jan 22 11:57:24 Is there any solution in sight? Jan 22 12:04:48 cs_nbp: I think there is a variable that allows you to not have package management on the rootfs Jan 22 12:06:48 likewise: cs_nbp even in that case I think you still need opkg-utils on the host Jan 22 12:07:16 so oe unusable for me at the moment? :.( Jan 22 12:07:24 eFfeM: hi, what does opkg-utils provide then? Jan 22 12:07:37 cs_nbp: why ? Jan 22 12:07:47 can't build rootfs Jan 22 12:08:03 likewise: iirc there was some script in it Jan 22 12:08:11 svn.openmoko.ord down, opkg-utils unobtainable Jan 22 12:08:24 but in each build it seems Jan 22 12:08:35 cs_nbp: known problem, feetch it e.g. from the angstrom mirror Jan 22 12:08:55 it is supposed to be a svn Jan 22 12:09:01 and looks for one Jan 22 12:09:24 i got a opkg-utils-native tar.gz Jan 22 12:09:37 but was unable to use that Jan 22 12:09:49 cs_nbp: download this and put it in your downloads dir (probably create also an md5 file for it: ftp://opensource.axon.tv/mirror/opkg-utils_svn.openmoko.org_.trunk.src.host._4747_.tar.gz Jan 22 12:09:56 i got a opkg-utils_svn.openmoko.org_.trunk.src.host._4747_.tar.gz Jan 22 12:09:58 a ok :) Jan 22 12:10:12 I do not have opkg-utils on my target, not even in deploy/ipkg directory. Jan 22 12:10:40 likewise: I think it was opkg_make-index that was needed, not fully sure (and actually seems pointless) Jan 22 12:11:18 likewise: in tmp/x86* or so ? btw is this oe classic or oe core? I know oe-classic needs it Jan 22 12:11:38 eFfeM: but why do you need it if package management is disabled? the only thing that I can think of is the post-install scripts Jan 22 12:12:01 eFfeM: ah it's a build-host-native package? Jan 22 12:12:39 yes, it is needed natively Jan 22 12:13:09 eFfem: what do I do with this? I already tried to incorporate a tar gz yesterday for like 3 hours, changed SRCURI to absolute path to the tar.gz, but that didn't work at all Jan 22 12:14:08 cs, no change in the recipe needed, just put the gz file in your downloads dir, make the md5 file there too Jan 22 12:14:21 do that by in downloads issue the cmd Jan 22 12:14:31 md5sum opkg-utils_svn.openmoko.org_.trunk.src.host._4747_.tar.gz > opkg-utils_svn.openmoko.org_.trunk.src.host._4747_.tar.gz.md5 Jan 22 12:14:48 cs_nbp: ^^ Jan 22 12:15:10 i dont know quite what happened. but somehow i managed to delete my deploy directory. strange Jan 22 12:15:18 cs_nbp: or add a mirror to a repo that has the sources (e.g. the angstrom one but that is sometimes slow) Jan 22 12:15:28 ipaqlor: deleted or accidently moved Jan 22 12:15:45 eFfeM: no, deleted. i got 4gb ram back :) Jan 22 12:15:50 eh, diskspace ;) Jan 22 12:16:04 ipaqlor: btw wrt your kernel being in jffs2: use keeprootfs or so in local.conf and see if it is there Jan 22 12:16:16 it might be deleted in a post rootfs step before the image is created Jan 22 12:16:42 eFfeM: thanks, ill try that. Jan 22 12:16:50 yw Jan 22 12:16:52 and have fun! Jan 22 12:19:57 wtf happened here. it worked some 10 hours ago. /DEVEL/oe-core/scripts/bitbake: line 91: /DEVEL/oe-core/build/tmp-eglibc/sysroots/x86_64-linux/usr/bin/pseudo: No such file or directory Jan 22 12:20:06 cant build nothing Jan 22 12:21:22 .... /DEVEL/oe-core/build/tmp-eglibc/sysroots/x86_64-linux/ is more or less empty Jan 22 12:22:18 great, someone broke opkg_svn recipe in oe-core, now has google code uri but doesn't work anyway, and now it became harder to put the tar.gz in because all version numbers changed-.- Jan 22 12:22:21 what have i done? is there some "cleanup" bitbake thingy i could have accidently thrown in? Jan 22 12:22:58 ok, ill need to take some time to check out what is going on. this is not looking good. Jan 22 12:22:59 cu later Jan 22 12:23:15 cs_nbp: oh, not too nice, you could decide to feth the old recipe from git and use that. Jan 22 12:23:39 last month I filed a bug for this for oe-core; it was closed but I just reopened it: Jan 22 12:23:40 ah Jan 22 12:23:40 http://bugzilla.pokylinux.org/show_bug.cgi?id=1867 Jan 22 12:32:17 yeehaw seems to have worked now Jan 22 12:32:24 txhalot Jan 22 12:34:55 yw Jan 22 12:35:45 oh i just discovered: through my egg sperryments yesterday there was a code.google.com dir in svn already, maybe that's why the new recipe failed? Jan 22 12:36:14 perhaps it doesn't fail when the dir doesn't exist, ill move mine away and try Jan 22 12:36:54 yes. when the dir is nonexistent the new recipe seems to work and fetch -.- Jan 22 12:37:29 now I have 2 opkg packages.,. yesterday I needed one :D Jan 22 12:38:53 :-) Jan 22 13:14:11 sry but still not working-.- got the opkg-utils tar.gz and the tar.gz.md5 in downloads but it still tries to download from svn.openmoko.org -.- Jan 22 13:14:23 going to search the recipe that refers to openmoko now Jan 22 13:15:45 of course its the oe-core/meta/recipes-devtools/opkg-utils/opkg-utils_svn.bb recipe Jan 22 13:26:03 eFfeM: when I bitbake opkg-utils the following happens: the opkg-utils-xyz.tar.gz renames to opkg-utils-xyz.tar.gz.lock, then it tries to contact svn.openmoko.org (which fails), then the build stops and opkg-utils-xyz.tar.gz is GONE Jan 22 13:35:29 eFfeM: it's looking for a 'r7' revision file, do you by any chance now the 'r' number of your ftp link? Jan 22 13:59:12 still havent got a jffs2 image mounted. anyone knows if nandsim is the way to go with recent kernels? Jan 22 14:07:04 ipaqlor: not sure, been a while since I tried Jan 22 14:07:10 ipaqlor: why do you need to? Jan 22 14:08:01 hi bluelightning :) Jan 22 14:08:05 hey Jan 22 14:08:23 bluelightning: i built some images. but they did not boot. i suspect there is no kernel inside :) Jan 22 14:09:04 bluelightning: i built a kexec and the core image. both are not triggerable by selecting boot from flash on the crl bootloader Jan 22 14:09:56 ipaqlor: there are a number of issues Jan 22 14:10:20 bluelightning: also, using hexdump i see a difference between the hh familiar jffs2 images and the one i created. some bytes shifted to the right.... Jan 22 14:10:54 bluelightning: yes, my main issue is that the bare basics like seeing what i just did arent available right now :) Jan 22 14:12:35 bluelightning: so my piorities right now: 1. get a look into what gets spitten out after mkfs.jffs2. 2. get a kexec image ready. (this flashing is killing me. erase, wrte, verify....) i cant debug other problems while my workflow is: compile, copy to fc, flash to nand, boot Jan 22 14:15:08 reset key+action, record, record(cause its a bit lame on getting the button action), record, select image, action, record, action :) Jan 22 14:15:59 ah effem is gone-.- Jan 22 14:16:19 makes me feel like one of the workers in metropolis. buttenpusher and clock hand handler Jan 22 14:17:50 ipaqlor: hmm, I don't know a lot about jffs2 Jan 22 14:18:18 but it could well be that there are some padding/alignment settings that are incorrect now Jan 22 14:19:02 bluelightning: me too, but its to conquer, cause its in the way :) Jan 22 14:19:12 FYI, if you've enabled meta-oe you'll need to set PREFERRED_VERSION_udev = "164" Jan 22 14:19:18 since svn.openmoko.org is down I got opkg-utils tarball from here ftp://opensource.axon.tv/mirror/opkg-utils_svn.openmoko.org_.trunk.src.host._4747_.tar.gz Jan 22 14:19:24 because 173 won't work with the old kernels we're using here Jan 22 14:19:31 bluelightning: yes, thats what i am looking into if i can verify that the kernel is in the jffs2 image Jan 22 14:19:49 I put it in downloads and put the corresponding .md5 file next to it Jan 22 14:20:02 still the fetcher tries to connect svn.openmoko.org Jan 22 14:20:11 bluelightning: a good hint. ill do that right away. Jan 22 14:20:34 ipaqlor: add a .done file; the .md5 is no longer used with oe-core (well, fetch2 to be precise) Jan 22 14:20:52 I tried PR = 'r7' ... 'r0', same result Jan 22 14:21:05 what could that be? Jan 22 14:21:13 bluelightning: pardon? Jan 22 14:21:36 ipaqlor: touch opkg-utils_svn.openmoko.org_.trunk.src.host._4747_.tar.gz.done Jan 22 14:21:57 in the downloads dir Jan 22 14:22:10 ah k ill try ;) Jan 22 14:22:30 bluelightning: mhm, not necessary. i have the tar+done file. Jan 22 14:22:53 bluelightning: why do you think i would need that? Jan 22 14:22:55 ipaqlor: er, sorry I got mixed up between you and cs_nbp :) Jan 22 14:23:00 a ok :) Jan 22 14:23:54 omg that didit, genius Jan 22 14:23:56 *dances Jan 22 14:24:19 ipaqlor: what I was going to say was, you can look in tmp/work/h3600*/(image)/rootfs to see the contents of the rootfs, no need to look in the jffs2 file Jan 22 14:24:41 should PREFERRED_VERSION_udev = "164" go into local.conf or into h3600 bb or into the 2.6.xxx hh kernel bb ? Jan 22 14:24:46 stylewise Jan 22 14:25:31 ipaqlor: I think in this instance it should go into h3600.conf... at least that's where I intend to put it Jan 22 14:25:46 bluelightning: i conquer that. there is definitly a need to look into the jffs2 :) Jan 22 14:26:24 the jffs2 is produced directly from that rootfs dir, so what's in that dir is in the jffs2 file Jan 22 14:26:37 brb Jan 22 14:26:41 bluelightning: i like to reduce the amount of blackboxes :) no need to create my own :) (jffs2) Jan 22 14:33:41 may someone please lend me a shell on a 8+ core i N linuxmachine with 30GB free diskspace ? *g* Jan 22 14:34:54 i imagine how a core-image compile would have been on the skiff machines :) a full native build... a month or something :)) Jan 22 14:37:45 ipaqlor: you forgot to ask for SSD disk space Jan 22 14:38:34 likewise: hehe, i add that as optional to my request ;) Jan 22 14:39:30 likewise: but 32GB rams and a big ramdisk would be ok for me too ;) Jan 22 14:40:15 likewise: its not that this is mission critical stuff that needs any safty nets :) Jan 22 14:40:30 ipaqlor: lol Jan 22 14:41:49 ipaqlor: the machine I use for builds here is a core i7 Jan 22 14:43:02 it doesn't mean you don't still wait around for things... Jan 22 14:43:19 bluelightning: lightningfast :) here its a t7200 and 4gb ram Jan 22 14:43:56 bluelightning: shure. 16 cores, 64gb ram, and a 32gb ramdisk within. this would roch the house! Jan 22 14:46:32 bluelightning: what makes my compile times even longer, is the fackt that the machine has heat issues, so i reduce the cpuspeed when over 85degrees celsius. which happens from time to time. Jan 22 14:48:58 but that isnt that bad, considering the 800mhz cores which still beat my old celeron dual prozessor machine at about 500mhz + 512mb ram Jan 22 14:49:24 that i had when i bought the ipaq :) Jan 22 14:50:30 those glorious bp6 times :) Jan 22 14:54:28 like now :( Core 0: +88.0°C (high = +100.0°C, crit = +100.0°C), cpu MHz: 800.000 Jan 22 14:55:04 i shouldnt type that much, this obviously must be the cause for the heatup ;) Jan 22 15:05:06 ipaqlor_afk: http://bluelightning.org/files/ipaqlinux/h3600-test/ Jan 22 15:05:34 these use the 2.6.29 kernel and udev 164 Jan 22 15:06:58 ipaqlor_afk: hmm, though it does look like there's no kernel in /boot... Jan 22 16:19:59 ipaqlor_afk: ok, I've fixed up the images to contain the kernel (as well as /boot/params) and reuploaded them Jan 22 16:28:56 morning all Jan 22 16:29:45 y 17pm 'morning' to u too Jan 22 16:32:08 17pm? crumbs, that must be 29:0, way late Jan 22 16:33:15 yeh:D rither 17 or 5 pm Jan 22 16:33:22 -r+i Jan 22 16:33:33 -i+e Jan 22 16:33:33 -.- Jan 22 16:36:26 Configuring eglibc-binary-localedata-en-us Jan 22 16:36:31 Unknown package 'locale-base-en-gb' Jan 22 16:37:41 do I need to enable something in addition to ENABLE_BINARY_LOCALE_GENERATION = "1" ? Jan 22 16:38:25 happened while do_rootfs Jan 22 16:44:00 checkout from oe-core is from 12 CET or so Jan 22 17:30:43 florian: hi Jan 22 17:31:08 florian: sitting at xmodmap and trying around to get some brackets and braces Jan 22 17:32:00 hi all Jan 22 17:32:22 cs_nbp: does touchscreen calibration work for you? Jan 22 17:33:57 florian: not on my present rootfs, and forgot to try on yours, but if you want I can give it a quick run Jan 22 17:34:55 cs_nbp: no need to hurry :) Jan 22 17:35:41 florian: you remember if fn on nbp kbd is the normal mode_switch, or what is going on there? Jan 22 17:36:07 florian: doesn't seem trivial since theres that pcon thing Jan 22 17:37:59 florian: and no xev yet Jan 22 17:38:29 cs_nbp: not offhand... but worst case we can check how the original driver handles it Jan 22 17:38:51 good point... for development evtest and xev would be good Jan 22 17:38:58 florian in psion lx you mean Jan 22 17:39:06 ^^ Jan 22 17:42:58 sure :) Jan 22 17:43:24 btw. xev works with remote displays as well Jan 22 17:44:17 ah k Jan 22 17:44:34 hrm... no xev in oe-core? Jan 22 17:44:37 but I have a file Käsebrot.txt w umlauts to test on the nbp Jan 22 17:45:21 but nice to know Jan 22 17:45:31 could have used that a year ago Jan 22 17:46:01 meta-demoapps... clever naming vor development tools Jan 22 17:46:50 psion lx does it w xstroke Jan 22 17:47:34 ahno :D Jan 22 17:47:38 sryrofl Jan 22 17:50:20 found lx modmap Jan 22 18:00:06 yeh was easy could have figured out looking at psion myself Jan 22 18:00:15 [][]{} Jan 22 18:31:00 florian: every time I do xmodmap netbookpro.xmodmap however, CPU usage of openbox rises to 84% Jan 22 18:32:36 stopped after 8 minutes o_O Jan 22 18:49:43 cs_nbp: oops Jan 22 18:51:39 florian: and .xmodmao is not taken at start of course Jan 22 18:52:35 florian: same w .profile and /etc/profile Jan 22 19:08:28 guys I can't seem to get it to find my recipe. I've got my layer listed in bblayers.conf and I've got the layer.conf file (set up per the link khem gave) in my layer's conf directory. I've compared to other layers and it looks in order and it will parse the recipe correctly if i do bitbake -b /path/to/recipe/myrecipie.bb. Any ideas? Jan 22 19:09:49 oh, it also lists my layer in the OE build configuration report after parsing recipes Jan 22 19:24:16 what is your BBPATH Jan 22 19:32:36 in my layer.conf: .=":${LAYERDIR}" , in bblayers ="${TOPDIR}" Jan 22 19:33:44 and what is topdir Jan 22 19:34:00 what is BBFILES Jan 22 19:35:05 you have to append to BBFILES as well Jan 22 19:36:42 in the local.conf: BBFILES += "${LAYERDIR}/recipes/*/*.bb ${LAYERDIR}/recipes/*/*.bbappend" Jan 22 19:37:54 TOPDIR is as it should be, it is set to the directory that has my top level sources, conf and build directories Jan 22 19:39:33 also bitbake -e | grep BBPATH shows that BBPATH does indeed include my layer directory Jan 22 19:41:54 ok look at BBFILES Jan 22 19:42:07 does it contain the dir your recipe is in ? Jan 22 19:42:13 yep! Jan 22 19:43:05 hmm Jan 22 19:43:43 well, it contains /home/baileyji/beagle/sources/m2fs-layer/recipes/*/*.bb, and my recipe is /home/baileyji/beagle/sources/m2fs-layer/recipes/m2fs-image.bb Jan 22 19:43:59 So I guess it is possible that wouldn't match Jan 22 19:46:16 * spacecolonyone tries adding recipes/*.bb Jan 22 19:47:28 hi, where should game be put (what layer, and what directory in the layer) ? Jan 22 19:47:34 *games Jan 22 19:47:56 like for instance meta-openembedded/meta-oe/recipes-games ? Jan 22 19:49:32 khem, no dice Jan 22 19:58:10 spacecolonyone: typically you would make an images/ dir underneath the recipes dir and put the image recipe in that Jan 22 19:59:39 spacecolonyone: there are see problem Jan 22 20:00:01 it should be /home/baileyji/beagle/sources/m2fs-layer/recipes/*.bb Jan 22 20:00:12 if you want /home/baileyji/beagle/sources/m2fs-layer/recipes/m2fs-image.bb to be found Jan 22 20:00:27 you have extra wildcard in there Jan 22 20:01:01 doh Jan 22 20:01:33 As, thanks for the tip blue Jan 22 20:33:27 hey, ive got an old Tungsten|E I'm trying to revive with some linux. im aware that OE has a palmtt profile that appears to include this unit, but im not 100% on where i should be going for model-specific info. any help would be hugely apprecieated!! Jan 22 20:41:43 florian: seems I got my build system back :) Jan 22 20:45:03 hawk-i, hi Jan 22 20:45:09 hello GNUtoo Jan 22 20:45:11 so what was your device already? Jan 22 20:45:20 palm Tungsten|E Jan 22 20:45:33 look in the layers listed here: http://www.openembedded.org/wiki/LayerIndex Jan 22 20:45:38 in board support layers Jan 22 20:45:45 if your device is supported Jan 22 20:46:14 like for instance in meta-handheld Jan 22 20:46:14 ok. thanks Jan 22 20:47:09 oooh, icic. ill start w/ the readmes and move into the gits in a sec Jan 22 20:47:52 omap 126mhz and 32mb ram? I suspect only opie will work on this hw Jan 22 20:48:13 yep. it's the ti omap cpu Jan 22 20:48:35 SoC. good good! palmtt.conf in meta-handheld/tree/conf/machine . it was reported on... i think angstrom's site that it should work Jan 22 20:48:52 oops Jan 22 20:49:04 that's wrong model... Jan 22 20:50:20 cs_nbp: very good Jan 22 20:50:53 hawk-i: http://palmtelinux.sourceforge.net/ Jan 22 20:51:07 ya, that's just the bootloader + kernel Jan 22 20:51:26 i have an old copy of familiar here, but the site is kaput, so I can't even grab the src anymore it seems :;/ Jan 22 20:51:41 florian: any idea as to why openbox might need that load of CPU just to apply xmodmap? Jan 22 20:52:18 florian: anyways now I will finally try that fontconfig/xft thing I'm dreaming of seeing fail Jan 22 20:52:27 so you should contact author, AFAIK tungsten E support was not mainlined Jan 22 20:52:31 hawk-i, then ask oe people to mirror the sources or fork them on gitorious or something like that Jan 22 20:53:05 ahhh ok Jan 22 20:53:12 ya, ill see if i can get ahold of em Jan 22 20:53:35 cs_nbp: no idea... I wonder if we can use matchbox with it Jan 22 20:54:07 hawk-i: at least you can get kernel patch and bootloader Jan 22 20:54:33 yep. ill have to see if i can find out where this specific copy of Opie came from so i can get the patches they used Jan 22 20:54:59 hopefully ill be able to get the patches so i have a place to start isntead of being completely terminal-bound or something Jan 22 20:56:15 hawk-i: patch is here: http://sourceforge.net/projects/palmtelinux/files/PalmTELinux/beta0/ Jan 22 20:56:50 you suppose that so long as the kernel is working i should be able to build other stuff for it without much issue> Jan 22 20:58:07 hawk-i: well, opie should work on it Jan 22 20:58:17 cool. this is a good start Jan 22 20:58:42 if i can, ill see about making a nice package for you guys to stick in your trees incase others wanna add this unit Jan 22 20:58:55 i gotta step out for a short bit, ill be back soon Jan 22 20:59:21 hawk-i: bye :-) Jan 22 21:02:44 florian: is there a recipe for matchbox already? Jan 22 21:03:31 cs_nbp: since poky has it I guess so Jan 22 21:58:19 florian: will try that as soon as my rsync backup finishes Jan 22 22:03:26 anyone know if there is a fairly complete recipe for python and related modules. It looks like python-core, python-modules, and task-python-everything have all been removed. Jan 22 22:09:27 there has to be, I just ended up with a heap of 74 python packages in my ipk Jan 22 22:09:45 incl python core Jan 22 22:11:23 no dice, I see it referenced from within some bb files, but bitbake python-core just gives: ERROR: Nothing PROVIDES 'python-core' Jan 22 22:12:05 I just meant I got python-core_2.7.2-r0.5_armv5te.ipk Jan 22 22:12:21 there doesn't have to be an app-recipe for that Jan 22 22:13:00 Ah, *looks* Jan 22 22:13:04 oe-core/meta/recipes-devtools/python Jan 22 22:15:16 try something like bitbake python-native Jan 22 22:15:30 yea, The trick is always figuring out the right one, eh? I added plain simple "python" to my image but when I run python on the target i get no module named traceback! Jan 22 22:15:58 that's why i search for the exact name before I put it in a recipe ;) Jan 22 22:18:41 well, I did, and which I saw python_7.7.2.bb and python-native_2.7.2 I figured the former what the correct one. I guess one develops a feel for this over time? Jan 22 22:19:11 dunnoh was just a lucky guess Jan 22 22:20:06 does it work at least? Jan 22 22:21:01 It builds something, I'll need to recreate my image and boot it to see if it is a full fledged python install Jan 22 22:22:01 if the python packages are not included in your rootfs collection then you will only find them in your deploy/ipk dir and will have to copy over and opkg install manually Jan 22 22:22:11 that's what I did to get python running Jan 22 22:22:58 my goal is to also figure out why EXTRA_IMAGE_FEATURES+="tools-sdk" isn't causing gcc to be included in the output image Jan 22 22:23:11 * spacecolonyone looks in rootfs Jan 22 22:23:18 did u look in ipk? Jan 22 22:25:07 I'm still leaning the whole directory structure for OE so it takes me a minute :D Jan 22 22:27:51 a bunch in deploy/ipk/armv7a Jan 22 22:27:53 good evening Jan 22 22:28:00 good evening! Jan 22 22:29:26 hay Jan 22 22:29:27 hi spacecolonyone Jan 22 22:29:33 250... ;) Jan 22 22:29:46 :P;) Jan 22 22:29:53 :) Jan 22 22:31:19 cs_nbp: I think the preferred way would then be to figure out how to make bb automagically add them to the rootfs collection. Jan 22 22:31:22 bluelightning: nice. ill check it out tomorrow. coreimage build here is done. but i need to reboot :))) to get jffs2 working :) Jan 22 22:31:49 spacecolonyone: think so too, but am also a naab Jan 22 22:32:05 started on 6.1. so :D Jan 22 22:32:43 6.1 <=> 1.6 ? Jan 22 22:32:55 6th jan Jan 22 22:33:07 like 2 weeks ago Jan 22 22:33:12 yea, twas my guess Jan 22 22:34:01 I just noticed bitbake has not created a rootfs directory in tmp, which isn't what the manual says! Jan 22 22:34:17 perhaps its named sysroots ;) Jan 22 22:37:32 the manual says they are two different things. and sysroots certainly does not have the full contents of the filesystem that is in the image in creates. well, I take that back it may, but it is spread out over a few subdirectories Jan 22 22:37:51 y Jan 22 22:38:09 but you have a manual for oe-core? Jan 22 22:38:18 spacecolonyone: the manual isnt the thing l( Jan 22 22:38:23 or u use classic? Jan 22 22:39:02 spacecolonyone: i wouldnt recommend using it for the base of "edge" oe Jan 22 22:39:07 Not sure how two answer these ?s I'm using what the angstrom setup scrips pulled, which is certainly MUCH more current than the manual Jan 22 22:39:23 y its oe-core and meta-oe Jan 22 22:39:47 but at least the manual talks about basics like .= ?=, etc. Jan 22 22:40:01 y its not a complete waste of time :D Jan 22 22:40:05 cs_npb do you mean why or yes? Jan 22 22:40:20 nm I now have context Jan 22 22:40:20 i mean yes and have grease on my fingers Jan 22 22:41:08 ipaqlor, can you explain what at a conceptual level is the difference in meta-oe and oe-core? Jan 22 22:41:21 mu Jan 22 22:41:30 it means: that's not a good question Jan 22 22:41:46 the one is part of the other as I understand it Jan 22 22:41:58 spacecolonyone: googling for mailinglists and irclogs had helped me more then the manual. nd of course this nice irc# :) Jan 22 22:42:16 yes, irc good *thumbsup Jan 22 22:42:27 mailinglist too ;) Jan 22 22:42:34 indeed, much nicer than some others Jan 22 22:42:45 yes, google is my fried! Jan 22 22:42:48 doh Jan 22 22:43:42 narf Jan 22 22:44:30 yeh lets hope hollywood doesn't censor this Jan 22 22:44:51 embiluniax Jan 22 22:45:17 that's the magic defense word? Jan 22 22:45:21 good to know Jan 22 22:45:54 yeah, its, like the "get out of jail" card from those famous gameses Jan 22 22:45:58 it actually works, tested it and it broke the screen of an iphone4 Jan 22 22:46:06 l) Jan 22 22:46:26 well, hard to find something that will not Jan 22 22:47:04 *nag Jan 22 22:50:13 whut? e-wm just compiled and installed without complaining Jan 22 22:54:15 spacecolonyone: mhm, conceptual, well. if you search for layers in the usermanual, you may get 0 hits :) besides that i havent yet exactly figured out in what direction this is going. still much patchwork for a while i guess. Jan 22 22:54:43 well theres like those 3-4 sentences on wiki Jan 22 22:54:49 I see! Jan 22 22:56:56 spacecolonye: the good side is, there is enough to do :) Jan 22 22:57:15 yeh. eg write a manual :D Jan 22 22:57:20 !!! Jan 22 22:57:24 just kiddin Jan 22 22:57:29 no, really Jan 22 22:57:37 Heh. Jan 22 22:57:39 yeh i guess its not 1st on the list Jan 22 22:57:50 a getting started manual is what i am about to write besides compiling Jan 22 22:57:58 Right now I'll settle for figuring out why EXTRA_IMAGE_FEATURES+="tools-sdk" isn't causing gcc to be included in the output image Jan 22 22:58:14 its no big thing, since ill like to comment my scripts anyhow Jan 22 22:58:28 so you don't even have the ipks or the ipks dont get automatically installed? Jan 22 22:58:45 right now i havent really started :( Jan 22 22:58:53 bootwise Jan 22 22:58:53 ipaqlor: if you need a collection of showstoppers ill be happy to supply Jan 22 22:59:25 cs_nbp: har, got mine in a regular basis. but its all just show. Jan 22 23:00:04 cs_nbp: there plenty of ipks, Jan 22 23:00:30 yes sure colonyone but is there also eg a gcc ipk? Jan 22 23:00:42 will there is a g++ Jan 22 23:00:52 anarsoul: btw I finally tested h1940 suspend just now - you're right, echo mem > /sys/power/state does work Jan 22 23:00:54 cs_nbp: besides that, just missing variable definitions that get -eq no valuerized, not matching md5sums, unfetchable sources, missing licenses.... and jffs2 imagesi built but can not take a look into cause of evil kernelconfigofmine :) Jan 22 23:00:55 and yea a gcc Jan 22 23:01:39 thing I'm not 100% what all I really need. I got a bunch installed with opkg on my beagle then when I tried to get media-ctl to build it wouldn't Jan 22 23:02:08 so it's only not packed. no idea why. copy your ipk archive to a uSD or SD card and install manually Jan 22 23:02:24 I can't remember the specifics now, but It made me think that there might be headers missingfor libraries Jan 22 23:08:20 i shutdown. cu tomorrow. Jan 22 23:08:24 ipaqlor: cya Jan 22 23:09:01 anyone know offhand what the difference between task-sdk-target and task-core-standalone-sdk-target is? the latter only libgcc libstdc++ and LIBC_DEPENDENCIES Jan 23 00:11:21 hmm, "cannot install package python native" Jan 23 00:50:24 sweet. media-ctl built, now on to figuring out why udev isn't creating /dev/media0 Jan 23 01:05:56 gn8 **** ENDING LOGGING AT Mon Jan 23 02:59:57 2012