**** BEGIN LOGGING AT Mon Jan 24 02:59:58 2011 Jan 24 03:01:09 03Korey Lu  07master * rf4cab5cb50 10openembedded.git/recipes/yaffs2/ (yaffs2-utils/mkyaffs2image-2.patch yaffs2-utils_cvs.bb): Jan 24 03:01:09 yaffs2-utils: Fix yaffs2 image generation fix Bugzilla/5483 Jan 24 03:01:09 Before making this change, a yaffs2 rootfs image generated by Jan 24 03:01:09 'IMAGE_FSTYPES = Jan 24 03:01:09 "yaffs2"' makes Linux kernel fail to find init. Jan 24 03:01:10 After makeing this change, the generated yaffs2 image would work. Jan 24 03:01:11 Signed-off-by: Khem Raj Jan 24 03:26:23 03Khem Raj  07master * r45fcdc2aed 10openembedded.git/recipes/udev/attr_2.4.44.bb: Jan 24 03:26:23 attr_2.4.44.bb: Dont add -lintl if doing native build and target is uclibc Jan 24 03:26:23 Signed-off-by: Khem Raj Jan 24 04:01:59 03Khem Raj  07master * r808193c1f2 10openembedded.git/recipes/linux/ (3 files in 2 dirs): Jan 24 04:01:59 gumstix-kernel_2.6.21.bb: Fix compile failure of sumversion.c fix Bugzilla/5142 Jan 24 04:01:59 Signed-off-by: Khem Raj Jan 24 05:59:31 Hi, I am having a problem with bitbake 1.10.1 & 1.10.2 regarding special character handling Jan 24 05:59:39 https://lists.berlios.de/pipermail/bitbake-dev/2011-January/000774.html Jan 24 05:59:55 Any opinions on this ? Jan 24 06:26:01 03Pau Espin Pedrol  07master * r47f6429db1 10openembedded.git/recipes/openmoko-3rdparty/emtooth2_svn.bb: Jan 24 06:26:01 emtooth2: upgrade to new revision 127 Jan 24 06:26:01 Signed-off-by: Martin Jansa Jan 24 06:49:29 gm Jan 24 08:36:50 03Khem Raj  07master * rbb4c101bb2 10openembedded.git/recipes/mc/mc_4.7.5.bb: Jan 24 08:36:50 mc_4.7.5.bb: Add recipe for 4.7.5 Jan 24 08:36:50 * This works equally well on uclibc systems. Older versions Jan 24 08:36:50 wont compile Jan 24 08:36:50 * Fixes bugzilla/1749 Jan 24 08:36:51 Signed-off-by: Khem Raj Jan 24 08:54:13 good morning Jan 24 08:55:53 good morning Jan 24 09:02:49 03Koen Kooi  07org.openembedded.dev * r1fec4ff6b5 10openembedded.git/conf/distro/ (2 files in 2 dirs): Jan 24 09:02:49 Revert "angstrom-2010.x.conf: Prefer binutils 2.21 and uclibc-git" Jan 24 09:02:49 This reverts commit e5274a6cceb56653971f355a4f19079831efb832. Jan 24 09:02:49 This was pushed without maintainer sign-off and breaks angstrom Jan 24 09:09:43 03Khem Raj  07master * rd10ef750c5 10openembedded.git/recipes/gcc/gcc-cross-kernel.inc: (log message trimmed) Jan 24 09:09:43 gcc-cross-kernel.inc: Install it in STAGING_BINDIR Jan 24 09:09:43 * After moving toolchains under native sysroot Jan 24 09:09:43 gcc-cross-kernels stopped working becaue it could Jan 24 09:09:43 not find the remaining gcc programs correctly. Jan 24 09:09:44 ${STAGING_DIR_NATIVE}${prefix_native}/bin/ is not Jan 24 09:09:44 same as STAGING_BINDIR for cross recipes Jan 24 09:34:44 gm Jan 24 09:50:28 anybody knows which recipe is responsible for mkfs.* tools? Jan 24 09:50:46 e2fsprogs for ext[234] Jan 24 09:50:55 dosfstools for mkfs.vfat etc Jan 24 09:52:41 hrw: thank you Jan 24 09:53:41 mckoan: dpkg -S `which mkfs.ext2` on debian systems will tell you under debian. for most situations it is same in OE Jan 24 09:57:57 hrw: good to know Jan 24 09:58:26 BTW I am unable to find the reciper for the command 'declare' Jan 24 09:59:27 let me guess - it is bash internal Jan 24 10:00:45 hi hrw and mckoan Jan 24 10:00:51 hi woglinde Jan 24 10:01:07 hi Henning Jan 24 10:03:12 hrw: so if I'm using busybox, I should not have declare Jan 24 10:04:32 mckoan: probably Jan 24 10:04:47 mckoan yes seems so Jan 24 10:04:58 try to avoid bashism where you can Jan 24 10:05:31 woglinde: I would like Jan 24 10:05:39 you cannt be sure on that on the other side dash, zsh or some other shell isnt run Jan 24 10:07:46 mckoan where is the problem? Jan 24 10:26:36 woglinde: I need the command 'declare' Jan 24 10:29:28 mckoan o.O Jan 24 10:31:39 good morning Jan 24 10:31:43 gm ericben|away Jan 24 10:31:53 gm ericben :-) Jan 24 10:31:54 hi ericben and likewise Jan 24 10:32:02 hi woglinde Jan 24 10:32:12 that's monday, OE's playground is open again ;-) Jan 24 10:33:07 just returned from holidays, so the playground is fresh and waiting for me :-) Jan 24 10:41:41 03Eric Bénard  07org.openembedded.dev * r900db4a837 10openembedded.git/recipes/busybox/ (5 files in 3 dirs): Jan 24 10:41:42 busybox: introduce git version Jan 24 10:41:42 which is not a default preference. Jan 24 10:41:42 Signed-off-by: Eric Bénard Jan 24 10:41:52 03Eric Bénard  07org.openembedded.dev * r51d31435b1 10openembedded.git/recipes/busybox/busybox_1.18.2.bb: Jan 24 10:41:52 busybox-1.18.2: add latest fix Jan 24 10:41:52 wc: fix swapped -c and -m Jan 24 10:41:52 1336f89d59fc571ef19f3cb74ced0580b3331a96 Jan 24 10:41:52 Signed-off-by: Eric Bénard Jan 24 10:50:25 hi djwillis Jan 24 11:18:52 woglinde: actually I can avoid declare ;-) so also bash **** BEGIN LOGGING AT Mon Jan 24 11:22:50 2011 Jan 24 11:24:37 mckoan *g* Jan 24 14:42:21 Hi guys, I'm getting checksum errors on coreutils-native_8.5.bb Jan 24 15:03:40 03Víctor Manuel Jáquez Leal  07org.openembedded.dev * re0ec546d18 10openembedded.git/recipes/u-boot/ (2 files in 2 dirs): (log message trimmed) Jan 24 15:03:40 u-boot-sakoman: fix break caused by the new binutils Jan 24 15:03:40 New assemblers need -march=armv7-a+sec on command line or .arch_extension sec Jan 24 15:03:40 inline to enable use of the smc instruction. Jan 24 15:03:40 This patch adds as-instr to check the latter to conditionally enable the Jan 24 15:03:41 former in AFLAGS for files that use smc. Jan 24 15:03:41 Signed-off-by: Víctor Manuel Jáquez Leal Jan 24 15:03:51 03Víctor Manuel Jáquez Leal  07org.openembedded.dev * r4f24435fbb 10openembedded.git/recipes/linux/ (3 files in 2 dirs): (log message trimmed) Jan 24 15:03:51 linux-omap4: fix break caused by the new binutils Jan 24 15:03:51 According to this discussion [1], a couple patches where required by the Jan 24 15:03:51 kernel using the latest binutils (2.21). Jan 24 15:03:52 This patch add those patches, which were posted in the Ubuntu's kernel Jan 24 15:03:52 repository. Jan 24 15:03:53 1. http://comments.gmane.org/gmane.comp.handhelds.openembedded/40650 Jan 24 15:03:53 03Víctor Manuel Jáquez Leal  07org.openembedded.dev * rf1c936b88f 10openembedded.git/recipes/linux/linux-omap4_2.6.35.3.bb: Jan 24 15:03:54 linux-omap4: push to a newer SRCREV Jan 24 15:03:54 Signed-off-by: Víctor Manuel Jáquez Leal Jan 24 15:03:55 03Koen Kooi  07org.openembedded.dev * r8b3c95117b 10openembedded.git/conf/machine/include/omap3.inc: Jan 24 15:04:37 omap3.inc: bump MACHINE_KERNEL_PR for kernel changes Jan 24 15:04:37 Signed-off-by: Koen Kooi Jan 24 16:43:14 hi Jan 24 16:46:48 hi blindvt Jan 24 17:04:27 hi all Jan 24 17:36:58 hey all Jan 24 17:37:41 is it possible to get the version of a used recipe from within another recipe? i thought i could use PREFERRED_VERSION_foo, but that doesn't seem to work unless it's set explicitly Jan 24 17:37:46 hi kergoth Jan 24 17:38:16 in the specific case, i'd like to use the version in my IMAGE_CMD Jan 24 17:38:21 nope, there's no way to do that. preferred version isn't enough Jan 24 17:38:27 to find the correct bootloader binary Jan 24 17:38:31 they could specify the rpeference against anything the recipe PROVIDES, for one Jan 24 17:38:40 for two, it could be selected via DEFAULT_PREFERENCE, not preferred version Jan 24 17:49:50 I keep running into build failures in do_rootfs with things like: resolve_conffiles: Existing conffile /mnt/sda1/oe/tmp/rootfs/myimage/etc/device_table is different from the conffile in the new package. The new conffile will be placed at /mnt/sda1/oe/tmp/rootfs/myimage/etc/device_table-opkg. Jan 24 17:50:08 what exactly is going on there and why is it an occasional build failure? Jan 24 17:52:00 it shouldn't be a failure afaik, it usually indicates that the conffile was modified by the user, so it's saving it for safe keeping -- might be worth checking to see what opkg's behavior is when a single conffile is "owned" by multiple packages though Jan 24 17:52:58 hmm Jan 24 17:53:57 ahh... Jan 24 17:54:13 yes your right, that is not the error - was misleading me and hiding the actual error Jan 24 17:54:22 :) Jan 24 17:54:34 'collected errors' at the end of the log file is a bit misleading ;P Jan 24 17:57:04 heh, i guess it didn't collect them all? Jan 24 18:03:31 so the actual 'occasional build issue' that I get is a seriels of errors in do_rootfs because of Unknown packages, which if I build them manually the image do_rootfs will work properly Jan 24 18:03:36 timestamp issues? Jan 24 18:20:10 03Tom Rini  07org.openembedded.dev * r46cc091e64 10openembedded.git/recipes/openssl/openssl.inc: Jan 24 18:20:10 openssl.inc: Simplify our check for a 'linux' OS. Jan 24 18:20:10 Signed-off-by: Tom Rini Jan 24 18:21:57 * khem yawns needs some tea mit milch Jan 24 18:25:23 ooh, i didn't realize rxvt-unicode version 9.10 has the 256 color patch applied with a configure option to enable it Jan 24 18:25:29 finally no more 88 color Jan 24 18:25:33 hrm Jan 24 18:25:36 * kergoth builds it Jan 24 18:25:42 kergoth: cool Jan 24 18:26:02 kergoth: I use putty on windoze and it has 256 color support Jan 24 18:26:29 i like rxvt-unicode, nice and simple. konsole and gnome-terminal are a bit too hefty for my taste, i use a tiled wm rather than a terminal with tabs Jan 24 18:26:36 * kergoth nods Jan 24 18:26:59 * khem used to use fluxbox Jan 24 18:27:10 now I use KDE Jan 24 18:27:13 me too, ages ago Jan 24 18:27:14 and konsole Jan 24 18:27:37 i use Awesome for a wm nowadays, along with manually starting all the gnome background bits Jan 24 18:27:52 hmm Jan 24 18:28:08 I like eyecandy stuff Jan 24 18:28:12 hmm... my Depends line in ipk files doesn't show library versions from shlibdeps anymore. is this intended behavior? is this happening to other distros, too? Jan 24 18:28:24 so I keep a system just like it is dump and ssh into real work horses Jan 24 18:28:30 i love tiled workspaces for terminals since you don't have to think about window positioning at all, it arranges them for you, and they're all visible at all times within that workspace -- but then on a different one it acts like a normal wm for things like the web browser Jan 24 18:28:34 nice Jan 24 18:28:51 do u have a screenshot Jan 24 18:29:10 I wonder what would it take to move to opkg-utils Jan 24 18:36:59 angstrom's ipks of 2010 don't have shlibdeps versions either, but 2008 has, e.g. http://www.angstrom-distribution.org/feeds/unstable/ipk/glibc/i686/gstreamer/gst-plugin-avi_0.10.25-r11.0.6_i686.ipk vs. http://www.angstrom-distribution.org/feeds/2008/ipk/glibc/i586/gstreamer/gst-plugin-avi_0.10.13-r1.1_i586.ipk Jan 24 18:37:19 interesting Jan 24 18:40:37 * Crofton slaps khem with a wet noodle Jan 24 18:48:28 hey Crofton Jan 24 18:50:39 XorA|sleep, http://theoatmeal.com/comics/legos Jan 24 18:50:41 hey Jan 24 18:53:40 * khem takes a stab at removing ipkg-utils Jan 24 18:55:22 I find several references to a OE/build/profile.. but where would i find it Jan 24 19:00:21 khem: import few changes from poky and you will be done in >50% Jan 24 19:04:47 hrw: yes I am done doing changes Jan 24 19:04:54 testing build now Jan 24 19:05:03 hrw: do u have any particular commits Jan 24 19:05:07 I should look at Jan 24 19:05:25 image.bbclass is one that needed change others were not that serious Jan 24 19:16:28 for US job seekers: http://jobs.linux.com/job/open-source-embedded-linux-developer-urbana-il-relocation-electronic-search-6749f65a2d Jan 24 19:39:55 kergoth: the semantics of explode_deps seem to have changed. package_ipk uses it in a different way than the comment in bitbake suggests Jan 24 19:40:22 hmm, in what way? Jan 24 19:40:53 package_ipk expects the version to be included Jan 24 19:41:03 its likely it was one of the commits merged from poky, i haven't touched that code directly.. ever i think, i think originally it was a copy/paste from portage or something Jan 24 19:41:38 ah, it probably should be using explode_dep_versions Jan 24 19:55:50 JaMa|Wrk: applied your EV_VERSION patch to upstream tslib, thanks Jan 24 19:56:20 JaMa|Wrk: there is of course a risk that the interface will change in an incompatible way and break it, but hopefully not ;) Jan 24 20:18:31 kergoth: thanks Jan 24 20:50:30 Heinervdm: hello, I think there is a problem with git server (http access) i get always connection time out. is the problem known ? Jan 24 20:51:20 nschle85: no idea Jan 24 20:52:42 Heinervdm: can you inform someone who administrates the git server please ? Jan 24 20:53:23 cbrake: do you know sth about connection problems to the git server? Jan 24 20:53:42 Heinervdm: no, checking Jan 24 20:53:52 thx Jan 24 20:54:22 Heinervdm: thank you Jan 24 20:54:37 cgit.oe.net have no http access iirc Jan 24 20:54:46 but not sure Jan 24 20:54:58 nschle85: Heinervdm: correct, I don't know that we have ever set up http access Jan 24 20:55:00 use git:// url Jan 24 20:55:09 however, we should probably be doing that Jan 24 20:55:25 iirc again we have mirror somewhere Jan 24 20:55:28 Heinervdm cgit *is* running on git.openembedded.org Jan 24 20:55:49 at the Yocto meeting, that was one of the big gripes from corp users--they need http to get past corp proxies Jan 24 20:55:50 .net? I can't vouch for. Jan 24 20:56:00 oh, org Jan 24 20:56:03 .net == .org for git I think Jan 24 20:56:09 it's the same ip Jan 24 20:56:10 my mistake Jan 24 20:56:18 for .org and .net Jan 24 20:56:21 okay I guess they fixed that Jan 24 20:56:25 and cgit should be running Jan 24 20:56:30 repo.or.cz from topic should have http Jan 24 20:56:31 * ka6sox checks ps Jan 24 20:57:17 ya, its working Jan 24 20:57:27 nschle85: you are trying to do a git pull, or are you surfing the website? Jan 24 20:57:38 fatal: http://git.openembedded.org/openembedded/info/refs not found: did you run git update-server-info on the server? Jan 24 20:58:07 Heinervdm: git pull, via git:// url, sorry no http Jan 24 20:58:28 that's working for me, and it seems it's working for the others too Jan 24 20:58:58 Heinervdm: i am not using ssh Jan 24 20:59:38 nschle85: i tried without ssh too, can you access the website? http://git.openembedded.org/ Jan 24 20:59:58 perhaps it's a routing problem of your provider Jan 24 21:00:36 HeinervdmOff: no, connection time out too Jan 24 21:01:06 look on traceroute/tracepath to check where connection is broken Jan 24 21:01:39 nschle85: git pull via http will not work unless there is some magic I don't know about Jan 24 21:02:31 is there another OE user channel or should i ask questions here ? Jan 24 21:02:43 this is it. Jan 24 21:02:48 k Jan 24 21:03:26 so lets say i want to create a new omapevm or beagleboard(xm) based on kernel 2.6.34 where would i begin Jan 24 21:03:59 asuming i also do not have a proper enviroment configured.. Jan 24 21:04:23 * unsolo guess getting a proper env going should be just following some guideline Jan 24 21:05:15 Jay7: 18 ae-21-52.car1.Seattle1.Level3.net (4.68.105.34) 193.298 ms 176.297 ms 175.067 ms Jan 24 21:05:15 19 * * * Jan 24 21:10:19 nschle85: what's the ip you are getting for git.openembedded.org? Jan 24 21:10:43 unsolo: start from Getting started on OE's wiki Jan 24 21:11:04 unsolo: setup build env, then build something already working for your board Jan 24 21:11:14 e.g. angstrom console-image Jan 24 21:11:24 then customise it Jan 24 21:17:09 Jay7: should i make a env script rather than exporting all the PATHS Jan 24 21:23:10 unsolo: it's doesn't any difference Jan 24 21:23:25 I have own custom shell script Jan 24 21:24:03 k Jan 24 21:37:09 NOTE: :EOL while scanning string literal (, line 1) while evaluating: Jan 24 21:37:12 ${TARGET_LDFLAGS} Jan 24 21:42:16 http://pastebin.com/PYmyEmHs has more Jan 24 21:55:21 hi, I just ran bitbake pull and have had issues because bitbake is now using gcc 4.3.3 versus what I had gcc 4.3.2 ... mainly OE can't find GCC Jan 24 21:55:57 am i better off rm'ing my temp directory and start from scratch or is there a way to salvage it? Jan 24 21:56:50 bitbaking gcc fails because glibc fails seems a bit recursive to me Jan 24 22:01:10 heh Jan 24 22:05:14 alfayez, what are you building for Jan 24 22:05:25 oh wait you are using angstrom 2008 Jan 24 22:05:32 hey Crofton Jan 24 22:05:36 hi Jan 24 22:05:46 you are beagle, right? Jan 24 22:05:48 is there a new Angstrom ? I've been out of the loop for a while Jan 24 22:06:03 yep ... Jan 24 22:06:15 I recommend using angstrom 2010 for omap3, it uses gcc 4.5+patches that help on omap3 Jan 24 22:06:27 ty Crofton Jan 24 22:06:47 argh then I probably should delete the temp folder and start from scratch Jan 24 22:06:52 yeah Jan 24 22:06:58 you could save the old one Jan 24 22:07:05 in case of an emergancy .... Jan 24 22:07:11 I was so close to porting to the latest dsplink :( Jan 24 22:07:29 you have any experience with using the omap3 dsp? Jan 24 22:07:45 btw i finished setting up the git repo I'm trying to make sure my stuff works for the "latest" releases Jan 24 22:07:49 yes pupnik Jan 24 22:07:54 cool Jan 24 22:07:55 0 Jan 24 22:08:08 I need to get dsplink or dspbridge on 2.6.34 Jan 24 22:08:20 i'm assuming what the E100 users are using is Angstrom 2010 with gcc 4.5 then? Jan 24 22:08:21 I'll disappear at any moment, when the fpga builds .... Jan 24 22:08:25 yeah Jan 24 22:08:45 i'd like any framework code so i can start making sounds with the dsp Jan 24 22:08:57 uhd fails on 2008 because something in the neon intrinsucs header breaks compiles on 4.3 :) Jan 24 22:08:58 great, no probs i'll email you the link as soon as i update my stuff Jan 24 22:09:02 thanks Jan 24 22:09:20 hah that explains why i never got uhd to work Jan 24 22:09:49 pupnik are you trying to integrate into some existing framework or a standalone app? Jan 24 22:10:13 i'd like to do fmopl emulation on the dsp to free-up cpu Jan 24 22:10:26 but first, a program that can output a tone... to alsa Jan 24 22:10:42 use the alsa-utils Jan 24 22:10:44 it has that Jan 24 22:10:52 what I've done is integrate dsplink into a GNU Radio Jan 24 22:11:04 basically gnuradio takes care of all the alsa stuff Jan 24 22:11:48 do you have to pass the pcm data from the dsp back to the cpu for output to the sound device? Jan 24 22:11:49 give me a sec to fetch up my git-hub link, though it's in shambles right now .. working to update everything to the latest angstrom dsplnk ..etc Jan 24 22:12:00 oh please no hurry for me Jan 24 22:12:00 alfayez: what kernel ? Jan 24 22:12:09 would love to be able to chat with you in the future Jan 24 22:12:13 32 Jan 24 22:12:16 k Jan 24 22:12:30 cool Jan 24 22:12:37 any code that is not in a shmables is stagnant :) Jan 24 22:13:16 lol that's a positive way of looking at it Jan 24 22:13:26 i guess retention and idle works with dsplink.. Jan 24 22:13:27 i Jan 24 22:13:37 i have heard dvfs doesnt Jan 24 22:15:01 which is why im considering dspbridge Jan 24 22:16:57 what does dspbridge do exactly? is it a clean codec-engine interface? Jan 24 22:17:09 alfayez: so they claim Jan 24 22:17:34 i found codec engine to be too confusing which is why i chose dsplink instead Jan 24 22:17:58 btw does dsplink allow for multiple instances to be context switched ? Jan 24 22:18:19 similar to a cell spe Jan 24 22:18:31 it depends you can make the dsp task based which means one thread at a time Jan 24 22:18:50 what do you mean by that Jan 24 22:19:04 or swi based, which is interrupt based and you can have multiple interrupts with different priorities Jan 24 22:19:14 basically dsp services one request at a time Jan 24 22:20:13 will it swap context in the dsp itself Jan 24 22:20:29 alfayez: so gnome-radio reads in pcm from the tuner card and outputs pcm to alsa? Jan 24 22:20:50 isnt it gnuradio ?= Jan 24 22:21:05 to my understanding yes, I have been meaning to move towards to interrupts as soon as I have a release done Jan 24 22:21:29 alfayez: cool Jan 24 22:21:30 gnuradio is used for software defined radio applications Jan 24 22:21:39 its not SCA Jan 24 22:21:54 nope that OSSIE that you're thinking about Jan 24 22:23:20 pupnik: gnuradio would read audio from alsa and you can modulate a signal and send it out a radio peripheral Jan 24 22:23:21 or all the other comercial ones Jan 24 22:23:52 unsolo: yeah being in academia i usually forget about industry :) Jan 24 22:25:36 hm Jan 24 22:26:34 pupnik: technically you can use it to do all the alsa stuff you want because you don't have to use a radio to use gnuradio, you can capture and output audio in PCM format in it Jan 24 22:26:39 if that's what you want to do Jan 24 22:28:26 im guessing that angstrom builds dsplink for the target.,. Jan 24 22:28:43 so it should be easy getting it working on 2.6.34 Jan 24 22:29:49 yeah and dsplink provides you with example projects to start with, that's how i startred playing around with the dsp Jan 24 22:30:24 cool Jan 24 22:30:45 unsolo: i need to disconnect now but we can continue talking next time i log in Jan 24 22:31:18 sure np Jan 24 22:31:24 cheers alfayez Jan 24 22:31:26 im on CET and ill be here for a long time ;) Jan 24 22:31:42 who needs sunlight, eh Jan 24 22:31:46 CET? Jan 24 22:31:53 central european time Jan 24 22:31:55 Central Eutopean Time Jan 24 22:32:14 aah cool now i feel like an idiot, later guys Jan 24 22:33:04 http://maltanar.blogspot.com/2010/05/building-dsplink-with-openembedded.html Jan 24 22:33:20 first i need to get a working OE xD Jan 24 22:33:49 i need also to build my own kernels and multi-image bootloader Jan 24 22:33:59 and am failing at bitbake Jan 24 22:35:33 hehe my bitbake is screaming but not the one from the angstrom git Jan 24 22:47:18 does anyone ever put a copy of all the handhelds.org dependent recipes sources ? tinderbox seems to be building the packages just fine Jan 24 22:47:57 this is a *huge* barrier to entry for new comers, not being able to build full distro images :-( Jan 24 22:55:22 aissen, try the angstrom source mirror Jan 24 22:56:44 or the openembedded mirror Jan 24 23:04:26 grg, Tartarus : thanks, looking for it in the source right now. Jan 24 23:18:18 I fetched the tarballs, but I still can't build the package — where should I put the tarball or the sources ? Jan 24 23:19:04 anyone have an idea what this is caused by ? http://pastebin.com/PYmyEmHs Jan 24 23:22:00 drive full? Jan 24 23:26:03 aissen, in your local.conf do.. Jan 24 23:26:50 SOURCE_MIRROR_URL = "http://sources.openembedded.org/" Jan 24 23:26:57 INHERIT += "own-mirrors" Jan 24 23:27:09 That will have OE fetch from sources.openembedded.org Jan 24 23:27:37 Tartarus: thanks ! apparently I might also need CVS_TARBALL_STASH Jan 24 23:27:38 pupnik: no not even remotely close to being full Jan 24 23:27:54 question regarding oe sources Jan 24 23:28:02 can they be locally mirrored easely ? Jan 24 23:28:20 Yes Jan 24 23:28:47 You just need sources in a place and then serve them up and use own-mirrors to fetch them Jan 24 23:29:00 (ie python -m SimpleHTTPServer &, or so) Jan 24 23:32:18 k tnx Jan 24 23:32:47 btw is there a "list" of scanned modules (like blackduck) im attempting to do some CYA (Cover Your ASS) Jan 24 23:38:24 alfayez you can clean the recipes and build its bit tricky but will work better to nuke the tmpdir when gcc version change happens Jan 24 23:39:19 alfayez is gone Jan 24 23:39:26 xD Jan 24 23:39:32 got my angstrom fetch working Jan 24 23:41:02 oebb.sh script makes nasty .oe in your home dir.. Jan 24 23:41:12 without letting you know it did that Jan 24 23:45:51 does bitbake have a cache for its files Jan 24 23:46:09 the whole handling of files takes some time Jan 24 23:46:18 yeah days here Jan 24 23:46:36 since its git i recon one could use the hash as a change detection compared to an index Jan 24 23:48:00 well its not that slow.. Jan 24 23:48:02 xD Jan 24 23:48:31 Error, TMPDIR has changed location. You need to either move it back to /Angstrom/angstrom-setup-scripts/build/tmp-angstrom_2010_x or rebuild Jan 24 23:48:58 so to rebuild i do what..exactly.. Jan 24 23:51:53 think i got it Jan 24 23:52:11 moving the build/tmp is not wise w/o renaming it Jan 25 00:04:18 hi khem Jan 25 00:04:33 i still see references to ipkg-make-index Jan 25 00:04:49 in package_ipk.bbclass Jan 25 00:07:46 grg: hmmm Jan 25 00:08:43 might be worthwhile grepping for each of ipkg-build ipkg-deb-unbuild ipkg-list-fields ipkg-unbuild Jan 25 00:08:43 ipkg-compare-versions ipkg-link ipkg-make-index ipkg-upload Jan 25 00:10:14 yeah I see Jan 25 00:10:20 I have fixed some of it already Jan 25 00:14:03 grg: do they all have opkg equivalents Jan 25 00:15:21 no Jan 25 00:16:38 ok so what would be equivalents Jan 25 00:16:47 opkg-link and opkg-upload are missing Jan 25 00:16:56 i have no idea what they are for Jan 25 00:17:14 i think i performed an ipkg-link removal in the past Jan 25 00:17:35 and i don't see any references to ipkg-upload so that shouldn't be a problem Jan 25 00:17:49 ok both of them are not used Jan 25 00:18:22 ipkg-make-index, ipkg-list-fields and ipkg-build seem to be the main ones and they all have opkg- equivalents Jan 25 00:19:28 i think {i,o}pkg-utils functionality is mostly already in oe classes in various forms Jan 25 00:19:47 make-index is probably the only part thats not Jan 25 00:20:13 how hard would it be to remove all dependence upon {i,o}pkg-utils? Jan 25 00:20:37 biggest issue is ipkg-build Jan 25 00:22:44 I reimplemented part of that funcionality for regression testing a few months ago: http://code.google.com/p/opkg/source/browse/trunk/tests/regress/opk.py Jan 25 00:23:03 (i wanted to play with python3 at the time) Jan 25 00:23:53 what would replace both of them if need be Jan 25 00:24:05 a small bbclass Jan 25 00:24:10 hmm Jan 25 00:24:40 seems better to have this in core classes than calling out to external, unmaintained, python scripts Jan 25 00:25:03 but, its unlikely that i would be the one to write the code, so that that with a grain of salt Jan 25 00:25:15 s/that that/take that/ Jan 25 00:31:16 what does RP think? Jan 25 00:33:10 * kergoth isn't RP, but thinks an opkg python module/package which does the work, and which both external scripts and oe classes can import would be a decent solution Jan 25 00:34:11 :) Jan 25 00:42:31 kergoth, did you end up looking at opkg on windows? Jan 25 01:10:25 grg: I'd seriously considered just copying the scripts into OE/Poky ;-) Jan 25 01:10:51 they're a little rough Jan 25 01:10:56 grg: Putting them somewhere where we could get patches applied to them would be great :) Jan 25 01:11:12 grg: They are but they appear to work... Jan 25 01:11:26 If someone is will to improve them, I won't stop them though Jan 25 01:11:33 that's the problem Jan 25 01:12:04 i've been willing to import opkg-utils into the opkg repository, but no one has been interested in maintaining it Jan 25 01:12:55 i think rewriting just the bits that are used by us would probably be the best way to have it maintained somewhere Jan 25 01:20:33 grg: There are certainly bits in there we could just drop Jan 25 01:21:05 grg: I'd be happy to see the OE/Poky patches we've collected up to those utils be applied in the opkg repo... Jan 25 01:21:12 i'm struggling to think of what is needed other than something to build an ipk and something to build the Packages list Jan 25 01:21:41 grg: I once used some of the python bits in pstage Jan 25 01:21:55 RP__, there are a bunch of patches against ipkg-utils that never were merged to opkg-utils too Jan 25 01:21:56 grg: but it turns out we need a package manager there like a hole in the head Jan 25 01:22:32 Is opkg-utils orphaned these days? Jan 25 01:22:44 or could we just step up and clean that out? Jan 25 01:23:06 tick was the last person to work on opkg-utils... back in 2008 Jan 25 01:23:45 the last i heard from him, was basically, "here, have commit access to opkg". Haven't had any correspondence with him since then Jan 25 01:24:01 presumably he moved on Jan 25 01:24:09 might be "we" just step up and maintain them? Jan 25 01:24:20 yeah Jan 25 01:24:34 i can import opkg-utils into the google code repo Jan 25 01:24:41 and apply patches Jan 25 01:25:21 or, those with svn.openmoko.org access could apply patches Jan 25 01:25:49 * RP__ could offer a repo on git.yoctoproject.org :) Jan 25 01:25:51 * RP__ hides Jan 25 01:26:14 * grg looks everywhere, but cant find RP__ Jan 25 01:26:23 github :) Jan 25 01:28:00 one day, i'd like to rewrite opkg Jan 25 01:28:18 You're not the first person to say that Jan 25 01:28:29 I think I've said it and I've heard several others say it too Jan 25 01:28:35 and i would hope that if/when that day comes, it won't be impossible to merge it into oe/yocto/poky/blah Jan 25 01:29:06 the opkg code base is small, but waaay overcomplicated Jan 25 01:29:24 I'd like to think Yocto can provide a home for embedded focused projects like that Jan 25 01:29:59 if i ever write some code towards it, i'll let you know Jan 25 01:30:44 perhaps reserve the name "ypkg" :P Jan 25 01:32:47 heh Jan 25 01:35:52 "sometimes-ypkg" -- isn't that how the vowels are listed? :) Jan 25 01:37:28 and sometimes i dont? Jan 25 01:37:42 or as in why-pkg? Jan 25 01:38:30 "a e i o u and sometimes y" <-- the list of vowels as taught when I was a kid Jan 25 01:43:37 * mrj10 got it Jan 25 02:28:28 03Khem Raj  07master * r61a8d80ccd 10openembedded.git/recipes/udev/ (16 files in 9 dirs): (log message trimmed) Jan 25 02:28:28 udev: Remove duplicate udev firmware loading rule Jan 25 02:28:28 A duplicate udev rule calling firmware.sh exists in Jan 25 02:28:28 50-udev-default.rules and Jan 25 02:28:28 udev.rules Jan 25 02:28:28 This causes the kernel to sometimes get the wrong the firmware due to a Jan 25 02:28:29 race between the two calls. Jan 25 02:29:08 anyone familiar with newish kernels on the beagleboard (xM)? Jan 25 02:29:47 i was just wondering if they expose an interface to the on-die temperature sensor on the OMAP Jan 25 02:30:27 i wrote a simple userspace driver (well, it cheats and peeks/pokes registers using /dev/mem) for it, and was wondering if there's already something else better out there Jan 25 02:30:51 and if not, maybe i could spin what i have into a little kernel module, or even part of the kernel. just wanted to ask some folks here before spamming the linux-omap mailing list Jan 25 02:34:30 03Jeff Lance  07master * r10a36187e0 10openembedded.git/recipes/ti/ (3 files): (log message trimmed) Jan 25 02:34:30 matrix-gui: Update to the latest SRCREV Jan 25 02:34:30 * update SRCREV to integrate in new platforms am35x, am180x and am181x. Jan 25 02:34:30 * Making support files identical across platforms Jan 25 02:34:30 * Adding new applications for PRU for am180x. Jan 25 02:34:30 * Adding application descriptions. Jan 25 02:34:30 * Adding support files updates for CCSv5. Jan 25 02:34:32 03Alessandro Sappia  07master * rbd00db501c 10openembedded.git/recipes/qt4/ (qt-demo-init/qtdemo-init qt-demo-init_0.1.bb): Jan 25 02:34:32 qt-demo-init: Corrected a typo in qtdemo-init script Jan 25 02:34:32 Signed-off-by: Alessandro Sappia Jan 25 02:34:33 Signed-off-by: Khem Raj Jan 25 02:34:33 03Jeff Lance  07master * r76bba8e429 10openembedded.git/recipes/ti/matrix-gui-common.inc: Jan 25 02:34:34 matrix-gui-common: Support SVG icons Jan 25 02:34:34 * Add support to install *.svg icons as well as *.png icons Jan 25 02:34:35 * Bump up the recipe revision Jan 25 02:34:35 Signed-off-by: Jeff Lance Jan 25 02:37:02 03Martin Panter  07master * r862f1dd152 10openembedded.git/conf/bitbake.conf: (log message trimmed) Jan 25 02:37:02 Fail if the cpio command is missing (v3) Jan 25 02:37:02 This one is rebased against today's git master, with an extra line for cpio.xz Jan 25 02:37:02 -Martin Jan 25 02:37:02 >From 388638cf72e6224fba4b96cda8db26fbc43bb20d Mon Sep 17 00:00:00 2001 Jan 25 02:37:02 From: Martin Panter Jan 25 02:37:02 Date: Sun, 16 Jan 2011 04:38:03 +0000 Jan 25 02:38:22 03Khem Raj  07master * r5876b9df76 10openembedded.git/recipes/opkg-utils/ (opkg-utils/index-ignore-filenotfound.patch opkg-utils_svn.bb): Jan 25 02:38:22 opkg-utils_svn.bb: Sync with yocto Jan 25 02:38:22 Signed-off-by: Khem Raj **** ENDING LOGGING AT Tue Jan 25 02:59:57 2011