**** BEGIN LOGGING AT Mon Aug 10 02:59:57 2009 Aug 10 03:03:27 Tartarus: then it is odd. Maybe I should dig up the log and file a bug? Aug 10 03:11:49 you're right, no one likes bugs Aug 10 04:54:00 03Holger Hans Peter Freyther  07org.openembedded.dev * rf20d7d2a7e 10openembedded.git/recipes/qt4/ (4 files in 2 dirs): (log message trimmed) Aug 10 04:54:00 qt4: Hack out the pg_config invocation on configure Aug 10 04:54:00 Currently the configure script of Qt is invoking pg_config Aug 10 04:54:00 which is a native application and not staged. Even though Aug 10 04:54:00 we pass the directory with the header files we will fail Aug 10 04:54:04 the postgres test as pg_config from the host will tell us Aug 10 04:54:06 something about /usr/include. Aug 10 04:54:08 03Graham Gower  07org.openembedded.dev * r4fe22eed50 10openembedded.git/recipes/portmap/ (portmap-6.0/no-pie.patch portmap_6.0.bb): (log message trimmed) Aug 10 04:54:11 portmap_6.0: Fix building on mipsel with gcc4.3 Aug 10 04:54:13 with gcc-4.4 and provides the following: Aug 10 04:54:15 | Aug 10 04:54:17 tmp/cross/mipsel/lib/gcc/mipsel-angstrom-linux/4.4.1/../../../../mipsel-angstrom-linux/bin/ld: Aug 10 04:54:19 portmap.o: relocation R_MIPS_HI16 against `__gnu_local_gp' can not be used when Aug 10 04:54:21 making a shared object; recompile with -fPIC Aug 10 04:54:23 03Holger Hans Peter Freyther  07org.openembedded.dev * rda247ac529 10openembedded.git/recipes/portmap/portmap-unslung_5-9.bb: Aug 10 04:54:26 portmap-unslung: Drop the recipe as it is obsolete Aug 10 04:54:28 According to Mike Westerhof it is obsolete and can be removed: Aug 10 04:54:34 http://patchwork.openembedded.org/patch/922/ Aug 10 04:54:36 03Graham Gower  07org.openembedded.dev * r2c748884ee 10openembedded.git/recipes/icu/files/rematch-gcc-bug.patch: Aug 10 04:54:39 icu: Fix compilation for mips using -fPIC on the rematch.o Aug 10 04:54:41 /home/grg/oe/tmp/cross/mipsel/lib/gcc/mipsel-angstrom-linux/4.4.1/../../../../mipsel-angstrom-linux/bin/ld: Aug 10 04:54:43 rematch.o: relocation R_MIPS_HI16 against `__gnu_local_gp' can not be used when Aug 10 04:54:47 making a shared object; recompile with -fPIC Aug 10 04:54:49 http://patchwork.openembedded.org/patch/920/ Aug 10 04:54:51 Signed-Off-By: Holger Hans Peter Freyther Aug 10 06:23:44 hi, i wonder if there is a script to generate a checksums.ini entry? Aug 10 06:24:43 i know there is at least a script to sort checksums.ini Aug 10 06:25:16 de_manuel_pi: tmp/checksums.ini holds the sums of "unknown" downloads Aug 10 06:25:28 you can copy them to the checksums.ini and then sort Aug 10 06:27:43 oh cool didnt find that info. so i can skip my own script :) thx Aug 10 06:28:21 de_manuel_pi: you could add this info somewhere to the wiki :} Aug 10 06:30:30 good idea! Aug 10 06:31:49 http://wiki.openembedded.net/index.php/Checksums Aug 10 06:32:04 oops Aug 10 07:27:11 another dump question :) Aug 10 07:28:06 whats the preferred way to get a small patch patch upstream? mailinglist, or mail to maintainer Aug 10 07:28:34 patch is just a package rename of recipe to forward version update Aug 10 07:32:09 de_manuel_pi: mailinglist Aug 10 07:32:55 ok Aug 10 08:05:48 *facepalm* Aug 10 08:06:23 so if i create the altboot-image for the zaurus, how do i make it into an image that is flashable on the sl-5600? Aug 10 08:12:59 good morning Aug 10 08:21:47 hrw|gone: ping Aug 10 09:14:28 morning Aug 10 09:14:56 hi woglinde Aug 10 09:16:02 03Koen Kooi  07org.openembedded.dev * rb218872e36 10openembedded.git/recipes/linux/linux-davinci_git.bb: linux-davinci: update git recipe Aug 10 09:16:08 03Koen Kooi  07org.openembedded.dev * r3ceaf72bda 10openembedded.git/contrib/angstrom/ (build-feeds.sh source-mirror.txt): Aug 10 09:16:08 Angstrom source mirror: don't mirror avis Aug 10 09:16:08 Angstrom feed builder: build more stuff Aug 10 09:17:16 hi ant Aug 10 09:20:21 03Koen Kooi  07org.openembedded.dev * r133a5c83d3 10openembedded.git/recipes/tasks/task-beagleboard-demo.bb: task-beagleboard-demo: make gnash armv7a specific since it's using boost atomic ops which are missing for pre-v7 arms Aug 10 09:31:24 good morning Aug 10 09:34:31 morning Aug 10 09:34:38 xcasex: póńg Aug 10 09:35:03 03Steve Sakoman  07org.openembedded.dev * r60f1515945 10openembedded.git/recipes/opencv/ (opencv-samples_svn.bb opencv_svn.bb): opencv: Bump SRCREV, explicitly build with ffmpeg Aug 10 09:35:03 03Steve Sakoman  07org.openembedded.dev * r76069e339e 10openembedded.git/ (4 files in 3 dirs): gnome-vfs: add 2.24.1, share patches with 2.22.0 Aug 10 09:35:04 03Steve Sakoman  07org.openembedded.dev * rb4fa8a2dee 10openembedded.git/ (conf/checksums.ini recipes/gnome/gvfs_1.3.3.bb): gvfs: add 1.3.3 Aug 10 09:36:14 hi hrw Aug 10 09:36:21 hrw: nm i figured it out Aug 10 09:36:50 ok Aug 10 09:36:54 xcasex: what was a problem? Aug 10 09:37:41 hrw: was trying to find a document detailing how to build flashable images, apparently i wasnt thinking straight and i figured it out eventually. but. do i use straight jffs2 images for the sl-5600 or do i use cramfs? Aug 10 09:39:59 xcasex: you want jffs2 Aug 10 09:40:07 xcasex: and you flash them with zaurus-updater Aug 10 09:40:15 ah yes :D Aug 10 09:40:23 xcasex: installing is described in openzaurus wiki: http://wiki.openzaurus.org/ Aug 10 09:40:37 *looks* Aug 10 09:40:49 tyvm hrw i realize this hw is way behind you guys Aug 10 09:42:06 hrw: xcasex: unfortunately we are experiencing some issues with jffs2 detection/mount in kexecboot ... mostly when coming from original Sharp rom...:p Aug 10 09:42:51 ant_work: luckily i'm still using the old bootloader Aug 10 09:43:42 xcasex: I do not care about zaurus devices anymore Aug 10 09:44:10 hrw: i realise, afaik you were the OZ guy right? Aug 10 09:44:26 yes Aug 10 09:44:27 hrw: i went from gumstix to openmoko and ran away screaming to this. Aug 10 09:44:29 bb in few Aug 10 09:44:37 np Aug 10 09:53:46 * XorA still has a pile of OE zaurus kit that needs to go somewhere where it can be loved Aug 10 09:54:09 xcasex: poodle is a rare beast nowadays, let us know how/whether it works Aug 10 09:56:47 XorA: that collie which won't boot...is stalling on jffs2 detection/mount? Aug 10 09:57:08 ant_work: no, I keep telling you, the SD driver doesnt work Aug 10 09:57:12 you see..I suppose there are issues with "flash_eraseall -j" Aug 10 09:57:55 XorA: how did tesing boot from sd? Aug 10 09:58:15 ant_work: no idea, but 2 collies, 4 SD card, no boot Aug 10 09:58:27 more than enough test! Aug 10 09:58:38 I did send the log, but I thinks he is a busy busy man# Aug 10 09:58:55 ant_work: rare in which way? :D Aug 10 09:59:03 XorA: job chase iirc Aug 10 09:59:27 anyway changes nothing, I shall not be doing any work on Zaurus since I have new shinies thanks to TI Aug 10 10:01:05 xcasex: not so many users asking about it Aug 10 10:01:15 neither here nor in OESF Aug 10 10:11:30 ant_work: aha Aug 10 10:11:42 ant_work: weelll from the looks of it i need to redo the altboot image i built Aug 10 10:11:59 do yourself a favor and forget altboot Aug 10 10:12:30 is at least deprecated Aug 10 10:13:00 :( Aug 10 10:15:02 hi ant_work Aug 10 10:15:38 pb__: hello, back to work? Aug 10 10:15:42 yup Aug 10 10:15:58 :{ Aug 10 10:16:19 how's 24/7 with baby? Aug 10 10:16:26 :P Aug 10 10:16:53 heh, baby is doing fairly well. she can sleep for up to nine hours at a stretch now. Aug 10 10:17:04 men do sometime appreciate office work... Aug 10 10:17:18 :-} Aug 10 10:17:19 pb__: feel lucky :) Aug 10 10:17:52 last night wasn't so good, I think she woke up twice, but normally she will sleep between about 19:00 and 05:00 fairly reliably. Aug 10 10:18:10 ok, just a minor tz drift... Aug 10 10:18:41 florian: right :-) I gather that some babies are far worse than this. Aug 10 10:20:17 re Aug 10 10:21:21 pb__: indeed... Aug 10 10:21:29 hrw: wb Aug 10 10:22:39 xcasex: poodle was not popular due to many reasons. first batches were using pxa250 so were slow, 32mb ram was step back compared to collie. same factor/screen as collie was also not nice thing especially when compared to c7x0 which were released after poodle Aug 10 10:23:13 hrw: to my defense, i paid 56$ for the unit :] Aug 10 10:25:10 too much ;D Aug 10 10:25:49 I paid 3/4 of my month salary to get collie (few years ago). then I did not bought zaurus Aug 10 10:25:54 since then Aug 10 10:27:18 * XorA dances the take the damn zaurus off my hands dance :-) Aug 10 10:27:45 ;D Aug 10 10:27:54 XorA: which one you have? Aug 10 10:28:30 hrw: tosa and collie that belong to OE Aug 10 10:28:41 hrw: and one collie which is mine which I shall keep for memories Aug 10 10:30:15 the tosa is really hard to find :( Aug 10 10:30:45 tosa has one nice thing Aug 10 10:30:47 screen Aug 10 10:31:30 the best of zaurus line Aug 10 10:32:49 :D Aug 10 10:42:17 re Aug 10 10:48:29 r Aug 10 10:48:31 er Aug 10 11:01:29 is anyone going to HAR this week? (har2009.org) Aug 10 11:02:03 koobe: some, not me though Aug 10 11:04:36 03Koen Kooi  07org.openembedded.dev * ra834b1cdfc 10openembedded.git/recipes/linux/linux-davinci/dm365-evm/defconfig: linux-davinci: add defconfig for dm365-evm Aug 10 11:09:39 I will also not be there Aug 10 11:25:46 koobe, I looked into going at the last minute Aug 10 11:25:56 har tickets would cost more than airfare from .us :) Aug 10 11:26:15 well, it would be close Aug 10 11:27:30 crofton it is sold out Aug 10 11:27:32 since some weeks Aug 10 11:27:57 yeah Aug 10 11:28:14 there is a gnuradio tent Aug 10 11:28:23 next year I will pay more attention Aug 10 11:28:41 crofton hm its every 4 years Aug 10 11:28:43 do you know how much tickets are when they go on sale? Aug 10 11:28:44 or so Aug 10 11:28:45 ah Aug 10 11:47:43 Is there a temporary way of telling bitbake to ignore all check sums? Aug 10 11:47:55 jic23: yes Aug 10 11:48:38 zecke: Excellent. How do you do it? Aug 10 11:49:11 jic23: check base.bbclass Aug 10 11:50:03 jic23: OE_ALLOW_INSECURE_DOWNLOADS Aug 10 11:52:25 zecke: Thanks. Aug 10 12:00:42 * RP yawns Aug 10 12:00:45 morning all Aug 10 12:01:21 hey Aug 10 12:02:03 morning Aug 10 12:02:21 RP: yawning at 13? Aug 10 12:02:27 hi rp Aug 10 12:02:34 hrw: I'm a bit tired ;-) Aug 10 12:02:48 rp vacations? Aug 10 12:03:10 Two days going around in circles on the road bike, then two days riding up muddy footpaths and through rivers :) Aug 10 12:03:44 with a stop at the seaside for ice cream, obviously :) Aug 10 12:13:24 03Koen Kooi  07org.openembedded.dev * r28ca6b31e0 10openembedded.git/conf/checksums.ini: checksums: add gvfs checksum Aug 10 12:25:35 03Koen Kooi  07org.openembedded.dev * re5389c3af5 10openembedded.git/conf/distro/include/angstrom-2008-preferred-versions.inc: angstrom: move to glib 2.21.4 Aug 10 12:57:58 ~curse perl in OE Aug 10 12:57:59 May the fleas of a thousand camels infest your most sensitive regions, perl in OE ! Aug 10 12:59:21 ~curse libtool-native-2.2.6a in OE Aug 10 12:59:22 May the fleas of a thousand camels infest your most sensitive regions, libtool-native-2.2.6a in OE ! Aug 10 13:00:56 lol Aug 10 13:01:03 libtool works fine Aug 10 13:01:05 for me Aug 10 13:01:14 methril|work: what is the problem? Aug 10 13:05:08 today it's working, after deleting the entire dev repo Aug 10 13:05:16 hm Aug 10 13:05:18 but yesterday it was complaining Aug 10 13:05:26 maybe some left overs from older versions Aug 10 13:05:42 maybe, i'm not complaining today :) Aug 10 13:05:51 but i see a curse and i like to put one hehehe Aug 10 13:05:57 lol Aug 10 13:07:44 hm...ibot has curses and larts, but no shoulder-pats/applauds ? Aug 10 13:09:05 well, it's a curse with a delay, but yesterday i was cursing libtool Aug 10 13:45:22 ~summon someone for Perl maintaining Aug 10 13:45:23 apt takes out 20 clean, identical-looking phones, some extra hands, and pretends to be a telemarketer for a large corporation, so he gets delivered a phonelist containing someone for Perl maintaining's coordinates Aug 10 13:45:40 we are able to build Perl. but we are not able to build anything which uses perl... Aug 10 13:45:56 hrw hm? Aug 10 13:46:06 whats going wron Aug 10 13:46:07 g Aug 10 13:47:15 woglinde: 1. we do not stage perl library Aug 10 13:47:27 oh Aug 10 13:47:31 why this? Aug 10 13:47:33 woglinde: 2. hostcc is used when perl stuff is built Aug 10 13:47:43 woglinde: wrong do_stage Aug 10 13:47:54 hm wee need a perl.bbclass? Aug 10 13:48:08 but even with staged library we do not have a way to build irssi/perl Aug 10 13:48:20 woglinde: it is just few lines so no bbclass for it Aug 10 13:49:01 we use perl from tmp/staging/HOST/ and get paths for host not target etc Aug 10 13:49:12 and I want irssi with perl on sheevaplug... Aug 10 13:49:42 hrw *g* the old think of personal itch Aug 10 13:49:46 thing Aug 10 13:49:48 args Aug 10 13:50:23 woglinde: yep, but it looks like I will gave up and will run irssi on debian systems instead Aug 10 13:50:43 hrw to mich work? Aug 10 13:50:52 looks like Aug 10 13:50:59 and I never was a fan of perl Aug 10 13:52:53 * xcasex pokes hrw in his zaurial sore spot Aug 10 13:53:33 zaurial? Aug 10 13:54:17 how to gain 0.5GB of space - "strip tmp/staging/HOST/usr/bin/*" Aug 10 14:07:23 woglinde: we use hostperl to ask for cflags and ldflags. it is normal in perl world. but fails on cross... Aug 10 14:07:27 fucking perl Aug 10 14:08:36 autotools is using hostperl anyway Aug 10 14:19:16 what's the key combination again to switch between GUI and console in a qemu session, again? Aug 10 14:19:40 alt-ctrl-1/2/3/4 Aug 10 14:29:49 hrw: thanks Aug 10 14:30:16 np Aug 10 14:30:17 Those things are interesting (and I've added that info to the OE wiki now) Aug 10 14:30:50 But I guess I was really looking for the key combo that brings me the guest OS console in an opie image Aug 10 14:30:53 morning Aug 10 14:31:06 Let me see if google gets me some info quickly Aug 10 14:33:01 http://lists.e2factory.org/pipermail/e2factory-users/2009-August/000010.html - how cool e2factory is Aug 10 14:34:10 other thing... Aug 10 14:34:27 which vars in configure.in contain host_sys and target_sys? Aug 10 14:35:01 http://wiki.openzaurus.org/ReleaseNotes/Oz3542 says Fn+Arrow Aug 10 14:35:18 what does that mean when running opie in qemu? Aug 10 14:35:28 hrw lol Aug 10 14:35:32 I'm pretty sure I've done it in the past, but can't remember how Aug 10 14:35:37 wyh they invented another crosstool Aug 10 14:35:51 Laibsch: no idea Aug 10 14:36:00 woglinde: if not less Aug 10 14:36:05 they could also pointed to the codesourcy-sdk Aug 10 14:37:38 Laibsch: sendkey +Left? Aug 10 14:37:55 that may work Aug 10 14:37:55 I can remember even Fn key code on akita Aug 10 14:38:20 you can tag the keynames Aug 10 14:38:23 no, can't :) Aug 10 14:38:23 aeh tab Aug 10 14:38:24 likely candidates? Aug 10 14:38:58 yes, can ;-) Aug 10 14:38:58 Laibsch: about angstrom and qemu - http://linuxtogo.org/gowiki/AngstromAndQemu/Akita Aug 10 14:39:14 Laibsch: I can't remember :) Aug 10 14:39:19 I gave up Aug 10 14:39:27 may be look into xmodmap on real device Aug 10 14:39:33 no more touching anything perl related Aug 10 14:41:40 Laibsch: 69 :) Aug 10 14:41:43 xev says Aug 10 14:41:50 may be 0x69 Aug 10 14:43:21 cool number ;-) Aug 10 14:44:15 http://picpaste.com/Screenshot_1.png is the tab completion for sendkey Aug 10 14:44:27 oh..about keyboard mismap on latest gpe on zaurus, well...try vi ...modifier was Cancel +i or Cancel+x Aug 10 14:44:30 was... Aug 10 14:45:03 0x45 is niceer Aug 10 14:46:42 yeah i am just not making much way here Aug 10 14:47:00 the 16mb opie image errors out @ the opie-taskbar-image Aug 10 14:47:08 none of the gpe targets launch the gui Aug 10 14:47:09 etc Aug 10 14:47:37 xcasex: most tested images are: console-image opie-image x11-image Aug 10 14:47:50 x11-image is gpe Aug 10 14:47:53 ant_work: did you test opie recently? Aug 10 14:47:56 yes Aug 10 14:48:01 btw, can anyone try to build x11-image for tosa? Aug 10 14:48:02 2 days ago Aug 10 14:48:06 no hang at calibration screen? Aug 10 14:48:15 no Aug 10 14:48:31 well, the qemuarm image I built does just that Aug 10 14:48:32 iirc opie is pre-calibrated isn't? Aug 10 14:48:37 it should be Aug 10 14:48:42 but this one isn't Aug 10 14:48:46 for whatever reason Aug 10 14:48:54 Jay7: I'll build one Aug 10 14:48:57 btw gpe is miscalibrated in all sl-c860 I've seen Aug 10 14:48:59 Laibsch: mouse axes can be rotated Aug 10 14:49:05 you just want to know if it builds or you want the image? Aug 10 14:49:11 Laibsch: I've error while building hal Aug 10 14:49:17 ok Aug 10 14:49:26 ant_work: aha Aug 10 14:49:37 ant_work: are there diff. kernel targets as well? Aug 10 14:50:07 no, same kernel Aug 10 14:50:51 iirc poodle has 2.6.24 Aug 10 14:51:17 btw, is collie now using vanilla kernel? Aug 10 14:51:30 hm. Aug 10 14:51:52 afaik tosa and collie both have mainstream support Aug 10 14:52:07 the kernel i've used so far (the kexec one as of the tips this morning) boots up no issues there. but the driver loading stops dead when loading hal Aug 10 14:53:01 Can I somehow get an opie image in qemu to just boot into console instead of opie? Aug 10 14:53:16 I tested a console image fine with qemu Aug 10 14:55:03 iirc no Aug 10 14:55:07 i can do that with my poodle Laibsch Aug 10 14:55:09 without a problem. Aug 10 14:55:11 every time! Aug 10 14:56:41 Well, I want to save myself from flashing thousands of times Aug 10 14:56:50 I want to test stuff with qemu right now Aug 10 14:57:03 Jay7: I think the metadata is currently broken Aug 10 14:57:27 Laibsch: imho there should be any way to boot w/o gui Aug 10 14:57:29 There seems to be some kind of Angstromism called SPLASH which is only defined for Angstrom Aug 10 14:57:33 may be some kernel parameter Aug 10 14:57:56 at least that is what it loooks like to me after a short inspection Aug 10 15:00:07 Laibsch: btw, may be just change runlevel? Aug 10 15:00:21 how? Aug 10 15:00:29 I have no access from keyboard or mouse Aug 10 15:00:37 Laibsch: in console-image there is psplash-zzap to get you back in console Aug 10 15:01:25 * Laibsch is trying to find a proper fix for this SPLASH thing Aug 10 15:04:03 so i'm supposed to be able to flash the sl-5600 with a 32mb image. when it itself complains of it being larger than 16mb *sigh* Aug 10 15:06:01 Laibsch: you can set the runlevel from the kernel commandline. might need to experiment to see if that has any useful effect though, it depends how opie is started. Aug 10 15:06:31 pb__: do you know any details? Aug 10 15:06:39 regarding runlevel? Aug 10 15:07:06 Laibsch: Just provide it as an argument to the kernel, init should inherit the argument and parse it. Aug 10 15:07:17 Laibsch: the default level is probably either 2 or 5. try changing it to maybe 3 or 4 and see what happens. Aug 10 15:07:47 yeah, I just found that all that should be needed is giving the number on the command line Aug 10 15:07:48 qemu --append "3" or suchlike Aug 10 15:07:50 thanks Aug 10 15:08:03 xcasex: no, the images are meant to be booted from SD nowaday (kexecboot) Aug 10 15:08:18 xcasex: you could keep a minimal distro in nand, however Aug 10 15:09:37 ant_work: aha. Aug 10 15:09:46 ant_work: any specific directory on said sd card? Aug 10 15:09:52 (to keep the images in?) Aug 10 15:10:24 decompress the **-image.tar.gz in thr root of the ext2/3 card Aug 10 15:10:41 one per partition if you have more Aug 10 15:10:46 ah. Aug 10 15:10:56 and format the card as ext* ? Aug 10 15:11:11 the .jffs2 images are for the nand Aug 10 15:11:16 yeah Aug 10 15:12:56 feel free to update http://linuxtogo.org/gowiki/ZaurusPoodle Aug 10 15:13:18 Can't seem to get qemu to start anything but runlevel 5 Aug 10 15:22:00 Laibsch: when you have 5 mins, look at zaurus-updater, pls. It is now 'universal' but there are still specific files for akita/poodle ? Aug 10 15:22:20 what should I look at/for? Aug 10 15:22:55 can you check the commit-date? I'm away from buildmachine Aug 10 15:23:05 tehre are too many updater.sh around Aug 10 15:23:19 now we have nandlogical and all the mess could be done from console Aug 10 15:23:53 I hardly ever touch that stuff Aug 10 15:24:02 I'm still not sure what to check Aug 10 15:24:16 I would delete the akita and poodle specific updater.sh files Aug 10 15:25:01 can you pls. check if these have been modified *after* the main updater.sh? Aug 10 15:25:21 gitweb doesn't help with old commits Aug 10 15:29:23 great. Aug 10 15:29:29 kexecboot works. but. Aug 10 15:29:36 it shows one entry and its the nand Aug 10 15:29:59 xcasex: do you have a kernel in /boot? Aug 10 15:30:15 How do I persuade bitbake to use a local get repostiory? Can't seem to figure out what the syntax would be. Aug 10 15:30:29 ant_work: yeah Aug 10 15:30:59 is anyone seeing anything breaking that uses python? Aug 10 15:31:06 git that is (not get) Aug 10 15:31:10 I ahd an email from someone trying to build gnuradio this AM Aug 10 15:31:16 ant_work: last change to akita updater.sh according to git log was in Nov 2007 Aug 10 15:31:19 I am working to confirm now Aug 10 15:31:20 Laibsch: the recipe 'zaurus-updater' is an interesting case of study about filepath: I suppose ${FILE_DIRNAME}/${PN} is parsed before ${FILE_DIRNAME}/${MACHINE} Aug 10 15:31:49 ant_work: podle updater.sh according to git log wasn't changed after 2005 Aug 10 15:31:56 if anyone knows of another poodle that is on the market i'll make that a dev machine foro e Aug 10 15:31:58 *oe Aug 10 15:32:04 last change to universal updater.sh was in 2008 Aug 10 15:32:16 ant_work: I hope that answers your question Aug 10 15:32:34 Laibsch: if my assumption about filespath is correct the files are cruft Aug 10 15:32:59 (the files in ${FILE_DIRNAME}/${MACHINE}) Aug 10 15:34:09 Crofton: I saw some non-fatal error messages a little while ago and an indication that python-native may not have been built. I aborted the build, baked python-native first and haven't seen that warning afterwards (but didn't pay much attention, either) Aug 10 15:55:55 xcasex: have you built the images or are you using prebuilt-ones? Aug 10 16:03:39 ant_work: i've built them using oe of course. Aug 10 16:04:28 xcasex: good Aug 10 16:04:47 silly question: is your SD card ext2/3/4 ? Aug 10 16:04:56 or plain fat? Aug 10 16:04:58 ext2 Aug 10 16:05:27 and if you boot form nand you can mount the SD? Aug 10 16:06:56 ant_work: yes. Aug 10 16:06:58 BUT! Aug 10 16:07:29 xcasex: hello mighty kergoth Aug 10 16:07:33 oops Aug 10 16:07:45 morning Aug 10 16:07:47 * kergoth sleepy Aug 10 16:07:50 xcasex: but ? Aug 10 16:07:52 the thing is. with the kernel i built it only boots as far as hal and then freezes the processes Aug 10 16:07:55 hello kergoth Aug 10 16:08:14 hail kergoth he who walks the circuits of trondom! ;) Aug 10 16:08:14 the staging revamping has _really_ tired you out it seems :-) Aug 10 16:08:57 naw, its just Monday Aug 10 16:09:02 tends to do that :) Aug 10 16:10:35 xcasex: so it boots, it finds the rootfs, processes the initscripts and get stuck somewhere? Aug 10 16:10:55 ant_work: right around the hal detection, problem found eh. Aug 10 16:11:06 this is gentoo/lfs all over again Aug 10 16:31:05 * kergoth ponders Aug 10 16:31:54 * kergoth thinks.. what if the mangling of staging files was actually a patch file instead of shell scripts.. hmm Aug 10 16:44:02 does anybody have a rough idea of how many GB a complete OE source mirror would be? Aug 10 16:44:16 think "bitbake -c fetchall world" Aug 10 16:46:57 Anyone have any suggestions for what might cause the following? Aug 10 16:47:02 Starting udevudevd-event[1575]: symlink(../ttyv3, /dev/char/3:99.udev-tmp) failed: No space left on device Aug 10 16:47:12 and there is plenty of room. Aug 10 16:47:51 ext2/3? are you out of inodes? Aug 10 16:47:55 or rather, ignore me Aug 10 16:47:58 /dev is a ram filesystem Aug 10 16:48:02 you ran out of memory, most likely Aug 10 16:48:13 * kergoth shrugs Aug 10 16:48:55 kergoth: jffs2 and seems unlikely ran out of ram as this happens on boot. Aug 10 16:49:13 kergoth: Other than a lot of error messages eveything seems to work fine. Aug 10 16:50:00 well, /dev is a tmpfs/ramfs, i don't see how else it would get that error there, unless I'm missing something Aug 10 16:50:29 hmm Aug 10 16:56:38 kergoth: Odd. it only does it for the first couple of boots. Ah well, that one may remain a mystery. Aug 10 17:00:09 kergoth: Thanks for suggestions. Wonder if there is any way of telling if it is a memory issue. hmm. Aug 10 17:00:15 good question :) Aug 10 17:00:31 add a cat /proc/meminfo or something in that startup script Aug 10 17:05:13 kergoth: 25M free just before udev startup runs. I'm somewhat disturbed if it uses that much! Aug 10 17:06:48 why? just because ram is being used by caches or what have you doesn't mean it won't be freed and used for your apps if you need it. there's no point having it sit there idle and useless, instead of making use of it Aug 10 17:07:18 I'll rephrase. I'm somewat disturbed if it needs that much and fails otherwise. Aug 10 17:17:48 Have to reflash anyway as I forgot to include usbnet so will see if it was just a one off issue or not. Aug 10 17:36:11 * kergoth wonders if bitbake should get include_next and require_next Aug 10 17:37:37 kergoth: what do they do ? Aug 10 17:38:00 get the next one in the bbpath Aug 10 17:38:01 (the inlcude_next and require_next) Aug 10 17:38:15 so conf/foo.conf could include the next conf/foo.next Aug 10 17:38:18 err, foo.conf Aug 10 17:38:29 kergoth: I dunno, I've never been a very big fan of that in gcc. Aug 10 17:38:43 i suppose it could lead to confusion in some cases Aug 10 17:39:01 it's kind of cute for allowing an "objecty" kind of overriding behaviour, but I've always found that debugging the cases when it does the wrong thing is a bit of a nightmare. Aug 10 17:39:21 * kergoth nods, makes sense Aug 10 17:39:26 iirc, there is one particular package in oe that keeps going wrong for that reason, some file gets installed in an unexpected place and disaster ensues Aug 10 17:39:39 i got bit by something like that Aug 10 17:39:54 some recipe was including like.. ../linux/linux.inc or something Aug 10 17:40:02 and it was grabbing one from a different collection instead of its own :) Aug 10 17:40:06 something like that, anyway Aug 10 17:40:06 heh Aug 10 17:40:20 no, i remember now. it did 'include recipes/linux/linux.inc' Aug 10 17:40:30 i changed it to do the relative bit instead to stay in that collection Aug 10 17:42:23 Jay7: FWIW, tosa x11-image built fine here. Aug 10 17:42:42 Laibsch: against current .dev? Aug 10 17:43:10 yes Aug 10 17:43:14 but minimal distro Aug 10 17:43:22 ah.. Aug 10 17:43:25 I had to manually specify the splash to use Aug 10 17:43:27 but that is about it Aug 10 17:43:28 I'm building angstrom x11-image Aug 10 17:43:37 I know Aug 10 17:43:46 look for the differences and you may find the problem Aug 10 17:43:55 pb_: I'm starting to think using multiple recipe files and includes was the wrong way to handle multiple versions/revisions/whatever of a thing Aug 10 17:44:00 usually, minimal should be more likely to run into problems Aug 10 17:46:50 Laibsch: http://tinderbox.openembedded.net/builds/22716/ Aug 10 17:46:55 that was my latest build Aug 10 17:47:15 probe-storage.c:47:19: error: blkid.h: No such file or directory Aug 10 17:48:59 kergoth: yes, I've been thinking that for a while now. something that's been on my list of projects for a few months is a way to implement a kind of "template" .bb file where you can have some variables (typically ${PV}) left unbound. Aug 10 17:49:05 Jay7: sounds like a missing build-time dependency, no? Aug 10 17:49:14 yeah.. Aug 10 17:49:31 but it is not -native package Aug 10 17:49:41 so it should be something in staging Aug 10 17:49:42 so, you'd just have a file named (literally) "foo_*.bb" and it would substitute ${PV} at build time. Aug 10 17:50:05 in the (relatively uncommon) case where different versions actually do need different recipes, you could either do the old thing or use a runtime conditional in the recipe to take the appropriate action Aug 10 17:50:22 btw, blkid.h is present in staging/i686-linux/usr/include/blkid/blkid.h Aug 10 17:51:32 turns out that this is moderately painful to implement in bitbake though, because the idea that ${PF} maps 1:1 to a PN-PV-PR tuple is quite deeply ingrained in its data structures. if you suddenly have a single PF which can produce multiple PVs with late binding then lots of the hash tables become a nuisance. Aug 10 17:54:08 not quite sure how to deal with that yet. I've been tempted to throw bitbake away and write a new tool, but that seems a bit of an excessive way to solve this problem. also considered trying to hire zecke or some such überhacker to do the work if I can find some budget for it. Aug 10 17:57:05 well, the runqueue and all uses the filename and fnid as its a main identifier, I'd expect we could alter the rest to do so as well Aug 10 17:57:18 hmm Aug 10 17:57:30 yeah, it's the filename that's the problem. Aug 10 17:57:39 I've wanted to replace bitbake forever, but rewrites never seem to go well Aug 10 17:57:48 ahh, right Aug 10 17:58:05 if you did "bitbake gcc-3.3 gcc-4.0 gcc-4.4", which isn't a completely unreasonable thing to do if you want to build a bunch of compilers, suddenly you need something better than the filename to disambiguate them. Aug 10 17:58:13 it needs to be the filename plus a bunch of variable bindings Aug 10 17:59:01 richard hacked in multiple "recipe" production from one for his virtualnative / bbclassextend stuff, but it's just that, a hack Aug 10 18:05:32 I'm thinking most of what we want to accomplish we can still get done via refactoring, deprecation of old ways of doing things, etc. I'd still kind of prefer to start from scratch, but that splits community effort and all Aug 10 18:07:19 one of the things that really, really bugs me, is how much the recipes can end up inadvertently depending upon how things work under the hood, both classes and bitbake. it makes it difficult to do such a transition. I'd like to get them stripped down to descriptive metadata & build steps for config, compile, and install, if at all possible Aug 10 18:07:22 kill the new tasks, etc Aug 10 18:07:45 the whole point of using a directly parsed file format instead of shell like portage was so we could switch formats if we chose to, but we're kind of locked in just now Aug 10 18:12:30 yeah Aug 10 18:13:57 of course, the trouble with a purely descriptive recipe format is that it's somewhat limiting, or at least that it forces you to put any behavioural changes in some other place (be that a special type of class, or bitbake itself, or...) Aug 10 18:14:13 not that that's necessarily a bad thing, but it'd be quite a difference compared to how things work now Aug 10 18:18:11 * kergoth nods Aug 10 18:18:59 I've been leaning toward completely declarative recipes, and classes being written directly in a programming language, like python, not metadata, though they could inject such if they chose to. just a shift in where the power resides Aug 10 18:19:04 * kergoth shrugs, lots of possibilities Aug 10 18:21:57 of course, we've been doing a good job builidng a community around the current structure Aug 10 18:22:18 so radical changes might not be widely adopted for a while Aug 10 18:24:25 that's true, though (luckily) most of the community contributions are things that would work fairly well in a declarative universe too. Aug 10 18:24:44 most simple recipes would be portable with little/no effort, and auxiliary files like patches would obviously be reusable too. Aug 10 18:26:43 the people who would be at most risk of losing would be the ones who have sunk a lot of effort into complicated recipes with either lots of programmatic sections, or lots of assumptions about the build system, and those tend to be more of a "power user"/core hacker thing anyway rather than something that gets contributed by relatively casual users. Aug 10 18:27:58 what do you mean by programmatic sections? Aug 10 18:28:11 embedded python in recipes is the most obvious example Aug 10 18:28:17 heh Aug 10 18:28:27 python is a dirty word Aug 10 18:28:32 i.e. anything that is imperative rather than declarative Aug 10 18:28:33 well Aug 10 18:29:22 I got an email that suggests the python 2.6.2 "upgrade" means you need to wipe tmp ... Aug 10 18:29:34 doh, what? Aug 10 18:30:01 not 100% certain Aug 10 18:30:06 clean build worked for me Aug 10 18:30:16 but someone building gnuradio had an issue Aug 10 18:30:26 gr uses python and swig Aug 10 18:33:27 kergoth: I'm not sure that a whole-hearted crusade against extra tasks is necessarily a good thing. So long as they're straightforward and don't poke at build system internals (which probably means written in shell) there doesn't seem any harm in allowing recipes to declare their own subtasks if they want to. Aug 10 18:34:14 we need one for gnurradio so we can build a private copy of libusb-0.12 Aug 10 18:34:15 kergoth: one example is the "do_munge_sources" thing that lots of recipes end up needing between unpack and configure, generally just a set of mv/cp/sed commands. can't really see any need to abolish that kind of thing. Aug 10 18:34:37 anybody else doing testing with qemu images? I start to get the feeling that qemu does not support the host's mouse. Could very well be the qemu call is wrong. But I use poky-qemu so that I would not have to worry about that. Aug 10 18:34:56 Is the mouse working for anybody that does GUI image booting in qemu? Aug 10 18:35:02 Jay7? Aug 10 18:35:14 Crofton|work: I'm not so sure about that, it seems like there are probably (in a utopian future with a well-managed staging) better ways of solving that problem. Aug 10 18:35:24 yeah, i suppose, though in that case I'd think itd be better to add the task in base and let users define that method if they want to use it, rather than the recipe adding the task itself, and having to know about what tasks run in what order and all Aug 10 18:35:26 but i udnno Aug 10 18:36:09 I guess. There isn't any real reason that you couldn't bolt those bits onto the front of do_configure, it's just easier not to. Aug 10 18:36:12 Crofton|work: the staging work I'm doing would let that recipe depend on libusb-0.12 and itd be used, but other recipes would get the other versions. the builds would work. the only question is how to handle packaging Aug 10 18:36:20 Laibsch: angstrom x11-image for akita is working Aug 10 18:36:42 you mean you could make libusb-0.12 and libusb-compat coexist? Aug 10 18:36:50 Laibsch: I think the mouse is meant to work in qemu, at least for x86. what target are you using? Aug 10 18:36:59 Crofton: yes, that would be the idea Aug 10 18:36:59 the filenames conflict Aug 10 18:37:26 right, that's why you need better management for staging (which really means per-package staging) Aug 10 18:37:42 pb_: qemuarm on x86 host Aug 10 18:38:00 you suggest x86 qemu image? Aug 10 18:38:12 Laibsch: ah, I'm not sure if qemu knows how to emulate any arm touchscreen hardware. might need to check the documentation. Aug 10 18:38:16 Laibsch: I'm using qemu-system-arm on x86 for testing angstrom x11-image for akita Aug 10 18:38:22 mouse is working ok Aug 10 18:38:32 ok, thanks Aug 10 18:38:40 qemu-system-arm -M akita .... Aug 10 18:38:50 Crofton: http://gist.github.com/164598 is the prototype I threw together this weekend. don't know if itll be worth keeping, its a proof of concept Aug 10 18:38:59 Jay7: let's see if it's a problem with poky-qemu Aug 10 18:39:01 use -show-cursor btw Aug 10 18:39:14 just to see mouse cursor :) Aug 10 18:39:40 sudo /usr/bin/qemu-system-arm -kernel zImage-qemuarm.bin -net nic,vlan=0 -net tap,vlan=0,ifname=tap0,script=/usr/share/poky-scripts/poky-qemu-ifup,downscript=/usr/share/poky-scripts/poky-qemu-ifdown -M versatilepb -hda x11-image-qemuarm.ext2 -usb -usbdevice wacom-tablet -no-reboot --append root=/dev/sda ro 2 console=ttyAMA0 console=tty0 ip=192.168.7.2::192.168.7.1:255.255.255.0 mem=64M Aug 10 18:39:48 Jay7: that is what poky-qemu calls Aug 10 18:40:02 -M versatilepb Aug 10 18:40:05 yes Aug 10 18:40:08 I was wondering about that Aug 10 18:40:11 what is it? Aug 10 18:40:11 did you building image for such board? Aug 10 18:40:19 not knowingly ;-) Aug 10 18:40:25 qemu-system-arm -M ? Aug 10 18:40:32 issue this and look :) Aug 10 18:41:11 qemu-system-arm -M akita -portrait -show-cursor -kernel zImage-kexecboot-akita.bin -mtdblock mtdblock-clean.img -serial file:log -append "console=tty1 console=ttyS0,115200n8 fbcon=rotate:1 logo.nologo debug" -sd sdcard.img Aug 10 18:41:22 my command line for kexecboot testing Aug 10 18:41:48 what machine do you build those images for? Aug 10 18:42:00 MACHINE="akita" Aug 10 18:42:04 OK Aug 10 18:42:11 I build for MACHINE=qemuarm Aug 10 18:42:47 versatilepb ARM Versatile/PB (ARM926EJ-S) Aug 10 18:43:35 hm.. not sure about this machine.. may be you should use pseudo-usb mouse in qemu Aug 10 18:44:00 Laibsch: you can use tosa too :) Aug 10 18:44:07 just use -M tosa Aug 10 18:44:13 but I've not tested this yet Aug 10 18:44:27 afk for some time Aug 10 18:46:31 Jay7: the images for real hardware don't build ext2 images by default :-( Aug 10 18:46:47 I'll see if I can't figure out the proper call for the qemuarm stuff Aug 10 18:57:49 Laibsch: you can override IMAGE_FSTYPES in local.conf if you want Aug 10 18:58:57 but, looking at your qemu invocation, it looks like it's set up to emulate a wacom usb tablet for the mouse. Aug 10 18:59:31 in that case, the most likely problem is that the xserver just isn't looking for the right kind of mouse device. might be that xserver-common doesn't understand qemuarm. Aug 10 19:00:17 or you might be missing some module, but I would have expected the versatilepb kernel to come with all the right bits built in Aug 10 19:02:37 03Klaus Kurzmann  07shr/import * rdd6c16cf50 10openembedded.git/recipes/tasks/task-shr-feed.bb: Aug 10 19:02:37 task-shr-feed: add python-xlib Aug 10 19:02:37 Signed-off-by: Klaus Kurzmann Aug 10 19:02:48 03Klaus Kurzmann  07shr/import * r4639bfb28d 10openembedded.git/recipes/libsdl/ (files/configure_tweak.patch libsdl-x11_1.2.11.bb): Aug 10 19:02:48 libsdl-x11: tweak configure_tweak.patch to include support for alsa Aug 10 19:02:48 Signed-off-by: Klaus Kurzmann Aug 10 19:35:21 hi again. I cancelled a build. Now when I restart "bitbake base-image", it says "ERROR: function do_stage failed" and aborts. I assume some invalid remnants. How can I continue? Aug 10 19:36:53 try bitbake -c rebuild Aug 10 19:36:57 which package is it Aug 10 19:37:13 qemu-native Aug 10 19:37:50 rebuild task I am not so sure but something like this bitbake -c clean qemu-native;bitbake base-image Aug 10 19:37:53 would work Aug 10 19:45:01 03Khem Raj  07org.openembedded.dev * rb4a6504890 10openembedded.git/conf/bitbake.conf: Aug 10 19:45:01 bitbake.conf: Initialize empty PACKAGE_EXTRA_ARCHS conditionally. Aug 10 19:45:01 Signed-off-by: Khem Raj Aug 10 19:45:01 Acked-by: Marcin Juszkiewicz Aug 10 19:56:49 meh Aug 10 20:04:31 *sigh* Aug 10 20:39:16 03Klaus Kurzmann  07shr/import * rcfb74c5621 10openembedded.git/conf/distro/include/shr-autorev-unstable.inc: Aug 10 20:39:16 shr-autorev-unstable: remove libgee as we use version tarballs now Aug 10 20:39:16 Signed-off-by: Klaus Kurzmann Aug 10 20:39:27 03Klaus Kurzmann  07shr/import * rff8603027e 10openembedded.git/recipes/tasks/task-fso-compliance.bb: Aug 10 20:39:27 task-fso-compliance: deactivate some connman plugins as they were creating problems Aug 10 20:39:27 Signed-off-by: Klaus Kurzmann Aug 10 20:39:27 03Klaus Kurzmann  07shr/import * r8f0d03c928 10openembedded.git/ (2 files in 2 dirs): Aug 10 20:39:28 shr-autorev.inc: update revs for testing image Aug 10 20:39:30 Signed-off-by: Klaus Kurzmann Aug 10 20:49:42 jo florian Aug 10 20:49:57 re **** ENDING LOGGING AT Mon Aug 10 21:05:15 2009 **** BEGIN LOGGING AT Mon Aug 10 23:46:18 2009 Aug 10 23:51:37 'nite Aug 10 23:53:13 morning :) Aug 10 23:53:32 ah global development Aug 10 23:54:05 keeps things interesting doesn't it Aug 10 23:54:45 so.... how about this weather we've been having? Aug 11 00:01:50 grg: you mean global warming, or some specific locale? trying to decide if i need to pass your comments through an i18n translator. Aug 11 00:02:25 i just assumed that you were having some weather too. Aug 11 00:02:35 and there's probably a lot of it Aug 11 00:03:27 i'm out in the SF Bay Area. feels like we're getting our last summer heat wave, about 80 degrees today Aug 11 00:04:32 what's yours like? Aug 11 00:05:44 Just a boring winter's day. No rain, not really that cold, but overcast. Aug 11 00:06:02 I'm in south australia Aug 11 00:06:10 Adelaide to be precise Aug 11 00:07:48 ah, i have good friends who live in adelaide. they have spoken very highly of its beauty. Aug 11 00:09:00 meh. its ok Aug 11 00:09:16 lol totally opposite reaction, alrighty :-) Aug 11 00:09:24 what brings you to OE? Aug 11 00:09:29 work Aug 11 00:09:47 i have a mipsel touchscreen device to hack on Aug 11 00:10:12 fun in a RISC-y sort of way Aug 11 00:10:22 yup Aug 11 00:10:33 not a very powerful unit, but fun Aug 11 00:10:50 how about you? Aug 11 00:11:04 same, but different platform. Aug 11 00:11:15 we're developing a board based on the Atmel AT91SAM9263 processor. Aug 11 00:13:24 it's my first experience with OE, but it's proving to be a very nice little system. Aug 11 00:13:36 needs help with documentation, but that's a common theme across FOSS Aug 11 00:14:10 unfortunately most people in a position to document stuff generally have more of a desire to code Aug 11 00:15:30 this chatroom has proven to be an invaluable resource for OE. having direct access to the principals helps tremendously. Aug 11 00:16:08 they probably get tired of it :-) but i'd gladly buy any of them a beer if they're ever in San Francisco. Aug 11 00:16:54 Hi! I try to build my own toolchain with 'bitbake meta-toolchain', but I get a file not found from this line: target_pkgs=`cat ${SDK_OUTPUT}/${prefix}/package-status | grep Package: | cut -f 2 -d ' '` Aug 11 00:17:03 which is from meta-toolchain.bb Aug 11 00:17:25 it seems that package-status is not created before Aug 11 00:17:48 but I have no idea what program should create it Aug 11 00:18:30 Jeff__: no clue, but have you checked the bugtracker (http://bugs.openembedded.org) for a related issue? Aug 11 00:19:44 yes, but I check again Aug 11 00:21:52 Jeff__: how recent is your GIT repository? Aug 11 00:31:13 kg4ysn|coding: I ran a git pull about 4 hours ago Aug 11 00:31:46 stupid question, but does your system have 'cat', 'grep', and 'cut' in its PATH? Aug 11 00:32:39 yes :-) Aug 11 00:32:42 Jeff__: also, do you know what the values of SDK_OUTPUT and prefix are / should be? Aug 11 00:33:28 I just checked the file manually and it is really not there Aug 11 00:33:31 the meta-toolchain.bb has these Aug 11 00:33:50 package-status sounds like something opkg related Aug 11 00:34:09 are there recommended values for this parameters? Aug 11 00:34:50 it looks like it should be ${WORKDIR}/sdk/image Aug 11 00:35:50 is there another file in there? perhaps called status? or packages? Aug 11 00:36:12 or is package-status somewhere else in ${WORKDIR}/sdk ? Aug 11 00:36:28 Jeff__: can you run the `cat ...` command on your system's command line, replacing ${SDK_OUTPUT} with ${WORKDIR}/sdk/image and ${prefix} with ...? i dunno, your toolchain prefix i assume? Aug 11 00:38:43 cat /..../package-status: No such file or directory Aug 11 00:39:20 grg: i ran a full find in my workdir, bur there is no package-status Aug 11 00:39:21 Jeff__: try doing an 'ls' of the directory above package-status to see if it exists Aug 11 00:39:32 kg4ysn|coding: same Aug 11 00:39:43 In my ipk directories I have a 'Packages' file. Aug 11 00:39:59 that would be the same format as what is expected of package-status Aug 11 00:40:13 (but I'm not building an sdk, and don't have that dir) Aug 11 00:40:54 Jeff__: what command are you giving to bitbake? specifically Aug 11 00:41:21 Jeff__: also, what's the target? Aug 11 00:41:35 bitbake meta-toolchain Aug 11 00:42:33 Jeff__: and what's the intended target? Aug 11 00:45:12 kg4ysn|coding: what do you mean by 'intended target' Aug 11 00:45:14 ? Aug 11 00:45:29 Jeff__: you're building a toolchain for what platform? Aug 11 00:46:45 Jeff__: what's in your local.conf? paste the contents here: http://paste.lisp.org/new/oe Aug 11 00:49:59 Jeff__ pasted "untitled" at http://paste.lisp.org/display/85169 Aug 11 00:50:54 and I'm building an i386-toolchain :-) Aug 11 00:51:26 Jeff__: i just kicked off a build of meta-toolchain on my system ... let's see if it errors. Aug 11 00:51:51 Jeff__, have you tried a non-parallel build? Aug 11 00:52:36 grg: good catch Aug 11 00:53:09 no I'll try Aug 11 00:53:13 and report back Aug 11 01:04:29 kg4ysn|coding: grg: there is no difference, but I have also tried to find any hint which command should create this file, but there was just no trace of it. also not in the opkg-source Aug 11 01:05:47 Jeff__, I suspect that if you find a file called 'Packages' that is in the correct place then you can use that file instead. Aug 11 01:06:18 the cat | grep | cut command should give a list of packages when run on that file Aug 11 01:06:56 who uses 'cat file | grep' anyway? whats wrong with 'grep file' Aug 11 01:07:43 or better yet, use awk. awk '/Packages:/ {print $2}' file Aug 11 01:07:47 much shorter Aug 11 01:08:15 who's the maintainer of the package? Aug 11 01:09:15 dunno. last commit came from Tom Rini Aug 11 01:10:58 might be worth an e-mail to him Aug 11 01:14:21 off to dinner. later mates! Aug 11 01:16:41 03Holger Hans Peter Freyther  07org.openembedded.dev * rd89b57a975 10openembedded.git/recipes/meta/meta-toolchain-qte.bb: Aug 11 01:16:41 meta-toolchain-qte: Fix the paths to the helper utilities Aug 11 01:16:41 They are directly copied into bin/ and not usr/bin. Change Aug 11 01:16:41 the script to point to the correct location. Aug 11 01:16:52 03Holger Hans Peter Freyther  07org.openembedded.dev * r16e1b32d69 10openembedded.git/recipes/ (meta/meta-toolchain-qte.bb qt4/files/g++.conf): Aug 11 01:16:52 qt4: Make it possible to set the QMAKE_QT_CONFIG variable Aug 11 01:16:52 this way we will find the qconfig.pri and all library names Aug 11 01:16:52 will be set correctly. This means that the SDK is usable to Aug 11 01:16:54 compile Qt applications out of the box. Aug 11 01:35:43 03Marc Olzheim  07org.openembedded.dev * r51221fe1ad 10openembedded.git/recipes/curl/ (curl_7.19.5.bb files/curl-add_all_algorithms.patch): Aug 11 01:35:43 curl-7.19.5: Fix for curl artifact 2825989 Aug 11 01:35:43 Add curl-add_all_algorithms.patch to 7.19.5 recipe. Fixes curl artifact id 2825989 Aug 11 01:35:43 https://sourceforge.net/tracker/index.php?func=detail&aid=2825989&group_id=976&atid=100976 Aug 11 01:35:45 http://cool.haxx.se/cvs.cgi/curl/lib/ssluse.c#rev1.229 Aug 11 01:35:47 http://patchwork.openembedded.org/patch/928/ Aug 11 01:35:49 Signed-Off-By: Holger Hans Peter Freyther Aug 11 01:38:20 03Rafael Campos Las Heras  07org.openembedded.dev * r7a1bea0c40 10openembedded.git/ (6 files in 4 dirs): Aug 11 01:38:20 Update the em-x270 machine to work with linux_2.6.23. Aug 11 01:38:20 * Use of generic linux_2.6.23 recipe Aug 11 01:38:20 * remove em-x270_2.6.23 recipe (ugly). Aug 11 01:38:20 * Add a patch that allows to compile with gcc 4.1 compiler. Aug 11 01:38:21 Signed-off-by: Rafael Campos Las Heras Aug 11 01:38:23 Signed-off-By: Holger Hans Peter Freyther Aug 11 01:52:21 03Graham Gower  07org.openembedded.dev * r7a7e665303 10openembedded.git/recipes/libmad/ (files/mad-mips-h-constraint.patch libmad_0.15.1b.bb): Aug 11 01:52:21 libmad_0.15.1b: fix libmad compilation with GCC-4.4 on mips Aug 11 01:52:21 Fix build failure due to removal of the h asm constraint in Aug 11 01:52:21 GCC-4.4 mips port. Aug 11 01:52:22 http://patchwork.openembedded.org/patch/921/ Aug 11 01:52:24 Signed-Off-By: Holger Hans Peter Freyther Aug 11 01:52:31 03Graham Gower  07org.openembedded.dev * r0d3f61c828 10openembedded.git/recipes/libgcrypt/ (files/mips-h-constraint.patch libgcrypt.inc): (log message trimmed) Aug 11 01:52:31 libgcrypt.inc: Fix libgcrypt compilation with GCC-4.4 on mips Aug 11 01:52:33 Patch taken from https://forum.openwrt.org/viewtopic.php?id=20520 Aug 11 01:52:35 Makes sense and seems to work ok. Aug 11 01:52:37 libgcrypt.inc has --disable-asm in EXTRA_OECONF, yet strangely the configure Aug 11 01:52:39 script does nothing with this information beyond parsing it. Assembler is Aug 11 01:52:41 certainly still being used. Aug 11 02:46:01 03Manuel Pietschmann  07org.openembedded.dev * r68233bb284 10openembedded.git/ (conf/checksums.ini recipes/python/python-sip_4.8.2.bb): Aug 11 02:46:01 python-sip_4.8.2.bb: Add new recipe for python-sip Aug 11 02:46:01 Signed-off-by: Khem Raj Aug 11 02:46:12 03Marc Olzheim  07org.openembedded.dev * r19a190ca2b 10openembedded.git/recipes/ (4 files in 2 dirs): Aug 11 02:46:12 eglibc-package.bbclass glibc-package.bbclass: Use ls -d in do_prep_locale_tree Aug 11 02:46:12 ls -d will list directories as plain files and then cp -R will copy the Aug 11 02:46:12 contents over. Aug 11 02:46:12 Signed-off-by: Khem Raj Aug 11 02:46:14 03Khem Raj  07org.openembedded.dev * r08020ed71f 10openembedded.git/: Merge branch 'org.openembedded.dev' of git@git.openembedded.org:openembedded into org.openembedded.dev Aug 11 02:46:14 denix, apparently the outage was planned Aug 11 02:47:10 Crofton|work: hmm, I guess a little note on the maillist or irc would have been nice... Aug 11 02:57:34 * khem concurs **** ENDING LOGGING AT Tue Aug 11 02:59:57 2009