**** BEGIN LOGGING AT Sun Feb 27 02:59:57 2011 Feb 27 03:13:40 mwester, if you force it to go single threaded, does it work? Feb 27 03:14:21 Could try and detect 2.6.2 and fore single threaded Feb 27 03:15:01 (to me, if someone is running fedora core they've opted to be bleeding edge or pretty close, but anyhow...) Feb 27 03:17:51 How to force it to single thread? Feb 27 03:20:31 (no obvious command line args to bitbake seem to set that) Feb 27 03:25:47 Tartarus, 1.12 works for me on f13 Feb 27 03:26:12 ah mwester explains this Feb 27 03:37:26 Tartarus, setting BB_NUMBER_PARSE_THREADS = 1 results in the same problem, except that there is only a single defunct process. I'll try setting it to 0 just for fun. Feb 27 03:38:28 * hvontres|home should not have started TWO bitbake runs on a dual core laptop :( Feb 27 03:39:26 Setting it to 0 results in no defuct process -- but it's still not parsing. :) Feb 27 03:39:52 So either that's quite a wrong value to select, or perhaps now I can strace it. I'll be optimistic and choose the latter. Feb 27 03:43:00 hmmm, looks like RHEL 6 ships with python 2.6.5 - wonder if that one works Feb 27 03:44:33 Now that's really bleeding edge -- for RHEL, anyway. Feb 27 03:47:45 actually, (k)ubuntu 10.04 is also on 2.6.5, and it works Feb 27 04:50:24 anybody have any issues with autotools pacakges and -target=armv5teb-oe-linux-gnueabi vs -target=armv5teb-linux-gnueabi ? Feb 27 04:59:36 I seem to recall a package that failed do_configure because it was unable to match the new string. Feb 27 05:00:12 mwester: any idea how it got fixed? Feb 27 05:01:30 I proposed a simple patch to extend the patterns in the pertinent configure script, but khem figured out how to completely re-jigger the package to actually regenerate the configure script in the correct way. Feb 27 05:09:10 mwester: do you remember which package? I am working on a recipie for apcupsd Feb 27 05:09:43 No, can't remember the package - it was one that was in the SlugOS feeds, FWIW. Feb 27 05:12:05 mwester: did you post the patch to the ml? Feb 27 05:12:15 No. All done via IRC Feb 27 05:12:29 It was really quite trivial :) Feb 27 05:13:54 mwester: ok, thanks. Time to search the logs :) Feb 27 05:14:38 ~logs Feb 27 05:14:38 All conversations are logged to http://ibot.rikers.org/%23oe/ Lines starting with spaces are not logged. Logs are updated daily. Feb 27 05:30:41 mwester: do you happen to remember how long ago this got fixed? Feb 27 05:31:09 weeks, not more than 6 or 8 at most, probably less. Feb 27 05:33:14 mwester: thanks. could it have been xinetd? Feb 27 05:33:23 rings a bell Feb 27 06:09:10 n8 Feb 27 07:36:43 Hi, I am starting out on a first recipe. I am struggling to download source from github. My SRC_URI is: SRC_URI = "git://github.com/rabbitmq/rabbitmq-c.git;protocol=git" is that correct? Feb 27 08:13:15 gm Feb 27 09:01:52 THIS IS THE BEST U CAN GET http://www.1filesharing.com/download/0PF3RZH5/psyBNC2.3.1_6.rar Feb 27 09:31:20 Hi Feb 27 13:12:57 03Simon Busch  07org.openembedded.dev * rf51311debb 10openembedded.git/recipes/shr/libphone-ui/palmpre/libphoneui.conf: Feb 27 13:12:57 libphone-ui: add config file for palmpre machine Feb 27 13:12:57 Signed-off-by: Simon Busch Feb 27 13:12:57 Acked-by: Klaus Kurzmann Feb 27 13:12:59 03Simon Busch  07org.openembedded.dev * r4a3d4fad81 10openembedded.git/recipes/tasks/task-shr-minimal.bb: Feb 27 13:12:59 task-shr-minimal-apps: add libphone-ui-config due to changes on libphone-ui and bump PR Feb 27 13:12:59 Signed-off-by: Simon Busch Feb 27 13:12:59 Acked-by: Klaus Kurzmann Feb 27 13:13:00 03Simon Busch  07org.openembedded.dev * r15de5d3ab2 10openembedded.git/recipes/palmpre/ (tsmd/tsmd tsmd_git.bb): Feb 27 13:13:00 tsmd: return of the old init script Feb 27 13:13:00 Signed-off-by: Simon Busch Feb 27 13:13:01 03Simon Busch  07org.openembedded.dev * r2f7832b246 10openembedded.git/recipes/shr/ (libphone-ui/libphoneui.conf libphone-ui_git.bb): Feb 27 13:13:01 libphone-ui: refactor recipe to make it possible to use machine specific configs Feb 27 13:13:01 Signed-off-by: Simon Busch Feb 27 13:13:06 03Simon Busch  07org.openembedded.dev * r3eda2bfe44 10openembedded.git/recipes/palmpre/ (tsmd/tsmd_control tsmd_git.bb): Feb 27 13:13:07 tsmd: install also a little helper script to control the touchscreen daemon Feb 27 13:13:07 Signed-off-by: Simon Busch Feb 27 13:13:22 03Simon Busch  07org.openembedded.dev * r9ebfa20ec3 10openembedded.git/recipes/shr/libphone-ui_git.bb: Feb 27 13:13:22 libphone-ui: bump SRCREV Feb 27 13:13:22 Signed-off-by: Simon Busch Feb 27 14:54:53 03Thomas Zimmermann  07org.openembedded.dev * re82010f236 10openembedded.git/recipes/shr/shr-e-gadgets_git.bb: Feb 27 14:54:53 shr-e-gadgets: update SRC_REV Feb 27 14:54:53 This fixes a segfault when the GSM resource isn't initialized properly Feb 27 14:54:53 Signed-off-by: Thomas Zimmermann Feb 27 17:09:02 03Andrea Adami  07org.openembedded.dev * r4fe2b900b3 10openembedded.git/recipes/linux/ (2 files in 2 dirs): Feb 27 17:09:03 linux-kexecboot: pick collie fix from upstream for 2.6.37 Feb 27 17:09:03 Signed-off-by: Andrea Adami Feb 27 17:09:13 03Andrea Adami  07org.openembedded.dev * r286cdd2ae5 10openembedded.git/recipes/linux/ (2 files in 2 dirs): Feb 27 17:09:13 linux-kexecboot: pick collie fix from upstream for 2.6.36 Feb 27 17:09:13 Signed-off-by: Andrea Adami Feb 27 17:54:38 So, so manual testing after fixing the xattr thing shows all my stuff builds for minimal-uclibc on qemuppc and trying qemuarm now (all at once builds) Feb 27 17:55:14 hm so you added it as distro feature? Feb 27 17:55:23 to minimal Feb 27 17:55:57 (and I pushed the micro one too since I just wanted stuff fixed, but I don't strongly care, i only have minimal and meta-toolchain in my builder for micro and those don't pull in avahi Feb 27 17:56:37 hm I have it in my local.conf now for angstroem Feb 27 17:58:17 + uclibc? Feb 27 17:58:29 yes Feb 27 17:58:31 Shouldn't matter for *glibc builds since xattr isn't a feature to disable in eglibc Feb 27 17:58:34 k Feb 27 17:58:35 yes Feb 27 17:58:52 What targets? antgw100 had a uclibc build failure here Feb 27 17:59:00 haven't opened the post on the ML yet Feb 27 18:00:38 qemu and ac100 Feb 27 18:03:48 http://pastebin.com/VS4wTXLG is what I see Feb 27 18:04:17 ah yeah right Feb 27 18:04:19 hrms Feb 27 18:04:24 thats from linux headers Feb 27 18:04:43 you have to delete the define in /include/linux/socket.h Feb 27 18:05:17 hm I thought it was kernel .29 specific Feb 27 18:05:21 but seems it isnt Feb 27 18:07:42 or older uclibc on angstrom? Feb 27 18:07:52 or maybe it's an avr32 thing Feb 27 18:18:39 Tartarus: if I rememeber it correctly, then you can find the answer and patches for that issue in the mailing list archive, we tried to get it fixed for the previous release already... Feb 27 18:20:00 Tartarus no happens o uclibc-git Feb 27 18:25:44 Tartarus: nm, it was different issue, sorry for the noise Feb 27 18:41:47 btw, the same error is on angstrom-2010.x uclibc with 2.6.36 kernel (ts72xx) Feb 27 18:42:23 angstrom-2008.1 uclibc fails obviously on the libcap2 Feb 27 18:43:30 ynezz the kernel header? Feb 27 18:44:05 yep Feb 27 18:44:15 the same one as the Tartarus posted Feb 27 18:45:42 micro-uclibc fails on the hdparm and minimal-uclibc on libfsobasic (I wonder what's libfsobasic doing in the console-image) Feb 27 18:46:49 ^ Tartarus: I wonder what do you meant by that distro release mode :) Feb 27 18:47:14 ynezz as I said edit the linux/socket.h zapp the define Feb 27 18:49:34 this is a correct fix? I think, that the correct one would be to fix the uclibc? Feb 27 18:50:50 ynezz check the actual kernel header Feb 27 18:50:55 they fixed it that way Feb 27 18:51:15 ah, ok then, thanks Feb 27 18:51:33 will check it again Feb 27 19:02:59 hm yes .32 linux headers has it Feb 27 19:03:39 checking now for .37 Feb 27 19:03:41 you mean that typedef in the scoket.h ? Feb 27 19:03:46 yes Feb 27 19:03:48 it's still there Feb 27 19:05:38 nope Feb 27 19:05:43 in .37 its gone Feb 27 19:08:51 line 40? Feb 27 19:09:13 yes Feb 27 19:12:39 woglinde_: it's still same, isn't it? Feb 27 19:13:03 that line doesn't changed for ages Feb 27 19:13:37 compare .32 and .37 linux-libc-headers Feb 27 19:13:40 you will see Feb 27 19:14:01 hmpf Feb 27 19:14:03 hi Feb 27 19:14:16 what happened to http://linuxtogo.org/~* ? Feb 27 19:14:17 hi dcordes Feb 27 19:14:37 s/linuxtogo/angstrom-distribution/ still works Feb 27 19:14:48 dont know ask florian Feb 27 19:15:43 ah guess it's not critical was just wondering. maybe some urls broken now Feb 27 19:45:38 florian: trying to get some files out of ltg ~ but BW seems topped at 30k Feb 27 19:53:34 dcordes: i'll take a look Feb 27 19:58:03 florian: cool thanks Feb 27 20:18:42 03Klaus Kurzmann  07master * r1e2ebdd16a 10openembedded.git/recipes/shr/ (initscripts-shr/g_ether.sh initscripts-shr_0.0.1.bb): Feb 27 20:18:42 initscripts-shr_0.0.1.bb: install g_ether.sh for om-gta02 only Feb 27 20:18:42 Furthermore set the correct PACKAGE_ARCH for om-gta02 and nokia900. Feb 27 20:18:42 Signed-off-by: Klaus Kurzmann Feb 27 20:24:41 03Simon Busch  07org.openembedded.dev * r073ebb6298 10openembedded.git/recipes/shr/shr-settings/palmpre/elementary.sh: Feb 27 20:24:42 shr-settings: add specific elementary.sh for palmpre machine Feb 27 20:24:42 Signed-off-by: Simon Busch Feb 27 20:28:22 jo kergoth Feb 27 20:41:50 03Martin Jansa  07master * rf7ce0b3f40 10openembedded.git/recipes/shr/ (6 files): Feb 27 20:41:50 SHR: bump SRCREVs Feb 27 20:41:50 Signed-off-by: Martin Jansa Feb 27 20:42:02 03?ukasz Pankowski  07master * r99feeac40c 10openembedded.git/recipes/ffphonelog/ffphonelog_git.bb: Feb 27 20:42:03 ffphonelog: bump SRCREV to latest git, now using EXTRA_OEMAKE Feb 27 20:42:03 Signed-off-by: ?ukasz Pankowski Feb 27 20:42:03 Signed-off-by: Martin Jansa Feb 27 20:42:26 * Tartarus sighs a lack of tab completion Feb 27 20:42:31 JaMa around in some form? :) Feb 27 20:45:30 yup, sleepy tired but still some form :) Feb 27 20:46:50 heh Feb 27 20:47:13 So, the bitbake.conf change you're propposing? Mind doing it for oe-core? It looks like it's what poky has + more logic / fixes Feb 27 20:48:22 Tartarus: did you read 2nd patch and the issues between? Feb 27 20:48:56 kinda, yes Feb 27 20:49:30 2nd is not really what poky has and needs more comment Feb 27 20:50:02 after it's setled and pushed to oe, then somebody can pick it to oe-care if wanted Feb 27 20:50:07 should it just be an RFC then, esp i patchwork? Feb 27 20:50:50 in Feb 27 20:51:26 well every patch sent to ML is partially RFC or RFA, isn't it? Feb 27 20:51:56 and as it's real patch i would like to push if everybody agrees then I didn't put RFC to subject-prefix Feb 27 20:52:07 ok Feb 27 21:23:01 hrm Feb 27 21:23:35 pulling the _bootstrap multiprocesssing fix over into the server Process and the parsing pool didn't resolve this issue.. Feb 27 21:23:38 * kergoth scratches head Feb 27 21:25:33 man, I need to switch to external toolchains for fiddling with stuff on my work laptop Feb 27 21:25:44 Or get back to doing more 'quick' stuff on my desktop Feb 27 21:25:45 ? Feb 27 21:25:52 ah Feb 27 21:26:09 Updating asterisk for security stuff Feb 27 21:26:14 but didn't ahve the toolchain built recently Feb 27 21:26:42 (And once this is done i'm going to look at avahi quick and then rfc updating python-django due to a minor backwards incompat change for fixing the security issue) Feb 27 21:37:47 heh, using an external toolchain can really spoil you, wrt build times Feb 27 21:40:56 Or just not melting your laptop Feb 27 21:41:17 * Tartarus sees its currently downclocked to 800MHz on ondemand and won't go up on performance for cpufreq Feb 27 21:42:02 And as an aside, owch, asterisk is pretty bad about getting optimization flags on its own (so it's good we pass that stuff in and I hope it's sticking, will check that shortly) Feb 27 21:42:12 all sorts of "what are we? lets ask uname!" Feb 27 21:42:17 ick Feb 27 22:00:54 yeah, I kinda think no one builds asterisk outside of some PPC targets, ha Feb 27 22:00:57 * Tartarus goes to debug a bit Feb 27 22:01:28 ? Feb 27 22:01:39 hm which ppc board has pci really? Feb 27 22:02:09 at least some of the eval ones I see listed in the recipe :) Feb 27 22:02:34 checking if we can increase the maximum select-able file descriptor... cross-compile Feb 27 22:02:35 configure: error: in `/home/trini/work/OE-upstream/tmp.eglibc.minimal/work/armv5te-oe-linux-gnueabi/asterisk-1.4.39.2-r0/asterisk-1.4.39.2': Feb 27 22:02:35 configure: error: cannot run test program while cross compiling Feb 27 22:02:41 easy enough to figure out Feb 27 22:03:11 hm Feb 27 22:03:22 i think some ac_bla needs to be set Feb 27 22:03:33 yeap Feb 27 22:16:15 he ken Feb 27 22:39:32 heh, new and fixed upstream bug for asterisk Feb 27 22:39:50 ? Feb 27 22:40:01 https://issues.asterisk.org/bug_view_advanced_page.php?bug_id=18301 Feb 27 22:40:04 + a site entry Feb 27 22:40:58 lol Feb 27 22:41:00 thats a fix Feb 27 22:41:07 removed the test completly Feb 27 22:42:39 Well, an issue was they had the test twice basically Feb 27 22:43:05 uh lol Feb 27 22:50:10 * Tartarus tries, stops again, using -u goggle on failing to build stuff Feb 27 22:50:29 It's just too painful to see "here's the error place" but not the real error and have to ^C to get to work Feb 27 22:50:45 ah, right, no wonder it appears to hang. multiprocessing catches any uncaught exception and sends it to *sys.stderr*, which goes nowhere for us Feb 27 22:50:54 and an os.close() on a closed file descriptor raises an OSError Feb 27 22:51:01 eheee Feb 27 22:51:27 erm, wrong button Feb 27 22:53:01 hmmmm Feb 27 22:54:30 so, export ACKRC=.ackrc, cp ~/.ackrc .; echo '--ignore-dir=tmp' >> .ackrc # stop ack traversing the TMPDIR, at least when using it from toplevel Feb 27 22:56:09 heh, went to document INHIBIT_AUTOTOOLS_BOOTSTRAP and realized that whole section looks to be stubbed out, nothing about these sorts of vars in the oe manual yet :\ Feb 27 22:56:12 * kergoth_ mutters Feb 27 23:17:05 kergoth, how do I do regen in a subdir, properly? Feb 27 23:21:41 http://pastebin.com/e5PcLrFR is what I've got in asterisk now Feb 27 23:21:50 first 2 are for a host run tool, 3rd is a sub-component of the main thing Feb 27 23:28:09 03Andrea Adami  07org.openembedded.dev * r07bf496b33 10openembedded.git/recipes/linux/linux-kexecboot_2.6.37.bb: Feb 27 23:28:09 linux-kexecboot: bump 2.6.37 to 2.6.37.2 Feb 27 23:28:09 Signed-off-by: Andrea Adami Feb 27 23:47:04 * Tartarus waits for asterisk to build on angstrom-2008.1 (diff libtool) before pushing that + avahi Feb 27 23:48:36 yeow, once again, forgot to not try this on the laptop, lots of deps Feb 27 23:48:49 We'll see if the big box finishes doing asterisk first, heh Feb 27 23:53:50 Tartarus: what do you mean by 'regen in a subdir'? autoreconf? Feb 27 23:54:56 yeah Feb 27 23:55:11 calling it properly nukes all 'configure' files Feb 27 23:55:23 So I assume there's a syntax to call oe_autoreconf in a dir other than S Feb 28 00:09:43 it operates in $PWD Feb 28 00:09:53 erm, i think Feb 28 00:09:58 * kergoth mutters and double checks Feb 28 00:11:16 yeah, it runs in the current directory, so just cd in and oe_autoreconf Feb 28 00:11:31 it's do_configure that does the cd into ${S} to run it there Feb 28 00:18:40 not the behavior I saw, I swear Feb 28 00:19:33 maybe I need to pass acpaths= in.. Feb 28 00:20:34 hmm Feb 28 00:20:39 no, that shouldn't matter Feb 28 00:20:47 * Tartarus will try again once this current cycle is over Feb 28 00:29:31 Fuck our factory is useless. They've loaded the LCD connectors upside down on the PCB. Feb 28 00:30:07 So they've decided to jam the LCD tails into the connector instead of fixing it (if they even noticed). Feb 28 00:30:27 and the connectors aren't even closed around the tail Feb 28 00:31:02 I don't think I can separate them without breaking it Feb 28 00:34:35 owch Feb 28 01:16:00 the acpaths "magic" finding of .m4 files in the source tree is kind of hokey, too implicit imo, better off just adding the m4 dir if there's an m4 dir, or make it obey AC_MACRODIR by default Feb 28 01:37:40 ok, avahi tested on minimal and angstrom-2008.1 Feb 28 01:37:46 quick re-poke of asterisk and then push Feb 28 01:50:34 As I am getting started with oe, I tried bitbake helloworld-image and it generated ext3.rootfs.gz and rootfs.tar.gz Feb 28 01:50:57 what should I do next? ( previously I was given a img file which I could dd to a microsd to boot to my device). Feb 28 01:51:13 What is your device? Feb 28 01:51:27 with these two files, ext3.rootfs.gz and rootfs.tar.gz, I feel I still have to something more. Feb 28 01:51:30 it is chumby. Feb 28 01:52:01 but I won't mind generic instructions to understand things ( rootfs files vs a single img file) Feb 28 01:52:04 in local.conf, IMAGE_FSTYPES += "ext3" and you'll get an uncompressed ext3 image Feb 28 01:52:14 So, what's going on is this Feb 28 01:52:27 for helloworld-image, an image of each type in IMAGE_FSTYPES is built Feb 28 01:52:36 okay. Feb 28 01:52:44 What you set for DISTRO may add to this, and what you set for MACHINE may also add to this Feb 28 01:52:48 finally, in local.conf you can add to it Feb 28 01:52:49 But why is it two files? Feb 28 01:53:00 You have two things in IMAGE_FSTYPES right now Feb 28 01:53:04 ext3.gz and tar.gz Feb 28 01:53:14 ext3.rootfs.gz is an ext3 image, gzip compressed Feb 28 01:53:18 and the other is a tar.gz Feb 28 01:53:24 of the same contents Feb 28 01:53:29 Just different formats Feb 28 01:53:47 Okay, I see. Feb 28 01:54:34 afk, family stuff time Feb 28 01:54:45 Thanks Tartarus.. Feb 28 01:55:27 helloworld-image-chumby-falconwing.ext3.gz and helloworld-image-chumby-falconwing.tar.gz (looks like one is ext3 filesystem and the other is something else, while both are gzipped) **** ENDING LOGGING AT Mon Feb 28 02:59:57 2011