**** BEGIN LOGGING AT Wed Feb 18 02:59:57 2009 Feb 18 04:06:27 hello Feb 18 04:07:02 is anyone having issues compiling bluez 4-4.19? Feb 18 05:46:00 03Mike Westerhof  07org.openembedded.dev * r2d25fe0db0 10openembedded.git/conf/distro/include/preferred-slugos-versions.inc: SlugOS: nail down preferred version for avahi and tiff (libtool crappage victims) Feb 18 07:56:17 morning Feb 18 08:40:04 hey Feb 18 09:37:28 03Koen Kooi  07org.openembedded.dev * rb13f1cef96 10openembedded.git/packages/dsplink/dmai_1.20.bb: dmai: comment out install, add staging Feb 18 09:37:39 03Koen Kooi  07org.openembedded.dev * r87360d6fd8 10openembedded.git/packages/dsplink/gstreamer-ti_svn.bb: gstreamer-ti: add more vars to make xdc a bit happier Feb 18 10:23:08 http://alexthunder.livejournal.com/309815.html Feb 18 10:23:14 morning Feb 18 10:25:23 morning hrw Feb 18 10:25:36 good morning :-) Feb 18 10:26:57 noob question: what is the difference between DEPENDS and RDEPENDS? Feb 18 10:27:50 first is build-time, second is runtime Feb 18 10:28:13 oh... that explain a few things :) Feb 18 10:28:40 so stuff in RDEPENDS get installed in the image? Feb 18 10:31:11 yes Feb 18 10:36:20 hrw: ok thanks a lot Feb 18 10:36:42 and this is different from IMAGE_INSTALL? Feb 18 10:46:03 http://pokylinux.org/doc/poky-handbook.html - read this Feb 18 10:50:38 * agateau reads Feb 18 10:58:40 hrw that link contains more useful information about OE than anything else I've seen! frame it ! Feb 18 10:59:20 topic it, sticky it, do something with it so that people attempting to use OE gets to see it :D Feb 18 10:59:30 redirect oe.org homepage :D Feb 18 11:00:21 indeed Feb 18 11:15:25 there is OE manual too Feb 18 11:24:45 hrw do you maintain poky ? is there a git tree, with tags for stables etc etc ? Feb 18 11:45:52 http://wiki.openembedded.net/index.php/Documentation is linked from the OE home page Feb 18 12:11:08 re Feb 18 12:12:09 BusError: git.pokylinux.org has git repository with stable branches. I do not maintain Poky but did that in past. Now I work on BUG Linux which is based on last stable release of Poky ('pinky' 3.1) Feb 18 12:12:52 BusError: and I am the only non-OH/Intel developer with r/w on that server. Not directly to Poky itself but to Poky-contrib repo which gets merged into Poky after review Feb 18 12:46:22 Hello! Feb 18 12:48:08 I have a HARD BUG, but developsers of OpenEmbedded ignoring it!!! http://bugs.openembedded.net/show_bug.cgi?id=5004 I know many people, which is have SAME problem!!! No solution for this problem! OpenEmbedded is totally unusable! Feb 18 12:48:51 * BusError wish there were more exclamation marks in the world, today Feb 18 12:50:55 fpga: do not shout Feb 18 12:51:29 fpga, way to make friends and influence people ... Feb 18 12:52:02 fpga: 5004 should be split into parts thats 1. Feb 18 12:52:45 fpga: 2. you did not gave idea which exactly MACHINE it is Feb 18 12:53:17 I assume that kernel is for OABI and rootfs is EABI - this is handled for OE generated kernels/rootfs Feb 18 12:53:36 fpga: I will close it as wontfix and wait for more info Feb 18 12:55:07 stop Feb 18 12:55:13 do not close Feb 18 12:57:28 Do not close without any comments - it's a very bad. Just write solution! OABI or EABI? I specified SUPPORTED machine type - with all known parameters. Why rootfs created wrong??? It's a definitely BUG. Supported machine type, I followed instructions strictly - result wrong = BUG Feb 18 12:59:37 fpga: lets start from beginning ok? Feb 18 12:59:49 and one problem at time Feb 18 13:00:40 03Koen Kooi  07org.openembedded.dev * rb46ca4b9bb 10openembedded.git/packages/linux/ (3 files in 3 dirs): linux-omap-pm: add patch to register all OPPs, tweak defconfig Feb 18 13:00:47 ok Feb 18 13:00:58 hrw: how are you? Feb 18 13:01:11 MACHINE=ep93xx is listed as maintained by Koen Kooi - ask him when last time he tried build on it Feb 18 13:01:19 zecke: good, thanks Feb 18 13:01:28 zecke: preparing to Mira's first birthday Feb 18 13:01:57 fpga: and tell me which kernel you got built Feb 18 13:03:04 I build my own kernel. But I tested with kernel from your site for my edb9302 - does not work Feb 18 13:03:10 same kernel panic Feb 18 13:03:48 Kernel + rootfs from your site - kernel panic Feb 18 13:04:01 'your site' == angstrom? Feb 18 13:04:46 fpga: show me config of your kernel - via pastebin.ca for example Feb 18 13:05:49 yes, angstrom Feb 18 13:08:05 * * OE Bug 5004 has been RESOLVED (WONTFIX) by openembedded(AT)haerwu.biz Feb 18 13:08:07 * * generated totally unusable rootfs Feb 18 13:08:09 * * http://bugs.openembedded.net/show_bug.cgi?id=5004 Feb 18 13:14:37 Sorry, what is wanted from me now? Now I reading "EABI/OABI incompatibility" article... Is it possibly source of my problem? Feb 18 13:15:10 yes Feb 18 13:15:20 thats why I am asking you for kernel config Feb 18 13:15:43 it should have EABI and Thumb support enabled. Feb 18 13:15:54 if it dies at init then it lacks them Feb 18 13:23:37 ok, 10x Feb 18 13:23:48 I'll try EABI... Feb 18 13:25:03 Where in kernel configuration I can find ABI type? I cant find it in menuconfig and xconfig... Feb 18 13:25:17 I sse it only in text... Feb 18 13:26:26 moment Feb 18 13:26:55 Kernel features -> Use the ARM EABI to compile the kernel (NEW) Feb 18 13:29:59 flo_lap, Crofton: The antenna arrived yesterday, thanks guys! Feb 18 13:30:49 alphaone: yw Feb 18 13:32:26 alphaone, awesome Feb 18 13:32:32 sorry for the delay :) Feb 18 13:45:02 Crofton: no problem, I didn't really need it until now. Feb 18 13:45:22 Had a Wifi antenna that was okay Feb 18 13:47:18 alphaone, almost any conductor works as an antenna :) Feb 18 13:47:49 Only question is, does it work well enough? :-) Feb 18 13:48:45 2 hrw: can I edit wiki http://wiki.openembedded.net/index.php/Getting_Started to add notes about EABI? Feb 18 13:49:39 I got many ultranegative emotions from openmebdedded without knowledge about ABI... Feb 18 13:56:31 fpga: you can of course. but what about writing separate page about it and adding note+link into getting started? Feb 18 13:56:57 i will definitely need to do that aswell Feb 18 13:57:28 with all the help i got getting OE up and running on ubuntu 8.10 Feb 18 14:14:35 03Koen Kooi  07org.openembedded.dev * ref3439f882 10openembedded.git/packages/linux/ (3 files in 3 dirs): linux-omap-pm: add patch to register all OPPs, tweak defconfig Feb 18 14:14:37 03Koen Kooi  07org.openembedded.dev * rcc7bd3b8bc 10openembedded.git/: Merge branch 'org.openembedded.dev' of git@git.openembedded.net:openembedded into org.openembedded.dev Feb 18 14:14:39 03Roger Monk  07org.openembedded.dev * r09b88241b5 10openembedded.git/packages/dsplink/ (dmai_1.16.bb dmai_1.20.bb): dmai: update 1.20, remove obsolete 1.16 Feb 18 14:14:40 03Roger Monk  07org.openembedded.dev * r9485db7d15 10openembedded.git/packages/dsplink/ (16 files in 2 dirs): codec-engine: fix lpm build and clean up old files Feb 18 14:30:32 hi! i want to cross compile some c code for my at91sam9260 to build a java bt api for my system.... have an build.xml that i must change to use the cross compiler Feb 18 14:31:44 thorsten1: it has been a hard path to get a solution to your problem :-( Feb 18 14:32:04 yes :-( Feb 18 14:32:43 but its the last try to get it running, then i prog it in c++.... Feb 18 14:32:51 thorsten1: luckily you're stubborn ;-) Feb 18 14:33:11 :P Feb 18 14:33:37 last time I used ant it support CC= environment variables so you can just CC="cross-compiler" Feb 18 14:34:18 when you havn't worked to all of that, you must be stubborn :D Feb 18 14:35:15 yes that i have done, cc=arm-none-linux-gnueabi, it compiled very well! but later it use ldd Feb 18 14:37:45 there is an LD= as well isnt there Feb 18 14:37:53 LD="cross-compiler" should work Feb 18 14:41:27 in the xml? Feb 18 14:45:12 i added LD=arm-none-linux-gnueabi-ld you mean that, but still doesnt run :-( Feb 18 14:46:48 thorsten1: hmmm, been a couple of years since I used ant Feb 18 14:47:02 thorsten1: maybe the variable name is wrong Feb 18 14:47:11 ok found something else, but now therse another error :D Feb 18 14:47:18 thorsten1: but I used to use ant to cross compile for smart cards and that worked somehow Feb 18 14:47:26 I booted Angstrom! Yyyesssssss... But I getting errors Starting udevnbd1: Attempted send on closed socket end_request: I/O error, dev nbd1, sector 0 FAT: unable to read boot sector nbd0: Attempted send on closed socket end_request: I/O error, dev nbd0, sector 0 FAT: unable to read boot sector and so on... But after this I see console and success login Feb 18 14:47:30 there wasnt use a variable, simply used ldd as executable Feb 18 14:48:02 fpga: nbd1 is remote network block device? Feb 18 14:48:37 I dont know :) Feb 18 14:48:56 Kernel command line: rootdelay=10 root=/dev/sda1 console=ttyAM,57600 Feb 18 14:49:06 fpga: show me dmesg? Feb 18 14:49:08 I'm using USB flash... ext2 Feb 18 14:49:10 ok Feb 18 14:50:53 At second boot I see new error messages: "udevd[241]: lookup_group: specified group 'scanner' unknown" and so on for nvram tss fuse kvm rdma Feb 18 14:52:07 dmesg is so long... may I put it here? Or can I use other resource? Feb 18 14:52:15 ~pastebin Feb 18 14:52:16 [~pastebin] A "pastebin" is a web-based service where you can paste anything over 3 lines without flooding the channel. Here are links to a few : http://www.pastebin.com , http://pastebin.ca , http://channels.debian.net/paste , http://paste.lisp.org , http://www.rafb.net/paste Feb 18 14:52:25 ok Feb 18 14:53:50 http://pastebin.com/d5c03ebd7 :) Feb 18 14:55:35 other errors: http://pastebin.com/d2cf5855d Feb 18 14:56:31 fpga, be aware you may be the only person here with that hw :) Feb 18 14:56:49 morning Feb 18 14:58:59 sorry Crofton, can't extract sense from this message, I'm russian, I dont speak english well Feb 18 14:59:30 lookup_group are harmless Feb 18 14:59:33 03Roger Monk  07org.openembedded.dev * r0acbb40cf1 10openembedded.git/packages/dsplink/ (dmai_1.20.bb files/dmai-unbreak-xdc-args.patch ti-paths.inc): dmai: remove xdc hack Feb 18 15:00:20 fpga: you can disable nbd in kernel probably ;d Feb 18 15:00:25 fpga: share kernel config Feb 18 15:00:32 * hrw -> food Feb 18 15:01:13 OK. How about http://pastebin.com/d2cf5855d ? Strange errors from udev... Feb 18 15:05:38 fpga: there are some russians ;) Feb 18 15:06:08 seems that you have no that groups in /etc/groups Feb 18 15:06:24 but that groups are in udev rules Feb 18 15:06:31 someone should fix that Feb 18 15:10:30 My kernel config : http://pastebin.com/d5e9dcf4 Feb 18 15:12:31 "someone should fix that" - how to fix, which files I shold edit, fix, etc? Feb 18 15:13:09 fpga: easy way is to add this groups to /etc/group Feb 18 15:13:26 long way - search and replace it in udev rules Feb 18 15:14:02 first way seems to be more correct Feb 18 15:14:12 Jay7 - I want to fix this 4ever. Where is the sources? Feb 18 15:15:13 fpga: hehe.. when you want fix it in OE then you should search what package supply /etc/groups Feb 18 15:15:14 For exmaple, somebody building this again and see same errors. How to remove them from "root source"? Feb 18 15:16:00 I'm unsure but this 'error' already known and may be already fixed in some distro's or images Feb 18 15:17:27 This is quite perrectal... OK, I'll fix it manually in my image... Feb 18 15:17:47 fpga: try to search in bugzilla and in oe-devel/angstrom-devel maillist Feb 18 15:18:00 may be in *-users@ too Feb 18 15:18:11 ok Feb 18 15:19:05 Openembedded: Mem: 7540K used, 21496K free, 0K shrd, 536K buff, 3140K cached = it's a very nice! Just 8 Mb RAM from system... nice... Feb 18 15:20:40 anyone messed with generating x86 vmware player images from OpenEmbedded yet? Feb 18 15:23:00 fpga: change CONFIG_BLK_DEV_NBD to module if you do not use nbd for booting Feb 18 15:23:51 fpga: btw, why do not use kernel from OE/Angstrom? Feb 18 15:23:52 fpga: update to recent OE - udev groups problem was solved on Friday Feb 18 15:35:47 03Koen Kooi  07org.openembedded.dev * r5fe2271f24 10openembedded.git/contrib/angstrom/local.conf: angstrom contrib: check in local.conf used on autobuilders Feb 18 15:40:12 03Felix Domke  07org.openembedded.dreambox * rbc3c61c5b9 10openembedded.git/packages/tuxbox/tuxbox-common.bb: tuxbox-common: update satellites.xml Feb 18 15:42:48 that udev fix, wouldn't it be easier just to add those missing groups? http://git.khz.cz/?p=openembedded-ts72xx;a=commitdiff;h=d6eb3351eb112126e36b91f3b2879c9b70efda83 Feb 18 15:43:11 I am trying to integrate qtopia in a rootfs, but bitbake does not like it: http://pastebin.com/m7de0b45f Feb 18 15:43:23 any idea what the correct way to do it is? Feb 18 15:43:54 ynezz, no it wouldn't Feb 18 15:44:25 don't have the rest of the userland stuff for dealing with that stuff mostly. In other cases it didn't really make sense Feb 18 15:49:08 hm, i have a recipe that builds from git, I download a zip snapshot, but the number in the current snapshot is lower than the previous one, so if one has installed the previous one, this one is considered to be older. What is the way around this? bump PR ? Feb 18 15:49:51 eFfeM: do you influence the numbers? Feb 18 15:50:24 zecke: no Feb 18 15:50:59 Tartarus: ok Feb 18 15:51:39 eFfeM: as a last resort you can bump the PE to get a upgrade path Feb 18 15:51:47 eFfeM: but where do you put the snapshot rev you get? Feb 18 15:52:19 zecke: sane-srcrevs Feb 18 15:53:06 "btw, why do not use kernel from OE/Angstrom?" - kernels from OE/Angstrom is obsolete... I do some Linux Kernel programming and I need ultra-mega-super=latest kernel from git Feb 18 15:53:13 nevermind, i think I made a mistake, wrong rev Feb 18 15:53:16 eFfeM: ah so this git repository is using rebase and is editing history? Feb 18 15:53:36 zecke no, i picked the wrong number Feb 18 15:54:39 sry about that Feb 18 15:55:08 :) Feb 18 15:55:57 fpga: create bitbake recipe for new kernel ;) Feb 18 15:56:06 when there is no one already Feb 18 16:04:10 zecke, apparently I did not goof up, while looking at the problem the 2nd time I made a mistake :-( Feb 18 16:04:26 I am trying to update midori from http://git.xfce.org/kalikiana/midori/ Feb 18 16:05:05 last commit is http://git.xfce.org/kalikiana/midori/commit/?id=3763990c3ad038013a3db7052f67c4c516a53560 and gives midori-3763990c3ad038013a3db7052f67c4c516a53560.tar.bz2 as download Feb 18 16:05:32 the 37... is from srcrevs but the previous version had a larger number Feb 18 16:12:08 this one looks nice but is 5.5v: http://alturl.com/2at build in battery. solar 1000 ma Feb 18 16:12:19 wrong chat Feb 18 16:12:21 eFfeM: ah "larger" as in comparing hashes Feb 18 16:12:59 eFfeM: yes, this is an outstanding issue in OE with srcrev on git... Feb 18 16:13:04 zecke yes, so if you install it says it cannot overwrite Feb 18 16:13:06 eFfeM: last resort is bumping PE :( Feb 18 16:13:20 I have to discuss this with RP Feb 18 16:13:26 RP: when do you have time? :) Feb 18 16:13:34 hm Feb 18 16:14:14 what about constructing the git names in such a way that you can just bump PR so have a name that looks like ${PR}_$(SRCREV} or so Feb 18 16:14:49 * eFfeM never bumped PE so not really an idea on where and on impact Feb 18 16:16:48 eFfeM: what people (e.g. koen) do is to put the PV into the PR... Feb 18 16:17:11 but that is bad... at Openmoko I solved this issue about 11 month ago by putting another number into the PV that counts upwards Feb 18 16:19:25 would PV="0.1" or so do the trick ? Feb 18 16:20:15 actually what I did was construct a name with the last published version + gitr Feb 18 16:26:33 hello Feb 18 16:26:39 anybody awake? Feb 18 16:26:52 anybody knows about bitbaking with OE for Nokia n810? Feb 18 16:27:02 I was advised to use chinook-compat Feb 18 16:27:14 but that is not in stable? Feb 18 16:31:41 you need .dev Feb 18 16:34:25 Updated my OE... "Using dash as /bin/sh causes various subtle build problems, please use bash instead." Feb 18 16:34:37 it's a sanity checker Feb 18 16:35:58 some apps use /bin/sh for bash scripts ;( Feb 18 16:36:13 we should really fix all of those. Feb 18 16:36:28 hmmm Feb 18 16:36:29 DISTRO 'chinook-compat' not found. Feb 18 16:36:37 kergoth: this time this is gtk+ Feb 18 16:36:37 even though I have .dev in local.conf Feb 18 16:36:56 udovdh: and in BBPATH? Feb 18 16:37:17 udovdh: DISTRO_NAME = "Chinook-compat" Feb 18 16:37:27 ahh.... Feb 18 16:37:28 thanks! Feb 18 16:37:30 org.openembedded.dev/conf/distro/chinook-compat.conf Feb 18 16:37:36 according to this ^^ Feb 18 16:37:55 Jay7: DISTRO_NAME is not user thing Feb 18 16:38:13 user needs to set DISTRO to name of file with distro definition. so 'chinook-compat' Feb 18 16:38:13 hrw: I mean name "Chinook-compat" Feb 18 16:38:22 not variable :) Feb 18 16:38:37 ah Feb 18 16:38:40 filename is the key Feb 18 16:38:43 yes Feb 18 16:39:20 "some apps use /bin/sh for bash scripts" - how to disable warning and start build? Feb 18 16:39:30 change symlink Feb 18 16:39:52 do I need to clean some cache before rebuild? Feb 18 16:40:02 DISTRO is in use, it does a bit more now Feb 18 16:40:23 fpga: not need Feb 18 16:40:54 "change symlink" warning does not show me which package causes this warning... I must search? Feb 18 16:41:44 fpga: go to /bin and look at /bin/sh Feb 18 16:41:56 it should be symlink to something like /bin/dash Feb 18 16:42:14 then relink it to /bin/bash Feb 18 16:42:41 it should be safe and nothing should break :) Feb 18 16:42:46 ok Feb 18 16:44:27 now: "sh -> /bin/bash" build started. Thanks :) Feb 18 16:44:27 hmmm Feb 18 16:44:29 ERROR: Feb 18 16:44:29 Dependency loop #1 found: Feb 18 16:44:30 etc Feb 18 16:44:33 hrw: how long you are working with OE? :) Feb 18 16:44:40 four of them! Feb 18 16:44:47 how to approach that? Feb 18 16:45:44 lemmu uncomment some -stable stuff Feb 18 16:47:40 Jay7: 5 years ago I bought Sharp Zaurus SL-5500 Feb 18 16:48:11 well.. 2 years ago I bought Sharp Zaurus SL-C1000 Feb 18 16:48:19 please have a look at: Feb 18 16:48:20 http://pastebin.com/m2c1c976f Feb 18 16:48:23 I have 3 years ;) Feb 18 16:48:26 for my dependency loops Feb 18 16:48:37 I took my local.conf for stable and h2200 Feb 18 16:48:43 changed it into -dev Feb 18 16:48:43 Jay7: my first submission was on 22 March 2005 Feb 18 16:48:55 and chinook-compat plus nokia800 Feb 18 16:49:25 Jay7: http://oe-devel.bkbits.net:8080/openembedded/?PAGE=cset&REV=405fe58a-SCbeZdn3UPShhP97OHQFg Feb 18 16:49:30 udovdh: I'm at your place will try to merge versions from beagleboard e.g. :) Feb 18 16:49:57 Jay7, euhm? Feb 18 16:50:06 also for larm and n800? Feb 18 16:50:10 you mean? Feb 18 16:50:26 http://pastebin.com/d341832c9 for local.conf Feb 18 16:50:49 udovdh: I mean that chinook-compat.conf needs more love :) Feb 18 16:50:58 ah Feb 18 16:51:05 so the loops might not be my 'fault'? Feb 18 16:51:10 well Feb 18 16:51:13 I can test! Feb 18 16:51:27 I have a working bitbake setup for h2200 and angstrom Feb 18 16:51:32 to build navit Feb 18 16:51:43 and I am aiming to build navit for n810 as well Feb 18 16:51:48 using the same setup Feb 18 16:52:01 q: what could I do at this stage? Feb 18 16:53:26 btw, who are Nokia IT's mentor? Feb 18 16:54:00 dunno Feb 18 16:55:48 Jay7: you mean chinook-compat maintainer? Feb 18 16:56:42 hrw: yes Feb 18 16:59:10 Robert Schuster created it and maintained, I did some updates in meantime Feb 18 17:01:50 khem: it turned out xserver-common sets up the ts calibration Feb 18 17:05:19 bye Feb 18 17:41:48 Hmmm Feb 18 17:42:00 ipk doesn't have size constraints does it? Feb 18 17:42:13 I'm seeing a bunch of files jsut not be installed Feb 18 17:42:24 That are in the .ipk Feb 18 17:42:29 (trying to install ltp) Feb 18 17:43:31 Tartarus: odd, i'm unaware of size constraints Feb 18 17:43:37 have you actually looked into the ipk ? Feb 18 17:43:43 yeap Feb 18 17:43:53 data.tar.gz has everything I expect Feb 18 17:43:56 very strange Feb 18 17:43:57 image does not Feb 18 17:44:24 i'm not an ipkg wiz, but isn't there an option to show what it does? Feb 18 17:45:42 opkg is badly broken :( Feb 18 17:45:50 mwester, in which way? Feb 18 17:46:21 i know -nogpg-nocurl wasn't building due to really wanting curl Feb 18 17:46:28 Tartarus: are you using opkg or ipkg Feb 18 17:46:31 Just look at the code. And if that isn't horrifying enough, run it through a static analysis tool (you might even try a lint-derived bit of code). Feb 18 17:46:39 It leads memory like a sieve. Feb 18 17:46:47 eFfeM, opkg Feb 18 17:46:48 It abuses/misuses /tmp Feb 18 17:46:55 * eFfeM noticed several memory leaks Feb 18 17:47:12 It decompresses the same control file up to three times during package install, for no apparent reason. Feb 18 17:47:14 not sure if opkg is an improvement above ipkg Feb 18 17:47:22 It least empty tmp directories behind. Feb 18 17:47:28 although okpg is not that nice either Feb 18 17:47:30 mwester, 3? That's a minimum Feb 18 17:47:32 It's abominable. Feb 18 17:47:35 haven't seen less than 10 in my images Feb 18 17:47:58 * Tartarus is gonna try switching to debs, since actual target side things aren't too important right now Feb 18 17:48:10 Tartarus: if you are adventurous: run opkg under strace Feb 18 17:48:55 Tartarus: I'm talking about tracing the execution -- if you install abc.ipk -- it will extract the control.tar.Z file at least twice, and decompress it -- which is pointless as it only needs to decompress it once, and I have seen cases where it has been extracted three times from the same package.ipk in a single opkg install command. Feb 18 17:49:28 ah Feb 18 17:49:38 All of that takes memory, and if your /tmp is tmpfs, then it comes out of your available RAM -- and the OOM killer gets you. Feb 18 17:50:03 And let us not even mention the return code. What the HECK is that?? "error code 590837843698738534798587395734985723498572439853289579857389579435987398" Feb 18 17:50:06 ?? Feb 18 17:50:18 Grrrrr Feb 18 17:51:01 How do I get qemu to emulate a spitz? (or something near enough to boot a spitz hdd image?) Feb 18 17:51:14 Basically, if it weren't for the fact that the Gods Of OE have DICTATED the use of opkg, I would have gone back to ipkg a looong time ago, and I would have saved many days of my life debugging and patch opkg. Feb 18 17:52:01 oh, yet one more reason not to actually use oe :} Feb 18 17:53:23 mwester: I don't think there's any particular reason you couldn't use ipkg with the current version of oe. Even if the current opkg bits are no longer compatible with ipkg it would be easy enough to fish the old ones back out of version control. Feb 18 17:53:36 No, that's not typical. In fact, it's atypical -- normally the OE core folks will err on the OTHER direction, and reject things that are not coded to a high-enough standard. In this case, I cannot think that anyone actually looked at the source. Feb 18 17:54:02 pb_: We went over this on the ML a long time ago; it's hardcoded into the class files themselves. Feb 18 17:54:26 mwester: which class files? package_opk. you mean? Feb 18 17:54:36 pb_: and more. Feb 18 17:54:58 which other ones? if opkg is hard-coded in any more fundamental classes than that, it is surely just a bug. Feb 18 17:54:59 I think it was also in kernel.bbclass as well. Feb 18 17:55:07 that does sound pretty bogus. Feb 18 17:55:18 No, apparently it is my thinking that is the bug, or so the thread generally went. Feb 18 17:55:27 opkg iis a travesty. Feb 18 17:55:52 eekee: try qemu-system-arm -M ? look at http://linuxtogo.org/gowiki/AngstromAndQemu Feb 18 17:56:02 agreed, opkg seems pretty bad. I'm not entirely convinced that ipkg is all that much better though. Feb 18 17:56:20 ipkg is totally unmaintained and buggy as hell Feb 18 17:56:27 I would love to share the results -- I had an opportunity to run a commercial C/C++ static analysis tool on the source code; it was appalling. Feb 18 17:57:08 yah, I can imagine. the thing I am not sure about, though, is how much of that horror was inherited from ipkg and how much is newly-created in opkg. Feb 18 17:57:22 But I am bound by a non-disclosure on that. It was enough to make me abandon plans to try to fix or patch opkg -- it is beyond fixing. Feb 18 17:57:47 dpkg works fine Feb 18 17:57:59 Interesting though. Feb 18 17:58:15 mr_nice: thanks! Feb 18 17:58:35 Hmmm... I wonder if trimming down dpkg to fit on the NSLU2 would be a vialble solution... that's a really good suggestion. Feb 18 17:58:47 Thanks XorA! Feb 18 17:59:08 eekee: qemu-system-arm -M spitz -kernel zImage -hda hdimage.img -mtdblock mtdblock.img Feb 18 17:59:14 eekee: something like this :) Feb 18 17:59:43 eekee: look at akita example and add -hda there :) Feb 18 17:59:58 mwester: I even merged in the poky fixes for dpkg Feb 18 18:00:00 eekee: it would be nice to add the infos to the wiki as well Feb 18 18:00:08 :) Feb 18 18:00:31 eekee: once time I have using this: qemu-system-arm -M spitz -show-cursor -portrait -mtdblock /dev/null -kernel zImage-spitz.bin -hda x11-image-spitz.ext2 -append "root=/dev/hda rootfstype=ext2" Feb 18 18:00:39 but it seems not very correct Feb 18 18:03:53 root should probably be /dev/hda1 Feb 18 18:04:24 that *.ext2 have no partitions inside Feb 18 18:04:39 this is one of wrong things Feb 18 18:04:46 ah I can take append options from /proc/cmdline of the same machine I took the disk image from :) Feb 18 18:06:06 you can try -drive file=rootfs.ext2 -append "root=/dev/sda" I use this for qemuarm Feb 18 18:07:08 hehe it boots! only trouble it it's sideways, & -landscape doesn't work Feb 18 18:08:33 good stuff though, I can definitely work with that Feb 18 18:08:55 append "fbcon=rotate:1" to cmdline Feb 18 18:09:22 btw, I have problems with spitz and mouse Feb 18 18:09:27 *had Feb 18 18:09:29 ah actually I should remove that :) Feb 18 18:10:32 maybe I should add -portrait? Wondering if it's kinda inverted logic Feb 18 18:10:51 try it all :) Feb 18 18:10:57 yeah :) Feb 18 18:12:37 I should probably run it with -nographic some of the time, seeing as I mostly want it for compiling Feb 18 18:13:48 hehe.. compiling in qemu is hardcore :) Feb 18 18:14:21 03Bernhard Guillon  07org.openembedded.dev * r9b2a4dfecb 10openembedded.git/packages/konqueror/ (3 files in 2 dirs): konqueror-embedded: fix acinclude and unset KDE_LDPATH, bump PR Feb 18 18:14:51 I grew up thinking I was a wimp, but I always seem to end up on the hardcore train LOL Feb 18 18:35:29 Hmm Feb 18 18:35:38 Anyone know the deb classes? Feb 18 18:36:18 dpkg-deb: parse error, in file `/OE/tmp/work/mipsel-linux/eglibc-2.9-r3/install/eglibc-localedata-gez-er+abegede/DEBIAN/control' near line 11 package `eglibc-localedata-gez-er+abegede': Feb 18 18:36:18 `Provides' field, invalid package name `glibc-localedata-gez_er@abegede': character `@' not allowed (only letters, digits and characters `-+._') Feb 18 18:37:30 hmm, maybe I can fix this :) Feb 18 18:38:56 legitimize_package_name needed more i think Feb 18 18:41:45 Tartarus: hmm @ should not be added Feb 18 18:42:45 well, ipk is OK wit @s, debs are not, it seems Feb 18 18:42:49 i think i got a fix now Feb 18 18:42:55 will push once i test Feb 18 18:56:03 re Feb 18 18:58:20 khem, OK, fixed it Feb 18 18:58:31 really just needed legitimize_package_name in the RPROVIDES stuff I added before Feb 18 18:58:49 I'll push soon, just in the middle of testing other stuff Feb 18 18:58:53 Bah, spoke too soon Feb 18 19:10:15 btw. I have a patch for eglibc cross locales (building ) so we can get rid of qemu but it needs testing Feb 18 19:12:08 * Tartarus moves his legitimize_package_name call around a bit Feb 18 19:15:49 Well what the heck Feb 18 19:16:07 bb.data.setVar('RPROVIDES_%s' % pkg, 'glibc-localedata-%s' % legitimize_package_name(group).lower(), d) Feb 18 19:20:32 thesing: Any news re collie? Feb 18 19:20:42 I managed to get past the boot screen Feb 18 19:20:48 * Tartarus tries another track Feb 18 19:21:06 thesing: I think the SD driver still does not like one of my cards, it never liked it. Feb 18 19:21:24 but booting the collie ultimately hangs Feb 18 19:21:36 Laibsch: hi Feb 18 19:21:44 I don't have the serial cable with so I don't know more Feb 18 19:21:52 thesing: Hi, how are you doing? Feb 18 19:22:02 s/with/with me/ Feb 18 19:22:21 Laibsch: does the plash got up to the middle? Feb 18 19:24:27 ? Feb 18 19:24:40 Oh, the splash bar? Feb 18 19:24:59 Yes, I think it hangs right about half-way through Feb 18 19:27:01 OK, maybe I just can't test this kind of fix with really rebuilding eglibc Feb 18 19:27:44 trying the rebuild gcc-cross-intermediate then rebuild eglibc trick Feb 18 19:27:52 then when that fails again :( blowing away Feb 18 19:29:33 Laibsch: then try editing /etc/init.d/bootmisc.sh and comment out the hwclock calls. Feb 18 19:32:33 Failed again, of course :( Feb 18 19:36:54 Hrm. How do I override sanity? Feb 18 19:37:11 The instructions say to touch sanity.conf but that causes sanity.bbclass to go nuts. Feb 18 19:41:23 Pah, SDK stuff is tied to opkg Feb 18 19:41:36 If dpkg works for the LTP issue, I'll tackle that next Feb 18 20:04:09 thesing: Thank you for the hint. I commented out both occurences of lines with hwclock in /etc/init.d/bootmisc.sh, the screen hung for a long time at about 60% progress. I'm presented with an X pointer after all! Feb 18 20:04:33 I can't go on from there because I think I brought the collie with the broken touchscreen :-/ Feb 18 20:04:49 The other one's at home Feb 18 20:05:27 Maybe you want to comment out those two lines in the repo? Feb 18 20:05:45 I disabled the rtc in the kernel. Feb 18 20:05:59 So It shouldn't hang with current images. Feb 18 20:15:25 What is current? Feb 18 20:15:37 I think the image is about a week old, max Feb 18 20:15:40 too old? Feb 18 20:28:05 Laibsch: yes I commited on Friday. Feb 18 20:28:16 aha Feb 18 20:28:17 thesing: 2.6.28 doesnt seem to fix my SD woes :-( Feb 18 20:28:24 good, thanks for the work Feb 18 20:29:08 XorA: so you card is detected by kexecboot, but you get errors later? Feb 18 20:29:21 thesing: yes Feb 18 20:29:34 thesing: it seems once it starts erroring it never recovers Feb 18 20:29:52 what errors do you get? Feb 18 20:30:27 hello, i'm having some trouble building popt with oe, it craps out with libtool saying "link: unsupported hardcode properties" Feb 18 20:31:35 the offending line is "/bin/bash ./arm-linux-uclibcgnueabi-libtool --tag=CC --mode=link arm-linux-uclibcgnueabi-gcc -march=armv5te -mtune=xscale -mno-thumb-interwork -mno-thumb -isystem/OE/zauros/tmp/staging/armv5te-linux-uclibcgnueabi/usr/include -fexpensive-optimizations -fomit-frame-pointer -frename-registers -O2 -L/OE/zauros/tmp/staging/armv5te-linux-uclibcgnueabi/usr/lib -Wl,-rpath-link,/OE/zauros/tmp/staging/armv5te-linux- Feb 18 20:31:37 uclibcgnueabi/usr/lib -Wl,-O1 -o libpopt.la -rpath /usr/lib popt.lo findme.lo poptparse.lo poptconfig.lo popthelp.lo -lintl" Feb 18 20:32:10 thesing: give me a few mins, need to write a new card Feb 18 20:32:37 if i remove -lintl, it continues but then i get missing references Feb 18 20:32:46 so i guess i have a problem with libintl? Feb 18 20:38:40 dromede: look at the config.log Feb 18 20:38:44 XorA: try to dd card with zeroes and then make newfs again Feb 18 20:39:28 Jay7: what does that gain? Feb 18 20:39:48 XorA: it helps me two times with my SD card Feb 18 20:40:01 but there was fat Feb 18 20:40:12 dromede: configure might not be able to find some of macro expansions like AM_ICONV etc Feb 18 20:40:22 Jay7: the errors are in the mmc layer, not filesystem Feb 18 20:40:55 XorA: kernel complains about bad sectors Feb 18 20:41:09 it's strange, because i can get it to compile if i do it by hand Feb 18 20:41:14 after dd'ing it continue working again Feb 18 20:41:45 that is, i did CC=my-cross-gcc ./configure --host=arm-my-target Feb 18 20:41:47 looks like some incompatibility between my card reader and Z's one Feb 18 20:41:49 dromede: thats possible, bitbake sets quite a few variables Feb 18 20:42:11 Jay7: well longshot I am trying it :-) Feb 18 20:42:15 so i was thinking i should just do it by hand and then add the do_compile stamp manually Feb 18 20:42:40 i'm using libtool 2.2.4 and automake 1.10 Feb 18 20:42:49 dromede: I would prefer to find the reason why it does not work with OE env Feb 18 20:43:26 you're right, but i'm out of ideas Feb 18 20:43:35 bit/win 12 Feb 18 20:43:51 dromede: ok, do you have your logs somewhere Feb 18 20:44:11 dromede: and which distro are you building Feb 18 20:45:23 oe is very new to me, i've only used it for a few days now Feb 18 20:45:35 so i'm trying to build minimal-uclibc for akita Feb 18 20:45:50 dromede: no worries, I can help to figure out your problem Feb 18 20:45:53 ok akita Feb 18 20:46:33 and i'm trying to build task-base Feb 18 20:47:50 so i think the problem is that libtool can't find my libintl Feb 18 20:47:58 could be Feb 18 20:48:06 or it doesnt like the "-lintl Feb 18 20:48:15 or it doesnt like the "-lintl" switch Feb 18 20:48:27 but the library is there Feb 18 20:49:57 dromede: is it popt ? Feb 18 20:50:06 yes Feb 18 20:52:27 oh dammit, i can't resume my task-base build Feb 18 20:52:40 and i don't know what i did :/ Feb 18 20:52:57 what does it say Feb 18 20:54:09 Missing or unbuildable dependency chain was: ['bluez-utils', 'gstreamer', 'popt', 'libintl'] Feb 18 20:54:16 ERROR: Required build target 'task-base' has no buildable providers. Feb 18 20:54:24 Missing or unbuildable dependency chain was: ['task-base', 'bluez-utils', 'gstreamer', 'popt', 'libintl'] Feb 18 20:54:31 i'll try to reset my env Feb 18 20:54:47 and i'll reset the cache Feb 18 20:55:56 XorA: if you don't get it working a log with MMC_DEBUG enabled would help Feb 18 20:56:45 thesing: need to dig out my serial cable as well, but gf is on her way, so not tonight Feb 18 20:57:28 XorA: understandable ;) Feb 18 20:58:30 * XorA hopes thre are not collies with different SD hardware to be annoying :-) Feb 18 21:03:45 both collies here exhibit same behaviour :-) Feb 18 21:05:44 ok, i got it running again Feb 18 21:06:25 and it still craps out on popt Feb 18 21:06:31 same libtool error Feb 18 21:09:04 how did u get it working Feb 18 21:09:26 i messed with popt DEPENDS and forgot about it Feb 18 21:09:55 hmm Feb 18 21:10:39 config.log wont help, do_configure works fine Feb 18 21:10:46 the do_compile stage fails Feb 18 21:11:07 i'll paste the offending line again Feb 18 21:11:16 dromede: I know :) it well Feb 18 21:11:58 dromede: still I need to see config.log Feb 18 21:12:12 just a sec Feb 18 21:13:53 here you go: http://pastebin.com/m1bbe272e Feb 18 21:18:18 dromede: OK so it finds libintl Feb 18 21:18:35 and puts it in linker commandline Feb 18 21:19:00 so it's a syntax problem? Feb 18 21:19:16 dromede: paste your compile log somewhere Feb 18 21:20:21 yes, fix worked :) Feb 18 21:21:32 http://pastebin.com/m6ab033b3 Feb 18 21:21:46 Tartarus: I have a fix for this dependency problem Feb 18 21:21:48 in eglibc Feb 18 21:21:57 I *think* it works Feb 18 21:22:45 the rebuild one? Feb 18 21:23:03 yes Feb 18 21:23:10 ah Feb 18 21:23:39 i'll push this compat fix in a sec Feb 18 21:25:17 anyways, i compiled popt by hand for now and my build continues Feb 18 21:26:04 dromede: if you have time then we can fix it Feb 18 21:26:49 i have time Feb 18 21:26:53 found something? Feb 18 21:30:35 khem, OK, pushed Feb 18 21:30:42 03Tom Rini  07org.openembedded.dev * rd7a5542cad 10openembedded.git/packages/eglibc/ (eglibc-package.bbclass eglibc_2.9.bb eglibc_svn.bb): Feb 18 21:30:42 eglibc: Rework the glibc compat packages logic again. Feb 18 21:30:42 We needed to use legitimize_package_name on our locale string. So Feb 18 21:30:42 we just take the eglibc package name that's been legitimized and change Feb 18 21:30:42 it from 'eglibc...' to 'glibc...' Feb 18 21:30:47 03Tom Rini  07org.openembedded.dev * r1e139dea63 10openembedded.git/packages/apt/ (5 files in 2 dirs): apt: Add 0.7.20.2 (current on debian mirrors) Feb 18 21:30:50 03Tom Rini  07org.openembedded.dev * rbc035e162e 10openembedded.git/packages/dpkg/ (3 files in 2 dirs): dpkg: Add 1.14.25 (current on mirrors) Feb 18 21:30:53 03Tom Rini  07org.openembedded.dev * r4e49aa53f5 10openembedded.git/conf/checksums.ini: checksums: Add apt-0.7.20.2, dpkg-1.14.25 Feb 18 21:35:22 dromede: I am able to reproduce your problem Feb 18 21:35:30 I will see if I can do something Feb 18 21:36:56 so is it a bug or is my config wrong? Feb 18 21:37:05 seems a bug Feb 18 21:37:20 can you tell me what the problem is? Feb 18 21:37:39 I have to look a bit deep Feb 18 21:37:51 oh :/ Feb 18 21:38:43 i have another problem, i cant download some of the source packages Feb 18 21:38:52 which ones Feb 18 21:39:03 dromede, DISTRO=angstrom-2008.1 or other? Feb 18 21:39:10 so my course of action is to find a mirror and put the info in my local checksums.ini Feb 18 21:39:11 angstrom has a mirror that helps with these problems Feb 18 21:39:24 is that a good way of dealing with this problem? Feb 18 21:39:39 tartarus: DISTRO=minimal-uclibc Feb 18 21:39:41 you don't have to update checksums.ini really Feb 18 21:39:44 Ah Feb 18 21:39:45 ok, sec.. Feb 18 21:40:11 in your local.conf add Feb 18 21:40:16 #use angstrom source mirrors to combat 404s Feb 18 21:40:17 INHERIT += "angstrom-mirrors" Feb 18 21:40:22 that will help with many of the problems Feb 18 21:40:30 thanks! Feb 18 21:40:36 for others, if you download into you DL_DIR you don't need to update the checksums.ini stuff Feb 18 21:40:41 it'll just check your dl vs that Feb 18 21:41:08 but if i dont update checksums.ini with md5 and sha256 it wont work Feb 18 21:41:20 it will, if you don't change the SRC_URI Feb 18 21:41:35 the checksum is vs the file it finds in your DL_DIR Feb 18 21:42:07 ok, i'll try that Feb 18 21:42:15 The way it kinda works is... Feb 18 21:42:39 bitbake uses the SRC_URI as a key to look into checksums.ini to find the md5/sha256 to use to check what it finds in DL_DIR Feb 18 21:43:00 doesn't matter where what's in DL_DIR comes from, so long as the checksum is right Feb 18 21:44:31 oh, i get it Feb 18 21:46:53 * mwester has wondered why the SRC_URI is used, instead of the filename, since there can only be a single filename in the DL_DIR anyway... Feb 18 21:48:30 ok, it works! Feb 18 21:48:32 I suppose there is the pyscopathic case where two project use the same file name Feb 18 21:48:38 so you need the url to go unique Feb 18 21:50:32 well, md5's are unique too aren't they? Feb 18 21:50:46 well, they are supposed to be :) Feb 18 21:51:02 why don't they just use md5 to identify source packages? Feb 18 21:51:13 but the point is, we are better off using the entire source URI as the index Feb 18 21:51:35 we want to make sure the suorce package does not get changed, without the file name changing Feb 18 21:52:41 why do we need complete URI for that Feb 18 21:54:25 crofton already told you, for the case where multiple packages have urls with the same file name Feb 18 21:54:34 hmmm Feb 18 21:54:58 is there such a case :) Feb 18 21:56:35 easily possible, consider patches for example Feb 18 21:58:12 kergoth, that would be an insane thing to do Feb 18 21:58:15 but Feb 18 21:58:19 people are crazy Feb 18 21:58:39 i don't see what's insane about it. consider a dir upstream with a set of quilt patches. consider one of them called version.patch or something Feb 18 21:58:54 at any rate.. better safe than sorry Feb 18 21:58:54 but we don't have checksums for patches :) Feb 18 21:58:58 right Feb 18 21:59:15 well, we store patches in the metadata now Feb 18 21:59:15 what do you mean? patches with http urls aren't checked? Feb 18 21:59:21 hmm Feb 18 21:59:27 I'm not sure about that Feb 18 21:59:40 i was going to say, there's no reason a package can't use a patch provided by upstream rather than pulling it local, if we dont need to modify it Feb 18 21:59:43 we would need to check Feb 18 21:59:45 what's a good example.. Feb 18 21:59:46 bash Feb 18 21:59:47 Crofton|work: not all patches are in metadata Feb 18 21:59:49 bash has a post release patchset. Feb 18 22:00:13 Crofton|work: some patches are outside of metadata Feb 18 22:00:14 busybox too Feb 18 22:01:07 is anyone familiar with the status of pxafb in recent kernels? Feb 18 22:12:32 Ug Feb 18 22:12:40 I don't believe XorA that dpkg works right now :) Feb 18 22:13:18 * Tartarus hacks a few things quick, wrt dpkg-native not placing perl stuff where apt-native expects and installing dpkg-architecture using host perl not installed perl Feb 18 22:14:58 eve... Feb 18 22:33:22 hmm, is there a sound logging utility that can trigger on a range of frequencies? Feb 18 22:37:06 never heard of such a thing Feb 18 22:37:20 but i bet you could whip something up in puredata Feb 18 22:37:36 why do you need that for if i may ask? Feb 18 22:37:53 i have a noise pollition issue. Feb 18 22:38:05 oops pollution Feb 18 22:39:11 i will investivate this thing though thx.. Feb 18 22:39:58 instead of manually logging incidents it makes sense to get a spare board to do it for me. Feb 18 22:40:30 so you need something that would identify the frequencies you consider noise? Feb 18 22:41:11 the issue is a bunch of chavs built a hot tub in an out house, it has a pump that vibrates my house intermittantly 24/7 Feb 18 22:41:46 its a good job firearms are generally not allowed here :) Feb 18 22:42:03 anyway, it kicks out the same low frequency Feb 18 22:42:20 i need to log it for environmental health officers Feb 18 22:42:56 well, a microphone and a spectrum analyzer would do i think Feb 18 22:43:14 cool. i think the firearms would work better :) Feb 18 22:43:52 it might, but you risk getting into trouble with the law so... :) Feb 18 22:43:59 or you could sabotage the pump Feb 18 22:44:11 * kergoth ponders ways to make it so changing a config file doesn't result in reparsing the recipes Feb 18 22:44:29 incremental parsing :) Feb 18 22:44:39 oh do think of something, it's annoying as hell :) Feb 18 22:44:55 ahh, i live near them so sabotage/gunning them down probably wont do me much good. Feb 18 22:45:01 :) Feb 18 22:45:19 plus, the only gun i have is a classic 80s potato gun. Feb 18 22:45:28 that hurts too Feb 18 22:47:43 ...and no ammo :) Feb 18 22:50:13 dear god, building task-base takes a long time... Feb 18 22:53:35 khem: been playing around with possible alternative apis for bb.data, since its so fugly.. http://kergoth.pastey.net/108517 still missing quite a bit, just a proof of concept Feb 18 22:55:09 * kergoth scratches head Feb 18 22:55:31 * khem looks Feb 18 22:57:02 it tracks variable references and dirties referrers on change and such. doesn't do the whole cowdict thing to fall back to the cfg data yet, and isn't associated with a file to monitor yet.. mainly its just a way to get me back in the game when it comes to my python.... heh Feb 18 23:05:41 Hmm Feb 18 23:19:33 i'm going off, thanks for your help people, bye! Feb 18 23:22:20 kergoth: hey, let me comment in a second Feb 18 23:23:36 np, no rush, just having fun playing around Feb 19 00:01:54 * kergoth wonders if there's any variables in a .conf that reference variables defined in the recipe/classes, or only the other way around Feb 19 00:23:47 03Khem Raj  07org.openembedded.dev * re4eb9aab1d 10openembedded.git/packages/popt/ (popt-1.7/mkinstalldirs.patch popt_1.7.bb): popt-1.7: Fix build with new libtool Feb 19 01:01:18 03Michael 'Mickey' Lauer  07org.openembedded.dev * r0aa925443d 10openembedded.git/conf/machine/ (7 files in 2 dirs): htc*.conf: factor out common include file for the msm7xxxx devices Feb 19 01:31:45 is there a upnp aware media player client for OE? Feb 19 01:44:49 Soopaman: http://djmount.sourceforge.net/ Feb 19 01:45:02 http://www.geexbox.org/ Feb 19 01:47:37 this dj mount is certainly interesting Feb 19 01:47:38 Soopaman: I dont think there is something there already in OE you might have to add recipe Feb 19 01:48:06 are there any online tutorials for putting together a recipe? Feb 19 01:48:27 I would advise to look at wiki Feb 19 01:49:26 cool Feb 19 01:49:51 so i figured out why i kept hitting my file limit the first times compiling OE Feb 19 01:50:23 ubuntu 8.10's implementation of the pulse audio module Feb 19 01:56:06 what was it Feb 19 01:56:13 pulse audio module ? Feb 19 01:56:25 hmmm how could it be related Feb 19 02:08:41 Hmm, who knows the deb classes well? I'm trying to figure out if we (and I think this is the case) build apt / dpkg to default to checking for conf files within ${STAGING_DIR} or use -c ${STAGING_DIR}/... to make sure our staging config file is read Feb 19 02:08:42 03Michael 'Mickey' Lauer  07org.openembedded.dev * r4c69f9c145 10openembedded.git/packages/linux/linux-msm7xxxx_git.bb: linux-msm7xxxx: bump kernel for raphael and diamond Feb 19 02:09:08 I'm pretty tempted right now to assume-provided apt/dpkg and make sure we enforce passing an arg to use the right conf file **** ENDING LOGGING AT Thu Feb 19 02:59:57 2009