**** BEGIN LOGGING AT Mon Jul 27 06:00:14 2009 **** ENDING LOGGING AT Mon Jul 27 06:02:16 2009 **** BEGIN LOGGING AT Mon Jul 27 06:03:37 2009 Jul 27 06:44:27 good morning Jul 27 07:05:00 morning Jul 27 07:08:12 how does one ASSUME_PROVIDED a kernel? Jul 27 07:08:51 i've tried virtual/kernel, linux, kernel and the linux_2.6.30.bb merrily tries to build regardless Jul 27 07:09:29 someone else asked this a while ago http://projects.linuxtogo.org/pipermail/openembedded-users/2007-February/000064.html but got no response Jul 27 07:13:27 grg: where did you try to add ASSUME_PROVIDED += ? Jul 27 07:13:36 in local.conf Jul 27 07:15:14 grg: hmmm looks like a not working option then Jul 27 07:16:30 so i have to create a working kernel recipe before i can build an OE image? Jul 27 07:17:57 03Graeme Gregory  07org.openembedded.dev * r721fc94773 10openembedded.git/ (conf/checksums.ini recipes/transmission/transmission_1.73.bb): transmission_1.73.bb : add new version of transmission Jul 27 07:17:58 03Graeme Gregory  07org.openembedded.dev * rf475c3fabb 10openembedded.git/recipes/powertop/powertop_1.11.bb: powertop_1.11.bb : add omapzoom2 support for building Jul 27 07:17:58 03Graeme Gregory  07org.openembedded.dev * rcadd730e45 10openembedded.git/recipes/libtool/ (6 files in 2 dirs): Jul 27 07:17:59 libtool_2.2.6a : add newest version of libtool disabled by default Jul 27 07:18:01 converted to use ${BPN} in the packages compared to old libtool Jul 27 07:20:34 grg: maybe, I suggest you to wait a couple of hours and ask again to someone else, maybe I'm wrong Jul 27 07:20:51 mckoan, i already asked a couple of days ago :( Jul 27 07:21:30 maybe i'll have more luck with the mailing list Jul 27 07:22:48 grg: yes please try on ML Jul 27 07:23:28 assume provided certainly works Jul 27 07:25:44 is this edit valid? http://wiki.openembedded.net/index.php?title=Getting_started&diff=1524&oldid=1511 Jul 27 08:02:27 good morning Jul 27 08:24:01 hi florian Jul 27 08:28:29 good morning Jul 27 08:32:00 hi florian Jul 27 08:33:08 hi rkirti Jul 27 08:38:09 florian: I need to dump 4 bytes from nand using klibc. The old code I have uses MEMREADLADDR ioctl (and includes linux/mtd/mtd.h). Now, I see there have been tons of changes wrt MTD. I'm a bit confused: glibc stages mtd-user.h while klibc has the regular mtd.h but lacks MEMREADLADDR. Should I use the new MEMREADOOB? Jul 27 08:39:07 ah, btw, wouldn't klibc's dd be able to dump the bytes? Jul 27 08:40:48 * florian is clueless... no idea about klibc yet Jul 27 08:49:49 hm... dd won't get the ECC information...one would need MEMREADOOB it seems Jul 27 08:52:35 but...if there are bad blocks there, the stupid bootloader would have failed..that's why there is a second copy of the params I'm looking for Jul 27 08:56:02 morning Jul 27 08:57:30 hi hrw Jul 27 09:03:59 florian: good morning Jul 27 09:04:37 hrw: happy birthday btw. :-) Jul 27 09:04:41 * florian knows he is late Jul 27 09:04:45 hi pb_ Jul 27 09:05:02 florian: thx Jul 27 09:05:07 morning. Jul 27 09:07:34 03Koen Kooi  07org.openembedded.dev * r179219dcc7 10openembedded.git/recipes/kexecboot/kexecboot_git.bb: kexecboot: switch git version to a stable cfgfiles SRCREV and make it the default Jul 27 09:07:34 03Koen Kooi  07org.openembedded.dev * r6eea2d81d1 10openembedded.git/: Merge branch 'org.openembedded.dev' of git@git.openembedded.org:openembedded into org.openembedded.dev Jul 27 09:08:58 hrw: it's your birthday? happy birthday then!! :) Jul 27 09:11:52 methril|work: was. yesterday. Jul 27 09:12:01 hrw: belated happy birthday :-) Jul 27 09:21:17 hi * Jul 27 09:21:36 Hrww: also from here hppy bdy :) Jul 27 09:36:03 HRW: Congrats! Jul 27 10:23:53 Hi, I'm getting this: "Multiple .bb files are due to be built which each provide virtual/kernel". I have created a new machine spec "xenclient-dom0.conf" and it has PREFERRED_PROVIDER_virtual/kernel = "linux-xenclient", also I created linux-xenclient-dom0_2.6.27.bb which has the git tree it needs to get the kernel from Jul 27 11:08:54 re Jul 27 11:11:40 I have a problem with dependencies (I think), how can I find out why it's trying to pull in the wrong kernel? My scenario is this: Jul 27 11:11:47 "Multiple .bb files are due to be built which each provide virtual/kernel". I have created a new machine spec "xenclient-dom0.conf" and it has PREFERRED_PROVIDER_virtual/kernel = "linux-xenclient", also I created linux-xenclient-dom0_2.6.27.bb which has the git tree it needs to get the kernel from Jul 27 11:12:37 but instead of using just my kernel it wants to pull in recipes/linux/linux-epia_2.6.19.2.bb as well Jul 27 11:14:03 th1_: so you have something inherited somewhere ;-). Want to pastebin your machine file? Jul 27 11:14:16 ok I'll pastebin both the kernel and machine config. Jul 27 11:15:48 http://pastebin.ca/1508669 Jul 27 11:16:15 the pastebin is down... Jul 27 11:16:35 http://pastebin.com/m268ab93d Jul 27 11:17:47 I based my machine spec out of one of the generic i686 ones Jul 27 11:21:11 th1_: minor things, you have DEFAULT_PREFERENCE = "-1" so it will look for another other suitable kernel 1st ;-), it's only one machine so no need for the brackets. Jul 27 11:21:49 I thought it would use 28 for that machine? Jul 27 11:22:00 the reason I put it there was because I don't want that kernel to be used for anything else Jul 27 11:22:30 if you have it just used for one machine type it will only be used by that. Jul 27 11:22:52 so should I eliminate the preferences lines altogether? Jul 27 11:23:41 You have other i686 kernels that are valid I think, also, why did you hack off a davinci config? You missing a few things such as any MACHINE_KERNEL_PR Jul 27 11:24:13 ok Jul 27 11:24:17 I hacked off one that used git ;) Jul 27 11:25:13 MACHINE_KERNEL_PR = "r1" in your machine file. Jul 27 11:25:53 th1_: your also lacking a specific SRCREV, that's not all that clean really. Jul 27 11:26:15 it's an internal repository only used for this project, always wanto to build from master Jul 27 11:27:19 are you sure? so you can't be 100% your kernel matches the rest of your system every time? Jul 27 11:27:27 Such as modules etc. Jul 27 11:28:08 th1_: I did miss your DEFAULT_PREFERENCE_xenclient-dom0 = "28", sorry about that, looks ok. Jul 27 11:28:11 if we make updates to the git repo we want it in the rootfs build Jul 27 11:28:23 well Jul 27 11:28:32 since it only supports that machine I guess it's unnecessary anyawy Jul 27 11:29:02 I would have thought the 'PREFERRED_PROVIDER_virtual/kernel = "linux-xenclient"' stanza would be enough Jul 27 11:29:07 But you won't get it if you don't bump MACHINE_KERNEL_PR as it wont pick up the need to rebuild something you have not told it about ;-) Jul 27 11:29:32 that's ok Jul 27 11:29:46 (for now) :) Jul 27 11:30:01 so if you try bitbake virtual/kernel what gets built? Jul 27 11:31:05 http://pastebin.com/m103f959 Jul 27 11:32:14 th1_: linux-xenclient-dom0 <> linux-xenclient ;-) Jul 27 11:32:55 ty :) Jul 27 11:35:02 that's got it, now it looks like it's not trying to build the other one any more Jul 27 11:36:16 now on to my next question .. I know the Fetchers are in Python but can I make one that's in Shell script? what I need is a git+guilt fetcher that will fetch my repos e.g. git://gitserver/kernel.git and then also fetch git://gitserver/kernel-pq.git which is a patch queue to go on top Jul 27 11:36:34 in my buildroot config I have a git-fetcher script to do just that but how can I add a hook into oe for it? Jul 27 11:49:12 th1_: why not just give two lines in SRC_URI for that? Jul 27 11:49:40 it will fetch one git, then second. you can then use patches from second one in SRC_URI too Jul 27 11:51:39 it will need to use guilt to apply the patches Jul 27 11:52:01 quilt is not enough? Jul 27 11:52:28 anyone here to review a gcc*.inc change? Jul 27 11:54:00 basically it needs to read the series file in pq and apply each of the patches listed, if guilt can do that then it's probably enough Jul 27 11:54:26 iirc guilt was 'quilt on top of git' Jul 27 11:54:46 yes Jul 27 11:56:00 I think it will break when we use branches since the pq tracks patches for different branches in different directories Jul 27 11:56:27 [ 3388.241923] BUG: Bad page state in process cc1 pfn:00006 Jul 27 11:56:29 shit Jul 27 11:59:02 problem is fixed in kernel git but to build it I need fix Jul 27 12:02:37 hi Cliff Jul 27 12:02:49 hrw: hello Marcin -- happy birthday :-) Jul 27 12:02:56 thx Jul 27 12:03:19 hrw: oh, congrats Jul 27 12:04:06 adding 4GB ram do system shown new kernel bug :( Jul 27 12:50:18 re Jul 27 12:56:49 hi Jul 27 12:59:59 memtest86 takes eons... Jul 27 13:31:43 morning Jul 27 13:31:47 hi chouimat Jul 27 14:27:56 re Jul 27 14:28:01 re Jul 27 14:59:48 morning all Jul 27 15:06:56 hey RP Jul 27 15:16:32 03Klaus Kurzmann  07shr/import * r8d92b97688 10openembedded.git/ (3 files in 2 dirs): Jul 27 15:16:32 new recipe python-phoneutils Jul 27 15:16:32 Signed-off-by: Klaus Kurzmann Jul 27 15:18:17 RP: hey! Jul 27 15:56:40 03Klaus Kurzmann  07shr/import * re22a881bd3 10openembedded.git/recipes/python/python-phoneutils_git.bb: Jul 27 15:56:40 python-phoneutils: use setuptools and not distutils Jul 27 15:56:40 Signed-off-by: Klaus Kurzmann Jul 27 16:18:03 03Klaus Kurzmann  07shr/import * r6331e4bfb5 10openembedded.git/recipes/pyphonelog/pyphonelog_git.bb: Jul 27 16:18:03 pyphonelog: add python-phoneutils as runtime dep Jul 27 16:18:03 Signed-off-by: Klaus Kurzmann Jul 27 18:52:24 03Stanislav Brabec  07org.openembedded.dev * r2e0fa4918d 10openembedded.git/recipes/libtool/ (libtool_2.2.4.bb libtool_2.2.6a.bb): libtool: Fixed paths to grep and sed. Jul 27 20:07:24 re florian Jul 27 20:14:45 re Jul 27 20:15:01 yes! 8gb ram passed memtest Jul 27 20:29:51 nice Jul 27 20:39:25 I am thinking of TLS and non TLS library problems in the build currently the autoconf tests are not sufficient in the cross build world to determine the TLS support on the target system. Jul 27 20:39:55 so I think one option we have is to put it in the bitbake.conf file Jul 27 20:40:44 bye Jul 27 21:36:57 I'm having a problem with "bitbake xorg-image" it can't find glibc-localedata-i18n, and when I try to bitbake that it says nothing PROVIDEs it Jul 27 21:37:04 any clues? Jul 27 21:40:43 th1: this should come from glibc Jul 27 21:41:04 http://pastebin.ca/1509371 is my machine config Jul 27 21:41:12 I did some stuff with the locales i.e. only enabled en_US Jul 27 21:41:15 could that be it? Jul 27 21:45:25 alternatively is there a way to say "just ignore that package and try to build the image anyway" Jul 27 22:16:56 argh Jul 27 22:17:27 I got rid of it from rootfs_ipk.bbclass and now it complains that there are 2 xservers Jul 27 22:18:22 I thought it was buildroot that was dependency hell but hey :) Jul 27 22:18:35 at least this prints an error instead of creating a crap build **** ENDING LOGGING AT Tue Jul 28 02:59:57 2009