**** BEGIN LOGGING AT Wed May 06 02:59:57 2009 May 06 05:36:55 03Angus Ainslie  07fso/milestone5.5 * rd0dc8de4d5 10openembedded.git/recipes/openmoko-projects/paroli_git.bb: paroli : ship dbus startup file May 06 06:09:55 how do i add a patch to a bitbake thing? May 06 06:10:39 the manual/docs seem to dismiss this section with "[to be done]" May 06 06:12:07 a package is broken for me. i could fix it, if only i knew how to get the build infrastructure to apply my patch May 06 06:14:20 * gumleef twiddles his thumbs May 06 06:40:13 good morning May 06 06:56:49 good morning May 06 10:07:02 florian_kc: good morning May 06 10:12:14 good morning May 06 10:18:21 morning all May 06 10:22:41 hi May 06 10:28:05 gm May 06 10:28:13 is there a way to rebuild all non-native packages with bitbake? May 06 10:28:27 and non-cross May 06 10:28:40 i.e. all TARGET_ARCH and MACHINE_ARCH packages May 06 11:08:24 woglinde: btw I asked that guy who also got error while compiling qt/embedded-4.5.1, and he used generic x86_64 ubuntu gcc 4.3 system. there is no embedded specific stuff in there May 06 11:08:44 woglinde: and have you tried to build qt4-embedded-4.5.1 with OE? May 06 11:11:58 booxter ups sorry May 06 11:12:02 didnt find the time May 06 11:12:06 hi rp May 06 11:12:14 woglinde: no problem :) May 06 11:12:43 maybee this evening May 06 11:12:47 hm or I try it here May 06 11:12:52 thats a good ide May 06 11:12:53 a May 06 11:15:01 jo mickeyl May 06 11:15:07 hi May 06 11:19:49 good morning May 06 11:20:07 hi methril May 06 11:20:33 03Robert Schuster  07org.openembedded.dev * rc57486c9bb 10openembedded.git/conf/checksums.ini: checksums.ini: New cmake 2.6.4 entry, removed unused cmake entries. May 06 11:20:43 03Robert Schuster  07org.openembedded.dev * rbdeb73a931 10openembedded.git/recipes/cmake/cmake-native_2.6.4.bb: cmake-native 2.6.4: New recipe. May 06 11:47:20 Can I use an console-image as ramdisk? (I'm trying, using a cpio.gz, but it doesn't seem to work) May 06 11:47:58 (I have made an u-boot image from the cpio.gz with the mkimage tool, manually) May 06 11:53:25 tsjsieb: what's the problem with it? May 06 11:55:09 that I don't know If I'm trying to create the ramdisk in te right way May 06 11:55:31 cpio.gz should work fine as an initrd May 06 11:55:42 with a recent kernel and approppriate kernel config May 06 11:56:33 I expect that you pass the initrd address correctly to u-boot bootm May 06 11:58:20 I'm not sure, but I have a working ramdisk example for my machine May 06 11:58:55 so I guess that kernel is usable and correct configured May 06 11:59:35 but when i try to load that kernel with my own 'cpio.gz in uboot mkimage file' May 06 12:00:17 uboot gives me a 'Verify Checksum.... Bad CRC', followed by a reset May 06 12:01:14 (wich it doesn't do, when i 'iminfo' the same adress, where I loaded the 'cpio.gz containing image' May 06 12:05:36 I also created a kernel, which does boot with the 'example ramdisk' I have, that's why I suspect there is something wrong with my own ramdisk May 06 12:14:01 Bad CRC doesn't have anything to do with the ramdisk contents May 06 12:14:43 you probably have not transferred the ramdisk to the device correctly or are trying to load it from a wrong address May 06 12:15:26 thebohemian: Good to see you were not eaten by the magic OE monster ;-) May 06 12:15:40 ibot: seen likewise? May 06 12:15:41 likewise was last seen on IRC in channel #oe, 4d 18h 19m 2s ago, saying: 'hello all'. May 06 12:18:41 hi ant May 06 12:19:00 hey woglinde May 06 12:22:05 florian: :) May 06 12:24:33 tsjsieb: if you are building an image generating a final ramdisk image I would be couruis to know which options are required in conf to do that, could you give me a clue? May 06 12:29:00 I don't think I understand you completly, but I'm trying to create a ramdisk to teskt if my created rootfs is ok, without using the NAND flash memory on my device. (as I'm suffering some problems there) May 06 12:29:28 ok May 06 12:29:59 and unfortunatly I'm not able to use NFS as my ethernet card doesn't get a valid MAC address when kernel boots May 06 12:30:11 so you tftp kernel.img $kerneladdr; tftp ramdisk.img $ramdiskaddr; bootm $kerneladdr $ramdiskaddr May 06 12:30:21 tsjsieb: my question is whether you are getting the ramdisk image automatically with a bitbake process or are you doing it by hand May 06 12:30:23 right? May 06 12:31:02 akheron: thats indeed the way I'm doing that May 06 12:31:11 thebohemian: I tried to build x11-gpe-java-image but it failed on classpath May 06 12:32:52 mckoan: I added cpio.gz to IMAGE_FSTYPES in my local.conf and I manually make a u-boot image, using the u-boot mkimage tool May 06 12:33:44 hi ant_work, woglinde, mickeyl May 06 12:36:13 akheron: but I changed the adresses a little, compared to the working example, because my cpio.gz is bigger then the expamle ramdisk.img May 06 12:37:06 and as they use 0x80a00000 for kernel and 0x80600000 I thought that my bigger cpio.gz might overlap over the kernel image May 06 12:37:27 (0x80600000 for the ramdisk.img) May 06 12:39:16 ok May 06 12:39:50 morning May 06 12:39:56 remember that kernel gets moved to another location, too, and that location should not overlap with your ramdisk May 06 12:40:54 hi hrw May 06 12:41:44 tsjsieb: I haven't noticed cpio.gz option, thank you May 06 12:41:47 pb_: I just tried gdb. Running in gb the same application dies from a segfault. May 06 12:42:44 tsjsieb: I always used ramdisk before using OE, what is your problem? which architecture? May 06 12:43:09 akheron: so both the kernel as the ramdisk get moved after loaded to the 1st adress (manually with tftp 'adress1' file) to the a second adress? May 06 12:46:53 mckoan: ARM (Compulab CM-X300) the problem is actually that I try to check if my rootfs is o.k., but I'm not able to use the NAND flash on the device, as it seems to suffer from a NAND flash driver bug, so now I'm trying to use the same rootfs als ramdisk May 06 12:47:13 jo raster May 06 12:47:33 woglinde: yooooooow! May 06 12:47:41 is it just me or has oe pretty much fallen of the rails? May 06 12:47:59 raster hm in which case? May 06 12:48:06 the only target angstrom+oe (or openmoko+oe for gta02) buils for is... overo (or i guess beagle) May 06 12:48:14 everything else is mired in gnu hash problems May 06 12:48:17 gcc segvs May 06 12:48:29 god knos how many other issues May 06 12:48:37 this happens May 06 12:48:40 raster, during builds? May 06 12:48:45 Crofton|work: yup May 06 12:48:46 if you leave normal ways May 06 12:48:52 which targets are you having problems with? May 06 12:48:54 and provide stuff for arm thumb May 06 12:49:00 i used to bea ble to build for about 5 or 6 targets May 06 12:49:02 now just 1 May 06 12:49:30 the gnu hash should be fixed May 06 12:49:37 Crofton|work: palmt650, nokia800, rokre6, om-gta02, c760 May 06 12:49:57 woglinde: i still have moe gnu has patches pending which i wont push until i can verify a full image build... which i cant May 06 12:50:09 raster where it fails with n800? May 06 12:50:20 I would have thought people cared about the gta02 May 06 12:50:22 for n800 you only can use chinook-compat May 06 12:50:47 woglinde: i forget. i'venuked my build dirs for all but omap3/overo May 06 12:51:16 but we/maemo has a gsoc student who tries to get a disbu ready which provides an image to falsh May 06 12:51:17 Crofton|work: not in oe. :( May 06 12:51:24 woglinde: no shit.. angstrom can be built for n8x0 too May 06 12:51:39 woglinde: i USED to be able to build angstrom for everyething May 06 12:51:39 hrw but its not compat May 06 12:51:41 hrw, what is the url for your buildbot? May 06 12:51:52 but of lae almsot every machine target except omap3 has barfed May 06 12:52:57 i am happy to help fix bugs where i can but when you get a mountain of them that requires 10+ minute test cycles May 06 12:53:05 Crofton|work: http://home.haerwu.biz:8010/ May 06 12:53:07 i kind of throw in the towel :( May 06 12:53:14 i was just wondering if it is just me May 06 12:53:18 or others see it too? May 06 12:53:39 http://home.haerwu.biz:8010/waterfall May 06 12:53:53 raster seems most people only building for beagleboard May 06 12:53:53 Crofton|work: yes May 06 12:53:54 raster, I suspect you are building for machines that are not actively maintained May 06 12:54:02 woglinde: thats what i was wondering May 06 12:54:08 I admit we need a way to document which machines have active maintinaers May 06 12:54:16 islike that the only target anyone builds for anymore? as its the only one that works :) May 06 12:54:22 Crofton|work: indeed i am May 06 12:54:25 raster, see hrw's buildbot May 06 12:54:30 Crofton|work: i am trying a token effort at maintaing them May 06 12:54:32 but its a bit much May 06 12:54:33 :( May 06 12:55:06 raster, can you post a list of known failing builds to the list May 06 12:55:29 I can burn some spare cycles running the builds and see if I can do anything May 06 12:55:41 sure! May 06 12:55:43 I'll need MACHINE, DISTRO, and image May 06 12:55:43 can do May 06 12:55:44 raster: builds for c7x0 (console-image, opie-image, x11-image) are ok May 06 12:55:54 x11-gpe-image perhaps May 06 12:56:02 x11-gpe-java-image no May 06 12:56:08 * hrw needs to patch x11 components ;( May 06 12:56:16 eh..hoe many similar images...who maintains these? May 06 12:56:16 (EE) FBDEV(0): FBIOPUT_VSCREENINFO succeeded but modified mode May 06 12:56:16 (EE) FBDEV(0): mode initialization failed May 06 12:56:22 i was just wondering if my uspicion that all thes targets are simply not maintained and thus breakages occure and are never found/fixed until i cycle arounda month later to try another build May 06 12:56:23 :) May 06 12:56:31 yeah May 06 12:56:40 ant_work: maybe its because i also build full evel tools May 06 12:56:44 like gcc for the target May 06 12:56:45 I'd like to ge this info documented so people know what to expect May 06 12:56:46 binutils May 06 12:56:48 gdb etc. May 06 12:57:00 most targets never build it May 06 12:57:01 raster: I didn't dare :) May 06 12:57:04 hahaha May 06 12:57:09 maybe thats why i get hit by it May 06 12:57:15 yeah May 06 12:57:15 BUT it id work a few months back nicely May 06 12:57:21 I was hit by the installer/strip issue May 06 12:57:35 still strange..it seems the host strip is called btw May 06 12:57:41 so its gonedownhill and i suspect oe needs a bit of spring cleaning just to get the housein order again after some ugly changes May 06 12:58:13 well, first we need to work out what is broken May 06 12:58:29 as you note, yuo are building some things that most people do not build May 06 12:58:30 raster because it worked dont establish that it was good May 06 12:58:41 I think that I will add some more targets in my buildbot - I am offline during long weekend (Thu->Mon) May 06 12:58:50 florian: that's weird. where does it segfault? May 06 12:58:50 Crofton|work: still disturbing how so many images are failing May 06 12:59:02 woglinde: hahahaha May 06 12:59:11 raster ;) May 06 12:59:21 well, I think raster is also building some stuff that most do not build May 06 12:59:21 Crofton|work: well it would be good togte moe things to build May 06 12:59:34 so it is possible there are only a few broken things May 06 12:59:39 agreed May 06 12:59:40 i am keen to try my hand at hacking up my g1 and re-using the android kerneland just replacing userspace with oe May 06 13:00:08 Crofton|work: they build for omap :) May 06 13:00:17 lets see May 06 13:00:24 i'm moving my oe tree to my xternal usb drive May 06 13:00:29 kind of out of space on internal May 06 13:00:35 damn these ssd's May 06 13:00:37 too small May 06 13:00:52 raster and mostly to slow May 06 13:01:23 woglinde: ssd's? May 06 13:01:25 hell no May 06 13:01:29 this baby vooms May 06 13:01:39 like a motherfucker May 06 13:01:42 intel one? May 06 13:01:45 boot time is gorgeously low May 06 13:01:52 e17 comes up in < 1 second May 06 13:01:54 we have here on an atom board May 06 13:01:57 yeah the x25m May 06 13:02:01 but maybee its the atom May 06 13:02:27 my previous hdd pulled a good 50m/sc sequential io May 06 13:02:30 it was a slowpoke May 06 13:02:38 but this x25 beats the pants offit May 06 13:03:40 Timing buffered disk reads: 674 MB in 3.00 seconds = 224.61 MB/sec May 06 13:03:55 also considering seek time is now basically non-existant May 06 13:03:56 :) May 06 13:04:24 hmmm May 06 13:04:33 what happened to FILESPATH? May 06 13:04:48 Crofton|work: kergoth deprecated it? May 06 13:05:03 does putting things in u-boot-git/beagleboard no linger work? May 06 13:05:37 should iirc May 06 13:05:41 hmm May 06 13:05:53 but u-boot-git needs to be added into FILESsomething May 06 13:07:02 hmmm May 06 13:07:24 I wonder how people have been moving stuff into files because of this? May 06 13:08:16 FILESPATHPKG += "u-boot-git" - try May 06 13:08:31 yeah May 06 13:08:56 collecting things in files, especially version specifc patches is evil May 06 13:12:32 <_cpo_> I upgraded from the old stable release to stable/2009. i always used in my env.sh file to set the OE base path and reference to this env variable in my local.conf May 06 13:12:53 <_cpo_> but this isnt working anymore. anyone can point me into the direction of this change in bitbake/oe? May 06 13:13:24 <_cpo_> another think is that when i refer to ${HOME} in a recipe, ${HOME} wont be expanded and is just empty May 06 13:15:53 the extra white list var May 06 13:15:59 hang ong May 06 13:17:07 _cpo_, http://wiki.openembedded.net/index.php/Advanced_configuration#Whitelist_Environment_Variables May 06 13:17:25 <_cpo_> thx May 06 13:18:39 guys: is there a device in OE which has more then 1 framebuffer? (other then BUG) May 06 13:24:15 hrw: never seen it, maybe a nintendo DS :-D May 06 13:24:30 mckoan: "in OE" May 06 13:24:45 :-D May 06 13:24:52 mckoan: BUG has 3 framebuffers May 06 13:25:48 hrw: human beings have only 2 eyes :-D May 06 13:25:59 <_cpo_> also i think i found a bug with devshell and newer versions of 'konsole'. the devshell will start console with the -T switch, but this one is no longer present in newer console versions May 06 13:26:36 tsjsieb: you have an idea of your memory map? May 06 13:26:55 tsjsieb: http://www.compulab.co.il/mediawiki/index.php5?title=Getting_started_with_Linux_on_CM-X300 May 06 13:27:28 tsjsieb: seems that linux kernel goes @ 0x80400000 May 06 13:28:17 tsjsieb: add your uImage size to it and the result (plus some space) is the address whete to place your ramdisk May 06 13:29:02 mckoan: thats not the kernel, but an u-boot script, wich contains the next steps to load the kernel and ramdisk May 06 13:30:14 this is the source of that script: http://rafb.net/p/L7OVpx64.html May 06 13:31:38 tsjsieb: is kernel.img a uImage type ? May 06 13:31:50 tsjsieb: what is kernel.img size (hex)? May 06 13:32:41 I think it is yes (and when I replace the kernel.img by a custom uImage, it works) May 06 13:33:24 It seemed a little bit strange to me that the location of the ramdisk in the script is lower than the kernel, but that's just a feeling May 06 13:33:28 hrw, FILESPATHPKG += "u-boot-git" does not seem to help May 06 13:33:49 I am getting somewhat annoyed with stuff like this May 06 13:34:04 rebuild no longer works as expected, python 2.5 going away etc May 06 13:35:24 tsjsieb: in any case you have to be sure that need the kernel and ramdisk size do not overlap May 06 13:36:58 mckoan: my custom uImage is 20208840 bytes (u-boot says) so I guess that's 0x1345CC8 (hex) to add to the 0x80a00000 address of the kernel.img) May 06 13:38:32 ( or should that be multiplied by 8, as the hex might be bits May 06 13:38:51 tsjsieb: fat kernel, and the ramdisk size ? May 06 13:39:31 tsjsieb cant you try to load kernel and initrd seperated May 06 13:40:57 mckoan: ramdisk I try to use is console-image cpio.gz of 7.4MB May 06 13:41:38 woglinde: you mean with u-boot, as in two different "tftp 'address' file" actions? May 06 13:44:46 hrw: fwiw u-boot_git for c7x0 takes the patch May 06 13:46:25 tsjsieb load the kernel to an adress and load the initrd to another not overlaping address May 06 13:47:40 ant_work: my c7x0 is rusting May 06 13:48:11 :) I don't see whj it shouldn't work for beagleboard May 06 13:58:46 woglinde: I think I do that, look in log: http://rafb.net/p/gl1u3C75.html both seem to be o.k when I check them with 'iminfo' on line 39 and 50, but when I 'bootm' the second one fails May 06 14:01:05 hm seems something wrong with your ramdisk May 06 14:02:45 hmm, or maybe the adress I use, I just tested it with the 'example ramdisk.img of compulab' on the same adress, and that fails in the same way May 06 14:03:12 and that ramdisk does work, when I load it to 0x80600000 May 06 14:03:34 and when you load your other ramdisk there? May 06 14:03:45 maybee ist beyound mem limit May 06 14:03:50 for the kernel May 06 14:04:21 its to fat for that place, it will override the kernel, there the kernel is on 0x80a00000 May 06 14:08:51 tsjsieb: Load Address: 00000000 ??? May 06 14:09:19 that it's to fat, might be possible, it's 7.4 MB (Angstrom console image) I'm currently rebuilding it with uclibc May 06 14:10:55 tsjsieb: time ago I found this post: http://lists.denx.de/pipermail/u-boot/2008-February/029672.html May 06 14:11:21 problem was CONFIG_INITRD_TAG May 06 14:11:52 mckoan: it is the same in the example ramdisk.img from compulab, wich does work. I'm not shure, but maybe it's only used when written to flash? May 06 14:13:54 the zaurus nand layout is useing a 7M(sys)/32m(root)/ - rest (home) layout. what image fits into the rootfs ( except the console-image ) ? May 06 14:14:11 pwgen: console-image with uclib May 06 14:14:13 http://ja.pastebin.ca/1173295?srch=noop May 06 14:14:34 was 2.2 mb May 06 14:14:42 and what sould i do if i want some graphical interface ? May 06 14:15:08 03Florian Boor  07org.openembedded.dev * rdc13ea62f2 10openembedded.git/recipes/gpe-conf/gpe-conf_svn.bb: gpe-conf: bump RV in svn recipe May 06 14:15:10 03Florian Boor  07org.openembedded.dev * rb182d05b9b 10openembedded.git/conf/distro/include/sane-srcrevs.inc: sane-srcrevs.inc: update gpe-conf May 06 14:15:17 iirc there are 5.7mb in total if you're using u-boot May 06 14:15:25 pwgen: ^^ May 06 14:16:32 pwgen: ah..sorry..you mean mtd2 May 06 14:17:06 opie-image, x11-image, x11-gpe-image will fit May 06 14:17:24 as i remember the "old" origianl image used something like 50% root/50% home . what was the reason to use 32mb>root/rest>home instead of useing all for root ? May 06 14:17:45 reson is bootloader May 06 14:18:06 Sharp bootloader doesn't allow repartitioning May 06 14:18:15 (ell..almost ;-) May 06 14:19:37 with kexec and uboot i can strip down the system area to approx 2 mb and can define it by kernel cmdline , ( who is using sharp bootloader ? ) May 06 14:19:42 tsjsieb: 7.4 MB ramdisk is not a problem, your problem is not its size May 06 14:20:24 pwgen: yep..you'll loose the 'nand backup' utilites provided by 2nd kernel May 06 14:20:25 tsjsieb: a lready asked you whether you have a device memory map May 06 14:20:40 tsjsieb: ram start and size May 06 14:20:41 pwgen: I never used these May 06 14:20:54 is really a waste of nand May 06 14:21:14 ant_work: nand backup is not in nand May 06 14:21:31 I mean the utility to save your 'home' May 06 14:21:32 mckoan: I don't know, where can I find that? May 06 14:21:37 ant_work: it is in diagnostics which is in prom May 06 14:21:48 tsjsieb: compulab May 06 14:21:52 hrw: there are two copies: one in PROM May 06 14:21:52 i can do nand backup useing the system menu , ( yes that works only if zou have no broken nand ) and providing a backup funtion to kexec will not be a problem May 06 14:22:01 hrw: Fn + D+M May 06 14:22:14 different is D+M May 06 14:22:20 tsjsieb: is hard to grope the way without hardware and without knowing exactly what are you doing there May 06 14:22:52 mckoan: I'm afraid there documentation is not very good May 06 14:23:38 tsjsieb: unfortunately I know I already faced to them May 06 14:26:46 and support / reaction time might be even worse May 06 14:27:31 btw...that patch for detecting UBOOT_LOAD_ADDRESS... May 06 14:27:42 someone tested it..I don't understand all :/ May 06 14:28:18 tsjsieb: let's supposse you have 'only' 64MB ram May 06 14:29:17 Memory: 64MB 64MB = 128MB total May 06 14:29:18 Memory: 122592KB available (3760K code, 492K data, 128K init) May 06 14:30:07 might this be usefull ( that's what linux kernel tells me, and it should be, by cm-x300 type number) May 06 14:30:52 hello May 06 14:31:51 I have compiled opie (bitbake opie-image) , enlarged the resulting opie-image.rootfs.ext2 and launched using http://linuxtogo.org/gowiki/Angstrom/qemu-system-arm parameters . but qemu writes out Bluetooth: RFCOM ver 1.8 and does not boot anything further (ie no graphics). how could i start seeing the opie desktop? May 06 14:32:56 mm I have to leave, thanks for help, will have to search on tomorrow again May 06 14:33:09 goodnight May 06 14:34:48 hello May 06 14:34:57 hi sz May 06 14:34:58 hi szsoftware May 06 14:35:47 I asked it yesterday, but the advice did not work for me. (I'm quite sure, I did it in the wrong way). May 06 14:36:30 How can I tell bitbake on the fly, which DISTRO I wanna bake, w/o defining in local.conf? May 06 14:36:57 define it in conf/auto.conf :D May 06 14:37:37 I wanna be able to bitbake in several consoles different distro/images May 06 14:37:45 the use different build dirs May 06 14:37:48 so do few build dirs May 06 14:37:50 :D May 06 14:38:14 I defined tmpdir with ${DISTRO} vars May 06 14:38:24 insane you are May 06 14:38:29 why? May 06 14:38:49 use few build dirs May 06 14:40:04 I guess it ought to work to pass DISTRO from the environment if you want. May 06 14:40:32 it does seem a bit strange to do that rather than having a build directory (and local.conf) per target DISTRO, but I don't think there's any real reason why it couldn't be done. May 06 14:40:42 I dont understand. Before I start bitbake I have to edit local.conf. I wanna avoid that, when I wanna bitbake a package for testing in a another distro cntext May 06 14:42:41 I wonder why locale-base-ar-kw contains all from libc6 here. May 06 14:42:48 did anyone managed to display the connman applet with illume ( enlightenment ) ? May 06 14:43:43 must leave now, thanks for help/advices May 06 14:47:05 /home/hrw/devel/buglabs/oe-dev-bug/build/tmp/cross/armv6/lib/gcc/arm-angstrom-linux-gnueabi/4.2.4/include/limits.h:122:61: error: limits.h: No such file or directory May 06 14:47:08 FSCK May 06 14:47:27 hrw, I've got glibc blowing up on some targets due to .debug junk May 06 14:49:56 I admit my ignorance, but I'm a bit lost with store_adr and load_adr.... May 06 14:50:04 http://www.pastebin.ca/1414261 May 06 14:50:18 same here May 06 14:51:03 how is 0xc0008000 => 0xA0008000.. remapped? May 06 14:58:51 hi,can serial works with just rx and ground if you only need to receive kernel messages? May 06 14:59:29 GnuToo, yes May 06 14:59:44 thanks a lot May 06 15:00:45 ant_work: looks like a CP15 MMU setting cached/buffered May 06 15:01:12 mckoan: is Zaurus pxa May 06 15:01:54 ant_work: CP15 ia an ARM register ;-) May 06 15:01:56 hi gnutoo May 06 15:02:13 mckoan: I thought was a machine...like CPM May 06 15:02:25 confused by the name :-) May 06 15:02:34 ant_work: np :-D May 06 15:02:56 ant_work: what exactly are you digging there? May 06 15:03:13 the pacth flow on the ML and I was testing it May 06 15:03:24 woglinde, hi May 06 15:03:28 just unsure about output..I expected other values May 06 15:04:11 perhaps I misunderstand: http://ja.pastebin.ca/1173295?srch=noop May 06 15:04:17 woglinde, by the way do you know BFD (the library used for the binary format description) ? May 06 15:04:40 mckoan: ## Booting image at a0800000 ... May 06 15:04:50 woglinde, because I have: ld: BFD (GNU Binutils) 2.19 assertion fail /home/embedded/oetmp_pma/work/arm-oabi-angstrom-linux-uclibc/binutils-cross-2.19-r2/binutils-2.19/bfd/elflink.c:2511 May 06 15:05:02 and I use oabi+hard float May 06 15:05:18 mckoan: vs. #store_adr = 000e0000 #load_adr = 00008000 May 06 15:05:36 iirc uboot won't boot XIP so copies the kernel in ram May 06 15:05:52 before launching it May 06 15:06:46 ant_work: looks ok May 06 15:11:56 I just can't see if the patch really detects the good value May 06 15:14:12 gnutoo ugh this has nobody tested May 06 15:14:17 oabi and hardfloar May 06 15:15:17 ant_work: look at this: May 06 15:15:19 17:14 hrw@home:boot$ /usr/local/poky/eabi-glibc/arm/bin/arm-poky-linux-gnueabi-objdump -p vmlinux-2.6.27.2 | sed -n 's/LOAD.*vaddr \([^ ]*\).*/\1/p'|wc -l May 06 15:15:22 2 May 06 15:15:31 should not return one? May 06 15:15:43 morning May 06 15:16:02 hi kergoth May 06 15:17:16 hi kergoth May 06 15:17:24 * hrw -> off May 06 15:19:35 woglinde, lol ok thanks May 06 15:19:49 hrw|afk: that's why I'm dubious May 06 15:20:22 woglinde, as I use a chroot and that the device rootfs is hard float I must use hard float...and the kernel is 2.4 so I must use oabi... May 06 15:30:32 ibot: seen dth? May 06 15:30:36 dth was last seen on IRC in channel #asterisk, 379d 15h 47m 4s ago, saying: 'ok,'. May 06 15:32:37 hi florian May 06 15:33:06 hi pwgen May 06 15:33:29 *G* bacula bb files allready added ? *G* May 06 15:34:04 * florian remembers having a question about it May 06 15:34:52 hi kergoth May 06 15:36:34 pwgen: right... wht was the reason to override do_configure? May 06 15:42:03 florian: i think i added the sqlite feature May 06 15:42:49 pwgen: ah that's what we have EXTRA_OECONF for. May 06 15:43:46 pwgen: There is a variable for the sourceforge mirrors too: SOURCEFORGE_MIRROR May 06 15:44:38 florian: i and mybe the static feature but thisis maybe not a common feature May 06 15:45:16 the openoffice build is a mess May 06 15:45:48 pwgen: should'n matter - you can put the whole list in there May 06 15:47:22 florian: by the way mtx is also needed ( mtx -> tape handling utilities ) May 06 15:50:04 florian: are the old preferred-gpe-versions-2.[6|7].inc still used by some distro? May 06 15:50:39 (I can't grep from here) May 06 15:51:27 ant_work: I don't think so... unless somone builds angstrom 2007.1 May 06 15:53:24 ant_work: your method gives 0xc0008000 for bug kernel and I use 0x80008000 for it May 06 15:53:50 hrw: 2 options: patch is broken or we don't understand May 06 15:53:53 :) May 06 15:56:13 florian: some older gpe recipes could be cleaned out if we keep the latest preferred-gpe-versions May 06 15:56:17 a lot May 06 15:57:15 ah..and what about sharprom-compatible...needing external toolchain May 06 15:57:27 for me, I'd remove it from OE May 06 15:57:45 ant_work: leave it May 06 15:57:57 * hrw|afk -> packing for weekend May 06 15:58:02 ant_work: do not clean up too much May 06 15:58:09 hrw|afk: enjoy May 06 15:58:09 so have a nice rest of week and see you on Tuesday May 06 15:58:30 bye hrw May 06 15:59:01 florian: some embedix kernels are now put in nonworking May 06 15:59:53 I doubt a build for c7x0 would have success May 06 16:00:02 (sharprom-compat) May 06 16:01:24 oh great.... buiding German locale files kills qemu May 06 16:04:20 ant_work: build for akita : task 6780 from 7238 ... in round about 1 hour i know it gpe-image works .. May 06 16:04:44 bad news for you..it failed 2 ays ago... May 06 16:05:06 x11-gpe-image was fine, though May 06 16:05:12 x11-image nod May 06 16:06:26 pwgen: btw do you see ERROR: '[]' RDEPENDS/RRECOMMENDS or otherwise requires the May 06 16:06:28 runtime entity 'virtual/arm-angstrom-linux-gnueabi-gcc-2.95' but it wasn't May 06 16:06:30 found in any PACKAGE or RPROVIDES variables May 06 16:06:33 still? May 06 16:07:09 ant_work: hmm i think i have seen something . ( last night ) May 06 16:07:19 see bug 4784 May 06 16:07:34 florian: sharp schrott May 06 16:08:30 *LOL* May 06 16:08:39 pwgen: it was cleaned by hrw for c7x0, can't remember for others Z May 06 16:09:02 ant_work: i build with bitbake -k gpe-image .. May 06 16:09:10 good luck May 06 16:09:12 ;) May 06 16:09:25 I can tell you where it broke...mom May 06 16:09:52 ant_work: i have managed to get illume runnin on my Z May 06 16:10:05 how is the 'splash'? May 06 16:10:35 people spent lot of time on it (beagle) May 06 16:10:45 ? the boot screen ? May 06 16:12:06 alpha-channels and so on... May 06 16:12:19 slow as expected . May 06 16:12:54 not so slow as tinderbox.... May 06 16:13:33 i have the problem, that key repeat isn't working May 06 16:14:50 and i needed some tweaks on xinit stuff to get it started by gpe-dm May 06 16:18:25 florian: btw gpe-today and gpe-nmf failing May 06 16:19:12 ant_work: both are quite obsolete by now :-( May 06 16:19:36 I was looking in the sources for 'install -s' May 06 16:20:03 never used those May 06 16:21:17 My todo list for gpe images is long... I should start to work on it some time. May 06 16:24:33 I was talking with rschuster before: after building gpe-image or x11-gpe-java-image I ended up with wrong gtk+ .pc file... Package 'gdk-linux-fb-2.0', required by 'GTK+', not found May 06 16:24:49 which package can have done this? May 06 16:33:18 bbl **** BEGIN LOGGING AT Wed May 06 16:47:47 2009 May 06 16:48:12 Hi I used to set this http://nopaste.org/p/a61QSKq7l in my own distribution file to disable udev, but since the update today this stopped working ... does anyone have an idea why it does not work anymore? (using stable/2009) **** ENDING LOGGING AT Sat May 09 06:23:54 2009