**** BEGIN LOGGING AT Fri May 21 02:59:57 2010 May 21 03:42:08 IgorK, you just prepend strace to your command, yes. There is no such thing as a cross strace, it functions in conjunction with a running binary May 21 03:42:53 you can then grep the strace output, for, e.g. wx_gtk2_aui May 21 04:21:00 grg: here is an strace output: http://pastebin.com/7FnHPHfu. Looks like it's not a wx_gtk2* problem. May 21 04:21:24 However I don't see what and where its failing.... May 21 04:23:57 This log is obtained from the bitbake' devshell May 21 04:24:46 It looks like the process has forked before you get the required information May 21 04:24:57 try adding the '-f' flag to strace May 21 04:25:55 btw, strace output is written to stderr, so you need to redirect with strace.... 2>/tmp/strace.log May 21 04:49:43 do I have to have angstrom installed someplace when I run bitbake to build things or can it download everything and install from scratch? (like gentoo, perhaps) May 21 05:04:06 grg: so strace -f ? May 21 05:04:20 and this is how I redirected the output... May 21 05:07:09 IgorK, yes. strace -f blah... May 21 05:13:14 grg: Here is the output: http://pastebin.com/uznquF14 May 21 05:14:30 And here is the version: GNU ld (Linux/GNU Binutils) 2.18.50.0.7.20080502 May 21 05:15:23 this is from "arm-angstrom-linux-gnueabi-ld --version" May 21 05:16:34 well clearly this file is missing: /home/arm/OE/temp/staging/armv4t-angstrom-linux-gnueabi/usr/lib/libwx_gtk2_aui-2.8.so May 21 05:17:26 IgorK, probably needs to be fixed in the wxwidgets recipe May 21 05:19:41 and the this file is clearly there. Look at the very bottom of pastebin output. ;-) May 21 05:21:19 go ahead and type ls -l /home/arm/OE/temp/staging/armv4t-angstrom-linux-gnueabi/usr/lib/libwx_gtk2_aui-2.8.so May 21 05:21:28 i don't think you have that file May 21 05:25:59 grg: As u can c here: http://pastebin.com/Ti4kASvS there is a libwx_gtk2_aui-2.8.so.0, but not libwx_*-2.8.so May 21 05:26:33 IgorK, yes. It looks like that file has not been staged for some reason. May 21 05:26:57 IgorK, you should post a message to the mailing list about this and hopefully someone will fix it. May 21 05:27:31 as an interim, you can create symlinks to the .so.0 files May 21 05:27:50 or use pkg-config, or wx-config (whatever that is) May 21 05:32:04 grg: it looks like there is a problem with the wx-config script for an OE May 21 05:32:27 However, I don't know if it's possible to have an easy fix May 21 05:33:03 and I don't know if there is a maintainer for wxwidgets in OE May 21 05:34:31 the file name I need to refer to is: "libwx_gtk2_aui-2.8-arm-angstrom-linux-gnueabi.so" May 21 05:34:58 http://cgit.openembedded.org/cgit.cgi/openembedded/log/?qt=grep&q=wxwidgets shows some people are making sure this recipe is more or less working May 21 05:35:34 However, since there are different kernels, different gcc versions the name of the libraries are different May 21 05:36:54 IgorK, the do_stage() function in the wxwidgets recipe needs to be modified to create symlinks to the expected .so files May 21 05:37:38 which means that on the different kernel this symbolic link will be named differently May 21 05:38:11 OK, but is there a way to get what name it should be? I guess from the gcc name... May 21 05:39:31 for some reason wxwidgets is adding -arm-angstrom-linux-gnueabi into the so name. Just create symlinks without that segment in the name. You probably need to do this for several .so files May 21 06:04:32 org.openembedded.dev/recipes/linux/linux-omap3_2.6.33.bb fetch failed May 21 06:04:59 i did the .bb file and found that http://www.angstrom-distribution.org/unstable/sources/git_www.sakoman.com.git.linux-omap-2.6.git_248cd24b4fd3376f41fd1b4e3554f4a8c4faa8b9.tar.gz May 21 06:05:07 trying to download May 21 06:05:19 but there is no file in that link May 21 06:06:02 Is there any alternative available for that linux-omap3_2.6.33.bb source link ? May 21 06:07:34 Any one came across this problem ? May 21 06:08:09 Where can i get the source package of linux-omap3_2.6.33 for overo board May 21 06:27:24 grg: yes, that is what I did, thank you for the help May 21 06:28:50 IgorK, make sure you post something to the mailing list so this can be fixed for the next person May 21 06:37:37 Tartarus: pong May 21 07:10:03 morning May 21 07:11:33 hrw: hello May 21 07:19:26 03Enrico Scholz  07org.openembedded.dev * rd98ded68ef 10openembedded.git/classes/packaged-staging.bbclass: May 21 07:19:26 packaged-staging: fixed deadlock May 21 07:19:26 Acked-by: Tom Rini May 21 07:19:26 Signed-off-by: Koen Kooi May 21 07:24:23 03Khem Raj  07org.openembedded.dev * r9bd07cffd2 10openembedded.git/recipes/binutils/ (binutils-cross-sdk_cvs.bb binutils_cvs.bb): (log message trimmed) May 21 07:24:23 binutils-cross-sdk_cvs.bb: Use special do_install for cvs version. May 21 07:24:23 * Remove do_stage for binutils_cvs May 21 07:24:23 * do_install is special for binutils cvs becasue May 21 07:24:23 binutils cvs checks out whole src tree May 21 07:24:24 hence it can not use autotools install funciton May 21 07:24:25 because that will install all tools that are in May 21 07:33:58 good morning May 21 07:59:27 ttb. good night May 21 08:25:46 good morning May 21 09:43:06 I did an IMAGE_INTSALL += "qt4-embedded-fonts" for fonts support module in qte for an image. What should i add for SVG suport module in qte?? May 21 09:45:23 please tell.... May 21 09:47:53 SPyzer : qt4-embedded-plugin-imageformat-svg ? May 21 09:52:24 hi May 21 09:52:53 03Koen Kooi  07org.openembedded.dev * r2c153f33b8 10openembedded.git/ (2 files in 2 dirs): linux-omap-psp 2.6.32: tweak reported powervalue for MUSB on beagleboard May 21 10:01:46 Hey everyone May 21 10:01:59 I just compiled a base-image using stable/2009 May 21 10:02:16 I'm wondering why ipkg and opkg are installed May 21 10:06:58 djdeath: check IMAGE_INSTALL in recipes/images/base-image.bb May 21 10:34:03 hi, when i am trying to an oe generated qemuarm image and try to generate a simple file by "cat > a" it says no space left on device?? I have already set "rw" as a kernel parameter and set the meory to 128 M and yet.... please help May 21 10:35:13 Spyzer: the ext2 image generator creates no spare inodes on the filesystem, so you cannot create files May 21 10:35:54 Spyzer: you can see this by typing df -i May 21 10:36:02 but then how do i create files on machine?? May 21 10:37:28 Spyzer: Spyzer: check resize2fs May 21 10:37:29 You have to tweak the parameters of genext2fs, just digging May 21 10:37:35 EXTRA_IMAGECMD_ext2 = "-i XXX" May 21 10:38:04 substitute XXX with some value, e.g. 300 May 21 10:38:31 genext2fs divides the filesystem size with this number and creates that many inodes May 21 10:38:38 ok well i am using ext3 images so is there any parameter for ext3 as well?? May 21 10:38:42 so XXX is basically the average file size in the file system May 21 10:38:48 akheron: good tip thanks, cleaner than resize2fs ;-) May 21 10:39:08 akheron: thanks from my side too :) May 21 10:39:24 Spyzer, ericben: np :) May 21 10:39:37 Spyzer: the same applies to ext3 as ext3 images are generated with genext2fs, too May 21 10:39:44 it's just EXTRA_IMAGECMD_ext3 then May 21 10:41:38 btw akheron i found that this EXTRA_IMAGE_CMD_ext3 is not present in every distro configuration(like its not there in angstrom). So can i safely introduce it from my own side?? May 21 10:43:19 ?? May 21 10:44:28 you should probably say EXTRA_IMAGECMD_ext3 += "-i foo" to not overwrite any distro specific tweaks May 21 10:44:40 i.e. use "+=" rather than just "=" May 21 10:45:15 though I don't see any distro use EXTRA_IMAGECMD_ext2 or 3 May 21 10:46:11 ohh well i thought that this is a standard parameter to be used in distro configuration May 21 10:46:24 so i can intriduce this in my own image file then?? May 21 10:47:48 ?? May 21 10:50:06 yes May 21 10:50:30 my guess is that many distros don't use ext2 or ext3 for production May 21 10:50:45 as other file systems are better suited for embedded devices May 21 10:51:45 yeah, but qemu supports ext3 as default that's why am going for ext3... May 21 11:12:36 how do i find the plugins names for qt e.g qt4-embedded-plugin-* or qt-embedded-plugin-* ?? May 21 11:12:49 i did a grep but no use May 21 11:14:25 i actually need qt4-embedded plugin for svg. Ericben : suggested me qte-embedded-plugin-imageformat-svg but even on adding this to IMAGE_INSTALL i am unable to view svg files through a simple qte app on my machine in the OE generated image?? May 21 11:14:51 The same app for my X11 desktop displays svg totally fine May 21 11:16:18 please help May 21 11:21:11 * anyone please tell me as to how can i build my_own_customized_image with svg support in qte ?? May 21 11:21:43 i have already tried qt4-embedded-plugin-imageformat-svg parameter to my IMAGE_INSTALL May 21 11:22:07 but the sample app provided by QT Demos (scribble) is unbale to load svg images May 21 11:22:11 please help May 21 11:25:04 can the problem in the svg image itself?? like i have specified encoding="UTF-8" in the xml of svg. Is it the case that UTF-8 is not being supported??? May 21 11:30:18 anyone may atleast please tell me as to where do i find what names do i need to use for any plugin for qte. I mean the qt4-embedded-plugins-* are these names specified anywhere?? May 21 11:39:51 please anyone.... May 21 11:42:33 do i have to install xml processing or anything as an extra plugin??? May 21 11:45:41 ericben: thx May 21 11:46:03 ericben: DISTRO_PACKAGE_MANAGER is set to ipkg ipkg-collateral May 21 11:46:11 ericben: so it's unclear to me May 21 11:46:28 ericben: I still don't see why opkg is installed May 21 11:50:09 ericben: please help me... May 21 11:50:27 lol May 21 12:41:59 djdeath: yes but I think the distro can override this, in example angstrom and minimal set IPKG_VARIANT to opkg and classes/rootfs_ipk.bbclass also set it to OPKG, so unless you explicitset it to ipkg I think opkg will be used as a default May 21 12:44:03 ericben: ok, and according to you, what could explain I have both of them installed May 21 12:44:04 ? May 21 12:46:25 djdeath: check /usr/lib/opkg/status file May 21 12:46:34 something depends on them May 21 12:48:02 hrw: I have both /usr/lib/opkg/status /usr/lib/ipkg/status May 21 12:48:11 hrw: there are identical May 21 12:48:47 hrw: nothing depends on both May 21 12:49:08 djdeath: pastebin it May 21 12:49:10 ~pastebin May 21 12:49:11 [~pastebin] A "pastebin" is a web-based service where you should paste anything over 3 lines so you don't flood 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://bin.cakephp.org/ , http://asterisk.pastey.net/ , or install pastebinit with yum or aptitude. May 21 12:51:02 http://pastebin.com/s5X2azBR May 21 12:51:30 djdeath: your right I've the same here in base-image May 21 12:53:47 djdeath: it seems DISTRO_PACKAGE_MANAGER is installed by base-image.bb (so ipkg) and IPKG_VARIANT is also installed by rootfs_ipk.bbclass (so opkg) May 21 12:54:30 so distro configuration needs fixing May 21 12:55:05 hrw!: in fact DISTRO_PACKAGE_MANAGER is not set anywhere but in recipes/images May 21 12:55:38 it's present in x11-sato-image base(image & nas-server-image. Isn't it something old which wasn't removed ? May 21 12:57:22 I lost track of .dev May 21 12:57:58 on which branch are you working ? stable ? May 21 12:58:43 stable 2009 May 21 12:59:02 maybe a little bit old ? May 21 12:59:39 no it's the same with .dev May 21 13:21:08 ericben: are you investiguating it ? May 21 13:24:09 what is the plugin for svg support in qte. I mean IMAGE_)INSTALL += [ which qt plugin for svg ] ??? May 21 13:25:17 djdeath: not actually, I don't have enough knowledge of this to be able to fix it now. A quick fix would be to set DISTRO_PACKAGE_MANAGER = "" in your conf file. May 21 13:26:02 ericben: ok, thx a lot May 21 13:54:07 any one here with avr32 toolchain ? May 21 13:54:44 and can try to build libiconv from scratch ? May 21 14:25:30 i am getting this error while building my_own_customized_image tune2fs: Could not allocate block in ext2 filesystem while trying to create journal file May 21 14:26:06 i have added EXTRA_IMAGECMD_ext3 += "-i 300" this as a parameter in my image.... May 21 14:28:50 bitbake is giving the error message do_rootfs failed May 21 14:28:51 ?? May 21 14:33:54 sorry i got disconnected, if anyone answered my previous problem kindly do it again.... May 21 14:36:43 i am getting this error while building my_own_customized_image tune2fs: Could not allocate block in ext2 filesystem while trying to create journal file May 21 14:36:45 i have added EXTRA_IMAGECMD_ext3 += "-i 300" this as a parameter in my image.... May 21 14:36:45 bitbake is giving the error message do_rootfs failed?? May 21 14:44:35 is there a switch to forcefully rebuild the cache in bitbake?? May 21 14:44:59 http://pastebin.com/t7VPJNB0 does anybody see such errors ? I have started to see it just today May 21 14:45:11 again some sync problem May 21 14:45:28 second time it builds May 21 14:46:03 khem: i am getting do_rootfs build error when building my image May 21 14:46:15 the error is tune2fs: Could not allocate block in ext2 filesystem while trying to create journal file May 21 14:46:49 and this started happening after some builds when i added this line in my_image.bb -> EXTRA_IMAGECMD_ext3 += "-i 300" May 21 14:48:35 what is the ideal value for --bytes-per-inode for 64M size ext3?? May 21 14:49:10 khem: I think it has something to see with this : http://www.mail-archive.com/openembedded-devel@lists.openembedded.org/msg06159.html May 21 14:58:26 is there any bitbake switch to forcefully rebuild the cache May 21 14:59:59 Spyzer: not that I know, but you can always "touch conf/bitbake.conf" to force a cache rebuild. May 21 15:00:39 mario: please take a look at my previous rootfs error May 21 15:01:52 mickey: good afternoon May 21 15:02:23 hey pb_, mickey May 21 15:03:26 g'day kergoth May 21 15:07:25 hi pb_, hey kergoth_ May 21 15:07:32 pb_: did you wrote that mail yet? May 21 15:08:00 this variable reference tracking shit is proving to be a rather large pain in my ass May 21 15:12:29 khem, oh no. starting udev. attempted to kill init May 21 15:12:49 ~lart udev May 21 15:12:49 * ibot forces udev to use Outlook Express May 21 15:12:54 use devtmpfs May 21 15:13:01 or are you on an old kernel? May 21 15:13:18 mickey: sorry, no, I've been distracted. I'll try to do it over the weekend. May 21 15:13:25 mickey, wait a sec I forgot to put the oe kernel May 21 15:13:34 pb_: ok, no hurries May 21 15:17:07 * kergoth_ realized yesterday that once a def'd function goes into the methodpool, we don't keep around its string version at all, anywhere, and you can't hash a python function object based on its value without disassembling it, and the bytecode won't be portable enough to use in a metadata hashing method May 21 15:17:16 khem, ok booted the minimal distro for htcleo with the gcc overrun line removed and still run into the same problem :( May 21 15:17:32 khem, it build with gcc 4.4.4 May 21 15:18:51 hello, the oe generated images were showing me "filesystem full, on creating any file on their fs in qemuarm machine" May 21 15:19:16 so i added EXTRA_IMAGECMD_ext3 += "-i 300" as a parameter to my_own_customized_image.bb May 21 15:19:42 but bcoz of that error started occuring like tune2fs: Could not allocate block in ext2 filesystem while trying to create journal file May 21 15:19:47 do_rootfs failed May 21 15:20:08 and if somehow the image even got build, in qemuarm the kernel was unable to mount the file system. May 21 15:20:43 Is this happening bcoz of EXTRA_IMAGECMD_ext3 += "-i XYZ" a particular XYZ must be selected ?? May 21 15:21:17 the bitbake process is failing for -i <500 and weird images are getting crated for -i >500 May 21 15:21:19 please help May 21 15:21:40 Spyzer, maybe just try removing EXTRA_IMAGECMD_ext3 what do you need it for ? May 21 15:22:23 i need it so that i can write data on the fs, otherwise when i load images in qemu and do a simple cat > a then it says no space left on device?? May 21 15:22:51 in this IRC was suggested to use the EXTRA_IMAGECMD_ext3 for preventing the error May 21 15:22:59 I see May 21 15:23:05 that's why, but it seems it gave birth to different errors May 21 15:23:13 what should i do?? May 21 15:24:39 Splat1, you could EXTRA_IMAGECMD_ext3 and see how it's done on other machines May 21 15:24:50 git grep May 21 15:29:42 Splat1, if it's general problem with qemu I bet you can find a 'good' -i value then May 21 15:29:45 in some overrun May 21 15:36:55 !oebug 5435 May 21 15:37:44 http://bugs.openembedded.org/show_bug.cgi?id=5435 May 21 15:37:50 mickey, any idea? May 21 15:39:16 khem, updated everything in 5435 May 21 15:39:17 bbl May 21 15:46:55 http://www.flickr.com/photos/hrwandil/4506971517 May 21 15:48:49 hrw: heh... what software do you run on it? May 21 15:49:05 florian: tests mostly now May 21 15:49:17 I just placed a BeagleBoard inside of a SGI 1600SW :) May 21 15:49:38 ;D May 21 15:50:19 hrw its great May 21 16:03:48 have a nice weekend May 21 16:08:50 !oebug 5435 May 21 16:10:16 ~oebug 5435 May 21 16:11:15 hmmm May 21 16:22:41 How hard is it to get oe to use a different compiler from gcc (i.e. llvm) when building binaries? May 21 16:24:44 very May 21 16:25:29 khem: what's involved? May 21 16:25:39 awozniak: unless the new compiler can compile open source software thats primarily built around gcc May 21 16:26:06 awozniak: the infrastructure itself allows to do this though May 21 16:26:10 thats not a problem May 21 16:26:59 khem: assume the alternate compiler will work. What recipes/env settings do I need to change to use it? May 21 16:28:07 if you have the toolchain already precompiled then you can look into external-toolchain recipes May 21 16:28:13 that will be the way to go May 21 16:29:51 khem: thank you. May 21 16:45:42 * khem off to work May 21 16:53:58 khem: while bitbaking an image can i specify the size for it?? I mean can i make 128 M or 256 M image according to my needs?? May 21 16:56:31 anyone?? May 21 17:02:15 is there any way to make a custom sized image like bitbake opie-image which creates an opie-image of 256 Mb ??? May 21 17:02:42 ~ugt May 21 17:02:43 methinks ugt is Universal Greeting Time. Created in #mipslinux, it is a rule that states that whenever somebody enters an IRC channel it is always morning, and it is always late when the person leaves. The local time of any other people in the channel, including the greeter, is irrelevant. http://www.total-knowledge.com/~ilya/mips/ugt.html May 21 17:03:20 ~ugt May 21 17:03:21 well, ugt is Universal Greeting Time. Created in #mipslinux, it is a rule that states that whenever somebody enters an IRC channel it is always morning, and it is always late when the person leaves. The local time of any other people in the channel, including the greeter, is irrelevant. http://www.total-knowledge.com/~ilya/mips/ugt.html May 21 17:05:21 ~ibot May 21 17:05:22 It's me May 21 17:05:40 !ibot May 21 17:06:03 ibot^2 May 21 17:13:18 hi all May 21 17:13:44 I want to make bitbake compile a specific version of recipes/ti/ti-dsplink May 21 17:14:02 if i do bitbake ti-dsplink it compiles version 1.64 May 21 17:14:20 there is a recipe .bb ti-dsplink_1.65.00.02.bb May 21 17:14:30 how can i tell him to compile this one? May 21 17:15:03 (I've tried bitbake ti-dsplink_1.65.00.02.bb but it says nothing provide it...) May 21 17:15:13 any idea? May 21 17:41:09 good morning May 21 17:41:30 is there a bug tracking database for openembedded? May 21 17:55:03 Yes May 21 17:55:07 bugs.openembedded.net, iirc May 21 17:57:30 03Klaus Kurzmann  07org.openembedded.dev * rd1b03808e6 10openembedded.git/recipes/linux/linux-openmoko-devel_git.bb: May 21 17:57:30 linux-openmoko-devel_git.bb: apply getline fix for scripts/unidef.c May 21 17:57:30 Signed-off-by: Klaus Kurzmann May 21 18:02:00 nobody for bitbake help? May 21 18:16:58 pepere_: what seems 2 b the problem? May 21 18:17:15 tartarus: r u sure? May 21 18:17:34 well I got the recipe in recipes/ti/ti-dsplink_1.65.00.02.bb May 21 18:17:38 maybe .org May 21 18:17:40 i forget May 21 18:17:44 or .com, heh May 21 18:18:05 but each time I launch bitbake ti-dsplink it compile the 1.64 version May 21 18:18:40 IgorK if I say: bitbake ../openambedded/recipes/ti/ti-dsplink_1.65.00.02.bb May 21 18:19:04 it says me nothing PROVIDES ti-dsplink_1.65.00.02.bb May 21 18:19:42 I'm wondering if it does not wait from me to downlod the tra.gz file from ti, so it could start its work? May 21 18:20:43 got this: May 21 18:20:44 ERROR: Nothing PROVIDES '../openembedded/recipes/ti/ti-dsplink_1.65.00.02.bb' May 21 18:20:51 ... :/ May 21 18:42:18 pepere_: denix might be able to help you May 21 19:07:09 ugh May 21 19:08:03 * kergoth_ just realized that FILESPATH / file:// are going to be an issue with the reference tracking. ideally we wouldn't want MACHINE pulled into every recipe's signature via FILESPATH / OVERRIDES, but we can't just blacklist them, because we really do want the unpack singature to change if a more specific version of a file shows up May 21 19:09:37 hmm, maybe we just need a way to flag usages of python snippets which depend upon the filesystem, and for those, fully resolve them and include that in the signature, rather than the components May 21 19:10:02 would make sense, hashing the output of things that poked the disk, rather than the components of it May 21 19:15:35 thoughts? May 21 19:21:20 kergoth_: you mean cases where we have file:// override for machines like in u-boot May 21 19:27:14 khem: yeah May 21 19:27:22 SRC_URI doesn't change, but the file used is different May 21 19:27:30 if we move to signatures instead of stamps, thats a problem May 21 19:28:08 re May 21 19:31:59 kergoth_: I see. Do we have info about file:// and override ? May 21 19:57:17 khem: see FILESPATHBASE, FILESPATHPKG, and FILESPATH in bitbake.conf. file:// just uses bb.which() to find it in FILESPATH May 21 20:04:11 hrm May 21 20:12:17 03Klaus Kurzmann  07org.openembedded.dev * r74228420b9 10openembedded.git/recipes/openbmap/ (2 files in 2 dirs): May 21 20:12:17 openbmap-logger_0.4.0.bb: add a patch to make it compatible with fsogsmd May 21 20:12:17 Signed-off-by: Klaus Kurzmann May 21 20:23:36 03Klaus Kurzmann  07org.openembedded.dev * r0609b025f4 10openembedded.git/recipes/freesmartphone/cornucopia.inc: May 21 20:23:36 cornucopia.inc: bump rev to get working fsonetworkd connection sharing May 21 20:23:36 Signed-off-by: Klaus Kurzmann May 21 20:24:03 * kergoth_ wonders how to get bitbake pushes to CIA May 21 20:28:32 kergoth_: heh call the feds May 21 20:29:24 heh :) May 21 20:29:45 this fucking variable reference tracking code just keeps getting more troublesome May 21 20:29:49 annoying May 21 20:42:29 I've got a simple question: what's the name of the file to be called by the "init" kernel parameter during angstrom's boot? May 21 20:42:53 in common linux distros that would be /sbin/init May 21 20:51:56 it should be same May 21 20:52:25 kergoth_: I am changing distro from minimal to minimal-uclibc and everything remains same bitbake gives me error in parsing May 21 20:52:50 its still looking for virtual/powerpc-oe-linux-uclibc-gcc May 21 20:53:09 but this time its eglibc so I wonder why May 21 20:53:28 huh, weird May 21 20:53:59 http://pastebin.com/PU0m9xpL May 21 20:54:47 hmmm where is TARGET_PREFIX set ? May 21 20:56:05 TARGET_PREFIX = "${TARGET_SYS}-" May 21 20:56:55 TARGET_SYS = "${TARGET_ARCH}${TARGET_VENDOR}${@['-' + bb.data.getVar('TARGET_OS', d, 1), ''][bb.data.getVar('TARGET_OS', d, 1) == ('' or 'custom')]}" May 21 20:57:35 so TARGET_OS May 21 20:58:56 hmmm May 21 20:59:44 kergoth_: how would it compute I guess when it encounters first TARGET_PREFIX he should go down evaluating other vars right ? May 21 21:00:18 the problem is that TARGET_OS = "${@compute_os_portion_of_target_triplet(d)}" May 21 21:02:29 strange.... I've built a jffs filesystem with the online builder, and it mounts correctly during boot, but I can't find neither /linuxrc (my old init), nor /sbin/init, so it panics warning unable to find init= May 21 21:04:17 hmm is it in .dev ? May 21 21:04:40 linuxrc should be a symlink to /bin/busybox May 21 21:05:09 I can't mount it as for now (I'm recompiling my kernel to support MTD and then I'll try to mount the image) May 21 21:05:33 I can build another image in .tar.bz2 format, which is most likely what I'll be doing to make sure May 21 21:05:56 yeah try that May 21 21:06:20 there are somewhat funny MTD drivers on linux May 21 21:06:32 some that allow you to use uncacheable RAM as a mtd device May 21 21:06:51 or ram that's beyond what kernel sees May 21 21:07:02 (that was the in-the-other-news section) May 21 21:12:38 hi i am bitbaking base-image for qemu arm but as soon as i start qemu it says root: mount: mounting rootfs on /failed: No such file or directory ?? May 21 21:12:45 why is this happening?? May 21 21:12:48 or how can i resolve this May 21 21:13:04 Spyzer: are you using script from contric/ May 21 21:13:07 contrib May 21 21:13:18 i didn't quite get u?? May 21 21:13:25 contrib/qemu/run-qemu.sh May 21 21:13:35 no it is my own.... May 21 21:13:38 is a script to ease May 21 21:13:44 it used to work a short time ago May 21 21:13:57 i have loaded many images since 2 months May 21 21:14:08 but just today nothing is loading?? May 21 21:14:27 Spyzer: show me your qemu invocation command May 21 21:14:59 ../qemu-system-arm -M versatilepb -usb -usbdevice wacom-tablet -show-cursor -m 64 -kernel zImage-2.6.25+2.6.26-rc4-r6-qemuarm.bin -drive file=Angstrom-qtehelloworld-glibc-ipk-2010.4-test-20100521-qemuarm.rootfs.ext3 -append "root=/dev/sda rw" May 21 21:15:30 and i am telling u this has loaded many images, even my_own_custozed_image was loading till today's afternoon May 21 21:16:30 afterwards i added EXTRA_IMAGECMD_ext3 += "-i 300" to my image and everything went wrong and even when i am doing bitbake -c clean qtehelloworld-image and again bitbake qtehelloworld-image May 21 21:16:34 the error is persisiting May 21 21:16:49 khem please hell me today in after noon is my project evaluation May 21 21:16:56 sword is on my neck :'( May 21 21:17:03 *hell help May 21 21:17:31 i just have 8 hrs to go if i go on without sleep boohooo May 21 21:18:24 Spyzer: and ./Angstrom-qtehelloworld-glibc-ipk-2010.4-test-20100521-qemuarm.rootfs.ext3 is present May 21 21:18:37 is there any command to clean recursively the dependencies of my_own_cutomized_image or any .bb package May 21 21:18:45 ofcourse May 21 21:19:45 khem: please listen the error started occuring when i added the param EXTRA_IMAGECMD_ext3 += "-i 300" May 21 21:19:53 and each image built by this command gave the error May 21 21:20:21 so now i removed this from my_own_customized_image did a bitbake clean and rebitbaked but yet the same error is occuring?? May 21 21:20:54 are you sure you regenerated images May 21 21:20:56 what exactly is going on?? May 21 21:21:03 and if they are then are u using them ? May 21 21:21:48 i mean i delete the images in the deploy folder what else can i do?? May 21 21:22:00 thats good May 21 21:22:02 and i am using the stuffs baked out of the deploy folder only May 21 21:22:48 khem: i am feeling as if i should jum of a cliff things are started to fall apart right when its showtime :( May 21 21:23:05 *jum jump May 21 21:23:36 is the cleaning task not done properly by me?? May 21 21:24:31 well do this May 21 21:24:49 clean one recipe May 21 21:24:55 a small one May 21 21:24:55 khem: not sure what you mean. unless you use :=, variables are expanded when they're used, not when they're encountered May 21 21:24:57 may be udev May 21 21:25:11 bitbake -c clean udev;bitbake May 21 21:25:42 where did you add EXTRA_IMAGECMD_ext3 bte May 21 21:26:56 here http://pastebin.com/rz4Bnepf is the code of my image-recipe and i added the EXTRA_IMAGECMD right after the first line May 21 21:27:21 khem how did bitbake-c clean udev can help it just cleaned the udev recipe May 21 21:27:28 built one i mean?? May 21 21:27:37 argh *?? !! May 21 21:27:47 it will rebuild the recipe and then regenerate image May 21 21:28:18 i mean so how shall it help in my case, can i delete the root folder where the file system is build somehow... May 21 21:28:52 bcoz the error was that mounting rootfs on / failed May 21 21:29:07 maybe the root fs has some corrupted things bcoz of my earliere parameter May 21 21:31:16 no the command you added should only affect final image creation May 21 21:31:32 hmmm May 21 21:31:41 but NOW what to do May 21 21:31:52 i am trying the thing u told lets hope for the best May 21 21:32:12 yeah dont worry all will be ok May 21 21:36:17 khem: oh no the same error occured again?? May 21 21:36:31 should i delete my tmp folder entirely, and recreate, that will take 4 hrs :( May 21 21:37:17 please suggest a thing May 21 21:37:18 hmm strange May 21 21:37:48 are you sure you removed this command May 21 21:37:55 from all places you added it to May 21 21:38:56 same results here :( May 21 21:39:10 doing online it is unable to find /sbin/init May 21 21:39:26 but in .tar.bz2 it's a symlink to busybox May 21 21:40:25 yes khem i just did a grep and its now where to be seen as far as han coding is concerned, sone distro's conf have it on by default May 21 21:41:41 khem: what dis u suggest should i risk it and delete my entire tmp directory and rebuild from scratch?? May 21 21:42:01 Spyzer: yeah if you have time then thats the best option May 21 21:42:08 dont do any updates or anything May 21 21:42:12 use same setup May 21 21:42:16 and just rebuild May 21 21:42:40 but I dont get it why it would build it differently May 21 21:42:53 you can also look into the logs of image generation May 21 21:43:16 and see if the command you added is still being used somehow May 21 21:43:25 khem: one more thing i remembered when i added that EXT@ param some new sources were donwloaded(one of them was new kernel) do u think some of a new version of any package is causing this?? May 21 21:43:50 don't be mad khem where do i check the image logs?? May 21 21:44:12 in your build tree May 21 21:44:42 * kergoth_ grumbles May 21 21:44:42 under qemuarm-oe-linux-gnueabi/ May 21 21:48:27 khem : if the error is bcoz of any new or upadte package then even if i delete evrything then also the error will persist May 21 21:48:56 yeah May 21 21:49:14 oh no, any suggestions then?? May 21 21:49:17 Spyzer: Oh wait a moment how did you get new packages May 21 21:49:21 did you do git pull May 21 21:49:48 no no git pull today, but yes a few days back i did that May 21 21:49:55 i even asked u how to stash May 21 21:50:03 and did it work after that ? May 21 21:50:10 so u told me the command which i have forgotten May 21 21:50:16 heh ok May 21 21:50:18 yeah the image the was working after that May 21 21:50:57 ok then it should keep working May 21 21:50:58 do u think i should do a git pull now, so that everything gets update?? May 21 21:51:03 *upadte updated May 21 21:51:13 Not needed I guess May 21 21:51:28 stay to minimal changes May 21 21:51:36 hmmm ok May 21 21:51:40 otherwise you will be in trouble May 21 21:51:45 then going from scratch rm -rf tmp :) May 21 21:51:51 ok May 21 21:51:55 Hey thar May 21 21:52:28 I've got a question on compiling quasar, if anyone would be willing to help? May 21 21:52:57 dont ask to ask just ask May 21 21:53:47 trying to compile quasar for angstrom for use on my archos 5IMT (armv7a) May 21 21:54:05 my build nerfs, and I've narrowed it down to the following error May 21 21:54:08 /home/jonez/usr/src/openaos/build/angstrom-dev/cross/armv7a/lib/gcc/arm-angstrom-linux-gnueabi/4.3.1/../../../../arm-angstrom-linux-gnueabi/bin/ld: warning: libstdc++.so.6, needed by /home/jonez/usr/src/openaos/build/angstrom-dev/staging/armv7a-angstrom-linux-gnueabi/qt3/lib/libqt-mt.so, not found (try using -rpath or -rpath-link) May 21 21:54:42 full pastebin error log here. May 21 21:54:42 http://pastebin.com/3HinBEP8 May 21 21:54:46 line 1631 May 21 21:55:10 How old is your snapshot May 21 21:55:14 of OE May 21 21:55:17 hm May 21 21:55:22 bout a month or so? May 21 21:56:02 * Spz0 is relatively new to bitbake and OE, so go easy. ;) May 21 21:56:26 Spz0: you should try after b362ce8d2330d6d8b2498ecbfe5f9ad8b183df68 commit id May 21 21:56:35 if you can do a git pull and retry May 21 21:56:54 actually you have to build from scratch when you pull beware May 21 21:57:04 hrrm May 21 21:57:42 the problem is so. It needs that fix May 21 21:59:31 k, I'll give it a shot. was hoping to not hav to build from scratch, but meh... :) May 21 22:07:55 git pull returned "already up to date" May 21 22:11:02 !oebug 5435 May 21 22:11:03 * * Bug 5435, Status: UNCONFIRMED, Created: 2010-05-19 14:26 May 21 22:11:04 * * lukas(AT)htc-linux.org: machine htcleo: gcc version etc. May 21 22:11:05 * * http://bugs.openembedded.org/show_bug.cgi?id=5435 May 21 22:11:10 khem, mickey|tv, thanks for the replies May 21 22:11:37 khem, disabling devtmpfs doesn't fix it with the stock udev version May 21 22:11:54 khem, next approach, setting the udev version to 151: http://tinderbox.openembedded.net/public/logs/task/5859777.txt May 21 22:16:18 Spz0: can you do git show b362ce8d2330d6d8b2498ecbfe5f9ad8b183df68 May 21 22:17:08 yes May 21 22:17:51 dcordes: add PREFERRED_VERSION_usbutils_local = "0.86" May 21 22:17:59 to your local.conf May 21 22:18:03 aight May 21 22:18:34 khem: hello May 21 22:18:41 khem, when I ran -c clean udev it removed some 151 stuff which confused me. May 21 22:18:59 ant_home: saluti May 21 22:19:00 khem: I'm testing the latest klibc now that the getline issue is solved May 21 22:19:12 (make_headers_install) May 21 22:19:12 ant_home: cool May 21 22:19:22 khem, should the patch to sane-toolchain.inc be added? May 21 22:19:31 which one ? May 21 22:19:36 armv7 ? May 21 22:19:49 not until you see it booting May 21 22:19:54 We have to get the complete set May 21 22:19:57 -PREFERRED_GCC_VERSION_armv7a ?= "4.4.1" May 21 22:20:01 ok. May 21 22:22:08 khem: bad news (though I remember smthg of you about that) May 21 22:22:12 /oe/build/tmp/work/c7x0-angstrom-linux-gnueabi/klibc-1.5.17-r14.0/klibc-1.5.17/usr/include/arch/arm/klibc/archsignal.h:107:37: error: asm-generic/signal-defs.h: No such file or directory May 21 22:22:23 khem: http://pastebin.com/Nd1apa2B May 21 22:22:27 from git show ...... May 21 22:22:37 this is klibc_1.5.17 vs. kernel 2.6.26 May 21 22:24:29 it's very nice to see the zaurus dog family still being supported so well May 21 22:28:56 console-image builds freesmartphone stuff. is that expected? May 21 22:32:56 dcordes: the biggest issue is the kernel (without having machine and qemu suppoort) May 21 22:44:14 ant_home, if you need any akita testing ping me May 21 22:47:04 dcordes: thx. I'd need more poodle/collie/tosa testers ;) May 21 22:47:21 clamshells are ok May 21 22:47:50 Spz0: can you paste the full build log of the failing package somewhere May 21 22:48:11 khem: as I supposed klibc broke support with 2.6.2x :/ May 21 22:48:15 nm you have it May 21 22:48:22 time to move on.... May 21 22:48:28 quasar failed build ::: http://pastebin.com/3HinBEP8 May 21 22:50:42 Spz0: the real problem is that its trying to link a C++ program with gcc where as it should have used g++ May 21 22:50:50 arm-angstrom-linux-gnueabi-g++ May 21 22:52:55 hmm May 21 22:53:04 Im using the OE quasar recipe May 21 22:53:04 quasar_0.9beta3.bb May 21 22:53:11 files/cross-compile.patch patch seems suspicious too May 21 22:53:27 rotate patch is one of mine. May 21 22:54:31 http://pastebin.com/8XQYiFmY May 21 22:54:39 contents of cross-compile.patch May 21 22:55:54 thanks for the help, btw. ;) May 21 22:56:42 yeah I can see that May 21 22:56:54 btw you can try one of those commands by hand May 21 22:57:01 and replace gcc with g++ May 21 22:57:17 e.g. one at line 1630 May 21 22:57:22 * khem needs to go May 21 22:59:02 k m8. thanks for the help. I"ll get with u later if I still can't resolve this today. :) May 21 23:23:47 khem, same crash with May 21 23:23:48 PREFERRED_VERSION_udev_local = "151" May 21 23:23:48 PREFERRED_VERSION_usbutils_local = "0.86" May 21 23:29:46 khem, how to go about getting rid of udev and using devtmpfs ? May 21 23:30:10 khem, it's clear how to enable devtmpfs in the kernel. but how can udev be removed? May 21 23:38:23 ~lart udev May 21 23:38:23 * ibot installs PocketPC on udev's PDA May 21 23:38:28 lol May 22 00:00:51 dcordes: yeah thats a good question May 22 00:01:08 dcordes: there are dependencies like util-linux-ng May 22 00:01:15 which pull in udev May 22 00:02:03 I think the init scripts need to be modified to disable the udev at startup May 22 00:02:14 that's what I thought May 22 00:02:29 is there any mechanism for this in oe? May 22 00:04:22 dcordes: as a hack you can delete udev/udev-151/init May 22 00:04:28 and have an empty one May 22 00:09:11 khem, ok and as a next step you think it's necessary to use an initramfs? May 22 00:11:52 could be. as I said you have to try it out May 22 00:13:54 khem, is it a problem to have the basic device nodes in /dev ? May 22 00:14:23 khem, when using CONFIG_DEVTMPFS CONFIG_DEVTMPFS_MOUNT May 22 00:17:22 bake&error May 22 00:41:10 hell no! May 22 00:41:23 INIT: PANIC: segmentation violation! May 22 00:42:14 sleeping for 30sec May 22 00:42:39 runs in a loop when enabling devtmpfs in kernel, removing udev from init May 22 00:42:55 after remounting rootfs May 22 00:43:32 what is the process init is spawning after remounting rootfs? (or where can I look it up) May 22 00:50:58 something low level is still wrong. May 22 00:51:03 illegal instruction May 22 01:07:35 dcordes: there might be more to it May 22 01:08:24 then just disabling udev May 22 01:10:24 khem, do you think the segmentation violation and illegal instruction errors are caused by non working /dev ? May 22 01:49:25 dcordes: I believe so May 22 01:49:51 dcordes: it could be generic too May 22 01:50:22 do you have details about the segfault May 22 01:50:29 does it give address and the intruction encoding May 22 01:52:36 khem, the problem is it showed only the first boot. I'm writing it up in the bug now May 22 01:52:52 khem, now I only get attempting to kill init panic after remounting rootfs May 22 01:54:47 ok May 22 01:54:49 ok I know why. it must be the huge root delay I added that got rid of it May 22 01:55:19 I will post dmesg with and without rootdelay May 22 01:55:27 and userdebug=-1 May 22 02:08:22 now look at that mess May 22 02:08:22 [ 43.772264] init (1): undefined instruction: pc=0000a654 May 22 02:12:24 khem, it must be generic low level problem right? May 22 02:25:53 !oebug 5435 May 22 02:25:54 * * Bug 5435, Status: UNCONFIRMED, Created: 2010-05-19 14:26 May 22 02:25:55 * * lukas(AT)htc-linux.org: machine htcleo: gcc version etc. May 22 02:25:56 * * http://bugs.openembedded.org/show_bug.cgi?id=5435 May 22 02:25:58 everything updated May 22 02:26:14 0000a654 is not a kernel address May 22 02:26:17 maybe somebody can give a hand switching from udev to devtmpfs in a clean way? May 22 02:26:21 nor userspace May 22 02:26:28 so something wonked May 22 02:26:30 earlier May 22 02:27:11 khem, hm don't quite get it. what does this mean? May 22 02:34:13 need some sleep bbl May 22 02:34:24 khem, again thanks for the help. **** ENDING LOGGING AT Sat May 22 02:59:57 2010