**** BEGIN LOGGING AT Thu Sep 29 02:59:56 2011 Sep 29 03:44:57 hi wondering if anyone is here to help? Sep 29 03:45:43 kaylessa: i have had better luck when asking in the day time (US day time) Sep 29 03:46:01 yea agree Sep 29 03:46:08 thanks i come back later Sep 29 04:15:33 i have the digraphs from bitbake. i wonder if somewhere in this is a clue on where/how to insert gcc so i get a working arm gcc on my output image Sep 29 04:16:54 upside, if i ever get to the bottom of this multi-week quest, i will "get" bitbake/oe Sep 29 05:52:16 hello Sep 29 07:34:27 good morning Sep 29 09:35:48 hi all Sep 29 09:36:12 I am building different machines from same core Sep 29 09:36:32 I assumed that one machine will pick sstate-cache of other machine Sep 29 09:36:39 but this did not happend Sep 29 09:37:01 I've noticed the same here, but didn't have time to debug yet Sep 29 09:37:08 can you compare sstate checksums? Sep 29 09:37:12 when I tool the diff of the siginfo file it said that Hash for dependent task /home/nahsan/SB/mel5/trunk/oe-core/meta/recipes-core/expat/expat_2.0.1.bb.do_install changed from 880bba4f07a918d05cbe13c082c5bc54 to 9cb6e1f751c7929ec68f231d7db4510d Sep 29 09:37:24 with bitbake-diffsigs Sep 29 09:37:42 how can I see what is changed in do_install function which changed the cache Sep 29 09:38:38 JaMa|Off: yes i used the bitbake-diffsig Sep 29 09:39:09 abut it only said "Hash for dependent task /home/nahsan/SB/mel5/trunk/oe-core/meta/recipes-core/expat/expat_2.0.1.bb.do_install changed from 880bba4f07a918d05cbe13c082c5bc54 to 9cb6e1f751c7929ec68f231d7db4510d" Sep 29 09:39:22 how can I see do_install changes? Sep 29 09:39:23 Noor: http://bugzilla.pokylinux.org/show_bug.cgi?id=1074 Sep 29 09:39:44 there should be siginfo file for do_install Sep 29 09:40:43 JaMa|Off: :O there is not siginfo file for do Sep 29 09:40:58 do_install inc sstate_cache Sep 29 09:41:09 let me go through the thread Sep 29 09:42:40 then something broke that bitbake patch later, because IIRC I've tested it and confirmed it to work right Sep 29 10:04:41 JaMa|Off: there is interest for unifying xserver recipes in oe-core. What's the status of your patch-queue? Sep 29 10:06:59 Noor: here it's maybe because of MACHINE_CLASS is used in the other machine and maybe it's not excluded from sigdata where it should Sep 29 10:07:24 ant_work: mostly to rework xserver-common/x11-common Sep 29 10:07:55 ant_work: I don't like /etc/Xusername used only in init.d script Sep 29 10:09:08 ant_work: but I don't have time to work on it this week Sep 29 10:09:20 where can I modify which packages are included in console-image? Sep 29 10:10:33 mar: in console-image recipe or tasks which are included in console-image IMAGE_INSTALL Sep 29 10:11:37 maybe i'm doing something wrong. I build console-image, etc. Now I want additional packages to be included. What is the best way to do it? Sep 29 10:15:00 JaMa|Off: what do u mean by MACHINE_CLASS ... BTW I have added MACHINE in BB_HASHBASE_WHITELIST Sep 29 10:17:51 Noor: it was used in oe.dev as extra override.. so I'm still using it ie for palmpre[12] MACHINEOVERRIDES .= ":${MACHINE_CLASS}" Sep 29 10:18:30 Noor: and you can also diff actuall do_install scripts in workdir/*/temp directory (if you can guess which one is which based on creation time) Sep 29 10:18:38 JaMa|Off: aahhh i mixed it with bbclass Sep 29 10:19:24 JaMa|Off: WELL i have already done it and I only see diff in {D} and build folder name Sep 29 10:19:55 let me see why signature of all the task are not dumpped may be that will help Sep 29 10:20:11 hicolor-icon-theme-native is missing :( Sep 29 10:20:31 Noor: why is D and buildfolder different? Sep 29 10:20:40 Noor: I guessed you have it in same directory Sep 29 10:21:10 Noor: if PV/PR or whatever was changed then it's expected to have different hash -> different sstate packages Sep 29 10:21:20 JaMa|Off: no its in the diff dir Sep 29 10:21:35 Zoxc: and what needs it? Sep 29 10:22:05 JaMa|Off: everything is same but i build p2020rdb in one build folder and build mpc8572 in other Sep 29 10:22:41 ['xfce-terminal-dev', 'exo', 'libxfce4util', 'xfce4-dev-tools-native', 'hicolor-icon-theme-native'] Sep 29 10:23:23 JaMa|Off: a fw days ago unexpectedly lot of stuff was rebuilt just changing machine c7x0 -> poodle (eglibc and co.). I expect same toolchain. Sep 29 10:25:53 ant_work: iirc this makes all sstate hashes invalid http://git.openembedded.org/cgit.cgi/openembedded-core/commit/?id=982373006a98cf2303514badd1cfb692108408c0 Sep 29 10:26:08 ant_work: so maybe you've switched machines just after pulling this change Sep 29 10:26:59 sigh.. this reminds me ccache cannot be easily turned off Sep 29 10:31:37 adding BBCLASSEXTEND = "native" to one of the hicolor-icon-theme_0.12.bb's makes it continue Sep 29 10:33:24 ah, nice, we have 2 recipes..again :/ Sep 29 10:33:46 signature of each task are being generated Sep 29 10:34:02 much of recipes-gnome seems to be duplicated between meta-oe and oe-core =P Sep 29 10:34:02 but those are generated in stamps folder Sep 29 10:36:51 or just librsvg, gtk+ and hicolor-icon-theme Sep 29 10:53:05 morning Sep 29 10:56:41 hi guys i am trying to build virtual/kernel for my board using mirco-uclibc Sep 29 10:58:13 but it has errors Sep 29 10:58:46 the error is as follows Sep 29 10:58:48 ERROR: 'virtual:native:/home/lkp/oe/openembedded/recipes/glib-2.0/glib-2.0_git.bb' while parsing /home/lkp/oe/openembedded/recipes/glib-2.0/glib-2.0_git.bb Sep 29 10:58:54 wondering if anyone could help? Sep 29 11:00:15 i switched from oe created toolchain to external codesourcery arm2009q1 toolchain in local.conf - toolchain seems working, but fakeroot gets no longer build.. how could that be? fakeroot is build by hostoolchain?! Sep 29 11:00:56 is there any problem downloading http://mirrors.openembedded.org//git-1.7.3.5.tar.bz2? Sep 29 11:01:17 i fail my build when i tries to download http://mirrors.openembedded.org//git-1.7.3.5.tar.bz2 Sep 29 11:16:18 hi wondering if there anyone that could help Sep 29 11:20:39 find some mirror of it, http://pkgs.fedoraproject.org/repo/pkgs/git/git-1.7.3.5.tar.bz2/8a8cd93b8a4dff0a03c0fdc77253af3e/git-1.7.3.5.tar.bz2 Sep 29 11:25:04 how to change it so that it download from a different mirror? Sep 29 12:33:58 fakeroot: command not found after switchting to external toolchain.. wtf Sep 29 13:06:00 hi i wondering how to solve this Sep 29 13:06:01 WARNING: Deprecated usage of 'patch' url param for '/home/lkp/oe/openembedded/recipes/binutils/binutils-2.20.1/binutils-uclibc-100-uclibc-conf.patch', please use 'apply={yes,no} Sep 29 13:07:42 how do I add mercurial? recipe for mercurial sits recipes/mercurial/mercurial-native_1.6.3.bb . What do I need to specify in IMAGE_INSTALL? Sep 29 13:07:54 I tried mercurial, mercurial-native, both doesn't work Sep 29 13:07:55 kaylessa: just remove the "patch=xx" parameter altogether Sep 29 13:08:10 may I know where excatly libpseudo.so is added in LD_PRELOAD Sep 29 13:08:31 kaylessa: from SRC_URI Sep 29 13:08:48 but where do i find the file Sep 29 13:09:02 its not inside binutils-2.20.1/binutils-uclibc-100-uclibc-conf.patch Sep 29 13:09:18 kaylessa: no, it's in the binutils recipe which is being built Sep 29 13:09:31 kaylessa: it's only a warning btw, it's not stopping your build from completing Sep 29 13:09:35 is it crucial? Sep 29 13:09:37 ok can Sep 29 13:10:09 thanks alot Sep 29 13:10:46 mar: if the recipe is mercurial-native*.bb then it's a native-only recipe and cannot produce packages to be installed on the target Sep 29 13:11:21 and what does it mean? Sep 29 13:11:33 that I have to copy files separately? Sep 29 13:12:24 or does it mean that I can only run it on my pc? Sep 29 13:12:34 mar: it means if you want mercurial on the target you'd need to create a new recipe Sep 29 13:12:54 presumably nobody has needed it up to now Sep 29 13:13:09 and what does native mean? Sep 29 13:13:21 native means for the build machine Sep 29 13:13:46 ok Sep 29 13:14:43 is there something like build-essential recipe? Sep 29 13:18:03 mar: I imagine task-sdk-native might be what you want Sep 29 13:18:19 (really, this word native is somewhat abused...) Sep 29 13:19:07 and what does it do? Sep 29 13:19:20 ah Sep 29 13:19:42 but it'll probably push all oe to device ? :) Sep 29 13:21:39 hi is there any good solution to fix checksum error during build? Sep 29 13:26:32 mar: not all of OE Sep 29 13:26:47 mar: just look at recipes/tasks/task-sdk-native.inc for the details Sep 29 13:28:48 virtual/kernel will create a uimage file for my machine right? Sep 29 13:31:20 kaylessa: if the machine is set up to create a uimage yes Sep 29 13:31:50 i.e. are the appropriate variables set in the conf/machine/machinename.conf Sep 29 13:32:18 yea my machine conf stated uimage Sep 29 13:32:30 hi i got this problem with checksum Sep 29 13:32:34 here is the error Sep 29 13:32:48 Expected MD5: 'b32b04148ecdd7344abc6fe8bd1bb021' and Got: 'e54d0094d3292f4d8f8b9b9b054b05fd' Expected SHA256: '13df57ab63a94e196c5d6e95d64e53262834fe780d5e82c28f177f9f71ddf62e' and Got: '6b7f63513a94e3d930860fd5d294045f2dded818e27bbcb1fa62a8cea7019e07' NOTE: Your checksums: SRC_URI[md5sum] = "e54d0094d3292f4d8f8b9b9b054b05fd" SRC_URI[sha256sum] = "6b7f63513a94e3d930860fd5d294045f2dded818e27bbcb1fa62a8cea7019e07" Sep 29 13:32:59 kaylessa: which recipe? Sep 29 13:33:04 do i edit .bb to suit the checksum? Sep 29 13:33:20 recipes/libtool/libtool-native_2.4.bb Sep 29 13:33:50 kaylessa: if you are sure that the upstream archive change is legitimate, yes Sep 29 13:33:59 the checksums are there for your protection Sep 29 13:34:34 how to know if it is legitimate Sep 29 13:34:52 sorry for asking so many question Sep 29 13:34:56 kinda new to oe still Sep 29 13:35:42 hey blue how to know if it is legitimate? Sep 29 13:38:30 kaylessa: that's a very good question... Sep 29 13:38:56 nvm i change and try if i can load the .bin means its legitmate Sep 29 13:38:57 haha Sep 29 13:39:23 maybe it is related to the other gnu package changes Sep 29 13:39:39 due to licensing concerns (missing source files) Sep 29 13:41:28 bluelightning, so I need to add task-sdk-native to IMAGE_INSTALL? Sep 29 13:41:36 mar: yes Sep 29 13:41:46 or, install it afterwards manually, one of the two Sep 29 13:48:17 ok thanks, trying to build now :) Sep 29 15:42:03 what could be the problem if building gcc it generates error: configure: error: C compiler cannot create executables Sep 29 15:42:10 in do_configure part Sep 29 15:49:05 mar: read config.log Sep 29 15:49:12 gcc_s.so Sep 29 15:49:16 is missing in linker Sep 29 15:51:16 anyone here have a response to my question on the ml about bluez4 pand/dund/hidd? Sep 29 15:52:04 is there a reason the oe-core mailing strips all the recipients…. grr Sep 29 15:52:48 bluelightning: execute! Sep 29 15:53:27 incandescant: yeah maybe I should just submit a patch to remove them... Sep 29 15:53:41 incandescant: morning btw :) Sep 29 15:53:48 top of the Sep 29 15:54:00 to be sure Sep 29 15:54:01 ;) Sep 29 15:54:32 bluelightning: bluez: seems ok to me Sep 29 15:58:56 and im not subscribed to oe-core three times for all my possible email aliases! yay Sep 29 15:58:59 now* Sep 29 17:07:59 Hi, can't create account in openembedded bugzilla - gives an error. Sep 29 17:07:59 So I'll report here: http://cgit.openembedded.org/cgit.cgi/openembedded/tree/recipes/gdb Sep 29 17:07:59 gdb versions are without 'a' postfix. Currently all tarballs are with 'a' (see http://ftp.gnu.org/gnu/gdb/), you know that GNU-is-GPL-violating situation. Sep 29 17:08:43 it's known Sep 29 17:09:15 send patch if you care about oe-classic to oe-devel ML, oe-core/meta-oe has those issues fixed already Sep 29 17:10:45 ok Sep 29 17:10:48 btw http://pastebin.com/ZwzWfAdK Sep 29 17:10:52 bugtracker output Sep 29 17:11:29 Hmm, it suggests to send email, will do:) Sep 29 17:15:04 Breaking mood... Sep 29 17:18:33 http://pastebin.com/djsRrb2b Sep 29 17:18:36 What the... Sep 29 17:18:58 Can't send email to bugzilla Sep 29 17:19:22 Me having a beta-tester day today. Sep 29 17:19:53 bugzilla should be read-only.. use oe-devel ML to send patches Sep 29 17:20:47 JaMa|Off: It told me: For help, please send mail to the webmaster (bugzilla@openembedded.net), giving this error message and the time and date of the error. Sep 29 17:21:01 I sent that error about not being able to register. Sep 29 17:21:08 And then I get another error:) Sep 29 17:21:22 I see, but just ignore bugzilla, ok? Sep 29 17:21:26 it's not used anymore Sep 29 17:21:30 Hmm Sep 29 17:22:43 Why not take it down then. Me not understand. Sep 29 17:22:58 Anyway Sep 29 17:51:38 anyone tried scipy on oe? Sep 29 18:18:10 I have added PREFERRED_PROVIDER_vim-syntax = "vim" to my image recipe, but I am still getting: Sep 29 18:18:18 NOTE: multiple providers are available for runtime vim-syntax (gvim, vim-tiny, vim) Sep 29 18:18:19 NOTE: consider defining a PREFERRED_PROVIDER entry to match vim-syntax Sep 29 18:18:48 Have I missed something here? Sep 29 18:18:56 davidlt: you could run bitbake -e recipe.bb and see if it's set to what you think Sep 29 18:19:17 davidlt: preferences are set in the configuration metadata, not recipes. set it in your local.conf. Sep 29 18:19:46 kergoth: same result, that's was the next place I tried Sep 29 18:24:05 msm: it is defined in the environment Sep 29 18:26:48 davidlt: its not obvious to me Sep 29 18:27:41 I just checked the output of bitbake with "-e" option and it was defined there Sep 29 18:35:52 msm: any other ideas why it complains? Sep 29 18:37:16 not off the top of my head - i have some of the sames issue to deal with Sep 29 18:37:17 I will try deleting the cache Sep 29 19:29:05 is there some sample system to research in how to do a shared preference service between some native and java application? via like an communication channel or what sort? instead of reading config files from java directly? Sep 29 19:29:17 my google foo is not good enough Sep 29 19:30:06 does anyone know how to attach built ext2 image to qemu? Sep 29 19:30:22 i mean i want to boot os with qemu Sep 29 19:30:35 mar: use the -hda switch? Sep 29 19:30:39 or whatever? Sep 29 19:30:59 i did try using -hda and kernel panics, says it can't find hda Sep 29 19:31:37 you need the appropriate drivers included in your kernel Sep 29 19:31:45 the qemu mang page tells you which hardware it supports Sep 29 19:31:56 or rather provides Sep 29 19:32:08 can I add some recipe to include those drivers in kernel? Sep 29 19:32:30 i am not an OE expert, you need to modify your kernel config Sep 29 19:32:51 so look at whatever builds the kernel Sep 29 20:10:09 does anyone have gcc built on qemu? Sep 29 20:12:37 wrong question, does anyone have gcc built on oe? :) Sep 29 20:12:52 I'm missing libgcc_s.so on my image.. Sep 29 20:18:46 mar: libgcc_s.so should be coming from gcc-cross Sep 29 20:20:29 imk installing task-sdk-native Sep 29 20:20:42 and running gcc on file produces linker error Sep 29 20:20:52 it happens even in my build machine after build Sep 29 20:34:05 Hi all, Sep 29 20:34:41 I noticed there is no bitbake recipe for the linux-omap3_2.6.34 kernel available in openembedded (at least not that I can find) Sep 29 20:35:08 I would like to create a custom recipe to build this version of the kernel Sep 29 20:35:53 I'm guessing I should change PV to be "2.6.39" and SRC_URI to be Sep 29 20:36:12 "git://www.sakoman.com/git/linux-omap-2.6.git;branch=omap-2.6.34;protocol=git Sep 29 20:36:35 sorry.. PV="2.6.34" Sep 29 20:37:06 but what is the meaning of SRCREV ?? does this try to match a particular revision of the source code? Sep 29 20:41:28 cbrake ping? Sep 29 20:41:34 ka6sox: Hey Tom Sep 29 20:41:49 hiya, go ahead..just close it up when you are done...thanks! Sep 29 20:41:59 ka6sox: ok, thanks Sep 29 20:44:30 cbrake: last report is a bit weird.. ie my commits to bitbake are listed in meta-yocto section Sep 29 20:45:00 JaMa: yes, they do something strange with Yocto where they merge a bunch of repos Sep 29 20:45:04 JaMa: so there are duplicates Sep 29 20:45:49 JaMa: hmm, but they are not in the bitbake section Sep 29 20:45:58 cbrake: yup only in meta-yocto Sep 29 20:46:59 JaMa: you changes are in last week's report for bitbake Sep 29 20:47:32 I found out that gcc doesn't create symlink libgcc_s.so -> libgcc_s.so.1 Sep 29 20:47:44 JaMa: so the merge of bitbake -> yocto most have crossed a week boundary Sep 29 20:48:05 but I don't know how to fix it Sep 29 21:46:59 hi, while testing qt 4.7.4 in oe-core using angstrom setup scripts, I rebased my branch and launched a new build. A few tools we compiled then qt tries to configure and fails with "stltest.cpp:47:20: fatal error: iterator: No such file or directory" : do you have an idea where this could comes from ? Sep 30 01:14:53 how does one do an RDEPENDS for /bin/sh? Sep 30 01:15:10 do I need to say bash? or really I just want busybox there as well **** ENDING LOGGING AT Fri Sep 30 02:59:57 2011