**** BEGIN LOGGING AT Fri Dec 21 02:59:59 2012 Dec 21 08:23:19 good morning Dec 21 08:54:20 morning all ! Dec 21 08:55:33 what's the purpose of the CONFFILES variable ? Dec 21 09:08:17 is it possible to print out a text message from a function in a recipe? I tried echo and bbnote but echo works only if the code in the function raises a failure Dec 21 09:09:20 mckoan hm I have something in my openjdk stuff Dec 21 09:09:21 mom Dec 21 09:10:07 morning all Dec 21 09:10:12 mckoan you mean in the logs or bitbake console output? Dec 21 09:10:33 bitbake console output, I need to debug what I am doing Dec 21 09:10:58 in case logs would be enough as well Dec 21 09:11:34 mckoan https://github.com/woglinde/meta-java/blob/master/recipes-core/openjdk/openjdk-7-common.inc Dec 21 09:11:44 look at echo "Configure with parallel-jobs: ${JDK_JOBS}" Dec 21 09:11:52 woglinde_: thx Dec 21 09:11:58 maybee its usefull for you Dec 21 09:12:02 it worked at least for me Dec 21 09:12:27 JDK_JOBS is JDK_JOBS = "${@get_jdk7_jobs(d)}" Dec 21 09:12:40 and get_ bla is python function Dec 21 09:13:08 woglinde_: I have been trying the same in do_configure_prepend() Dec 21 09:13:19 echo "~~~~~~~~~~~~~~~~~~~~~~~~~" Dec 21 09:13:37 but it isn't shown on the console output Dec 21 09:13:52 only if do_configure_prepend() fails Dec 21 09:34:42 I discovered that bbwarn and bbnote works only if I add die after them Dec 21 09:35:28 and echo as well Dec 21 09:39:39 hehe Dec 21 09:49:25 mckoan: is that some sort of flushing issue? Dec 21 09:53:26 florian: ping Dec 21 09:53:30 morning all Dec 21 09:53:41 hrw: pong Dec 21 09:53:43 hi hrw Dec 21 09:54:25 florian: any idea why linuxtogo.org is not responding? Dec 21 09:55:12 hrw: not really... let me see Dec 21 09:55:36 I have customer for my sheevaplug but need to restore it first to working condition Dec 21 10:16:05 jackmitchell: I don't know but looks like bitbake works in such way Dec 21 10:17:08 hrw: I also have 2 sheevaplug to sell :-) Dec 21 10:17:45 mckoan: I have client for just one ;) Dec 21 10:18:10 hrw: did you use ebay? Dec 21 10:18:20 mckoan: does something like: sys.stdout.flush() help on the off chance? Dec 21 10:18:44 mckoan: rather than die/echo Dec 21 10:22:53 mckoan: no Dec 21 10:23:56 | * satisfy_dependencies_for: Cannot satisfy the following dependencies for busybox: Dec 21 10:23:59 | * update-alternatives-cworth * update-alternatives-cworth * Dec 21 10:24:10 bit bitbake it == ERROR: Nothing PROVIDES 'update-alternatives-cworth' Dec 21 10:26:10 opkg has it. Dec 21 10:36:50 When I build my custom distro I get a .ext3 file and a .tar.gz of the root file system. Is there a way I can make oe generate a complete image that I can just dd to my sdcard ? Dec 21 10:39:31 kroon?????? Dec 21 10:39:49 making a partition an dd ext3 to it works Dec 21 10:43:29 Ok, so the .ext3 is an image I should be able to pass to qemu aswell ? Dec 21 10:44:08 easier to pass untared .tar.gz over nfs Dec 21 10:45:10 Sorry, the real problem I'm trying to solve is how to boot my distro with qemu. I pass the zImage and ext3 file to qemu, but qemu crashes with "qemu: hardware error: integratorcm_write: Unimplemented offset 0x6900" Dec 21 10:45:31 Somewhere I read that this was because of an incorrect hd-image being passed Dec 21 10:51:12 ext3 should work Dec 21 10:51:27 kroon look at the oe-core qemu-image starting scripts Dec 21 10:53:51 woglinde, ok will do Dec 21 10:57:59 i wrote a little class to relocate /etc configuration files to /var/etc for read-only rootfs, if someone is interested feel free to comment or review it, it's there : http://pastebin.com/4DKZBNLd Dec 21 10:59:53 the same one with less typos :) http://pastebin.com/7yVv9tx1 Dec 21 11:08:17 afournier mail it to the oe-core ml for review Dec 21 11:10:41 woglinde: you think it's worth it ? Dec 21 11:10:54 hm dont know Dec 21 11:11:28 it's more like a hack to me Dec 21 14:46:51 when you specify a second tar.gz in the SRC_URI, is possible to extract it in a subdirectory ? Dec 21 14:47:16 ${S} instead of ${WORKDIR} Dec 21 14:55:55 mmh... with recent -core, locale files are not installed anymore in the image :( Dec 21 14:56:07 probably caused by ff01518c6667e1b6e87df4e5a435a701d0f38fa7 Dec 21 14:57:20 TARGET_VENDOR should be quoted there Dec 21 15:02:14 is there any documented policy to submit patches to oe-core? Dec 21 15:02:34 yes Dec 21 15:03:28 any link please ;-) Dec 21 15:04:01 http://www.openembedded.org/wiki/How_to_submit_a_patch_to_OpenEmbedded is still valid? Dec 21 15:05:11 http://bit.ly/XUnoek Dec 21 15:05:40 lol Dec 21 15:06:48 you can also find this http://www.openembedded.org/wiki/Commit_Patch_Message_Guidelines with the same great tool called google Dec 21 15:08:29 I love lmgtfy Dec 21 15:09:59 JaMa: nice thx :-D Dec 21 15:10:40 is basically always the same ad per oe-classic Dec 21 15:11:18 I asked that because I see in ML a more structured header for the patches Dec 21 15:11:38 ad actually that is not descriver anywhere Dec 21 15:11:58 I'll try to guess the best solution Dec 21 15:12:52 (when one want to bee too polite may appear silly) sorry for my question Dec 21 15:13:31 mckoan: I guess the best solution is to adopt the guideline for commit messages Dec 21 15:13:31 mckoan: it's described in one of those 2 links, did you read them? Dec 21 15:14:22 in ML, there are some patches that are not as good as described in these two documents. Please don't take ML as a reference to submit a patch Dec 21 15:16:10 I have just realized that when people reference a yocto bugzilla with [YOCTO #12345], the document suggests to include the full URL but no-one does that :) Dec 21 15:16:31 s/no-one/no one/ Dec 21 15:20:28 JaMa: don't you think that commit messages (headers) should use present tense? Dec 21 15:25:09 ops Dec 21 15:28:29 JaMa: ping Dec 21 15:33:12 eren: i prefer commit message in the past, so the ChangeLog is in the past i.e. "Fixed" Dec 21 15:37:24 afournier: git is a distributed tool, and a lot of patches circulate in mailing lists Dec 21 15:38:08 eren: the present is an illusion Dec 21 15:38:11 rather than seeing the commits as a history (fixed that, added foo), I think a commit should be seen as an instruction that changes something Dec 21 15:38:12 eren: pong, sorry my connection died Dec 21 15:38:16 16:21:41 < JaMa> yes and they usually are IMHO Dec 21 15:38:17 16:22:27 < JaMa> but it's true, some people use past tense Dec 21 15:38:35 so, it makes sense to use present tense Dec 21 15:40:18 http://365git.tumblr.com/post/3308646748/writing-git-commit-messages Dec 21 15:41:05 "Git is a distributed version control system where there are potentially many places to get changes from. Rather than writing messages that say what you’ve done; consider these messages as the instructions for what applying the commit will do" Dec 21 15:41:40 he references git rebase and merge using present tense as well Dec 21 15:41:54 http://git.kernel.org/?p=git/git.git;a=blob;f=Documentation/SubmittingPatches;h=ece3c77482b3ff006b973f1ed90b708e26556862;hb=HEAD Dec 21 15:43:05 afournier: I guess we are saying the same think? Dec 21 15:43:11 yes Dec 21 15:43:40 indeed there is a guy here (http://stackoverflow.com/questions/3580013/should-i-use-past-or-present-tense-in-git-commit-messages) that clarifies it well Dec 21 15:43:45 you just pasted a guideline that says "use imperative, past tense" :) Dec 21 15:44:11 at the same time, you said "present is an illusion" Dec 21 15:44:19 i said i personnaly prefer using the past, but the git guidelines say use the present Dec 21 15:44:33 git commit are optional => present, svn commit are history => past Dec 21 15:44:35 oh okie Dec 21 15:45:17 JaMa: my point is that why do we suggest using past tense im commit message headers in Guideline? Dec 21 15:45:20 I'm referring to wiki Dec 21 15:45:51 the examples uses past tense: " rpm: Adjusted the foo setting in bar Dec 21 15:46:08 "foobar: Adjusted the foo setting in bar", etc Dec 21 15:46:19 maybe oe started on svn ? :) Dec 21 15:46:28 or cvs... Dec 21 15:46:53 eren: ask someone who wrote that guildelines Dec 21 15:47:11 alright sir :) Dec 21 15:47:19 it was discussed in lenght in TSC iirc Dec 21 15:48:19 fray: ping Dec 21 15:49:31 ensc|w: I'm looking at your commit http://patchwork.openembedded.org/patch/39177/, isn't pipes.quote deprecated? Dec 21 15:50:25 JaMa: dunno; what's the recommended way to quote things in oe? Dec 21 15:52:45 dunno :) Dec 21 16:32:43 morning Dec 21 17:09:28 JaMa: morning Dec 21 17:09:39 JaMa: https://github.com/meta-qt5/meta-qt5/commit/1ecc46ac42a685057c94e1bd30147ea81ab925bf Dec 21 17:09:53 JaMa: that did noty break the build? Dec 21 17:15:22 we cannot touch STAGING_DATADIR Dec 21 17:15:42 and there are only those 2 issues I've pasted before Dec 21 17:18:28 JaMa: right, but did it still work? Dec 21 17:18:34 i.e. qtbase Dec 21 17:19:39 JaMa: oh wait... after making that change the qtjsbackend broke... or... Dec 21 17:25:26 levonmaa: it's the same file, isn't it? Dec 21 17:25:49 so it's just fixed in source before it's installed to sysroot Dec 21 17:26:16 this is also invalid: Dec 21 17:26:17 ### Remove the old pri files Dec 21 17:26:18 62 61 rm ${STAGING_DATADIR}/${QT_DIR_NAME}/mkspecs/*.pri Dec 21 17:27:21 and this too: Dec 21 17:27:22 mkdir -p ${STAGING_DATADIR}/${QT_DIR_NAME} Dec 21 17:27:22 cp -r ${STAGING_DATADIR_NATIVE}/${QT_DIR_NAME}/mkspecs ${STAGING_DATADIR}/${QT_DIR_NAME}/ Dec 21 17:27:44 you'll see a lot of warnings from sstate which will be changed to fatal later Dec 21 17:28:20 we just need to find better fix Dec 21 17:40:24 JaMa: righty-o... I guess I have to think abt that while enjoying x-mas ;) Dec 21 17:59:33 levonmaa: I hope to find some time for qt5 between xmas and year++ Dec 21 22:30:09 JaMa: ping Dec 21 22:31:37 JaMa: is there a way to have version numbers in the DEPENDS line, i.e. DEPENDS += " pkg-name >= x.y" Dec 21 22:32:54 levonmaa no Dec 21 22:33:07 woglinde: crap ;) Dec 21 22:33:20 sorry Dec 21 22:33:28 any ideas on how one would mimic that Dec 21 22:34:14 there is Dec 21 22:34:20 wait.. not DEPENDS Dec 21 22:34:23 but runtime you can.. Dec 21 22:34:56 that's too late ;) Dec 21 22:35:16 it's the responsibility of the distribution configuration to ensure a sane system (version wise included) Dec 21 22:47:01 levonmaa only via preffered_version **** ENDING LOGGING AT Sat Dec 22 02:59:58 2012