**** BEGIN LOGGING AT Thu Apr 11 02:59:59 2013 Apr 11 05:34:48 morning @ all Apr 11 06:29:31 Hi~ How bitbake setup the dependency for a specified task ? Apr 11 06:30:19 I mean where is the python code to parse the addtask keyword. Apr 11 07:29:50 good morning Apr 11 07:49:33 good morning mckoan ;) Apr 11 08:21:03 morning all Apr 11 08:22:57 hi bluelightning, mattnie, all Apr 11 08:25:51 morning ;) Apr 11 08:26:13 bluelightning: morning -- you have read it? http://lwn.net/Articles/546900/ Apr 11 08:26:48 silviof: yes saw that late last night Apr 11 08:51:42 Hi everybody, hi otavio. I'm having a problem in X startup using the latest build from master for nitrogen. I get this log: Apr 11 08:51:51 http://pastebin.com/JjzDCpEH Apr 11 08:52:33 It's stays there on "VIVANTE(0): test Initializing EXA" and doesn't move. Apr 11 08:54:03 It used to work after the patches to revert to xserver 1.11, but doesn't work from up-to-date master (even before moving to the boundary kernel) Apr 11 08:54:51 so I think there's nothing wrong in meta-fsl-arm Apr 11 08:55:20 I meant meta-fsl-arm-extra Apr 11 08:57:32 I thing the problem might be between http://git.yoctoproject.org/cgit/cgit.cgi/meta-fsl-arm/commit/?id=666485e1835aaa34518464265980acfee250783d excluded and http://git.yoctoproject.org/cgit/cgit.cgi/meta-fsl-arm/commit/?id=ea6848fca73128a3e39c4488145cb306e43d5c47 included (even I'd exclude that patch, as it's related only to sabrelite) Apr 11 08:58:37 Can anybody try to help me debug the issue? Apr 11 08:59:06 (screen is blank) Apr 11 11:02:08 is there an "easy" way to resolve dependency issues in kernel? e. g. with deactivating networking? at the moment it is boundless... Apr 11 11:37:51 hi, I'm packaging tclreadline (recipe: http://pastebin.com/5vFTjdhS) but I can't get so shared libs in the resulting core-image-minimal Apr 11 11:39:06 the files are in the tclreadline-dev rpm but I'm not sure I have to explicitly add it in the IMAGE_INSTALL_append Apr 11 11:39:32 of course I'm a newbie, so I'd be happy to RTFM if you point me to the right direction:) Apr 11 11:42:27 david-e_: the libs shouldn't be in the -dev, just the .so symlinks Apr 11 11:43:44 rburton: that's the way it ended up without any explicit rule. I've tried to FILES_${PN} = "${libdir}/*.so" with no success Apr 11 11:44:11 david-e_: that's not correct though, .so symlinks should not be in the main package, that's on purpose Apr 11 11:44:40 david-e_: as i said, .so symlinks are meant to be in -dev. the main package should contain the library itself. Apr 11 11:44:44 (applications are supposed to link to the versioned libraries so that things don't break) Apr 11 11:44:52 (break on upgrades, I mean) Apr 11 11:45:16 ok about the symlink, but why /usr/lib/libtclreadline-2.1.0.so is not there then? Apr 11 11:45:50 rpm -qlp tmp/deploy/rpm/ppce500v2/tclreadline-dev-2.1.0-r0.ppce500v2.rpm shows it Apr 11 11:48:15 because its only for development. if you want development packages installed, then install them explicitly or use the dev-pkgs image feature to get all the relevant ones installed automatically. Apr 11 11:51:42 I've changed IMAGE_INSTALL_append to " tcllib tclreadline tclreadline-dev" and I get dev files also from other packages Apr 11 11:52:51 things like ./usr/include/bits/stdlib-float.h are nor in the image Apr 11 11:52:51 s/nor/now/ Apr 11 11:59:59 hey, im trying to build a simple clutter application for the core-image-clutter but when i try to build it i get an error saying it cannot find some libraries ("cannot find -l*" (clutter-glx-1.0, cogl-pango, json-glib-1.0, cogl). pkg-config finds clutter-1.0 but it doesn't find these libraries. what am i doing wrong? Apr 11 12:19:30 greakv: build on the target, or build in bitbake? Apr 11 12:27:58 bitbake Apr 11 12:28:19 but i think i managed to fixed the problem Apr 11 13:02:47 https://wiki.yoctoproject.org/wiki/Help:Editing looks broken Apr 11 13:06:51 Zagor: thanks, I've filed a bug against the wiki Apr 11 13:47:50 Is yocto@yoctoproject.org subscribing open, broken and/or just slow? Apr 11 13:48:14 mcfrisk: people have been saing that delivery is slow at times Apr 11 13:49:00 tried to send patch, heard nothing, noticed subscriber only postings, tried to subscribe, still waiting :) Apr 11 13:50:56 * Crofton|work beats halstead Apr 11 13:51:28 Hopefulyl he says this and checks into the mailing lists Apr 11 14:06:47 is anyone else seeing spikes of really high sys load during builds? I took this nmon snapshot the other day: http://i.imgur.com/xHd997Q.png Apr 11 14:07:35 I'm trying to figure out what it is. it doesn't appear to be context switches. sar doesn't show an unusual number when this happens. Apr 11 14:08:25 https://wiki.yoctoproject.org/wiki/Yocto_1.4_Milestone_Test_Report#Yocto_1.4_M5_RC3_Build_Test_Report Apr 11 14:08:54 https://wiki.yoctoproject.org/wiki/Yocto_Project_v1.4_Status#Milestone_5 Apr 11 14:09:01 Song_liu: that was Ema Apr 11 14:10:58 Song_liu: 3672 Apr 11 14:25:34 oh, 1.4 is coming, brace yourselves Apr 11 14:26:18 well 1.4M5... Apr 11 14:26:19 ;) Apr 11 14:26:52 :) Apr 11 14:27:13 Song_liu: https://wiki.yoctoproject.org/wiki/Performance_Test Apr 11 14:31:18 Song_liu: No objections to what RP said here Apr 11 14:35:36 RP_: can yup please merge that icecc or should I send it for backport later? Apr 11 14:39:16 JaMa: ah, yes Apr 11 14:39:24 thank you Apr 11 14:39:45 another reason for me to propose upgrade from 1.3 to 1.4 :) Apr 11 14:40:13 btw: did anyone benchmark scalability and performance of PR service? Apr 11 14:40:59 or is there some plan to provide "public" PR service at least to see how it will behave with 100+ builders start accessing it in parallel? Apr 11 14:41:16 JaMa: have you a link handy to the icecc patch? Apr 11 14:41:26 JaMa: I think this is why I keep forgetting :/ Apr 11 14:41:58 http://patchwork.openembedded.org/patch/46369/ or do you prefer git hash from contrib? Apr 11 14:43:13 JaMa: that is great thanks Apr 11 14:55:50 mcfrisk, I'm happy to look into that for you. Can you please share which e-mail you used and if you subscribed via e-mail or the web form? Apr 11 14:56:16 halstead, thanks Apr 11 14:56:39 Crofton|work, Any time. :) Apr 11 14:57:01 halstead: used web interface to subscribe, email is mikko.rapeli@iki.fi Apr 11 14:59:49 halstead: also did not receive any bounces or such when sending to yocto@ without being subscribed, I also send email about this to yocto-owner Apr 11 15:01:01 rburton: would you mind maintaining https://wiki.yoctoproject.org/wiki/Releases also on OE wiki? I think many users search "releases" there too and the results are not very accurate :/ Apr 11 15:01:50 mcfrisk, I see one post from you caught for moderation and a notice sent to yocto-owner. The issue is with our DNS and the iki.fi record. Fixing that. Apr 11 15:31:47 Zagor: that is interesting. We have made some parts more efficient and that could mean we're starting to see points where it does hammer the system Apr 11 15:33:48 Zagor: All sys load I notice which is interesting Apr 11 15:34:01 Zagor: any idea which tasks are triggering it? Apr 11 15:34:38 it's during heavy times of building, when multiple devs are hammering the machine with wide parallell builds. the top cpu users are usually gcc and python Apr 11 15:35:06 I am clueless what the kernel is busy doing Apr 11 15:35:15 Zagor: io? Apr 11 15:35:41 but does io really show up as sys time like that? Apr 11 15:35:56 shouldn't it rather be "wait" time? Apr 11 15:36:05 Zagor: right :/ Apr 11 15:36:15 Zagor: could it be gcc piping data between processes? Apr 11 15:36:39 Zagor: python is probably something different too Apr 11 15:37:14 Zagor: obviously gcc is do_compile, any idea what python was doing? do_package? Apr 11 15:37:44 I know when we've tested do_package with -P, you see it spending a lot of time in stat calls Apr 11 15:37:56 no. probably just general bitbake running Apr 11 15:38:47 I'm struggling with ways to measure/log/instrument this Apr 11 15:39:59 message passing (log data) to the UI? Apr 11 15:40:58 I can't see it being something so trivial. we are talking 64 2.5GHz processors spending 90% of their time in the kernel. it's... a lot of work Apr 11 15:51:28 when pseudo is being used over large data sets, it can generate a huge number of messages.. Apr 11 15:51:52 large data sets, i.e. 100,000 files+ .. lots of 'stat' operations.. etc Apr 11 15:59:17 ugh, I think bitbake's repochanged handling in bb.fetch.git is broken. it sets a flag when it updates the repository, and uses that to deterimne when to write the tarball, but what if the task gets interrupted between those two steps? Apr 11 15:59:28 now you have an updated git2 repo but an out of date mirror tarball Apr 11 15:59:49 i have a setup in that state now, the repo is updated but the tarball is not Apr 11 16:02:28 Also seems potentially racy when you have multiple bitbake builds sharing a DL_DIR, unless that entire block is protected by a lock (haven't checked) Apr 11 16:02:47 I was under the impression it -was- protected... Apr 11 16:03:07 * kergoth shrugs, if so there's just the interruption to worry about Apr 11 16:11:57 I suspect that there's also issues with switching the value of BB_GENERATE_MIRROR_TARBALLS. if the repo was updated in a prevoius build with it disabled, but you had a tarball from the past existing, then the build after enabling it wouldn't recreate the tarball even though its out of date Apr 11 16:12:15 so, yeah, multiple problems with this logic afaict Apr 11 16:12:20 whee! Apr 11 16:12:33 OT: Is there a page for end products that use Yocto/Poky or any Yocto compatible product? Any possibility to have my sysmoBTS added to this page? Apr 11 16:12:35 * kergoth adds todo to open bug Apr 11 16:12:46 s/bug/a bug/ Apr 11 16:13:09 zecke, there is YP compliance program.. part of that program allows you to be a registered "participant".. you can then register your usage/products that use it Apr 11 16:13:20 * fray goes to find hte link Apr 11 16:13:56 https://www.yoctoproject.org/ecosystem/yocto-project-participants Apr 11 16:14:10 mcfrisk, I've tracked it down to a problem with our DNS forwarders and DNSKEY on iki.fi. Apr 11 16:15:02 halstead: great, should I repost or try subscribing? Apr 11 16:15:38 mcfrisk, Probably not. It will go out once this is resolved. Apr 11 16:16:29 fray: thanks Apr 11 16:16:59 halstead: thanks! Apr 11 16:21:29 mcfrisk, Confirmation e-mail was accepted for delivery. Apr 11 16:34:57 halstead: thanks, received all the emails now Apr 11 16:38:05 halstead: i'm being moaned at when i push to poky-contrib to run "git prune" Apr 11 16:38:59 rburton, That is supposed to happen automatically on the server. I'll see what's up. Apr 11 17:20:54 is there a port convention for prserv? Apr 11 17:21:44 ka6sox: are you goint to set that public instance? Apr 11 17:22:34 JaMa, I was thinking about but wanted to be standard and not go off and setup things that break other builds. Apr 11 17:23:33 ka6sox: ok, good Apr 11 17:25:30 pidge, is there a prserv standard port? Apr 11 17:38:13 ka6sox: I don't believe so. It's whatever is set in PRSERV_HOST. Apr 11 17:40:06 pidge, since we are going to use a public server I'd like to have a standard port (usable outside too) Apr 11 17:40:19 would you like me to find a suitable port? Apr 11 17:43:40 ka6sox: ping halstead on port numbers. What are the details on this? I'm not in the loop and I'm curious what the plan is. Apr 11 17:45:06 ka6sox, Please let me know what port number you'd recommend. Apr 11 17:46:58 halstead, will do...let me do some research. Apr 11 18:18:51 I'd really appreciate anyone with a centos 5 build host testing the fix in this bug: https://bugzilla.yoctoproject.org/show_bug.cgi?id=4245 Apr 11 18:18:52 Bug 4245: normal, Medium, 1.4 M6, laurentiu.palcu, IN PROGRESS IMPLEMENTATION , wayland fails to compile on Centos 5.x Apr 11 18:36:44 rburton, poky-contrib should be all clean now. Apr 11 22:23:57 zedii: online? Apr 12 00:03:25 zeddii_home: there ? Apr 12 00:03:52 I am looking for packaging up initramfs into kernel itself, do we have support for such a thing in yocto/oe-core Apr 12 00:04:02 last time I used it was with oe.classic Apr 12 00:04:07 it worked there Apr 12 00:04:13 kernel was built couple of time Apr 12 00:04:14 s Apr 12 00:04:30 hi khem Apr 12 00:04:43 mranostay: man howdy Apr 12 00:04:49 you won the cape prize Apr 12 00:04:56 we should party Apr 12 00:05:10 heh Apr 12 00:06:25 * mranostay looks at pollen chart and cries Apr 12 00:10:59 oh so its the smog thats killing me Apr 12 00:11:08 I could not go to work today Apr 12 00:11:23 felt so sick Apr 12 00:11:32 wb zenlinux Apr 12 00:11:45 it's good to be back ;) Apr 12 00:11:55 * zenlinux just switched computers Apr 12 00:13:52 zenlinux: how it up there on the farm? :P Apr 12 00:14:32 you could say things are going swimmingly Apr 12 00:16:55 im getting a sanity check problem... it says it cant download test data from network Apr 12 00:17:28 but I checked the CONNECTIVITY_CHECK_URIS variable in my distribution and I could download both files using wget Apr 12 00:17:39 any thoughts? **** ENDING LOGGING AT Fri Apr 12 02:59:58 2013