**** BEGIN LOGGING AT Wed May 07 02:59:58 2014 May 07 05:02:29 khem, http://pastebin.com/J4kAd25B May 07 06:06:06 is there a way to clean all kernel related packages ? May 07 07:17:04 good morning May 07 07:36:58 sync May 07 07:37:04 sory May 07 07:47:38 qt-x: no need to be sorry, its important to be in sync! May 07 07:48:10 :D May 07 07:53:58 and zipped May 07 08:40:34 morning all May 07 08:40:47 hi bluelightning, all May 07 08:46:15 hi mckoan May 07 10:53:28 hi all May 07 10:56:24 hi matteo May 07 14:41:06 top-o-the-moin-in... May 07 14:57:17 * darknighte yawns May 07 14:57:28 it's 10am.. slacker May 07 14:58:52 * kergoth yawns also May 07 14:59:08 lol, dunno what timezone you are in.. May 07 14:59:12 I do know Seans.. ;) May 07 14:59:35 all this yawning is contagious May 07 14:59:42 not good when i've got a call with my manager in 30 seconds! May 07 15:00:48 * nerdboy yawns May 07 15:00:52 * BCMM yawns May 07 15:01:00 stop that May 07 15:01:05 i think you fell asleep for a moment. May 07 15:01:37 darknighte: pointer coming up... May 07 15:02:54 http://www.youtube.com/watch?v=4n99CFX_CVY <= dave and mike doing an actual video May 07 15:05:44 that's dave with the nice telecaster May 07 15:06:36 fray: was a long night... May 07 15:07:00 uhh-ohh May 07 15:07:08 * nerdboy not a country guy but it's still good... May 07 15:09:00 and mike's cowboy dance is.. May 07 15:09:09 not sure what that is May 07 15:20:14 * nerdboy 's wife complains about too much country music... May 07 15:25:24 hi there. I'm building a Poky based system using apt for package management, and I would like to have an apt repo where I can put updates, so the devices can get them. I can do the device side of things (adding a sources.list entry should be enough), but is there any support in Poky for building the repo in apt format from the deb files? May 07 15:25:38 (I tried the #poky channel but got no answer there) May 07 15:27:19 PulkoMandy: if you host your repository on a debian machine you can use tools like apt-ftparchive or reprepro to create the apt index files. May 07 15:28:09 yes, just wanted to know if there was some built-in way to do it, otherwise I'll write some scripts myself May 07 15:28:37 You will need to generate 'deb' packages in your Poky configuration.. then as andhe said, use a debian machine (or another suitably equiped one) to serve the feeds May 07 15:28:57 the feed management is not provided currently within OE/YP.. May 07 15:29:07 only package generation and on-target updates May 07 15:31:46 ok, thanks! May 07 15:31:57 PulkoMandy: the deploy/ipk dir actually does work as a feed (i serve it up with apache) May 07 15:32:09 haven't tried debs though May 07 15:32:53 what happens when you point your browser to .../deploy/deb or whatever? May 07 15:33:23 er, need more coffee... May 07 15:33:58 tried setting it in an opkg feed and serving it with apache? May 07 15:34:06 er, apt feed May 07 15:34:47 * nerdboy should try debs or rpms again... May 07 15:39:54 it doesn't look like a standard apt repo May 07 15:41:27 PulkoMandy: the image is built using apt, so it's a standard repo May 07 15:41:58 try serving it with apache/nginx/whatever May 07 15:42:23 ok, I'll try that May 07 15:57:36 PulkoMandy: the ipk on looks like this May 07 15:57:55 http://www.gentoogeek.org/bbb-ipk/ May 07 15:58:13 not sure if bbb is visible yet May 07 15:58:35 http://www.gentoogeek.org/rpi-ipk/ should work May 07 16:00:22 nerdboy: if you or someone you know has made a distro for BeagleBone, I recommend you register on http:/beagleboard.org/project May 07 16:00:53 already did May 07 16:01:18 someone i know is me May 07 16:01:56 nerdboy: I'd be concerned if you didn't know yourself ;) May 07 16:02:21 about the only person i know well... May 07 16:05:28 fray: fyi, looks like you can use sstate-cache-management.sh with —stamps-dir= instead of an sstate mirror to pare down SSTATE_DIR, but you still need to do a second build, to get a stamps dir which is from a from-sstate build (e.g. no do_package stamps). https://gist.github.com/kergoth/3d73977975b17ea272ac — only a slight improvement over the second build + sstate_mirrors configuration, but it's slightly less setup. going to be altering our May 07 16:05:29 automated builds to use that process, most likely May 07 16:21:19 does someone know why curl compiles against gnutls? May 07 16:21:40 not offhand. we need to improve the ability to select the ssl implementation in general May 07 16:21:47 it's not hte only case wher eopenssl or gnutls are options May 07 16:22:03 but packageconfig isn't ideal, since this is isn't exactly boolean May 07 16:23:31 curl-native compiles against (open)ssl while the default builds against gnutls May 07 16:24:09 would be nice to switch mostly to one or the other, to try to pare down the dependency graph as much as possible May 07 16:24:41 could be license related, openssl has an exception, iirc May 07 16:24:43 * kergoth shrugs May 07 16:27:20 Hi, From fresh install, branch master or daisy with MACHINE ?= "beaglebone" having elfutils_0.155.bb, do_compile) failed with exit code '1' any ideas? May 07 16:30:38 kapare: is that elfutils or elfutils-native failing? and can you pastebin the actual error? May 07 16:31:09 bluelightning, ERROR: Function failed: do_compile (log file is located at /home/knight/yocto-beaglebone/build/tmp/work/x86_64-linux/elfutils-native/0.155-r0/temp/log.do_compile.15992) May 07 16:31:28 so elfutils-native May 07 16:31:39 that still doesn't tell me a lot... May 07 16:34:19 bluelightning, sorry delete my build folder and relaunch core-image-minimal will let you know but let me gives you my output... May 07 16:34:24 The default qemu build seems to create more as needed: .ext3, .tar.bz2, modules* May 07 16:34:44 kapare: you don't need to delete anything, that won't solve this May 07 16:35:30 bluelightning, http://pastebin.com/xWpMhyyz May 07 16:37:00 bluelightning, my new build gave me the same error. The error is poping fast May 07 16:37:17 kapare: yes, as I said rebuilding won't fix this May 07 16:37:27 kapare: what host distribution are you running? May 07 16:37:58 Ubuntu 12.04 May 07 16:38:02 * nerdboy enjoys creative commons licensed jazz May 07 16:38:27 kinda like free beer tastes better May 07 16:38:43 kapare: that is odd, because AFAIK we test on that... May 07 16:38:48 bluelightning, more precisely Ubuntu 12.04.4 LTS May 07 16:38:50 especially when the beer is *really* good May 07 16:39:59 bluelightning, the only thing I found on the web is http://patchwork.openembedded.org/patch/58625/ May 07 16:40:51 I don't think that's the same error May 07 16:40:52 bluelightning, but this seem old and the patch is in not exactly the same but there is a fix-build-gcc-4.8.patch May 07 16:41:28 bluelightning, That's what I thought and why I'm asking ;) May 07 16:45:22 bluelightning, hmm forget about that could be my problem: gcc (GCC) 4.10.0 20140415 (experimental) May 07 16:46:25 ah yeah that wouldn't be the standard compiler for Ubuntu 12.04.4 LTS ;) May 07 16:55:07 bluelightning, back to gcc (Ubuntu/Linaro 4.6.3-1ubuntu5) 4.6.3 ;) and except couple of warning everything seem to compile nicely thx! May 07 16:56:29 kapare: great :) May 07 16:57:28 * nerdboy performs another cross-pollination experiment May 07 16:57:45 let's see if anyone takes the bait... May 07 16:59:09 bluelightning, just curious when you test do you run other compiler version as for example gcc 4.7, 4.8 4.9 ? May 07 17:01:39 kapare: we almost always run the host distribution's gcc version May 07 17:01:48 for some distros that is 4.8.x May 07 17:19:17 Hi folks - upgrading to the latest release, I'm getting the following when building -c populate_sdk; is this a known issue? May 07 17:19:18 WARNING: QA Issue: ELF binary ‘/path/to/yocto/build/tmp-eglibc/work/x86_64-nativesdk-foo-linux/nativesdk-qemu/1.7.0-r0/packages-split/nativesdk-qemu/opt/goose/0.1/sysroots/x86_64-foo-linux/usr/bin/qemu-system-i386' has relocations in .text May 07 17:25:14 wotte: it's a known issue yes: https://bugzilla.yoctoproject.org/show_bug.cgi?id=6104 May 07 17:25:15 Bug 6104: normal, Medium+, 1.6.1, cristian.iorga, ACCEPTED , Possible false positive QA warnings while building May 07 17:26:05 Thanks bluelighting. I'll disable that QA check then for the time being. May 07 17:27:51 FWIW, it is just a warning, not an error - so it won't be preventing your build from successfully completing May 07 17:30:13 bluelightning, did you see these? May 07 17:30:14 http://community.validation.linaro.org/dashboard/image-charts/OpenEmbedded-CURRENT-bbb May 07 17:30:29 http://community.validation.linaro.org/scheduler/alljobs?search=Open&length=25#table May 07 17:30:40 their current tests appear to assume debs though May 07 17:31:59 Crofton: I hadn't, thanks May 07 17:32:18 we need to fix the tests May 07 17:32:22 or help them fix them May 07 17:32:39 akbennett, was also asking about what x86 images could be tested May 07 17:36:04 Crofton, yeah, we ran into some issues trying to boot the x86 iso images, hanging, waiting for removable media on virtualbox/kvm etc... Want to make sure we are starting with the right images May 07 17:36:45 pidge might know, but looks like she isn't here May 07 17:37:37 I believe the ISOs don't currently support running under emulation May 07 17:37:38 https://bugzilla.yoctoproject.org/show_bug.cgi?id=2549 May 07 17:37:39 Bug 2549: enhancement, Low, 1.7, dvhart, NEW , [USE CASE] Support live images on emulated targets (hddimg for kvm) May 07 17:40:29 and this is the thanks i get... May 07 17:40:32 bluelightning, https://secure.flickr.com/photos/linuxfoundation/14081148681/in/set-72157644040774060/ May 07 17:41:01 i bumped it in yocto a few months back May 07 17:41:01 nerdboy: traitor May 07 17:41:01 poopy May 07 17:41:12 heh May 07 17:43:52 hmm, formosa streets is about 15 minutes by bike from ehre May 07 17:46:41 isn't a bit early for formosas? May 07 17:47:19 just checking locations May 07 18:14:08 * nerdboy checks for Crofton's joke ciruit May 07 18:40:40 Crofton, are there specific high-value tests that we should target to get running in LAVA May 07 18:42:12 I'm not sure May 07 18:42:24 knowing the image boots is a good start May 07 18:42:29 checks for kernel regressions May 07 18:42:54 internally I'd like to use it for testing my current project to watch for regressins May 07 18:43:04 yeah, a good start May 07 18:43:25 Bill setup a list for testing May 07 18:43:32 and you've got specific hw May 07 18:43:36 I shoudl send the LAVA dashbaod for the bone to it May 07 18:43:43 does that run daily now? May 07 18:43:45 yep May 07 18:43:58 unreleased, but a "normal" daul core A9 May 07 18:44:02 it is running, but doesn't look like a build showed up. May 07 18:44:20 pidge sent an email about the builds being done earlier, maybe they were late May 07 18:46:24 hw integrations are more controlled by the boot procedure, network support in the bootloader, console stability, power on behavior. If that's all standard, then it's a bit more straight forward. May 07 19:02:30 crap, forgot about that too May 07 19:02:59 lava,toaster.. toaster,lava.. May 07 19:03:11 * nerdboy flips a coin May 07 20:07:47 Hi! I', trying to build the moc music player for my raspberry pi but I get an error regarding libltdl. I'm not sure what to do. May 07 20:08:19 | libtoolize: putting libltdl files in LT_CONFIG_LTDL_DIR, `libltdl'. | libtoolize: `COPYING.LIB' not found in `/home/parallels/raspy/yocto/poky/build/tmp/sysroots/raspberrypi/usr/share/libtool/libltdl' | autoreconf: libtoolize failed with exit status: 1 | ERROR: autoreconf execution failed. | WARNING: exit code 1 from a shell command. | ERROR: Function failed: do_configure (log file is located at /home/parallels/raspy/yocto/poky/b May 07 20:08:36 i'll post the error on paste.bin ,... May 07 20:09:17 http://pastebin.com/wUK7CZT1 May 07 20:10:07 it seems that COPYING.LIB could not be found... how do I get it? May 07 20:16:29 its likely you need to add libtool to your DEPENDS May 07 20:20:06 well that was too easy ;-( May 07 20:21:01 is there any change to get of this annoying "Licensing Error" thing? May 07 20:21:50 currently I have to download all archives by hand and look for the license files and md5sum them May 07 20:22:36 and add the LIC_FILES_CHKSUM to the recipe May 07 20:24:39 the archives are already in your download directory for one. for two, they're already extracted by bitbake as well. see bitbake -e yourrecipe | grep '^S=' to get the path to the extracted sources. alternatively, use bitbake -c devshell recipe to drop yourself into a shell in its source tree May 07 20:25:06 hmm May 07 20:31:21 kergoth, thanks for the command but it seems that there are files missing in the directory May 07 20:31:35 I'm missing files like README, NEWS, ... May 07 20:32:20 there are only *.c and *.h files May 07 20:33:10 maybe the recipe grabs the archive from another mirror, but the recipe only links to sourceforge May 07 20:43:03 Crofton, can you point me to the location where you believe we assume debs? I'm not finding any in my review of the tests we put on the bbb. May 07 20:44:20 hang on May 07 20:45:31 no problem, just found some time to look into them May 07 20:46:10 /lava/bin/lava-installed-packages: line 5: dpkg-query: command not found May 07 20:48:17 this may be a red herring May 07 20:49:54 yeah, we use that to query the installed software packages on a debian system during an executed test, not as important for prebuilt image like oe. not a failure though May 07 20:50:17 the test toolchain.yaml is more or less useless though May 07 20:50:38 I've removed it from future execution May 07 21:09:39 akbennett, getting a set of tests that passes is a good start May 07 21:09:48 hopefully we can get more eyes on it May 07 21:10:03 and people thinking about creating tests specific to our needs May 07 23:42:43 If I'm trying to add a kernel config fragment, do I have to do anything other than put it in SRC_URI in linux-yocto_3.10.bbappend? May 07 23:43:04 maxtothemax: yes, you have to set FILESEXTRAPATHS May 07 23:43:25 did that May 07 23:43:57 here's my append file: May 07 23:43:58 FILESEXTRAPATHS_prepend := "${THISDIR}/files:" May 07 23:43:58 SRC_URI_append = " file://ftdi_sio.cfg" May 07 23:45:48 should be fine as long as you put ftdi_sio.cfg in the right place (in the files/ subdir in the dir the bbappend is in). what's the behavior? an error, or just not applying your change? May 07 23:46:20 module doesn't wind up in the image May 07 23:46:51 trying again hold on May 07 23:48:57 it would show up in build/tmp/work/minnow-poky-linux/linux-yocto/3.10.35+gitAUTOINC+7df9ef8ee4_cee957655f-r0/linux-minnow-standard-build/.config right? May 07 23:50:11 maxtothemax, look in the kernel source dir under .meta May 07 23:50:23 you'll find some logs that tell you specifically what was applied May 07 23:50:30 it's possible you are missing a config dependency May 07 23:50:39 in 1.7 we'll restore these warnings to the console May 07 23:50:49 oh, you have to include config dependencies as well? May 07 23:50:53 they were deemed too noisy earlier, but now... now you get this May 07 23:50:59 that's almost certainly my problem May 07 23:51:21 maxtothemax, of course, it's not possible to do that automatically (as many depends are logical ORs) May 07 23:51:48 last time I compiled a kernel was probably 3-4 years ago May 07 23:51:52 ok, I'll try that May 07 23:52:33 interesting, FTDI_SIO depends on CONFIG_EXPERIMENTAL May 07 23:53:31 for example: May 07 23:53:42 /build/yocto/daisy/intel-corei7-64_20140429134804/build/tmp/work/corei7-64-intel-common-eywa-linux/linux-yocto/3.14.0+gitAUTOINC+6201e4c7c6_c8de8c8856-r0/linux/.meta May 07 23:54:09 and the rest is here: May 07 23:54:16 /build/yocto/daisy/intel-corei7-64_20140429134804/build/tmp/work/corei7-64-intel-common-eywa-linux/linux-yocto/3.14.0+gitAUTOINC+6201e4c7c6_c8de8c8856-r0/linux/.meta/cfg/standard/intel-corei7-64 May 07 23:54:54 right, I found those files, but I'm going to scattershot guess that the deps were the problem without looking at them yet :) May 07 23:55:00 files like redefinition.txt, invalid.txt, unknown.txt, and mismatch.txt will help May 07 23:55:05 yup May 07 23:55:09 very likely May 07 23:58:32 yep, it's now showing up in .config May 08 00:09:46 I think linux-yocto's do_package could be parallelized a bit, assuming that also packages up the modules May 08 00:12:51 maxtothemax, watch where you step, that there is a *minefield* May 08 00:13:19 my favorite May 08 00:13:19 but yes, it's been being slowly improved over time, and still on the list for improvements in 1.7 May 08 00:14:47 yay, I have an ftdi-sio-xxx.rpm May 08 00:16:37 great May 08 00:18:22 is this list of improvements actually written down somewhere? May 08 00:19:33 Yes, several in bugzilla, several in the 1.7 kernel planning page, and a few queued up on me to get added to each. May 08 00:28:08 * nerdboy looks around for belen May 08 00:28:13 that was too easy... May 08 00:28:33 definitely not obtuse enough May 08 00:29:58 * nerdboy grumbles about too easy designs taking all the hair-pulling out of it... May 08 02:54:04 zhenhua: May 08 02:54:14 zhenhua: May 08 02:54:20 https://wiki.yoctoproject.org/wiki/Yocto_1.7_Schedule **** ENDING LOGGING AT Thu May 08 02:59:58 2014