**** BEGIN LOGGING AT Tue Apr 29 02:59:58 2014 Apr 29 08:38:17 morning all Apr 29 11:01:12 HI all Apr 29 11:01:59 i add qt5 layer in my BSP and i am getting error after adding "qtwebkit " Apr 29 11:02:10 Error: qtwebkit-qmlplugins not found in the base feeds (lec3517 armv7a-vfp-neon armv7a-vfp armv7a armv6-vfp armv6 armv5e-vfp armv5e armv5-vfp armv5 armv4 arm noarch any all) Apr 29 13:17:17 Hi. Any reason why the default kernel for the genericx86 and genericx86_64 is held back to 3.10 in daisy? Apr 29 13:18:28 Hi. I'm trying to write a recipe for dhcpcd (http://roy.marples.name/projects/dhcpcd/home). It builds, but it builds using gcc instead of arm-poky-linux-gnueabi-gcc as defined by my bitbake variables. It does this because it can't find arm-poky-linux-gnueabi-gcc during ./configure. Is there a way to set PATH in do_configure like it's set during do_compile? Apr 29 13:18:41 zeddii: ^ Apr 29 13:19:04 Denwid: it should already be set correctly... Apr 29 13:19:20 @Den bitbake -c devshell dhcpd Apr 29 13:23:09 well my devshell is totally buggy, gotta fix that first Apr 29 13:25:36 bluelightning: You're right. It switches to gcc instead of arm-poky-linux-gnueabi-gcc somewhere later in the process... Apr 29 13:27:10 it could be there is some hardcoded reference to gcc somewhere in the dhcpcd code Apr 29 13:27:21 well, its build scripts that is Apr 29 14:46:50 halstead: ping Apr 29 14:48:00 Good morning scottrif1 Apr 29 14:48:57 halstead: hi - I was talking to RP the other day and there is some Ontario Canyon work things (infrastructure) I would like to discuss with you if you have a moment. private? Apr 29 15:03:17 Hmm, seems not all files have correct ownership in an image i build out of daisy. e.g. hwdb, fontconfig cache, and one of the lists in var/lib/opkg/lists/ Apr 29 15:06:25 kergoth: I seem to recall someone reporting a similar issue on the mailing list, can't find it at the moment Apr 29 15:06:37 the conclusion was that it was opkg related Apr 29 15:07:30 ah found it: http://lists.openembedded.org/pipermail/openembedded-core/2014-April/091779.html Apr 29 15:10:16 ah, thanks Apr 29 15:45:02 * kergoth wonders if anyone has looked into creating ${PN}-src packages ala sourcepkg/sourceipk but using archiver to do the majority of the work, out of curiosity Apr 29 16:09:07 hi can someone help me find bitbake recipes for python-scipy? Apr 29 16:09:27 I was able to bitbake python-numpy with yocto out of the box but i can't seem to find scipy anywhere Apr 29 16:09:30 thanks a lot Apr 29 16:10:57 dtran11: http://www.openembedded.org/Layers_FAQ#I_can.27t_find_a_recipe_I.27m_looking_for._What_can_I_do.3F Apr 29 16:11:19 thanks Apr 29 16:12:54 actually if you refresh that I just updated the link to the manual section Apr 29 17:22:38 How often are changes to the openembedded core merged into yocto/poky? E.g.: https://www.mail-archive.com/openembedded-core@lists.openembedded.org/msg50176.html Apr 29 17:22:55 And how often is the git code 'frozen' to a tar.gz? Apr 29 17:23:43 it's created whe you set BB_GENERATE_MIRROR_TARBALLS = "1", though it doesn't always write out the updates to the tarball when it should in my expeirence Apr 29 17:24:35 volker-: for the first question, immediately Apr 29 17:25:22 kergoth: sorry, with "frozen" to a tar.gz I meant the tar's you can find here: http://downloads.yoctoproject.org/releases/yocto/yocto-1.5.1/ Apr 29 17:30:28 bluelightning: can't verify that. the change made it to the master branch but not to the dora branch Apr 29 17:32:03 volker-: you asked about oe-core to poky, not about the master branch to a stable branch Apr 29 17:32:53 not all changes from master will be backported automatically to stable releases Apr 29 17:33:06 see https://wiki.yoctoproject.org/wiki/Stable_branch_maintenance Apr 29 17:34:57 bluelightning: ok, thank you. Apr 29 17:36:33 bluelightning: "Point release" is the creation of the tgz? Apr 29 17:39:47 packetpub.com still lists the yocto book as "expected April 2014" Apr 29 17:52:44 volker-: point release as in 1.x.y, where y changes Apr 29 17:54:22 * kroon has a problem using wandboard-dual machine in oe-core master, sanity check fails Apr 29 17:55:13 "Error, the PACKAGE_ARCHS variable does not contain TUNE_PKGARCH (cortexa9t2hf-vfp-neon)." Apr 29 17:56:18 Somehow related to the arm .inc-file updates from today. I guess something in meta-fsl-arm needs updating.. Apr 29 18:03:24 kroon: weird it shouldn't need any update in BSPs, can you send list of your PACKAGE_ARCHS? Apr 29 18:06:09 JaMa, I first reverted "tune-cortexr4.inc: Add thumb and arm to TUNE_FEATURES" - still the same error, then I reverted "feature-arm-thumb.inc, arch-armv4.inc: Add "arm" to TUNE_FEATURES", now bitbake runs Apr 29 18:07:30 one sec Apr 29 18:07:54 This is after reverting so I can run bitbake: PACKAGE_ARCHS="all any noarch armv5hf-vfp armv5ehf-vfp armv6hf-vfp armv7ahf-vfp armv7ahf-vfp-neon cortexa9hf-vfp cortexa9hf-vfp-neon cortexa9hf-vfp-neon-mx6 wandboard_dual" Apr 29 18:08:00 and please send me link to MACHINE config you're using Apr 29 18:09:22 JaMa, https://github.com/Freescale/meta-fsl-arm-extra/blob/master/conf/machine/wandboard-dual.conf Apr 29 18:09:31 but its nasty, with a couple of includes... Apr 29 18:09:46 well, maybe not nasty Apr 29 18:10:43 And it uses includes from meta-fsl-arm layer I believe Apr 29 18:11:31 yes looks like it, what's DEFAULTTUNE in your conf? Apr 29 18:11:58 the one I've found in imx-base.inc doesn't have thumb enabled Apr 29 18:12:10 JaMa, DEFAULTTUNE="cortexa9hf-neon" Apr 29 18:12:40 can you pastebin how TUNE_PKGARCH was constructed (the output from bitbake -e)? Apr 29 18:14:01 maybe I see the issues Apr 29 18:14:16 can you try to change: Apr 29 18:14:19 -ARMPKGSFX_THUMB .= "${ARM_THUMB_SUFFIX}" Apr 29 18:15:06 +ARMPKGSFX_THUMB .= "${@bb.utils.contains('TUNE_FEATURES', 'thumb', '${ARM_THUMB_SUFFIX}', '', d)}" Apr 29 18:15:08 ? Apr 29 18:15:19 http://pastebin.com/0z2tJktd Apr 29 18:16:29 kroon: this pastebin is with the change reverted, right? Apr 29 18:16:37 yes Apr 29 18:16:59 if the change above works for you, then please send patch to oe-core ML Apr 29 18:17:03 ah sorry, I see I can run bitbake -e anyway Apr 29 18:17:25 I'm at ELC and won't be able to test it properly for few more hours Apr 29 18:18:12 Is there a way to install all python-packages? I don't see a meta package here (like I did with the kernel modules) Apr 29 18:18:42 JaMa, this with restored master branch: http://pastebin.com/RRuuZuVz Apr 29 18:19:58 kroon: yes looks like the ARMPKGSFX_THUMB suffix is enabled even without thumb in enabled features, so the change above should fix it Apr 29 18:20:15 should I send the patch so that you'll just reply with tested tag? Apr 29 18:21:45 JaMa, indeed it did let bitbake run, but I got one helluva lot of warnings of style "WARNING: Recipe 'linux-cfa' selects ARM_INSTRUCTION_SET to be 'None', but tune configuration overrides it to 'arm'" Apr 29 18:23:20 patch sent, gtg Apr 29 18:23:47 ok, thanks Apr 29 18:27:54 that was fast :-) Apr 29 18:30:34 * kergoth wonders why core-image-minimal-initramfs is used for image-live and includes installation stuff, rather than having a separate image for this Apr 29 19:20:17 I am trying to install am rpm package and getting an error no package provides .so but the package itself provides the .so Apr 29 19:21:13 I even tried to create a separate package with the .so and did a RDEPENDS on that package but still getting the same error Apr 29 19:21:20 Any ideas why ?? Apr 29 19:23:37 lobalaku: you're building an image (i.e., bitbake core-image-foo) with RPM as your package manager? Apr 29 19:23:54 yes Apr 29 19:24:23 lobalaku: maybe the SOLIBS setting? Apr 29 19:24:28 lobalaku: and you have some package foo that depends on a library file that is in package bar, and foo has RDEPENDS_${PN} += "bar"? Apr 29 19:25:13 @garibaldi yes Apr 29 19:26:12 @volker How do I change the setting ? Apr 29 19:28:05 lobalaku: this happens only for one package? Apr 29 19:28:29 lobalaku: try SOLIBS=".so" (the default is ".so.*" from my understanding) Apr 29 19:28:32 seems SOLIBS might affect a lot of packages, but I'm not sure Apr 29 19:28:48 ahh Apr 29 19:29:12 as of now this is happening for only one package Apr 29 19:29:37 I see what volker- is saying, it is by default .so.* Apr 29 19:29:42 meta/recipes-support/db/db_5.3.21.bb:SOLIBS = "-5*.so" Apr 29 19:30:05 volker : Thanks. I will give it a try Apr 29 19:30:21 might be good to set it to ".so*" to catch both .so and .so.1 Apr 29 19:30:34 (if your package has both) Apr 29 19:30:47 ok Apr 29 19:31:29 I am not sure if SOLIBS will fix it. FILES_${PN} might be the fix Apr 29 19:33:46 SOLIBS looks fine. Apr 29 19:34:05 I will try FILES_${PN} too Apr 29 19:35:07 lobalaku: "looks fine" = worked without any error so far? Apr 29 19:35:45 volker: No… SOLIBS setting is already set as "-5*.so" Apr 29 19:36:15 and your file has the style "-5*.so"? Apr 29 19:36:44 How do I verify that? Apr 29 19:37:09 sorry I am relatively new to this.. Just beginning to learn Apr 29 19:37:25 http://www.yoctoproject.org/docs/current/dev-manual/dev-manual.html#platdev-working-with-libraries Apr 29 19:38:20 volker: Thank you so much.. Apr 29 19:38:33 lobalaku: thank me by telling me what fixed it ;-) Apr 29 19:38:50 Sure… :) Apr 29 20:06:26 Hi! Apr 29 20:30:26 Right now I definie all files I want to copy to the build directory via SRC_URI. Is there a way to copy entire folders or using wildcards? Apr 29 20:32:30 both work fine Apr 29 20:32:44 you can specify a directory, or a wildcard. take your pick Apr 29 20:32:53 also, you could have just tried and seen it work ;) Apr 29 20:37:00 hi otavio Apr 29 20:41:51 hmpf, youre right, file://* works, I trhink I just should have increases PR Apr 29 20:44:34 SRC_URI = "file://*;subdir=testfiles does not like to move * to ${WORKSPACE}/testfiles/ Apr 29 23:01:08 halstead: ping Apr 29 23:01:23 scottrif, Good afternoon. Apr 29 23:01:42 halstead: private **** ENDING LOGGING AT Wed Apr 30 02:59:58 2014