**** BEGIN LOGGING AT Thu May 07 02:59:58 2015 May 07 05:50:08 hi all May 07 05:50:34 i have some error when im trying build yocto with my package May 07 05:51:36 https://www.dropbox.com/s/52npwut3lx13jn5/Screenshot%202015-05-07%2011.51.06.png?dl=0 May 07 05:52:50 shakalaka: well the message says it all, doesn't it? May 07 05:53:15 i don't know how can i fix it =) May 07 05:53:19 shakalaka: and why that roundtrip screenshot->crop->dropbox when you can just dump the log into a pastebin? May 07 05:53:55 shakalaka: i would point you at the relevant snippets, but as i can't c&p them out... May 07 05:53:56 i have autotool for taking screenshots in dropbox May 07 05:54:14 ok, 1 sec May 07 05:55:06 http://pastebin.com/h5TpF861 May 07 05:55:31 thats the point: ERROR: Function failed: cool: LIC_FILES_CHKSUM points to an invalid file: /home/valentin/Desktop/Projects/Yocto/poky/build/tmp/work/i586-poky-linux/cool/1.0-r0/cool-1.0/COPYING May 07 05:56:07 shakalaka: you probably set your LIC_FILES_CHECKSUM to a file that doesn't exist in your sources. May 07 05:56:18 recipe http://pastebin.com/kENh8dNm, layer.conf http://pastebin.com/EAyjAh5t, bblayers.conf http://pastebin.com/zWDAE3hE May 07 05:56:34 nope, it exists May 07 05:57:26 oooh May 07 05:57:30 its alias May 07 05:57:51 shakalaka: :) May 07 05:58:41 shakalaka: plus, if your recipe is called cool-1.0, the tarball should be called cool-1.0.tar.gz or something similar, and should extract to a directory called cool-1.0 May 07 05:58:58 makes it easier for everyone to handle May 07 05:59:15 ok, thanks for this info May 07 06:20:41 hello warthog9, pleasant evening to you May 07 06:23:02 * warthog9 waves at sri May 07 07:00:31 I'm confused about how BBMASK works.. I have a distro which sets BBMASK = "mask/dir/a/|mask/dir/b/" and then I want to mask another recipe so I use BBMASK_append = |mask/certain/recipe*".. Looking at bitbake -e myimage | grep ^BBMASK the resulting BBMASK looks ok but my additional recipe isn't masked.. If I instead add the exact same thing I append in local.conf to the original BBMASK assignment in the distros conf it works and the recipe is masked.. May 07 07:01:26 Good morning folks. Another day with yocto development. Had a larger pause now... May 07 07:01:46 the output of bitbake -e myimage | grep ^BBMASK is identical in both cases, but only one result in the recipe being masked. May 07 07:01:58 morning May 07 07:07:20 I have an QA warning: "ERROR: QA Issue: ivi-logging: Files/directories were installed but not shipped". How can i set an ignore for the file it complains. It should really not be shipped. May 07 07:10:54 irontia: either you can do_install_append() { rm ${D}/files/not/to/ship } or set INSANE_SKIP_${PN} = "installed-vs-shipped" May 07 07:11:21 the last option disables that particular QA check May 07 07:22:28 irontia: if it's the doc and cmake stuff you can put then in the -dev package using FILES_${PN}-dev += " ${libdir}/cmake/ ${docdir}/ " May 07 07:23:42 good morning May 07 07:36:39 any folks running yocto-autobuilders in here? I'm wondering what people do to prevent the deploy dir getting out of hand in terms of space used May 07 07:36:51 am I missing some "delete old files" switch? May 07 07:38:10 joshuagl: we just do a periodical rm -fR and start over from scratch May 07 07:38:23 LetoThe2nd: hah, I feared that would be the response May 07 07:38:36 you could always just throw more drives at it :| May 07 07:39:12 joshuagl: actually it has the good sideeffect that you know for sure that you actually *can* rebuild from scratch May 07 07:39:37 joshuagl: is it the images or packages that is the problem? There's an option for removing previous image when building a new one: RM_OLD_IMAGE = "1" May 07 07:39:54 and there's RM_WORK too May 07 07:40:02 erbo: oooh, that sounds like just what I'm after May 07 07:40:10 it's the images and we're already using rm_work May 07 07:42:17 erbo: thanks! May 07 08:02:20 thank you erbo May 07 08:03:20 morning all May 07 08:04:19 bluelightning, RP: do you know if it's intentional that e.g. BBMASK_append in local.conf is added after BBMASK is used? I have a distro which sets BBMASK = "somthing" and I cannot find a way to add to that from local.conf May 07 08:04:44 erbo: += should work May 07 08:05:04 hmm May 07 08:05:10 actually it won't in that case May 07 08:05:19 I think you may be stuck... May 07 08:06:34 erbo, have you tried just BBMASK alone, the doc says its only parsed from local.conf May 07 08:06:50 no distro.conf is overriding local.conf, but I was hoping that _append/_prepend would work but that only happens after it is used.. bitbake -e something | grep ^BBMASK show the result I want, but during parsing it's the wrong value May 07 08:07:38 redengin: I guess the doc is unclean then, distro.conf overrides the value in local.conf :/ May 07 08:07:53 *unclear* May 07 08:08:26 http://www.yoctoproject.org/docs/current/ref-manual/ref-manual.html have you tried .= ? May 07 08:09:20 looks like it has a filter syntax as opposed to a list syntax May 07 08:09:21 redengin: yes, .= is evaluated immediately and then it's just overridden by distro.conf May 07 08:10:07 ah, I see, yeah, then you're hosed by your distro, how about you just change the distro to use .= as well? May 07 08:10:17 it would work if the distro.conf used .=, but that is out of my control currently.. I can always ask if then can change it for future releases though May 07 08:10:30 they May 07 08:11:12 btw, why are you trying to avoid building stuff that the distro wants? May 07 08:12:17 it doesn't sound like the outcome would be good May 07 08:12:21 I want to use bluez5 and bluez5 only, but it seems a bit tricky to get that working reliable so I want to mask out bluez4 to avoid the issue completely May 07 08:13:26 so you've set bluez5 and still see it use other versions? May 07 08:13:47 erbo: rather than BBMASK, you could use PNBLACKLIST for that May 07 08:14:19 honestly BBMASK should almost never be used May 07 08:15:54 bluelightning: oh, that looks promising. thanks! May 07 08:20:17 redengin: yeah some packages seems to pull in bluez4 instead of bluez5.. May 07 08:21:21 bluelightning: the PNBLACKLIST way seems to work, bitbake bluez4 failed but bitbake seems to resolv all dependencies anyway so everything should use bluez5 now. May 07 08:21:22 sounds like they may have hardcoded dependencies then, have you tried just updating the distro recipe locally to see what happens? May 07 08:22:11 erbo: at long last in fido+ we have a proper mechanism for selecting bluez5 over bluez4, unfortunately that was a long time coming :/ May 07 08:24:53 bluelightning: that's great news! I'm stuck in daisy in this project for now, but atleast I have a nice workaround now. May 07 09:20:02 erbo: _append probably doesn't work since overrides probably aren't expanded before BBMASK is used by bitbake. If you file a bug for that, we could have a look into it May 07 09:47:50 LetoThe2nd: we enabled the debugging in cupsd and there is only one filter, so is has to do something with the driver of our printer? May 07 09:48:18 LetoThe2nd: the package for printer drivers => gutenprint is not buildable (https://wiki.linaro.org/Platform/DevPlatform/CrossCompile/CurrentPackageCrossBuildStatus) May 07 09:49:46 limotec-dev: hehe, i'm no printer specialist by any means, was just pointing to something i found on the interwebs. but yes, lack of some driver thingy sounds like what i found yesterday when googling for your problem. May 07 09:54:26 LetoThe2nd: so are we, we thought with cups package it would work, but there has to be someone who wanted to print in yocto on a ARM board May 07 09:58:12 limotec-dev: probably. but sorry, no ideas, here. May 07 09:58:38 LetoThe2nd: thanks for trying tho May 07 15:01:00 Hi, I'm trying to produce the meta-qt5 toolchain but I'm getting this error : http://pastebin.com/HVfX8hnn Is somebody can help me ? May 07 15:20:06 kergoth: there's a bitbake patch i just sent to the list that you might want to have an opinion on... May 07 15:22:41 * kergoth would love it if we versioned properly, provided stability guarantees, and released to pypi, but given that's not likely to happen anytime soon, he's fine with the removal May 07 15:24:48 kergoth: so RP is aware, can you put that in writing. i had the task to send the patch and check with you :) May 07 15:25:52 rburton: I did see the feedback here May 07 15:26:14 kergoth: I agree FWIW but I also can't see it happening for various reasons :/ May 07 15:48:35 * kergoth nods May 07 15:49:09 it's tied so tightly to oe-core, it's painful enough making changes that need to be in sync between them as is May 07 15:52:59 How do I rotate the screen in Poky distro? Do I have to write my own xorg.conf. There doesn't seem to be one in rootfs.img May 07 16:23:11 Did anyone ever figure out a script to find unused patches/etc in a layer? May 07 16:32:38 kergoth: I'm sure I could knock something up... May 07 16:33:33 yeah, doesn't seem like it'd be too terrible, i know it's been asked about at one point, wasn't sure if anyone did it yet though May 07 16:33:36 might be useful for cleanup May 07 16:33:38 I thought I already had, but then I remembered I was verifying a patchset to drop unused patches rather than actually figuring out which ones were unused, and I just did that with a few greps May 07 16:33:51 * kergoth gets more caffeine May 07 16:38:18 * moto-timo orders a double espresso May 07 19:08:18 I have a question about supplying a full defconfig in recipes-kernel linux. May 07 19:08:20 Is it necessary that this file be named defconfig in order for the build to recognize it as the default .config? May 07 19:09:57 Is there any variable that I can use in the bblayers.conf file that points to the poky directory? I'd like to make things relative so I can reuse the same file in other machines. Something that could allow me to specify layers like this: "${POKY_ROOT}/meta-name" May 07 19:11:08 rschaef: yes, if you are using the kernel.bbclass configuration step. May 07 19:12:21 DS_: no, but you can use ##OEROOT## in the bblayers.conf.sample, which the setup scripts will replace May 07 19:14:09 and that comes into play by "inherit kernel". So unless I would like to remake the compile logic for kernels, it would just be easiest to name the .config defconfig? May 07 19:14:47 yep. May 07 19:15:11 you can extend it via config fragments later, when/if you feel the need. May 07 19:16:14 technically you don't have to remake the cmpile logic to roll your own write of the .config, but you really should stick to the existing conventions to safe headaches and be consistent May 07 19:16:21 s/safe/save/ May 07 19:16:26 * zeddii nods May 07 19:22:28 if I have multiple boards with similar full configs already defined, would the easiest/best way be to have an empty defconfig file, and then KERNEL_FEATURES_append_${MACHINE_NAME} = ${MACHINE_NAME}.cfg ?? May 07 19:22:44 kergoth thanks! May 07 19:23:47 hey May 07 19:23:54 my kernel problem is finally solved \o/ May 07 19:24:01 rschaef: have a base configuration fragment, call it defconfig (or not if you use the fragments, you could call it base.cfg), and then just add fragments to the SRC_URI that are machine specific. May 07 19:24:47 okay, thanks! May 07 19:49:01 and is there a certain spot where the final .config file is placed before compiling the kernel? just to make sure I did this correctly. May 07 19:50:01 rschaef, it goes into ${B}, the kernel build directory. May 07 20:30:30 okay, sanity check: if I have a recipe at recipes-kernel/linux-mykernel_3.14.bb, I should place my defconfig at recipes-kernel/linux-mykernel/defconfig, and then from within linux-mykernel_3.14.bb I should put SRC_URI += "file://defconfig" May 07 21:13:34 Does anyone know what "QA Issue: ELF binary 'path' has relocations in .text [textrel]" mean? May 07 21:19:47 realBigFoot, check these docs out: http://www.yoctoproject.org/docs/current/ref-manual/ref-manual.html#qa-issue-textrel May 07 21:20:46 sayguh, nice thanks! May 07 21:26:18 I created meta/recipes-bsp/formfactor/files/genericx86-64/machconfig with this the content of this: http://pastebin.com/dNUbkKMM in hopes this would rotate the touchscreen panel and enable touch. Unfortunately it doesn't seem to take. Any thoughts why? May 07 21:33:40 realBigFoot: you can have a glimpse using i.e. scanelf -qT May 07 21:35:06 ant_home, you mean bitbake scanelf command ? May 07 21:36:01 I mean to build it and use it on target device May 07 21:37:12 ant_home, ohh i got it... i will do it **** ENDING LOGGING AT Fri May 08 02:59:59 2015