**** BEGIN LOGGING AT Mon Jun 18 03:00:07 2018 Jun 18 06:09:31 Other than number of setscene tasks shown is there any other way to confirm the usage of SSTATE_MIRRORS? Jun 18 06:09:47 Can anyone please help? Jun 18 07:36:11 RP: I'll try to fix things, but operating out of a hotel room is challenging :) I guess I can use the auh machine for builds. Jun 18 07:36:32 RP: I'm now literally between jobs Jun 18 08:10:36 kanavin_home: no problem, feel free to use that machine or one of the AB workers for builds Jun 18 08:11:08 kanavin_home: I know the feeling on jobs, the hotel room certainly won't help though! Jun 18 08:12:24 kanavin_home: in case of emergency, i can donate an AWS instance for a couple of hours while you're between jobs Jun 18 08:30:27 New news from stackoverflow: How to configure which BOOST library files are included in a Yocto SDK Jun 18 09:03:38 Dear all, Jun 18 09:03:56 Is it possible to replace a single function defined in a bbclass file? Jun 18 09:04:18 For example I do need to add single line in a single function in a kernel-fitimage.bbclass Jun 18 09:04:41 (it is OK for me to copy 10 lines of the overiden function) Jun 18 09:05:04 I've created ./classes/kernel-fitimage.bbclassappend Jun 18 09:05:09 with the single function Jun 18 09:06:44 lukma: bbclassappend is not a thing... so that particular way won't work (I'm not good enough at yocto to tell you how to do it, sorry...) Jun 18 09:07:36 The problems is: Jun 18 09:07:53 - I need to add single line Jun 18 09:08:04 is there a bitbake switch which only show me the verbose gcc compile outputs ? Jun 18 09:08:19 and I thought if there is a mechanism similar to bbappend for recipes Jun 18 09:08:26 to extend the bbclass Jun 18 09:08:47 without the need to create my own class -> like kernel-fitimage-fixed.bbclass Jun 18 09:08:59 which would inherite the original one and the function to be replaced Jun 18 09:30:21 hi everyone. I try to add a file to the rootfs describing the image that is currently built. So I added SRC_URI and FILES to the image recipe and also a do_install() but that is never called. I fear that this is due to the "inherit core-image" and that my way isn't possible at all. Has anyone an idea? Jun 18 09:30:41 New news from stackoverflow: Adding python internal modules (cgi.py) to yocto image || Bitbake recipes - Simple file copy Jun 18 09:34:59 rob_w: log.do_compile will have the compiler output Jun 18 09:38:09 yes i know .. that was my question to get rid of always looking at that file but getting those outpouts in my cmdline Jun 18 09:38:28 but if thats the way , no problem Jun 18 09:38:49 rob_w: wind river linux has a patch to do that iirc Jun 18 09:38:57 its was just to ease up hunting down my compiler warnings Jun 18 09:39:36 make them fatal :) Jun 18 09:40:44 hmm that is definitly worth a thought ! Jun 18 09:41:15 my collegues will love that ;) but then i save my time keep telling them to hunt the warnings Jun 18 09:56:49 Hi, i am developping an application using eclipse ADT yocto. I create a yocto ADT autotools project and I am using some libraries like libxml2 but when i include some header file from this library, the yocto autotools build doesn't see them even if in my makefile.am i added the library include in the CXXFLAGS and the library in LDFLAGS. Does anyone have faced this kind of issue to help me fix this? Jun 18 10:32:37 What is the process for switching maintainer on a repository at git.yoctoproject.org? Jun 18 10:36:18 gunnarx, which one? Jun 18 10:40:00 meta-ivi. is the process different for different repos? ;-) Jun 18 10:41:16 there's a new maintainer... just wondering how to set up ssh key or whatever credentials (better than taking over the previous maintainer's credentials) Jun 18 10:49:02 mostly involves talking with existing maintainees Jun 18 10:49:08 I'm not familiar with that one Jun 18 10:49:18 halstead, should be able to help with keys Jun 18 10:49:39 who is existing maintainer? Jun 18 10:50:39 Well it's not abandoned, it's an agreed transfer. The existing maintainer Yong-Il Yoh can confirm, we can have an email exchange for example. Jun 18 10:51:25 However, granted he has not updated this for a while, only the GitHub copy. That will change if we transfer it. Jun 18 10:52:50 halstead: can we sort out the remaining details via email? Jun 18 10:53:28 Because I need to loop in someone else also (the new maintainer, it's not me) Jun 18 10:54:22 halstead, is US west coast. He'll likely want emails with particulars. Clearly we can't transfer maintainship to random people in irc :) Jun 18 10:54:35 but it sounds like everything should go smoothly Jun 18 10:54:39 "random" :) Jun 18 10:54:45 :) Jun 18 10:54:51 yes, as long as I get a contact, we'll make it work Jun 18 10:56:07 I messaged his email to you Jun 18 10:56:32 excellent, thanks for your heop Crofton|work Jun 18 10:56:35 *help Jun 18 11:05:24 hi Jun 18 11:39:28 hmm, nobody has a clue, maybe I should try another approach: has someone an idea how I can get information about a built image into the rootfs it builds? Jun 18 11:40:00 ROOTFS_POSTPROCESS_COMMAND, probably. Jun 18 11:40:11 using a dedicated recipe for that purpose and include it just in that image would surely work, but the natural way imho would be to integrate that very information into the image recipe itself and getting it into the rootfs somehow Jun 18 11:40:44 oh, great, thanks for that hint, LetoThe2nd, I have a look at that right away Jun 18 11:54:39 LetoThe2nd: perfect, that did the trick, thanks a bunch! Jun 18 11:56:08 RP: thanks, also my top priority now is finding a place to live! in Berlin that is not at all easy. once that is sorted I can think again about bug fixes and patches :) Jun 18 11:56:11 LetoThe2nd: ^^^ Jun 18 11:57:11 kanavin_home: good luck, then! Jun 18 11:57:54 kanavin_home: like i said, i'll be loosely around for the rest of the week, and in cases of emergency can provide AWS computing power. after that, i'm on leave for considerable time. Jun 18 12:01:44 kanavin_home: that sounds like the right priority! :) Jun 18 12:02:13 absolutely agreed. Jun 18 12:04:06 I'll keep an eye on the mailing lists, telling people what's wrong with their patches is often not complicated at all ;) Jun 18 13:02:34 kanavin_home, congrats on the ove to Berlin Jun 18 13:02:40 I'd love to move there :) Jun 18 13:02:52 espceially before it became popular Jun 18 13:03:08 Crofton|work: yep, rents are skyrocketing Jun 18 13:03:27 Crofton|work: for various reasons the chance to move came up only now Jun 18 13:08:39 All the cool people are there! Jun 18 13:11:15 Crofton|work: locals are protesting against google opening an office in Berlin Jun 18 13:11:31 Ah Jun 18 13:11:35 Crofton|work: they don't mind google specifically, but more of what it represents in places like SF or London Jun 18 13:11:40 If in doubt, blame Google Jun 18 13:12:45 specifically that those places became playgrounds for the rich, no one can afford to own where they live, and unless you have a high salary, you cannot afford to rent either Jun 18 13:13:08 a global problem Jun 18 13:13:17 and yet, someone still has to clean the streets and teach the kids Jun 18 13:13:54 Crofton|work: nordic countries have managed to contain it, but it doesn't come cheap. I donated a full 1/3 of my salary to taxation Jun 18 13:14:00 in Finland Jun 18 14:57:51 * kergoth yawns Jun 18 15:02:35 kanavin_home: income tax is higher in germany than in finland.. be prepared Jun 18 15:06:49 * fray didn't know though.. I thought Finland was one of the highest Jun 18 15:06:55 'er.. know -that- Jun 18 15:11:20 mcfrisk: I used the 'unofficial tax calculator' to see where we'd end up, I think the ends will meet quite ok even with one of us not working for some time Jun 18 15:12:08 how to prevent map_depends_variable from adding multilib prefix? I have busybox bbappend which adds RDEPENDS_${PN} += "/bin/bash" but I don't want /bin/bash to be prefixed as lib32-/bin/bash and if I try to use RDEPENDS_${LIB32_PREFIX}${BPN} then the original RDEPENDS_${PN} replaces it; and if I try to append with virtclass-multilib-lib32 then it's ignored even when it's in OVERRIDES Jun 18 15:14:40 JaMa: I don't think you are even supposed to put binary paths in there, why can't you use a package name? Jun 18 15:15:51 kanavin_home: this is intentionally /bin/bash to work around https://bugzilla.yoctoproject.org/show_bug.cgi?id=9217 Jun 18 15:15:52 Bug 9217: normal, Medium, 2.99, Martin.Jansa, ACCEPTED , Many unsolveable QA warnings from build-deps and file-rdeps Jun 18 15:22:15 JaMa: if this can be reproduced with plain oe-core, please tell how in that bug Jun 18 15:22:42 JaMa: or a minimal way to do it with meta-oe Jun 18 15:23:56 it's written in the ticket already Jun 18 15:26:54 JaMa: where? I am not seeing it. Jun 18 15:28:46 around 30th line in 2nd to last comment Jun 18 15:29:12 'The test case is busybox bbappend' thing? that is not plain oe-core :-) Jun 18 15:29:12 RP: kicked a new mut, should fix the drm problems and will also error out if the i586 problem happens on other arches, so i'm pretty much expecting some failures there Jun 18 15:29:51 geez, not so positive for a Monday Jun 18 15:30:22 kanavin_home: it's minimal way how to reproduce it there Jun 18 15:31:39 RP, stable/sumo-next is ready to go. clean build over weekend Jun 18 15:31:54 JaMa: but it's simply not right to put /bin/bash to RDEPENDS/PROVIDES in the first place, why do you need to do that? Jun 18 15:32:36 JaMa: if that is to silence the QA warnings, then you should dig deeper into why the warnings happen Jun 18 15:35:33 kanavin_home: yes, it's to silence thousands of QA warnings Jun 18 15:36:09 JaMa: yep, so how can the warnings be triggered? that's the reproducer I meant Jun 18 15:38:03 JaMa, RP, kergoth: do any of you remember why we have a mktemp recipe when both busybox and coreutils have it? Jun 18 15:38:25 i'm guessing at some point busybox mktemp was too minimal for something? Jun 18 15:40:14 the only difference between busybox mktemp and coreutils mktemp now is --suffix Jun 18 15:43:01 kanavin_home: http://paste.ubuntu.com/p/ZYx2V8xk7Z/ Jun 18 15:43:35 kanavin_home: build e.g. libtool in TMPDIR with and without bash built before Jun 18 15:45:21 I am trying to build a systemd service but my recipe can't find a systemd include (https://hastebin.com/iwutaxeyif.hs) I am an occasional yocto user so things might not be quite right overall. Is there anything blatantly wrong in my recipe? any suggestions/thoughts much appreciated Jun 18 15:45:22 kanavin_home: in 1st case, there won't be any warning, but libtool will rdepend on proper bash (even when user wanted to use busybox as a provider) in 2nd case you'll get the QA warning http://errors.yoctoproject.org/Errors/Build/61824/ Jun 18 15:48:03 kanavin_home: the work around I've sent to oe-core before fixes both these issues, but fails when we enable multilib as well (because of the lib32- prefix being applied) Jun 18 15:49:44 kanavin_home: to reproduce the 2nd part (about e.g. libmali.so) might be more tricky with oe-core, because oe-core neither meta-oe provide alternative prebuilt virtual/*gl providers Jun 18 15:50:38 kanavin_home: possibly reproducible with old vc-graphics recipe from meta-raspberrypi, but nobody is using that Jun 18 15:50:52 JaMa: if you can put all of this information (about busybox/libtool issue) into a single comment in the bug, that'd be appreciated. The configuration snippet, and specific steps. Jun 18 15:51:59 fwiw: anonymous python is also triggered too early to replace lib32-/bin/bash with /bin/bash Jun 18 16:15:02 Anyone tried to install the YP plug in for Eclipse Oxygen lately (sumo). None of this stuff works for me Jun 18 16:31:43 khem: huh, https://github.com/yrnkrn/zapcc seems interesting Jun 18 16:33:19 hi, I'm trying to excute a python script to generate a custom image format in a image_type.bbclass. The python script is properly installed in recipe-sysroot-native/usr/bin/ but fails to load some python modules such as "from Crypto.Cipher import AES ImportError: No module named 'Crypto'". I tried to add EXTRANATIVEPATH += "python3-native" without success. ./usr/lib/python3.5/crypt.py is in Jun 18 16:33:25 recipe-sysroot-native of the image. How can I get a python3-native tool to find the modules ? Jun 18 16:54:48 rburton: no memory of that Jun 18 16:54:50 rburton: i have no clue Jun 18 16:59:33 awww my build :( Jun 18 21:01:51 I built the fsl-image-gui target for an NXP imx7dsabred board I received and can boot, connect etc. In comparing against the sample image that came on the SD card with the board though, there is substantially less functionality. I'm having trouble finding which layers/recipes I might add in to get some of it back before I start making my own image. Suggestions where I might find a list of software that comes on the sample SD card images? Jun 18 21:24:33 dennism: you should talk to nxp support Jun 18 21:24:45 or may be look into fsl layers Jun 18 21:25:10 I'm searching through the various fsl-image* available now. Jun 18 21:25:24 the layers are here https://github.com/freescale Jun 18 21:25:46 but the SDKs might be different than what is published who knows Jun 18 21:25:57 perhaps one of their multimedia layers is what they used for the sample sd. Jun 19 00:33:04 dennism: have you checked the release notes for the fslc releases? http://freescale.github.io/doc/release-notes/2.4/ **** ENDING LOGGING AT Tue Jun 19 03:00:03 2018