**** BEGIN LOGGING AT Fri May 04 03:00:05 2018 May 04 05:31:56 New news from stackoverflow: RPC package for my Yocto Linux build May 04 06:22:27 Hello, does anyone happen to know Juro's handle in here? May 04 07:15:21 aehs29: Hi, what is "Juro's handle"? May 04 07:18:37 nayfe: probably the nickname of some person with a real life first name of "Juro" May 04 07:18:48 * LetoThe2nd points at the /who feature of irc May 04 07:20:19 LetoThe2nd: oww ok I was far from that May 04 07:20:23 thx May 04 08:01:33 LetoThe2nd: thanks,haha now I just have to loop through all the names May 04 08:09:43 otherwise, ask him by mail? :D May 04 08:18:35 Hi folks! What *-initial recipes certainly do? For example, we have glibc recipe and glibc-initial. Why? May 04 08:19:27 i have set up SSTATE_MIRRORS but no setscene tasks run at all. how can i investigate why? May 04 08:21:07 acrap: AFAICS this is related to the bootstrapping stage, primarily of the toolchain May 04 08:23:05 acrap: yes, initial recipes provide basic functionality for the toolchain to be built May 04 08:24:05 aehs29: thanks a lot! May 04 08:24:15 LetoThe2nd: thanks! May 04 08:25:17 acrap: you can get more info if you look at the manual, on the cross development toolchain section May 04 08:30:18 aehs29: i can't find it. Could you get me the link, please? May 04 08:30:56 aehs29: i mean the link to "cross dev toolchain section" May 04 09:05:32 Hi Just a quick beginner question. I am trying to fix a problem on my build. I need to make a symlink like similar to "ln -s /sys/bus /dev/bus " in a recipe. So I added ln -sf /sys/bus ${D}/dev/bus in my do_install_append section and FILES_${PN}-dev += "/dev/bus". There is no warning when i run bitbake but the symlink is missing in the final image May 04 09:07:06 jesand: you cannot make symlinks there, as /dev, /sys, /proc and /tmp usually are only virtual filesystems May 04 09:08:58 LetoThe2end: ok. I see.. Do you have any suggestion on how to make the link then since when it is missing libusb will not work? May 04 09:10:09 jesand: find the *real* source of the problem. what does libusb expect, why and where? is your image even providing it? if it is something in /dev, maybe it is a missing kernel module or udev rule May 04 09:12:13 LetoThe2nd: I will try do that, Thank you for pointing me in the right direction :-) May 04 09:12:26 jesand: good luck May 04 10:32:53 New news from stackoverflow: To build custom BSP layer using Yocto for a bare metal board May 04 10:59:23 hello, has anyone used the meta-qt5 master branch recently? what is the best way to test it? is it compatible with the current poky sumo branch? May 04 11:02:59 New news from stackoverflow: Can't connect to free wifi network with iwconfig in Linux embedded May 04 11:07:45 eduardas_m: right now, sumo and master are identical May 04 11:09:18 rburton_: btw, out of pure interest - wasn't sumo slated as release like 2 weeks ago? May 04 11:26:52 if the same variable is defined in both local.conf and site.conf, which wins? May 04 11:28:30 osse: bitbake -e tells you :) May 04 11:29:37 LetoThe2nd: bear with us while we sort the fallout from https://lists.yoctoproject.org/pipermail/yocto-ab/2018-April/002178.html May 04 11:30:42 kanavin: ah, didn't notice :-( May 04 11:30:54 kanavin: and i'll bear with anything, i was just wondering. May 04 11:31:06 LetoThe2nd: it gets worse https://lists.yoctoproject.org/pipermail/yocto-ab/2018-April/002212.html May 04 11:31:13 LetoThe2nd: if you can help in any way, do May 04 11:31:59 LetoThe2nd: good idea. seems it's local May 04 11:33:52 this is unfortunate May 04 11:34:25 and really messed up May 04 11:34:58 one of the advantages of corporate communities is supposed to be shit like this not happening May 04 11:37:36 kanavin: jsut catched up on the thread... whew May 04 11:38:50 LetoThe2nd: there's also the may message https://lists.yoctoproject.org/pipermail/yocto-ab/2018-May/002216.html May 04 11:39:28 kanavin: yeah, saw it May 04 11:40:47 with Wind River moved out of the Intel umbrella, maybe they could hire Richard? May 04 11:41:42 Son_Goku: I think for everyone's benefit it's best that RP operates independently May 04 11:41:56 sure, but it seems LF doesn't care enough May 04 11:42:47 Son_Goku: WR is still in flux as the sale hasn't gone through, so nobody really knows exactly what is happening there yet May 04 11:42:56 * Son_Goku waves at rburton_ May 04 11:43:15 well, there's only one more month left in 2Q 2018 May 04 11:43:21 Son_Goku: we need to wait and see, I believe there is an important YP advisory board meeting coming up May 04 11:43:24 It's most likely closing in the next few weeks May 04 11:48:34 LetoThe2nd: anyway the main reason for 2.5 being a little late was that m3 was really late (read the status reports to see why). we caught up a bit with the last milestone being short, the first RC is in qa now May 04 11:50:11 LetoThe2nd: the RP situation actually has nothing to do with sumo release, but there will be changes during the 2.6 cycle (intel is having a bit of a shuffle, WR sale, etc) May 04 11:50:27 I'm trying to debug (unsing `-e`) why bitbake evaluates a variable (`COMPATIBLE_MACHINE`) in the way it does. However, I cannot do so, because it abborts because the provided `MACHINE` is not part of `COMPATIBLE_MACHINE` -.-' May 04 11:51:58 rburton_: i see, thx May 04 11:57:09 anyone here uses mercurial with bitbake May 04 11:57:33 fetcher has problems getting specified revisions May 04 12:03:47 kanavin: anuj's gst upgrade fails in g-i if you fancy having a lot whilst he is asleep May 04 12:04:27 rburton_: I was *just* about to go ride a bike :) I'll look at it, is there a branch where I can pull from? May 04 12:04:36 kanavin: go ride a bike :) May 04 12:04:46 I mean I will look at it later ;) May 04 12:04:55 ross/mut is justed pushed and breaks if you build gstreamer-plugins-base May 04 12:06:04 rburton_: at some point someone has to do the heavy lifting of meson switchover :-/ May 04 12:07:53 rburton_: it can be done piecemeal though May 04 12:09:00 rburton_: ERROR: Task (virtual:native:/home/ak/development/poky/meta/recipes-devtools/flex/flex_2.6.4.bb:do_compile) failed with exit code '1' May 04 12:09:35 rburton_: | Makefile:1687: recipe for target 'stage1scan.c' failed May 04 12:09:35 | make[2]: *** [stage1scan.c] Segmentation fault (core dumped) May 04 12:09:51 nuts May 04 12:10:10 rburton_: to be exact, May 04 12:10:10 | ./stage1flex -o stage1scan.c ../../flex-2.6.4/src/scan.l May 04 12:10:10 | Makefile:1687: recipe for target 'stage1scan.c' failed May 04 12:10:10 | make[2]: *** [stage1scan.c] Segmentation fault (core dumped) May 04 12:10:42 reverting for now May 04 12:10:53 I will stop and go riding when I reach the gst failure ;) May 04 12:10:58 so it's all nice and ready May 04 12:11:10 kanavin: yeah revert the flex upgrade May 04 12:11:27 not sure how that worked for me May 04 12:30:04 rburton_: -next merged to master May 04 12:33:16 New news from stackoverflow: Arm-poky-linux-gnueabi-gcc cross compile issue May 04 12:48:50 RP: \o/ May 04 12:50:23 hello all May 04 12:51:53 Our company yocto is not happy with my debian unstable distro. And it exit for errors. It requires an ubuntu of a certain version. Is there a way i can install an ubuntu here without going to virtualbox ? May 04 12:52:07 if my package needs autotools to build, i don't have to specify anything in the .bb, am i right? May 04 12:52:14 other than inherit autotools May 04 12:52:16 varjag: 'inherit autotools' May 04 12:52:25 angelo_ts: the usual way is to use a container. May 04 12:52:35 rburton_: *nod* but the rest should happen automagically? May 04 12:52:38 neverpanic, sounds acceptable May 04 12:52:46 neverpanic, if any guide, welcome May 04 12:53:03 angelo_ts: if it requires a specific ubuntu then your company yocto is a bit broken. should work on debian but unstable means new which means you get to discover the bugs in your company yocto first. May 04 12:53:18 varjag: right, that does configure/make/make install for you May 04 12:53:25 because bitbaking the project fails, it tries to go with make right away May 04 12:53:39 varjag: sounds like you need to read the configure log to see why May 04 12:53:44 angelo_ts: https://www.yoctoproject.org/docs/latest/mega-manual/mega-manual.html#setting-up-to-use-crops May 04 12:53:53 rburton_: there's no configure log anywhere May 04 12:54:39 "mbdaemon/1.0-r0/build/config.log: No such file or directory" May 04 12:54:39 May 04 12:54:49 in my log.do_configure May 04 12:54:59 are my paths botched up somehow, or May 04 12:55:51 varjag: the log.do_configure in the temp/ directory May 04 12:56:05 are you *sure* this is using autoconf? May 04 12:56:19 yep that one, see the quote above from it May 04 12:56:26 rburton_: quite sure, i wrote that package myself May 04 12:56:29 sounds like the configure is bust May 04 12:56:37 rburton_, building a recipe i get May 04 12:56:42 try using autotools-brokensep and see if that fixes it May 04 12:56:45 i can build it via ./configure and make manually no issues May 04 12:56:46 WARNING: Host distribution "debian" has not been validated with this version of the build system; you may possibly experience unexpected failures. It is recommended that you use a tested distribution. May 04 12:56:50 if it does then your configure is broken May 04 12:56:53 hmm May 04 12:57:09 i probably better go with the container ? May 04 12:57:17 angelo_ts: sure, you're not using a supported distro, so if it breaks then you get to fix it. its just a warning, i turn that warning off :) May 04 12:57:22 at the end shouldn't lose in performance May 04 12:58:21 no, it makes no difference May 04 12:59:07 varjag: well your configure must be doing something odd as it clearly works for the 463 recipes in oe-core that inherit autotools :) May 04 12:59:25 it does not run configure May 04 12:59:29 for whatever reason May 04 12:59:37 so that's not the issue May 04 13:00:16 the class will show you where it looks for the configure, if its not available in ${S} May 04 13:00:41 i don't have your recipe or sources so your on your own a bit. proper autotools just works with the single inherit, yes. May 04 13:01:54 do i have to specify ${S} manually, or it should work with some default May 04 13:02:09 ../work//projectname May 04 13:03:02 it defaults to WORKDIR/PN-PV, so if your tarball doesn't unpack to PN-PV then yes you need to set it May 04 13:05:05 hmm.. that could be it May 04 13:05:06 thanks May 04 13:09:12 kanavin: anuj found the problem May 04 13:10:05 rburton_: should I re-send this patch? May 04 13:10:33 anujm: please do May 04 13:10:43 anujm: didn't realise you where here. why are you still here? :) May 04 13:13:19 rburton_: ok, will do in a minute. was watching a movie, sword of the stranger, nice anime :) May 04 13:20:25 rburton_: it *was* the wrong S May 04 13:20:47 it all snowballed from me using hg instead of git like every other damn project on the net May 04 13:21:28 so fetch unpacks it under workdir/projectname May 04 13:33:30 New news from stackoverflow: how to get an older version of gcc in yocto beside the default one May 04 14:03:34 New news from stackoverflow: Dockerfile to run a package revision (PR) server May 04 14:05:27 if I want to set something like image-prelink permanently, can I just put a USER_CLASSES in my distro.conf or something like that? May 04 14:16:05 Hi, does yocto master branch now support multiple kernels per machine? Any pointers to documentation? May 04 14:39:35 oh, USER_CLASSES is just goes into plain old INHERIT May 04 15:03:46 New news from stackoverflow: cannot execute binary file: Exec format error for a custom Yocto layer May 04 15:05:45 rburton: i'm sorry to hear you've been disabled May 04 15:05:54 https://yoctoproject.biterg.io May 04 16:10:37 tlwoerner: that data looks wrong :/ I don't get a mention! :) May 04 16:14:38 halstead, when is the maint window on the .io AB ? May 04 16:15:12 armpit, Usually start around now. May 04 16:15:33 armpit, I'll ping you once it's complete! :) May 04 16:15:38 k. thanks May 04 16:34:03 New news from stackoverflow: How to set python version to 2.7 for Yocto project work in Ubuntu 18.04? May 04 16:37:19 armpit, Build away! May 04 16:37:32 halstead, cool. thanks May 04 17:53:12 armpit: I got two update branches for stable queue May 04 17:53:19 armpit: https://github.com/OSSystems/oe-core/compare/rocko...rocko-next May 04 17:53:35 armpit: https://github.com/OSSystems/oe-core/compare/sumo...sumo-next May 04 18:06:49 thanks May 04 18:10:49 armpit: is there I way I can make it easier? or using the branches are good ? May 04 18:11:12 branches are easy for me. thanks May 04 18:19:52 zeddii_home: http://git.openembedded.org/openembedded-core-contrib/commit/?h=kraj/master&id=22b2280f4978df41d9bd546773dce1b3538a35ca May 04 18:20:28 khem: I’ll grab it now and integrate it to the tree. May 04 18:20:37 zeddii: we need these fixes for ppc to work with gcc8 May 04 18:20:46 I think first one is good to sent to lkml too May 04 18:21:02 zeddii_home: I need similar fix for mips May 04 18:21:18 will let you know once I push it May 04 18:21:24 gotcha. I’ll track the one upstream and reference it in the commit. May 04 18:23:11 armpit: can you please take https://patchwork.openembedded.org/patch/149165/ to rocko as well? it's waiting for couple months May 04 18:25:56 JaMa, thats currently included in the stable/rocko-next bring built. it got caught up in the some the uninative issues May 04 18:43:19 armpit: there are few changes coming for ssp changes that I am getting into oe-core soon May 04 18:43:29 meta-openembedded intake is too slow May 04 18:44:13 are you waiting on maintainers too long ? May 04 18:55:41 zeddii_home: http://git.openembedded.org/openembedded-core-contrib/commit/?h=kraj/master&id=0e85aea10984c1cefe8712ca5ed643daf854cbef May 04 18:55:44 this is mips fix May 04 18:56:27 ack'd May 04 19:04:32 New news from stackoverflow: Can't connect to free wifi network with iwconfig in Linux embedded [on hold] May 04 19:47:10 zeddii_home: pushed another fix for mips64 as well :) May 04 21:43:21 RP: I have 3 major changes 1) ssp 2) gcc8 3)drop nis,nsl (obsoleted glibc features), I ran image tests on core-image-sato all looks good on all arches for both musl/glibc May 04 21:44:01 RP: one thing occured to me, that spe backends are on a lifeline in gcc 8 and can only be enabled with --enable-obsolete option to you know where they are headed May 04 21:44:22 so I wonder if we should also plan to drop ppc from first class citizen in 2.6 May 04 22:03:10 khem is ssp required for PPC support? I thought it was only useful on some of the e500/e5000 lines.. May 04 22:03:45 (and yes, I realize that e5000 is probably one of the last available PPC chips...) I'm also not sure if there are ANY PowerPC codes listed for 'new' designs anymore May 04 22:04:03 ssp does not work on ppc May 04 22:04:26 ok, then I'm confused May 04 22:04:43 ssp work I did is common it has nothing to do with ppc May 04 22:05:26 we were using gcc provided impls and libc provided impls in a non deterministic way May 04 22:05:47 the ssp rework ensures that we use libc provided impls always May 04 22:06:17 except for mingw32 where we need gcc provided libssp May 04 22:07:11 powerpcspe as a backend is deprecated and might disappear from gcc in 9.x time frame next year May 04 22:07:37 ya, I'm not surprised.. it's probably not getting much love.. May 04 22:07:58 so We have to start deprecating as well May 04 22:08:24 I am sure we can have year or two but it will be pain to maintain it during this time May 04 22:08:25 keeping support, but deprecating and explaining why would be fine with me.. if NXP (or someone else) shows up, we can reverse that May 04 22:08:53 our qemu is not using spe so we can keep usingit May 04 22:08:59 * fray goes back to demolishing the central chimney to the house.. :P May 04 22:09:02 its 74xx May 04 22:09:05 http://kernel.crashing.org/fray/chimney May 04 22:09:26 khem, ya... I would expect NXP to step up if they want further e* support.. May 04 22:09:47 (but I think we've moved to a dualopoly of IA and ARM right now... (I'm not seeing any new mips designs either) May 04 22:10:05 RISC V -might- be something to promote when it's ready, but it seems to be still in it's infancy... May 04 22:10:09 * fray really goes.. abck in a bit May 04 22:10:12 fray: dont forget riscv :) May 04 22:10:54 risc-v is coming in big way into non intel markets thats my prediction for next 5 years May 04 22:34:01 khem: I'm aware we have several series outstanding, thanks for the heads up May 04 22:34:21 khem: we should perhaps ask about ppc on the arch list, see who if anyone is intersted May 04 23:35:05 Re powerpc keep in mind there's a lot of stuff on long term support wrt hardware May 04 23:35:26 Someone stepped up for ppc8xx for uboot May 04 23:35:47 And they have to support it until 2020? 2022? I forget off hand May 04 23:36:26 And it's also true that old designs need to get supported with newer SW for security reasons May 04 23:36:55 I'm also not saying it needs to be done without finding people to help May 04 23:37:11 But I'm thinking it should be possible to find people May 04 23:37:42 Maybe poke some ppc lists too May 05 00:07:51 Tartarus: I agree, but I think its a professional ask :) May 05 00:08:20 no cool toys run ppc at my home so I have no incentive to keep it going May 05 00:14:59 RP: sent a v2 of the pull request, if there is slot give it a go on master-next May 05 00:15:30 I am holding off gcc8 for subsequent pull request after this lands in master safely May 05 00:35:29 New news from stackoverflow: Which module to install to fix "module QtQuick.Controls.Styles is not installed" error for Yocto? May 05 01:32:22 fray: that's quite a pile of bricks you laid in the bathroom! May 05 01:32:33 https://www.urbandictionary.com/define.php?term=laying%20a%20brick May 05 02:32:57 tlwoerner goal is tomorrow putting them in buckets and getting them out of the house.. :) then finish removing the chimney to the floor of the bathroom.. (there is about 3 1/2 feet of brick left.) **** ENDING LOGGING AT Sat May 05 03:00:03 2018