**** BEGIN LOGGING AT Wed Jun 21 03:00:03 2017 Jun 21 07:17:38 Hi there, I'm trying to add apache2 and php support for apache2 in my image (working on morty branch), both are working separately, but php doesn't work on apache. I built php-modphp for php support on apache, and bbappend the recipe to add PACKAGECONFIG_append_class-target = " apache2", but still does not work Jun 21 07:18:40 also, the env var PHP5 is not set, so the conf 70_mod_php5 is not load Jun 21 07:18:53 https://github.com/openembedded/meta-openembedded/blob/morty/meta-oe/recipes-devtools/php/php/70_mod_php5.conf Jun 21 07:19:08 someone already deal with apache and php ? Jun 21 07:38:44 how come I get a QA error (expanded-d) when using ${D} in pkg_preinst? the docs say you shouldnt but not why Jun 21 08:36:01 rburton_: do you want me to do something with mut and mut2 builds from last night? Jun 21 08:43:47 from what i gather, devtool does not offer an equivalent to -c clean/-c cleansstate - so how could i force a full recipe rebulid there properly? Jun 21 09:19:00 LetoTheII: bitbake -c clean Jun 21 09:19:44 LetoTheII: I don't see any usecase for cleansstate with devtool, but use bitbake -c cleanstate if you need to Jun 21 09:20:46 Hello! I’m trying to add oracle java to my image but I’m getting “HTTP request sent, awaiting response... 401 Authorization Required” during the fetch. Anyone who knows how to fix this? Maybe I’m missing something? Jun 21 09:29:14 luc4: have you read the layer readme? it's probably oracle being oracle with regards to licenses... Jun 21 09:30:28 jku: yes, I did that already but it seems it still tries to download Jun 21 09:30:39 jku: maybe I’m missing that .done file? Jun 21 09:38:46 Hey guys. Here's my entry for "weird thing of the day". I've been trying to debug for a while I can't access https addresses (same problem with curl, wget or whatever) up until the uptime of my board was between 10 and 15 minutes. http addresses always work fine, https only after that wait time. Investigations brought the suspects towards gnutls using /dev/random, and /dev/random being *very* slow at initializing. After 10- Jun 21 09:38:48 15 minutes in fact I get: https://pastebin.com/ya4fY1GY Jun 21 09:40:17 Question is: could my core-image-minimal be too minimal to generate enough entropy at boot? Is there a good entropy generator in poky (or other layers)? Jun 21 09:42:15 diego_r: are you sshing to the board? Jun 21 09:42:46 the gnutls error immediately follows "random: nonblocking pool is initialized" Jun 21 09:43:15 rburton_: happens regardless of me ssh-ing or not Jun 21 09:43:33 keyboard input is a prime source of entropy you see Jun 21 09:43:42 so a minimal image over ssh has very little entropy Jun 21 09:44:29 "sysctl kernel.random.entropy_avail" will tell you how much is in the pool Jun 21 09:44:35 to verify if its a lack of entropy itself Jun 21 09:44:43 rburton_: indeed /proc/sys/kernel/random/entropy_avail is 0 as soon as I root login via command line Jun 21 09:45:27 yep, 0 also checking with sysctl of course Jun 21 09:45:46 theres a number of entropy daemons you can use Jun 21 09:47:11 your board may have a hardware RNG generator that you can enable Jun 21 09:47:45 rburton_: yep, my i.MX6 seems to provide /dev/hwrng Jun 21 09:48:14 so you probably want to use rng-tools to feed that into the pool Jun 21 09:48:25 or something, ask freescale Jun 21 09:48:54 jku: haven't looked at mut2 yet how bad was it? :/ Jun 21 09:49:17 rburton_: thanks for your precious help Jun 21 09:50:08 rburton_: well... I haven't checked extensively so not sure if the same failures, but lots of red anyway Jun 21 09:50:26 sanity tests and build failures Jun 21 09:51:38 I'll have a closer look Jun 21 09:55:00 ah, kanavin was right: qt4 is hosed Jun 21 09:56:30 might be time for retirement then unless someone steps up to fix it Jun 21 09:59:33 Hi ! I checked that the VPU from I.MX6 have two version of driver, one for linux-imx, one for linux-mainline. How can I be sure of the one is used besides the fact of looking for kernevel version? Like to see if there is backport... Jun 21 10:00:12 kanavin: have you seen the dnf sanity failures: https://autobuilder.yocto.io/builders/nightly-x86-64/builds/333/steps/Running%20Sanity%20Tests/logs/stdio Jun 21 10:17:36 kanavin : if you remember, yesterday night i asked to know which kernel I have. So I followed your tips Jun 21 10:17:55 kanavin : and I get linux-mainline_4.1.36-phy3.bb Jun 21 10:18:46 kanavin : But when i do uname -a it tells that is a linux-imx Jun 21 10:21:28 kanavin : so what does uname -a read? Jun 21 10:22:35 has some add an engine to OpenSSL on Yocto? Jun 21 10:22:53 I am running into a lot of problems to have it compile (it is compiling externally with the toolchain generated by Yocto) Jun 21 10:37:55 rburton_: any idea about this one https://autobuilder.yocto.io/builders/nightly-arm/builds/376/steps/Running%20ESDK%20Sanity%20Tests/logs/stdio Jun 21 10:38:23 that is not useful Jun 21 10:39:47 oh its because the step before failed Jun 21 10:39:55 so yes it didn't write the images Jun 21 10:40:59 oh right Jun 21 10:42:12 so the fact that it tries to run that step is probably a bug somewhere? Jun 21 10:44:50 would be nice if the steps would be able to have explicit deps and can skip run-tests if build-images worked Jun 21 10:48:29 how do i create a task within bbclass that runs after do_image and only lists in `bitbake -c listtasks core-image-*` aka not every single recipie. I want to take rootfs, kernel, dtb, and create a uboot image. And I could just INHERIT += 'kernelitb' like mender does it for its .sdimg in DEPLOY_DIR_IMAGE Jun 21 10:51:28 is Leo (sandoval) here btw? Jun 21 10:52:14 jku: lsandov is marked as away Jun 21 10:53:05 it's not even 06:00 for him Jun 21 10:53:34 oh yes :) Jun 21 10:53:38 ah thanks, was about to look up the time :) Jun 21 10:53:43 irc really needs localtime support Jun 21 11:08:14 ChrysD: man 2 uname Jun 21 11:08:26 jku: I probably succeeded in downloading the package manually. The .done file is mandatory it seems. However I’m getting this now: https://pastebin.com/9sMB4MBn. It seems like it expects Java 8, but the package in the layer is named ejre-7u${PV_UPDATE}-fcs-b${BUILD_NUMBER}-linux-arm-vfp-hflt-client_headless-07_may_2014.tar.gz. Isn’t this weird? Jun 21 11:10:30 jku: top work on SWATing Jun 21 11:10:54 jku: as i see your mails i've been reverting patches, so shout when you think you've looked at the main failures and i can refire mut2 Jun 21 11:11:51 jku: nope - I guess it's because of dnf update to latest upstream. Strange that the same error does not occur when building images (which I did test) Jun 21 11:12:00 rburton_: ^^^ Jun 21 11:12:52 rburton_: thanks, I think I'm done now. the qt4 failure you saw already Jun 21 11:13:12 cheers Jun 21 11:13:31 jku: rburton_: the qt4 failure should be dealt with by removing qt4 builds from the autobuilder Jun 21 11:13:34 yes Jun 21 11:13:40 enough is enough :) Jun 21 11:13:42 yeah agreed Jun 21 11:13:50 for now i'll remove the upgrade, and file a bug for joshuagl :) Jun 21 11:14:10 I'll look into dnf stuff now, as I have nothing else to do before holiday starts at the end of the week Jun 21 11:15:09 jku: I wanted to try the boot2qt image on one of your NUCs, but I guess that will wait Jun 21 11:15:38 kanavin: feel free to take any Jun 21 11:15:45 jku: they do not support qemu (because of required opengl support, I think) Jun 21 11:16:13 jku: sure, just would be better to have you present too Jun 21 11:16:24 ok Jun 21 11:27:21 kanavin : already done... it's just say that it give system information Jun 21 11:27:54 +1 for remove qt4 and add qt5! Jun 21 11:28:12 Crofton|work: adding qt5 is not going to happen :) Jun 21 11:28:44 sato forever! Jun 21 11:29:31 Crofton|work: qt folks have their own distro called boot2qt, which I'm now investigating Jun 21 11:29:40 hmm, interesting Jun 21 11:29:44 who are these people? Jun 21 11:29:48 or url? Jun 21 11:30:00 Crofton|work: they're not particularly forthcoming about its content, because it forms the base for their commercial embedded offering, but the layer is open source Jun 21 11:30:12 Crofton|work: http://code.qt.io/cgit/yocto/meta-boot2qt.git/tree/README Jun 21 11:30:30 kanavin : AH ok i understand !! You asked me to do "man 2 uname "... I didn't understand that is a the command... but sometimes like... man to uname.. Jun 21 11:30:54 ChrysD: yes, uname is a command and a kernel syscall Jun 21 11:31:07 man uname will give you the command, man 2 uname will give you the syscall details Jun 21 11:31:16 kanavin : i didn't know about "man 2 " Jun 21 11:31:45 kanavin : i was doing man uname as usually ahahah Jun 21 11:31:58 ChrysD: manual pages are split in sections Jun 21 11:32:55 lots of bbappends Jun 21 11:32:56 weird Jun 21 11:33:24 it would be really nice to have more modern images suitbale for tradeshow demos Jun 21 11:33:31 Crofton|work: yeah, it's not a work of beauty, and latest supported release is morty Jun 21 11:33:39 lol Jun 21 11:33:46 very diplomatic :) Jun 21 11:34:06 appends are better than copy/pastes :) Jun 21 11:34:12 Crofton|work: also, they do not support qemu, if you want emulator, it's virtualbox Jun 21 11:34:13 very diplomatic :) Jun 21 11:34:30 emulaors are so last decade :) Jun 21 11:34:46 kanavin : even morty is far new when you say the average commercial meta versions... Jun 21 11:34:57 kanavin : see* Jun 21 11:34:59 ChrysD: I live in a oe-core maintenance bubble Jun 21 11:35:11 ChrysD: 99,5% of my time is spent on the master branch Jun 21 11:35:23 kanavin : and me understanding all that stuff Jun 21 11:35:25 kanavin xD Jun 21 11:35:33 lol Jun 21 11:35:52 kanavin : I'm a big noob in everything and trying to make a report of how to use yocto and why the meta bsp of the vendor choosen is good/bad Jun 21 11:35:53 a pox on vendors with ancient rubbish :) Jun 21 11:36:35 ChrysD, if you fix version and stay on it Jun 21 11:36:45 you lose support from the larger project after a year Jun 21 11:36:46 kanavin : and when I see that the BSP vendor use a kernel with the wrong drivers for VPU and with the wrong gstreamer plugins... I realize that with linux embedded systems, is the software that drives the hardware choose ahahah Jun 21 11:37:02 so you need internnal CVE tracking team to keep your stuff up to date (wrt to CVE's) Jun 21 11:37:36 so you can use resources to maintain your code to current releases, or maintain other peoples code to old releases Jun 21 11:37:45 Crofton|work : So a big team that will do ton of backport from linux-mainline and make it working with linux-imx and pray that enegineer from that company works well and fastly to get features ahahah Jun 21 11:38:13 or supstrream as much as you can Jun 21 11:38:23 typically for non business specific items Jun 21 11:38:31 and focus on your area of expertise Jun 21 11:39:00 Crofton|work : Yeah but when the bsp vendor is providing a kind of out of box meta which include lot of fork, backport and copy/paste Jun 21 11:39:13 Crofton|work : you are very dependant from them Jun 21 11:39:31 but how long to they provide support? Jun 21 11:39:44 Crofton|work : the problem isn't support, the problem is tom ake it working Jun 21 11:39:45 and do they support all layers Jun 21 11:39:54 heh Jun 21 11:40:07 basically, invest time and efoort now or later Jun 21 11:40:24 Crofton|work : They are only on krogoth version and says that video works well... But they are using the wrong gstreamer plugins with the wrong driver and the wrong kernel. Jun 21 11:40:33 also, vendors need feedabck that customers see this problem Jun 21 11:40:48 Crofton|work : Yeah I have done a big report of my investigation Jun 21 11:40:50 also they need feedback their work is junk :) Jun 21 11:41:32 I find that the typical vendor engineer knows they have iussues, but their management tells them customers are not complaining Jun 21 11:42:01 Crofton|work : but you know, i'm a young guy in a complex world so how reliable I can be... that's the question I faced. Jun 21 11:42:19 Crofton|work : So I make an effort to understand what are the philosophy inside and making test... Jun 21 11:44:16 Crofton|work : Like for exemple, when you have an Issue, sometimes you know where to start. Like when I get many gstreamer problems, I thougth that the fault was more because of me and how I use the tool and maybe not the software made by a team of engineers that "supposely and surely" know what they are doing. Jun 21 11:44:59 ChrysD, I completely understand your situation Jun 21 11:45:05 Crofton|work : I started to contact maintener of the codes and tools directlty... And it's from there that I understand why it "couldn't " work Jun 21 11:46:20 Crofton|work : I had lot of support from the #yocto channel #gstreamer channel and ##linux channel and helped me because when you check for information on net with multiple document it make me like a bit confuse. Using something and undesrtanding something is quite different. Jun 21 11:47:04 in a way, open project hurt themselves because the amount of work is invisible to managers Jun 21 11:47:14 Crofton|work : When you have as only background skill, using as an user Ubuntu.. You don't really know how linux works ahahah. So it make complicate to have a kind of workflow to how to debug. Jun 21 11:47:24 Crofton|work : i have started to go deep into gstreamer code Jun 21 11:47:41 Crofton|work : When somebody tells me just that question " What's the gstreamer you are using, which kernel are you using". Jun 21 11:48:19 well good luck Jun 21 11:48:24 Crofton|work : and the plugins ' that the bsp vendor tell me to use ' wasn't made for the drivers that is inside the kernel i'm using and the development of the tool ' at the moment of the version of the kernel' wasn't good. Jun 21 11:49:05 Crofton|work : So know I know a little more of what are the question i should ask my self to try to debug... Jun 21 11:49:07 feels like vendot made one specific set of stuff owrked, declared victory and moved on Jun 21 11:49:33 Crofton|work : But the plugins was made to use a driver of the linux-mainline... but they are using linux-imx. So couldn't work ahah Jun 21 11:49:49 Crofton|work : But the commercial says that its works... As supposely it "hardwarely" can work. Jun 21 11:50:09 I suspect the gstreamer guys can't support all vendors special kernels Jun 21 11:50:36 Crofton|work : and as my team come from the microcontrollers worlds ans RTOS worlds which most of time they do 100% of the software jobs in their own... for them buying a board is only matter of the hardware and the software doesn't really matter Jun 21 11:51:48 Crofton|work : So when you see, oh , there is a video decoder and it's compatible with krogoth version... it should work as the linux open source world is really rich.... Until you know that the darkside of it is most of time bsp vendor fork project and you don't have this rich ecosystem. Jun 21 11:51:50 ChrysD: if you want to use vendor hardware then you tend to need to use the entire stack. you can kick a fuss if its ancient, but at least if you use what they support then you can get help from them if it doesn't work Jun 21 11:52:07 heh Jun 21 11:52:28 assuming they have an effective support team .... Jun 21 11:52:48 Crofton|work : so the consideration, the choice you made on linux embdded system are a bit different when you choose a microcontroller or a board with RTOS support... Jun 21 11:53:37 Crofton|work: yeah big assumption Jun 21 11:54:18 Crofton|work : Like for example when you make the choice of the board, you shall now before buying, what you would like ... So " i want video streaming". Ok let's check to what the linux ecosystem give... Ok gstreamr cool ! So it should work... But no warning ! Look if in the software of the board, it's using the good version with the good kernel... etc etc etc Jun 21 11:55:01 the evaluation should be look at the vendors sw practices and compare with best practices Jun 21 11:55:14 Crofton|work : yeah Jun 21 11:55:14 in a perfect world Jun 21 11:55:38 annoyingly, sometimes the vendor with the best hardware ahsd the worst software practices Jun 21 11:55:39 Crofton|work : but when you didn't know what linux is.... And you think that yocto is the name of an Ikea furniture... You need to experience once. Jun 21 11:56:10 lol Jun 21 11:56:20 Yocto is Ikea furntiture Jun 21 11:57:05 Crofton|work : And it's the point where I am. So i can tell the team here that the compared to " microcontroller and RTOS world" the software is a big chooce. Jun 21 11:57:21 Crofton|work : Because before in industry you get the work of developer, but more and more you have the work of integrator Jun 21 11:57:36 yep Jun 21 11:57:55 Crofton|work : and that's what I understood so that you can drive your project only by the hardware capabilities Jun 21 11:58:26 Crofton|work : you can't * Jun 21 11:59:44 Crofton|work : So it was very rich for my internship. Jun 21 11:59:51 I suspect this culture transition is partly why the hardware vendors are getting away with vendor kernels Jun 21 12:00:08 the attitude you own the entire stack for your hw Jun 21 12:01:12 Crofton|work : and the problem is also the buyer... That it doesn't really know how you can judge a linux embedded system vendor Jun 21 12:01:41 Crofton|work : Because if there is a HW decoder on your board, supposely, they support well that. Jun 21 12:04:10 rburton_: rng-tools much improved the situation. Still it takes something like a minute (compared to 10-15 mins previously) to get "nonblocking pool is initialized". Strange is entropy is (seemingly) high since when login prompt is ready kernel.random.entropy_avail = 3095. Jun 21 12:08:58 diego_r : What does "nonblocking pool is initialized" means? I have it lot of time. Jun 21 12:11:09 ChrysD, https://www.google.com/search?q=nonblocking+pool+is+initialized&oq=nonblocking+pool+is+initialized&aqs=chrome..69i57.650j0j4&sourceid=chrome&ie=UTF-8 Jun 21 12:11:16 ChrysD: it means the random.c driver has enough entropy to provide /dev/random Jun 21 12:13:01 diego_r : so that mean that software which depends on /dev/random have to wait until that message. Jun 21 12:13:51 ChrysD: https://forums.xilinx.com/t5/Embedded-Linux/random-nonblocking-pool-is-initialized/m-p/739788#M17915 Jun 21 12:15:14 diego_r : interesting! Jun 21 12:15:34 ChrysD: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux-stable.git/tree/drivers/char/random.c?h=v4.1.15#n658 Jun 21 12:15:59 Interestingly the printout has been removed (or moved) in recent kernels Jun 21 12:18:18 diego_r : thanks Jun 21 12:31:10 Hi. Does anyone know why do_kernel_configcheck is so slow? If a several fragments remove various lines from the .config file, it spends more than 20 minutes to complete the task Jun 21 12:38:43 Crofton|work: we've been getting some less fragmentation mantaining a common fslc kernel base Jun 21 12:38:56 good :) Jun 21 12:39:00 Crofton|work: few vendors are using it as base for their bsp Jun 21 12:39:14 Crofton|work: not perfect but a step further Jun 21 12:39:26 we need to make progress :) Jun 21 12:39:36 Crofton|work: Phytec, is not of one using it ;-) Jun 21 12:40:07 linux-yocto is not perfect, but using it + patches at leasts show delta from mainline Jun 21 12:40:31 or lts kernel Jun 21 12:40:38 hallo Jun 21 12:40:50 btw, thanks otavio for your wic integration work in meta-freescale-3rdparty, I'm using it a lot! Jun 21 12:41:02 Crofton|work: right but usually ARM-based BSP has a bunch of patches Jun 21 12:41:12 Crofton|work: so not easy to put Jun 21 12:41:17 diego_r: :-) Jun 21 12:41:40 diego_r: check the 'use-mainline-bsp' override ;-) Jun 21 12:42:04 conditioning vendors to build recipe with linux-yocto + patches, then upstream patches so next release has fewer patches is a path to easy maintenance Jun 21 12:42:45 Crofton|work: agreed but not many vendors does it; also their upgrade roadmaps does not match with yp ones Jun 21 12:43:02 yeah, this is a headache Jun 21 12:43:05 Crofton|work: as their delta is big, it is not trivial Jun 21 12:43:12 need a work flow that moves them in the right direction Jun 21 12:43:21 otavio: I've seen that too, VERY useful thanks Jun 21 12:43:24 Crofton|work: i've been trying Jun 21 12:43:45 diego_r: :-) Jun 21 12:44:32 I need help. I want to overwrite a device tree with another one in a bbappend. I want to copy this dt in the do_build_prepend() task and use the variable $S. But this points to the work directory. But the kernel sources are in the work-shared directory. Is there another variable that could I use to point to the right kernel sources? Jun 21 12:45:20 eddi: prepare a patch Jun 21 12:45:28 eddi: best practice Jun 21 12:47:42 Ok. I try this. Thanks. Jun 21 12:49:10 Crofton|work : I guess sometimes you made some quite of decisions as an engineer between what's the delay the leaders give you to do and develop so that did not give you enough time and making a kidn of " make it working" Jun 21 12:53:55 Hi, I'm building a new dts for my use and have to add it to the makefile but the makefile in the /arch/arm/boot/dts gets rebuilded when I start bitbake. From where does bitbake get the information on how to build the makefile? And how may I change that or get my value to display there? Jun 21 12:59:18 Can I get ${DEPLOYDIR} within do_image_custom? Jun 21 13:01:31 jku: rburton: updated patchset for dnf ready Jun 21 13:02:50 cool Jun 21 13:02:51 thanks alex Jun 21 13:05:46 rburton: check mesa one pls Jun 21 13:13:33 rburton: I am going on holiday for 3 weeks after tomorrow Jun 21 13:13:52 rburton: ad hoc car trip in bavaria :) Jun 21 13:14:05 the car is booked, but nothing else Jun 21 13:14:19 sounds like that could be fun Jun 21 13:14:20 have a good time Jun 21 13:14:34 i'll throw the dnf series at the ab shortly to see what happens Jun 21 13:15:13 cheers Jun 21 13:16:29 kanavin: https://autobuilder.yoctoproject.org/main/tgrid <-- just master+your 3 patches, so it should be fast and green right? :) Jun 21 13:17:22 Hello! Anyone else getting this when trying to add oracle java to an arm image: https://pastebin.com/9sMB4MBn? It seems like rules are for java 8, while the package is for java 7. Any idea? Jun 21 13:18:06 @luc4 seems that the paste has been removed Jun 21 13:18:14 jku: the checkuri failure on the mut2 respin was transient, it appears to work now Jun 21 13:18:14 rburton: I can neither confirm nor deny anything whatsoever :) Jun 21 13:18:49 https://en.wikipedia.org/wiki/Glomar_response Jun 21 13:18:51 alexgeorgescu: weird, I just opened it…. this is another one https://pastebin.com/yXYN78t0 Jun 21 13:34:09 * kanavin is *super* impressed with boot2qt Jun 21 13:34:18 as expected, it's light years ahead of sato :) Jun 21 13:35:10 bluelightning: ^^^ Jun 21 13:35:32 kanavin: cool - sounds like I should try it out Jun 21 13:35:51 bluelightning: if you have a NUC around, I can give you a link to the image I built Jun 21 13:36:29 kanavin: congrats you're now in charge of demo machines at the next conf Jun 21 13:36:54 kanavin : is it ironical? Jun 21 13:37:04 kanavin: I don't unfortunately Jun 21 13:37:21 kanavin: i do link please Jun 21 13:39:35 rburton: just a sec, first I send bluelightning a screenshot :-) Jun 21 13:39:42 sure Jun 21 13:43:30 bluelightning: you should have a screenshot at least :) Jun 21 13:45:38 kanavin: thanks, looks neat Jun 21 13:47:06 kanavin: may I see it? Jun 21 13:47:13 * otavio is curious Jun 21 13:48:50 otavio: the image size is 1.4 gigabytes, but I can forward you the screenshot, sure Jun 21 13:50:09 kanavin: it is fine; I am on a good internet now ;-) Jun 21 13:50:11 LOL Jun 21 13:51:49 yay for broadband in Pelotas :-) Jun 21 13:53:40 otavio: I am uploading the image now to google drive Jun 21 13:54:00 surprisingly not very slow Jun 21 13:58:25 ;-D Jun 21 13:58:29 diego_r: \m/ Jun 21 13:59:16 https://drive.google.com/open?id=0B_LMOpJvp7Vrc1NtUWxYZUpybWs Jun 21 13:59:28 otavio: rburton: bluelightning: ^^^ Jun 21 14:01:56 kanavin: screenshot? Jun 21 14:02:26 otavio: that's the image for the NUC Jun 21 14:02:35 kanavin: i got the image; will get a NUC to try Jun 21 14:02:44 kanavin: dd for usd? Jun 21 14:02:49 otavio: yes Jun 21 14:02:55 good Jun 21 14:02:56 usb Jun 21 14:03:45 otavio: forwarded you the screenshot by email too Jun 21 14:06:58 kanavin : why so happy about that boot2qt? Jun 21 14:07:44 maybe 'cause sato is so horrible :-D Jun 21 14:08:28 ChrysD: because Yocto needs a 'real UI' Jun 21 14:28:57 How can I tell bitbake that I want him to update the dts files? I build one it compiled but now that I updated it he just skips it cause he knows it. Jun 21 14:48:05 kanavin : what do you mean by yocto need an UI ? You mean driving a yocto project through a UI instead of commandline by having a kind of image that give a live system to computer for building systems? Jun 21 15:03:27 Duffy__: I forced it to rebuild the DTB by deleting the one in the ../build/tmp/work-shared/ directory. I forget where exactly it lives, but find should be able to pull it out for you. Jun 21 15:03:44 I imagine there's a better way, but I'm dumb. Jun 21 15:07:31 Thats smarter than what I did :-) Just naming them by higher numbers can get me into a lot of trouble. Ty Jun 21 15:22:56 whats the intended usage of image-container.bbclass? i set my trivial test case to inherit it, but all i get then is a bunch of json files Jun 21 15:26:27 i'm trying to build the cross-compile toolchain as described here: http://variwiki.com/index.php?title=Yocto_Toolchain_installation&release=RELEASE_MORTY_V1.0_DART-6UL Jun 21 15:26:38 i've setup the yocto environment Jun 21 15:27:37 but the second command in section 2, MACHINE=imx6ul-var-dart DISTRO=fslc-x11 build_x11, fails with "failed, build_x11 not found" Jun 21 15:28:12 what am i missing? Jun 21 15:29:20 yates: this build_x11 seems to be some variscite specific script Jun 21 15:30:09 most probably you missed to source the environment that brings the path for it Jun 21 15:30:17 see also http://variwiki.com/index.php?title=Yocto_Build_Release&release=RELEASE_MORTY_V1.0_DART-6UL#Build_X11_GUI_demo_image Jun 21 15:37:37 thanks LetoThe2nd Jun 21 15:38:10 OpenSSL is not installing the engines in /usr/lib/ssl/engines Jun 21 15:38:12 how to fix that? Jun 21 15:46:39 LetoThe2nd: it's not that i missed it, it's that they omitted (accidently, no doubt) it from that step of their wiki page. i.e., wrong directions... Jun 21 16:06:41 ChrysD: no, what I mean is that when you boot a yocto image, what you see is either a cold, uncaring command line, or the sato interface, which is limited and outdated Jun 21 16:07:21 ChrysD: we need something graphical that showcases common embedded use cases in a modern, pretty UI, and boot2qt fills that gap fantastically Jun 21 16:09:31 "cold, uncaring command line" Jun 21 16:09:42 * Crofton|work takes offense! Jun 21 16:12:58 Crofton|work: if you want to play with boot2qt, Iposted the link to the image above :) Jun 21 16:15:03 kanavin : Are you speaking about a splash screen? Jun 21 16:27:54 kanavin, for NUC? I suppose I should spin a build for my rpi-3 Jun 21 16:28:13 although I do not want to unseat myself as demo master for YP stand Jun 21 16:34:05 Crofton|work: yep, it's a NUC image Jun 21 17:00:09 I don't have a NUC Jun 21 17:11:57 how come I get a QA error (expanded-d) when using ${D} in pkg_preinst? the docs say you shouldnt but not why Jun 21 18:42:51 whats the intended usage of image-container.bbclass? i set my trivial test case to inherit it, but all i get then is a bunch of json files Jun 21 18:59:43 * Crofton|work studies meta-boot2qt to figure out how it works without all the stuff to make it "easy" Jun 21 19:13:54 Crofton|work: you might find this helpful http://doc.qt.io/QtForDeviceCreation/qtee-custom-embedded-linux-image.html Jun 21 19:14:29 Crofton|work: exporting MACHINE is important or nothing will work as advertised :) Jun 21 19:14:46 heh, I figure I can work it out from first principles Jun 21 19:15:05 I'll lol hard if meta-poky is required :) Jun 21 19:15:46 Crofton|work: they hide that part, they do their own layer fetching, and you don't need to know or care Jun 21 19:15:57 ah, but I do care :) Jun 21 19:17:16 actually i second Crofton|work. to me theres no point in them doing another set of magic tricks to hide the oe mechanisms Jun 21 19:17:56 it does look like I add to my checout and add a couple of layers to bblayers, then work out required settings in local.conf for distroless build Jun 21 19:18:39 why cant' they just pour stuff into a layer and properly document its usage? Jun 21 19:18:57 everyone is special :) Jun 21 19:19:24 interesting Jun 21 19:19:30 of course. especially me, i'm the most special of all! Jun 21 19:19:37 they have a distro.conf and use meta-poky Jun 21 19:20:04 LetoThe2nd: because it's primarily a commercial shrinkwrapped product right now, they will ship you a big blob, and everything is done via Qt Creator IDE Jun 21 19:20:47 cloning meta-boot2qt and working with it from command line works, but is not supported or encouraged in any way Jun 21 19:21:15 it's a bit of luck that layer is even published, as they could as well keep it hidden Jun 21 19:21:17 kanavin: which makes only sense if you are the hardware vendor and want to sell the platform. in their case, i jsut find it annoying for people who already knwo their way around OE Jun 21 19:21:57 LetoThe2nd: this is how they market the stuff https://www.qt.io/qt-for-device-creation/ Jun 21 19:22:07 "request a free trial" Jun 21 19:22:20 hooray Jun 21 19:22:53 i found the documentation and proceeding so scary that i shied away from giving it a try Jun 21 19:22:55 if we use it to promote OE, we need to understand how it works and fits in to what we want to promote Jun 21 19:23:44 I'm all for peopl etryin gto profit, but I expect them to be good community memebrs also Jun 21 19:23:54 at the moment this is to me totally not what i would recommend to anybody lloking for a ui platform that fits into the oe ecosystem/mindset Jun 21 19:24:17 as it seems to just feast on the ecosystem and ignore the mindset. Jun 21 19:25:19 anyways, can anybody shed some light on the intended usecase of https://github.com/openembedded/openembedded-core/blob/master/meta/classes/image-container.bbclass for me? Jun 21 19:25:51 look like it should be an additional inherit after inherit image, and just trigger some warning if you don't use linux-dummy Jun 21 19:26:03 LetoThe2nd: maybe, but the UI is light years ahead of sato, so I suggest we talk to them and find common ground :) I can imagine they would be very happy to become the officially endorsed UI of oe/yocto Jun 21 19:26:22 Crofton|work: ^ Jun 21 19:26:45 kanavin: do we know whos behind it? Jun 21 19:27:06 LetoThe2nd: the Qt Company of course - same people making Qt itself Jun 21 19:27:19 hi yocto Jun 21 19:28:01 kanavin: yeah, but some mastermind specifically? like jama for meta-oe, koen/khem/xora for angstrom, RP/rburton for us... ;-) Jun 21 19:28:28 LetoThe2nd: layer's README does list a person :) Jun 21 19:29:00 http://code.qt.io/cgit/yocto/meta-boot2qt.git/tree/README Jun 21 19:29:49 and I can see that he has maybe half the commits in the repo Jun 21 19:30:03 I am working on creating a symlink to a library to another directory. ex.. /lib64/libxyz -> /lib Jun 21 19:30:24 kanavin: sounds like we should try to get in touch Jun 21 19:30:36 I found IMAGE_POSTPROCESS_COMMAND is not useful Jun 21 19:30:42 LetoThe2nd: yes, of course Jun 21 19:30:59 i'M GLAD THEY ARE USING META-QT5 Jun 21 19:31:02 der Jun 21 19:31:39 what I really need is a tool to let me fiddle with bblayers :) Jun 21 19:31:54 because it put the full path from system Jun 21 19:32:04 * LetoThe2nd hands Crofton|work a vim binary Jun 21 19:32:09 here you go Jun 21 19:32:18 lol Jun 21 19:32:56 Is there any efficient method to make symlink within rootfs Jun 21 19:33:01 vickylinuxer: this sounds like a xy problem. you're doing x, and think y is a way to archieve your goal. you're stuck doing y, and ask us about it. why not just ask about x directly? Jun 21 19:33:34 vickylinuxer: any recipe can do symlinks in the install stage. but like i said - it sounds like xy Jun 21 19:34:17 Crofton|work: except they use their own fork of meta-qt5 :) I really suggest you trust their layer fetching script :) Jun 21 19:34:31 that's absurd Jun 21 19:34:44 But I want do it in image postprocess stage Jun 21 19:35:15 None of my profesional dealings with Qt have given me the warm fuzzies.. unless you are willing to pay them Jun 21 19:35:50 Crofton|work: but if you want to be dealing with mysterious build failures or an image that does not boot, be my guest... Jun 21 19:35:51 and pisses me off as they are pulling some of my work into their fork :) Jun 21 19:38:53 vickylinuxer: Why specifically image postprocess stage? Jun 21 19:41:18 I dont want to add any bbappend for the recipe Jun 21 19:42:10 I just wanna add it in some_image.bbappend / local.conf entry Jun 21 19:44:12 : Sorry I am really dont know how to exactly reply to you. Newbie to IRC :-) Jun 21 19:48:23 vickylinuxer, you are doing better than most ... Jun 21 19:49:02 vickylinuxer: Me too ;) AFAIK, you just type in the username + ":" manually before the message Jun 21 19:50:53 jpew + ":" oh.. Thank You Jun 21 19:51:42 vickylinuxer: lol, almost. just jpew: Jun 21 19:51:57 jpew: Sorry. I messed up. Now it should be fine Jun 21 19:52:16 vickylinxer: anyway, it is a little odd that you want to append an image but not a recipe Jun 21 19:53:25 jpew: yes, I know.. But really I want to do this way. Jun 21 19:54:50 vickylinuxer: Fair enough. I think you might hev 2 options: 1) make a new recipe that make the symlink, or 2) use IMAGE_POSTPROCESS_COMMAND which *should* work... what issue did you have? Jun 21 19:57:02 jpew: I tried 2nd option. The symlinks are absolute path of from host (Ex., /home/user/..../libxyz). Not in a fakeroot environment Jun 21 19:58:06 jpew: I want it like ( /lib64/libxyz -> /lib/xyz) Jun 21 20:01:14 vickylinuxer: what does the "ln" command that you are using look like? Jun 21 20:03:10 jpew: ln -s ${IMAGE_ROOTFS}/lib64/libxyz ${IMAGE_ROOTFS}/lib Jun 21 20:04:04 vickylinuxer: Ah, I think you need to omit the first ${IMAGE_ROOTFS}: ln -s /lib64/libxyz ${IMAGE_ROOTFS}/lib Jun 21 20:08:04 hey everyone.. I'm trying to build my first layer (following a tutorial), which builds the nano text editor.. However, I'm getting an error I can't debug.. Can someone give me a hand? This is the error: https://pastebin.com/q46uLSsT this is the error's log: https://pastebin.com/tLr8RV1v this is the recipe: https://pastebin.com/DEsg8wUt I think the error is due to the do_fetch function in nano.bb, but I can't figure out why Jun 21 20:09:15 jpew: no words.. I hope this should workkkkkk.. Thanks a lot.. Jun 21 21:46:24 http://git.yoctoproject.org/cgit/cgit.cgi/meta-cloud-services/tree/meta-openstack/recipes-devtools/python/python-pep8_1.6.2.bb seems to have a mismatched checksum in it Jun 22 00:25:18 so who can get me one of this https://minnowboard.org/minnowboard-3 **** ENDING LOGGING AT Thu Jun 22 03:00:01 2017