Monday, 2020-03-09

*** robert_yang <robert_yang!~robert@> has quit IRC00:11
*** robert_yang <robert_yang!~robert@> has joined #yocto00:11
*** robert_yang <robert_yang!~robert@> has quit IRC00:31
*** robert_yang <robert_yang!~robert@> has joined #yocto00:32
yoctiNew news from stackoverflow: Yocto image on raspberry pi 3 [closed] <>00:40
*** robert_yang <robert_yang!~robert@> has quit IRC00:41
*** stew-dw_ <stew-dw_!~stew-dw@2607:fb90:a22f:7e8:f4ce:b8a0:81f1:7321> has joined #yocto00:41
*** robert_yang <robert_yang!~robert@> has joined #yocto00:42
*** stew-dw <stew-dw!~stew-dw@> has quit IRC00:45
*** camus1 <camus1!~Instantbi@> has joined #yocto00:46
*** kaspter <kaspter!~Instantbi@> has quit IRC00:48
*** camus1 is now known as kaspter00:48
*** net_wayfarer <net_wayfarer!net_wayfar@gateway/shell/ircnow/x-innvqzcygohnpwsy> has quit IRC00:57
*** agust <agust!> has quit IRC01:18
*** dev1990 <dev1990!> has quit IRC01:59
*** net_wayfarer <net_wayfarer!net_wayfar@gateway/shell/ircnow/x-uvcjakscoannnkef> has joined #yocto02:00
*** d_s_e <d_s_e!~d.s.e@> has quit IRC02:24
*** sj69 <sj69!> has quit IRC02:33
*** robert_yang <robert_yang!~robert@> has quit IRC02:44
*** robert_yang <robert_yang!~robert@> has joined #yocto02:45
*** stew-dw_ <stew-dw_!~stew-dw@2607:fb90:a22f:7e8:f4ce:b8a0:81f1:7321> has quit IRC02:56
*** otavio <otavio!~otavio@debian/developer/otavio> has quit IRC03:06
*** otavio <otavio!~otavio@> has joined #yocto03:08
*** otavio <otavio!~otavio@debian/developer/otavio> has joined #yocto03:08
yoctiNew news from stackoverflow: Added meta-oe recipe failed <>03:10
*** stew-dw <stew-dw!~stew-dw@> has joined #yocto03:11
*** stew-dw <stew-dw!~stew-dw@> has quit IRC03:14
*** otavio <otavio!~otavio@debian/developer/otavio> has quit IRC03:15
*** stew-dw <stew-dw!~stew-dw@2607:fb90:a22f:7e8:f4ce:b8a0:81f1:7321> has joined #yocto03:16
*** otavio <otavio!~otavio@> has joined #yocto03:17
*** otavio <otavio!~otavio@debian/developer/otavio> has joined #yocto03:17
*** otavio <otavio!~otavio@debian/developer/otavio> has quit IRC03:23
*** otavio <otavio!~otavio@> has joined #yocto03:25
*** otavio <otavio!~otavio@debian/developer/otavio> has joined #yocto03:25
*** sj68 <sj68!> has joined #yocto03:28
*** sj68 <sj68!> has quit IRC03:29
*** sj52 <sj52!> has joined #yocto03:30
sj52How to build nfs in yocto? Is there any command in yocto to build nfs? Basically, I don't want to create rootf.tar.gz everytime, instead I want to just update the unzipped root file system that I point to my nfs server. How can I do so?03:38
*** vineela <vineela!vtummala@nat/intel/x-iggbghpmkyittpdw> has joined #yocto03:51
*** NiksDev2 <NiksDev2!~NiksDev@> has quit IRC04:05
*** NiksDev2 <NiksDev2!~NiksDev@> has joined #yocto04:05
*** nerdboy <nerdboy!~sarnold@gentoo/developer/nerdboy> has quit IRC04:17
*** mrc3 <mrc3!~mrc3@linaro/mrc3> has joined #yocto04:29
erbosj52: I don't think there's anything that automatically does this. But if you really want to do it with bitbake it shouldn't be too hard to add a simple image-type for this. See
*** bluelightning <bluelightning!~paul@pdpc/supporter/professional/bluelightning> has quit IRC05:33
*** hyper_dave <hyper_dave!~quassel@> has quit IRC05:35
sj52erbo - Thanks, but I don't want to build the rootfs.tar.gz all the time. In fact, I want to just build an unextracted rootfs which gets updated everytime I run a build command. Basically, this unextracted rootfs is my nfs root. I want to point this rootfs there and don't want to change the rootfs source location. I will just reboot my device to get05:37
sj52the updated contents from rootfs.05:37
sj52Is there any way to do so?05:37
*** leeboby <leeboby!777b87a8@> has joined #yocto05:38
*** bluelightning <bluelightning!~paul@pdpc/supporter/professional/bluelightning> has joined #yocto05:39
erboI guess yes, if you 1. add a new new type that handles the copying of IMAGE_ROOTFS to a location exposed by NFS. 2. Remove other image types (like tar.gz) from your IMAGE_FSTYPES.05:40
erboYou mean uncompressed and not unextracted, right?05:41
*** hyper_dave <hyper_dave!~quassel@> has joined #yocto05:43
sj52Sorry, I mean uncompressed05:54
sj52$ls -l /home/shreyas/rootfs_microzed/05:55
sj52root    root        4096 Mar  9 15:50 etc05:55
sj52root        4096 Mar  9 09:19 run05:55
*** pharaon2502 <pharaon2502!> has joined #yocto06:01
*** ibinderwolf <ibinderwolf!> has joined #yocto06:13
*** ibinderwolf <ibinderwolf!> has left #yocto06:14
*** ibinderwolf <ibinderwolf!> has joined #yocto06:14
*** hamis_lt_u <hamis_lt_u!~irfan@> has joined #yocto06:20
*** sno <sno!> has quit IRC06:26
*** m1ster_r0b0t <m1ster_r0b0t!> has quit IRC06:28
*** vineela <vineela!vtummala@nat/intel/x-iggbghpmkyittpdw> has quit IRC06:30
*** AndersD <AndersD!> has joined #yocto06:34
*** goliath <goliath!> has joined #yocto06:48
*** panfir <panfir!~crs@> has joined #yocto06:55
*** sno <sno!~sno@> has joined #yocto07:01
*** guerinoni <guerinoni!> has joined #yocto07:02
*** agust <agust!> has joined #yocto07:14
*** jobroe <jobroe!~manjaro-u@> has joined #yocto07:16
*** frsc <frsc!> has joined #yocto07:17
*** rob_w <rob_w!~bob@unaffiliated/rob-w/x-1112029> has joined #yocto07:21
*** pohly <pohly!> has joined #yocto07:26
*** sstiller <sstiller!> has joined #yocto07:31
*** lucaceresoli <lucaceresoli!> has joined #yocto07:37
*** kroon <kroon!~kroon@> has joined #yocto07:38
yoctiNew news from stackoverflow: Yocto: add musl library giving build error <>07:41
*** vermaete <vermaete!> has joined #yocto07:42
vermaeteAnybody online who has experience with re-generating the python manifest?07:44
*** guerinoni <guerinoni!> has quit IRC07:44
*** guerinoni <guerinoni!> has joined #yocto07:44
*** AndersD <AndersD!> has quit IRC07:47
*** guerinoni <guerinoni!> has quit IRC07:50
*** fl0v0 <fl0v0!> has joined #yocto07:50
*** guerinoni <guerinoni!> has joined #yocto07:53
*** yacar_ <yacar_!> has joined #yocto07:59
*** leon-anavi <leon-anavi!~Leon@> has joined #yocto08:04
*** hamis_lt_u <hamis_lt_u!~irfan@> has quit IRC08:08
*** goliath <goliath!> has quit IRC08:09
*** guerinoni <guerinoni!> has quit IRC08:14
*** guerinoni <guerinoni!> has joined #yocto08:15
*** gtristan38 <gtristan38!6e0be3b1@> has joined #yocto08:15
gtristan38Hi, so I've successfully addressed the dozens of issues in my client's project which caused the builds to break when you enable sstate caching (mostly typical stuff, like naively linking into build directories of other recipes, or installing directly into the staging sysroot instead of the ${D} install directory)08:17
*** leeboby <leeboby!777b87a8@> has left #yocto08:17
gtristan38And now that I have builds that dont break, I'm faced with a new problem I'm not entirely sure how to debug08:17
gtristan38That is: Some recipes are getting rebuilt from one clean build to another, even though the sources have not changed *at all* between builds and the same sstate cache is present08:18
gtristan38Any tips on how to debug this brand of problem would be appreciated, what to watch out for etc08:19
gtristan38For one thing: It would be nice to be able to track what inputs are going into the sstate cache keys, and ensure that cache keys are stable (unchanging between invocations of bitbake when the sources don't change)08:20
gtristan38Any tips on what bitbake voodoo I might invoke to track what inputs are being considered in the calculation of an sstate cache key, and/or just "bitbake show-key recipe-name" to just simply view the key for a given recipe ?08:21
LetoThe2ndbitbake diffsigs maybe, or whatsitcalled?08:21
gtristan38LetoThe2nd, that appears helpful thanks :)08:24
gtristan38it's a program called bitbake-diffsigs here08:24
*** hamis_lt_u <hamis_lt_u!> has joined #yocto08:24
gtristan38Ah that is quite helpful indeed, I'll have to double check this all, but it managed to explain to me precisely what changed (with a bit of digging)08:31
vermaeteHi,  I've some questions about re-generating the python3 manifest.08:31
vermaeteWhen I do in my project, what's on the Zeus release, the output without changing the input is a bit different.08:31
vermaeteAnd my rootfs gives an error.08:31
vermaeteThe error seems to be related with the fact that python3-pkgutil is with the manifest before re-generating in an own package.08:32
* gtristan38 also discovers bitbake-whatchanged08:32
vermaeteAnd after re-generating it is in python3-core.  And python3-pkgutils seems to be empty.08:33
*** mihai <mihai!~mihai@unaffiliated/mihai> has joined #yocto08:35
*** TobSnyder <TobSnyder!> has joined #yocto08:39
*** jobroe <jobroe!~manjaro-u@> has quit IRC08:41
*** jobroe <jobroe!~manjaro-u@> has joined #yocto08:45
*** LocutusOfBorg <LocutusOfBorg!~LocutusOf@> has joined #yocto08:46
*** LocutusOfBorg <LocutusOfBorg!~LocutusOf@ubuntu/member/locutusofborg> has joined #yocto08:46
*** PinkSnake <PinkSnake!51ff1123@> has joined #yocto08:49
*** mckoan|away is now known as mckoan08:49
mckoangood morning08:49
*** yann <yann!> has quit IRC08:51
LetoThe2ndmckoan: howdy!08:52
*** Bunio_FH <Bunio_FH!> has joined #yocto08:52
PinkSnakeMorning guys08:53
*** Bunio_FH <Bunio_FH!> has quit IRC08:57
*** PaowZ_ <PaowZ_!~Vince@> has joined #yocto09:04
Dracos-Carazzahello, I've a strange issue. for some reason my sdk exports also a huge set of source files (.c/.cpp)09:05
*** nopeman <nopeman!d54413ba@> has joined #yocto09:05
nopemansmall question - I'm trying to flash Renesas R-Car M3 board and when I'm going into SCIF Download mode it's locked on some weird memory address like here:
nopemanany idea how to reset it?09:05
*** kroon <kroon!~kroon@> has quit IRC09:05
*** PaowZ <PaowZ!~Vince@> has quit IRC09:05
Dracos-Carazzanopeman, what power supply do you use? I had the same issue on my H3 when my power supply didn't deliver enough amps09:06
*** florian_kc <florian_kc!~florian_k@Maemo/community/contributor/florian> has joined #yocto09:06
*** NiksDev <NiksDev!~NiksDev@> has joined #yocto09:08
*** gtristan38 <gtristan38!6e0be3b1@> has quit IRC09:08
nopemanaxio met Ax-3010ds 0-30V 0-10A09:08
nopemanim giving it 5v and it's taking below 1A09:09
qschulzJPEW: I doubt you want to append to MACHINEOVERRIDES as the default value of MACHINEOVERRIDES is ${MACHINE} and I don't think it quite makes sense to have something more specific than the name of the machine.09:10
*** goliath <goliath!~goliath@> has joined #yocto09:11
qschulzJPEW: MACHINEOVERRIDES is "read" from rightmost to leftmost. Though having ${MACHINE} as default value for MACHINEOVERRIDES means that you can only prepend to it. The trick is that you want to prepend before inheriting other files doing the same otherwise your order is messed up (we messed it up twice here already)09:12
*** Bunio_FH <Bunio_FH!> has joined #yocto09:13
qschulzJPEW: also, MACHINEOVERRIDES is only a part of the overall OVERRIDES mechanism, and I don't know if it's used alone somewhere. So I'd have a look at the content of OVERRIDES and see if you have something with higher priority than your newly added MACHINEOVERRIDES (e.g. DISTROOVERRIDES has higher priority than MACHINEOVERRIDES ;) )09:14
*** vermaete <vermaete!> has quit IRC09:14
*** kroon <kroon!~kroon@> has joined #yocto09:15
*** Guest5287 <Guest5287!a5e14925@gateway/web/cgi-irc/> has joined #yocto09:28
Guest5287Hi, i want to create a custom ubi layout with ubinize. Is there a way to do this with a special image? I inherit from core-image.09:39
*** yann <yann!~yann@> has joined #yocto09:41
*** LocutusOfBorg <LocutusOfBorg!~LocutusOf@ubuntu/member/locutusofborg> has quit IRC09:41
LetoThe2ndGuest5287: if it involves seperate partitions, this sounds very much like something to look into with wic09:42
Guest5287i use only one MTD partition and multiple ubi volumes09:42
LetoThe2ndGuest5287: yeah but what do the volumes represent?09:42
qschulzGuest5287: ubi in IMAGE_TYPEFS and declare MKUBIFS_ARGS and UBINIZE_ARGS in your machine configruation file?09:43
Guest5287ubi and ubifs are already in my IMAGE_FSTYPES UBINIZE_ARGS is new to me...09:44
LetoThe2ndqschulz: another option, but i think it depends very much on the usecase, in the end.09:45
Guest5287Oh no, i have all of them already declared...09:45
Guest5287But this is only the layout of the flash?09:45
Guest5287LetoThe2nd ubi is managing them ?09:46
LetoThe2ndGuest5287: ?09:46
*** florian_kc is now known as florian09:46
Guest5287LetoThe2nd or what did you mean with your question?09:46
*** vermaete <vermaete!> has joined #yocto09:47
LetoThe2ndGuest5287: no, i meant what the ubi volumes are meant to repesenr. parts of the userland? or just additional partitions that are initially empty?09:47
Guest5287LetoThe2nd they represent userland, kernel and multiple DTBs. And that two times, beause of the update concept09:48
paulbarkerI just did a pull on poky master and rebuilt my image. I'm seeing lots of noise about unihash, is that now normal?09:48
LetoThe2ndpaulbarker: how does the noise sound if you pipe it to your stereo?09:49
*** gtristan <gtristan!~tristanva@> has quit IRC09:49
paulbarkerFor example: `Task virtual:native:/home/pbarker/Projects/Yocto/poky/meta/recipes-support/popt/ unihash changed to c6b6a3568c70e1d9ad48d85001273b1d52f83785ec8113265186cb86ac6b66cb`09:55
paulbarkerAlso `Already covered setscene for /home/pbarker/Projects/Yocto/poky/meta/recipes-core/images/ so ignoring rehash (remap)`09:56
paulbarkerAnd `Reported task /home/pbarker/Projects/Yocto/poky/meta/recipes-core/images/ as unihash 1af3abb29e5c8dec9ee74818bf4e8d3e3ce69e8313c331fa2e2613b3643a7d23 to u09:56
paulbarkernix:///b/Yocto/poky/build/hashserve.sock ({'taskhash': 'd67ee50cce5a58d71e17e129455d27286f360a7c8e0be0faf6cb4af2dff8ecfb', 'method': 'oe.sstatesig.OEOuthashBasic', 'unihash': '1af3abb29e5c809:56
paulbarkerThere's hundreds of lines of this output so it buries any other messages09:57
*** LocutusOfBorg <LocutusOfBorg!~LocutusOf@2001:b07:5d32:c012:d535:27a9:d5e0:e97d> has joined #yocto09:59
*** LocutusOfBorg <LocutusOfBorg!~LocutusOf@ubuntu/member/locutusofborg> has joined #yocto09:59
*** bluelightning <bluelightning!~paul@pdpc/supporter/professional/bluelightning> has quit IRC10:00
vermaeteas shameless promotion of my python manifest issue:
*** Guest5287 <Guest5287!a5e14925@gateway/web/cgi-irc/> has quit IRC10:09
warpme_guys: can somebody point me which WIC part (.py module) is responsible for writing grub core.img to disk image?10:17
paulbarkerwarpme_: It'll be one of the source plugins, likely `bootimg-*.py`10:20
paulbarkerIn `scripts/lib/wic/plugins/source/`10:21
warpme_for sure :-) but I scanned those files carefully multiple times and probably i'm blind as I can find it.... simply can't understand how wic is doing this stage in EFI bootable image preparation....10:24
paulbarkerwarpme_: What's your target MACHINE?10:30
*** emrius <emrius!> has joined #yocto10:30
emriusGood day everyone, I have a question regarding good practice when extending an existing layer (meta-oe) with a third party recipe ( Should I `cp` or `ln` the recipe's subdirectories content (`recipes-devtools`) into the meta-oe layer`s subdirectory `meta-python` because it's a python package (scipy) I'm10:38
emriusactually targeting? That seemed odd to me. Thus I created a dedicated layer `meta-scipy` to only contains to recipe of that github link. But this also feels a bit add, to have a layer with only one recipe... Any hint on where to put that recipe?10:38
warpme_target is x86_64....but maybe some context: i'm using part of Yocto (mainly wic) in my project ( So far with great success (bootable SD card images for allwinner/rockchip/amlogic/broadcom; x86 BIOS bootable live USB keys). Issue I have is adopting WID to produce EFI bootable images.  So far missing part is instaltion of GRUB part on already prepared image (EFI & root partitions seems10:38
warpme_to be OK; just ESP nor core.img seem to be missing).10:38
qschulzemrius: if you feel like oe-scipy should be part of meta-python or meta-oe, please send a patch to make it part of the master version. If you want to work with old branches of the layers or can't update them, you need a separate layer. Do NOT touch anything that isn't maintained by you is the rule of thumb10:39
emriusOk, that agrees with my guts :)10:40
* LetoThe2nd agrees with qschulz and therefore transitionally with emrius' guts.10:40
paulbarkerwarpme_: Ok. Which specific MACHINE are you using when you see the issue though?10:41
emriusOk, but for now, I would like to test that recipe. It contains patches for cmake and lapack. Is it best to throw it into it's own layer so that it is kinda isolated from everything else?10:43
qschulzemrius: bbappends for cmake and lapack and you're good10:43
qschulzin your own layer10:44
emriusqschulz: Thanks! I'll try that right away10:47
emriusBut wait.10:47
emriusI face a naming convention issue/question here.10:48
emriusthe repo is called oe-scipy. Is that a bad name in the first place? Should that rather be meta-scipy?10:49
emriusThis might seem pedantic but I just want to make sure I don't mess things up in the first place10:49
emriusoe-scipy then contains recipes-devtools10:50
emriusAnd I thought that it would be best to take the convention: `meta-mylayername/meta-mypackage/recipes-recipetype/content`10:51
emrius`meta-mypackge` would then coincide with `oe-scipy`. Am I overthinking things?10:52
LetoThe2ndemrius: you are a bit wrong indeed.10:52
LetoThe2ndemrius: the repo should be named meta-something.10:53
LetoThe2ndemrius: and the structure roughly this: meta-something/recipes-something/something/something_x.y.z.bb10:53
emriusthanks a lot for the clarification! That charms my guts yet again10:53
*** vermaete <vermaete!> has quit IRC11:04
*** Zajc <Zajc!> has quit IRC11:04
*** Zajc <Zajc!> has joined #yocto11:05
qschulzemrius: the layout you suggested is more for repos having multiple layers in them. I'd say it's rather rare to have such needs as YP users?11:07
warpme_paulbarker: I'm using WIC to create EFI bootable image for x86. For WIC log (and to see how far i'm with this) - pls see my msg in this IRC posted 6:33PM 8/03. There I put also .wks file i'm using to create image. Basically: wic ends without any error - but during inspecting created image I'm getting impression there is no ESP partition with grub core.img nor any log from wic about embedding grub core.img into11:13
warpme_MBR<->part_table gap in mage....11:13
warpme_paulbarker: I'm using WIC to create EFI bootable image for x86. For WIC log (and to see how far i'm with this) - pls see my msg in this IRC posted 6:33PM 8/03. There I put also .wks file i'm using to create image. Basically: wic ends without any error - but during inspecting created image I'm getting impression there is no ESP partition with grub core.img nor any log from wic about embedding grub core.img into11:13
warpme_MBR<->part_table gap in image....11:13
emriusqschulz: ah ok, you mean large projects like e.g. meta-openembedded, right?11:15
emriusIs there a good resource that wraps layouts and good practices up?11:16
*** LocutusOfBorg <LocutusOfBorg!~LocutusOf@ubuntu/member/locutusofborg> has quit IRC11:17
*** guerinoni <guerinoni!> has quit IRC11:19
qschulzemrius: #yocto on freenode :D11:19
emrius;D (y)11:20
qschulzemrius: the thing is, most of the time "it depends" (TM LetoThe2nd)11:21
*** ctlnwr_ <ctlnwr_!~catalin@> has quit IRC11:21
*** ctlnwr <ctlnwr!~catalin@> has quit IRC11:21
emriushmm okay. Not very surprising but not very satisfying, either =(11:22
qschulzemrius: but layout in a layer is bound to BBFILES in conf/layer.conf. The layout (if matched to BBFILES) does not really matter.11:22
qschulzit's not really used AFAIK?11:22
*** tgamblin <tgamblin!~tgamblin@> has joined #yocto11:27
*** yacar_ <yacar_!> has quit IRC11:35
*** rburton <rburton!rburton@nat/intel/x-cgyoktjdtuddmrkf> has joined #yocto11:35
*** LocutusOfBorg <LocutusOfBorg!~LocutusOf@ubuntu/member/locutusofborg> has joined #yocto11:38
*** vermaete <vermaete!> has joined #yocto11:42
*** khem <khem!~khem@unaffiliated/khem> has quit IRC11:48
*** berton <berton!~berton@> has joined #yocto11:52
*** NiksDev2 <NiksDev2!~NiksDev@> has quit IRC11:59
*** berton <berton!~berton@> has quit IRC12:00
*** berton <berton!~berton@> has joined #yocto12:01
*** khem <khem!~khem@unaffiliated/khem> has joined #yocto12:14
*** fitzsim <fitzsim!> has quit IRC12:15
*** mattis <mattis!> has joined #yocto12:16
mattisHi. I am trying to understand how to develop a program external to yocto which can run on my newly developed yocto build. Typically I would download the toolchain from arm e.g arm-none-eabi-g++ or similar. Can I build such a toolchain with yocto? I read some places, that its found in opt/<target>/sdk but I am not sure what that means exactly...12:20
*** Ilya_S <Ilya_S!bc788b40@> has joined #yocto12:24
LetoThe2ndmattis: i think #3 is on sdks and how to develop stuff :)12:25
LetoThe2ndmattis: the nutshell version: you can generate a sdk for a given image (which is the target to run on, then.)12:25
LetoThe2ndmattis: and that sdk traditionally installs to said /opt path. but if you don't have the image/app developer split, its way easier to just use the build directly (#10 or #11 on the videos, on app developemnt)12:27
mattisLetoThe2nd: Thanks I am watching it right now. So I assume this is going to be explained.12:29
LetoThe2ndmattis: well certainly not *EVERYTHING* is being explained, but certainly enough to get started.12:29
Ilya_SHey, I am trying to make an SDK based on my image that will be used for cross compiling. One of my recipes downloads some files from git and I want them to be included in the SDK. Currently the SDK has some of the header files from that recipe, but lack other headers and several source files that I need for cross compilation. I tried to search for12:33
Ilya_Sa way to do this but nothing worked for now. Anyone know a way to do this?12:33
LetoThe2ndIlya_S: depends (TM)12:33
LetoThe2ndIlya_S: this can either be a badly packaged thing, something that doesn't have a proper DEPENDS on it, or something that is compile-time only. in the latter case you probably want to look at TOOLCHAIN_TASK_HOST (or pretty similar, i can't ever get the var name right)12:34
Ilya_Sessentially I just want to know if there is a specific folder that I could copy these files and the SDK build will just grab them along12:34
Ilya_Stried TOOLCHAIN_TARGET_TASK_append12:35
LetoThe2ndno. it does not work like "just grab and copy". it works like "package properly and get your dependencies sorted out"12:35
Ilya_STOOLCHAIN_TARGET_TASK_append = " osdk-staticdev"12:35
LetoThe2ndIlya_S: then maybe that thing does not contain what you acutally want/need. look at the packages.12:35
LetoThe2ndplus, shouldn't it be TOOLCHAIN_HOST_TASK?12:36
LetoThe2ndi mean you want the things to run on the host? but i might be mistaken.12:36
*** yacar_ <yacar_!> has joined #yocto12:36
Ilya_SI think these files are not part of the compilation. They are just there as sample code with helper functions, so they won't be in any depends12:36
LetoThe2ndIlya_S: like i said then, fix the packaging.12:37
LetoThe2ndmake a whatever-examples package out of them, or something similar.12:37
Ilya_SI'm not sure I understand what you mean by "packaging"?12:38
*** RobertBerger <RobertBerger!> has joined #yocto12:39
LetoThe2ndIlya_S: a recipe does generate a number of packages. ${PN}, ${PN}-dev, ${PN}-dbg... so if you think the recipe should provide somthing, then it has to go into a pacakge.12:39
LetoThe2ndIlya_S: one of the live coding sessions covers this, i think its called depends, redepends, packaging or something close. IIRC, #512:40
Ilya_SAh, yeah, I tried to do that too. Are all these ${PN} included in the SDK?12:40
LetoThe2ndthose which you either depend on or explicitly ask for are included.12:40
Ilya_SHow can I ask them to be included? If it's covered in the coding session, I can try to check that out12:41
LetoThe2ndIlya_S: again, included means "TOOLCJHAIN_HOST_TASK"12:42
Ilya_SAh, I see, thanks. I'll try to figure it out12:43
*** hamis_lt_u <hamis_lt_u!> has quit IRC12:48
fullstopI'm trying to have Yocto generate my ubi, but my needs are different than that of the ubi / ubifs image class.  I attempted to make my own image class but can't seem to get bitbake to actually make the thing.12:52
*** hamis <hamis!~irfan@> has joined #yocto12:52
fullstopI added IMAGE_CLASSES += " myimagetype " and the bbclass is included in the layer configuration.12:53
*** gtristan <gtristan!~tristanva@> has joined #yocto12:57
*** pharaon2502 <pharaon2502!> has quit IRC12:59
*** pharaon2502 <pharaon2502!> has joined #yocto13:00
*** hyper_dave <hyper_dave!~quassel@> has quit IRC13:00
*** yacar_ <yacar_!> has quit IRC13:02
*** emrius <emrius!> has quit IRC13:06
qschulzfullstop: aren't you more intestered in IMAGE_CMD_ instead? Add your own IMAGE_TYPES and define an IMAGE_CMD_mytype for it?13:07
qschulz(see image_types.bbclass)13:07
fullstopqschulz: that is in my bbclass13:07
qschulzfullstop: missing IMAGE_FSTYPES in your image recipe?13:09
fullstopperhaps, let me look once this build fails13:10
fullstopqschulz: where is the best place to put IMAGE_FSTYPES?  local.conf?  distro.conf?13:10
qschulzfullstop: I can see that's it's both in image recipes and in machine conf files... So... /me shrugs?13:11
qschulzfullstop: there are types specific to an image and other specific to a machine, your call13:12
*** JaMa <JaMa!~martin@> has joined #yocto13:12
fullstopI understand13:12
* fullstop moves this stuff to machine conf13:13
*** pharaon2502 <pharaon2502!> has quit IRC13:17
*** locutus_ <locutus_!~LocutusOf@2001:b07:5d32:c012:d535:27a9:d5e0:e97d> has joined #yocto13:18
*** vermaete <vermaete!> has quit IRC13:19
*** LocutusOfBorg <LocutusOfBorg!~LocutusOf@ubuntu/member/locutusofborg> has quit IRC13:20
*** pharaon2502 <pharaon2502!> has joined #yocto13:21
*** stephano <stephano!> has joined #yocto13:22
fullstophey, it's doing something.  Thanks, qschulz!13:24
qschulzfullstop: nice :)13:26
*** ebail <ebail!> has joined #yocto13:29
*** leon-anavi <leon-anavi!~Leon@> has quit IRC13:36
*** rcw <rcw!~rcw@> has joined #yocto13:36
*** leon-anavi <leon-anavi!~Leon@> has joined #yocto13:36
yoctiNew news from stackoverflow: systemd: Freezing execution in Yocto <>13:42
*** Scoutboy <Scoutboy!~quassel@> has joined #yocto13:49
*** Scoutboy <Scoutboy!~quassel@> has quit IRC13:54
*** Scoutboy <Scoutboy!~quassel@> has joined #yocto13:54
*** mattis <mattis!> has quit IRC13:55
*** RobertBerger1 <RobertBerger1!> has joined #yocto13:58
*** RobertBerger <RobertBerger!> has quit IRC13:59
*** hpsy <hpsy!~hpsy@> has joined #yocto14:01
*** ericch <ericch!> has joined #yocto14:03
*** vermaete <vermaete!> has joined #yocto14:05
rburtonokay feeling dumb14:09
rburton(no smart jokes LetoThe2nd)14:09
rburtonso if i run a binary from a devshell from the native sysroot i get cannot open shared object file: no such file or directory14:10
rburtonbut if I ldd the binary inside the devshell, it resolves to /usr/lib64/
LetoThe2ndrburton: is it the multilib stuff hitting hard again?14:10
*** ssajal <ssajal!~ssajal@> has joined #yocto14:10
mihaiwhich ldd?14:10
rburtonstruggling to understand how ldd says something different considering its just a wrapper around the loader14:11
RPrburton: does it pick the same loader?14:11
* RP doesn't trust ldd14:11
LetoThe2ndrburton: there was a couple of things related lately. wrong loader?14:11
LetoThe2ndRP was faster.14:11
rburton        /scratch/poky/sysroots-uninative/x86_64-linux/lib/ => /usr/lib64/
rburtonright if i use the sysroot ld directly i get the error14:15
rburtonso for some reason the uninative isn't even attempting to look in the host paths14:18
vermaeteCould somebody shine a bit of light into how the python manifest works?14:19
rburtonvermaete: what bit of it?14:20
*** RobertBerger <RobertBerger!> has joined #yocto14:20
vermaeteI've *just* recreated it without modify anything in it.  And it looks like the python3-pkgutil is not in the python3-core.14:20
vermaeteWhat's probably not correct.14:20
vermaeteI have this at Zeus and master.14:20
*** RobertBerger1 <RobertBerger1!> has quit IRC14:21
vermaeteBecause of this, my rootfs is not building anymore.14:21
rburtonmaybe that class doesn't get explicitly mentioned in the dependency trees, might need special casing14:22
vermaeteThis because the few recipes that are using python3-pkgutil are not able to include in the image.14:22
vermaete@rburton: fine, but every release this manifest if probably regenerated, not?14:22
vermaeteWhy is it different at my side when re-generating it?14:23
vermaeteBecause I'm using different layers?14:23
rburtonvermaete: only regenerated when needed14:24
rburtonand the generated onemight be fiddled by hand further14:24
vermaeteokay, I've got this.14:24
rburtonso LD_DEBUG=all shows that it only looks in my sysroot and not on the host14:25
*** sstiller <sstiller!> has quit IRC14:27
*** nopeman <nopeman!d54413ba@> has quit IRC14:29
*** hamis <hamis!~irfan@> has quit IRC14:29
*** mattis <mattis!> has joined #yocto14:30
mattisNot sure what is happening, bitbake core-image-minimal works, but bitbake core-image-minimal -c populate_sdk fails, should this not work for the same reason as the original?14:31
mattisERROR: Task (/home/mattis/development/cgm_yocto/build/build/poky/meta/recipes-devtools/gdb/ failed with exit code '1'14:32
paulbarkermattis: gdb-cross-canadian is only needed for the SDK so that wouldn't cause the image build to fail14:32
paulbarkerCan you post the full bitbake output via a pastebin?14:33
RPrburton: I can't remember how we designed it to work :/14:36
rburtonme neither!14:36
mattispaulbarker: Yes but its the sdk I want to build and this is failing.14:40
paulbarkermattis: That is a strange one. I'd recommend dropping meta-freescale temporarily and seeing if you can build gdb-cross-canadian on its own for a similar machine (maybe qemuarm). That should show whether the issue is in the Freescale BSP or in oe-core14:45
Crofton|roadneed to run the yocto check layer script against meta-freescale14:51
*** rob_w <rob_w!~bob@unaffiliated/rob-w/x-1112029> has quit IRC14:52
smurraymeta-freescale has some oddness around packaging wrt tuning14:53
*** dreyna <dreyna!> has joined #yocto14:54
mattispaulbarker: ok. I had some issues earlier with some out of memory on hdd, since I am using docker, I will first do a clean install.14:54
smurrayCrofton|road: it has a significant number of bbappends in dynamic-layers on top of the ones in the layers, I'm pretty sure it violates the yocto layer reqs14:54
smurrayCrofton|road: it even has a dynamic bbappend for AGL that we have to BBMASK out, as it's wrong :/14:55
Crofton|roaddamn, no wonder BSP's make people angry14:56
fraynot sure it's in the YP guidelines, but any package 'tweaked' by a BSP has to be declared machine specific.. what a mess14:57
mattissmurray: and others did I do something wrong?14:59
*** AndersD <AndersD!> has joined #yocto14:59
armpiteverything make me angry : )14:59
*** AndersD <AndersD!> has quit IRC14:59
smurraymattis: no, you might just be stumbling upon someone else's brokenness14:59
smurraymattis: I'd recommend trying what paulbarker suggested, try building the SDK without meta-freescale in the mix, to hopefully narrow down where the problem lies15:01
mattissmurray: I will, just starting off fresh now.15:01
vermaeteIf you have a small moment.15:02
vermaeteThat manifest file, is this something that is input to generated some output.  And that output is imported in git.  Or is this parsed and used by bitbake itself.15:03
mattisvermaete: Are you asking me?15:04
vermaetesorry, was for rburton.  If he has time.15:05
*** WillMiles <WillMiles!~Will@> has joined #yocto15:21
*** guerinoni <guerinoni!> has joined #yocto15:23
*** yacar_ <yacar_!> has joined #yocto15:27
*** jobroe <jobroe!~manjaro-u@> has quit IRC15:34
rburtonvermaete: the manifest is used to build the package and dependency lists15:35
*** vermaete <vermaete!> has quit IRC15:43
*** mckoan is now known as mckoan|away15:56
*** goliath <goliath!~goliath@> has quit IRC15:57
*** vineela <vineela!vtummala@nat/intel/x-aqdfpimmxssrgyla> has joined #yocto15:58
*** TobSnyder <TobSnyder!> has quit IRC16:17
*** RobertBerger <RobertBerger!> has left #yocto16:21
JPEWWhich mailing list do I use for doc updates to bitbake?16:27
kroonJPEW, bitbake-devel last time I checked16:30
kroonwith a suggested "bitbake-user-manual:" heading16:31
*** kroon <kroon!~kroon@> has quit IRC16:32
*** goliath <goliath!> has joined #yocto16:37
*** panfir <panfir!~crs@> has left #yocto16:42
yoctiNew news from stackoverflow: systemd: Freezing execution in Yocto [closed] <>16:43
*** fl0v0 <fl0v0!> has quit IRC16:44
*** yacar_ <yacar_!> has quit IRC16:51
*** sno <sno!~sno@> has quit IRC16:52
*** mihai <mihai!~mihai@unaffiliated/mihai> has quit IRC16:56
*** robbawebba <robbawebba!~rob@> has joined #yocto17:06
*** bluelightning <bluelightning!~paul@pdpc/supporter/professional/bluelightning> has joined #yocto17:06
*** yann <yann!~yann@> has quit IRC17:09
*** Bunio_FH <Bunio_FH!> has quit IRC17:12
*** ericch <ericch!> has quit IRC17:13
*** ericch <ericch!> has joined #yocto17:14
*** m1ster_r0b0t <m1ster_r0b0t!> has joined #yocto17:22
*** goliath <goliath!> has quit IRC17:28
*** florian <florian!~florian_k@Maemo/community/contributor/florian> has quit IRC17:32
*** nerdboy <nerdboy!~sarnold@> has joined #yocto17:42
*** nerdboy <nerdboy!~sarnold@gentoo/developer/nerdboy> has joined #yocto17:43
*** kiwi_29 <kiwi_29!> has joined #yocto17:49
*** Bunio_FH <Bunio_FH!> has joined #yocto17:53
kiwi_29hello ... I have a custom distro which installs many applications using IMAGE_INSTALL .17:56
kiwi_29When I run. "bitbake customdistro -c populate_sdk". I get error like this17:58
kiwi_29do_populate_sdk: Unable to install packages. Command   <PATH>/0-r0/recipe-sysroot-native/usr/bin/apt-get  install --force-yes --allow-unauthenticated  <LIST OF CUSTOM APPS>  target-sdk-provides-dummy <MORE LIST OF CUSTOM APPS>18:01
kiwi_29Reading package lists...Building dependency tree...Reading state information...Some packages could not be installed. This may mean that you haverequested an impossible situation or if you are using the unstabledistribution that some required packages have not yet been createdor been moved out of Incoming.The following information may help to18:01
kiwi_29resolve the situation:The following packages have unmet dependencies: target-sdk-provides-dummy : Conflicts: coreutilsE: Unable to correct problems, you have held broken packages.18:01
*** goliath <goliath!> has joined #yocto18:01
paulbarkerkiwi_29: Could you post the full bitbake output via a pastebin?18:02
*** stephano <stephano!> has quit IRC18:03
*** stephano <stephano!> has joined #yocto18:03
*** Bunio_FH <Bunio_FH!> has quit IRC18:05
*** rcw <rcw!~rcw@> has quit IRC18:07
kiwi_29Loading cache: 100% |#################################################################################################################################################################################| Time: 0:00:00Loaded 3357 entries from dependency cache.NOTE: Resolving any missing task queue dependenciesBuild Configuration:BB_VERSION           =18:07
kiwi_29"1.44.0"BUILD_SYS            = "x86_64-linux"NATIVELSBSTRING      = "universal"TARGET_SYS           = "x86_64-poky-linux"MACHINE              = "genericx86-64"DISTRO               = "customdistro"DISTRO_VERSION       = "0.0.1"TUNE_FEATURES        = "m64 core2"TARGET_FPU           = ""meta                 meta-poky            meta-yocto-bsp       =18:07
kiwi_29"cpe-yocto3.0-poky:94f6b31befda5c496f65e863a6f8152b42d7ebf0"meta-customdistro          = "wncagent:dc4822f9484d1ab8b66c7b25164f8ace4a40046b"meta-oe              meta-python          meta-networking      = "zeus:e855ecc6d35677e79780adc57b2552213c995731"Initialising tasks: 100%18:08
kiwi_29|############################################################################################################################################################################| Time: 0:00:02Sstate summary: Wanted 184 Found 183 Missed 1 Current 949 (99% match, 99% complete)NOTE: Executing TasksNOTE: Setscene tasks completedERROR:18:08
kiwi_29customdistro-distro-1.0-r0 do_populate_sdk: Unable to install packages. Command '/home/kiwiuser/projects/targethardware/src/customdistrodistro/poky/build/tmp/work/genericx86_64-poky-linux/customdistro-distro/1.0-r0/recipe-sysroot-native/usr/bin/apt-get  install --force-yes --allow-unauthenticated <LIST OF PACKAGES, one of the package is coreutils>18:08
kiwi_29target-sdk-provides-dummy <MORE LIST OF PACKAGES>' returned 100:Reading package lists...Building dependency tree...Reading state information...Some packages could not be installed. This may mean that you haverequested an impossible situation or if you are using the unstabledistribution that some required packages have not yet been createdor been18:08
kiwi_29moved out of Incoming.The following information may help to resolve the situation:The following packages have unmet dependencies: target-sdk-provides-dummy : Conflicts: coreutilsE: Unable to correct problems, you have held broken packages.ERROR: Logfile of failure stored in:18:08
kiwi_29/home/kiwiuser/projects/targethardware/src/customdistrodistro/poky/build/tmp/work/genericx86_64-poky-linux/customdistro-distro/1.0-r0/temp/log.do_populate_sdk.19581ERROR: Task (/home/kiwiuser/projects/targethardware/src/customdistrodistro/poky/meta-customdistro/recipes-core/images/ failed with exit code18:08
kiwi_29'1'NOTE: Tasks Summary: Attempted 3941 tasks of which 3940 didn't need to be rerun and 1 failed.Summary: 1 task failed:  /home/kiwiuser/projects/targethardware/src/customdistrodistro/poky/meta-customdistro/recipes-core/images/ There was 1 ERROR message shown, returning a non-zero exit code.18:08
*** rcw <rcw!~rcw@> has joined #yocto18:08
kiwi_29I wish the indentation was good18:08
kiwi_29one of the packages in the list is coreutils18:08
kiwi_29when I do "bitbake core-image-minimal -c populate_sdk" does not throw any error18:09
paulbarkerkiwi_29: Use a pastebin please18:10
*** stephano <stephano!> has quit IRC18:10
*** kiwi_29 <kiwi_29!> has quit IRC18:12
paulbarkerkiwi_29: Are you using deb packages?18:14
*** stephano <stephano!> has joined #yocto18:15
*** kiwi_29 <kiwi_29!> has joined #yocto18:15
paulbarkerkiwi_29: Are you using deb packages?18:15
*** locutus_ <locutus_!~LocutusOf@2001:b07:5d32:c012:d535:27a9:d5e0:e97d> has quit IRC18:15
kiwi_29paulbarker yes18:16
kiwi_29when I removed coreutils from list inside "IMAGE_INSTALL"18:16
paulbarkerkiwi_29: I'd recommend switching to ipk or rpm as they're much better tested and supported18:16
kiwi_29The bitbake command succeeds18:16
kiwi_29The bitbake command is "bitbake customdistro -c populate_sdk"18:17
paulbarkerkiwi_29: Ok. Can you try this with ipk or rpm packages and see if that works?18:18
paulbarkerWith coreutils still included18:18
*** florian <florian!~florian_k@Maemo/community/contributor/florian> has joined #yocto18:18
kiwi_29populate_sdk seems to injecting target-sdk-provides-dummy which conflicts with coreutils18:18
kiwi_29I have to use deb unfortunately for some time18:19
paulbarkerkiwi_29: At least try it without so you can narrow down where the issue is18:19
paulbarkerIf the bug report is "works with ipk but not deb" we've got a much smaller area to search for the underlying problem18:20
kiwi_29true... also    it mentions to go away from deb18:20
*** crs_ <crs_!~crs@> has joined #yocto18:21
*** kiwi_29 <kiwi_29!> has quit IRC18:23
*** Ilya_S <Ilya_S!bc788b40@> has quit IRC18:25
*** kiwi_29 <kiwi_29!> has joined #yocto18:28
kiwi_29btw paulbarker. ... I wanted to install all the development tools gcc, ldd etc on target system and my understanding is that  "-c populate_sdk" accomplishes that ..please let me know if my understanding is correct18:29
kiwi_29I basically want to debug an application produced as part of yocto build. I want to debug on the target18:29
kiwi_29so I want gcc, gdb, ldd, readelf etc tools on target machine.18:30
kiwi_29is "-c populate_sdk" the right way of doing it18:30
paulbarkerkiwi_29: No, that creates an SDK for use on the host. If you want those on the target just add them to your image18:31
kiwi_29I populate_sdk creates an sdk for cross-compilation environment where development/compilation and even debugging happens on host18:33
kiwi_29Is there a single recipe that provides all these devtools so that it compiles as part of target18:35
*** sno <sno!> has joined #yocto18:35
*** roussinm <roussinm!> has quit IRC18:36
*** roussinm <roussinm!> has joined #yocto18:37
paulbarkerkiwi_29: Maybe a packagegroup but I can't recall the name18:37
roussinmI have an image that is packages inside a ``, but some windows client, when unzipping can't "see" the `wic` when using a tool to burn the image to an sdcard. I would like to intervene after the image creation but before the zip task. I would like to rename to .wic image to an .img so that the window tool can "see" the file, because the combobox has "*.img"18:41
*** sno <sno!> has quit IRC18:43
roussinmI think what I'm trying to accomplish looks like : addtask makezip before do_build after do_image_complete. Inside the makezip task I would rename the file and execute the CONVERSION_CMD_zip.18:44
*** sno <sno!> has joined #yocto18:44
*** sj52 <sj52!> has quit IRC18:52
*** sno <sno!> has quit IRC18:52
*** crs_ <crs_!~crs@> has quit IRC18:53
*** tgamblin <tgamblin!~tgamblin@> has quit IRC18:53
*** guerinoni <guerinoni!> has quit IRC18:54
*** locutus_ <locutus_!~LocutusOf@> has joined #yocto18:55
JaMawget in ubuntu-20.04 doesn't like the ssl on
JaMaConnecting to (||:443... connected.18:59
JaMaOpenSSL: error:1425F102:SSL routines:ssl_choose_client_version:unsupported protocol18:59
frayit's a Let'sEncrypt cert.. so maybe affected by the recent problems there?19:01
fraybut even so, both Safari and Chrome load it fine fo rme19:01
neverpanicFirefox also complains it's using weak crypto19:03
neverpanicECDHE_RSA_AES_256_CBC_SHA. Not exactly the most modern cipher suite, and not AEAD.19:04
*** rcw <rcw!~rcw@> has quit IRC19:04
*** rcw <rcw!~rcw@> has joined #yocto19:04
*** florian <florian!~florian_k@Maemo/community/contributor/florian> has quit IRC19:06
*** florian <florian!~florian_k@Maemo/community/contributor/florian> has joined #yocto19:08
*** ferry <ferry!> has joined #yocto19:09
*** rcw <rcw!~rcw@> has quit IRC19:10
*** rcw <rcw!~rcw@> has joined #yocto19:12
kiwi_29I get "No such file or directory" when I run an application on yocto target19:17
kiwi_29the application is present on the path , has execute permission for user19:17
neverpanickiwi_29: /lib64/ --list $binary, I bet a library is missing.19:19
neverpanic(adjust for your architecture, of course)19:19
kiwi_29this is the output19:22
*** locutus_ <locutus_!~LocutusOf@> has quit IRC19:22
*** ferry <ferry!> has quit IRC19:23
neverpanicseems fine.19:25
mcfriskkiwi_29: runs strace of the binary. maybe a config file, dlopen() or something else is failing. The file path from strace output will give hints.19:27
*** aidanh_ <aidanh_!~aidanh@unaffiliated/aidanh> has joined #yocto19:29
*** aidanh <aidanh!~aidanh@unaffiliated/aidanh> has quit IRC19:30
*** aidanh_ is now known as aidanh19:30
*** kiwi_29 <kiwi_29!> has quit IRC19:39
*** PinkSnake <PinkSnake!51ff1123@> has quit IRC19:45
*** kanavin_home <kanavin_home!~ak@2a02:2450:1011:56f:238d:84be:b349:9184> has quit IRC19:48
*** tgamblin <tgamblin!> has joined #yocto19:48
*** gsalazar <gsalazar!> has joined #yocto19:50
*** kiwi_29 <kiwi_29!> has joined #yocto19:56
kiwi_29mcfrisk neverpanic this is a go binary ..just fyi19:56
*** kanavin_home <kanavin_home!~ak@2a02:2450:1011:56f:238d:84be:b349:9184> has joined #yocto19:58
LetoThe2ndkiwi_29: let me guess. you did not crosscompile this through bitbake?19:58
LetoThe2ndkiwi_29: its just another case of using the wrong loader.19:58
LetoThe2ndfor reasons this is the new cool s**t.19:59
*** LocutusOfBorg <LocutusOfBorg!~LocutusOf@> has joined #yocto20:04
*** LocutusOfBorg <LocutusOfBorg!~LocutusOf@ubuntu/member/locutusofborg> has joined #yocto20:04
*** gsalazar <gsalazar!> has quit IRC20:08
kiwi_29LetoThe2nd. the binary is generated as part of the Yocto build . Yocto generates the distro and also compiles this binary. Host is x86-64 (Xeon).  and target is x86-64(celeron)20:11
LetoThe2ndkiwi_29: ah. well there was something really similar lately which ultimately boiled down to multilib. don't remember for sure if it was also host contamination involved, but thats the way to look.20:13
kiwi_29how do I get multilib support on target?20:14
kiwi_29btw ..output of strace
kiwi_29mcfrisk neverpanic20:15
LetoThe2ndkiwi_29: i've poked the person who can share enlightenment. i'm personally not in touch with the topic, sorry.20:16
kanavin_homekiwi_29: what does 'file /usr/bin/yourapp' say?20:19
*** guerinoni <guerinoni!> has joined #yocto20:24
*** kiwi_29 <kiwi_29!> has quit IRC20:26
*** kiwi_29 <kiwi_29!> has joined #yocto20:28
kiwi_29 ELF 64-bit LSB executable, x86-64, version 1 (SYSV), dynamically linked, interpreter /lib64/, Go BuildID=Z5k_gYnNQnexc0-pOqvx/l4ovWnRBgEF7GKYjAxmT/cU6SkJFIAj-Opt94rLKw/vXNKfIKTs_g7zuGJayny, not stripped20:28
kanavin_homekiwi_29: does  /lib64/ exist?20:29
kiwi_29no... the output of dependecies is.
kanavin_homekiwi_29: what language is the app written in?20:34
kanavin_home(e.g. what was used to compile it)20:34
khemkiwi_29: it seems golang compiler is not aware of ldso path20:34
khemkiwi_29: it seems a bug to me in OE, we should be aligning the ldso paths across compilers be it c/c++ or rust or golang20:35
kiwi_29I see..however it is found as we see        no ?20:35
khemis that a cmd you ran manually ?20:36
kanavin_homekiwi_29: the interpreter is the one you see in the 'file' output, and if it's not there, exec() will immediately fail with a file not found error20:36
kanavin_homewhat libs is the binary linked with is irrelevant then20:36
kanavin_homecan you show the recipe?20:37
khemits not recipe problem,20:37
kanavin_homeusually this happens when ppl transfer something they compiled in ubuntu, and expect it to 'work'20:38
khemkiwi_29: on your build host, go to builddir of the app and find the binary and run readelf -e on it and it should show you what interpreter is encoded into the binary, if that exact path is missing on target then we know the proble20:38
kanavin_homekhem: see above - file says 'ELF 64-bit LSB executable, x86-64, version 1 (SYSV), dynamically linked, interpreter /lib64/, Go BuildID=Z5k_gYnNQnexc0-pOqvx/l4ovWnRBgEF7GKYjAxmT/cU6SkJFIAj-Opt94rLKw/vXNKfIKTs_g7zuGJayny, not stripped'20:39
kiwi_29here is recipe khem kanavin_home.
khemkiwi_29: right so does /lib64/ exist on target ?20:43
kiwi_29khem. no20:43
kiwi_29how do I get it?20:43
khemright so thats the problem20:43
khemwe have to teach the go compiler to use right paths that match OEs norms20:43
kiwi_29my custom distro is based on core-image-minimal20:43
kiwi_29should I base it on core-image-multilib20:44
khemthats ok, you are doing all well20:44
kanavin_homekiwi_29: I think your recipe might be using the host go compiler20:44
kanavin_homethis 'go this' 'go that' stuff looks wrong20:44
kiwi_29I have used inherit go in beginning of recipe20:44
kiwi_29without that go command does not work20:44
*** peacememories <peacememories!~textual@2a02:8388:8982:e900:cc71:2343:725c:153> has joined #yocto20:45
kanavin_homeyeah, but the correct form is ${GO} , not go20:45
kiwi_29I see20:45
kanavin_homeand the inherit go should set up the compilation already, no need to overwrite it20:45
*** peacememories <peacememories!~textual@2a02:8388:8982:e900:cc71:2343:725c:153> has quit IRC20:46
khemyeah you should defer as much work as possible to templates provides in go.bbclass20:46
khemso cross compiling can work as expected20:46
kiwi_29so what should go inside do_compile ?20:47
khem${GO} build ABCagent.go20:47
kiwi_29trying it out right now...will update you in a moment20:48
kanavin_homekiwi_29: nothing20:48
*** mauz555 <mauz555!~mauz555@2a01:e0a:56d:9090:3092:7169:2a22:9e8d> has joined #yocto20:48
khemkiwi_29: Infact your recipe could be cleaned up please look at recipes-extended/go-examples/go-helloworld_0.1.bb20:50
kiwi_29kanavin_home  I see your point... how do I make sure I m compiline MYAPP.go.  which variable I use in beginning of recipe to specify that it is the20:50
kheme.g.  you can define GO_IMPORT and GO_INSTALL20:50
khemand thats pretty much you would need besides boiler plate code20:50
* RP finally finds the locale relocation bug in glibc20:52
*** sno <sno!> has joined #yocto20:53
*** LocutusOfBorg <LocutusOfBorg!~LocutusOf@ubuntu/member/locutusofborg> has quit IRC20:54
*** kiwi_29 <kiwi_29!> has quit IRC20:55
*** kriive <kriive!> has joined #yocto21:00
JaMaRP: memory check for you, it's about commit from 2011/2009 :) do you know why we have 2 different CACHE settings in default setup?21:02
JaMameta/conf/bitbake.conf:CACHE = "${TMPDIR}/cache${@['', '/' + str('MACHINE', d, 1))][bool('MACHINE', d, 1))]}${@['', '/' + str('SDKMACHINE', d, 1))][bool('SDKMACHINE', d, 1))]}"21:02
JaMameta/conf/distro/defaultsetup.conf:CACHE = "${TMPDIR}/cache/${TCMODE}-${TCLIBC}${@['', '/' + str('MACHINE', d, 1))][bool('MACHINE', d, 1))]}${@['', '/' + str('SDKMACHINE', d, 1))][bool('SDKMACHINE', d, 1))]}"21:02
JaMa2nd was introduced in 2011 with
JaMaany objection to update the 1st one and drop 2nd?21:03
JaMa2009 was last "functional" change in the 1st one (other than getVar API change updates later twice)
RPJaMa: presumably its because TCMODE and TCLIBC are nodistro specific21:04
*** yann <yann!> has joined #yocto21:05
RPJaMa: we've had wider adoption of that convention since21:05
RPJaMa: if someone had a summy nodistro conf file, it would break with TCMODE/TCLIBC not in bitbake.conf21:05
RPJaMa: no object to simplifying this though21:06
JaMaok, will do that in 3.221:06
JaMaTCLIBC is used in quite a few places now, TCMODE only in meta/conf/local.conf.sample.extended + defaultsetup.conf21:07
RPJaMa: right, TCLIBC is pretty much required now21:07
*** kiwi_29 <kiwi_29!> has joined #yocto21:07
*** WillMiles <WillMiles!~Will@> has quit IRC21:09
RPtgamblin: you've seen my coreutils hack in master-next right?21:12
kiwi_29khem kanavin_home. ..thanks .. I m working on it and will update21:12
kiwi_29will ping back with results21:13
*** rcw <rcw!~rcw@> has quit IRC21:14
*** pharaon2502 <pharaon2502!> has quit IRC21:16
*** rcw <rcw!~rcw@> has joined #yocto21:17
*** berton <berton!~berton@> has quit IRC21:21
*** kiwi_29 <kiwi_29!> has quit IRC21:21
*** sheelba <sheelba!> has joined #yocto21:25
*** frsc <frsc!> has quit IRC21:26
*** ningauble <ningauble!> has quit IRC21:27
*** rcw <rcw!~rcw@> has quit IRC21:29
*** mattis <mattis!> has quit IRC21:35
*** pohly <pohly!> has quit IRC21:35
*** kanavin_home <kanavin_home!~ak@2a02:2450:1011:56f:238d:84be:b349:9184> has quit IRC21:43
*** kanavin_home <kanavin_home!~ak@2a02:2450:1011:56f:238d:84be:b349:9184> has joined #yocto21:44
*** guerinoni <guerinoni!> has quit IRC21:55
*** rcw <rcw!> has joined #yocto21:55
RPJaMa: its tempting to consider removing nodistro entirely and mandate setting MACHINE/SDKMACHINE, simplify the code21:56
RPI think reasons for now setting those are now historical21:56
RPJPEW: with the new logging, does it capture hashequiv to the on disk logs by default?21:57
JaMaRP: agreed, e.g. meta-updater settings also expects MACHINE to be set and causes more ugly parsing error when it isn't, having more simple base and error early enough when not set would be nicer22:00
tgamblinRP: I see it now22:04
*** sheelba <sheelba!> has quit IRC22:06
neverpanickiwi_29 could try /lib/ /usr/bin/APPNAME, but I guess at this point it's pretty much established that the loader path is the problem.22:07
RPtgamblin: it gets us part way towards merging22:08
RPtgamblin: gcc-plugin reproducibility issue remains22:08
*** kriive <kriive!> has quit IRC22:09
tgamblinhmm ok22:13
xyzzy42I've got a recipe, "A", that wants files to be owned by a user account created by recipe "B".  B works, the account appears in its sysroot and the image.  "A" DEPENDS on B, but the sysroot for A doesn't have the accounts.  It does have other files provided by B.22:15
*** rcw <rcw!> has quit IRC22:16
RPtgamblin: just be glad it resolves the multilib one ;-)22:17
*** stephano <stephano!> has quit IRC22:19
xyzzy42I see there was a bug related to this, but it looks like it was fixed long ago.  I'm on zeus.
tgamblinRP: Whatever works I suppose. The nuclear option would've been to just remove that assert test22:20
RPtgamblin: is there just one test using that module?22:21
RPtgamblin: that would solve the problems and let us merge if it were disabled22:21
tgamblinRP: it's a single test that's failing on qemux86, from what I can tell (and from what Alex was saying)22:22
tgamblinI did a few test runs earlier today but got sidetracked from actually writing a patch to nuke that particular test case22:22
RPtgamblin: oh, there is that and the second issue of gcc being added to the build chain and gcc-plugin not being reproducible22:22
tgamblinAh. Hmm, let me have a look22:23
RPgcc comes from the libperl-module-build dependency22:23
RPtgamblin: there was a third problem of gcc multilib conflicts but I fixed that22:23
tgamblinRP: Ah, right, if I remember correctly, removing libmodule-build-perl causes 9ish test cases to fail22:24
* tgamblin starts a build to confirm the number22:25
RPtgamblin: right and they'd not fail deterministicly either as it would depend on image contents22:25
RPtgamblin: libperl-module-build has ptests so is installed in ptest images so wouldn't get "seen" by our autobuilder22:25
kanavin_hometgamblin: RP: the test fails on both arm and x86 - ['tests/tail-2/']22:26
kanavin_homeit has to be core-image-sato-sdk-ptest though if you want to reproduce22:26
RPkanavin_home: it fails on our autobuilder too22:31
RPat least its consistent22:31
tgamblinMy bad, I'm adding a lot of extra work for the release :/22:32
*** kanavin_home <kanavin_home!~ak@2a02:2450:1011:56f:238d:84be:b349:9184> has quit IRC22:33
RPtgamblin: its not the only thing holding things up22:33
*** kanavin_home <kanavin_home!> has joined #yocto22:35
neverpanicxyzzy42: Ah, yes, the dreaded user management thing. The relevant code is around here:
neverpanicI suggest you always choose statically configured uids and gids when re-using users across recipes, because otherwise you may get different uids and gids depending on when tasks are re-run, and you may get different uids when creating packages and at install time on target22:39
*** adelcast <adelcast!~adelcast@> has quit IRC22:39
tgamblinRP: kanavin_home: make that 63 tests failing without libmodule-build-perl - I had that number mixed up with my older total for when I was mulling over adding/not adding valgrind + strace22:39
RPtgamblin: ok, so we do need it, thanks22:40
*** JaMa <JaMa!~martin@> has quit IRC22:40
neverpanicat $work, we've avoided dependencies between recipes for users to prevent triggering rebuilds of everything when we were adding or changing users, and instead just had an include file that contained the statically assigned UID/GID of those users, and had a single recipe that would create them all.22:41
neverpanicEvery other recipe just includes the list of UIDs/GIDs and chowns files to the UIDs directly, without having the users in sysroot.22:41
xyzzy42Looks like do_populate_sysroot is what's supposed to be adding them.  From the log, it looks like it gets skipped.  "NOTE: Installed into sysroot: []" then  "NOTE: Skipping as already exists in sysroot: ['package-with-users', ...]"22:43
neverpanicwell, you'll have to clean to see the code that should add the user run.22:44
*** leon-anavi <leon-anavi!~Leon@> has quit IRC22:44
xyzzy42yep, cleansstate before I did that22:44
neverpanicI would assume it's the recipe-specific sysroot task for recipe A that would add the user, though.22:45
xyzzy42yes, I see the dependency of do_preapre_recipe_sysroot has package_with_users:do_populate_sysroot in it22:45
neverpanicbut as I said, writing a file that contains user_B_UID = "1234" user_B_GID = "3456", including it from both recipes and doing chown ${user_B_UID}:${user_B_GID} in recipe A may be the simpler solution, and in our experience, has fewer problems.22:46
neverpanicAlso pacakge_using_user:do_prepare_recipe_sysroot does depend on package_with_users:do_populate_sysroot, but package_with_users:do_populate_sysroot would create the user in the separate sysroot staging area for package_with_users, but not in the recipe-specific sysroot for package_using_user22:50
neverpanicthe do_populate_sysroot can't do that, as the recipe specific sysroot doesn't exist yet, and the /etc/passwd file in package_with_users's sysroot can't simply be copied either (and usually isn't, as only specific subfolders are staged into the sysroot)22:50
neverpanicSo there needs to be some additional logic running in package_using_user's do_prepare_recipe_sysroot to get this right.22:51
*** agust <agust!> has quit IRC22:51
xyzzy42I thought that extra logic was the purpose of this,
neverpanicSeems so. Maybe you're running into the filter in line 122?22:54
xyzzy42hmm, exit 0, odd choice of exit code, maybe that's it22:55
xyzzy42another thing I thought of, is that I do reference the group systemd-journal in a package, which is created by a systemd recipe, and that does work22:56
neverpanicseems meta/recipes-core/base-passwd/ takes care of running the postinst scripts.22:56
xyzzy42the group is created by the systemd recipe, the recipe that uses the systemd-journal group doesn't create, but does DEPENDS on it, and that worked22:56
neverpanicHm, or meta/classes/staging.bbclass:583 in extend_recipe_sysroot called by do_prepare_recipe_sysroot.23:00
xyzzy42I see recipe-sysroot/usr/bin/postinst-useradd-{systemd,my-package-withuser} both exists.  But the users/groups created by the former are present in sysroot/etc/passwd, while the latter are not.23:02
*** adelcast <adelcast!~adelcast@> has joined #yocto23:03
*** bluelightning <bluelightning!~paul@pdpc/supporter/professional/bluelightning> has quit IRC23:08
*** ericch <ericch!> has quit IRC23:08
*** bluelightning <bluelightning!~paul@pdpc/supporter/professional/bluelightning> has joined #yocto23:09
*** goliath <goliath!> has quit IRC23:12
*** dbro <dbro!> has joined #yocto23:20
*** florian <florian!~florian_k@Maemo/community/contributor/florian> has quit IRC23:22
*** dev1990 <dev1990!> has joined #yocto23:28
dev1990I'm having problem with "git pack-redundant" it's called by git fetcher under some conditions, it gives me always sigfaults23:36
kergothit's using git on your host..23:36
dev1990yes, I just want to ask if someone having same problem, I'm having this with 2.15.123:37
*** rburton <rburton!rburton@nat/intel/x-cgyoktjdtuddmrkf> has quit IRC23:37
dev1990ups I mean 2.25.1, I also have this with 2.14.1 and 2.13.1. Not sure it's my custom host configuration, I ruleout memory corruption and disk corruption23:43
dev1990ok, not my day, affected version 2.25.1, 2.24.1 and 2.23.1. Probably my distribution fault somehow, this sigfault not crashing bitbake I just found that Yocto builds giving me sigfauls in dmesg from time to time23:48
JPEWRP: No. the AB needs to set BB_LOGCONFIG... Which maybe should be in the env whitelist23:59

Generated by 2.17.2 by Marius Gedminas - find it at!