**** BEGIN LOGGING AT Wed Mar 11 03:02:27 2020
Mar 11 03:20:14 <yocti> New news from stackoverflow: Can't add dmidecode into yocto <https://stackoverflow.com/questions/60628772/cant-add-dmidecode-into-yocto> || bitbake-layers add-layer meta-python meta-raspberrypi failed <https://stackoverflow.com/questions/59127987/bitbake-layers-add-layer-meta-python-meta-raspberrypi-failed>
Mar 11 03:50:19 <yocti> New news from stackoverflow: What is the use of symbolic links in Yocto deploy folder <https://stackoverflow.com/questions/60628908/what-is-the-use-of-symbolic-links-in-yocto-deploy-folder>
Mar 11 04:29:51 <sj52> I have build python 2.7.15 using yocto. However, I get the following error - a = "\\x3cdiv\\x3e" b = a.decode("ascii") Traceback (most recent call last):
Mar 11 04:42:12 <sj52> The encoding folder is empty
Mar 11 04:42:14 <sj52> python2.7/encodings$
Mar 11 04:50:29 <yocti> New news from stackoverflow: Can't find encoding in python - ascii? <https://stackoverflow.com/questions/60628205/cant-find-encoding-in-python-ascii>
Mar 11 05:12:00 <Marex> hey uh, is setting IMAGE_FSTYPES in an image recipe (not local.conf or so) valid ?
Mar 11 05:18:48 <khem> Marex: yeah
Mar 11 05:19:08 <khem> you can do for initramfs images e.g.
Mar 11 05:21:21 <Marex> khem: ah ok, for some reason I was under the impression IMAGE_FSTYPES should only be set in local.conf / machine.conf
Mar 11 05:24:12 <khem> Marex: as a DISTRO maintainer thats what I want
Mar 11 05:33:02 <Marex> khem: errr, what is it that you want, per-image-recipe IMAGE_FSTYPES ?
Mar 11 05:43:43 <khem> I want to be able to generare same fstype for all machines as a distro person e.g.
Mar 11 05:46:14 <Marex> khem: but then you set the IMAGE_FSTYPE in your local.conf , right
Mar 11 05:47:08 <Marex> khem: but my question is the other way around -- is it valid for image recipe, e.g. foo-image.bb , to set IMAGE_FSTYPES to e.g. ubifs , and generate only ubifs if I bitbake foo-image ?
Mar 11 05:47:28 <Marex> khem: no matter what IMAGE_FSTYPES is set in e.g. machine-bar.conf ?
Mar 11 05:50:41 <yocti> New news from stackoverflow: Bitbake fails at do_rootfs : none of the providers can be installed <https://stackoverflow.com/questions/60503613/bitbake-fails-at-do-rootfs-none-of-the-providers-can-be-installed> || meta-virtualization rocko branch unable to locate docker package <https://stackoverflow.com/questions/49517065/meta-virtualization-rocko-branch-unable-to-locate-docker-package>
Mar 11 06:53:59 <xtron> hi all, is there any key comparison mechanism to pop a warning message when user use default keys for signing etc.. other way around is comparing the md5sum with default key's md5sum
Mar 11 07:24:52 <hopser> good morning. with the tipps of yesterday 'bitbake core-image-minimale' ends with "ERROR: Task (/workdir/poky/meta/recipes-devtools/qemu/qemu-system-native_4.1.0.bb:do_compile) failed with exit code '1'".
Mar 11 07:25:01 <hopser> I'm on yocto-3.0.1 in a docker container. (https://github.com/crops/poky-container/blob/master/README.md)
Mar 11 07:25:47 <erbo> hopser: you need to put the actual error in a pastebin somewhere so we can look at it
Mar 11 07:26:56 <hopser> erbo: ok, haven't done this before. I'll figure it out. Thank you
Mar 11 07:28:00 <erbo> It's usually a bit further up from what you gave us, and usually bitbake also gives you a log file name where you can find the error.
Mar 11 07:29:26 <ben24> Hello yocto users,I'm trying to build multiple kernels with yocto and I have a question about out-of-tree modules (kernel-module-xxx recipes): how can I handle them?They are not included in the module.tgz and they are not built as rpm or ipk.
Mar 11 07:34:00 <hopser> erbo: this is the error message -> https://pastebin.com/V38UVSpY
Mar 11 07:36:38 <erbo> hopser: "/workdir/build/tmp/hosttools/ld.bfd: final link failed: Input/output error", that doesn't really look like a yocto error to me. Any chance something went wrong before which could have left some generated file in a weird state?
Mar 11 07:38:02 <erbo> hopser: are you using some network filesystem or so that could cause IO issues?
Mar 11 07:38:58 <hopser> erbo: i'm on a laptop with arch linux. bitbake is run inside docker
Mar 11 07:39:23 <hopser> erbo: my partitions are ext4
Mar 11 07:39:54 <erbo> ok, and you bind the directory where output it stored into the docker container?
Mar 11 07:40:24 <erbo> If yes, that's how I always do and it's a quite common way.
Mar 11 07:40:33 <hopser> erbo: i'm working under /home
Mar 11 07:41:36 <hopser> erbo: this is how I run docker: "docker run --rm -it -v /home/hopser/projekte/spielwiese/yocto/01-quick-build:/workdir crops/poky --workdir=/workdir"
Mar 11 07:42:10 <erbo> hopser: then that shouldn't be the problem
Mar 11 07:42:33 <hopser> erbo: good
Mar 11 07:43:08 <hopser> erbo: what should i look after next?
Mar 11 07:44:10 <erbo> To me it looks like errors I mostly see when a generated binary ended up being truncated or so for some reason, but the build system thinks it's completed so it's not rebuilt.
Mar 11 07:44:38 <erbo> So I'd try to "bitbake -c cleansstate qemu-system-native" and then try the build again.
Mar 11 07:44:57 <hopser> erbo: thank you. i'll try that
Mar 11 07:46:34 <hopser> erbo: ERROR: Task do_cleanstate does not exist for target qemu-system-native
Mar 11 07:47:22 <hopser> erbo: next step would be to remove all folders and start over again?
Mar 11 07:48:58 <erbo> it's two "s"
Mar 11 07:49:03 <erbo> cleansstate
Mar 11 07:49:29 <hopser> erbo: :)  thanks
Mar 11 07:49:35 <erbo> :)
Mar 11 07:50:50 <hopser> erbo: wow, that additional 's' is magic
Mar 11 07:51:40 <hopser> erbo: i'm looking forward to use some additional 's' for my coding
Mar 11 07:52:37 <erbo> whens in doubtsss add somes ss
Mar 11 07:52:49 <hopser> lol, yeah
Mar 11 08:04:58 <hopser> erbo: thank you! 'bitbake -c cleansstate qemu-system-native' lead to a clean base. no the image was build
Mar 11 08:05:18 <hopser> erbo: /s/no/now/g
Mar 11 08:10:37 <erbo> hopser: ok, so then problably some file ended up corrupted for some reason. I usually only see that then the OOM killer have stepped in, or I have killed a build process in a not so nice way.
Mar 11 08:11:03 <erbo> But good that it works now
Mar 11 08:11:53 <hopser> erbo: thank you. i've learned how to handle this in the future :)
Mar 11 08:16:10 <LetoThe2nd> note please, -c clean should also work and is the recommended way.
Mar 11 08:17:56 <hopser> LetoThe2nd: alright, i'll use them both ;)
Mar 11 08:19:59 <erbo> Yeah I guess any truncated binary would have been generated by the failing task, so clean should work. Using cleansstate means that the other tasks run before, like configure etc, would also be re-run which probably wasn't neccessary here.
Mar 11 08:24:39 * LetoThe2nd declares this to be the new YP anthem https://www.youtube.com/watch?v=2nX6qGeyaGM
Mar 11 08:24:47 <LetoThe2nd> (until i declare something else)
Mar 11 08:28:30 <RP> wow, a green full build. Makes a nice chance to wake to that
Mar 11 08:28:49 <RP> albeit with two warnings which JPEW knows about :)
Mar 11 08:29:22 <LetoThe2nd> RP: click link, crank up to 11 and add that to the builg green feeling - awesome!
Mar 11 09:29:01 <hopser> LetoThe2nd: how do run runqemu? on my host i use 'qemu-system-x86_64 -kernel bzImage -hda core-image-minimal-qemux86-64.ext4 -append "console=ttyS0 root=/dev/hda"'
Mar 11 09:30:08 <hopser> LetoThe2nd: when using docker container for bitbake, i can't use runqemu. and on my host the env-settings are missing
Mar 11 09:31:38 <LetoThe2nd> hopser: you totally can, with a few minor restrictions.
Mar 11 09:32:31 <LetoThe2nd> hopser: i'm not a huge qemu user anyways, but "runqemu qemux86-64 nographic slirp" should suffice for simple tests inside a docker container.
Mar 11 09:35:21 <hopser> LetoThe2nd: ah, i see, thanks. inside the container 'runqemu qemux86' fails. i haven't seen your parameters so far.
Mar 11 09:36:16 <LetoThe2nd> hopser: hint "fails" is one of the most useless error descriptions ever. if you really care about getting help, make sure to include at least the most obvious error messages.
Mar 11 09:36:54 <LetoThe2nd> a.k.a. "it doesn't work!" - "what does that mean? does it sit at home all day and watch tv?"
Mar 11 09:38:13 <erbo> That's what my kid told the daycare personell, "My dad is at home all day watched TV".
Mar 11 09:38:25 <hopser> LetoThe2nd: that's right. you solved the problem with running qemu inside docker. so i didn't want to spam the chat
Mar 11 09:38:38 <erbo> s/watched/watching/
Mar 11 09:40:47 <LetoThe2nd> hopser: for the next time :)
Mar 11 09:41:06 <hopser> LetoThe2nd: for completeness: in docker, 'runqemu qemux86' leads to 'runqemu - ERROR - /workdir/build/tmp/deploy/images/qemux86 not a directory valid DEPLOY_DIR_IMAGE'
Mar 11 09:41:29 <hopser> LetoThe2nd: ...and 'runqemu - ERROR - Command 'ls -t /workdir/build/tmp/deploy/images/qemux86/*.qemuboot.conf' returned non-zero exit status 2.'
Mar 11 09:41:53 <LetoThe2nd> hopser: probably because you build for qemux86-64 :) but thats totally not docker related.
Mar 11 09:43:36 <hopser> LetoThe2nd: we had the story about slow working brains, yesterday... ;)
Mar 11 09:45:05 <LetoThe2nd> hopser: anyways, this was good information anyways because it removes any doubt about problems in containers, and concludes to a simple user error. so however annoying it might be for you, it is helpful for us to know what actually went wrong, and that (in this case) we do not have to care further.
Mar 11 09:45:47 <hopser> LetoThe2nd: good :)
Mar 11 09:45:59 <hopser> LetoThe2nd: and thank you. :)
Mar 11 09:46:24 <LetoThe2nd> :)
Mar 11 09:47:57 <LetoThe2nd> between Ceyboard and Khair? ;-)
Mar 11 09:48:34 <nextPEBCAK> LetoThe2nd: for me it's Chair and Keyboard :-P
Mar 11 09:48:55 <nextPEBCAK> LetoThe2nd: is there a copyright on your version?
Mar 11 09:50:01 <LetoThe2nd> rule #1) leto is always right.
Mar 11 09:50:15 <LetoThe2nd> rule #2) if leto is not right, rule #1 applies.
Mar 11 09:50:43 <LetoThe2nd> rule #3) RP, kergoth and rburton can override #1 and #2 :)
Mar 11 09:55:48 <nextPEBCAK> LetoThe2nd: if put that into my readme
Mar 11 09:56:21 <LetoThe2nd> btw.
Mar 11 09:56:23 <nextPEBCAK> '/s/if/i've/g'
Mar 11 09:56:24 <LetoThe2nd> RP: https://twitter.com/TheYoctoJester/status/1237674451304660992
Mar 11 09:56:45 <RP> LetoThe2nd: :)
Mar 11 09:57:24 <nextPEBCAK> wow, congrats to tha AO
Mar 11 10:00:25 <LetoThe2nd> RP: i just noticed that i basically delivered a prime impostor example. claiming "we break" as somebody who doesn't have a single patch in YP/OE is... funny!
Mar 11 10:01:12 <RP> LetoThe2nd: you could send a patch? :)
Mar 11 10:02:08 <qschulz> LetoThe2nd: well at least, you didn't break a dot release of thud with your first patch :)
Mar 11 10:02:13 <LetoThe2nd> RP: i will, once i figure out my build failures with freebeer_1.0.bb
Mar 11 10:03:41 <LetoThe2nd> qschulz: want some PR for that?
Mar 11 10:05:15 <qschulz> LetoThe2nd: lots of beer to forget
Mar 11 10:07:46 <LetoThe2nd> qschulz: ++
Mar 11 10:07:57 * LetoThe2nd will have beer and loud music tonight.
Mar 11 12:20:27 <Ninic0c0> Hello, I'm trying to add a do_rootfs function, all seems ok but i get an error about timestamp (https://pastebin.com/nUGP6xfA) it's probably because i have created a sub dir inside deploy dir ? Thx!
Mar 11 12:50:11 <zino> hi, I can't see the service file in the resulting image: I got this from another working recipe. any hints? https://pastebin.com/8ehFv2gG
Mar 11 12:51:58 <LetoThe2nd> zino: FILES_${PN} need the path in D, not the filename
Mar 11 12:54:14 <zino> LetoThe2nd: sorry (i'm quite new to this), I don't really get it
Mar 11 12:54:43 <zino> I have gcom.service sitting in "files" folder
Mar 11 12:55:06 <rburton> doesnt the systemd class do that for you anyway?
Mar 11 12:55:11 <zino> and gcom.service ends up in the rpm
Mar 11 12:55:26 <rburton> and your do_install should use ${includedir} instead of making up incdir
Mar 11 12:55:45 <LetoThe2nd> rburton: yeah, i think so too. still the line is buggy :)
Mar 11 12:56:44 <LetoThe2nd> zino: it ends up in the rpm, but doesn't get isntalled, but the whole rest of the rpm does?
Mar 11 13:01:26 <zino> good question: I get gcom-package-1.0-r0.corei7_64.rpm, gcom-package-dbg-1.0-r0.corei7_64.rpm, gcom-package-dev-1.0-r0.corei7_64.rpm, gcom-package-staticdev-1.0-r0.corei7_64.rpm
Mar 11 13:02:01 <zino> the service file is only in the first rpm and I don't think it gets installed
Mar 11 13:02:46 <LetoThe2nd> zino: don't think?
Mar 11 13:03:06 <zino> if i "bitbake -c devshell gcom-package" I can see the service file in sysroot-destdir
Mar 11 13:03:18 <LetoThe2nd> zino: how about any other contents of the rpm? and, maybe it just doesn't get installed to where you thing?
Mar 11 13:03:21 <LetoThe2nd> s/thing/think/
Mar 11 13:03:42 <LetoThe2nd> zino: thats not the way to check, you really have to look at the rpm file itself.
Mar 11 13:03:59 <zino> LetoThe2nd: "don't think" since it contains two files that I don't see int the image, so it's actualli "no"
Mar 11 13:04:29 <rburton> so the problem is that your image doesn't contain the package
Mar 11 13:04:35 <rburton> do you add the package to the image?
Mar 11 13:04:37 <LetoThe2nd> yeah.
Mar 11 13:05:01 <LetoThe2nd> zino: have you just done bitbake gcom, or have you actually added it to your image recipe and rebuilt it?
Mar 11 13:05:23 <ben24> hello everybody,i'm trying to use yocto multiple kernel capability but i don't understand how to manage out-of-tree modulesis there a way to build them for alternative kernels or add them to alternative kernels module.tgz?
Mar 11 13:09:15 <zino> LetoThe2nd: gcom declared as dependency of another package. I've tried "bitbake -c cleanall gcom-package; bitbake core-image-minimal"
Mar 11 13:10:34 <LetoThe2nd> zino: DEPENDS meant that it has to be around during build time of the other package, not exactly at run time. so if that other package actually relies on gcom to be installed and running, you need to add to to RDEPENDS_${PN} of the other on.
Mar 11 13:10:38 <LetoThe2nd> +e
Mar 11 13:12:32 <zino> LetoThe2nd: thank you, I'm adding that. could this be the reason for the missing service file?
Mar 11 13:12:56 <LetoThe2nd> zino: "missing" in the sense of "nobody knew you wanted it to be installed"? yes, probably.
Mar 11 13:20:41 <qschulz> I don't remember exactly, does Yocto enforce the python version for task? or does it take /usr/bin/python and it takes whatever the host distro linked it to?
Mar 11 13:21:17 <LetoThe2nd> qschulz: for task?
Mar 11 13:21:44 <LetoThe2nd> qschulz: AFAIK it uses the host one, but checks the minmal version/requirements.
Mar 11 13:22:07 <qschulz> LetoThe2nd: could you elaborate on what's after the coma in your sentence?
Mar 11 13:22:55 <LetoThe2nd> qschulz: bitbake itself runs on the one that the host provides. it checks the host provided one for the needed minimal version.
Mar 11 13:23:13 <LetoThe2nd> e.g. 3.5 currently, if my brain works.
Mar 11 13:24:09 <qschulz> LetoThe2nd: where is this "needed minimal version" set?
Mar 11 13:25:19 <LetoThe2nd> qschulz: https://git.openembedded.org/bitbake/commit/?id=1412dcc077b1bea10b72fc8b525d6258dca46d97
Mar 11 13:26:28 <zino> LetoThe2nd: I've added RDEPENDS_${PN} = " gateway" and rebuilt but I still can't see the service file
Mar 11 13:26:29 <qschulz> LetoThe2nd: I'm asking because we're apparently using some python2 code and some clients have python3 as default in their host system which breaks the code.
Mar 11 13:26:58 <LetoThe2nd> huh?
Mar 11 13:27:21 <LetoThe2nd> zino: the one that depends ON gcom has to have gcom in its RDEPENDS.
Mar 11 13:27:46 <LetoThe2nd> zino: if that is given, then the next step is to look at the manifest
Mar 11 13:27:53 <LetoThe2nd> qschulz: :)
Mar 11 13:28:16 <zino> LetoThe2nd: thank you for your patience
Mar 11 13:28:30 <LetoThe2nd> zino: i am highly impatient!
Mar 11 13:29:02 * LetoThe2nd cranks up Hammerfalls "Patient Zero".
Mar 11 13:29:55 <LetoThe2nd> qschulz: you could maybe use meta-python2, depend on python2-native and do some nasty hacks to (ab)use that?
Mar 11 13:30:00 <Ninic0c0> Stupid Q: I have set  MYVAR="avalue" inside do_rootfs_append (inside bbclass) and added do_cust_function() inside the same class. do_cust_function is not able to get the value of VAR. In yocto variables are not always expanded in bash ? Thx
Mar 11 13:31:24 <qschulz> LetoThe2nd: no, definitely no. It's just a hack we have in place so clients only have to set one variable to set the root password.
Mar 11 13:31:30 <LetoThe2nd> Ninic0c0: they do not leak outside of functions, AFAIK.
Mar 11 13:31:46 <LetoThe2nd> qschulz: very good. hi5
Mar 11 13:32:00 <qschulz> LetoThe2nd: https://git.openembedded.org/bitbake/tree/lib/bb/__init__.py?h=1.40
Mar 11 13:32:09 <qschulz> I'm puzzled, would this mean it's using python3?
Mar 11 13:32:20 <qschulz> (we're using thud, so bitbake 1.40 apparently)
Mar 11 13:32:38 <Ninic0c0> LetoThe2nd so what is the tricks to share a variable between 2 function in bbclass ? o.0
Mar 11 13:32:49 <qschulz> Ninic0c0: set it outside :) ?
Mar 11 13:34:06 <zino> LetoThe2nd: that did the trick: thank you very much
Mar 11 13:34:15 <Ninic0c0> qschulz Thx to replyy too, in fact the variable is based on a time stamp, the do_rootfs_append save the value and do_cust_function need this variable latter in the build process :S I'm stupid ? o.0
Mar 11 13:36:00 <qschulz> is the do_cust_function only called from your do_rootfs task?
Mar 11 13:36:17 <LetoThe2nd> zino: listening to metal in general and hammerfall in particular often does the trick, i agree.
Mar 11 13:36:43 <qschulz> I mean, does it have to be a task or can it be part of do_rootfs? In which case, just make it a normal shell function and call it from do_rootfs and send whatever variable you want as argument to the function?
Mar 11 13:36:47 <qschulz> Ninic0c0: ^
Mar 11 13:36:53 <emrius> Hey, I'm trying to write a recipe to fetch a resource from github which I think should be absolute standard. Still I'm failing here and I have no fuc*** clue why. I have this `SRC_URI = "git://git@github.com/iris-edu/libmseed.git;protocol=git;branch=2.x"``
Mar 11 13:37:03 <emrius> and `SRCREV = "67f0bee9e3b3e55be94e3b34944203270fac4677"`
Mar 11 13:37:19 <emrius> which I is the commit I want to checkout. Anything wrong so far?
Mar 11 13:37:42 <qschulz> emrius: why git@?
Mar 11 13:38:04 <emrius> grrrrrrr
Mar 11 13:38:12 <qschulz> SRC_URI = "git://github.com/iris-edu/libmseed.git;branch=2.x"?
Mar 11 13:38:21 <emrius> stupid copy-paste from github.com
Mar 11 13:38:25 <LetoThe2nd> emrius: SRC_URI = "git://github.com/fedora-modularity/libmodulemd;protocol=https;branch=1.x-maint something
Mar 11 13:38:38 <LetoThe2nd> emrius: straight out of just grepping poky/meta
Mar 11 13:39:25 <emrius> thanks. I should really have another coffee
Mar 11 13:39:34 <LetoThe2nd> s/coffee/beer/
Mar 11 13:41:46 <tgamblin> LetoThe2nd: or Bailey's in the coffee
Mar 11 13:42:33 * LetoThe2nd gives tgamblin the creepy looks.
Mar 11 13:49:28 <LetoThe2nd> tgamblin: i do some speaking now and then for a specifc conf organizer, and they have the habit of setting up a speaker dinner the evening before. every time i'm with them, after the dinner i order an espresso, pour some scotch from my flask in and declare the conference officially open :)
Mar 11 13:55:32 <emrius> Hmm... It's downloadin something but there is no license file...
Mar 11 13:55:44 <emrius> Am I missing something here: https://pastebin.com/vB2QPyrA
Mar 11 13:55:46 <emrius> ?
Mar 11 13:56:20 <emrius> I mean neglecting the checksums
Mar 11 13:57:01 <emrius> Hang on, let me add the checksums
Mar 11 13:57:11 <emrius> probably it just looked as if something was pulled
Mar 11 13:57:13 <emrius> ?!
Mar 11 13:57:23 <LetoThe2nd> emrius: license TODO and a non-set checksum won't work
Mar 11 13:58:26 <tgamblin> LetoThe2nd: an excellent choice
Mar 11 13:58:34 <emrius> Why doesn't it raise an exception complaining about that then but instead says `LIC_FILES_CHKSUM points to an invalid file: /home/marius/mender-qemu/build/tmp/work/core2-64-poky-linux/libmseed/1.0-r0/libmseed-1.0/LICENSE [license-checksum]``
Mar 11 13:58:46 <emrius> confusing
Mar 11 13:59:03 <LetoThe2nd> emrius: bevore i do so, have you actually looked at poky for examples?
Mar 11 13:59:12 <LetoThe2nd> tgamblin: they keep booking me for unknown reasons :)
Mar 11 14:07:24 <emrius> Same result also with checksums: https://pastebin.com/kGXFw9n6
Mar 11 14:09:08 <emrius> BTW: checksums are not correct, I know. I waited for bb to check them before when writin a new recipe from scratch and corrected them then according to the verification. That worked before. However, it seems like bb doesn't even reach the point of verification.
Mar 11 14:11:34 <qschulz> emrius: S = "${WORKDIR}/git" is missing
Mar 11 14:12:05 <emrius> Thanks! shouldn't that be default?
Mar 11 14:12:13 <qschulz> no
Mar 11 14:12:21 <bittin-> Listening to talk about Yocto at WSLConf #1 Day 2 now :)
Mar 11 14:12:33 <LetoThe2nd> bittin-: WSLConf?
Mar 11 14:12:35 <emrius> ok, I will (y)
Mar 11 14:12:44 <emrius> ah no
Mar 11 14:13:11 <qschulz> emrius: http://git.yoctoproject.org/cgit.cgi/poky/tree/meta/conf/bitbake.conf#n372
Mar 11 14:13:18 <emrius> sorry thought that was where I would find why it's not default.
Mar 11 14:13:37 <bittin-> LetoThe2nd: https://www.wslconf.dev/ Microsoft and Canonicals conference about Windows 10 and Ubuntu, was supposed to be at Microsofts offices in Seattle but its now online over Bluejeans Canonicals instance video conference due to Coronavirus
Mar 11 14:13:59 <LetoThe2nd> bittin-: ah interesting.
Mar 11 14:14:22 <emrius> @qschulz: alright, but if the fetcher is git, data will be cloned into `git`. So can't the git fetcher simply set S = ...
Mar 11 14:14:23 <emrius> ?
Mar 11 14:14:24 <bittin-> https://twitter.com/DonPezet/status/1237742232817500160
Mar 11 14:14:33 <emrius> wouldn't that make sens?
Mar 11 14:14:37 <emrius> sense
Mar 11 14:14:38 <qschulz> emrius: patches welcome :)
Mar 11 14:14:46 <emrius> (y)
Mar 11 14:15:38 <emrius> If this image will finally be working one day, I will dive into the actual code of bitbake
Mar 11 14:15:46 <qschulz> The thing is, you can have multiple git repo which aren't the main source (where make should be run for example, etc.) in SRC_URI
Mar 11 14:15:56 <qschulz> and that would fuck up S
Mar 11 14:16:19 <emrius> I see
Mar 11 14:17:07 <rburton> bittin-: 'unleash your iot development tools' i presume?
Mar 11 14:17:25 <bittin-> rburton: yep
Mar 11 14:18:08 <bittin-> still for another 13 minutes, and if you don't watch it live it will be on Youtube on Tuesday next week
Mar 11 14:19:32 <LetoThe2nd> bittin-: whats the executive summary?
Mar 11 14:19:49 <rburton> presumably that iot azure sphere is awesome and yocto powered
Mar 11 14:20:01 <rburton> MS would say the first bit obviously :)
Mar 11 14:21:14 <bittin-> rburton: actually they haven't said anything about Azure Sphere yet
Mar 11 14:22:45 <bittin-> they are mostly talking about that you can use Yocto to setup an own Linux for your smart device and connect to it via Visual Studio/Visual Studio code to program it
Mar 11 14:22:56 <bittin-> talk just ended 8 minutes for questions now
Mar 11 14:23:13 <LetoThe2nd> bittin-: so that brought you to use?
Mar 11 14:25:04 <bittin-> LetoThe2nd: yeah don't really work with embedded much however, but tough i would notice you
Mar 11 14:26:02 <LetoThe2nd> bittin-: ah. thanks, it is much appreciated.
Mar 11 14:37:04 <JPEW> RP: Thoughts on the latest comment on https://bugzilla.yoctoproject.org/show_bug.cgi?id=13813
Mar 11 14:37:05 <yocti> Bug 13813: normal, High, 3.1 M4, JPEWhacker, IN PROGRESS REVIEW , bitbake hashequiv messages in the console are too verbose
Mar 11 14:38:23 <JPEW> Not sure if we should leave it, make it only print once, remove it completely, or make it only print if any setscene tasks actually ran (e.g. "NOTE: 5 setscene tasks completed")
Mar 11 14:38:31 <RP> JPEW: we need to hide that one somehow
Mar 11 14:38:59 <RP> JPEW: A summary of the number of tasks completed might be best?
Mar 11 14:39:26 <RP> JPEW: that way the user sees progress? Hiding may be easiest. Not sure :/
Mar 11 14:39:55 <RP> JPEW: we had warnings in last nights build again. My current plan is to wait for the fixes for this, then build M3
Mar 11 15:16:21 <fullstop> I have a systemd bbappend just for a config file.. If I change this file, which has nothing to do with the compilation of systemd, a considerable amount of things will be rebuilt.  Is there a way to avoid this?
Mar 11 15:18:47 <LetoThe2nd> fullstop: make it a standalone recipe that just RDEPENDS on systemd, if the config file is freestanding.
Mar 11 15:19:20 <qschulz> fullstop: no I don't think so. Since a file from SRC_URI has changed, do_fetch is run again, and since it's been re-run, the tasks depending on it will be re-run as well
Mar 11 15:19:36 <fullstop> I think that it will conflict with the systemd package itself.
Mar 11 15:19:54 <LetoThe2nd> fullstop: then its more complicated (TM)
Mar 11 15:19:55 <fullstop> I want to modify journal.conf
Mar 11 15:20:10 <LetoThe2nd> on the other hand, how often are you gonna do that?
Mar 11 15:20:26 <fullstop> According to my statistics, twice.
Mar 11 15:21:32 <LetoThe2nd> if my problems consist of caring about building systemd twice, i'll consider myself a happy man.
Mar 11 15:22:34 <fullstop> haha
Mar 11 15:22:43 <yocti> New news from stackoverflow: Yocto install a script along with a kernel module <https://stackoverflow.com/questions/60638969/yocto-install-a-script-along-with-a-kernel-module>
Mar 11 15:22:56 <fullstop> It's the other stuff which depends on systemd that I care about.
Mar 11 15:24:11 <fullstop> aha, I can put the stuff in /etc/systemd/journald.conf.d/*.conf
Mar 11 15:40:47 <fullstop> Ah, that's much better.  :-)
Mar 11 17:10:32 <JPEW> RP: Ok, I think I have it sorted out. I'll work on getting the patches sent today
Mar 11 17:10:59 <RP> JPEW: thanks! Think we're nearly there with M3 :)
Mar 11 17:11:06 <JPEW> RP: Good
Mar 11 17:11:34 <RP> JPEW: FWIW "bitbake-selftest -r tinfoil" seems to reproduce those warnings
Mar 11 17:12:17 <JPEW> RP: Ok, I'll run that locally to make sure I fixed them
Mar 11 17:13:29 <RP> JPEW: I was starting to think about looking but it sounds like you have it in hand
Mar 11 17:15:27 <JPEW> RP: Weird. I didn't see them locally. Is it on a specific AB node? I wonder if it's specific to a version of Python
Mar 11 17:16:43 <JPEW> RP: Oh, nevermind. I see it when I add -j1
Mar 11 17:17:12 <RP> JPEW: I didn't have -j locally. Seen on centos7, ubuntu1804 and ubuntu1904 but not on some other workers
Mar 11 17:17:26 <RP> JPEW: Its intermittent somehow
Mar 11 17:17:37 <RP> which is never a good sign :/
Mar 11 17:23:48 <rpi_poky_newbie> HelloWorld!!
Mar 11 17:24:13 <rpi_poky_newbie> im tring to expand the rootfs at boot like the raspbian distro but it seems to fail..... with raspbian it adds a init script to cmdline.txt to repartition then systemd  calls resize2fs.... with poky after boot i cant access cmdline.txt so using raspbian scripts result in a bootloop... any ideas?
Mar 11 17:25:46 <jrdn> have you tried setting IMAGE_ROOTFS_EXTRA_SPACE_append in your image.bb?
Mar 11 17:26:42 <jrdn> that will just append some empty space to your image.  do you need it to happen at boot?
Mar 11 17:27:04 <rpi_poky_newbie> wanting to happen at boot
Mar 11 17:28:03 <rpi_poky_newbie> but cmdline.txt is missing form /boot after image is running so the way raspbian does it fails
Mar 11 17:29:24 <rpi_poky_newbie> creates a bootloop because it cant remove itself from the init=some.sh off of the cmdline.txt
Mar 11 17:33:52 <rpi_poky_newbie> using the fs-init recipe in the yocto index just says the "partition is at x already, nothing to do" it seems to want the partition to be resized with fdisk before running resize2fs to expand
Mar 11 17:34:40 <JPEW> RP: Did you ever write up that howto on tracemalloc? Looks like it tracks resource warnings also :)
Mar 11 17:37:51 <RP> JPEW: I don't remember if I did :/
Mar 11 17:38:46 <RP> JPEW: 18th Feb, bitbake-devel, "bitbake memory usage debugging using tracemalloc"
Mar 11 17:39:40 <JPEW> RP: Got it. Thanks
Mar 11 17:44:38 <rpi_poky_newbie> so any ideas? would it be better off to create a 3rd data partition and expand it?
Mar 11 17:48:30 <rpi_poky_newbie> anyone know where cmdline.txt goes while running?
Mar 11 18:44:23 <rpi_poky_newbie> so many peeps and no one talks is this thing loaded with bots?
Mar 11 18:48:52 <khem> rpi_poky_newbie: there is a package called 96boards-tools in meta-96boards if you add it to recipe it will autoexpand the paritions, default is to expand the default rootfs partition
Mar 11 18:52:27 <rpi_poky_newbie> thx i will check now
Mar 11 18:54:38 <rpi_poky_newbie> heres a stupid one? whats the right way to include a package thats been split and not installed? bbappend, RDEPENDS_pn what has been working but is that right?
Mar 11 18:59:34 <rpi_poky_newbie> khem : thx i found it will try it now
Mar 11 19:23:22 <dbro> Has anyone had success using the meta-openembedded gRPC recipe and CMake's find_package when cross compiling? The oe gRPC recipe installs gRPCConfig.cmake only into the recipe-sysroot-native dir, and when CMake finds that file it is capable of finding only the corresponding recipe-sysroot-native libs, though libgrpc++.so is installed in
Mar 11 19:23:23 <dbro> recipe-sysroot
Mar 11 19:26:46 <dbro> I can make it work by manually copying the missing cmake artifacts (and grpc binaries which a target would never need but gRPC's cmake files require) in do_compile_prepend, but that's wrong
Mar 11 19:28:04 <rpi_poky_newbie> https://github.com/openembedded/meta-openembedded/issues/162
Mar 11 19:28:19 <rpi_poky_newbie> did you see that?
Mar 11 19:34:03 <dbro> I hadn't, and don't have the patch. Thank you! My issue is specific to target recipe so not sure it's related but trying now....
Mar 11 19:34:54 <dbro> e.g: I think the cmake dir was "shipped" to sysroot-native fine (don't fully grok the "shipped" terminology though)
Mar 11 19:48:03 <mauz555> is there a way to override HOST_SYS for a specific layer ?
Mar 11 20:09:53 <dbro> :rpi_poky_newbie That patch fixes the cmake files not installed, but not the grpc binaries (e.g: grpc_cpp_plugin). A start! Thank you again
Mar 11 20:19:47 <rburton> mauz555: why would you do that?
Mar 11 22:18:11 <mauz555> i have troubles building a rust app for aarch64 using `meta-rust` layer (https://github.com/meta-rust/meta-rust) anyone could help ?
Mar 11 23:29:55 <JPEW> RP: Just pushed the next set of bitbake logging patches
Mar 11 23:30:10 <JPEW> Err, send them to the mailing list that is
Mar 11 23:31:36 <khem> JPEW: does this address some of logging issues I see with docker
Mar 11 23:31:49 <JPEW> khem: Depends; what are you seeing?
Mar 11 23:31:57 <khem> e.g. when output is piped to a file then it has tonne of junk in it
Mar 11 23:32:27 <JPEW> khem: Like when you do 'bitbake > log.txt' ?
Mar 11 23:32:43 <khem> I have to disable -t and then it works
Mar 11 23:32:49 <khem> -t to launch docker
Mar 11 23:33:13 <khem> right
Mar 11 23:33:31 <JPEW> khem: No it won't fix that. Basically, it's because bitbake thinks its running with a tty and doesn't know it's being redirected
Mar 11 23:33:35 <khem> and bitbake a wrapper to launch bitbake under docker
Mar 11 23:33:46 <khem> yeah ok
Mar 11 23:33:55 <khem> not a big issue but it is annoying
Mar 11 23:34:05 <JPEW> Ya, FWIW, we do this: https://github.com/garmin/pyrex/blob/master/pyrex.py#L476
Mar 11 23:34:36 <JPEW> It's not perfect, i.e. it breaks down if you try to do stdout/stderr seperately
Mar 11 23:35:26 <khem> right yeah I know
Mar 11 23:37:22 <JPEW> khem: But, the patches I have pending will allow you to explicitly log certain domains to a separate file more easily
Mar 11 23:37:39 <JPEW> Basically, you can do anything that Python's structured logging will allow :)
Mar 11 23:44:16 <kergoth> khem: wehn i wrap bitbake under docker i check if stdout is a tty, and if not, add "| cat" to the end of the in-docker bitbake command ;)
Mar 11 23:44:45 <khem> kergoth: yeah that works too
Mar 11 23:44:56 <khem> but I was hoping it to be seemless
Mar 11 23:46:51 <RP> JPEW: will that new code fix tinfoil log warnings?
Mar 11 23:47:01 <JPEW> RP: It did in my testing
Mar 11 23:47:21 <RP> JPEW: ok, thanks. I didn't think knotty was active in that case :/
Mar 11 23:48:09 <JPEW> RP: Huh, weird. There was some previous code in that spot that closed the loggers that I removed b/c I thought it was unnecessary
Mar 11 23:48:12 <JPEW> Apparently not
Mar 11 23:48:18 <khem> kergoth: can you share the logic ?
Mar 11 23:56:03 <rpi_poky_newbie> hey anyone made a bsp package for "Orange pi 4g-iot"  yet?
Mar 12 00:29:26 <RP> JPEW: I'll run it, see what happens :)
Mar 12 00:29:31 <RP> JPEW: thanks!
Mar 12 02:55:01 <yocti> New news from stackoverflow: wic image is EFI but need legacy <https://stackoverflow.com/questions/60646899/wic-image-is-efi-but-need-legacy>
**** ENDING LOGGING AT Thu Mar 12 02:59:57 2020