<DvorkinDmitry>
how can I make /var/log directory non-volatile? I'm using systemd, its recipe adds /etc/tmpfiles.d/00-create-volatile.conf file with "... /var/volatile/log ..." record inside. May I just override this file with .bbappend and my own 00-create-volatile.conf file?
demirok has quit [Quit: Leaving.]
Estrella has quit [Ping timeout: 260 seconds]
Estrella_ has joined #yocto
nemik_ has quit [Ping timeout: 256 seconds]
nemik_ has joined #yocto
nemik_ has quit [Ping timeout: 260 seconds]
nemik_ has joined #yocto
florian_kc has quit [Ping timeout: 260 seconds]
sakoman has quit [Quit: Leaving.]
xmn has quit [Quit: ZZZzzz…]
kpo has quit [Read error: Connection reset by peer]
kscherer has quit [Quit: Konversation terminated!]
mark_L has quit [Quit: mark_L]
mark_L has joined #yocto
mixfix41 has quit [Ping timeout: 246 seconds]
xmn has joined #yocto
jclsn has quit [Ping timeout: 252 seconds]
jclsn has joined #yocto
xmn has quit [Ping timeout: 260 seconds]
xmn has joined #yocto
jdavid75 has quit [Quit: Client closed]
aleksandarsimono has quit [Quit: Client closed]
invalidopcode has quit [Remote host closed the connection]
invalidopcode has joined #yocto
Guest70 has quit [Quit: Client closed]
amitk has joined #yocto
sakoman has quit [Quit: Leaving.]
tor has joined #yocto
kpo has quit [Read error: Connection reset by peer]
kpo has joined #yocto
camus has joined #yocto
camus has quit [Client Quit]
camus has joined #yocto
camus has quit [Client Quit]
mark_L has quit [Ping timeout: 260 seconds]
camus has joined #yocto
camus has quit [Ping timeout: 255 seconds]
camus has joined #yocto
alessioigor has joined #yocto
rstreif has quit [Ping timeout: 260 seconds]
mattes-b_ has quit []
Starfoxxes has joined #yocto
Starfoxxes has quit [Max SendQ exceeded]
pbergin has quit [Quit: Leaving]
<LetoThe2nd>
yo dudX
tomzy_0 has joined #yocto
leon-anavi has joined #yocto
xmn has quit [Ping timeout: 268 seconds]
gho has joined #yocto
rob_w_ has joined #yocto
mckoan|away is now known as mckoan
<mckoan>
good morning
<tomzy_0>
morning
camus1 has joined #yocto
* alessioigor
waves all
camus has quit [Ping timeout: 260 seconds]
camus1 is now known as camus
invalidopcode has quit [Remote host closed the connection]
invalidopcode has joined #yocto
Starfoxxes has joined #yocto
Starfoxxes has quit [Max SendQ exceeded]
<tomzy_0>
Is there anywhere good practices / methodology how to most effectively update Yocto release in custom project?
<tomzy_0>
Yesterday on yocto summit chat there was a little discussion about that
<tomzy_0>
and I was wondering if there is some kind of guidelines how to do it best
<JaMa>
I prefer to do it on ongoing basis, that you adapt to new changes as soon as they land in master, because then it's easier to see what caused those issues and to find some examples how similar issues were fixed in other components and other layers, but I guess you're already long past that
<LetoThe2nd>
tomzy_0: yeah was just about to say - the best option probably is to have a branch surfing on main and fixing that as you go along and see breakage.
<JaMa>
I also prefer to have our internal builds at least buildable with each release (when upgrading across couple releases at the same time), because it gives me at least some "bisect" points when debugging something
<tomzy_0>
interesting point of view
<JaMa>
and I don't try to make it perfect in runtime, just because we have a lot of prebuilt binaries from other teams and other companies and some of them are expensive to get rebuilt
<JaMa>
so the full set of prebuilt binaries compatible with new ABI is available only for the final release we're upgrading to
Estrella_ has quit [Ping timeout: 256 seconds]
Estrella has joined #yocto
<tomzy_0>
Yeah, runtime is other thing.. but I guess for that we should always have a set of tests that checks needed functionalities
zpfvo has joined #yocto
<LetoThe2nd>
for the record, i am currently looking into a very similar topic here
camus has quit [Remote host closed the connection]
<JaMa>
I had a presentation about it on Wed, but specific for your LGE use
camus has joined #yocto
ArgaKhan has joined #yocto
<JaMa>
after doing 7 major yocto release upgrades there since 2012 :)
<tomzy_0>
LetoThe2nd which is?
<LetoThe2nd>
tomzy_0: how to continually support and check bsps and integrations.
mvlad has joined #yocto
<tomzy_0>
JaMa sorry but what is LGE? Will definitely check your pres once they will be available on youtube :)
<ArgaKhan>
Hello, this is my first time joining this channel. Actually I'm having a problem and while searching for a solution I found it here.
<tomzy_0>
wow
<JaMa>
s/your/our/g in my earlier message
<tomzy_0>
Did not know that webOS could be built with yocto
<LetoThe2nd>
ArgaKhan: welcome! state your case of emergency and wait for rescue!
<LetoThe2nd>
tomzy_0: hehe webos IS YOEP
<tomzy_0>
JaMa yea, makes more sense now :D
<JaMa>
webOS as well as PalmOS were always built with OE
<ArgaKhan>
Thank you, then I will write my problem.
<ArgaKhan>
This is the error I'm having:
<ArgaKhan>
Exception: subprocess.CalledProcessError: Command '['rm', '-rf', '/build/tmp/work/armv7vet2hf-neon-mepos-linux-gnueabi/qtmultimedia/5.15.3+gitAUTOINC+185f37e68e- r0/sstate-build-package/']' died with <Signals.SIGABRT: 6>.
nemik_ has quit [Ping timeout: 268 seconds]
<JaMa>
there just wasn't Yocto Project yet when PalmOS was actively developed :)
<tomzy_0>
LetoThe2nd well, looks like there is lot for me to learn in this world
<LetoThe2nd>
tomzy_0: for all of us.
<ArgaKhan>
If I compile from vscode about 5 seconds after this error, vscode crashes, if I compile from terminal, terminal crashes.
nemik_ has joined #yocto
onkelpit has joined #yocto
<LetoThe2nd>
ArgaKhan: compile from vscode? please explain your use case a little more. is this when you run bitbake? or when trying to use an sdk?
<ArgaKhan>
LetoThe2nd It happens when I use bitbake. When I say compiling from vscode, I edit vscode files and run bitbake from the integrated terminal.
amitk_ has joined #yocto
<ArgaKhan>
Sorry for sending you a message. I was trying to ping you like you ping me. I am newbie in this irc stuff.
<phako[m]>
ERROR: Layer meta-chimeratk depends on meta-oe and isn't found. -> I have added meta-oe using bitbake-layers add-layer - I don't really understand why yocto-check-layer keeps yelling at me
<d-fens_>
hi, anybody has some script laying around with the jq magic to extract unpached issues from the cve log?
<qschulz>
phako[m]: meta-oe is not a valid name for a layer IIRC
<qschulz>
if you're looking at expliciting the dependency on meta-openembedded/meta-oe, I believe it is openembedded-layer you're supposed to add
<phako[m]>
er ok. how would I find that out for any random layer?
<qschulz>
you look into BBFILE_COLLECTIONS in the conf/layer.conf
<phako[m]>
thanks!
nemik_ has quit [Ping timeout: 256 seconds]
nemik_ has joined #yocto
nemik_ has quit [Ping timeout: 268 seconds]
nemik_ has joined #yocto
EwelusiaGsiorek[ has quit [Quit: You have been kicked for being idle]
d-fens has joined #yocto
RobW has joined #yocto
d-fens_ has quit [Ping timeout: 268 seconds]
<rburton>
ArgaKhan: rm aborting sounds a lot like your disk is dead. check dmesg for errors.
rcw has quit [Ping timeout: 268 seconds]
manuel1985 has joined #yocto
<LetoThe2nd>
ArgaKhan: sorry, i had to leave for a few minutes. yeah see rburtons advice first, then you might need to be more specific e.g. which layers, and a log. I'm not aware of anything in core that would directly call rm that way (but i might of course be wrong)
invalidopcode has quit [Remote host closed the connection]
<rburton>
that's just bitbake purging old builds, but rm shouldn't crash.
invalidopcode has joined #yocto
<LetoThe2nd>
rburton: ok
<rburton>
thus, my theory that its getting I/O errors and aborting
florian has joined #yocto
Guest13 has joined #yocto
<Guest13>
Hi everyone. I'm having a problem with activating ov5645(coral camera) support. I opened a topic like this:
kpo has quit [Read error: Connection reset by peer]
manuel1985 has joined #yocto
kpo has joined #yocto
<qschulz>
Guest13: you might have to enforce a lower I2C frequency for the whole i2c2 bus since Omnivision sensor usually only support Full Speed (400KHz IIRC)
<qschulz>
Guest13: clock-frequency = <400000>;
<qschulz>
Guest13: then if that's fixed, then check if the camera appears on i2c by using i2cdetect -y 2 from userspace
<qschulz>
check the MCLK is running correctly and at the appropriate rate, check the power up sequence, etc...
<qschulz>
scope will be *really* helpful
<Salamandar>
Hey guys, I got a (custom) recipe that needs to install its files to a partition as a tarball (dd if=tarball of=mtd) (not my choice).
<Salamandar>
What's the best way to "require" this recipe into my build ?
<Salamandar>
Is it a good idea to add it to `IMAGE_INSTALL:append` even though it doesnt install anything into the rootfs ?
Estrella has quit [Ping timeout: 260 seconds]
<Salamandar>
I'm planning to use do_deploy to generate the tarball and put it in the deploy dir
Estrella has joined #yocto
<LetoThe2nd>
Salamandar: "partition as a tarball"? so what is the actual outcome? a binary image of the partition? a tarball? something in your root file system?
<Salamandar>
A tarball that will be dd-ed "raw" on the partition
<Guest13>
qschulz thank you, i will try.
<Salamandar>
so, a tarball :)
<LetoThe2nd>
Salamandar: sounds like you actually want a custom image type
thomasd13 has joined #yocto
<Salamandar>
maybe yeah. I've never written image recipes yet :D
<LetoThe2nd>
Salamandar: not image recipe. IMAGE_TYPE
<Salamandar>
ok, thank for the pointer. I'll read about it :)
<LetoThe2nd>
Salamandar: those can essentially define the form that your output has. we have tar.gz, we have ext4, ... wic is then a postprocessing step that you can also use.
<thomasd13>
Good morning guys, I'm still working on the "yocto-sdk gdb .so issue". I've noticed some warnings when building the sdk at package "external-arm-toolchain" which may seems related.
<thomasd13>
external-arm-toolchain-2019.12-r0 do_package_qa: QA Issue: external-arm-toolchain: /sbin/ldconfig is owned by uid 1000, which is the same as the user running bitbake. This may be due to host contamination [host-user-contaminated]
<LetoThe2nd>
Salamandar: from the descriptiion, you kinda want an IMAGE_TYPE .bin.tar :-)
<Salamandar>
LetoThe2nd: Interesting. So my recipe just installs normally in a specific folder, and the custom image type picks the files there ?
<thomasd13>
Can someone tell me where I can start my research about this? If that is really an issue in my case?
<LetoThe2nd>
Salamandar: if its only a directory, then you might need to also use wic. but thats a bit beyond me, sorry.
<thomasd13>
I've got alot "host contamination" warnings at this package
prabhakarlad has joined #yocto
<Salamandar>
yes it's only a directory, but it _has_ to be a tarball. My vendor un-tars the partition itself, so it's beyond my control.
<LetoThe2nd>
Salamandar: as mentioned then. IMAGE_TYPES and/or wic, possibly a combination thereof
<Salamandar>
Thank you ! :)
<qschulz>
Salamandar: also FYI, wic can create images with multiple partitions
<qschulz>
so you can dd the whole thing to the /dev/sdc for example and it'll flash the partition table and everything
<Salamandar>
Yes, that's what I would like to do once i manage to generate this partition
<Salamandar>
For now, the vendor's layer manually concatenates the images…
<jclsn>
Took me some time to find out that a defconfig in oe-local-files is used when building from the workspace...
manuel1985 has quit [Ping timeout: 265 seconds]
amitk_ has quit [Ping timeout: 246 seconds]
florian_kc has joined #yocto
aleksandarsimono has joined #yocto
amitk_ has joined #yocto
Frank48 has joined #yocto
alessioigor has quit [Quit: alessioigor]
alessioigor has joined #yocto
kpo has quit [Read error: Connection reset by peer]
kpo has joined #yocto
Guest49 has joined #yocto
<Guest49>
Hello all, i have a quick question. I want to have my image and my initramfs image using same DISTRO but one (image) using systemd and one (initramfs) using sysvinit.
<Guest49>
For that, i include DISTRO_FEATURES_append = " sysvinit systemd" in my distro.conf file
<Guest49>
In image recipe, i set VIRTUAL-RUNTIME_init_manager = "sysvinit" dor initramfs and VIRTUAL-RUNTIME_init_manager = "systemd" for image.
<Guest49>
Problem is that, if i check installed packages in buildhistory folder, i can see that systemd is installed on my initramfs image.
<Guest49>
How can i have sysvinit on my initramfs without installing systemd ?
<qschulz>
Guest49: the init system is a distro configuration, you need two different distros
<Guest49>
qschulz ok, but when i build my image, it automatically builds my initramfs image. If i need two distros, i need to review this process (INITRAMFS_IMAGE = "my-initramfs"), right ? Because if i'm not wrong, we cannot set a distro in image recipe
<qschulz>
Guest49: I assume this would be a case for a multiconfig setup which is definitely more complex than the typical septup
<qschulz>
I have never used it so can't say if that would fit the requirements but that's what comes to mind
<qschulz>
Guest49: but yes you're right, recipe data is local, configuration data is global. Changing the distro in the image will only apply to the image and no other recipe
<qschulz>
nice, didn't know we had doc for that :)
<mcfrisk>
we have all kinds of really cool docs nowadays, just need to keep the page open, refresh and search :)
<Guest49>
you are right. sorry for the inconvenience
<Guest49>
i will read the f.... manual. my apologies
_lore_ has quit [Remote host closed the connection]
Guest49 has left #yocto [#yocto]
_lore_ has joined #yocto
Guest2911 has joined #yocto
guest13215 has left #yocto [#yocto]
<qschulz>
mcfrisk: yeah "RTFM" can be a bit aggressive :/
Frank48 has quit [Quit: Client closed]
starblue has quit [Ping timeout: 260 seconds]
thomasd13 has quit [Ping timeout: 256 seconds]
Guest13 has quit [Quit: Client closed]
gho has quit [Quit: Leaving.]
starblue has joined #yocto
Frank47 has joined #yocto
gho has joined #yocto
gho has quit [Client Quit]
<mcfrisk>
qschulz: yea, meant to comment on that. I did not mean f...ing but FINE as a compliment to all who contribute to the docs. Sadly Guest49 left too fast, but hopefully will be back if anything is still unclear. Even after years with yocto I still don't know much and mostly just have the doc index page open and search there, as well as git grep..
<ArgaKhan>
As rburton said, I ran bitbake and when it crashed, I ran dmesg, but there was no log added.
<ArgaKhan>
I don't know exactly what meta it came from. It works from within meta-qt5/ecipes-qt/qt5. And the name of the package is qtmultimedia.
<ArgaKhan>
And sorry for the late reply, it was a lunch break.
gho has joined #yocto
rob_w_ has quit [Quit: Leaving]
aleksandarsimono has quit [Quit: Client closed]
PhoenixMage has quit [Ping timeout: 256 seconds]
Guest13 has joined #yocto
amitk__ has joined #yocto
<DvorkinDmitry>
may I set VOLATILE_LOG_DIR in the image file?
amitk_ has quit [Ping timeout: 256 seconds]
amitk__ has quit [Ping timeout: 256 seconds]
<Saur[m]>
DvorkinDmitry: No, `VOLATILE_LOG_DIR` needs to go in your configuration as it is used by multiple recipes.
Guest13100 has joined #yocto
Guest13 has quit [Ping timeout: 260 seconds]
Guest82 has joined #yocto
<DvorkinDmitry>
Saur[m], where can I put it instead of local.conf? in mymachine.conf ?
Frank47 has quit [Quit: Client closed]
Frank47 has joined #yocto
amitk_ has joined #yocto
<LetoThe2nd>
DvorkinDmitry: usual good places are either machine or distro, it depends a bit.
aleksandarsimono has joined #yocto
<Saur[m]>
DvorkinDmitry: It is typically a distro variable since if you change it, a couple of recipes will be built differently.
<Guest2911>
*** wkawka <wkawka!~wkawka@84-10-27-202.static.chello.pl> has joined #yocto 13:26
<Guest2911>
wkawka o/ 13:27
<Guest2911>
wkawka I am having a nice battle with xorg-xserver recipe. Before there were outdated patch files in meta-freescale which I have removed, but now meson has a problem with dependency, to be exact: 13:32
<Guest2911>
wkawka | Run-time dependency gl found: NO (tried pkgconfig and system) 13:32
<Guest2911>
wkawka | 13:32
<Guest2911>
wkawka | ../xorg-server-21.1.3/glx/meson.build:35:4: ERROR: Dependency "gl" not found, tried pkgconfig and system 13:32
<Guest2911>
wkawka | 13:32
<Guest2911>
wkawka | A full log can be found at /work/build-x11/tmp/work/cortexa9t2hf-neon-mx6qdl-poky-linux-gnueabi/xserver-xorg/2_21.1.3-r0/build/meson-logs/meson-log.txt 13:32
<Guest2911>
wkawka | ERROR: meson failed 13:32
<Guest2911>
wkawka | WARNING: exit code 1 from a shell command. 13:32
<ArgaKhan>
As rburton said, I ran bitbake and when it crashed, I ran dmesg, but there was no log added.
<ArgaKhan>
I don't know exactly what meta it came from. It works from within meta-qt5/ecipes-qt/qt5. And the name of the package is qtmultimedia.
jmk1 has quit [Quit: Leaving.]
<JaMa>
ArgaKhan: qtmultimedia doesn't do anything special which would cause this (unless you have .bbappend with something nasty in other layers), so I don't think it's specific to this recipe
<LetoThe2nd>
ArgaKhan: I am not using meta-qt(5), so i can't comment
aleksandarsimono has quit [Quit: Client closed]
<ArgaKhan>
JaMa How can I find out what is causing it? Is there a way?
<JaMa>
ArgaKhan: can you reproduce it every time or was it just one-of
<ArgaKhan>
LetoThe2nd I'm sorry, I seem to have posted it again, but while I was trying something, I sent the same article again. I'm not used to irc yet :D
<ArgaKhan>
JaMa Every time I run bitbake, it first gives this error and then I get a crash.
<JaMa>
ArgaKhan: did you check your disks (as rburton suggested before) and run fsck on that FS?
prabhakarlad has quit [Quit: Client closed]
<Guest2911>
I see: "Pending Moderation" on it...
Guest2911 has quit [Quit: Client closed]
Guest294 has joined #yocto
Guest294 is now known as PaolaMG
prabhakarlad has joined #yocto
<ArgaKhan>
JaMa I did what rburton said. I used dmesg command before and after running bitbake but even after crash no new entries were added to dmesg. now i tried fsck and got an output like this:
<ArgaKhan>
JaMa: I ran it and it gave the following output. Since the output is originally in Turkish, I converted the Turkish places to English:
<ArgaKhan>
sudo fsck -f /dev/sdb1
<ArgaKhan>
fsck on util-linux 2.37.2
<ArgaKhan>
1st pass: checking nodes, blocks and lengths
<ArgaKhan>
e2fsck 1.46.5 (30-Dec-2021)
<ArgaKhan>
Node 51065254 extent tree (at level 1) could be narrower. Optimize<y>? Yes
<ArgaKhan>
Node 51065255 extent tree (at level 1) could be narrower. Optimize<y>? Yes
<ArgaKhan>
Node 51065259 extent tree (at level 1) could be narrower. Optimize<y>? Yes
<ArgaKhan>
Node 51065261 extent tree (at level 1) could be narrower. Optimize<y>? Yes
<ArgaKhan>
Node 51396967 extent tree (at level 2) could be narrower. Optimize<y>? Yes
<ArgaKhan>
Pass 1E: Optimizing extent trees
<ArgaKhan>
Pass 3: Checking for directory connectivity
<ArgaKhan>
Pass 2: Checking the directory structure
<ArgaKhan>
Pass 4: Checking the application count
sakoman has joined #yocto
<mckoan>
ArgaKhan: please use pastebin
<ArgaKhan>
okay
<JaMa>
ArgaKhan: can you delete that directory manually from command line?
<ArgaKhan>
JaMa I tried, it was deleted. I'm running Bitbake again.
xmn has joined #yocto
<barath>
Hi all. Possibly a stupid question, but is it possible to remove ${PN} from PACKAGES for a recipe which builds a bunch of packages not named PN? When I remove PN from PACKAGES, I get the error that the recipe file still "RDEPENDS or otherwise requires it"
<ArgaKhan>
JaMa I didn't get any error this time, Thank you. But it still crashed.
<JaMa>
barath: yes you can, but then nothing can depend on it, the error message should show what depends on $PN package and you need to fix it there
<JaMa>
ArgaKhan: vscode crashed, right?
<ArgaKhan>
JaMa Yes, now I'm also trying from terminal.
<qschulz>
barath: I think ${PN}-dev RDEPENDS on ${PN} too
<qschulz>
barath: The easiest is probably to entirely redefine PACKAGES variable instead of only removing ${PN} from it?
<qschulz>
because I assume you' wouldn't want ${PN}-dev ${PN}-dbg and ${PN}-src either?
<JaMa>
why? most recipes without PN in PACKAGES still use other default packages and only issue I'm aware of is the need to set empty RDEPENDS:${PN}-dev (which causes issues only when PN-dev is installed in image with populate_sdk or e.g. PN-dev in IMAGE_FEATURES
CharlesKwiatkows has joined #yocto
<barath>
ah yes, that all makes sense, thanks. Pretty sure that's my general problem. After removing all of those default packages as well I don't have that dependency anymore! I should've realized that was the problem 🤪
<barath>
thanks all!
<JaMa>
where do you package /usr/src/debug files now?
<JaMa>
or .debug directories?
<barath>
those are supposed to go into other-package-dbg packages which are built by the recipe, so I was going to remove the ${PN}-dev, ${PN}-dbg etc packages anyways, I just didn't try that immediately
<barath>
JaMa: you're right, dhcp.inc for instance, which removes $PN from PACKAGES, seems to then set ${PN}-dev and ${PN}-staticdev to empty
<barath>
*sets those RDEPENDS to empty
<JaMa>
I was migrating from multiple -dbg packages in single repository to just PN-dbg a while ago (when automatic packageging of all .debug directories was added)
<barath>
yeah I guess I'm not sure whether we'll split them or not. some of the packages dont make sense to install on the same hardware in our case
<ArgaKhan>
JaMa I think I understood the problem, and the solution. The error I gave you crashes the first time, but when I run it for the second time, it prints that error on the screen. Now, for example, I did not get that error from qtmultimedia, but a different package gave that error. Likewise, I deleted that file manually and continue. If it is not as I thought, it will crash after a while without any error and then I will write it here, okay?
<ArgaKhan>
vscode or terminal doesn't matter.
<JaMa>
did you get another SIGABRT in rm when running in terminal without vscode being involved? I can imagine vscode crashing and taking rm child with it, but rm aborting in regular bitbake looks really strange
<qschulz>
OOM not happy with vscode running bitbake maybe?
<JaMa>
that should show up in dmesg as well
<JaMa>
not sure how vscode would close if it hits "ulimit -m"
<JaMa>
moto-timo: +
<ArgaKhan>
yes i bought it in another package. I'm currently running from the terminal. If you want to let me know the progress.
Frank47 has quit [Quit: Client closed]
Frank47 has joined #yocto
<ArgaKhan>
Ok, this was quick but now I'm running it from terminal and it crashes directly without any error and the window closes.
Frank47 has quit [Client Quit]
mattes-bru has joined #yocto
d-s-e has quit [Quit: Konversation terminated!]
mattes-b_ has joined #yocto
<JaMa>
ah you mean running vscode in terminal, not bitbake build in terminal?
mattes-bru has quit [Ping timeout: 260 seconds]
<ArgaKhan>
no. I am using the bitbake command directly in the terminal.
<JaMa>
sorry, I'm lost, what window closes then?
<JaMa>
and are you still talking about SIGABRT in rm or what kind of crash?
<ArgaKhan>
terminal window closes suddenly.
GNUmoon has quit [Ping timeout: 255 seconds]
kpo has quit [Read error: Connection reset by peer]
<ArgaKhan>
No, I don't get any errors anymore. It crashes while the processes are running. The window closes and does not show any errors.
kpo has joined #yocto
<JaMa>
run another bash inside that terminal window, so that the window doesn't close when bash ends, but it's still very strange
cp- has joined #yocto
<JaMa>
ArgaKhan: also check if there is something interesting at the end of bitbake-cookerdaemon.log
mattes-b_ has quit [Ping timeout: 256 seconds]
<ArgaKhan>
where is the bitbake-cookerdaemon.log ?
<JaMa>
TOPDIR
GNUmoon has joined #yocto
<ArgaKhan>
JaMa There is so much here (bitbake-cookerdaemon.log). Where would it be appropriate for me to send it to you? can you check it?
<JaMa>
empty this file, run it in child bash and if it fails again and you don't see the error, upload (now smaller) bitbake-cookerdaemon.log somewhere
<JaMa>
sorry, doesn't make much sense to me, why would bitbake close not only its bash (e.g. with set -e), but also the parent bash as well? I'm running out of ideas, but this definitely isn't normal, can you run it in console (without wm)?
<JaMa>
or in GNU screen session or something if it's the terminal window getting killed for whatever reason, not the bash itself
Herrie has quit [Ping timeout: 265 seconds]
<vvn>
Can a sub package ("${PN}-foo") package files already packaged by the main ${PN}?
<JaMa>
vvn: files are assigned to packages in the order how they are listed in PACKAGES variable
<Saur[m]>
No, the first FILES:<package> that matches wins.
<JaMa>
so if ${PN}-foo is before ${PN} in PACKAGES, then FILES:${PN}-foo will win over FILES:${PN}
<vvn>
OK so basically you cannot have FILES:${PN} = "file1" and FILES:${PN}-foo = "file1"
<JaMa>
you can and you often have as default FILES:* are relatively wide
<ArgaKhan>
JaMa I just thought of something that might be related to the problem. This build uses a lot of ram. I have 8 gb ram on my computer, I allocated 144 gb as swap. The build process uses 7.5gb of regular ram and 14gb of swap. then it kills and the terminal closes. Could it have something to do with it?
<JaMa>
that's why you usually have to prepend extra packages to default PACKAGES, not to append them
<JaMa>
ArgaKhan: OOMK would be shown in dmesg
<JaMa>
144g swap was a typo?
<vvn>
So if I want a file to be a two packages, I have no choice but to fork the pn.bb recipe, right?
<Saur[m]>
vvn: If you want the same file to appear in multiple packages, you would have to duplicate it in the do_install() step. This of course means you cannot have it in multiple packages in the same path.
<JaMa>
it cannot be in two packages at the same time
<ArgaKhan>
JaMa nope :D
<JaMa>
ArgaKhan: so you still had 130g of swap available, right?
<vvn>
Saur[m]: JaMa: let's say I want a variant of the btrfs-tools package without python support for usage in the initrd (without the hassle of using a different distro/multiconfig), I must add a btrfs-tools-initrd_1.2.3.bb package, with require recipes-devtools/btrfs-tools/btrfs-tools_${PV}.bb and RCONFLICTS:${PN} = "btrfs-tools". Correct?
<vvn>
(as well as the PACKAGECONFIG tweaks obviously)
<JaMa>
ArgaKhan: I don't see anything wrong in dmesg, you can also check tmp-glibc/log/cooker/MACHINE-YOU-BUILD/console-latest.log
<ArgaKhan>
xfce4-full-image was skipped: 'x11-base' in IMAGE_FEATURES is not a valid image feature. Valid features: allow-empty-password allow-root-login bash-completion-pkgs dbg-pkgs debug-tweaks dev-pkgs doc doc-pkgs empty-root-password lic-pkgs overlayfs-etc package-management post-install-logging ptest-pkgs qtcreator-debug read-only-rootfs read-only-rootfs-delayed-postinsts splash src-pkgs stateless-rootfs staticdev-pkgs
Frank66 has joined #yocto
<Saur[m]>
vvn: I don't know anything about btrfs-tools, so I do not know what happens when you enable the "python" PACKAGECONFIG for it. However, if all that config does is install more tools, then you could package them in a separate package, e.g., ${PN}-python. Then have the main package recommend the ${PN}-python package. That way you can avoid installing it by setting a BAD_RECOMMENDATION.
<ArgaKhan>
JaMa but i alredy use this command: bitbake core-image-minimal-xfce
<ArgaKhan>
this error is old
<vvn>
Saur[m]: I understand your point for that particular use case, and indeed that'd be the way to go for such package. But let's take an example of a package foo which embeds the support as part of the binary, not extra files. Would I need to do what I described with a foo-bar_1.2.3.bb recipe?
rstreif has joined #yocto
prabhakarlad has quit [Quit: Client closed]
prabhakarlad has joined #yocto
tomzy_0 has quit [Quit: Client closed]
Frank66 has quit [Quit: Client closed]
tor has quit [Quit: Leaving]
kpo has quit [Read error: Connection reset by peer]
prabhakarlad has quit [Quit: Client closed]
kpo has joined #yocto
alessioigor has quit [Quit: alessioigor]
florian_kc has quit [Ping timeout: 268 seconds]
florian has quit [Quit: Ex-Chat]
<Saur[m]>
Well, if you have a recipe that produces a binary that you want to be different for some images, you would other have to do it through a distro feature (which seems a bit excessive), make the recipe build multiple binaries that can the be packaged separately or use multiple recipes.
<Saur[m]>
vvn: ^^^^
jmk1 has joined #yocto
Guest55 has joined #yocto
nemik_ has quit [Ping timeout: 256 seconds]
vvn has quit [Quit: WeeChat 3.7.1]
vvn has joined #yocto
Guest55 has quit [Quit: Ping timeout (120 seconds)]
Guest55 has joined #yocto
nemik_ has joined #yocto
<vvn>
Saur[m]: got it. The context was that I want less dependencies and smaller binaries in the initrd and while this is easily done with another distro/multiconfig, it doesn't rely justify the hassle for it. A few package improvements seem just enough to provide a simple initramfs image with adding extra complexity to the existing build.
<vvn>
without*
gho has quit [Quit: Leaving.]
demirok has joined #yocto
Guest41 has joined #yocto
leon-anavi has quit [Quit: Leaving]
mckoan is now known as mckoan|away
jmk1 has left #yocto [#yocto]
zpfvo has quit [Quit: Leaving.]
Guest55 has quit [Quit: Client closed]
Guest41 has quit [Quit: Connection closed]
yann has quit [Ping timeout: 268 seconds]
Frank75 has joined #yocto
vmeson has quit [Ping timeout: 252 seconds]
Frank75 has quit [Quit: Client closed]
Starfoxxes has joined #yocto
vmeson has joined #yocto
prabhakarlad has joined #yocto
mattes-bru has joined #yocto
florian_kc has joined #yocto
invalidopcode has quit [Remote host closed the connection]
invalidopcode has joined #yocto
alessioigor has joined #yocto
<vvn>
humm the usage of RRECOMMENDS is confusing. It mixes the notion of availability (like a kernel module either built-in or packaged) and the notion of extension for extra features.
ErRandir has joined #yocto
alessioigor has quit [Quit: alessioigor]
<vvn>
Should we build empty packages for built-in modules and start using RDEPENDS instead of RRECOMMENDS for kernel-module-*?
florian_kc has quit [Ping timeout: 260 seconds]
<ErRandir>
To get around a firewall (and save bandwidth) I've set FETCHCMD_git to do an rsync from another internal server, but when I build it always reports "no output". I think it's cause by the progress checker that it runs. Making rsync more verbose does not help. Any ideas how I can make it work?
<JaMa>
why not use PREMIRROR instead of rsync?
prabhakarlad has quit [Quit: Client closed]
<ErRandir>
When I've configued that it always tries to get a tarball, and does not attempt a git clone.
<JaMa>
yes you need to enable the other internal server to generate tarballs
<JaMa>
but then you get everything from the PREMIRROR in nice controlled way, not just git recipes through hacked FETCHCMD_git
mattes-bru has quit [Ping timeout: 252 seconds]
<ErRandir>
At the moment the other server does not have any of the tools installed, but I can give that a try. Thanks!
mark_ has joined #yocto
mark_ is now known as mark_l
sakoman has quit [Quit: Leaving.]
<mark_l>
I brought in the dhcp 4.4.1 recipe in our own meta layer which should have the highest override priority, the recipe requires dhcp.inc that has PACKAGES += "dhcp-libs dhcp-server dhcp-server-config dhcp-client dhcp-relay dhcp-omshell", but in build the build system still fallback to use dhcp-relay_4.4.3.bb from meta-openembedded. Any insights what am I missing?
florian_kc has joined #yocto
kpo has quit [Read error: Connection reset by peer]
kpo has joined #yocto
amitk has quit [Ping timeout: 256 seconds]
amitk has joined #yocto
sakoman has joined #yocto
<mark_l>
My coworker advised me that I should wipe out my build directory after switching package provider. Trying that now
manuel1985 has joined #yocto
Vonter has quit [Quit: WeeChat 3.7.1]
Minvera has joined #yocto
amitk has quit [Ping timeout: 260 seconds]
mattes-bru has joined #yocto
<Saur[m]>
mark_l: The order of the layers in your bblayers.conf file makes a difference in this case. You need to make sure your layer is listed before meta-networking.
mattes-bru has quit [Ping timeout: 268 seconds]
<mark_l>
Saur: Yes, I checked that, our layer is ahead of that. It's still trying to build the meta-oe one after I wiped the tmp directory. Should I add RPROVIDES:${PN} += "${PN}-relay" to the recipe, I tought having that in "PACKAGES" should be enough
florian_kc has quit [Ping timeout: 256 seconds]
<mark_l>
On a side note, the same recipes work fine under Dunfell, as the dunfell version of meta-oe didn't have dhcp-relay recipe. Do I have to split the dhcp recipe to dhcp-relay.bb to override it?
onkelpit has quit [Quit: leaving]
<mischief>
when i run bitbake -u ncurses i just get a traceback with "bb.BBHandledException". is this expected?
<vvn>
what's the best way to customize fstab per image?
sakoman has quit [Quit: Leaving.]
florian_kc has joined #yocto
kiwi_29_[m] has joined #yocto
sakoman has joined #yocto
<kiwi_29_[m]>
Hello, I use go toolchain version 1.14 from pok(meta/recipes-devtools/go). My whole project is based on dunfell. Now, the changes to my main app written in go requires me to upgrade go to 1.19, which is available in poky on branch langdale. This would mean I will have to checkout langdale tag for the whole poky repo. To avoid this, I created recipes-devtools/go folder in my custom layer and copied go 1.19 recipes from langdale tag of poky.
<kiwi_29_[m]>
Made appropriate changes based on the errors I received when I did "bitbake go". After all the errors were taken care of, I see that running bitbake go compiles go version 1.14 from meta/recipes-devtools/go instead of mycustomlayer/recipes-devtools/go . what am I missing here
<kiwi_29_[m]>
mycustomlayer has higher priority than meta and also go recipe in mycustomlayer has higher version
florian_kc has quit [Ping timeout: 256 seconds]
Guest41 has joined #yocto
Guest41 has quit [Client Quit]
mvlad has quit [Remote host closed the connection]
Minvera has quit [Remote host closed the connection]
<kiwi_29_[m]>
ok..here is the solution : I found that meta layer has PREFERRED_PROVIDER_go defined in one of the files . That basically led yocto to compile go version 1.14. I overrode it by writing PREFERRED_PROVIDER_go = "1.19.3" in my local.conf. With this change, my latest 1.19.3 version of go is compiling