bluelightning has quit [Remote host closed the connection]
bluelightning has joined #yocto
camus1 has joined #yocto
camus has quit [Ping timeout: 264 seconds]
camus1 is now known as camus
ecdhe has quit [Ping timeout: 246 seconds]
rber|res has joined #yocto
ecdhe has joined #yocto
RobertBerger has quit [Ping timeout: 258 seconds]
ecdhe has quit [Read error: Connection reset by peer]
ecdhe has joined #yocto
ecdhe_ has joined #yocto
Net147 has quit [Ping timeout: 260 seconds]
ecdhe has quit [Ping timeout: 260 seconds]
Net147 has joined #yocto
Net147 has joined #yocto
Net147 has quit [Changing host]
camus has quit [Ping timeout: 264 seconds]
camus has joined #yocto
<roussinm>
Is there a documented PEG for bitbake?
<roussinm>
Wondering if I could write a language server for it, maybe the easiest would be to bake it inside the bitbake repository...
<roussinm>
s/bake/write/
camus1 has joined #yocto
camus has quit [Ping timeout: 264 seconds]
camus1 is now known as camus
ecdhe_ has quit [Read error: Connection reset by peer]
ecdhe has joined #yocto
pgowda_ has joined #yocto
troth has quit [Ping timeout: 264 seconds]
troth has joined #yocto
Tokamak has joined #yocto
rmmr has quit [Ping timeout: 258 seconds]
rmmr has joined #yocto
sakoman has quit [Ping timeout: 260 seconds]
fray has quit [Ping timeout: 265 seconds]
fray has joined #yocto
sakoman has joined #yocto
mranostaj has quit [Remote host closed the connection]
mranostaj has joined #yocto
davidinux has joined #yocto
prabhakarlad has quit [Quit: Client closed]
Schlumpf has joined #yocto
alessioigor has joined #yocto
alessioigor has quit [Client Quit]
Tokamak has quit [Ping timeout: 264 seconds]
rob_w has joined #yocto
camus1 has joined #yocto
camus has quit [Ping timeout: 260 seconds]
camus1 is now known as camus
roussinm has quit [Quit: WeeChat 3.3-dev]
leon-anavi has joined #yocto
<banana_smoothie[>
ERROR: Multiple versions of python3-native are due to be built (virtual:native:/home/poky/meta-custom/recipes-devtools/python/python3_3.9.6.bb virtual:native:/home/poky/meta/recipes-devtools/python/python3_3.8.11.bb). Only one version of a given PN should be built in any given build. You likely need to set PREFERRED_VERSION_python3-native to select the correct version or don't depend on multiple versions.
<manuel1985>
moto-timo, took a look on your kas snippet https://www.irccloud.com/pastebin/evPnFvan/ now, looks good! Exactly how I assumed it to interact. Good to see it in a real world example.
MWelchUK has joined #yocto
zpfvo has quit [Ping timeout: 264 seconds]
zpfvo has joined #yocto
zpfvo has quit [Ping timeout: 260 seconds]
zpfvo has joined #yocto
<qschulz>
tlwoerner: we need a new honister branch on meta-rockchip I guess :)
<qschulz>
tlwoerner: is there any particular reason for not having one branch for >= dunfell BTW?
goliath has joined #yocto
<Guest47>
RP good, thanks!
zpfvo has quit [Ping timeout: 268 seconds]
zpfvo has joined #yocto
bps2 has joined #yocto
bps has quit [Ping timeout: 264 seconds]
<tlwoerner>
qschulz: oops. i usually create branches for all releases
bps2 has quit [Remote host closed the connection]
bps2 has joined #yocto
zpfvo has quit [Ping timeout: 258 seconds]
zpfvo has joined #yocto
vicale has joined #yocto
zpfvo has quit [Ping timeout: 260 seconds]
zpfvo has joined #yocto
chrfle has quit [Ping timeout: 260 seconds]
<tlwoerner>
qschulz: oh wait, why do we need a branch for honister? meta-rockchip builds fine against oe-core master
<tlwoerner>
is this the u-boot thing? or the openssl thing?
<tlwoerner>
qschulz: A) mostly resources. i build-test against master and LTS every night for every machine. i don't have the resources to test every combination of machine vs 3 oe-core branches to ensure that everything >= dunfell builds on just one meta-rockchip branch
wwilly has quit [Quit: Leaving]
camus has quit [Ping timeout: 260 seconds]
<tlwoerner>
B) i think it's confusing for users. let's say a user wants to build gatesgarth. the easiest thing to tell them is "checkout gatesgarth on every layers you want to use and you're good to go"
camus has joined #yocto
<tlwoerner>
but instead many BSP layers don't create branches for every oe-core release, they only create branches when they have to so now the simple instruction from above has to be amended to say something like:
zpfvo has quit [Ping timeout: 260 seconds]
<tlwoerner>
"for every layer maybe try master if it doesn't have a <branch> branch, or maybe look for the last branch before <branch> and try that?"
zpfvo has joined #yocto
<tlwoerner>
if meta-rockchip's gatesgarth layer doesn't have a feature or machine that they want from master, they're welcome to backport it, test it, and submit a pull request
<tlwoerner>
and that's something an advanced user can do very easily, but trying to explain obtuse branching to a person new to OE/YP will just confuse and discourage them
zpfvo has quit [Ping timeout: 260 seconds]
<tlwoerner>
so that's why, contrary to others, i believe it's a good practice to create branches that mirror oe-core. but i do understand the other use-case
zpfvo has joined #yocto
jmiehe has joined #yocto
zpfvo has quit [Ping timeout: 258 seconds]
zpfvo has joined #yocto
<qschulz>
tlwoerner: we need a honister branch exactly for the arguments you exposed above :) There's a honister release available and master oe is now for 3.5 (kirkstone IIRC?)
<qschulz>
tlwoerner: meta-swupdate is doing that IIRC, but they actively maintain each branch
<qschulz>
basically backporting every patch from their master branch to other branches
<qschulz>
even on EOL oe branches :o!
<qschulz>
anyway, was just wondering, I'd have seen actively maintained dunfell and hardknott/honister branches or a single branch compatible with all branches. But I don't have maintainership experience :)
ecdhe has quit [Ping timeout: 264 seconds]
camus has quit [Ping timeout: 260 seconds]
zpfvo has quit [Ping timeout: 264 seconds]
<qschulz>
tlwoerner: basically, I think I'll send a new machine for meta-rockchip but I probably want to support dunfell in addition to honister
camus has joined #yocto
zpfvo has joined #yocto
<qschulz>
if you're open to contributions to older branches, that'd be awesome :) otherwise I'll just maintain this on my own :)
ecdhe has joined #yocto
<tlwoerner>
meta-rockchip is open to all contributions :-)
<tlwoerner>
am i using the wrong openembedded-core? (git://git.openembedded.org/openembedded-core.git) my oe-core/meta/conf/layer.conf still says 'LAYERSERIES_CORENAMES = "honister"'
<RP>
tlwoerner: master hasn't moved forward yet but it will... :)
<RP>
tlwoerner: i.e. nothing incompatible happened as yet
<RP>
you could say the openssl/python changes are enough to move that to kirkstone actually
<qschulz>
tlwoerner: looking forward to sending you a patch for adding our SoM to meta-rockchip and using upstream as our BSP :)
<qschulz>
though I need to complain already... there are warnings :p
<tlwoerner>
qschulz: ah that. that has nothing to do with 3.4 vs 3.5
<qschulz>
tlwoerner: there was no justification behind 3.4 vs 3.5 except thatr you have a branch per release and there is a new release :)
ecdhe has quit [Ping timeout: 260 seconds]
<tlwoerner>
that's because we're using the in-kernel default arch/arm64/configs/defconfig to configure our arm64 kernels
<tlwoerner>
and the default, in-kernel defconfigs are… (hmm… how would zeddii put this…)
<tlwoerner>
(haha lol)
<qschulz>
for aarch64 it's not great to say the least :)
<tlwoerner>
on the one hand there's way too much stuff enabled, and on the other hand many rockchip boards are missing features that should be enabled
<tlwoerner>
zeddii has done a great job laying the ground-work for generating really amazing, properly tuned kernel configs
<tlwoerner>
i'm working on extending that into meta-rockchip so we can have nice kernel configs too :-)
<qschulz>
:)
<tlwoerner>
as an example, the default in-kernel defconfig doesn't work for the roc-rk3328-cc, so it doesn't boot out of the box
<tlwoerner>
i could just slap a tweaked defconfig in there which would allow me to add that board to the BSP, but it would be much nicer to add once the new kernel config stuff is in
<tlwoerner>
wow! a new raspberry pi zero is coming out
<tlwoerner>
rpi zero v1 had a bcm2835 which is an ARM11 generation
<tlwoerner>
rpi zero v2 will have a bcm2710 which is an arm cortex-a53 64-bit
<tlwoerner>
that's quite the upgrade!
<qschulz>
yup
<qschulz>
though I'm not sure about the choice of requiring a 2.5A power supply over micro-USB
* RP
is remembering why he dislikes trying to change/improve/fix bits of the gcc recipes
Schlumpf has quit [Ping timeout: 256 seconds]
kiran has joined #yocto
camus1 has joined #yocto
camus has quit [Remote host closed the connection]
camus1 is now known as camus
<Guest47>
When I build yocto sdk the toolchain gets named gnueabi. Any way to change that? SDK_VENDOR does not help.
<banana_smoothie[>
RP rburton A python3 on a 64 arch can read a third-party module, but a python3 on a 32-bit arch can't. A self built python3 on the same 32-bit arch also can read that third-party package... Could you explain me why a 32-bit arch cannot read that third-party package that a 64-bit arch or a simple built python3 on that same target can?
pgowda_ has quit [Quit: Connection closed for inactivity]
ecdhe has joined #yocto
troth has quit [Ping timeout: 258 seconds]
Schlumpf has joined #yocto
<MWelchUK>
Hmm...
<MWelchUK>
`IMAGE_QA_COMMANDS` run from `do_image_qa`, which is added with:
<MWelchUK>
`IMAGE_PREPROCESS_COMMAND` run from `do_image`, which is added with:
<MWelchUK>
addtask do_image after do_rootfs
<MWelchUK>
addtask do_image_qa after do_rootfs before do_image
<MWelchUK>
So the Image QA tests are performed prior to the preprocessing (let alone the post processing) which could easily alter the contents of the image...
codavi has joined #yocto
rob_w has quit [Quit: Leaving]
zpfvo has quit [Ping timeout: 260 seconds]
zpfvo has joined #yocto
troth has joined #yocto
argonautx has joined #yocto
ar__ has joined #yocto
zpfvo has quit [Ping timeout: 260 seconds]
zpfvo has joined #yocto
camus has quit [Ping timeout: 260 seconds]
troth has quit [Ping timeout: 245 seconds]
camus has joined #yocto
codavi has quit [Ping timeout: 260 seconds]
codavi has joined #yocto
ar__ has quit [Read error: Connection reset by peer]
ar__ has joined #yocto
chrfle has joined #yocto
ar__ has quit [Read error: Connection reset by peer]
ar__ has joined #yocto
codavi has quit [Ping timeout: 264 seconds]
ar__ has quit [Read error: Connection reset by peer]
<RP>
MWelchUK: you'd hope that things changing content would be part of the rootfs process but we probably don't enforce that like we should
<RP>
banana_smoothie[: I have no idea offhand, you'd just need to debug it :(
vd has quit [Quit: Client closed]
tre has quit [Remote host closed the connection]
vd has joined #yocto
<banana_smoothie[>
RP: I tried without success. Could you help me the debug if I attach the details?
<RP>
banana_smoothie[: I can take a look at details you share but I don't really have the time to step you through debugging it :(
manuel1985 has quit [Quit: Leaving]
Schlumpf has quit [Ping timeout: 256 seconds]
Tokamak has joined #yocto
jmiehe has quit [Quit: jmiehe]
roussinm has joined #yocto
camus has quit [Ping timeout: 260 seconds]
camus has joined #yocto
zpfvo has quit [Ping timeout: 260 seconds]
zpfvo has joined #yocto
zpfvo has quit [Ping timeout: 258 seconds]
zpfvo has joined #yocto
RobW has quit [Quit: Leaving]
zpfvo has quit [Ping timeout: 258 seconds]
zpfvo has joined #yocto
JJ14 has joined #yocto
<JJ14>
Hey folks, I've got a question, I've got a variable in local.conf that points to a file in my build environment, can I add the content of the file the path points to to the input checksum of a task?
<JJ14>
that is in my local.conf I've got VARIABLE = "/path/to/file" and I'd like to rerun a task if the content of "file" changes
<qschulz>
JJ14: [file-checksums] varflag for the task that needs to be rerun
<qschulz>
and you add your path to there
<qschulz>
-to
<qschulz>
e.g. `do_compile[file-checksums] += "${VARIABLE}:True"` IIRC
<qschulz>
if I read externalsrc.bbclass correctly
<qschulz>
might not need the :True
<JJ14>
alright thank you I'll play around with that one a bit
<RP>
qschulz: you need the :True
<RP>
(it represents whether the file exists)
<RP>
(we have to track files which now exist that didn't too)
<JJ14>
the file might or might not exist though, which depends on the build environment
<JJ14>
same for the variable it might not be defined on other systems
<RP>
JJ14: then you need to set file-checksums accordinly
<qschulz>
JJ14: then you'll need some inline python to set this value accordingly I guess? a simple os.path.exists(d.getVar(VARIABLE)) ? ":True" : ":False" ?
nerdboy has quit [Ping timeout: 260 seconds]
<qschulz>
ah, ternary does not exist in python, then the ":True" if d.getVar(VARIABLE) else ":False" or something like that, you'll figure something out :)
<JJ14>
most likely I've got some good pointers to work with now
zpfvo has quit [Ping timeout: 260 seconds]
zpfvo has joined #yocto
zpfvo has quit [Ping timeout: 258 seconds]
zpfvo has joined #yocto
nerdboy has joined #yocto
nerdboy has quit [Changing host]
nerdboy has joined #yocto
dev1990 has joined #yocto
<jaskij[m]>
is `${D}` removed or otherwise cleaned between `do_compile` and `do_install`?
roussinm1 has joined #yocto
florian has quit [Quit: Ex-Chat]
zpfvo has quit [Ping timeout: 260 seconds]
zpfvo has joined #yocto
florian_kc has quit [Ping timeout: 265 seconds]
dev1990 has quit [Quit: Konversation terminated!]
zpfvo has quit [Ping timeout: 260 seconds]
zpfvo has joined #yocto
rfuentess has quit [Remote host closed the connection]
zpfvo has quit [Ping timeout: 265 seconds]
mckoan is now known as mckoan|away
zpfvo has joined #yocto
leon-anavi has quit [Quit: Leaving]
Guest47 has quit [Quit: Client closed]
zpfvo has quit [Ping timeout: 252 seconds]
stefan_schmidt has quit [Quit: Client closed]
dev1990 has joined #yocto
JJ14 has quit [Quit: Client closed]
<dev1990>
hi, for some time I have problems with pseudo on my host machine
<dev1990>
RP: should yocto use my host findutils in do_install?
michaelo_ has quit [Quit: Lost terminal]
camus has quit [Ping timeout: 265 seconds]
camus has joined #yocto
<dev1990>
I undestand correctly when compiling findutils with -D_FILE_OFFSET_BITS=64 (on my host)
<dev1990>
I create find that is not compatible with pseudo somehow, difference in structure(?) that is used by some calls
<dev1990>
this might be future bug for yocto/oe, if more distros will use -D_FILE_OFFSET_BITS=64 for findutils
fitzsim has quit [Read error: Connection reset by peer]
vd has quit [Quit: Client closed]
vd has joined #yocto
<vd>
what is the maximum number of overlayfs layers?
fitzsim has joined #yocto
florian_kc has joined #yocto
babb0 has joined #yocto
<abelloni>
RP: should we rename the oe-core layer facebook ?
<RP>
dev1990: yes, we use find from the host. Tracking it down to a util helps a lot
akiCA has joined #yocto
<RP>
dev1990: we need to track down what this means in terms of function calls
<marc1>
hi, looking for advices on bug 14375. Not really sure how I can fix it. I'm currently providing a different serial-getty@.service file which reverts to the default systemd settings. Any tips appreciated!
<RP>
abelloni: I don't understand
florian_kc has quit [Ping timeout: 260 seconds]
<babb0>
Hi, I have a question about config fragments. While upgrading from dunfell to honister I have problem with some kernel config options that won't apply from fragment. One example is CONFIG_NETFILTER_ADVANCED. It is enabled by default, I disable it in menuconfig, do diffconfig to get the fragment and put it in SRC_URI. Then running clean and menuconfig it is still enabled. Same issue with some other options while others seems to work as
<babb0>
expected. Any ideas about this? Workarounds?
<zeddii>
RP: I heard you had joined Facebook, so we rebranding.
<zeddii>
s/we/were/
<RP>
zeddii: I joined facebook?
<jaskij[m]>
are you guys making fun of the new name?
<zeddii>
heh. no. Just making up rumours.
<zeddii>
jaskij[m], yes.
<jaskij[m]>
RP: FB is rebranding and their new name is "Meta"
<zeddii>
RP Just facebook's new "meta" name for the overally company
<zeddii>
and hence oe-core/meta is in the line of fire!
<kanavin_>
so they officially admit the facebook brand is tainted?
Tokamak has quit [Ping timeout: 264 seconds]
<RP>
zeddii: ah :)
<vd>
but the layer directory is "meta" but its name really is "core"
* vd
is just nitpicking
camus1 has joined #yocto
camus has quit [Read error: Connection reset by peer]
camus1 is now known as camus
behanw has joined #yocto
<RP>
dev1990: are you on a 32 bit system?
<dev1990>
64bit
Tokamak has joined #yocto
florian_kc has joined #yocto
mvlad has quit [Remote host closed the connection]
kiran has quit [Ping timeout: 245 seconds]
<RP>
dev1990: I'm trying compiling findutils with and without that option, then comparing the function list of functions that changed with those in pseudo
<RP>
dev1990: list looks like fstatfs64 lstat64@GLIBC_2.33 stat64@GLIBC_2.33 fstatat64@GLIBC_2.33 fopen64 openat64@GLIBC_2.4 getrlimit64 fseeko64 open64 fcntl64@GLIBC_2.28 readdir64 lseek64 fstat64@GLIBC_2.33
<RP>
dev1990: with the most likely being the 2.33 ones
<RP>
dev1990: my guess is fcntl64
<RP>
dev1990: so we need to figure out a wrapper in pseudo for that
<RP>
tlwoerner: did you want a hopefully relatively easy pseudo issue? :)