vthor has quit [Remote host closed the connection]
jmiehe has quit [Quit: jmiehe]
xmn has quit [Ping timeout: 268 seconds]
GNUmoon2 has quit [Remote host closed the connection]
GNUmoon2 has joined #yocto
davidinux1 has quit [Ping timeout: 268 seconds]
davidinux1 has joined #yocto
chep has quit [Read error: Connection reset by peer]
chep has joined #yocto
jclsn has quit [Ping timeout: 255 seconds]
tec has quit [Quit: bye!]
jclsn has joined #yocto
tec has joined #yocto
mbulut has joined #yocto
ablu has quit [Ping timeout: 255 seconds]
ablu has joined #yocto
Daanct12 has quit [Read error: Connection reset by peer]
Daanct12 has joined #yocto
absurdist has quit [Remote host closed the connection]
absurdist has joined #yocto
Xagen has joined #yocto
Xagen has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
amitk has joined #yocto
jmd has joined #yocto
Minvera has quit [Ping timeout: 268 seconds]
<mbulut>
on kirkstone i have a recipe that only inherits setuptools3 to fetch and build a python package from a private git repo. the package uses setuptools-scm for dynamic versioning. when i query the version at runtime using importlib.metadata.version, i only get '0.0.0' which seems like the fallback version having kicked in... building the package standalone and querying using same method i do get a proper version, so in principle it
<mbulut>
seems to work. any idea what could be the reason why it's not working when built in yocto?
<mbulut>
i'm not fetching the sources as tarball but via git, so i'd expect setuptools-scm to have the git repo available at build time
vthor has joined #yocto
<mbulut>
oh, i found the problem... i had python3-setuptools-scm in DEPENDS, where it should be python3-setuptools-scm-native
nerdboy has quit [Ping timeout: 246 seconds]
leon-anavi has joined #yocto
rob_w has joined #yocto
rfuentess has joined #yocto
sukbeom has quit [Quit: Ping timeout (120 seconds)]
sukbeom has joined #yocto
nerdboy has joined #yocto
nerdboy has quit [Changing host]
nerdboy has joined #yocto
frieder has joined #yocto
jmd has quit [Remote host closed the connection]
luc4 has joined #yocto
jmd has joined #yocto
mckoan_ has joined #yocto
mckoan|away has quit [Ping timeout: 264 seconds]
enok has joined #yocto
alessioigor has joined #yocto
<khem>
yep
<khem>
these are build time ops which means you need packages on the build machine. Hence the native packages
enok has quit [Read error: Connection reset by peer]
<Ad0>
I have my own program which I want to include in the distro. I wonder if I should do a git pull from another repository or just include it in the yocto recipes. what's the most usual way (given that the program is small and not part of another solution) ?
prabhakalad has quit [Quit: Konversation terminated!]
prabhakalad has joined #yocto
<LetoThe2nd>
Ad0: not sure I understand the question, but in a nutshell it is "I have this program which I want to add to my build", right? Then the answer is as usual: write a recipe, add it to your layer.
PaowZ has joined #yocto
geoffhp has quit [Quit: Leaving]
<Ad0>
LetoThe2nd, thanks. I am thinking if I should just include it directly and maintain it in the yocto repository or pull it from a seaparate repository
<LetoThe2nd>
Ad0: "in the yocto repository", you mean somewhere in poky? then the answer is a clear no. just think about how you would manage the resulting sources.
<LetoThe2nd>
Ad0: or is it "do I add the sources to my layer, instead of a separate repository"? then the answer is: technically possible, but I've found that cumbersome for two reasons: the source code changes clutter your git log, and the day will come when you want to use the tool outside a yocto context. then its really hard to access if its hidden in a layer repo.
ablu has quit [Ping timeout: 255 seconds]
florian has joined #yocto
enok has joined #yocto
<Ad0>
thanks!
<Ad0>
yes I mean in my layer
<Ad0>
ok external repository it is
<Ad0>
oh and in case I have multiple programs, should they be in separate repositories or should I have them in one? Then I would probably need to do sparse clones
<mckoan>
I have a recipe managing a custom library I created. It installs the .so correctly in the final image and in the SDK when I build with populate_sdk. Hoever in the SDK I don't have the .h provided by the library. Do I need any particular setting to add .h to the SDK ?
jmd` has joined #yocto
<KanjiMonster>
Ad0: do the programs are related to each other/need to be kept in sync? If yes then a single repository might make sense. A recipe can provide multiple packages, so you can just split them up within the recipe and don't need multiple recipes
<Ad0>
ok
<Ad0>
thanks
<Ad0>
I will collect them in one recipe
florian has quit [Quit: Ex-Chat]
leon-anavi has quit [Ping timeout: 246 seconds]
jmd` has quit [Remote host closed the connection]
<rburton>
RP zeddii: FYI, binutils master is advertising relocation support which means the kernel turns on CONFIG_RELR but on arm64 the relocation is broken so you end up with kernels that don't actually work. toolchain people are fixing binutils but just a headsup if you try that combination of pieces for any reason.
<jstephan_>
hello all! I would like to change a variable defined in bitbake.conf to append the PV of a given recipe. I tried to register an handler inside sanity.conf that would register itself on bb.event.ParseCompleted
<jstephan_>
my handler is called, I can see debug output, but when using bitbake -e I don't see the change on the variable
<Ad0>
also how do I add my own environment variables properly?
<rburton>
jstephan_: probably because you're hitting the wrong datastore in your handler. what do you actually want to do?
enok has quit [Read error: Connection reset by peer]
<jstephan_>
rburton: I found a bug in kernel.bclass do_symlink_kernsrc function when bsp supports several kernel version, because kernel sources are moved (and symlinked) to a shared place (STAGING_KERNEL_DIR)
<jstephan_>
so first compilation of let's say 5.15 works, first compilation of let's say 6.6 is ok, but when switching back to 5.15, the symlink of S point to STAGING_KERNEL_DIR that contains 6.6 source code ...
<jstephan_>
so my idea was to create kernel-source-5.15 and kernel-source-6.6 inside STAGING_DIR and when Parse is complete, set STAGING_KERNEL_DIR to either kernel-source-5.15 or kernel-source-6.6
<rburton>
using event handlers seems like a terrible fragile way of doing that
<jstephan_>
ok :)
leon-anavi has joined #yocto
<rburton>
zeddii: ^^^
MrCryo has joined #yocto
<jstephan_>
I was also thinking of creating a symlink from kernel-source to the correct kernel-source-xxx to not change STAGING_KERNEL_DIR, but I guess the event handler is also a bad way to do this
mbulut has quit [Ping timeout: 240 seconds]
sakoman has quit [Ping timeout: 246 seconds]
<RP>
rburton: thanks for the warning!
eminboydak has joined #yocto
eminboydak has quit [Client Quit]
tec5 has joined #yocto
tec has quit [Ping timeout: 256 seconds]
tec5 is now known as tec
florian has joined #yocto
eminboydak has joined #yocto
<rburton>
RP: can you pick 'setuptools: drop 2to3' into next?
sakoman has joined #yocto
jmd` has joined #yocto
tec has quit [Quit: bye!]
tec has joined #yocto
Daanct12 has quit [Quit: WeeChat 4.3.2]
goliath has quit [Quit: SIGSEGV]
<RP>
rburton: queued
<RP>
rburton: do the other patches in -next look ok?
<eminboydak>
Hi, can i install openjdk-11 or openjre-11 with meta-java layer? I cant see in recipes.
<eminboydak>
ldywicki already included in bblayers.conf file
alessioigor has quit [Quit: Client closed]
<kanavin>
eminboydak, meta-java only has recipes for openjdk-8
<kanavin>
if you need 11, you need to write them yourself i'm afraid, and I do not envy you. It might be easier to get it from a host distro and pass in via HOSTTOOLS, if you only need it at build time and not on the targets.
<kanavin>
or you can convince your management that Java is a dead technology, relying on proprietary tools written in it was a big mistake, and you need something future proof ;-)
<eminboydak>
kanavin oh ok, thank you for your information
<ldywicki>
err, you can try to fetch amazon build of openjdk from meta-aws
<ldywicki>
@kanavin your assessment of java is a bit unfair, can you recognize oracle hotspot from openjdk?
<ldywicki>
@eminboydak I've worked once at fork of meta-zulu-java, which is based on binary builds of Azul stuff: https://github.com/connectorio/meta-zulu-java. This is a bit old, but I think it still should work if you lift layer compatibility. I've made also some work on making generic meta-jvm layer which would integrate other openjdk builds i.e. temurin.
<ldywicki>
above layer covers x86_64 and arm architectures
<rburton>
RP: i feel like writing a test case for the flags change
pivi has quit [Ping timeout: 252 seconds]
florian has quit [Ping timeout: 264 seconds]
<zeddii>
rburton: not really a use case that I'm interested in hacking in .. I'd just make sure that the source was removed, unpacked and re-linked if the version changed.
pivi has joined #yocto
<jstephan_>
zeddii: of course user can always use a cleansstate when changing the kernel version, but I feel it's too bad to recompile everythin while previous kernel was already compiled. And compiling the kernel usually takes several minutes ...
<zeddii>
I didn't say cleasstate by the user
* zeddii
knows a few things about kernel builds
<jstephan_>
so what would you suggest?
<zeddii>
so spare my sanity with the pandering comments
<jstephan_>
:)
<zeddii>
do a clean task, just like any other recipe with source code, this is just a bit special where we put it all that time ago. you don't keep multiple versions of source around for any recipe, why would we do it for the kernel.
<zeddii>
I've got 10 year old bugs on external modules not being rebuilt when the kernel source or kernel itself is changed, it's the same missing dependency. I've just never had time to look into it.
vquicksilver has quit [Ping timeout: 255 seconds]
enok has quit [Read error: Connection reset by peer]
Minvera has joined #yocto
luc4 has quit [Quit: Konversation terminated!]
florian has joined #yocto
sotaoverride is now known as Guest2483
Guest4001 is now known as sotaoverride
<jstephan_>
zeddii: not sure to fully understand your reply about a clean task?
tec has quit [Ping timeout: 240 seconds]
tec has joined #yocto
Net147 has quit [Quit: Quit]
tec has quit [Client Quit]
Net147 has joined #yocto
Net147 has quit [Changing host]
Net147 has joined #yocto
goliath has joined #yocto
mbulut has joined #yocto
<jstephan_>
zeddii: I am sure you knows a lot more than me about kernel builds and yocto in general, but I have time and I really want to contribute to the YP. Maybe you can give me pointers on how to properly fix this?
<leon-anavi>
tlwoerner, nice, I didn't see this in meta-rockchip. Anyway, the idea is to keep meta-rauc-community as an aggregation of different examples.
<tlwoerner>
leon-anavi: i litterally *just* pushed it
<leon-anavi>
oh, was it pushed today?
<tlwoerner>
15 minutes ago
<tlwoerner>
leon-anavi: thank you for meta-rauc-community, it's nice having concrete examples to examine
<leon-anavi>
feel free to open GitHub pull request to meta-rauc-rockchip. I literally did this on my spare time on Sunday and after that my colleagues were OK to wrap it and write this short article
<leon-anavi>
I don't have any modern Rockchip apart for Rock Pi 4 B so this is the only device I tested on.
<rfs613>
are there cases where HOST_ARCH and TARGET_ARCH would differ? The documentation only says
<rfs613>
meta/conf/documentation.conf:HOST_ARCH[doc] = "The name of the target architecture. Normally same as the TARGET_ARCH."
<rburton>
rfs613: normally only in the fun situation where you're building cross compilers
<rfs613>
so does the terminology follow from autoconf?
<rfs613>
eg. it is for canadian cross situation
<rburton>
yes
<kanavin>
yes, except we confuse host and target in many places, and it's not pressing enough to get terminology entirely correct
<kanavin>
but otherwise yes, it's set by GNU and we follow
<kanavin>
it's also becoming less relevant with modern toolchains, e.g. rust which can build for any target, and doesn't have to be specifically configured for a particular one
<kanavin>
(I wish gcc could do that...)
CrazyGecko has quit [Remote host closed the connection]
rfuentess has quit [Remote host closed the connection]
<rfs613>
rburton: kanavin: thanks for clarification.
eminboydak has quit [Quit: Client closed]
eminboydak has joined #yocto
<rburton>
kanavin: i don't suppose you've looked at upgrading librsvg to the latest release that use meson instead of autotools?
<kanavin>
rburton, I guess not, what does latest AUH say?
<kanavin>
I've been trying to sort my OSS presentation, and then get back to making devtool aware of lockstep-upgrades for multiple recipes
<kanavin>
"The year 2038 problem is a well known integer overflow issue on many 32 bit platforms, some of which will be still in use on the day when it happens: January 19 2038. In this talk I would like to present where the problem comes from, what the Yocto project has done to address the issue, which base work in kernel and libc has been utilized to avoid a total system collapse, how to test a system's readiness for that date, and which
<kanavin>
further issues this has uncovered in common open source components. I hope this prompts an interesting discussion and further ideas to ensure the world does not go down in 14 years. "
khem has joined #yocto
<moto-timo>
yocton: I missed seeing your nick on this channel, we can continue the any discussion here about the triage topic
<yocton>
moto-timo: (following here) Thx! I'll try to see if the "Upstream-status" of RSS is related to the "devtool latest-version" the AUH uses.
<yocton>
moto-timo: I feel like there is a thread we can pull here to help the deprecation policy
<moto-timo>
yocton: I agree. Let's keep bouncing ideas around.
<moto-timo>
yocton: And it could be adding some kind of plugin to devtool to output something that can be consumed by a script to populate the database in layerindex.
<moto-timo>
yocton: this is what I did for the Yocto compatible layers.
<moto-timo>
yocton: so some similar concept will create the data we need and ingest it. Then it becomes a cronjob script or something like that to automate it, as far as possible. Plus some kind of new UI to display the information.
Kubu_work has quit [Quit: Leaving.]
<yocton>
moto-timo: Thanks for all this info. I'll need to disgest a little... How about creating a bug to keep this?
<moto-timo>
yocton: excellent idea :)
<moto-timo>
yocton: can you add the enhancement request with what you have in mind and then I'll add links and commentary?
geoffhp has joined #yocto
smooge has quit [Quit: I have resigned. I will not be pushed, filed, stamped, indexed, briefed, debriefed, or numbered! My life is my own.]
<yocton>
vmeson: I believe that is you who asked about the AUH run on meta-oe : on meta-oe sublayer, on ~1000 recipes, 24.1% failed the version checking, (then 17.1% failed the attempted upgrade)
<moto-timo>
yocton: about a year ago I guess I was running meta-python, meta-perl and meta-oe (only python and perl recipes). Any given week there were about 50 recipes in meta-python that needed an upgrade. Around a dozen in meta-perl. Folks tend to ignore the ones that fail (some of them, especially in meta-perl are non-trivial to fix).
<moto-timo>
sadly, two of the build machines I was using have since gone off-line for various reasons.
<moto-timo>
I'd love to see the ones where version checking fails get more attention. It is easy to ignore in the current way AUH runs, as far as I recall.
<moto-timo>
Since we have core-image-ptest-<recipe> (and similar in other layers), we could upgrade recipes and run ptests on reduced content images all in one workflow. Probably adds a lot of build time for all those extra images though? not sure
<moto-timo>
I mostly run AUH for my own recipes, so not the full blown "all" recipes run.
jmd has quit [Remote host closed the connection]
jmd has joined #yocto
tepperson has joined #yocto
<tepperson>
I'm trying to write a recipe for a go program, I've got a recipe put together but it builds a *.a output instead of a binary. Outside of yocto it builds a binary no problem. Any ideas what I might be doing wrong? I am specifically wanting to build https://github.com/OpenPrinting/ipp-usb
<tepperson>
I figured it out, I needed to make a recipe for the dependent go module and DEPEND on it.
Starfoxxes has quit [Remote host closed the connection]
dankm has quit [Remote host closed the connection]
eminboydak has quit [Quit: Client closed]
dankm has joined #yocto
vquicksilver has joined #yocto
enok has quit [Ping timeout: 264 seconds]
<kanavin>
if openjdk provides prebuilt binary tarballs for x86-64, aarch64 and ppc64, then it should be done in same way as go and rust
<kanavin>
both *used* to be bootstrapped from a different language, but that's been abandoned ages ago
<kanavin>
now the first step is to pull prebuilt binaries and use them as 'native' compilers
jmd has quit [Remote host closed the connection]
dmoseley has joined #yocto
florian_kc has quit [Ping timeout: 260 seconds]
<vvn>
can we have concurrent accesses to DL_DIR? e.g. two build container sharing a DL_DIR downloading the same package at the same time
olani- has joined #yocto
florian_kc has joined #yocto
florian_kc is now known as florian
Guest11 has joined #yocto
<Guest11>
Hello. I created the recipe. I added this recipe to the DEPEND of another recipe. The created files of first recipe were placed in recipe-sysroot. I added "recipe-name-native" to DEPENDS to second recipe. The files did not appear in recipe-sysroot-native. What am I doing wrong?
olani- has quit [Remote host closed the connection]
<RP>
vvn: yes
olani- has joined #yocto
lexano has joined #yocto
Xagen has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
geoffhp has quit [Quit: Client closed]
MrCryo has quit [Ping timeout: 256 seconds]
<JPEW>
rburton: Pushed a patch to ML to automatically pull the SPDX licenses from the JSON data
geoffhp has joined #yocto
geoffhp has left #yocto [#yocto]
Guest11 has quit [Quit: Client closed]
leon-anavi has quit [Remote host closed the connection]
ecdhe has quit [Read error: Connection reset by peer]
ecdhe has joined #yocto
Guest36 has joined #yocto
|Xagen has joined #yocto
<Guest36>
Hi, is there a simple way to build all recipes in a layer?
mbulut has quit [Ping timeout: 264 seconds]
olani- has quit [Remote host closed the connection]