camus has quit [Read error: Connection reset by peer]
xmn has joined #yocto
camus has joined #yocto
camus has quit [Ping timeout: 268 seconds]
camus has joined #yocto
Estrella___ has joined #yocto
pgowda_ has joined #yocto
Estrella_ has quit [Ping timeout: 252 seconds]
amitk has joined #yocto
nemik has quit [Ping timeout: 252 seconds]
nemik has joined #yocto
nemik has quit [Ping timeout: 252 seconds]
nemik has joined #yocto
goliath has joined #yocto
vladest has joined #yocto
vladest has quit [Ping timeout: 252 seconds]
manuel1985 has joined #yocto
goliath has quit [Quit: SIGSEGV]
goliath has joined #yocto
starblue has joined #yocto
goliath has quit [Client Quit]
Guest53 has joined #yocto
<Guest53>
Hi guys, there is a typo I guess in the Yocto project docs
<Guest53>
3.23.3.1 Making Sure the Packaging is Done, there it says 'lighttpd', but probably 'libhttpd' is meant
<Guest53>
that's it
ptsneves has joined #yocto
<LetoThe2nd>
Guest53: i think its correct. lighttpd is a httpd server package, which is being split into packages.
<Guest53>
Okay, thanks
manuel1985 has quit [Ping timeout: 244 seconds]
<LetoThe2nd>
Guest53: thanks for the heads up!
<Guest53>
yep
alejandrohs has quit [*.net *.split]
npcomp has quit [*.net *.split]
bryanb has quit [*.net *.split]
JudgeMaygarden[m has quit [*.net *.split]
grma has quit [*.net *.split]
mux[m] has quit [*.net *.split]
kergoth has quit [*.net *.split]
frosteyes has quit [*.net *.split]
rperier has quit [*.net *.split]
eggman has quit [*.net *.split]
shivamurthy has quit [*.net *.split]
stacktrust_ has quit [*.net *.split]
rburton has quit [*.net *.split]
stacktrust_ has joined #yocto
rperier has joined #yocto
bryanb has joined #yocto
kergoth has joined #yocto
shivamurthy has joined #yocto
alejandrohs has joined #yocto
frosteyes has joined #yocto
eggman has joined #yocto
rburton has joined #yocto
npcomp has joined #yocto
<Guest53>
Does bitbake use an embedded Python interpreter to process python-style tasks?
mux[m] has joined #yocto
JudgeMaygarden[m has joined #yocto
<Guest53>
I want to use a specific interpreter (e.g. python3.5 and higher), can I do that? Or is there only one fixed embedded python interpreter that comes with the OE version?
mborzecki has left #yocto [#yocto]
<LetoThe2nd>
Guest53: I don't know, but what is the rationale behind it?
mborzecki has joined #yocto
<Guest53>
Just not being sure if I can execute complex logic from a Python task inside a recipe, or should I better write the logic to a script and call that from recipe task
gsalazar has joined #yocto
<Guest53>
I think the second option is the way to go..
<LetoThe2nd>
Guest53: you can do pretty complex stuff in there, the question is more, where does it correctly fit in.
<LetoThe2nd>
Guest53: if you look at the sdk-related bbclasses for example, you can see that it is quite non-trivial.
<LetoThe2nd>
Guest53: please note that I just looked at master. if you are using something outdated, then ... *shrug*
<Guest53>
I can check that in the sources. Thanks, you made things more clear for me!
RichardGolledge[ has joined #yocto
denisoft81[m] has joined #yocto
mborzecki has joined #yocto
marex has joined #yocto
gsalazar has quit [Read error: Connection reset by peer]
gsalazar_ has joined #yocto
zpfvo has joined #yocto
leon-anavi has joined #yocto
frieder has joined #yocto
goliath has joined #yocto
gsalazar_ has quit [Quit: Leaving]
gsalazar has joined #yocto
nemik has quit [Ping timeout: 252 seconds]
nemik has joined #yocto
manuel1985 has joined #yocto
ardo has quit [Read error: Connection reset by peer]
nemik has quit [Ping timeout: 245 seconds]
nemik has joined #yocto
ardo has joined #yocto
mvlad has joined #yocto
<Guest53>
LetoThe2nd is there something special that is worth knowing in bitbake's Python? Or is it just a regular py3 with certain libraries loaded?
<Guest53>
or 'module', not libraries
<Guest53>
that would be more correct, in terms of python
<Guest53>
e.g. utils etc
<LetoThe2nd>
Guest53: no idea.
<Guest53>
Yes, I am just scared of doing something, hence a lot of asking. Let's start coding :)
<LetoThe2nd>
good plan, try and find out. :-)
<qschulz>
Guest53: python 3.6 is the minimal required version for Bitbake for a few releases now
<qschulz>
ah well, that's what LetoThe2nd pointed at
<Guest53>
yes, but thanks
<Guest53>
I use a lower actually. we have an old release
<Guest53>
3.4 I guess. But still good, I can work with that I think
<Guest53>
the only solution is upgrading actually, I will have one more argument to suggest an upgrade
<qschulz>
3.4 has this commit
<qschulz>
but it's EOL anyway
<qschulz>
so please upgrade to kirkstone (4.0)
<Guest53>
Okay, thanks
<LetoThe2nd>
qschulz: time to celebrate! I successfully answered a python question today!
<qschulz>
it's an LTS (2 years support right now, maybe more, depends if companies put money into the project for longer maintenance)
<Guest53>
need to make up new topic for the streams
<qschulz>
LetoThe2nd: yay \o/
<Guest53>
Thanks qschulz
<LetoThe2nd>
Guest53: heh, sadly no streaming sessions these days. would love to do, but time...
<qschulz>
Guest53: I don't know if the python interpreter used in Bitbake allows to use any python module/extension in its python tasks
<qschulz>
(I assume it could/should but it's also not THAT great if you're using non-core modules because you'd then require your users to install additional python packages on their host
OnkelUlla has joined #yocto
<qschulz>
and have to deal with versioning etc...)
<Guest53>
Okay, maybe calling a script is a good option
<qschulz>
Guest53: script would be the same if it's called by Bitbake during recipe parsing/building
<Guest53>
yes, that was the idea
<qschulz>
actually... not true
<qschulz>
if you use it as a script and call it as a script directly from the recipe, you could create a native recipe for this python module/extension you need and use it
<Guest53>
yes
<qschulz>
but that needs to happen after do_prepare_recipe_sysroot (likely do_configure, do_compile, do_install or that kind of stuff)
<Guest53>
I see you are more bitbake/oe oriented than me. I was thinking simple. Just call subprocess.call("python3 script.py")
<Guest53>
like this
<qschulz>
in which case, I think you wouldn't even need to care about the minimal version because the native python recipe would be used instead of the host (IIRC, mnot entirely sure)
<qschulz>
Guest53: you could simply call in a shell task ./script.py (Maybe), or ${PYTHON_NATIVE} ./script.py (or something like that0
<Guest53>
PYTHON_NATIVE is a var for that, okay, didn't know thanks
<qschulz>
that's not the correct name, need to look it up
<Guest53>
yes, shell can also be used. Okay, thanks for ideas. Can try them now
<qschulz>
${PYTHON} should be enough
gsalazar has quit [Ping timeout: 245 seconds]
gsalazar has joined #yocto
<qschulz>
but I don't see it being used
<qschulz>
so wondering if just inheriting (inherit) python3native in your recipe should be enough
<Guest53>
let me check that too
<qschulz>
since the bbclass exports a PYTHON environment variable, maybe that does the trick
<JaMa>
yes just python3native and then call python3 from shell task or shebang in the script
<Guest53>
cool, thanks for solution
<JaMa>
if any python3 from host is good enough for your script (no fancy extensions) then you don't need the inherit
<JaMa>
python3 is in PATH from HOSTTOOLS
<JaMa>
and I would do separate script, because if the logic is complex then you might need to debug your script separately and that's easier by hand in your source repo, than debuging complex python code inside task code
<Guest53>
exactly
wkawka has joined #yocto
grma has joined #yocto
<JaMa>
vmeson: is Aryaman Gupta here? I'm running some benchmarks with /proc/pressure limits and I might have some questions, e.g. why cpu and io, but not memory? In too many of my builds it not only gets slower, but then fails with OOMK and ram usage is IMHO a bit less predictable than cpu/io, so BB_PRESSURE_MAX_MEMORY seems like good idea to me, but maybe there were technical reasons not to do that
ardo has quit [Read error: Connection reset by peer]
ardo has joined #yocto
wkawka has quit [Quit: Client closed]
ardo has quit [Read error: Connection reset by peer]
ardo has joined #yocto
Guest53 has quit [Ping timeout: 252 seconds]
florian has joined #yocto
wkawka has joined #yocto
<wkawka>
Hi, is there a proper channel with kas-container? i didn't find it
Guest81 has joined #yocto
Guest81 has left #yocto [#yocto]
<rburton>
there's no kas channel, no
<rburton>
some people here use kas so might be able to help, but if you want to talk to the maintainers then you need to use the google group
xmn has quit [Quit: ZZZzzz…]
<ptsneves>
@rburton regarding the CMake passing absolute paths to flex, even if we could get CMake to pass relative paths(which we cant) there is no way the #line would be useful without also setting some prefix for the target which from the source it is not possible.
prabhakarlad has joined #yocto
<ptsneves>
perhaps patch flex to accept a map for the #line entry?
nemik has quit [Ping timeout: 252 seconds]
nemik has joined #yocto
<wkawka>
I have a strange problem, kas cannot resolve any host, event if i can clone them. I tried reinstalling, rebooting etc. nothing helps
<wkawka>
even*
<rburton>
ptsneves: is this in the context of the -src package?
<ptsneves>
@rburton yes and the fact it triggers buildpaths qa
nemik has quit [Ping timeout: 240 seconds]
<rburton>
ptsneves: if so i'm 50/50 on patching flex to take a map vs just post-processing the -src with a search-and-replace
Sona has joined #yocto
nemik has joined #yocto
<ptsneves>
@rburton lol me too that is why i ask. I guess the tipping point is whether this is expected to happen in other packages, or has been found already
<ptsneves>
on the other hand i do not see this as an issue in actual binaries due to gcc itself having the map feature
Guest53 has joined #yocto
<qschulz>
wkawka: you ran the git clone *manually* *inside* kas-container?
gsalazar has quit [Read error: Connection reset by peer]
gsalazar has joined #yocto
<wkawka>
it was outside it, i will try to enter it
<rburton>
ptsneves: it's entirely a generated source problem, but also hits other tools like the glib enum/marshal generators
<rburton>
wkawka: what specifically can't access any host
<wkawka>
it cannot clone even poky
<rburton>
what is it
<qschulz>
rburton: kas-container I assume
<rburton>
what *specifically* are you trying to do.
<rburton>
literally just docker into the kas-container then git clone?
<qschulz>
wkawka: don't you have a company proxy or something in place? have you tried with *any* container (like, debian:bullseye or something like that)
<wkawka>
i'm trying to run kas-container shell with ssh keys set
<qschulz>
wkawka: let's first make sure poky clones inside it, and it does not require ssh keys to do so (except if you have a mirror which can only be access with ssh keys?)
<qschulz>
accessed*
<wkawka>
how can i manually run the container, i tried something like `docker run ghcr.io/siemens/kas/kas:3.0.2 shell /path/to/kas/file`
<wkawka>
I can't do this in the standard way, it cannot reslove hosts and then I cannot enter shell
<Guest53>
Is anyone able to download yocto-docs from the server?
<qschulz>
Tyaku: ptsneves: ^ add a DEPENDS += "maven-native" in your recipe I guess?
<ptsneves>
yep
<qschulz>
(and add this layer or copy the relevant files (recipe, include files, additional files, etc..)
<ptsneves>
yep :)
prabhakarlad has quit [Quit: Client closed]
zpfvo has joined #yocto
<Tyaku>
Thanks qschulz, i'm watching it now.
ardo has quit [Read error: Connection reset by peer]
GNUmoon2 has quit [Ping timeout: 268 seconds]
ardo has joined #yocto
GNUmoon2 has joined #yocto
<rburton>
ptsneves: bitbake -e will show you the data store.
paowz_ has quit [Ping timeout: 255 seconds]
florian_kc has joined #yocto
brazuca has joined #yocto
florian has quit [Ping timeout: 268 seconds]
hcg has joined #yocto
ardo has quit [Read error: Connection reset by peer]
paowz_ has joined #yocto
ardo has joined #yocto
sakoman has joined #yocto
nemik has quit [Ping timeout: 240 seconds]
nemik has joined #yocto
brazuca has quit [Quit: Client closed]
nemik has quit [Ping timeout: 240 seconds]
nemik has joined #yocto
florian_kc is now known as florian
<ptsneves>
why does systemd's systemd.pc root_prefix empty even when using usrmerge?
kriive has quit [Remote host closed the connection]
<ptsneves>
fixing the ROOTPREFIX_NOSLASH would remove meta/recipes-core/systemd/systemd/0001-Move-sysusers.d-sysctl.d-binfmt.d-modules-load.d-to-.patch
alessioigor has joined #yocto
alessioigor has quit [Client Quit]
rsalveti has joined #yocto
<Tyaku>
Thanks qschulz it works like a charm. I put DEPENDS += "maven-native openjdk-8-native" and then in configure i'm able to do this: export JAVA_HOME=${WORKDIR}/recipe-sysroot-native/usr/lib/jvm/openjdk-8-native ${WORKDIR}/recipe-sysroot-native/usr/bin/mvn package. Then it's only install stuff to put the JAR where I want.
goliath has quit [Quit: SIGSEGV]
<qschulz>
Tyaku: :+1:
manuel1985 has quit [Ping timeout: 240 seconds]
frieder has quit [Remote host closed the connection]
Guest14 has joined #yocto
Guest14 is now known as brazuca
<ramacassis[m]>
Hi everyone, the project I am working on needs to build a recipe (u-boot) in both a dev and release version.
<ramacassis[m]>
Looking for best practices:
<ramacassis[m]>
it is better to create two distinct recipes "u-boot-dev.bb" and "u-boot-rel.bb" that set some specific variables before calling a common file "u-boot.inc".
<ramacassis[m]>
Or to keep a unique recipe "u-boot.bb", with default variables, and use local.conf to override some of them ?
<ramacassis[m]>
In the end I do either:
<ramacassis[m]>
OR, ./compile-u-boot.sh dev (or rel) # the script add relevant vars in local.conf, then calls $ bitbake u-boot
<ramacassis[m]>
bitbake u-boot-dev (or -rel) # the relevant vars are already set in recipe
<qschulz>
ramacassis[m]: you should use two different distro
florian has quit [Ping timeout: 268 seconds]
<ramacassis[m]>
that's the keyword I was looking for, way better now. Thanks qschulz !
brazuca has quit [Quit: Client closed]
florian has joined #yocto
Guest14 has joined #yocto
Guest14 is now known as brazuca
kanavin has quit [Quit: Leaving]
amitk has quit [Ping timeout: 245 seconds]
kanavin has joined #yocto
nemik has quit [Ping timeout: 268 seconds]
nemik has joined #yocto
Tyaku has quit [Quit: Lost terminal]
k4wsys[m] has quit [Quit: You have been kicked for being idle]
goliath has joined #yocto
florian has quit [Ping timeout: 268 seconds]
nerdboy has quit [Ping timeout: 252 seconds]
zpfvo has quit [Ping timeout: 252 seconds]
GLumen has joined #yocto
GLumen has quit [Client Quit]
nerdboy has joined #yocto
nerdboy has joined #yocto
nerdboy has quit [Changing host]
florian has joined #yocto
brazuca has quit [Quit: Client closed]
<vmeson>
JaMa: it's a holiday in Ontario, Canada today so please ask again tomorrow. IIRC, we have lots of memory on our builders and the memory pressure only happens if you start to swap.
<JaMa>
vmeson: ok, thanks, I have 128G in 64thread builder and it used to be OK, but now it triggers OOMK (without lots of swap) if /proc/pressure can back-off a bit before this happens it would be great
<vmeson>
JaMa: Using too low a limit (cpu/io) can slow down an isolated build and if there is other activity causing io/cpu pressure, then regulating your own pressure can *really* make a build slower.
<JaMa>
I also plan to test the ninja approach from agherzan, because chromium is usually the reason for too much presure in my builds (and if I happend to run 2 chromium do_compile at the same time) then bitbake not starting any other task won't be enough to prevent swap+OOMK
<JaMa>
while ninja can react faster and on more fine grain level inside the do_compile
<vmeson>
JaMa: interesting. we've been running on 24/56 core systems with 64 / 128 GB RAM respectively.
<JaMa>
so -j 24/-j 56 without hyperthreading? or double that?
<vmeson>
JaMa: 2x chromium builds is just asking for trouble! ;-) We'll see what we can figure out. Note that Aryaman is a summer intern and his term is over in 2 weeks.
<JaMa>
nodejs+chromium at the same time is similar story
<vmeson>
JaMa: we've done a variety of builds from -j24/56 respectively to -j 8/12 when running 3 builds at once. I'll get Aryaman to share some of our perf data tomorrow.
<JaMa>
I'm updating the test-oe-build-time results now as well
<vmeson>
JaMa: yeah, we need ninja/make to have a mode where they limit the jobs issued. There is already -l for ninja/make and while that helps it's very slow to respond due to the 1 minute averaging.
* vmeson
goes back on his "socialist" summer holiday. ;-)
<JaMa>
thanks and enjoy :)
* JaMa
had load >420 yesterday and it was more than a minute, but not sure if ninja would get enough attention from cpu to actually notice that :)
pgowda_ has quit [Quit: Connection closed for inactivity]
GNUmoon2 has quit [Write error: Connection reset by peer]
sakoman has quit [Quit: Leaving.]
GNUmoon2 has joined #yocto
seninha_ has joined #yocto
seninha has quit [Ping timeout: 268 seconds]
<JPEW>
Has anyone seen errors like `FileNotFoundError: [Errno 2] No such file or directory: '/mnt/sstate/99/sigtask.wq6gkrm7' -> '/mnt/sstate/99/sstate:hdparm::9.48:r0::3:990d8b39c15752a2dd68369cb3609283_unpack.tgz.siginfo'` when using sstate over NFS?
<JPEW>
I think maybe I have so many builds there is a small chance that the unique temporary filename (wq6gkrm7) is not unique and it races with another builder
<JPEW>
(esp. since all the sigtasks must be unique for a given directory)
mihai has quit [Quit: Leaving]
<JPEW>
Hmm, `mkstemp()` generates the "unique" portion by using the current timestamp, so I think this is entirely possible to conflict
nemik has quit [Ping timeout: 240 seconds]
nemik has joined #yocto
florian has quit [Ping timeout: 268 seconds]
nemik has quit [Ping timeout: 240 seconds]
nemik has joined #yocto
<rburton>
but surely mkstemp is race-free as it returns a fd?
<JPEW>
Ah, but we don't leave the FD open while renaming it, so it's not atomic
<JPEW>
(and I'm not sure that would work properly over NFS anyway)
<rburton>
i've been told it should work over nfs, but i've been told many things about nfs
<JPEW>
rburton: Ya. It's worth a try. The rename-while-open and O_EXCL both have to work over NFS for it to work properly
<JPEW>
Sent a patch
vvn has quit [Quit: WeeChat 3.6]
ptsneves has quit [Ping timeout: 244 seconds]
vvn has joined #yocto
nemik has quit [Ping timeout: 252 seconds]
nemik has joined #yocto
seninha_ has quit [Quit: Leaving]
seninha has joined #yocto
Guest14 has joined #yocto
Guest14 is now known as brazuca
<Xagen>
is there a way to tell which package provided a package listed in the manifest?
<Xagen>
the one i'm looking for is a dependency for something, i'm just not sure what specifically
nemik has quit [Ping timeout: 268 seconds]
nemik has joined #yocto
seninha has quit [Remote host closed the connection]
seninha has joined #yocto
florian has joined #yocto
mvlad has quit [Remote host closed the connection]
seninha has quit [Remote host closed the connection]