simonew has quit [Remote host closed the connection]
simonew has joined #yocto
rob_w has quit [Remote host closed the connection]
zpfvo has joined #yocto
rfuentess has joined #yocto
jmd has quit [Ping timeout: 268 seconds]
Ad0 has joined #yocto
Net147 has quit [Quit: Quit]
Net147 has joined #yocto
Net147 has quit [Changing host]
Net147 has joined #yocto
enok has joined #yocto
mvlad has joined #yocto
Kubu_work has joined #yocto
<thomas_34>
Good morning, does someone know whats the "safest" way to check if a variable is empty or not set? Does oe.utils provide a function for that?
prabhakalad has quit [Ping timeout: 260 seconds]
prabhakalad has joined #yocto
khem has quit [Remote host closed the connection]
enok has quit [Ping timeout: 240 seconds]
mihai has quit [Quit: Leaving]
enok has joined #yocto
jpuhlman has quit [Ping timeout: 268 seconds]
enok has quit [Ping timeout: 260 seconds]
florian has joined #yocto
jpuhlman has joined #yocto
splatch has joined #yocto
splatch is now known as ldywicki
enok has joined #yocto
ldywicki has quit [Changing host]
ldywicki has joined #yocto
<mckoan>
thomas_34: do you mean in a recipe?
rudynotfound has joined #yocto
<rudynotfound>
hey there i am looking forward to be a yocto developer , how can i get started?
<usvi>
get a devkit, good book and start hacking
<rudynotfound>
hacking what
<usvi>
writing and changing code, compiling, running and debugging
<rudynotfound>
also making my own independent distro and using yocto in the kernel so i cannot take a risk of it not being secured so i planned to join the project itself to be sure of its security and stuff
<rudynotfound>
chat has been really quiet
rudynotfound has quit [Quit: Client closed]
<usvi>
...
<ldywicki>
strange ;-)
rudynotfound has joined #yocto
<usvi>
its the nature of this chat that there are sometimes more messages and sometimes less
<rudynotfound>
irc very complicated
<ldywicki>
@rudynotfound I do use arch, but never really made any build of it beyond several packages. Do you use arch as a base for your distro?
<rudynotfound>
no i would use lfs
<ldywicki>
to be precise I use manjaro
<rudynotfound>
gonna create its own base
<rudynotfound>
i work at endeavour os community edition hyprland
<rudynotfound>
but i use vanilla arch
Michael_Guest has joined #yocto
<ldywicki>
I see, irc was a great place in the past, nowadays most of chats moved to more modern tools, but irc is still great. It is much less distracting than stuff with mobile apps.
<rudynotfound>
discord replaced it seriously
<rudynotfound>
imma go install arch again lol
Vivith has joined #yocto
Vivith has quit [Client Quit]
<ldywicki>
per say you do not need arch, you can try to conduct build using container to spin bitbake
<ldywicki>
I use also a tool called kas which is quite handy to assemble build config
<ldywicki>
anyhow, you probably know more about linux development than myself, throwing tooling just so you have idea :)
<rudynotfound>
its what i use i do not wanna use windows but my school got ms office so i cannot even leave it
<rudynotfound>
i sometimes feel like erasing disk and installing arch like i love linux and tux is friendly with me since my drivers are opensource even the gpu and the processor
<ldywicki>
you can run, you can hide, ms office will find you
<rudynotfound>
my classmates even barely know how to update windows
<rudynotfound>
i am like the hacker in the class
Tyaku has quit [Quit: leaving]
<rudynotfound>
fedora would be better for these stuff ig? because it got more options for development?
<rudynotfound>
because thats what torvalds use
<ldywicki>
not sure, most of linux distros will provide you a way to pull development tools, and with bitbake you need mostly standard stuff as all heavy lifting (like cross compilation) happens inside of the build.
opalfruit has joined #yocto
<rudynotfound>
okai
enok has quit [Ping timeout: 252 seconds]
<opalfruit>
Hi, just wondering if anyone can point me in the right direction, i can see aarch64-poky-linux-gcc exists, which i believe i'm using to build the kernel/rootfs programs, but i'm trying to also install arm-poky-linux-gcc so the host can also compile armhf programs, which i was trying to do using DEPENDS, but i keep getting close matches only listing aarch64-poky-linux-gcc, anyone got any ideas?
<rudynotfound>
what is your yocto env?
<opalfruit>
i'm using kirkstone, if that's what you mean? i did just noticed it also lists virtual/arm-none-eabi-gcc-native, but not sure if that would be the right compiler to compile u-boot
florian_kc has joined #yocto
<rudynotfound>
you are doing everything right
<thomas_34>
mckoan, I mean in a config. For example to append something to OVERRIDES
enok has joined #yocto
rudynotfound has quit [Quit: Client closed]
<ldywicki>
@opalfruit this might depend on MACHINE variable, the arm-poky-linux-gcc will require different target than aarch64
Guest72 has joined #yocto
<opalfruit>
ldywicki: thanks, could i possibly add a different MACHINE variable to the particular .bb file i want to use armhf, rather than using multiple conf files and multiconfig?
starblue has quit [Ping timeout: 256 seconds]
Kubu_work has quit [Read error: Connection reset by peer]
<ldywicki>
@opalfruit this I don't know, but you can swap variable without changing checkout directory, it will cause recompilation to new architecture and re-use of some state (and downloads), the work directory is smart enough to distinguish both.
manuel__ has quit [Ping timeout: 268 seconds]
starblue has joined #yocto
<ldywicki>
you will end up with another rootfs image, but work directory will contain also packages (if you configured package managed) to be deployed onto this architecture
Guest72 has quit [Quit: Client closed]
enok has quit [Ping timeout: 256 seconds]
<alessioigor>
Should we stop to use tarballs?
ldywicki has quit [Quit: Updating details, brb]
ldywicki has joined #yocto
thomas_34 has quit [Ping timeout: 250 seconds]
manuel__ has joined #yocto
<Fahad>
can yocto generated image use package repo of other linux distro having compatible format?
<Fahad>
for example, if build system is configured to use deb package format, can Ubuntu's apt packages repo be used?
ehussain has joined #yocto
Guest5 has joined #yocto
Guest5 has quit [Client Quit]
Fahad has quit [Quit: Client closed]
Saur_Home85 has quit [Quit: Client closed]
Saur_Home85 has joined #yocto
<ldywicki>
generated packages are compatible with other distros till some degree (their format is the same), however dependency paths often do not match. When I generated apt package with yocto I could use it with ubuntu/raspbian but had to force install due to "unresolved" dependencies. For me it was not an issue (package not bound to architecture), but for other cases it won't be that easy.
jbo_ is now known as jbo
tgamblin has quit [Ping timeout: 264 seconds]
lexano has joined #yocto
DvorkinDmitry has joined #yocto
<DvorkinDmitry>
I found that i need to manually disable a couple of classes in openembedded-core/meta/classes/ by moving they filenames names to ...off because some other layer is replacing them unsuccessfully. Is there are any other possibility to do the same in OE/Yocto settings variables?
Kubu_work has joined #yocto
<Saur>
DvorkinDmitry: If you want a bbclass to actually not do anything, you can put an empty version of it in a layer you control. Do note that the order bbclasses are searched for is based on BBPATH so typically your layer needs to be specified earlier in BBPLAYERS than the layer which bbclass you want to override.
<DvorkinDmitry>
Saur, there are rust* and cargo* classes are redefined in additional layer. I need to switch off the same classes from OE completely.
<Saur>
DvorkinDmitry: bitbake will only read a bbclass once, so either it will read it from the additional layers or it will read it from OE. Which one is read depends on the order in BBLAYERS.
<DvorkinDmitry>
Saur, aha! and... is there are any mechanism to avoid reading several .bbclass files from OE layer?
<Saur>
DvorkinDmitry: If the classes are read from your additional layers, they will not be read from OE.
Guest80 has joined #yocto
<DvorkinDmitry>
Saur, ok. so only layers order
<JaMa>
doesn't BBMASK work for bbclasses as well?
<Saur>
DvorkinDmitry: Well, technically the order in BBPATH, but since most layers append to it, it mostly matches the order in BBLAYERS.
<DvorkinDmitry>
thank you, Jama and Saur! I'll try
<Saur>
JaMa: No idea, I had not considered that. If it does, that is great.
<Guest80>
This is my first time here - I am seeking a place to ask a yocto build question...Do I just post it here ? what is the correct protocol
<JaMa>
I really don't know, but someone used that in one of our layers (to my surprise), trying now
<JaMa>
and looks like BBMASK doesn't work for bbclasses
<Saur>
Guest80: You are in the correct place. Just go ahead and ask.
<Saur>
JaMa: :(
<Saur>
Guest80: If you want to post longer stuff like logs, then make sure to use pastebin or similar.
Kubu_work1 has joined #yocto
<JaMa>
DvorkinDmitry: Saur: yes the documentation and code is quite clear that BBMASK is only for recipes and bbappends, sorry for confusion
<Guest80>
I am learning and very very new.... I was to able to build the poky distribution without issue. I then want to build a 3.1 Dunfell version in support of the IMX6 from Variscite. So I am following the build procedures from them... I am using a VM on Mac and Ubuntu as the host OS. I am receiving bitbake errors and I think I am missing something
<Guest80>
very basic. Any chance someone may know how to point me in the right direction. here are the errors.....
<JaMa>
Guest80: it was fixed in newer bitbake, check git log and backport necessary commit (or better upgrde to supported version of OE)
<JaMa>
e2be6defb compat.py: remove file since it no longer actually implements anything
<JaMa>
included since bitbake 1.48.0
enok has joined #yocto
<Guest80>
Ok I created a pastebin account for next time - thanks
Guest28 has joined #yocto
Guest28 has quit [Client Quit]
<Guest80>
Ok I will try it with a docker container and 18.04 - thanks for the guidance.
xmn has joined #yocto
sev99 has joined #yocto
Xagen has quit [Ping timeout: 264 seconds]
opalfruit has quit [Quit: Connection closed for inactivity]
Guest80 has quit [Quit: Ping timeout (120 seconds)]
florian_kc has quit [Ping timeout: 268 seconds]
Michael_Guest has quit [Quit: Client closed]
tgamblin has joined #yocto
thomas_34 has quit [Quit: Client closed]
thomas_34 has joined #yocto
<thomas_34>
A short question: Why is always "=." used, when OVERRIDES or MACHINEOVERRIDES are modified? Why not OVERRIDES:append = "foo" ?
<mckoan>
thomas_34: in case "=." is a prepend
Xagen has joined #yocto
rudynotfound has joined #yocto
<thomas_34>
Okay... I thought maybe the variable expansion time matters here
coelebs has quit [Quit: Client closed]
rudynotfound has quit [Quit: Client closed]
sakoman has joined #yocto
Guest12 has joined #yocto
linfax has quit [Ping timeout: 256 seconds]
<Guest12>
Can someone explain to me why the variable SOURCE_DATE_EPOCH in the reproducible_build.bbclass class is added to the BB_HASHBASE_WHITELIST? -> In the recipe for systemd, the variable is used to set the time epoch for the package during the configure task. Since the SOURCE_DATE_EPOCH variable is added to the BB_HASHBASE_WHITELIST, the systemd recipe
<Guest12>
is not rebuilt when the SOURCE_DATE_EPOCH is changed.
<neverpanic>
SOURCE_DATE_EPOCH changing alone is typically not considered enough to trigger a rebuild of a package. If none of the other inputs changed, why rebuild just because SOURCE_DATE_EPOCH changed?
<neverpanic>
On the other hand, SOURCE_DATE_EPOCH should usually be derived from the source tree, so if SOURCE_DATE_EPOCH changes, didn't your source tree also change, which should trigger the rebuild?
<RP>
neverpanic: SOURCE_DATE_EPOCH clamps the output timestamps so it can and does affect the output
<Guest12>
SOURCE_DATE_EPOCH is set globally in the reproducible_build.bbclass and I want to use the variable to set the time globally for a build of an image. My problem is that in a build with a changed SOURCE_DATE_EPOCH systemd is not rebuilt and is used from my sstate cache.
<Guest12>
if the SOURCE_DATE_EPOCH changes and systemd is configured with a different time epoch, then this affects the systemd application, i.e. the binary and behaviour changes
<neverpanic>
RP: I never claimed it doesn't. I'm just saying, it shouldn't be an input on its own.
<RP>
neverpanic: right, agreed. I think I misread some of the context
<gmorell>
.w 2
<RP>
Guest12: for a given set of inputs, the output should always be deterministic (reproducible). In general we automatically set SDE to something that is deterministic so it isn't used as an input. It isn't intended to be set manually
<neverpanic>
Guest12: Sounds like you'reholding S_D_E wrong, then. The entire point is to avoid a dependency on the current build time by using the value from S_D_E instead. If you are changing S_D_E, but the other inputs to the systemd build do not change, that breaks the entire purpose of S_D_E.
<neverpanic>
Also, it would be a major footgun if changing SOURCE_DATE_EPOCH caused systemd to rebuild, which then in turn caused every dependent of systemd to rebuild. You'd be recompiling half of your userland for a timestamp.
rcw has joined #yocto
<Guest12>
For me, a reproducible build means exactly that for a given set of inputs, the output should always be deterministic (reproducible). Which is clearly violated by changing S_D_E or time_epoch of systemd. Wouldn't it make sense, if S_D_E is not suitable, to use another variable so that the build of systemd is deterministic?
<Guest12>
The rebuild of systemd and recompiling half of userland recipes is a logical consequence of this and also happens with every other minor change to systemd.
simonew has quit [Remote host closed the connection]
simonew has joined #yocto
<neverpanic>
Fortunately there's a great definition of what it means for a build to be reproducible alrady, which I know because I helped write it: https://reproducible-builds.org/docs/definition/
<neverpanic>
Note that it says "given the same build environment", and "The relevant attributes of the build environment […] are defined by the authors or distributors."
<RP>
Guest12: The key thing is the SDE isn't a user set input, the system deterministicly determines it
<neverpanic>
Yocto (and everybody else) just happens to be defining this as SDE is part of the build environment, if you want to reproduce a build, you'll have to use the same SDE.
<neverpanic>
Maybe if you want an image with a newer /etc/timestamp, you should write a systemd.bbappend that deletes it from systemd and add a separate recipe that creates it from scratch, one that will always be rebuilt.
<neverpanic>
That would solve your problem, while still leaving the systemd recipe reproducible, and at the same time, you'll avoid the userspace rebuild.
<Guest12>
Ok all right thanks, then I misunderstood the definition of SDE. I will make a workaround for my "problem"
<JaMa>
meta/classes/image-buildinfo.bbclass is also useful to provide timestamp of the image build and even there it's questionable if the image should re-run do_rootfs just because of the timestamp (if you're using reproducible build artifacts names which don't want to rebuild kernel and images every second)
rfuentess has quit [Remote host closed the connection]
<yudjinn>
hello, I'm having an issue while trying to setup the layerindex on an internal network, can someone help?
<yudjinn>
we have some protected repos that we'd like to use in it, and I've added an ssh-key to the container, but it still gives a permissions issue. I've gone and attached to the layerapp container using docker exec and ran the exact same commands (just importing the utils.runcmd that the fetch_repo uses) and it works as expected, but it fails when run from within the web-ui with "Could not read from remote
<yudjinn>
repository.."
* ldywicki
@khem nice story, do we have a patch in oe? ;)
khem has quit [Quit: WeeChat 4.2.1]
khem has joined #yocto
enok has joined #yocto
sev99 has quit [Quit: Client closed]
thomas_34 has quit [Ping timeout: 250 seconds]
locutusofborg has quit [Ping timeout: 268 seconds]
Haxxa has quit [Quit: Haxxa flies away.]
Haxxa has joined #yocto
enok has quit [Ping timeout: 252 seconds]
Fahad has quit [Quit: Client closed]
sakoman has quit [Quit: Leaving.]
ptsneves has joined #yocto
Saur_Home85 has quit [Quit: Client closed]
Saur_Home85 has joined #yocto
leon-anavi has quit [Quit: Leaving]
mvlad has quit [Remote host closed the connection]
prabhakalad has quit [Ping timeout: 268 seconds]
prabhakalad has joined #yocto
vladest has quit [Quit: vladest]
prabhakalad has quit [Ping timeout: 252 seconds]
prabhakalad has joined #yocto
ptsneves has quit [Ping timeout: 252 seconds]
ehussain has quit [Quit: ehussain]
ehussain has joined #yocto
vladest has joined #yocto
prabhakalad has quit [Ping timeout: 268 seconds]
prabhakalad has joined #yocto
alessioigor has quit [Ping timeout: 256 seconds]
Xagen has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]