tepperson_ has quit [Remote host closed the connection]
nemik has quit [Ping timeout: 248 seconds]
nemik has joined #yocto
tepperson_ has joined #yocto
nemik has quit [Ping timeout: 268 seconds]
nemik has joined #yocto
invalidopcode1 has quit [Remote host closed the connection]
invalidopcode1 has joined #yocto
tepperson_ has quit [Ping timeout: 255 seconds]
Thorn has joined #yocto
thomasd13 has joined #yocto
tepperson_ has joined #yocto
thomasd13 has quit [Remote host closed the connection]
thomasd13 has joined #yocto
Joel76 has joined #yocto
<Joel76>
Good morning! I have noticed mesa and cairo being built as part of my core-image-minimal. How do I figure out why? I don't want or need graphics in my build!
<Joel76>
Actually, I'll rephrase... I want glibc-locale but not the rest of glibc. How can I have this?
thomasd13 has quit [Ping timeout: 246 seconds]
thomasd13 has joined #yocto
warthog9 has quit [Read error: Connection reset by peer]
eLmankku has joined #yocto
tepperson_ has quit [Ping timeout: 255 seconds]
mcfrisk has joined #yocto
warthog9 has joined #yocto
jclsn has quit [*.net *.split]
karl has quit [*.net *.split]
grma has quit [*.net *.split]
beastie has quit [*.net *.split]
eLmankku_ has quit [*.net *.split]
mcfrisk_ has quit [*.net *.split]
rfs613 has quit [*.net *.split]
fray has quit [*.net *.split]
mlaga97 has quit [*.net *.split]
zwelch has quit [*.net *.split]
warthog9 is now known as 020AAPQ8E
jclsn has joined #yocto
karl has joined #yocto
beastie has joined #yocto
rfs613 has joined #yocto
zwelch has joined #yocto
fray has joined #yocto
mlaga97 has joined #yocto
tepperson_ has joined #yocto
thomasd13 has quit [Ping timeout: 256 seconds]
mvlad has joined #yocto
Poppy has joined #yocto
<Poppy>
I like to use variables and flags in Bitbake, all works from python function ( d.getVarFlags('USERS_VALUES_%s' % val) ) but I don't understand how to read variable flagged from bash code... Any tips ?
<Poppy>
For the story I would like to create new users from list ( I have 50 users to set) so I would like to loop over the flags :)
thomasd13 has joined #yocto
<thomasd13>
Good morning guys
<LetoThe2nd>
yo dudX
dShapo has joined #yocto
<LetoThe2nd>
Poppy: the standard expansion ${FANCY_VARIABLE_NAME} doesn't work?
mckoan|away is now known as mckoan
<mckoan>
good morning
<Poppy>
LetoThe2nd for the username list yes, but I don't understand how to use this var as flag. Maybe I'm wrong in something but the goal is to do something like that:
<Joel27>
Moin! where does python3-codecs come from in ./recipes-devtools/python/python3-can_4.0.0.bb? Also, I added a dependency on that but it did not install a bunch of Asian codecs
<Joel27>
I mean, I can't find python3-codecs for some reason
goliath has joined #yocto
chep has joined #yocto
<LetoThe2nd>
anybody have a short pointer for me on how to exclude -native things from an INCOMPATIBLE_LICENSE = "GPL-3.0*" catchall?
prabhakarlad has joined #yocto
<LetoThe2nd>
or a specific -native thing.
grma has joined #yocto
meego has joined #yocto
meego has quit [Remote host closed the connection]
vladest has quit [Remote host closed the connection]
vladest has joined #yocto
d-s-e has joined #yocto
<LetoThe2nd>
INCOMPATIBLE_LICENSE_EXCEPTIONS seems to not be able to understand "make-native:GPL-3.0-only"
Spinola has joined #yocto
tepperson_ has quit [Ping timeout: 252 seconds]
tepperson_ has joined #yocto
Joel27 has quit [Quit: Client closed]
rob_w has joined #yocto
ptsneves has joined #yocto
xmn has quit [Ping timeout: 264 seconds]
<Spinola>
[rburton] I build a new SDK with TOOLCHAIN_HOST_TASK:append = " nativesdk-gmp", now I can install the sdk again, but I get the same problem regarding a missing gmp.h when I'm executing the 'make scripts' command. I've added the target and host manifest files of the sdk to my pastebin https://pastebin.com/0Biz4CN8https://pastebin.com/N8UAW78c
<LetoThe2nd>
by the way, the "building without GPL3"-note here https://docs.yoctoproject.org/ref-manual/images.html#images seems to be incorrect, the build falls over with Missing or unbuildable dependency chain was: ['core-image-minimal', 'meta-environment-extsdk-qemux86-64', 'zlib', 'make']
Spinola_62 has joined #yocto
olani- has quit [Ping timeout: 264 seconds]
olani- has joined #yocto
meego has joined #yocto
<qschulz>
LetoThe2nd: the build may fail if you have gplv3 source code to be compiled in, so maybe something in the depednency chain?
<qschulz>
michaelo: ^
<LetoThe2nd>
qschulz: yeah thats true. but a) the documentation states it is supported, and b) i'm pretty sure that core-image-minimal does not ship make. so something is wrong here.
<qschulz>
LetoThe2nd: what about INCOMPATIBLE_LICENSE:class-native = "" ?
<LetoThe2nd>
qschulz: nopes.
meego has quit [Remote host closed the connection]
<LetoThe2nd>
which brings up the interesting question, how to set up the non-gplv3 build configuration as this definitely is not the image by default.
goliath has quit [Quit: SIGSEGV]
<d-s-e>
LetoThe2nd: It's usually better to use INCOMPATIBLE_LICENSE in your image recipe.
florian has joined #yocto
<LetoThe2nd>
d-s-e: good point.
<Entei[m]>
While specifying "package-management" as one of the `EXTRA_IMAGE_FEATURES`, poky installs `dnf` and `rpm` for installing rpm packages(the selected package class). But it doesn't seem to install `mock`, 'drpm`, `createrepo` etc. to be able to set up repository.
<Entei[m]>
Any solutions to this other than creating custom recipes and layers?
<LetoThe2nd>
Entei[m]: why should those be installed on the target?
seninha has quit [Remote host closed the connection]
seninha has joined #yocto
starblue has quit [Ping timeout: 268 seconds]
<Entei[m]>
<LetoThe2nd> "Entei: why should those be..." <- Coz setting up repo is requiring me to create a repmd.xml file, which can be done with createrepo utility, which has dependencies on drpm
<LetoThe2nd>
Entei[m]: but you don't run the repo on the target, you run it on some development host.
starblue has joined #yocto
Poppy has quit [Quit: Client closed]
<dacav>
Hi. I'm looking at a sysroot of a 'inherit deploy' recipe that doesn't do any compilation. I see it is populated with headers and such. What should I inherit to make it minimal? `allarch` seems reasonable, ...but headers are still here. Headers do not break anything by being there, it is more of a "I like minimality" matter.
<dacav>
*minimalism
<qschulz>
dacav: check if you have something in DEPENDS variable, but otherwise I wouldn't care too much tbh
<dacav>
I do. It is a native binary that I need, but perhaps the recipe that provides it is broken in some way
<dacav>
Thanks qschulz
thomasd13 has quit [Ping timeout: 268 seconds]
<qschulz>
dacav: if you have a DEPENDS in there, then the sysroot of the DEPENDS and **all** its recursive DEPENDS sysroot are brought in
<rburton>
Entei[m]: what LetoThe2nd the said. mock is a tool for building RPMs, createrepo is for managing repos that you're hosting. none of those are needed on the target, you just need dnf/rpm, which you'll get with package-management.
<qschulz>
(in the native sysroot though, for native recipes)
<qschulz>
s/recipes/dependencies
<dacav>
Thanks qschulz, that might explain it.
<dacav>
Another question, if I can :) I've got a recipe r1 which is supposed to provide multiple packages. One of them is -native only. Is it correct to use 'PACKAGECONFIG ??= "foo"' and 'PACKAGECONFIG:native ??= "foo bar"? I also have PROVIDES = "foo bar", but I'm not sure about it
<dacav>
In the recipe r2 that needs bar-native, I wrote DEPENDS = "bar-native"
<Entei[m]>
<LetoThe2nd> "Entei: but you don't run the..." <- Well true that...
<qschulz>
PACKAGECONFIG:native will not work for starters
<qschulz>
class-native will though
<qschulz>
but PACKAGECONFIG is not for defining packages
<dacav>
Mh, maybe I'm doing that wrong yes. I believe I'm thinking of subpackages in RPM
Thorn has quit [Ping timeout: 246 seconds]
<qschulz>
dacav: IIRC, native recipes do not have a specific knowledge of packages
<qschulz>
target recipes do though
<qschulz>
RDEPENDS in other recipes use target packages and DEPENDS use recipes
<dacav>
Weird... I always pictured them as a symmetric concept. But then is it feasible to have a single recipe that produces a number of sub-packages?
<qschulz>
dacav: yes, use PACKAGES
<dacav>
Thanks, I'll have a look there :)
<qschulz>
dacav: it's actually the default, we create plenty of packages already (-dev, -dbg, -src, etc...)
<qschulz>
dacav: PACKAGECONFIG is for configuring your software (so maybe it shouldn't actually be called PACKAGECONFIG but RECIPECONFIG?)
<dacav>
Yes, as I understand it it is for propagating stuff like `--with-xxx` in autotools. In my case, cmake, I use -DFOO=ON etc
<dacav>
The reason why I thought of using PACKAGECONFIG is that the "bar" package above is native-only. It will not compile for the target at all
<qschulz>
dacav: I don't think native has a notion of packages
<qschulz>
dacav: what you could do though is separate your recipe in two
<qschulz>
one for target and one for native
<qschulz>
and specify PACKAGECONFIG varflags in it
<qschulz>
e.g. PACKAGECONFIG[bar] will only exist in the target recipe
<qschulz>
so that even if you select it for the native recipe, it won't do anything
<qschulz>
the limitation is that you cannot override varflags depending on the machine/recipe type AFAIR
<qschulz>
so you have to use something else instead
<dacav>
I don't know what varflags is (yet), but it sounds good
florian_kc has joined #yocto
<dacav>
Any recommended way to avoid the duplication of the recipe?
<dacav>
I mean, of the file content
<dacav>
Somewhat similar to a class, but not as complex
xcm has joined #yocto
<dacav>
But qschulz, maybe I'm overthinking: what if I just allow the build of 'bar' even for the target, even knowing it will fail? I don't expect anyone to DEPEND on bar, they should only DEPEND on bar-native. If they do, the build will just fail.
<qschulz>
dacav: bar in PACKAGECONFIG[bar] is a varflag
Schlumpf has quit [Quit: Client closed]
<qschulz>
dacav: you can use .inc files that you require
<qschulz>
see what u-boot is doing for example
<qschulz>
and then you only need do define what is different in each recipe
<qschulz>
so you would have a common.inc, target.bb and native.bb
<qschulz>
(obviously not named like that :) )
<dacav>
Ah, nice! I forgot about .inc! Sorry, too many things to think :D I'm not that pro yet
prabhakarlad has quit [Quit: Client closed]
<qschulz>
dacav: that's also a possibility indeed, up to you
d-s-e has quit [Ping timeout: 265 seconds]
<dacav>
I aim at being pro. It seems to be a really good career investment
dv_ has quit [Ping timeout: 255 seconds]
ArgaKhan has joined #yocto
camus has quit [Quit: camus]
<ArgaKhan>
I get the following error at 99% during compilation. Since I compile the kernel separately, I don't want to compile it with yocto. I don't know how to compile the latest version with yocto to our device anyway. Here is the error:
<ArgaKhan>
The following packages have unmet dependencies:
<ArgaKhan>
E: Unable to correct problems, you have held broken packages.
<ArgaKhan>
linux-dummy-dev : Depends: linux-dummy (= 1.0-r1) but it is not installable
prabhakarlad has joined #yocto
<kroon>
Anyone else having problems with grub reading ext4 images created using the latest e2fsprogs v1.47.0 ?
<kroon>
(using master-as-of-today recipes of grub and e2fsprogs)
<rburton>
ArgaKhan: maybe you've got dev-pkgs enabled and nobody tested that path with the dummy kernels. try adding ALLOW_EMPTY:${PN} = "1" to linux-dummy.bb
<DvorkinDmitry>
can I use RRECOMMENDS:${PN} += "kernel-module-..." in my kernel recipe? I have two kernel versions. One needs external kernel-module-x0, another needs internal kernel-module-x1. Depends on kernel version one or another should be installed into the image
<rburton>
sure
dv_ has joined #yocto
<DvorkinDmitry>
but when I add RRECOMMENDS:${PN} += "kernel-module-x0" into my kernel_v0.bbappend it doesn't work
Thorn has joined #yocto
<ArgaKhan>
rburton: Unfortunately I added this but nothing changed. I am still getting the same error.
goliath has joined #yocto
prabhakarlad has quit [Quit: Client closed]
Spinola_62 has quit [Quit: Leaving]
<dacav>
qschulz: I had a good lunch and split the recipe direclty into foo.bb and bar-native.bb, both including foobar.inc. Looks perfect, thanks!
Spinola has quit [Quit: Leaving]
jmiehe has quit [Quit: jmiehe]
kscherer has joined #yocto
hcg has joined #yocto
<kroon>
mkfs.ext4 from e2fsprogs v1.47.0 has a new feature enabled by default, "metadata_csum_seed". Unless I remove it, grub 2.06 is unable to read the ext4 filesystem
meego_ has joined #yocto
meego_ has quit [Remote host closed the connection]
seninha has quit [Remote host closed the connection]
seninha has joined #yocto
Poppy has joined #yocto
seninha has quit [Remote host closed the connection]
seninha has joined #yocto
<Poppy>
I'm trying to update FILES:${PN} from anonymous python function ( with d.appendVar('FILES:${PN}', "foo bar") ) but i got some warning about override variables (WARNING: Variable key FILES:${PN} ( foo bar ) replaces original key FILES: (${bindir}/* ${sbindir}/* ${libexecdir}/* ${libdir}/lib*${SOLIBS} .....)) there is a best way to update FILES from
<Poppy>
python function ?
JM20 has joined #yocto
tepperson_ has quit [Remote host closed the connection]
d-s-e has joined #yocto
tepperson_ has joined #yocto
kroon has quit [Quit: Leaving]
pidge has joined #yocto
sakoman has joined #yocto
tepperson_ has quit [Remote host closed the connection]
mrpelotazo has quit [Read error: Connection reset by peer]
mrpelotazo has joined #yocto
<mcfrisk>
how reproducible are master branch initramfs images? are the file timestamps handled reproducibly? I'm having problems getting TPM checksums of the initramfs reproducible. local and CI builds come with different values..
<rburton>
mcfrisk: a initramfs shouldn't be any different from any other image, so should be reproducible
<mcfrisk>
rburton: thanks for confirmation, I will diff the binaries.. something makes them build differently
<dacav>
I've got a old recipe here, whose do_install() creates a tempdir (via mktemp -d) to place some temporary files. Wouldn't it be just ok to use ${WORKDIR}?
<rburton>
probably
<rburton>
mktemp -d would put the files in /tmp by default, so if they're large you're potentially introducing cross-device copies too
<dacav>
The recipe adds a 'trap' to remove it. And it adds a subshell in order to avoid yocto to mess with the trap
<ArgaKhan>
rburton: I tried what you said but it didn't work. My question was a bit above, but if you don't remember, I can share it again. Any other ideas?
<yocton>
Hi! On the future Bugzilla 5 : One option that is really weird when using it for the first time is the one that advance to the "next" bug when submitting changes. Is it possible to have the "After changing a bug" option set to "Show the updated bug" by default?
tepperson has joined #yocto
<tepperson>
how do i configure ufw to be active on startup with a bbappend?
<DvorkinDmitry>
can I use RRECOMMENDS:${PN} += "kernel-module-..." in my kernel recipe? I have two kernel versions. One needs external kernel-module-x0, another needs internal kernel-module-x1. Depends on kernel version one or another should be installed into the image. but when I add RRECOMMENDS:${PN} += "kernel-module-x0" into my kernel_v0.bbappend it doesn't work - module is not included into the image
prabhakarlad has joined #yocto
<mischief>
does anyone know why setting INITRAMFS_IMAGE_BUNDLE causes the whole kernel to be rebuilt single-threaded? it seems like in the shell function in the kernel class, PARALLEL_MAKE isn't passed. i think that's fine, but the worse problem is that the whole kernel gets rebuilt just to bundle initramfs..
nemik has quit [Ping timeout: 268 seconds]
nemik has joined #yocto
<rburton>
mischief: sounds like you should file a bug with reproducers
<michaelo>
qschulz: thanks for notifying me about this documentation issue
amitk has quit [Ping timeout: 268 seconds]
nemik has quit [Ping timeout: 256 seconds]
nemik has joined #yocto
amitk has joined #yocto
<tepperson>
any idea why ufw would tell me "Extension limit revision 0 not supported" when trying to do ufw enable ?
Poppy has quit [Quit: Client closed]
<rburton>
tepperson: you might want to speak to the ufw people about that
Guest58 has joined #yocto
Guest58 has quit [Client Quit]
amitk has quit [Quit: leaving]
<RP>
rburton: nice :)
<rburton>
RP: actually surprised :)
ptsneves has quit [Ping timeout: 255 seconds]
florian_kc has quit [Ping timeout: 248 seconds]
<khem>
abelloni_: I see you dropped the opensbi and vte patches from master-next from your tree, did you see anything problematic with them ?
<JaMa>
17.58s without docs and whatever else you've enabled there
Vonter has quit [Ping timeout: 255 seconds]
kevinrowland has joined #yocto
nemik has quit [Ping timeout: 255 seconds]
nemik has joined #yocto
agrue_ has joined #yocto
agrue has quit [Ping timeout: 255 seconds]
nemik has quit [Ping timeout: 260 seconds]
<tepperson>
rburton: who do i contact regarding problems with ufw?
nemik has joined #yocto
<kevinrowland>
Hey, can I create a new task and give it access to `base_do_compile`? When I look at `run.do_compile` I see `base_do_compile` defined, but with a new task no such luck. I'd like to run some static analysis that essentially runs as a wrapper to the build command, and the easiest way to do that is to just call the existing `base_do_compile`
<kevinrowland>
function. I see that `do_compile` is autogenerated so I'm guessing this is difficult
<DvorkinDmitry>
can I use RRECOMMENDS:${PN} += "kernel-module-..." in my kernel recipe? I have two kernel versions. One needs external kernel-module-x0, another needs internal kernel-module-x1. Depends on kernel version one or another should be installed into the image. but when I add RRECOMMENDS:${PN} += "kernel-module-x0" into my kernel_v0.bbappend it doesn't work - module is not included into the image
hcg has quit [Quit: Client closed]
invalidopcode1 has quit [Remote host closed the connection]
invalidopcode1 has joined #yocto
Minvera has quit [Remote host closed the connection]
kscherer has quit [Quit: Konversation terminated!]
seninha has quit [Quit: Leaving]
kevinrowland has quit [Quit: Client closed]
seninha has joined #yocto
<rburton>
zeddii: turns out you can't build a big-endian kernel
<rburton>
zeddii: eg ERROR: linux-yocto-6.1.9+gitAUTOINC+d7393c5752_ccd3b20fb5-r0 do_package_qa: QA Issue: Endiannes did not match (1, expected 0) in /lib/modules/6.1.9-yocto-standard/kernel/fs/nls/nls_cp869.ko [arch] after setting a big-endian tune on qemuarm
<rburton>
hm should have expected 0x02, surely
<rburton>
but 0x01 is little-endian
prabhakarlad has quit [Quit: Client closed]
<tepperson>
where do i find documentation for "kernel config fragments" ?
dlan has quit [Quit: leaving]
dlan has joined #yocto
<zeddii>
rburton, yah. we've never tested that. I recall musing about doing a feature that switches in a differnt config frag, or a completely different qemu machine. but we never did complete it.
<rburton>
tepperson: the kernel section of the docs
rob_w has quit [Read error: Connection reset by peer]