GNUmoon has quit [Remote host closed the connection]
GNUmoon has joined #yocto
otavio has quit [Ping timeout: 272 seconds]
otavio has joined #yocto
Ram-Z has quit [Ping timeout: 276 seconds]
qschulz has quit [Read error: Connection reset by peer]
qschulz has joined #yocto
tangofoxtrot has quit [Remote host closed the connection]
tangofoxtrot has joined #yocto
tangofoxtrot has quit [Remote host closed the connection]
otavio has quit [Ping timeout: 272 seconds]
tangofoxtrot has joined #yocto
mrpelotazo has joined #yocto
seninha has quit [Remote host closed the connection]
mrpelotazo has quit [Ping timeout: 244 seconds]
starblue1 has quit [Ping timeout: 276 seconds]
starblue1 has joined #yocto
mrpelotazo has joined #yocto
Ram-Z has joined #yocto
sakoman has quit [Quit: Leaving.]
seninha has joined #yocto
sakoman has joined #yocto
seninha has quit [Remote host closed the connection]
seninha has joined #yocto
GNUmoon has quit [Read error: Connection reset by peer]
GNUmoon has joined #yocto
peoliye has joined #yocto
Circuitsoft has joined #yocto
seninha has quit [Ping timeout: 240 seconds]
seninha has joined #yocto
nemik has quit [Ping timeout: 260 seconds]
nemik has joined #yocto
nemik has quit [Ping timeout: 276 seconds]
nemik has joined #yocto
seninha has quit [Remote host closed the connection]
nemik has quit [Ping timeout: 268 seconds]
nemik has joined #yocto
nemik has quit [Ping timeout: 272 seconds]
nemik has joined #yocto
amitk has joined #yocto
sakoman has quit [Quit: Leaving.]
goliath has joined #yocto
peoliye has quit [Quit: Client closed]
qvwwsy has joined #yocto
<qvwwsy>
hello, after switching to kirkstone I get several kernels in the /boot/ dir: zImage, fitImage, uImage which wasn't the case with the honister. oe-pkgata-util says that they come from linux-yocto recipe which isn't much of a help. Any hints?
ptsneves has joined #yocto
ptsneves has quit [Ping timeout: 244 seconds]
ptsneves has joined #yocto
xmn has quit [Ping timeout: 276 seconds]
rob_w has joined #yocto
rob_w has quit [Remote host closed the connection]
tlwoerner__ has joined #yocto
<LetoThe2nd>
qvwwsy: do a bitbake -e on linux-yocto and look at the KERNEL_IMAGE_TYPES variable (or kinda similar)
<LetoThe2nd>
and, yo dudX
tlwoerner_ has quit [Ping timeout: 244 seconds]
paulg has quit [Ping timeout: 244 seconds]
paulg has joined #yocto
kami has joined #yocto
rob_w has joined #yocto
frieder has joined #yocto
amitk has quit [Ping timeout: 272 seconds]
amitk has joined #yocto
mateuszmar2 has joined #yocto
wkawka has joined #yocto
mckoan|away is now known as mckoan
<wkawka>
Hi, can I navigate what and how affects rootfs size? I have a strange situation where container which is even not on rootfs partition takes 30k blocks
<mckoan>
good morning
<wkawka>
I'm trying to investigate it
<LetoThe2nd>
wkawka: maybe enable buildhistory, and inspect the output? plus, look at the contents, which packages put the stuff into your resulting thing.
<LetoThe2nd>
wkawka: side note: your question is contradictory in itself. "what affects rootfs size"-"even not a rootfs partition"
dev1990 has joined #yocto
Schlumpf has joined #yocto
davidinux has quit [Ping timeout: 268 seconds]
goliath has quit [Quit: SIGSEGV]
LetoThe2nd has quit [Quit: WeeChat 3.5]
davidinux has joined #yocto
LetoThe2nd has joined #yocto
denisoft81 has joined #yocto
<denisoft81>
Hi all, I received an error message when I installed a python module on yocto hardknott with pip3. Kindly, how can I do? Thanks in advance https://pastebin.com/V2kkAM0r
<Tyaku>
Hi, What is the best practices with rsyslog and yocto ? For example, currently I would like to implement in my custom image "IMAGE_INSTALL += rsyslog" and configure rsyslog to transmit the logs to a fixed IPV4 address, but only if the image is built "on debug". Is it the good way to do it ?
<Tyaku>
Also, is it possible in a custom image (image.bb) to put a condition like "if (XXX) IMAGE_INSTALL" where XXX comes from local.conf
<khem>
we do have a variable called DEBUG_BUILD when set to 1 it build debug version of image but the notion of debug is different for different people
<khem>
you might be better off by defining separate image for your debug version
<khem>
and add it in there
dacav has joined #yocto
<LetoThe2nd>
Tyaku: you can "inject" information into the image recipe from local.conf. from my experience, the most maintainable approach is to package you dev-time modifications up as a custom DISTRO that you hold in a personal layer.
<khem>
I think its not dev thing that is being asked here, its more like a deployed image with and without rsyslog,
<LetoThe2nd>
khem: i interpreted the "if the image is built on debug" as a dev version. if we're talking about deployed stuff, then you're rightä
<Tyaku>
So, to build an image with/without rsyslog depending if it's production firmware or not,
<Tyaku>
The best option is to have custom-image-prod.bb and custom-image-dev.bb ?
<LetoThe2nd>
Tyaku: its the most common approach. I would suggest the -dev version being not independent, but directly require-ing the -prod one and only extend, but thats a minor thing.
LetoThe2nd has quit [Quit: WeeChat 3.5]
LetoThe2nd has joined #yocto
LetoThe2nd has quit [Client Quit]
LetoThe2nd has joined #yocto
<Tyaku>
LetoThe2nd I was watching it, but as the -prod is a .bb I can't put inherit custom-image-prod
<LetoThe2nd>
Tyaku: not inherit. require. see core-image-minimal-dev.bb for the blueprint :-)
<qschulz>
glembo[m]: depending on the scenario, you could use do_python_task[postfunc] += "some_sh_func;" if I'm not mistaken
<qschulz>
but still no argument though
florian_kc has joined #yocto
<qschulz>
i guess you could create a variable in the python function and put it in the datastore, and access it from the shell function with ${MYVAR} ?
<qschulz>
just throwing ideas right now
<qschulz>
(d.setVar for adding a variable to the datastore)
camus1 has joined #yocto
camus has quit [Read error: Connection reset by peer]
camus1 is now known as camus
<rburton>
glembo[m]: bb.build.exec_func()
<rburton>
literally hundreds of examples in oe-core
<rburton>
you'll need to throw values into the datastore to "pass arguments" though
nemik has quit [Ping timeout: 244 seconds]
nemik has joined #yocto
nemik has quit [Ping timeout: 268 seconds]
nemik has joined #yocto
<fleg>
Hi, I'm trying to use the `npmsw://` url in my SRC_URI, but some packages are in registries that require auth tokens. At first I thought that having a properly configured .npmrc would solve the problem, but after digging deeper I've found that retrieving dependencies stored in the shrinkwrap file is done "by hand" in fetch2/npmsw.py. Is there an
<fleg>
obvious way to use the auth token there, or do I have to get my hands dirty and work around it somehow? ;)
Schlumpf has quit [Quit: Ping timeout (120 seconds)]
manuel1985 has quit [Ping timeout: 244 seconds]
starblue1 has quit [Ping timeout: 272 seconds]
TobiQS has joined #yocto
starblue1 has joined #yocto
Circuitsoft has quit [Quit: Connection closed for inactivity]
<ptsneves>
RP: > To make this work you need to know "where it looked but didn't find" files
<ptsneves>
The thing is that with this new logic it will not even proceed to full parsing and it will not do so until a valid file is given for parsing. This seems to me exactly the same logic i implemented by making the not finding a file a fatal error.
<ptsneves>
I am sorry i did not see your message before :(
Schlumpf has joined #yocto
<ptsneves>
The shortcutting of parsing into a fatal is actually highlighted by the message Alexandre sent on the original conversion: It broke a poky test which tested a missing file on the download step. That poky test which now fails at the file checksumming step, and the exception is different and the reason the test now fails. Maybe this is the key behavior change to evaluate if you want to proceed with my change.
kami has joined #yocto
<RP>
ptsneves: There is no such thing as "full parsing"
seninha has joined #yocto
<RP>
ptsneves: There is a piece of the puzzle I don't think you understand here :/
<RP>
ptsneves: bitbake doesn't reparse recipes every time, each time you run it. Instead it forms a cache of the data it needs and detects changes to the recipes to know what to reparse
<ptsneves>
@RP Ok, i am aware of that. And what bitbake wants to know is if the files changed to know if needs to retrigger do-fetch
<RP>
ptsneves: Imagine FILESPATH = "a:b:c", SRC_URI = "file://myfile" and c/myfile exists and you run bitbake -p. If you then touch b/myfile, we need to reparse this recipe for the new checksum
<ptsneves>
@RP i follow :)
<RP>
ptsneves: it isn't even about running it, it is knowing when to reparse. To make things work, we not only need to know the checksum of c/myfile but also that b/myfile didn't exist
<RP>
if b/myfile exists, we know something changed and we have to reparse
kami has quit [Ping timeout: 252 seconds]
<LetoThe2nd>
does anybody happen to know a variable which has a nice evaluation history in a vanilla poky setup? nice in terms of, good for showing.
<ptsneves>
ok still following. Let us consider that:
<ptsneves>
When SRC=URI="file://myfile" and it does not exist we call this State A.
<ptsneves>
When SRC_URI="file::/myfile" exists and is modified we call State C.
<ptsneves>
When SRC_URI="file://myfile" and it starts existing we call State B.
<ptsneves>
I argue that with the current patch all combinations of states and their changes are observable or irrelevant, but appropriate parse cache invalidation will happen
<ptsneves>
The result of get_checksump_file_list:
<ptsneves>
file list@STATE A = ?? // the program is aborted due to the fatal. No parse cache matters and current state is a dont care
<ptsneves>
file list@STATE B = ['file://myfile'] // The contents are now checksummable and the file exists. The list has a new element
<ptsneves>
file list@STATE C = ['file://myfile'] // The contents changed and the checksuming results will decide whether the cache is invalidated. The list does not change.
<ptsneves>
STATE A -> STATE B/C => observable due to list becoming available
<ptsneves>
STATE B <-> STATE C => list not changing, checksumming to decide if file changed
<ptsneves>
STATE B -> STATE A => not observable for cache purposes but irrelevant since bitbake dies of fatal error
<ptsneves>
oh damn my client broke this down. Sorry guys
jpuhlman has quit [Remote host closed the connection]
jpuhlman has joined #yocto
<RP>
ptsneves: I don't think that does match what I'm talking about, it doesn't work like this
<ptsneves>
@RP Ok that is fine. If you wish to revisit the topic let me know
<RP>
ptsneves: go back to the scenario I mentioned. c/myfile existed and was parsed. We then create b/myfile. How does bitbake know to reparse things and that things have changed?
<RP>
ptsneves: we can clean up the code, we just can't drop this piece of functionality
<ptsneves>
@RP totally agree :) I have been on the biting end of parsing cache not being invalidated when it should :) I will try to understand it a bit better but i doubt the patch can stand if you are right
fitzsim has quit [Remote host closed the connection]
<ThomasRoos[m]>
I know I can use ipkg instead of rpm - then the error is gone.
<landgraf>
ThomasRoos[m]: you should rebuild corretto without X11 support (either use configure and/or some kind of packageconfig/flag). again --skip-broken will not help and corretto will not be installed
<landgraf>
rpm adds autogenerated dependency, I'm not sure if it can be filtered using yocto
<landgraf>
oh. it's binary . well...
<ThomasRoos[m]>
yes...
<ThomasRoos[m]>
Rebuilding is the best solution, but not achievable without big efforts.
goliath has quit [Quit: SIGSEGV]
<ThomasRoos[m]>
I wonder why there is no simple way to disable this check - I just want to install files and dnf shouldn't care about if they can be executed...
<landgraf>
ThomasRoos[m]: It's DNF's job to handle dependencies. if you want to install files use rpm --nodeps
<landgraf>
ThomasRoos[m]: another option - just extract the rpm somewhere (rpm2cpio <rpmfile> | cpio -id) and use it
<ThomasRoos[m]>
landgraf: yes, how to pass this when creating a image?
<ThomasRoos[m]>
I was hoping that I can set a env variable or bitbake var to disable this DNF check - enable (-skip-broken)
<rburton>
ThomasRoos[m]: use the bin_package class to repackage the rpm and delete the bits that are causing you problems
jmiehe has quit [Ping timeout: 272 seconds]
JohnSmith[m] has joined #yocto
<ThomasRoos[m]>
<rburton> "Thomas Roos: use the bin_package..." <- I don't use the rpm version of corretto - I just use the binary package as a zip file. Sorry - not specific about this. So my only chance to create an image without using package_rpm and use package_ipk instead or delete the libs that causing problems.
qvwwsy has quit [Ping timeout: 268 seconds]
<rburton>
ThomasRoos[m]: delete the libs causing problems
<rburton>
that's not a workaround, it's the solution
<ThomasRoos[m]>
yes, I like this the best - but how to do this automatically? Try and error?
<LetoThe2nd>
ThomasRoos[m]: rather "seek and destroy"
<rburton>
well, then you've no excuse for not doing it properly ;)
jpuhlman has quit [Remote host closed the connection]
<ThomasRoos[m]>
...yes
<rburton>
it's just a jdk, how hard can it be to build <runs manically>
<rburton>
meta-java does build a jdk from scratch, so it's not impossible, just a... challenge.
<ThomasRoos[m]>
I know
<ThomasRoos[m]>
;)
tgamblin has joined #yocto
Tyaku has quit [Quit: Lost terminal]
jpuhlman has joined #yocto
<ptsneves>
ThomasRoos[m]: funny thing. ipk has a EXTRAOPKGCONFIG variable to add extra parameters to opkg but rpm does not have. You can append to ROOTFS_PKGMANAGE if you want but this will break if you change package manager in the future, given aguments are likely not compatible
<rburton>
don't work around this with rootfs tweaks
<rburton>
just fix the packaging
<ptsneves>
:) sorry only caught up with the conversation now
tgamblin has quit [Ping timeout: 260 seconds]
<ptsneves>
by the way, how do you guys extract the patches from emails to apply on your trees? Do you copy paste them?
<RP>
ptsneves: save the email and use git am
<ptsneves>
thanks
<rburton>
some people use lore/b4
<rburton>
or patchworks
<rburton>
i had a tool which meant i could tag emails in my mailer and then run a script which would fetch the mails and apply them
<JPEW>
I have an embarssing `git mutt` alias :)
goliath has joined #yocto
seninha has joined #yocto
<ptsneves>
Yes. I just tried saving the email in .txt from thunderbird and the file is not recognizable by git :(
<ptsneves>
i guess i will use lore.
<qschulz>
ptsneves: right click, save selected messages, plain text format
<qschulz>
should be enough, worked for me (for other mailing lists than oe/yocto though)
rob_w has quit [Remote host closed the connection]
<JaMa>
saving from mutt works fine for me and if I want to apply it "remotely", then downloading mbox format from patchwork is 2nd best option for me
<denisoft81>
Hi all, I would need to install a python module on the yocto image to run a test script and if everything works correctly I append the appropriate recipes. When installing using pip3 I get this error. How can I fix it? https://pastebin.com/8A1N2Hpi
<ptsneves>
@JaMa that seems to work :) thanks
<qschulz>
ptsneves: tried with this exact patchset (I assume) and it exported it fine? "works on my machine"TM :)
<qschulz>
denisoft81: are you running pip3 on your target directly?
TobiQS has quit [Remote host closed the connection]
<ptsneves>
@qschulz obviously we need to containerize thunderbird to make sure your machine is the ruling one
<denisoft81>
Hi qschulz, thanks for your reply. Yes, I run pip3 directly on target
<qschulz>
denisoft81: going to give you some headaches :/
<qschulz>
denisoft81: here it complains you don't have some gcc on your target to compile the python module
manuel1985 has joined #yocto
<qschulz>
ptsneves: there's a flatpak on flathub :)
<ptsneves>
@qschulz :D
<ptsneves>
oh damn i cannot apply bitbake's paches directly into poky
<ptsneves>
bah i give up i will butcher the patch manually
<qschulz>
ptsneves: git am --directory=bitbake ? (just from the manpage, haven't tried)
<denisoft81>
thanks qschulz
nemik has quit [Ping timeout: 272 seconds]
nemik has joined #yocto
kscherer has joined #yocto
sakoman has joined #yocto
nemik has quit [Ping timeout: 240 seconds]
nemik has joined #yocto
<ptsneves>
maybe. I just manually edited the patch.
<RP>
qschulz: that sounds a lot simpler than some of the stuff I end up doing
<qschulz>
ptsneves: in the worst case, I think patch should be able to handle those too, just go into bitbake and then patch -p 1 < my.patch ?
<ptsneves>
@qschulz yep patch has -p1 but here it is the other way around i think. I would need the equivalent of patchdir :)
<ptsneves>
i have another patch i need to bring to poky so i will try the --directory
TobiQS has joined #yocto
<qschulz>
ptsneves: you always need -p 1 because of a/ and b/ in the paths IIRC? since patch patches relative to the current directory, if you cd into bitbake dir in poky git dir it should be fine
<ptsneves>
your are right
tjqurm has joined #yocto
<ptsneves>
@qschuz it worked! Another tool in the bag :)
<qschulz>
ptsneves: nice! :muscle:
manuel1985 has quit [Ping timeout: 276 seconds]
zeddiii is now known as zeddii
vladest has quit [Ping timeout: 255 seconds]
mrpelotazo has quit [Read error: Connection reset by peer]
mrpelotazo has joined #yocto
ptsneves has quit [Ping timeout: 272 seconds]
TobiQS has quit [Remote host closed the connection]
fenrig has joined #yocto
<fenrig>
Hi, please refer me to other channel (#meta-oe?) if I'm asking in the wrong location. Im trying to add lighttpd-module-authn_file in kirkstone
<fenrig>
So I'ved expanded the packageconfig to incorporate "nettle" and add the -Dwith_nettle=true compile flag. It compiles authn_file (checked in the temp/log.do_compile). On top of that I have the mod_authn_file.so in <image>/usr/lib/lighttpd/
<fenrig>
I believe this line should pick that module up an create the yocto package (lighttpd-module-authn_file), but when I create my image I still get "nothing provides lighttpd-module-authn_file needed by ..."
<qschulz>
fenrig: there might be some issues around the filename being named authn_file
<qschulz>
underscores are invalid characters for packages
<fenrig>
oh so the code to generate the filename wont work as well
<fenrig>
"the line"
<qschulz>
so either it's messed up or we have some magic stuff that replaces the underscore by a dash
<qschulz>
oe-pkgdata-util list-pkgs lighthttpd should give you the list of packages created by the recipe
<fenrig>
it doesnt list the modules in fact of lighttpd
<fenrig>
Im only getting lighttpd from that package
<fenrig>
though if I do bitabke lighttpd-module-authn_file it says this "lighttpd RPROVIDES lighttpd-module-authn_file"
<qschulz>
mmm, --recipe lighttpd maybe instead of just lighttpd?
<fenrig>
lighttpd-module-authn-file
<fenrig>
oh it works indeed :D i just need to put - instead of _
<fenrig>
I guess I took that "line" too literal
<fenrig>
ty for your help
kranzo has joined #yocto
<qschulz>
fenrig: well, if you don't know underscore is an invalid character for package name and that there is some magic in do_split_packages, that was actually a safe assumption :)
<qschulz>
very welcome :)
<fenrig>
I forgot about the _ , indeed I did not know of this magic :D but it works now, super happy
tnovotny has quit [Quit: Leaving]
seninha has quit [Quit: Leaving]
<kranzo>
Hi all,
<kranzo>
does a cleanall the sstate for the recipe or just the specific configuration active right now?
<rburton>
all sstate
<RP>
kranzo: use cleansstate, not cleanall and it is for the recipe
tjqurm has quit [Ping timeout: 244 seconds]
<kranzo>
thats good thx, now the tricky question, i'm facing a basehash change because of some tinkering with env vars.
<kranzo>
can i apply the env var to the taskhashes of a single recipe instead of all (as it seems to work when exported in the *.conf)
jpuhlman_ has joined #yocto
jpuhlman is now known as Guest642
jpuhlman_ is now known as jpuhlman
Guest642 has quit [Killed (copper.libera.chat (Nickname regained by services))]
<kranzo>
fenrig: but you read it :)
otavio has quit [Ping timeout: 240 seconds]
otavio has joined #yocto
borsnw has joined #yocto
otavio has quit [Ping timeout: 244 seconds]
otavio has joined #yocto
<borsnw>
so, I still want to get rid of the kernels be put into the /boot. linux-yocto recipe generates kernel-image-fitimage, kernel-image-uimage and kernel-image-zimage. I want to get rid of these package. bitbake -e linux-yocto shows KERNEL_IMAGETYPES="zImage uImage fitImage". I found in the meta/poky/meta/classes/kernel.bbclass the following https://pastebin.com/jZNAeXWF
<borsnw>
it says - allow machine to override deps if kernel image is not wanted. So how do I override?
<borsnw>
# printf "!pe"
<rburton>
borsnw: don't install kernel-image (assuming a recent release). PACKAGE_EXCLUDE = "kernel-image-*" works in an image recipe
florian_kc has quit [Ping timeout: 240 seconds]
<borsnw>
yes, kirkstone
<borsnw>
should this be bbappend to the linux-yocto.bb ?
<rburton>
no, image
<borsnw>
what image? sorry, not familiar with yocto
florian has quit [Quit: Ex-Chat]
<rburton>
the image you're building
seninha has joined #yocto
otavio has quit [Read error: Connection reset by peer]
<sakoman>
/home/pokybuild/yocto-worker/no-x11/build/build/tmp/work/core2-64-poky-linux/ruby/3.1.2-r0/recipe-sysroot-native/usr/bin/ruby: error while loading shared libraries: libcapstone.so.4: cannot open shared object file: No such file or directory
<RP>
sakoman: well, that explains a few things
geoffhp has joined #yocto
<RP>
and raises other questions too
<sakoman>
RP: indeed it does :-(
geoffhp has quit [Client Quit]
* sakoman
sees continuing wack-a-mole in our future
xmn has quit [Ping timeout: 276 seconds]
* sakoman
queues a build, sees it won't start for quite some time, goes for food
nemik has quit [Ping timeout: 272 seconds]
nemik has joined #yocto
Maxxed has joined #yocto
kscherer has quit [Quit: Konversation terminated!]