qschulz has quit [Remote host closed the connection]
nemik has quit [Ping timeout: 240 seconds]
nemik has joined #yocto
qschulz has joined #yocto
nemik has quit [Ping timeout: 240 seconds]
nemik has joined #yocto
wkawka has quit [Ping timeout: 252 seconds]
nemik has quit [Ping timeout: 252 seconds]
nemik has joined #yocto
starblue has quit [Ping timeout: 276 seconds]
starblue has joined #yocto
<denix>
🤦
seninha has quit [Quit: Leaving]
sakoman has quit [Quit: Leaving.]
nemik has quit [Ping timeout: 256 seconds]
nemik has joined #yocto
nemik has quit [Ping timeout: 252 seconds]
nemik has joined #yocto
sakoman has joined #yocto
_whitelogger has joined #yocto
<vvn>
whatever I do, qtwebengine recompile every times. It's now compiling qtdeclarative-5.15.3+gitAUTOINC+32b37e20f7-r0 for example. Am I doing something wrong?
Tokamak has quit [Read error: Connection reset by peer]
Guest8962 has quit [Read error: Connection reset by peer]
dev1990 has joined #yocto
adrian__ has joined #yocto
OnkelUlla has joined #yocto
goliath has joined #yocto
bps has joined #yocto
bps has joined #yocto
bps has quit [Changing host]
mckoan|away is now known as mckoan
<mckoan>
good morning
leon-anavi has joined #yocto
frieder has joined #yocto
bps has quit [Ping timeout: 240 seconds]
cmd has quit [Remote host closed the connection]
pbergin has joined #yocto
florian_kc has joined #yocto
Rolle79 has joined #yocto
olani has joined #yocto
<Rolle79>
I am cross-compiling a toolchain for arm. I have specified TARGET_OS=linux-gnueabi and TCLIBC=glibc, but still the SDK ends up with a musl and a musl-eabi toolchain in addition to the linux-gnueabi. How can I disable this?
Vonter has quit [Ping timeout: 246 seconds]
thomas__ has joined #yocto
thomas_ has quit [Ping timeout: 240 seconds]
nemik has quit [Ping timeout: 256 seconds]
nemik has joined #yocto
Schiller has joined #yocto
Vonter has joined #yocto
florian_kc has quit [Ping timeout: 276 seconds]
Guest0815 has joined #yocto
Bardon_ has joined #yocto
Bardon has quit [Ping timeout: 276 seconds]
bps has joined #yocto
bps has joined #yocto
<Guest0815>
Hello, I noticed a call of bitbake does not use all cores for compiling in parallel, instead it seems it's only a single-core compiling. As my PC is quadcore, bitbake would be much more faster with parallel compiling. Is there a bitbake option for that?
<Guest0815>
(release is honister)
<landgraf>
Guest0815: THere's bitbake option for that. However default value is to use parallel compiling
<landgraf>
Guest0815: BB_NUMBER_THREADS and PARALLEL_MAKE
<florian>
derRichard: I had a similar issue a few days ago... (from germany).
wkawka has joined #yocto
nemik has quit [Ping timeout: 252 seconds]
nemik has joined #yocto
kroon has joined #yocto
wkawka has quit [Quit: Client closed]
florian_kc has quit [Ping timeout: 248 seconds]
xmn has quit [Quit: ZZZzzz…]
<kayterina[m]>
pyinstaller has a COPYING.txt which was added by devtool in the recipe. It says that pyinstaller is GPLv2+ , so I also add LICENSE = "GPLv2+" ?
<kayterina[m]>
changed Uknown to GPLv2+ is ok?
kroon has quit [Quit: Leaving]
<qschulz>
kayterina[m]: not sure to follow what you mean by "COPYING.txt which was added by devtool"
<kayterina[m]>
devtool add found a copying text and added that to the recipe with th ecomment I should check the licenses
<kayterina[m]>
the text is from the package
<kayterina[m]>
*copying.txt
R0b0t1 has joined #yocto
<R0b0t1>
How do I use bitbake-layers to add an existing openembed layer to my project?
<R0b0t1>
I'm trying to modify the distro files from stmicro
<qschulz>
kayterina[m]: unfortunately, it is not GPL-2.0 :/
<kayterina[m]>
a,ok. So I keep their license file copying.txt in the recipe and it is ok?
<qschulz>
kayterina[m]: I would actually set it to GPL-2.0-or-later-with-bootloader-exception and then use NO_GENERIC_LICENSE to add the COPYING.txt to the license files
<qschulz>
kayterina[m]: yup, that's usually what should be done
<qschulz>
you always point to in-tree (in the sources) files for the license
davidinux has joined #yocto
<qschulz>
if it's not a common license, I think you need to play with NO_GENERIC_LICENSE or BitBake might complain
<qschulz>
R0b0t1: bitbake-layers --help ? :)
<R0b0t1>
yes but open questions for that command are what do I name the directory, why, and what do I put in it
<R0b0t1>
I did find that meta-openembedded was already in the project with all contents, including meta-networking; I need to edit the recipes to include pppd, how do I do that?
<R0b0t1>
I don't mind reading the docs but finding what exactly I need to read and why is proving to be a big issue
amahnui1 has quit [Quit: Connection closed for inactivity]
<qschulz>
R0b0t1: you need to clone the git repo for meta-openembedded manually in one directory
<qschulz>
then bitbake-layers add-layer $SOMEPATH/meta-openembedded/meta-networking should do it?
<R0b0t1>
ok, good, that is what I came up with before I noticed the project included meta-openembedded
<R0b0t1>
last part hmm
<R0b0t1>
how do I select which things in a layer I want
<R0b0t1>
the entirety of meta-networking is pretty big
<R0b0t1>
looks like meta-networking is already added from show-layers but pppd is not selected
<qschulz>
if it does not, bitbake will complain it cannot find the file
<qschulz>
R0b0t1: you shouldn't need to import only a selection of recipes from a layer
<qschulz>
R0b0t1: note that including a layer does not mean building the recipes
<qschulz>
it means the recipes will be parsed and are available if they need to be built
<qschulz>
if you need to add something to your image
<qschulz>
you need: 1) a recipe (probably already aailable in some layer youi need to add), 2) add it to your image one way or the other (usually via IMAGE_INSTALL in your image recipe)
<R0b0t1>
so where do I specy it's install? add my own layer, and that layer pulls it in?
<wkawka>
Hi, I'm trying to integrate previously builded u-boot image with builded kernel version. Both of them are working if they are alone, but when I simply copy u-boot on card, system cannot boot throwing: libfdt fdt_check_header(): FDT_ERR_BADMAGIC. I have to discover how can I rebuild either of those to get them work together.
<wkawka>
When I'm trying to boot it manually, using booti command I'm getting : Bad Linux ARM64 Image magic!
<wkawka>
I'm new to Yocto and I am getting confused, it is so much to learn there.
davidinux has quit [Ping timeout: 252 seconds]
JPEW has quit [Read error: Connection reset by peer]
JPEW has joined #yocto
Ch^W has quit [Ping timeout: 240 seconds]
nemik has quit [Ping timeout: 252 seconds]
nemik has joined #yocto
Ch^W has joined #yocto
nemik has quit [Ping timeout: 240 seconds]
nemik has joined #yocto
nemik has quit [Ping timeout: 240 seconds]
nemik has joined #yocto
nemik has quit [Ping timeout: 246 seconds]
nemik has joined #yocto
GillesM has joined #yocto
GillesM has quit [Client Quit]
seninha has joined #yocto
seninha has quit [Remote host closed the connection]
seninha has joined #yocto
nemik has quit [Ping timeout: 256 seconds]
nemik has joined #yocto
nemik has quit [Ping timeout: 240 seconds]
nemik has joined #yocto
vvn has quit [Quit: WeeChat 3.5]
vvn has joined #yocto
thomas__ has quit [Ping timeout: 256 seconds]
<dvorkindmitry>
I have public GIT repo at github. What is 100% correct syntax to use in SRC_URI? I've got error on fetch with SRC_URI="git://github.com/me/myrepo.git;protocol=https;branch=master". Github says something about port that should not be used
<qschulz>
dvorkindmitry: can you actually git clone with https protocol outside of yocto?
<dvorkindmitry>
qschulz, what url should I use for clone to do exactly like Yocto does?
<dvorkindmitry>
no, It says, I need to use publickey
<qschulz>
dvorkindmitry: then you probably need to enable https access on your github repo for this to work
<qschulz>
otherwise, you could use protocol=ssh but you'll likely need to handle different ssh pubkeys and whatnot, usually not the best
<qschulz>
especially when it's a public git repo
<dvorkindmitry>
qschulz, oh! I made a mistake. When U was clonning like this:
xmn has joined #yocto
<dvorkindmitry>
git clone https://github... it cloned correctly.
<dvorkindmitry>
but when I try "git clone github.com:me/myrepo.git" It said about the key.
<qschulz>
yeah that's normal
<qschulz>
maybe bitbake -DDDDDD my-recipe -c fetch -f could give you the exact command that is run
<qschulz>
so you can do it again on your PC locally and check if it works
<dvorkindmitry>
but my SRC_URI is git://github.com/me/myrepo.git;protocol=https;branch=master; and it looks like it really use the git:// method
<rburton>
what does bitbake actually say?
<rburton>
guessing the problem from an error we can't see is tricky
rob_w has quit [Quit: Leaving]
<dvorkindmitry>
rburton, give me 5 mins...
<Schiller>
does anyone have experience with customizing the schedulers.py from the yocto-controller?
leon-anavi has quit [Quit: Leaving]
otavio has quit [Quit: leaving]
otavio has joined #yocto
sakoman has joined #yocto
Sukru has quit [Ping timeout: 252 seconds]
pbergin has quit [Quit: Leaving]
<Schiller>
basically i just want to add a REVISION to the <branchdefaults> List in the schedulers.py. Can't find the right syntax.
florian_kc has joined #yocto
amahnui1 has joined #yocto
Tyaku has joined #yocto
seninha has quit [Quit: Leaving]
<kayterina[m]>
I have built three recipes, pyinstaller, altgrapch and pyinstaller-hooks-contrib. There is a pyinstaller executable in WORKDIR/image/usr/bin/pyinstaller . How can I use it in my-recipe.bb:do_compile()?
<qschulz>
kayterina[m]: you need to DEPENDS on pyinstaller-native in my-recipe
<qschulz>
then simply use pyinstaller directly
Schiller has quit [Ping timeout: 252 seconds]
nemik has quit [Ping timeout: 248 seconds]
nemik has joined #yocto
nemik has quit [Ping timeout: 240 seconds]
nemik has joined #yocto
ptsneves has quit [Ping timeout: 252 seconds]
<kayterina[m]>
pyinstaller-native is the one that I made a recipe for,? Because I named it just "pyinstaller" and DEPENDS = "pyinstaller" gives the same error.
<dvorkindmitry>
rburton, Fetcher accessed the network with the command LANG=C git -c core.fsyncobjectfiles=0 clone --bare --mirror "https://github.com/mmm/kkk.git
<qschulz>
kayterina[m]: if you need to run something at build time, it needs to be compiled for the host. That is a native recipe.
<qschulz>
If you don't need pyinstaller on the target, ever. You can just name your recipe pyinstaller-native.bb and have an inherit native in there
<rburton>
dvorkindmitry: that's not the whole log, or the real src_uri
<dvorkindmitry>
rburton, yes. Trying to generate good log for you. It is not easy, case I have multiconfig and several recipes that uses private repos with key at ~/.ssh/. If I drop the key from ~/.ssh/, some other recipes stop my build before fetch is done for the target recipe...
davidinux has joined #yocto
<qschulz>
dvorkindmitry: bitbake my-recipe -c fetch -f bypasses all this
<dvorkindmitry>
qschulz, I do like this: bitbake -DDDDDD -c fetch -f mc::myrecipe. (:myarch: is set to default in local.conf) Is it wrong?
<dvorkindmitry>
it try to refetch another recipes, that have no dependency to the target one. and fails case I dropped ssh KEY
fitzsim has quit [Ping timeout: 240 seconds]
<kayterina[m]>
if I understand correctly, an inherit native in the recipe does not need a BBCLASSEXTEND = "native"
<rburton>
correct
<rburton>
kayterina[m]: feel free to share the recipe so we can check it makes sense
<rburton>
personally i'd say use bbclassextend as it's more generic. someone might want a nativesdk version, or a target version.
davidinux has quit [Ping timeout: 248 seconds]
davidinux has joined #yocto
nemik has quit [Ping timeout: 252 seconds]
nemik has joined #yocto
nemik has quit [Ping timeout: 246 seconds]
nemik has joined #yocto
ilunev has joined #yocto
<ilunev>
Hello! I'm trying to run "devtool modify linux-yocto" in a setup where there's 2 bbappends to the recipe. For some reason, devtool tries to find odroid-kmeta/bsp/odroid-h2/bsp/common-pc-64/common-pc-64-cpu.cfg , all after odroid-h2 being garbage, I think, from generic x86_64 config. Why does it look for kmeta stuff in a recipe folder and why it even attempts to do so? There are 2 kmeta in SRC_URI, machine is odroid-c4.
<ilunev>
ah. odroid-h2.scc contains "include bsp/common-pc-64/common-pc-64.scc" but I think it wants to read it from generic linux-yocto kmeta, but there isn't such a file in meta-odroid.
<fray>
Is there a way to dump the hash equivalency data for a few recipes/tasks? We're trying to do some valiation that two different builds produced the same contents and this seems to be th eeasiest way for us
pgowda_ has quit [Quit: Connection closed for inactivity]
<dvorkindmitry>
If I drop the key from ~/.ssh/ I can't build my target recipe althrough it has no dependency for another private recipes that is used in the image. What would you recommend?
<sveinse>
Are there any ways to restrict/shape the resources for particular recipe when building? E.g. reduce the number of parallell jobs? I have problems building nodejs which is by far the biggest package in my setup, and the linux oom kills the build from time to time requiring me to start over.
<rburton>
you can use overrides in local.conf to set PARALLEL_MAKE per-recipe
<rburton>
or use -l instead of -j, which looks at the load
<rburton>
assuming gnu make, that is.
<rburton>
it would be great if all the build tools supported a common way of saying 'spawn up to N threads whilst the system load is less than L', but they don't
<rburton>
gnu make can't even agree in a single build with -j10 not to spawn 10*10 jobs
<sveinse>
yeah, i'm seeing linux loads of 40-80, which is by normal standard extremely high
<sveinse>
rburton: I think maybe the issue with make is when you have make->bash->make constructs, where the outer job context isn't passed through the shell layer to the inner make andj job controls goes havok
<rburton>
then they're not using the same recipe you are
<rburton>
as they don't have protocol=https
<dvorkindmitry>
rburton, thank you!
davidinux has joined #yocto
<sveinse>
I normally don't experience gcc and building to be particularely memory intesive. We installed 128G ram in another build server, but linux is using most of it for fs buffering and not for compiling. It's seldom to see gcc consuming more that 1G of mem. In fact putting build/tmp inside tmpfs had a much larger effect than anything else. So even ending up with a tmp tree of 50Gb it doesn't get starved for
<sveinse>
memory.
<sveinse>
Nodejs seems to be the notable exception evidently.
<rburton>
anything involving JS :)
Gambino has joined #yocto
<JaMa>
sveinse: depends on what you're building I guess, I can easily trigger OOMK with -j 64 build of chromium; 2G per cpu core is no longer enough for some C++ code
<sveinse>
JaMa: jep. I'm building vanilla nodejs in kirkstone
<sveinse>
Even Qt isn't as heavy
<JaMa>
qtwebengine triggers OOMK as well
<sveinse>
ok, that I haven't seen. But I'm building Qt a lot thou
<JaMa>
now I'm lowering P_M for webruntime/qtwebengine to about 2/3 cpu threads just to avoid OOMK getting triggered
<sveinse>
Has there been any initatives to look at the parallell tasks vs parallell jobs recently?
<JaMa>
it was fine with gcc-10 I think, since 11 it's worse, not sure how much worse 12 (but -j 40 still works with langdale)
<sveinse>
I directed a test a year ago for compiling yocto in azure vm, where I tried many permutations on CPU from 1 through 64 VCPUs to see what gave the most efficient result. I think we landed on that for that setup, approx 16 VCPUs seems to give the best efficiency.
<sveinse>
Efficiency = cost vs time
<JaMa>
the job server and -l was mentioned few times, but I don't remember seeing it recently, I assume it waits for you to volunteer :)
<JPEW>
sveinse: That matches my observations. We build on 16 core 32 GB RAM VMs
<JaMa>
it depends highly in what you typically build (and how much sstate you reuse) in my experience
<sveinse>
The -l can be tricky, as there are so many disparate build systems in the phlethora of different packages we build
<sveinse>
JaMa: yes, we try to maxiumize sstate sharing and reusage. But that test, specifically, were done without any sstate to compete on equal terms.
nemik has quit [Ping timeout: 260 seconds]
<JaMa>
16c is nice for "smaller" images, but once chromium gets rebuilt in most of your builds (for whatever reason) then icecc or higher cores really help rebuilding it 1000 times a day
nemik has joined #yocto
<sveinse>
Yocto build is somewhere inbetween being CPU bound and IO bound, not heavy optimization on either doesn't really give that much benefit
<JaMa>
but I'm also biased and spoiled from having 10+, 60+core machines just for my personal builds until recently, building in cloud and paying per-core would force me to do fewer builds :)
<sveinse>
We needed to reinvest in new server, and at the volume of builds that we're doing (handful a day), our estimates seems to indicate that it will be cheaper to use cloud computing for this rather than investing in new metal. It will have an impact on build times, but that is acceptible.
frieder has quit [Remote host closed the connection]
davidinux has joined #yocto
<RP>
sveinse: it does depend a lot on your build. Memory does help and the system is more cpu bound than IO on most systems I've looked at but it does depend on what you build
tgamblin has joined #yocto
<JaMa>
it also depends on how many engineers are blaming slow CI for not being productive :)
<sveinse>
RP: yes. As I mentioned earlier, we added 128G to a server which didn't use it all. Rather putting build/tmp on tmpfs had a much much larger impact on performance, despite the 30G-60G drainage on available memory for builds.
davidinux has quit [Ping timeout: 240 seconds]
mckoan is now known as mckoan|away
<RP>
sveinse: you might be able to better tune the filesystem to match the tmpfs fwiw
<sveinse>
RP: what fs do you normally use for build tmps?
<JaMa>
barrier=0 (commit=10000)
wkawka has quit [Quit: Client closed]
kevinrowland has quit [Quit: Client closed]
<RP>
sveinse: ext4
ilunev has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<RP>
and noatime
<sveinse>
How is it with journal, is that impeding performance in high fs activity?
gsalazar has quit [Ping timeout: 256 seconds]
xmn has quit [Ping timeout: 260 seconds]
<RP>
sveinse: I've not found it to be an issue, bigger factors are the commit time and barriers as JaMa mentions
<sveinse>
thanks
florian_kc has quit [Ping timeout: 240 seconds]
<JaMa>
the biggest pain with long commit is that when you finish a build and then e.g. reboot, the umount takes long time and when it's writting all that data you don't plan to use after the reboot to your precious NVME drive, then tmpfs is better option (especially for builds where you don't plan to use the artifacts for anything and only want to see if it builds OK and maybe the cooker log from it)
<JaMa>
often I hope that rm_work removed most of the useless data before it got written to actual drive, but tmpfs makes it more explicit that I don't care about that data later
Rolle79 has quit [Quit: Client closed]
florian_kc has joined #yocto
Tokamak_ has quit [Ping timeout: 246 seconds]
Tokamak has joined #yocto
nemik has quit [Ping timeout: 260 seconds]
nemik has joined #yocto
Tokamak_ has joined #yocto
Tokamak has quit [Ping timeout: 256 seconds]
nemik has quit [Ping timeout: 240 seconds]
nemik has joined #yocto
goliath has joined #yocto
starblue has quit [Ping timeout: 246 seconds]
starblue has joined #yocto
nemik has quit [Ping timeout: 240 seconds]
nemik has joined #yocto
Telgareith has joined #yocto
<Telgareith>
what's a good starting point for yocto for raspberry pi? The only source I can find is a 2016 book, and I imagine things are quite different than then
<Telgareith>
does it even have to be raspberry pi, it'd be nice to be able to throw an image into vmware and test it
<Telgareith>
also, will there be recording of the lectures of the summit?
<Telgareith>
dream job is giving bonus points for knowing yocto... And I'll probably be the only person who knows what it is nevermind is learning how to use it (I need to learn it for another project anyway)
<moto-timo>
This is in a console for a Digital Ocean droplet.
nemik has quit [Ping timeout: 276 seconds]
nemik has joined #yocto
<Telgareith>
do yocto builds ususally consume very large amounts of ram? I'm on a 16C/32T machine and its currently using 95gb ram
<Telgareith>
running from inside WSL2
bps has joined #yocto
bps has joined #yocto
goliath has quit [Quit: SIGSEGV]
seninha has joined #yocto
ilunev has quit [Read error: Connection reset by peer]
ilunev has joined #yocto
nemik has quit [Ping timeout: 240 seconds]
nemik has joined #yocto
florian_kc has joined #yocto
nemik has quit [Ping timeout: 260 seconds]
nemik has joined #yocto
ilunev has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<LetoThe2nd>
Telgareith: sounds about okay, if your box has that amount of ram
<Telgareith>
it does
<Telgareith>
but the builds over and its still hovering there
<LetoThe2nd>
Telgareith: I guess you're looking at the windows process manager and see WSL2 still taking about that amount, but thats a peculiarity of the hyper-v that underlies WSL2
<Telgareith>
ah, it's being used by fs buffer
<LetoThe2nd>
Telgareith: once bitbake is over, "yocto" consumes exactly no ram.
nemik has quit [Ping timeout: 240 seconds]
nemik has joined #yocto
nemik has quit [Ping timeout: 240 seconds]
nemik has joined #yocto
<denix>
LetoThe2nd: are you sure about that? one can never forget "yocto" and it stays in your memory forever!
<LetoThe2nd>
denix: my words were, (as always!) carefully chosen - thats why i used "ram", and not "memory". because my memory is random, but certainly not accessible in a robust way.
<LetoThe2nd>
;-)
* derRichard
expected scripts/contrib/convert-overrides.py to fix my layer and not to rape it ;-\
<derRichard>
the script even replaced underscores in shell scripts *grrrr*
<LetoThe2nd>
yeah i've had that effect too at one occasion
<RP>
derRichard: there is a reason we couldn't do it automatically :(
<derRichard>
and it is an implicit dos2unix, it messed up all kind of new lines
<RP>
I didn't realise anyone was using dos newlines in OE metadata so that was never even tested
<derRichard>
RP: you don't work a lot with customers that develop on windows? ;-)
<RP>
derRichard: there is a patch that looks a bit like that I just merged
goliath has joined #yocto
fitzsim has joined #yocto
<JaMa>
I run dos2unix from time to time, just to teach my co-workers to use linux or at least configure git on windows correctly :)
<JaMa>
but it's rarely an issue with bitbake metadata, the .patch files for our internal components are much worse, e.g. when I forget to use --keep-cr with git am
adrian__ has quit [Ping timeout: 252 seconds]
guysoft42 has quit [Ping timeout: 240 seconds]
<derRichard>
ah, convert-overrides.py "fixed" also an src uri here ;-\
<derRichard>
and having a colon in a file name is something bitbake/lib/bb/cache.py does not like
mvlad has quit [Remote host closed the connection]
seninha has quit [Remote host closed the connection]
florian has quit [Ping timeout: 260 seconds]
nemik has quit [Ping timeout: 240 seconds]
nemik has joined #yocto
<JaMa>
derRichard: yeah, I had quite a few cases like this as well, I undid them in separate commit so it's easy to re-run the script and then re-fix again
<derRichard>
JaMa: yep. just did so. :-)
<derRichard>
hm, commit ebfa1700f41b ("bitbake: fetch2/cooker: Fix source revision handling with floating upstreams") is also problematic. i have a recipe which uses AUTOREV, it works perfectly fine. but when i try to build it after running devtool modify, bitbake thinks i use floating upstreams
<derRichard>
bitbake foo # works
<derRichard>
devtool modify foo # hack hack
<derRichard>
bitbake foo # fails
<derRichard>
after running devtool reset -n foo i can build it again ;-\
nemik has quit [Ping timeout: 246 seconds]
nemik has joined #yocto
seninha has joined #yocto
bps has quit [Ping timeout: 252 seconds]
<kevinrowland>
Here's a weird one. We've got a `do_install()` task that runs a script that runs a `git` command to find the root of the repository, but the `git` command fails with "fatal: unsafe repository ('...' is owned by someone else)".. so we added one line in `do_install()` to just call `/usr/bin/whoami` and it says `root`! Why is `do_install()` running as
<kevinrowland>
root? Is it in a fakeroot? TBH I don't fully understand fakeroot/pseudo
<RP>
kevinrowland: yes, it is run under pseudo which is a fakeroot environment
<RP>
kevinrowland: which release? There are some fixes queued for dunfell/kirkstone
<kevinrowland>
Hardknott
nemik has quit [Ping timeout: 252 seconds]
<RP>
kevinrowland: you'll have to port the fixes from one of the other branches then
nemik has joined #yocto
<kevinrowland>
Cool, thanks for responding. Seems like an intermittent issue though, I haven't been able to repro and it only reared its head on another dev's machine and only after we moved some layers around
<RP>
kevinrowland: it depends whether git has a CVE fix backported to it or is a very recent version