Xagen_ has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<khem>
ac_slater: are you trying it with master branch ?
<khem>
if not try it there first, since it will have latest setuptools
kpo_ has quit [Ping timeout: 245 seconds]
<ac_slater>
khem: thanks! I have a vendor version of yocto so it's a bit tough. I can do the normal bisecting. I was just curious if there was a known issue with python packages - this is my first time adding one
goliath has quit [Quit: SIGSEGV]
alejandrohs has quit [Ping timeout: 260 seconds]
Daanct12 has joined #yocto
alejandrohs has joined #yocto
<khem>
Yeah usually you try with master and if it fails there then you have an upstream problem we can also chime in here
davidinux has quit [Ping timeout: 252 seconds]
davidinux has joined #yocto
alimon has quit [Ping timeout: 240 seconds]
xmn has quit [Ping timeout: 264 seconds]
alimon has joined #yocto
jclsn has quit [Ping timeout: 268 seconds]
kpo_ has joined #yocto
ablu has quit [Read error: Connection reset by peer]
jclsn has joined #yocto
ablu has joined #yocto
vvmeson has quit [Ping timeout: 260 seconds]
raghavgururajan has quit [Ping timeout: 276 seconds]
jonesv has quit [Read error: Connection reset by peer]
tleb has quit [Ping timeout: 256 seconds]
jmd has joined #yocto
<ac_slater>
thanks khem
sstiller has joined #yocto
alessioigor has joined #yocto
alessioigor has quit [Remote host closed the connection]
<Average_Joe>
I've build an image for RISC-V. The crosscompiler assumed that the processor has the A-extension available, but unfortunately my implementation hasn't got (full) support (yet).
<Average_Joe>
How can I change the machine for which yocto builds ? When I'm doing baremetal, I use the arguments: -march=rv32gc
<Average_Joe>
All help is welcome!!
<Average_Joe>
Thanks!
Guest18 has quit [Ping timeout: 250 seconds]
alessioigor has quit [Quit: alessioigor]
alessioigor has joined #yocto
lexano has quit [Ping timeout: 246 seconds]
vladest1 has joined #yocto
vladest has quit [Ping timeout: 260 seconds]
vladest1 has quit [Ping timeout: 268 seconds]
alperak has joined #yocto
sakman has quit [Ping timeout: 268 seconds]
Saur has quit [Remote host closed the connection]
lexano has joined #yocto
Saur has joined #yocto
<LetoThe2nd>
yo dudX
<LetoThe2nd>
Average_Joe: create a new machine conf based on the closest that is available, and then change the tune as you need it.
sakman has joined #yocto
g0hl1n has joined #yocto
prabhakarlad has joined #yocto
<Average_Joe>
Oh nice ... feedback from the juggler him-self ... Thanks!!
<Average_Joe>
(I was just watching you on youtube :))
<Average_Joe>
Kudos on those videos ... really helpful!
<LetoThe2nd>
Average_Joe: thanks, glad to hear
<LetoThe2nd>
Average_Joe: for getting started, maybe look through the machines listed on the layer index, there almost certainly is some modified-riscv starter template.
<RP>
/home/pokybuild/yocto-worker/qemux86-world-alt/build/build/tmp/work/core2-64-poky-linux/dnf/4.18.2/recipe-sysroot/usr/bin/postinst-docbook-xsl-stylesheets-xmlcatalog: line 5: xmlcatalog: command not found
rob_w has quit [Remote host closed the connection]
Chaser has joined #yocto
rob_w has joined #yocto
Saur has quit [Remote host closed the connection]
Saur has joined #yocto
Guest55 has joined #yocto
Saur has quit [Remote host closed the connection]
<Guest55>
How do I enable address sanitizers(-fsanitize=address) in yocto
Saur has joined #yocto
Guest95 has joined #yocto
<alessioigor>
Guest55: Install the libraries (libasan libubsan and so on) into your image
* alessioigor
waves all
Saur has quit [Remote host closed the connection]
Saur has joined #yocto
Saur has quit [Remote host closed the connection]
Saur has joined #yocto
<Guest55>
I tried adding IMAGE_INSTALL:append = libasan libubsan but in the target when I run a program it gives me this error
<Guest55>
cannot find libasan_preinit.o: No such file or directory
<Guest55>
cannot find -lasan: No such file or directory
<alessioigor>
Are you verified that libraries are present?
<rburton>
RP: no thats totally from my changes, i'll dig
<Guest55>
No, I haven't verified it. But when we append for those libraries in local.conf it has to be present right?
<Guest55>
I could see that those libraries are installed
<KanjiMonster>
how do I tell bitbake to create packages for every transitive dependency? I have the issue that if I do bitbake foo it will create packages for foo and foo's dependencies, but not for the dependencies of foo's dependencies
xmn has joined #yocto
<RP>
KanjiMonster: see the --runall option to bitbake?
<Guest55>
Even if those libraries(libasan and libubsan) are present why do I see these errors: cannot find libasan_preinit.o: No such file or directory
<Guest55>
cannot find -lasan: No such file or directory
<KanjiMonster>
RP: I'm already using --runall=do_package_write_ipk, which helps for the immediate dependencies, but not the transitive ones. I also noticed these don't show up in pn-buildlist
<RP>
KanjiMonster: try do_build ?
<KanjiMonster>
let me test ...
<KanjiMonster>
RP: nope, doesn't help :/
<alessioigor>
Guest55: Perhaps you have to install -dev packages...
<rburton>
yeah you'll want the -dev forms if you want to link to them on the target
<Guest55>
Let me try that out..
<RP>
KanjiMonster: what are you defining as a transitive dependency? That option should cover it but we may have different definitions I guess
amitk has joined #yocto
<KanjiMonster>
RP: e.g. openvswitch depends on python3-twisted which depends on python3-appdirs, which depends on python3-click. And python3-click does not get packaged, making the generated openvswitch package uninstallable due to missing dependencies
<KanjiMonster>
more or less I look for the right command to "build this package and everything that's needed to install it so I can put the result in a feed"
<KanjiMonster>
this is kirkstone FWIW
<RP>
KanjiMonster: can you quickly try master? You shouldn't need a full build, just a bitbake -g and look at task-depends.dot
<RP>
KanjiMonster: it is possible there is a bug in there. It sounds like what you want should work but I can't be sure without looking into it in detail
<KanjiMonster>
RP: will do
JerryM has joined #yocto
<rburton>
KanjiMonster: fwiw, python3-appdirs doesn't depend on click
<rburton>
ah the problem might be that these are only in RDEPENDS and not in DEPENDS
<rburton>
rdepends being basically packaging metadata and not actually the build time recipe dependency graph
alessioigor has quit [Quit: alessioigor]
alessioigor has joined #yocto
<RP>
rburton: "--runall RUNALL Run the specified task for any recipe in the taskgraph of the specified target (even if it wouldn't otherwise have run)."
<JaMa>
JPEW: I was sharing your "Hash Equivalence at Scale" slides with someone and was surprised that there is no mention of BB_HASHSERVE/BB_HASHSERVE_UPSTREAM if you ever refresh these it might be useful to show how to configure the bitbake build to use it
<RP>
rburton: it should technically follow into that
g0hl1n has quit [Quit: Client closed]
<KanjiMonster>
rburton: RP: It feels there is a depth limit or so. E.g. if I build a packagegroup rdepending on openvswitch, the packaging stops at python3-twisted(-core). If I build openvswitch, it goes one level more and builds the twisted-core RDEPENDS, but not their's
<rburton>
RP: why does ab-helper's kirkstone branch have "master" all over the repo-defaults block? i'd have expected it to say 'kirkstone' everywhere...
rob_w has quit [Remote host closed the connection]
<rburton>
presumably something else is switching the defaults
<RP>
KanjiMonster: right, there were bugs like that which I have fixed. I just can't remember when, hence asking if master has the issue
<RP>
rburton: I don't think ab-helper's branch config does much
<RP>
rburton: the intent was there originally but I think it all comes in too late
g0hl1n has joined #yocto
<RP>
rburton: yocto-ab2's code is usually what wins
<KanjiMonster>
meh, pastebin complains that it's too large.
g0hl1n has quit [Client Quit]
g0hl1n has joined #yocto
<KanjiMonster>
(spoiler: there is python3-click-native, but no python3-click in there)
<RP>
KanjiMonster: what is interesting is that python3-click isn't here at all
<RP>
KanjiMonster: did you disable debian renaming?
<KanjiMonster>
all I did was checkout poky, added meta-openembedded and meta-virtualization, did . oe-init-buildenv and then ran bitbake -g openvswitch. no code changes or even local.conf changes
jmiehe has quit [Quit: jmiehe]
<rburton>
so it's not pulling the names from RDEPENDS and turning them into explicit DEPENDS
g0hl1n has quit [Quit: Client closed]
<JaMa>
RP: I have maybe interesting side-effect of "Fix EXPORT_FUNCTIONS bug" change in bitbake, do you have 2 mins to see if it's expected?
<JaMa>
RP: .bb file defines do_install(), then there is .bbappend which adds cmake inherit, before this change the do_install in recipe is ignored and only cmake_do_install is called in do_install, after this change the do_install from .bb is used and cmake_do_install isn't called at all
<RP>
KanjiMonster: ok. I'll have to try and look into it, I don't know what to advise
<RP>
JaMa: I assume cmake uses EXPORT_FUNCTIONS?
* RP
checks and it does
<RP>
JaMa: that is expected, the previous behaviour is a bug
<RP>
JaMa: I did worry we may find some of these
Guest95 has quit [Quit: Client closed]
Guest55 has quit [Ping timeout: 250 seconds]
prabhakarlad has quit [Quit: Client closed]
vladest1 has joined #yocto
vladest has quit [Read error: Connection reset by peer]
vladest1 is now known as vladest
<KanjiMonster>
RP: thank you. Feel free to ping me if you want me to test anything. For now I'll go down the rabbit hole and call bitbake on dependencies until I get no errors from opkg anymore
<JaMa>
RP: ok will check what changed in the image after this change, sofar I know about 2 cases which resulted in build failures, but there might be more where it silently changed the content
<rburton>
KanjiMonster: a workaround would be to bbappend some DEPENDS in, like add click DEPENDS to incremental
<RP>
JaMa: EXPORT_FUNCTIONS was always defined as "do not override any existing recipe function" so the bug was clear, if annoying. It doesn't make sense to behave another way
<JaMa>
RP: OK, so the right fix is to define do_install calling cmake_do_install after the inherit? that's what I did now
<Xogium>
is there a way to make yocto use poky to build a minimal ramdisk image ? As in, a ramdisk that gets bundled with the kernel, but with it being the whole rootfs ? I.e: no pivoting to a full rootfs later
<JaMa>
in other case the do_install and inherit is in the same .bb, so I'll just reorder them
<rburton>
Xogium: there's INITRAMFS_BUNDLE or similar. you control the file system so just don't have a pivot in it.
goliath has joined #yocto
<Xogium>
hmm, good point
<RP>
JaMa: I think so
<rburton>
Xogium: the only important thing about an initramfs image is that it needs to be a cpio. if you want to have it so its a self-contained system and doesn't pivot, then make it do that.
<Xogium>
righto
<rburton>
Xogium: though a true bundled initramfs is a pain as you need to rebuild the kernel every time you change it. if your bootloader supports separate kernel and initrd then that is nicer
<Xogium>
yeah that's also true
<Xogium>
here's just for a quick run in qemu ;) trying to develop a hardware tts and I just need something lightning fast for this purpose
<Xogium>
I could make my own cpio manually, but I figure, might as well learn yocto while doing this :D
vladest has quit [Quit: vladest]
vladest has joined #yocto
alperak has quit [Quit: Client closed]
alperak has joined #yocto
zpfvo has quit [Ping timeout: 264 seconds]
zpfvo has joined #yocto
Guest41367 has quit [Remote host closed the connection]
Daanct12 has quit [Quit: WeeChat 4.1.2]
jmiehe has joined #yocto
jmiehe has quit [Client Quit]
sotaoverride is now known as Guest2709
Guest2709 has quit [Killed (iridium.libera.chat (Nickname regained by services))]
sotaover1ide is now known as sotaoverride
zpfvo has quit [Ping timeout: 256 seconds]
zpfvo has joined #yocto
luc4 has joined #yocto
ctraven has joined #yocto
luc4 has quit [Client Quit]
jonesv has joined #yocto
<KanjiMonster>
RP: I wrote a shell script to check the Packages lists for any missing packages from Depends, and added those to the RDEPENDS of our packagegroup (likely a concidence, but they were all python3 packages)
tleb has joined #yocto
raghavgururajan has joined #yocto
prabhakarlad has joined #yocto
<LetoThe2nd>
ü
sakoman has quit [Remote host closed the connection]
sakoman has joined #yocto
<KanjiMonster>
LetoThe2nd: ä
<LetoThe2nd>
KanjiMonster: yep
tleb has quit [Ping timeout: 260 seconds]
jonesv has quit [Ping timeout: 260 seconds]
raghavgururajan has quit [Ping timeout: 260 seconds]
jmiehe has joined #yocto
Average_Joe has quit [Quit: Leaving]
davidinux has quit [Quit: WeeChat 3.5]
davidinux has joined #yocto
raghavgururajan has joined #yocto
jonesv has joined #yocto
zhmylove has joined #yocto
vmeson has joined #yocto
raghavgururajan has quit [Ping timeout: 264 seconds]
<KanjiMonster>
RP: that was quick! will give it a test once I'm out of the current meeting
<JPEW>
JaMa: ah ya. That didn't change so I didn't think to put it in my presentation
<RP>
KanjiMonster: I've tweaked it a bit and sent to the list for review
Vonter has quit [Ping timeout: 276 seconds]
Vonter has joined #yocto
florian_kc has joined #yocto
vladest has quit [Ping timeout: 240 seconds]
jmiehe has quit [Quit: jmiehe]
simonew has joined #yocto
jmiehe has joined #yocto
<rburton>
RP: ah for bonus fun the xsl thing is only breaking in dnf because it depends on libdnf which has to depend on a _target_ gtk-doc because cmake is terrible and that then pulls in the target xsl and that is the one that breaks. or something.
vladest has joined #yocto
<rburton>
RP: ok xmlcatalog fix sent which sorts out dnf for me
Chaser has quit [Quit: My Mac has gone to sleep. ZZZzzz…]
zhmylove has quit [Quit: Leaving]
vladest has quit [Remote host closed the connection]
Guest95 has joined #yocto
vladest has joined #yocto
<KanjiMonster>
RP: partial success, at least with kirkstone. building openvswitch triggers python3-click, but building a packagegroup that RDEPENDS on openvswitch does not. Now trying master (will take a bit)
jmiehe has quit [Quit: jmiehe]
JerryM has quit [Quit: Konversation terminated!]
Estrella has quit [Quit: No Ping reply in 180 seconds.]
<RP>
khem: did you drop the :class-native overrides that were no longer needed?
<khem>
but thats for native bb which is separate its target one
<RP>
khem: ah, right
<RP>
khem: I wonder if the do_configure:append is now not doing what it did before
sakoman has quit [Quit: Leaving.]
<RP>
khem: instead of an :append it may want to explictly call autotools_do_configure ?
<khem>
let me try
florian_kc has joined #yocto
<moto-timo>
hmmm... when we dropped distrodata.bbclass (warrior?) we lost the do_distro_check task. i wonder where that would be appropriate to go now?
<moto-timo>
I suspect we are no longer testing that Debian/Fedora/Ubuntu comparison code from lib/oe/distro_check.py so some oe-selftest is also in order ;)
alessioigor has quit [Quit: alessioigor]
florian_kc has quit [Ping timeout: 264 seconds]
alessioigor has joined #yocto
GNUmoon2 has quit [Ping timeout: 240 seconds]
GNUmoon2 has joined #yocto
<tlwoerner>
strange behaviour from bmaptool: half the time it see the queue/scheduler as [bfq] and the other half the time it sees [none]
<tlwoerner>
and that's with leaving the usb stick plugged in continuously
<tlwoerner>
distilling the code that performs this read, i can run that 30 times in a row and it always reads [none]
<tlwoerner>
([none] is the correct value, i can 'cat' that file all day long and it always reads [none])
alperak has quit [Quit: Client closed]
Chaser has quit [Quit: My Mac has gone to sleep. ZZZzzz…]
<rburton>
moto-timo: honestly, i'd be tempted to rip it out. don't care what other distros ship, we should be shipping latest (or have a good reason not to)
<moto-timo>
rburton: we have historically had this in RRS and in the branch comparison for layerindex
<moto-timo>
rburton: it is an aid in getting people to leave other distros and switch to Yocto Project
<moto-timo>
(or vice versa, but I stick to my own definition)
<rburton>
yeah i remember
<rburton>
still: meh
frieder has quit [Ping timeout: 260 seconds]
florian_kc has joined #yocto
amitk has quit [Ping timeout: 240 seconds]
<RP>
moto-timo: its original purpose was to compare licenses cross distro but that was a long time ago!
<RP>
moto-timo: some of the Intel specific reasons for it no longer exist afaik
<moto-timo>
I think it is still a valid use case to for instance be looking at a BeagleBoard Debian release or NVidia Ubuntu release when moving to Yocto Project.
<moto-timo>
We have had customers in similar situations in the last two years.
<moto-timo>
anyway, it seems to be close to functional as a standalone distro-check.bbclass
mvlad has quit [Remote host closed the connection]
LDericher has quit [Ping timeout: 268 seconds]
LDericher has joined #yocto
prabhakarlad has joined #yocto
vladest has quit [Quit: vladest]
jmd has quit [Remote host closed the connection]
Kubu_work has joined #yocto
Kubu_work has quit [Client Quit]
alessioigor has quit [Quit: alessioigor]
vladest has joined #yocto
Xagen has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
xmn has quit [Ping timeout: 276 seconds]
xmn_ has joined #yocto
prabhakarlad has quit [Quit: Client closed]
xmn_ has quit [Quit: xmn_]
xmn has joined #yocto
astlep550401 has quit [Ping timeout: 256 seconds]
astlep550401 has joined #yocto
florian_kc has quit [Ping timeout: 260 seconds]
<khem>
RP: is there a change of behaviour in autotools with inherit_defer ?
<khem>
wait a moment, this does not have any automake stuff in it, it needs cmake
<khem>
how did it work
<khem>
it does inherit cmake so I guess now cmake bbclass should be inherit_defer'ed too
<khem>
this is going to be quite confusing when to use inherit and when to use inherit_defer, the side effects needs to be well documented/explained