Tokamak has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
florian_kc has joined #yocto
florian_kc has quit [Ping timeout: 240 seconds]
sakoman has quit [Quit: Leaving.]
Tokamak has joined #yocto
Tokamak_ has joined #yocto
Tokamak has quit [Read error: Connection reset by peer]
Vonter has joined #yocto
goliath has joined #yocto
xmn has quit [Quit: ZZZzzz…]
xmn has joined #yocto
Tokamak_ has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
Guest47 has joined #yocto
Guest47 has quit [Client Quit]
mariusz1 has joined #yocto
Wouter0100 has quit [Read error: Connection reset by peer]
Wouter0100 has joined #yocto
ziga__ has joined #yocto
xmn has quit [Ping timeout: 256 seconds]
GNUmoon has quit [Ping timeout: 276 seconds]
Vonter has quit [Ping timeout: 256 seconds]
Vonter has joined #yocto
dlan has quit [Ping timeout: 240 seconds]
Vonter has quit [Read error: Connection reset by peer]
Vonter has joined #yocto
dlan has joined #yocto
Vonter has quit [Ping timeout: 256 seconds]
dti has joined #yocto
dtometzki has quit [Ping timeout: 240 seconds]
rob_w has joined #yocto
GNUmoon has joined #yocto
frieder has joined #yocto
mckoan|away is now known as mckoan
<mckoan>
good morning
frieder has quit [Ping timeout: 240 seconds]
zpfvo has joined #yocto
zpfvo has quit [Ping timeout: 256 seconds]
zpfvo has joined #yocto
frieder has joined #yocto
zpfvo has quit [Ping timeout: 256 seconds]
zpfvo has joined #yocto
Guest24 has joined #yocto
<coldspark29[m]>
Morgain
zpfvo has quit [Ping timeout: 240 seconds]
zpfvo has joined #yocto
Schlumpf has joined #yocto
zpfvo has quit [Ping timeout: 256 seconds]
zpfvo has joined #yocto
Schlumpf has quit [Client Quit]
Schlumpf has joined #yocto
Vonter has joined #yocto
vladest has quit [Quit: vladest]
Vonter has quit [Ping timeout: 240 seconds]
florian has joined #yocto
vladest has joined #yocto
Vonter has joined #yocto
Guest24 has quit [Quit: Client closed]
zyga-mbp has joined #yocto
mvlad has joined #yocto
leon-anavi has joined #yocto
zyga-mbp has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
zyga-mbp has joined #yocto
zyga-mbp has quit [Client Quit]
Vonter has quit [Read error: Connection reset by peer]
Vonter has joined #yocto
Vonter has quit [Ping timeout: 256 seconds]
kroon has joined #yocto
<kroon>
I see noise in buildhistory when doing: bitbake -c cleansstate shadow-native && bitbake shadow-native && bitbake -c clean shadow-native && bitbake shadow-native
<kroon>
The permissions for /var/spool/mail/ jumps between 755 and 775
<kroon>
I can see that the tar archive in sstate has "correct" permissions (775) for the dir, but when unpacking from sstate it seems the perms are changed to 775
<kroon>
sorry: ".. are changed to 755"
<kroon>
The target package does not seem to have this issue, only the native one
jmk1 has joined #yocto
dev1990 has quit [Quit: Konversation terminated!]
tre has joined #yocto
zpfvo has quit [Ping timeout: 252 seconds]
frieder has quit [Remote host closed the connection]
zpfvo has joined #yocto
frieder has joined #yocto
rr12zer has quit [Remote host closed the connection]
zpfvo has quit [Ping timeout: 256 seconds]
zpfvo has joined #yocto
Guest71 has joined #yocto
gsalazar has joined #yocto
Guest71 is now known as bomba5
kroon has quit [Remote host closed the connection]
kroon has joined #yocto
zpfvo has quit [Ping timeout: 256 seconds]
zpfvo has joined #yocto
gsalazar_ has joined #yocto
zpfvo has quit [Ping timeout: 240 seconds]
zpfvo has joined #yocto
Harun has joined #yocto
gsalazar has quit [Ping timeout: 256 seconds]
rr12zer has joined #yocto
<Harun>
Hello:) , how can I add glibc-locale to an image?
<Harun>
if done with IMAGE_INSTALL_append it fails with an error that nothing provides glibc-locale.
<Harun>
Currently the image has locale but no charmaps.
<Harun>
I am using dunfell
florian_kc has joined #yocto
<RP>
Harun: the locales are packaged separately so you probably need to be more specific
<qschulz>
Harun: you have glibc-charmaps package which brings all locales AFAIU
lucaceresoli has joined #yocto
asdf78 has joined #yocto
jmlemetayer[m] has left #yocto [#yocto]
camus has quit [Quit: camus]
camus has joined #yocto
asdf78 has quit [Quit: Client closed]
kroon has quit [Quit: Leaving]
<Harun>
qschulz: Thank you, that was it :)
BCMM has joined #yocto
<kanavin>
RP: just so you don't miss it, exclude_packages is now [ ] in repro test :)
<RP>
kanavin: I did notice :D
<RP>
kanavin: feels like we cheated with ruby docs but nice to get there! :)
<RP>
kanavin: thanks for the work on those last pain points!
<kanavin>
RP: yes, it's not perfect, but it doesn't reduce test coverage or runtime support, so I can live with it :)
<kanavin>
and the upstream bug report should be enough for upstream to look into it, if they have capacity and interest
<RP>
kanavin: right, we did try on it
mariusz1 has quit [Ping timeout: 240 seconds]
Guest13 has joined #yocto
Guest13 has quit [Client Quit]
florian_kc has quit [Ping timeout: 256 seconds]
florian_kc has joined #yocto
zpfvo has quit [Ping timeout: 252 seconds]
zpfvo has joined #yocto
<paulg>
was trying to build on ubu 16.04 using buildtools + environment-setup-x86_64-pokysdk-linux ; pseudo-native failed final link with missing pow/exp etc until I added "-lm" to link.
<paulg>
Seemed odd. pseudo doing math ops? Not sure anyone cares enough it is worth investigating/patching?
zpfvo has quit [Ping timeout: 240 seconds]
zpfvo has joined #yocto
xmn has joined #yocto
zpfvo has quit [Ping timeout: 240 seconds]
zpfvo has joined #yocto
sakoman has joined #yocto
zpfvo has quit [Ping timeout: 256 seconds]
<RP>
paulg: which buildtools? libm was merged into libc now so this issue probably goes away soon
zpfvo has joined #yocto
<RP>
or perhaps it was caused by a buildtools which has the merge?
<paulg>
I didn't explicitly choose a buildtools, this is what it gave me yesterday.
<paulg>
(needed buildtools since ubu-16.04 doesn't have py3 v3.6+ fwiw)
<RP>
3.1 is dunfell I think so technically still supported and would be pre glibc merge
<RP>
20200923 does sound old though
<paulg>
this is a machine that hasn't been in use for a while ; all repos should be on master and were updated yest.
<RP>
oh. That is bad then :/
<paulg>
it was idle long enough that I had to do the append dance on local.conf and change conf version from 1 to two ; then blew away sstate tmp cache
<RP>
I was going to say, even dunfell uses a buildtools of autobuilder.yocto.io/pub/non-release/20210214-8/buildtools/x86_64-buildtools-extended-nativesdk-standalone-3.2+snapshot-7d38cc8e749aedb8435ee71847e04b353cca541d.sh
<paulg>
I'm not really familiar with buildtools or how it chose what to wget, but I think I can scroll up and see.. lemme check.
<smurray>
paulg: heh, never realized you're up north of Toronto, what are you stuck with for internet access up there?
<paulg>
smurray, same as when I'm over in Ottawa - cable. Obviously not as all-encompassing over in this area, esp. if you are off the beaten path.
<smurray>
paulg: yeah, I bet
* paulg
watches "Parsing recipes" contribute to global warming for the 1000th time...
<RP>
paulg: well, my build with 3.4 worked :/
* paulg
is up to sqlite3 building...
<paulg>
old hardware keeps you honest. You care about where your CPU cycles are going.
<paulg>
Once I manage to kill rust with fire and plague, maybe I can make use of 8G machines again.
<RP>
paulg: you can blame vmeson for rust ;-)
<paulg>
oh believe me, I do.
<paulg>
Regularly.
<paulg>
Nope. Still blows up. I nuked everything but conf and downloads.
<paulg>
Must be triggered by my conf and layer selection somehow?
<paulg>
or installed pkgs on host?
<paulg>
Dunno.
<paulg>
Will look at it again after lunch - need to go do some other overdue crap now.
<RP>
paulg: there is something we're both missing here :/
<RP>
paulg: are you using poky or nodistro? What is odd is your link line is different to mine in pseudo-native
<RP>
mine has --alow-shlib-undefined on it
<RP>
which is a uninative addition
<RP>
poky enables uninative by default, nodistro doesn't
zpfvo has quit [Ping timeout: 256 seconds]
<paulg>
I've got the overc layer - which is basically a real-metal version of what was buildappliance, and I think it defines its own distro vars....
zpfvo has joined #yocto
<RP>
paulg: I think uninative is either not enabled at all or being disabled
<paulg>
let me grep around
<paulg>
unless zeddii can remember...
tre has quit [Remote host closed the connection]
mariusz1 has joined #yocto
<zeddii>
paulg, remember what ? They day of the week ? Maybe.
<zeddii>
but nothing special should be in the overc definitions around uninative
<paulg>
RP, my config has roots going back years. So if we added something to the generic boiler plate template for local.conf around uninative, I might not have it?
<paulg>
like zeddii - I refuse to abandon my config and build directory on a daily basis as the sstate cult seems to do.
zpfvo has quit [Ping timeout: 256 seconds]
zpfvo has joined #yocto
<zeddii>
my current build dir is 6 years old!
Schlumpf has quit [Quit: Client closed]
zpfvo has quit [Ping timeout: 256 seconds]
zpfvo has joined #yocto
dev1990 has joined #yocto
<rburton>
RP: "I'm sending it out after this call"
<JPEW>
RP: Didn't get to that hashequiv issue. I have another meeting, then lunch, but I should be free in about 2 hours if you still want to go over it today
<paulg>
RP, so I created a new dir (build-sato), used the default conf dir that oe-init-build-env makes and pseudo-native built ok. (same shell, same buildtools env)
<paulg>
it would seem I found a way to break buildtools with my conf/layers.
<paulg>
I'll diff the conf files and look for suspects
mariusz1 has quit [Ping timeout: 256 seconds]
florian_kc has joined #yocto
otavio has quit [Remote host closed the connection]
<bluelightning>
anyone happen to know what would influence whether or not ${B}/tools/* gets installed to ${STAGING_KERNEL_BUILDDIR}/tools ?
Habbie has joined #yocto
xmn has quit [Ping timeout: 256 seconds]
<mischief>
why are both dunfell and kirkstone marked apr 2024 LTS when kirkstone is released 2 years later?
<JPEW>
mischief: dunfell was extended, kirkstone hasn't been
<JPEW>
mischief: dunfell was originally also 2 years FWIW
<LetoThe2nd>
yoyo dudX
<JPEW>
LetoThe2nd: Little late (or early) for you?
<LetoThe2nd>
JPEW: ~10pm. ok-ish
<mischief>
JPEW: thanks.
<JPEW>
LetoThe2nd: Ugh, 24 hour clock threw me off :)
<zeddii>
bluelightning: do_shared_workdir() (but I'm sure you found that already). but we are currently pretty selective about what gets copied over
<bluelightning>
zeddii: the weird thing is I have the results of one build where stuff has been installed, and one were it hasn't, and I can't tell what would be different between the two at the moment
<bluelightning>
still digging into it
<zeddii>
ah. that's different. task racing ? or is it objtool ? That's conditional on a kernel config
<bluelightning>
in one build just the objtool binary is there, in the other (older build) I have bpf/ and objtool/, and objtool/ contains sources as well
<zeddii>
same machine both builds ? That is odd.
<bluelightning>
yeah... could be that someone has hacked in something somewhere in here to copy those files over
florian_kc has quit [Ping timeout: 240 seconds]
<RP>
paulg: try adding INHERIT += "uninative" into local.conf. We do this in poky but overc won't pull that in
<RP>
paulg: that should work around the issue but what the real fix is, not sure
<RP>
JPEW: sorry, I was afk, long day. I'm back around now. Did you have any ideas on a fix?
<JPEW>
RP: From a high level, all I could think of was that the unihash for all the dependencies (RDEPENDS?) needs to be in the output.... which doesn't seem fun
kanavin has quit [Remote host closed the connection]
kanavin has joined #yocto
<paulg>
yeah - I'd fired off another new dir, and copied in my old overc conf dir and still got the fail. I'll look at it some more in an effort to understand wtf and what crimes uninative does, now that you've made me curious.
<JPEW>
RP: That would be the only "automated" method that would get all the cases.... otherwise it's just manually annotating cases we find
<JPEW>
At least AFAICT
<RP>
JPEW: that is basically what I was thinking. Either the unihash of dependencies or of things with shlib dependencies
<RP>
we could inject those into the outhash
<RP>
JPEW: I don't like it
<JPEW>
even shlibs is insufficient; imagine a bash script that execs zstd
<RP>
JPEW: right, I know
<paulg>
(or I could move to 18.04 and do a "didn't see it ; didn't happen"...)
<RP>
paulg: I look forward to seeing your reaction when you realise what it does :)
<paulg>
:-P
leon-anavi has quit [Quit: Leaving]
<RP>
paulg: it basically lets us use binaries built on one system on any other system regardless of distro
<JPEW>
RP: Perhaps.... do_populate_sysroot_setsecene() can add a file with the deps an their unihash to the sstate archive?
<sgw>
RP: send you and zeddii and another reply, I think am in a maze of twisty little passages, all alike.
<JPEW>
That's the really... ugly... way to force the issue
<RP>
JPEW: that was kind of what I was thinking. I wish I'd come up with somethig better
<JPEW>
sgw: The way out of any hedge maze is to always turn left (or right if you prefer).... takes a while though :)
<sgw>
xyzzy works well also, but not sure it will with the kernel.bbclass and package.bbclass
mariusz1 has quit [Ping timeout: 256 seconds]
<RP>
sgw: I think you'll have to tweak do_strip to create a copy and then deploy to use that copy, or move the strip and size check into a larger do_deploy
<paulg>
RP - ok but it does, as you hint, leave one wondering how that bleeds into pseudo-native link flags... presumably some "configure" test...
<RP>
paulg: no, that much is clear to me. the stuff uninative does changes LDFLAGS (see uninative.bbclass) so the options passed in will mean it sees a modern libc with libm merged
<RP>
paulg: my main question is how we'd fix the non-uninative case
<paulg>
Oh, oh, I know! We add "-lm".
* paulg
runs
<paulg>
aside from kooks like me, how big is the expected user base for buildtools w/o native unicorns?
<sgw>
RP: yeah, I was thinking about the copy part, I was not sure of changing the ordering of strip/size check wrt install/deply, since install happens first. Thx
<RP>
paulg: it isn't needed with any recent glibc now so it is a backwards step
<paulg>
so was systemd, but that happened anyway. :-P
<RP>
paulg: that space is rather small else we'd have a bug report before now
* RP
is ignoring conversations that bring systemd in
<paulg>
kinda what I figured 1st thing this AM - probably nobody is hitting it and largely nobody cares.
<paulg>
[01/11/22:08:53] <paulg> Seemed odd. pseudo doing math ops? Not sure anyone cares enough it is worth investigating/patching?
<moto-timo>
let the shouting and gnashing of teeth begin
<moto-timo>
I'll put on my shield belt, hopefully no one brings a lasgun to the party
<tlhonmey>
I always wondered why they didn't just do remote-controlled lasguns that could do the shooting from a safe distance and let the potential explosion be irrelevant...