Thursday, 2018-09-06

*** chandana73 <chandana73!~ckalluri@> has quit IRC00:01
*** chandana73 <chandana73!~ckalluri@> has joined #yocto00:09
*** jkprg <jkprg!~jkprg@> has quit IRC00:36
*** JaMa <JaMa!~martin@> has joined #yocto00:40
*** nighty- <nighty-!> has joined #yocto00:48
*** jkprg <jkprg!~jkprg@> has joined #yocto00:50
*** linuxjacques <linuxjacques!~jacques@nslu2-linux/jacques> has quit IRC01:10
*** ntl <ntl!> has quit IRC01:13
*** dc13ff <dc13ff!uid190567@gateway/web/> has joined #yocto01:17
*** Net147 <Net147!~Net147@unaffiliated/net147> has quit IRC01:17
*** Net147 <Net147!~Net147@unaffiliated/net147> has joined #yocto01:18
*** rburton <rburton!~textual@> has quit IRC01:23
*** rburton <rburton!> has joined #yocto01:23
*** Son_Goku <Son_Goku!~King_InuY@fedora/ngompa> has joined #yocto01:50
*** jkprg <jkprg!~jkprg@> has quit IRC01:51
*** rburton <rburton!> has quit IRC01:59
*** jkprg <jkprg!~jkprg@> has joined #yocto02:01
*** Son_Goku <Son_Goku!~King_InuY@fedora/ngompa> has quit IRC02:04
*** stephano <stephano!~stephano@> has quit IRC02:10
*** Son_Goku <Son_Goku!~King_InuY@fedora/ngompa> has joined #yocto02:24
*** jynik <jynik!> has quit IRC03:00
*** rburton <rburton!> has joined #yocto03:03
*** Hooloovo0 <Hooloovo0!> has quit IRC03:05
*** Son_Goku <Son_Goku!~King_InuY@fedora/ngompa> has quit IRC03:08
*** Son_Goku <Son_Goku!~King_InuY@fedora/ngompa> has joined #yocto03:09
*** Hoolootwo <Hoolootwo!> has joined #yocto03:10
*** jkprg <jkprg!~jkprg@> has quit IRC03:15
*** arielmr <arielmr!> has quit IRC03:16
*** jkprg <jkprg!~jkprg@> has joined #yocto03:16
*** jkprg <jkprg!~jkprg@> has quit IRC03:25
*** tgraydon <tgraydon!~textual@> has quit IRC04:15
*** dc13ff <dc13ff!uid190567@gateway/web/> has quit IRC04:16
*** Hoolootwo is now known as Hooloovo004:38
*** Son_Goku <Son_Goku!~King_InuY@fedora/ngompa> has quit IRC04:54
*** armpit <armpit!~armpit@2601:202:4180:c33:45cc:4239:c235:fa42> has quit IRC04:58
*** chandana73 <chandana73!~ckalluri@> has quit IRC05:09
*** armpit <armpit!~armpit@2601:202:4180:c33:a9a3:ce31:52a9:bd5c> has joined #yocto05:10
*** pohly <pohly!> has joined #yocto05:51
*** hundeboll <hundeboll!> has joined #yocto05:52
*** agust <agust!> has joined #yocto06:12
*** rob_w <rob_w!~bob@unaffiliated/rob-w/x-1112029> has joined #yocto06:21
*** frsc <frsc!> has joined #yocto06:24
*** lusus <lusus!~lusus@> has joined #yocto06:34
*** tasslehoff <tasslehoff!~Tasslehof@> has joined #yocto06:50
*** fl0v0 <fl0v0!> has joined #yocto06:53
*** Kitsok <Kitsok!~kitsok@2a02:6b8:0:506:d681:d7ff:fe6d:d499> has joined #yocto06:59
KitsokHello all!06:59
KitsokNeed help plz. I'm upgrading from the old yocto (probably 2.2) and the dependencies are broken. When I try to bake openresty, I've got error that the library it depends on (luajit) is not found. As I understand this is due the switch to package specific sysroots. I've read explanation in the reference guide but still can't understand how to point to luajits' sysrot in openresty recipe?07:03
*** gtristan <gtristan!~tristanva@> has quit IRC07:08
*** miwa <miwa!~miwa@unaffiliated/miwa> has quit IRC07:09
*** varjag <varjag!> has joined #yocto07:13
*** falk0n_ <falk0n_!> has quit IRC07:14
*** georgem_home <georgem_home!uid210681@gateway/web/> has quit IRC07:22
nayfeKitsok: when you put DEPENDS =  "luajit", it should every needed files in recipe-sysroot07:31
nayfeof your recipe07:32
*** mckoan|away is now known as mckoan07:32
*** grumble <grumble!~grumble@freenode/staff/grumble> has quit IRC07:32
Kitsoknayfe: is it copied to openresty sysroot from luajit sysroot?07:33
*** grumble <grumble!~grumble@freenode/staff/grumble> has joined #yocto07:33
Kitsoknayfe: hmhm. Thank you, let me check07:34
nayfekitsok you can pastebin your recipe is you want07:34
KitsokOops, yes, it's there but wrong version. openresty needs 2.1, but I've got 2.0. The OE has 2.0.5 version, and I have 2.1.0-beta3 recipe in my layer, somehow it was not built by bitbake07:35
*** grahamgelding <grahamgelding!ca86f31a@gateway/web/freenode/ip.> has quit IRC07:36
KitsokIt's strange. bitbake sees my luajit recipe but in unknown layer:  ?                    2.1.0-beta307:38
KitsokSo somehow I didn't describe my layers07:38
nayfewhat is your current yocto version?07:41
KitsokIt's derived project, how to check? I think it's 2.507:41
KitsokOK, I've got openresty built, just copied my recipe to OE layer and it got compiled. The question is why bitbake ignored it in my layer?07:42
*** rburton <rburton!> has quit IRC07:54
*** rburton <rburton!> has joined #yocto07:56
KitsokWell, looks like my layers is misconfigured, it's not shown in bitbake-layers show-layers, but bitbake sees recipes placed in my layer. How can it be?07:58
*** boucman_work <boucman_work!~jrosen@wesnoth/developer/boucman> has quit IRC07:59
*** T_UNIX <T_UNIX!uid218288@gateway/web/> has joined #yocto08:08
*** boucman_work <boucman_work!~jrosen@wesnoth/developer/boucman> has joined #yocto08:15
*** armpit <armpit!~armpit@2601:202:4180:c33:a9a3:ce31:52a9:bd5c> has quit IRC08:23
*** BCMM <BCMM!~BCMM@unaffiliated/bcmm> has joined #yocto08:25
*** Son_Goku <Son_Goku!~King_InuY@fedora/ngompa> has joined #yocto08:29
*** dv_ <dv_!> has quit IRC08:37
*** florian_kc <florian_kc!~florian_k@Maemo/community/contributor/florian> has joined #yocto08:39
*** gtristan <gtristan!~tristanva@> has joined #yocto08:46
*** dv_ <dv_!> has joined #yocto08:50
*** florian_kc is now known as florian08:56
*** gtristan <gtristan!~tristanva@> has quit IRC09:01
*** gtristan_ <gtristan_!~tristanva@> has joined #yocto09:02
*** nighty- <nighty-!> has quit IRC09:35
*** mystictot <mystictot!~shyam@> has joined #yocto09:39
*** awe00 <awe00!~awe00@unaffiliated/awe00> has joined #yocto09:40
*** bluelightning <bluelightning!~paul@pdpc/supporter/professional/bluelightning> has quit IRC09:45
*** bluelightning_ <bluelightning_!~paul@pdpc/supporter/professional/bluelightning> has joined #yocto09:45
*** georgem <georgem!~georgem@> has quit IRC09:54
*** georgem <georgem!~georgem@> has joined #yocto09:54
*** mystictot <mystictot!~shyam@> has quit IRC09:55
*** falk0n <falk0n!> has joined #yocto10:00
*** mystictot <mystictot!~shyam@> has joined #yocto10:05
ak77 which are tested (host) distributions for yocto?10:12
kanavinak77: meta-poky/conf/distro/poky.conf has the list10:20
kanavinak77: note that it differs from one yocto release to another10:20
ak77kanavin: thank you10:22
ak77colegue has strange error on his machine (ubuntu 16.04), total rebuild (no downloads) breaks on configure for m4-native:configure: error: cannot run C compiled programs.10:26
mckoanak77: which Yocto version?10:29
mckoanak77: did you properly setup the build machine?10:29
ak77mckoan: using openembedded-core and meta-openembedded master (don't have the revision at hand, but we all use same), this did work before, it just broke out of the sudden. yes all host dependencies were installed,10:31
ak77mckoan: oe-build.. sourced.10:32
ak77mckoan: we'll check gcc versions, for native packages host's compiler is used i guess10:32
rburtonread config.log in the m4-naive build directory to see why it can't run gcc10:32
rburtonit will say clearly what the error is10:34
*** Kitsok <Kitsok!~kitsok@2a02:6b8:0:506:d681:d7ff:fe6d:d499> has quit IRC10:37
ak77rburton: of course it said. thank you10:47
rburtonkanavin: is your patchbomb in a branch?10:53
*** nighty- <nighty-!> has joined #yocto10:56
yoctiNew news from stackoverflow: error while loading shared libraries: cannot open shared object file: No such file or directory <>10:57
*** rob_w <rob_w!~bob@unaffiliated/rob-w/x-1112029> has quit IRC11:01
*** Son_Goku <Son_Goku!~King_InuY@fedora/ngompa> has quit IRC11:11
kanavinrburton: yes, the classic
*** Guest82133 <Guest82133!> has quit IRC11:15
*** Son_Goku <Son_Goku!~King_InuY@fedora/ngompa> has joined #yocto11:17
ak77 anyone uses combo-layers? we are trying google's repo tool - not sure it's the right tool for the job11:27
ak77openembedded-core's sanity.conf on master demands bitbake 1.39.1 which is not present in bitbake repo11:36
rburtonRP: ^ bitbake needs a push?11:39
rburtonak77: the version exists, just not the tag11:39
rburtonRP:  g-ir-scanner-lddwrapper: prelink-rtld: not found11:40
*** Kitsok <Kitsok!~kitsok@2a02:6b8:0:506:d681:d7ff:fe6d:d499> has joined #yocto11:42
*** tasslehoff <tasslehoff!~Tasslehof@> has quit IRC11:44
kanavinRP: is that for me-ß11:44
kanavinrburton: ^^^11:45
KitsokHow to debug why bitbake ignores recipe in my layer with the newer version of lldpd?11:49
*** bluelightning_ <bluelightning_!~paul@pdpc/supporter/professional/bluelightning> has quit IRC11:52
*** ant_work <ant_work!> has joined #yocto11:59
*** davenporten <davenporten!8b55c117@gateway/web/freenode/ip.> has joined #yocto12:02
*** gtristan_ <gtristan_!~tristanva@> has quit IRC12:03
rburtonkanavin: no, with master12:06
kanavinrburton: is that related to prelink update patches?12:12
kanavinI rebased my branch on current master before submitting, but test-built it with earlier master12:12
*** lumag <lumag!~lumag@> has joined #yocto12:14
*** lumag <lumag!~lumag@> has left #yocto12:14
*** RyanMeulenkamp <RyanMeulenkamp!> has joined #yocto12:18
*** eduardas_m <eduardas_m!~eduardas@> has joined #yocto12:18
RyanMeulenkampHi guys! Is it possible that there is some kind of bug in NAT/iptables. I'm experiencing poor network stability.12:19
RPkanavin: just seen the prelink thing in my local build. Am reverting Khem's prelink change12:20
RP(to test)12:21
RPkanavin: confirmed that reverting Khem's change but keeping mark's works12:22
RPI think that patch from Khem is wrong and should be reverted12:22
kanavinRP: particularly as it doesn't explain what the 'fix' means12:25
RPkanavin: right12:25
kanavinone commit id is changed to a different commit id...why?12:25
RPkanavin: khem told me it was broken and we needed the fix, I just trusted he was right12:26
kanavinRP: I just sent my bomb as you have seen, sorry it's a bit late, but you know my circumstances :)12:26
zeddii_homeRP: as usual, poking at something small resulted in 4 hours of work .. but I have the qemux86 tiny build .. but I’m at home today. You mentioned some other boards warning, you don’t happen to recall them ? I can’t get at my work IRC log :D12:26
kanavinI took out the more serious upgrades for now - webkit, and the dnf stack has seen a major rewrite... again :/12:27
RPzeddii_home: sorry about that, I know the feeling all too well, you're describing my last few days!12:29
RPzeddii_home: MACHINE=qemux86 shows config warnings with 4.14 too (CONFIG_IRQ_REMAP and CONFIG_X86_X2APIC) and also qemumips with 4.1412:30
RPzeddii_home: that is the summary12:30
zeddii_homelooking at the logs, I found out that all of my symbol resolution library was completely broken, so I had to update it and modify it to a new API. wheee. took until 1:30 AM this morning, so I am not braving the office.12:30
ant_workhello, I can't find googling but there is a problem with qemuppc and python12:30
zeddii_homeI’ll fix those up and get them in my queue today. I have some other pending config changes from others to merge, so I’ll do it all at once.12:31
ant_workkhem, said it is hidden by security flags12:31
zeddii_homeant_work: a build problem or runtime ?12:31
ant_workno, no, it is unknom triple12:32
ant_workfails on configure12:32
zeddii_homeI built qemuppc recently, so this must trigger with some different distro features or layers ?12:33
ant_work| checking for the platform triplet based on compiler characteristics... powerpc-linux-gnu12:34
ant_work| configure: error: internal configure error for the platform triplet, please file a bug report12:34
RPzeddii_home: The autobuilder highlighting warnings as orange builds is good but means we need some cleanup to get to green builds, I do appreciate the help with that!12:34
RPzeddii_home: I already switched tiny to 4.18 btw12:34
ant_workzeddii_home, thus ERROR: Task (/oe/oe-core/meta/recipes-devtools/python/ failed with exit code '1'12:34
ant_workRP: about the warnings in 4.14 LTS, I was chatting with arnd about backports, he said eventually these warnings (since in 4.4 LTS afais) will be fixed12:37
ant_worksnprintf() etc iirc12:37
ant_workzeddii_home, afais poky and khem's autobuilder do use security flags. nodistro not yet12:39
ant_workso just fire a build for qemuppc/nodistro12:39
ant_workI am using muls libc fwiw12:39
ant_workeh musl12:40
ant_workzeddii_home, ah, fwfw I'd have tried qemuppc64 but it is not in oe-core...12:44
*** fdanis_away is now known as fdanis12:44
RPant_work: we're talking about different warnings12:50
zeddii_homeRP: I can see my the audit warning, and my fixed up symbol code works!12:55
zeddii_homeeasy to see what that’s invalid for qemux86 in that output :DS12:55
RPzeddii_home: much easier! :)12:57
rburtonRP: the gdb upgrade is missing checksums13:06
ant_workzeddii_home, looking at, the platform triplet is obviously valid, it must be the $MULTIARCH test failing13:09
ant_workI'll look at it later, no pc here13:09
* zeddii_home nods13:11
*** distsys <distsys!926cc862@gateway/web/freenode/ip.> has joined #yocto13:12
distsysHello Yocto community13:12
distsysI am getting from my customer the kernel, booloader, etc, plus I have my set of drivers as well the application packages i need13:14
distsysnow i want to build my own image from thses componnents using Yocto13:14
distsysie: I dont likt my image to be based on pocky distribution13:15
distsysbut a complete custon one13:15
distsysany idea how to proceed, from where to start... or does Yocto support only building dstors based on Pcky ?13:16
distsysAny suggestion please ?13:16
rburtondistsys: no, poky is an example13:16
rburtonmake your own distro13:16
ant_workzeddii_home, suspect about  tweak-MULTIARCH-for-powerpc-linux-gnuspe.patch13:16
rburtondistsys: you start by setting DISTRO=mydistro in local.conf, and putting whatever you want to change into mydistro.conf13:17
distsysrburton: ok got it. But the thing is that all tutorials are dealing with how to add layers, recipes, etc to exisitng images, but how to buil one from scatch13:18
rburtondistsys: start with an image that is close to what you want, copy it, edit it13:19
kanavindistsys: also, what are you getting from the customer exactly? prebuilt binaries?13:19
distsyshmmm sources as well as prebuolt libraries13:20
distsysok, so now i have to options:13:20
distsys1)start with an empty image and proceed with adding my packages one by one 2)start with and a minimal image (close to what i want)13:22
distsysi am trying to find tutorial about the first option, but couldnt find... I you know one such good tutorial good you point me to the url13:23
RyanMeulenkampdistsys: Start with minimal image and get it to boot13:23
rburtonRP: would you/anyone else object to removing the pid from the bitbake progress info?13:26
rburtonits not like it's actually useful13:26
yoctiNew news from stackoverflow: Quick rebuild of device tree only with Yocto/bitbake? <>13:28
*** Kitsok <Kitsok!~kitsok@2a02:6b8:0:506:d681:d7ff:fe6d:d499> has quit IRC13:28
mcfriskis there an easy way to tell which ptests are failing in yocto releases like sumo?13:32
RPrburton: it was there as when problems occur with the threading, you could tell what was where13:33
rburtonmcfrisk: was asking qa that earlier, it's all manual right now.13:34
rburtonRP: sounds like "no" to me ;)13:34
rburtonmcfrisk: i was saying we should have a way to run the ptests automtically and collate regressions before release13:34
mcfriskrburton: ok. i need to analyze each test failure then. we have a test suite scripting layer sitting above test execution scripts where we can flag mandatory, flaky and skipped tests. then QA manages the suite so that the mandatory set remains always (barring system instabilities) passing and developers can easily get a 0/1 or failed/pass result.13:37
RPrburton: I've actually been considering adding something sensible like elapsed time13:47
rburtonRP: hm yeah you only get that if the task has a progress bar don't you13:48
rburtonno, thats a lie13:48
rburtondo you mean total elapsed time?13:48
rburtonwould be neat in the top line, i've a class that adds a total time for super-basic benchmarking13:49
RPrburton: ah, it depends which output you're talking about :)13:50
rburtoni've patched out the pid output locally13:50
rburtonCurrently  1 running tasks (3761 of 3770)  99% |############################################################################ |0: gdb-8.1.1-r0 do_compile - 165s13:50
rburtonmeant to be a newline in there13:51
rburtonoh why the enforced padding in the Currently %2s runnin tasks13:52
rburtonwhy not just %d13:52
RPrburton: dropped gdb, reverted prelink, added your patches and refired -next13:52
RPloving the new AB stop/start :)13:52
nayfeHi I have a question about WIC. I'm trying to add swupdate with A/B scheme and I was wondering how this WKS works: . It has two "part / --source rootfs" with different labels, so it creates two entries in fstab?13:55
*** Kitsok <Kitsok!~kitsok@2a02:6b8:0:506:d681:d7ff:fe6d:d499> has joined #yocto13:57
*** davenporten <davenporten!8b55c117@gateway/web/freenode/ip.> has quit IRC13:58
*** mystictot <mystictot!~shyam@> has quit IRC14:01
KitsokHi again! Have a questing regarding FILES_${PN}. I have two packages that give me a conflict on do_rootfs stage. /etc/default conflicts between attempted installs of busybox-syslog-1.27.2-r0.armv6 and openrack-1.0-r1.armv6. The last one writes to /etc/default/openrack, the first one to other file. In openresty recipe I indicate this file in FILES_${PN}. What am I doing wrong?14:02
nayfeyou can't have two recipes providing the same file installed in the same time14:04
Kitsoknayfe, but it's not the same file, it's different files in the same directory14:04
nayfeit shoould be install involcation14:04
nayfeinstall permissions14:04
KitsokSo it checks "install"?14:04
KitsokNot FILES_${PN}?14:05
nayfeI mean, if busybox installs /etc/default with 0644 perm and yours with 700 it will conflict14:05
KitsokAha, I see, checking14:05
Kitsokhmmm.... /etc/default permissions are the same14:06
KitsokIs it correct to check the permissions like in this directory? build/tmp/work/armv6-openbmc-linux-gnueabi/busybox/1.27.2-r0/image/etc/default14:07
nayfebusybox installs it with install -d ${D}${sysconfdir}/default14:09
nayfemaybe share your recipe?14:09
*** armpit <armpit!~armpit@2601:202:4180:c33:a9a3:ce31:52a9:bd5c> has joined #yocto14:11
nayfeyour default file is in .../openrack/root?14:16
Kitsokopenrack/files/openrack/root/etc/default/openrack, top openrack is the recipes' directorfy14:17
KitsokAha, I have g+w in all this tree14:17
KitsokFixing permission in the files tree fixed the problem, thank you, nayfe14:20
*** ryansturmer <ryansturmer!> has joined #yocto14:21
nayfenp :)14:21
ryansturmerHello all14:22
ryansturmerI am struggling with something that I suspect is very simple - I have a recipe that I am trying to rebuild from scratch, but I can't seem to provoke it to "clean"14:22
ryansturmerI am trying to clear the sstate cache (which I understand will provoke a re-fetch and build)14:23
Kitsokryansturmer, did you try bitbake -c cleansstate?14:23
ryansturmerso I run bitbake -c cleansstate14:23
ryansturmerexcuse me14:23
ryansturmerbitbake -c cleansstate myrecipe14:23
ryansturmerand then bitbake myrecipe14:24
ryansturmerand bitbake myrecipe doesn't re-run the recipe14:24
ryansturmerit just passes through and says "0 tasks had to be re-run"14:24
Kitsokryansturmer, met this many times :)14:25
Kitsokgive me a minute14:25
ryansturmerBless you, Kitsok14:25
RPkanavin: I squashed a fix for meson into -next14:26
Kitsokryansturmer, what I do and it usually helps is -c unpack -f, then -c compile -f14:28
*** distsys <distsys!926cc862@gateway/web/freenode/ip.> has quit IRC14:28
*** lukma <lukma!> has left #yocto14:33
*** ant_work <ant_work!> has quit IRC14:39
*** AbleBacon <AbleBacon!~AbleBacon@unaffiliated/ablebacon> has joined #yocto14:47
*** JPEW_ <JPEW_!cc4da369@gateway/web/freenode/ip.> has joined #yocto14:50
*** varjag <varjag!> has quit IRC14:52
seebsso, the .NET/pseudo lockups are finally diagnosed, and I understand them, and fixing them is going to be potentially *really* hard.14:56
seebsProblem: Unlike everything else in the entire world, an open of a named pipe, as opposed to any other kind of file, for RDONLY or WRONLY (as opposed to read/write), can block.14:56
seebspseudo assumes that open doesn't block.14:56
RPseebs: ah.14:58
seebsWe could hack in O_NONBLOCK on opens of pipes, or just in general, and it would break some things, because opens of pipes for writing-only would fail instead of blocking. Actually fixing this is hard.14:58
JPEW_seebs: Is it a problem because of locked state in pseudo?14:58
seebsPseudo has locked a mutex because it doesn't want to take interrupts, etcetera, while in the middle of its highly-magical code path.14:59
JPEW_seebs: After you open the pipe with O_NONBLOCK can you poll() or select() after all the state is unlocked?14:59
seebsBasically, there's a whole setup/locking/signal-blocking, etcetera, path which needs to be reversed on exit. I guess we could do a fancy thing where the code inside the handler unblocks everything and releases the mutex, calls the actual syscall, then re-acquires the mutex and resets everything. Maybe.15:00
seebsSure, but you *can't* open the pipe WRONLY with O_NONBLOCK unless a reader's already present.15:00
seebsThe reader can open it with O_NONBLOCK and just block on reads, but the writer will actually *fail*.15:00
seebsOkay, it might be possible to hack this up, but it'd require some special casing. We'd need to restore the signal mask and so on.15:02
seebsAnd actually I think that code may be running in the wrong order.15:02
seebsright now, it's pseudo_sigblock(); pseudo_getlock(); /* actual syscall */ pseudo_droplock(); pseudo_sigunblock();15:03
seebseh. either answer's arguably wrong. we can't actually be sure we controlled the signals if we do them outside the lock, but if we start the lock and then take a signal, that's *also* bad.15:03
*** eduardas_m <eduardas_m!~eduardas@> has quit IRC15:03
seebsanyway, it may be fixable, but it'll be hard. in the named pipe case, we actually don't *care* about blocking much, because the open won't create a file, but we do need to have the lock when we process the open because we need it to get a file descriptor in our table.15:05
seebsso approximate outline of a fix is: we create a template flag or something for "unblock to call the actual syscall", and set that for the open family, possibly only in cases where we know this would apply.15:07
seebsthis causes dropping/reclaiming the lock around the direct syscall in the antimagic case.15:07
seebsthen in the actual internals of open, we have the same thing.15:08
*** joaocfernandes <joaocfernandes!~Joao@> has joined #yocto15:08
seebsthis might or might not break stuff, but it's probably (???) safe.15:08
*** Son_Goku <Son_Goku!~King_InuY@fedora/ngompa> has quit IRC15:09
RPseebs: that sounds like a possibility, messy but could work...15:12
RPseebs: a summary of this into the bug would be very useful btw! :)15:12
RPrburton: do we want an llvm qemu test case added to qa-extras?15:14
*** jacques <jacques!~jacques@nslu2-linux/jacques> has joined #yocto15:17
*** jacques is now known as linuxjacques15:18
ryansturmerKitsock I think my problem may be that I am using a custom fetcher15:20
rburtonRP: yeah i think so15:24
*** RyanMeulenkamp <RyanMeulenkamp!> has quit IRC15:26
*** Kitsok <Kitsok!~kitsok@2a02:6b8:0:506:d681:d7ff:fe6d:d499> has quit IRC15:30
mcfriskhi, I'm getting tired of various tools doing IO to disk while building. image and package managers tend to call fsync()/sync() all the time and slow things down when plenty of RAM is available. I use eatmydata wrapper around bitbake but I'd like to add it to bitbake shell and python tasks too. I was thinking of writing a small recipe from Debian version eatmydata so that it's available for native,15:30
mcfrisknativesdk and target. then maybe hook up to bb run() to prefix all executions with eatmydata. Any better ideas?15:30
*** frsc <frsc!> has quit IRC15:33
*** lusus <lusus!~lusus@> has quit IRC15:35
*** kanavin <kanavin!~kanavin@> has quit IRC15:35
*** boucman_work <boucman_work!~jrosen@wesnoth/developer/boucman> has quit IRC15:47
*** boucman_work <boucman_work!~jrosen@wesnoth/developer/boucman> has joined #yocto15:47
*** varjag <varjag!> has joined #yocto15:55
*** mystictot <mystictot!~shyam@> has joined #yocto16:01
*** linuxjacques <linuxjacques!~jacques@nslu2-linux/jacques> has quit IRC16:02
khemrburton: gdb is fixed in
khemthats the only change on top of master-next16:03
*** mckoan is now known as mckoan|away16:05
kergothmcfrisk: if your wrapper uses LD_PRELOAD, that's not going to work for all tasks, only those which aren't fakeroot/pseudo, as those use LD_PRELOAD already16:05
kergothafaik, anyway16:05
*** chandana73 <chandana73!~ckalluri@> has joined #yocto16:10
mcfriskkergoth: eatmydata works with either LD_PRELOAD which is good for bitbake shell tasks, but it can also work with eatmydata command wrapper which could be used with run() and other functions.16:16
*** awe00 <awe00!~awe00@unaffiliated/awe00> has quit IRC16:16
*** mystictot <mystictot!~shyam@> has quit IRC16:17
*** fdanis is now known as fdanis_away16:21
*** armpit <armpit!~armpit@2601:202:4180:c33:a9a3:ce31:52a9:bd5c> has quit IRC16:23
kergothdoes the command wrapper not just use preload internally? that's what most commands like that do, fakeroot included16:24
*** joaocfernandes <joaocfernandes!~Joao@> has quit IRC16:27
*** mystictot <mystictot!~shyam@> has joined #yocto16:27
*** tgoodwin <tgoodwin!> has joined #yocto16:35
*** fl0v0 <fl0v0!> has quit IRC16:35
*** stephano <stephano!stephano@nat/intel/x-iommzvcrrydyyabw> has joined #yocto16:45
*** mystictot <mystictot!~shyam@> has quit IRC16:46
RPrburton: restarted -next again16:47
khemRP: missed gdb ?17:05
*** tgraydon <tgraydon!~textual@> has joined #yocto17:11
*** jae1 <jae1!~jaewon@> has quit IRC17:31
*** awe00 <awe00!~awe00@unaffiliated/awe00> has joined #yocto17:39
*** arielmr <arielmr!> has joined #yocto17:45
*** jacques <jacques!~jacques@nslu2-linux/jacques> has joined #yocto17:49
*** jacques is now known as linuxjacques17:49
*** WillMiles <WillMiles!> has joined #yocto18:08
*** armpit <armpit!> has joined #yocto18:13
*** T_UNIX <T_UNIX!uid218288@gateway/web/> has quit IRC18:15
*** rburton <rburton!> has quit IRC18:34
*** rburton <rburton!> has joined #yocto18:35
*** georgem_home <georgem_home!uid210681@gateway/web/> has joined #yocto18:37
*** arielmr <arielmr!> has quit IRC18:40
*** rburton <rburton!> has quit IRC18:46
*** rburton <rburton!> has joined #yocto18:47
*** florian <florian!~florian_k@Maemo/community/contributor/florian> has quit IRC19:13
*** florian_kc <florian_kc!~florian_k@Maemo/community/contributor/florian> has joined #yocto19:13
aehs29Thanks for the much need python+pgo patches rburton19:14
*** ntl <ntl!> has joined #yocto19:16
*** ntl <ntl!> has quit IRC19:23
*** awe00 <awe00!~awe00@unaffiliated/awe00> has quit IRC19:41
*** joaocfernandes <joaocfernandes!> has joined #yocto19:44
*** chandana73 <chandana73!~ckalluri@> has quit IRC19:53
*** awe00 <awe00!~awe00@unaffiliated/awe00> has joined #yocto19:55
*** chandana73 <chandana73!~ckalluri@> has joined #yocto19:57
*** joaocfernandes <joaocfernandes!> has quit IRC20:04
*** joaocfernandes <joaocfernandes!> has joined #yocto20:04
*** bluelightning_ <bluelightning_!~paul@pdpc/supporter/professional/bluelightning> has joined #yocto20:04
jdeldo I need to do anything special to to get pkg-config .pc files to be provided by a recipe properly?20:09
*** bluelightning_ is now known as bluelightning20:12
*** joaocfernandes <joaocfernandes!> has quit IRC20:16
aehs29jdel: I believe theres an inherit needed that will take care of that20:18
jdelthere's 'pkgconfig'20:22
jdelbut I think that's a "client" of pkgconfig20:22
jdelie its for a recipe whose configure script wants to use pkgconfig to find libraries20:23
*** ntl <ntl!> has joined #yocto20:23
jdeli *think* i just need to plop the .pc files into place20:23
jdelbut i'm having trouble getting them to populate in another recipe-sysroot directory20:23
*** AbleBacon_ <AbleBacon_!~AbleBacon@unaffiliated/ablebacon> has joined #yocto20:26
*** AbleBacon <AbleBacon!~AbleBacon@unaffiliated/ablebacon> has quit IRC20:29
*** AbleBacon_ is now known as AbleBacon20:29
*** ntl <ntl!> has quit IRC20:30
jdelah, I had RDEPENDS not DEPENDS in my consuming recipe20:45
*** nayfe <nayfe!uid259604@gateway/web/> has quit IRC20:47
RPkhem: yes, sorry. Will try and remember for the next one. There is a PREFERRED_VERSION for gdb we probably should just delete too20:49
*** tgoodwin <tgoodwin!> has quit IRC20:50
*** BCMM <BCMM!~BCMM@unaffiliated/bcmm> has quit IRC20:54
*** ntl <ntl!> has joined #yocto20:57
RPfray: If A.rpm depends on B.rpm and B rrecommends A, would the postinst of A run before B?21:02
*** armpit <armpit!> has quit IRC21:06
*** likewise <likewise!~leon@> has joined #yocto21:08
*** arielmr <arielmr!> has joined #yocto21:08
*** Bunio_FH <Bunio_FH!> has quit IRC21:13
khemRP: in the latest patch on kraj/master I have bumped that pin to 7.2 as well21:17
RPkhem: ok, thanks21:20
*** ant_home <ant_home!> has joined #yocto21:21
likewiseCan I remove my patch from patchwork? I only see the option to archive.21:22
khemdeleting would need an admin21:22
likewisearchive means patch is no longer consider? I screwed up the v2 subject header..21:23
*** WillMiles <WillMiles!> has quit IRC21:24
khemyou can just mark it superceded21:25
*** AbleBacon <AbleBacon!~AbleBacon@unaffiliated/ablebacon> has quit IRC21:25
likewisekhem: tnx21:28
*** varjag <varjag!> has quit IRC21:31
*** varjag <varjag!> has joined #yocto21:31
*** pohly <pohly!> has quit IRC21:35
RPThis looks like an rpm bug :(21:43
RPthat really doesn't help me solve the build/test issues though21:43
*** stephano <stephano!stephano@nat/intel/x-iommzvcrrydyyabw> has quit IRC21:45
*** awe00 <awe00!~awe00@unaffiliated/awe00> has quit IRC21:49
*** dc13ff <dc13ff!uid190567@gateway/web/> has joined #yocto21:49
*** jae1 <jae1!~jaewon@> has joined #yocto22:03
*** varjag <varjag!> has quit IRC22:12
tlwoernershould yocto-builds mailing list be more of an announce-only list? i.e. closed to posting from anyone but the build servers?22:12
tlwoernerhalstead: ^22:13
halsteadtlwoerner: I thought it was for build output.22:14
tlwoernersome people have been asking general build questions there (their confusion is understandable)22:14
halsteadtlwoerner: the list description says it is for discussion too. I don't think I've seen it used that way though.22:15
likewisenite all22:15
*** likewise <likewise!~leon@> has quit IRC22:15
*** gtristan_ <gtristan_!~tristanva@> has joined #yocto22:18
rburtonhalstead: can you change the description so it says output of autobuilder logs?22:32
halsteadrburton: yes. Good idea.22:32
*** gtristan_ <gtristan_!~tristanva@> has quit IRC22:33
RPhmm, removing the sysprof upgrade means its broken, probably due to the glib upgrade22:57
* RP wonders what removing glib will break :/22:57
* RP is going to have to sleep22:57
RPkhem: the sysprof upgrade from kanavin broke sysprof, haven't looked into why but its looking like fixing it may be easier than unentangling other upgrades :(22:58
RPkhem: er, I mean broke on musl23:00
* RP really is going to sleep23:00
aehs29RP: night23:01
rettichschnidiI am wondering if devtool was ever meant to be used on .bbappend files23:33
*** chandana73 <chandana73!~ckalluri@> has quit IRC23:33
rettichschnidiWhen I execute "devtool modify xyz" it always starts off with even when there is a xyz.bbappend23:33
*** armpit <armpit!~armpit@2601:202:4180:c33:7da7:aee9:aa08:ff3c> has joined #yocto23:35
kergothwhat do you mean?23:39
kergothdevtool modify operates on SRC_URI as it is, with appends applied23:39
rburtonRP: got a fix for sysprof23:43
*** ant_home <ant_home!> has quit IRC23:44
*** gnac <gnac!> has joined #yocto23:47

Generated by 2.11.0 by Marius Gedminas - find it at!