Friday, 2019-03-22

*** warthog9 <warthog9!> has quit IRC00:01
*** warthog9 <warthog9!> has joined #yocto00:02
*** learningc <learningc!> has joined #yocto00:37
*** User__ <User__!> has joined #yocto00:40
*** learningc <learningc!> has quit IRC00:45
tzanybody have any idea how to debug init scripts without changing behaviour? I've got a bunch of scripts that seemingly don't execute at startup, but if I add "#!/bin/bash -ex\nexec >> /var/log/$0-init 2>&1" at the top they all work fine :s01:14
kergothsounds like you made them run under bash instead of /bin/sh, which could be an entirely different shell01:15
tzyep, the exec line wasn't working under /bin/sh01:16
tzalthough this is a later iteration, might rollback to it to see if it works01:16
tz(the exec line with sh)01:16
tzthe ones that are failing are things like networking, which shouldn't really be an issue afaik01:16
tzand syslog... which makes things a bit annoying obviously01:17
tzstill works with /bin/sh and the exec line, hmmm01:18
*** BCMM <BCMM!~BCMM@unaffiliated/bcmm> has quit IRC01:25
*** nighty- <nighty-!> has joined #yocto01:31
*** chandana73 <chandana73!~ckalluri@> has quit IRC02:04
*** dv_ <dv_!~dv@> has quit IRC03:20
*** dv_ <dv_!> has joined #yocto03:34
*** armpit <armpit!~armpit@2601:202:4180:c33:d417:1425:3c72:9f78> has quit IRC04:00
*** armpit <armpit!~armpit@2601:202:4180:c33:45a8:35fd:6c1a:9e02> has joined #yocto04:13
*** tardyp <tardyp!sid45259@gateway/web/> has quit IRC04:31
*** ribalda <ribalda!sid306640@gateway/web/> has quit IRC04:31
*** paulbarker <paulbarker!sid269702@gateway/web/> has quit IRC04:31
*** justinsg <justinsg!sid296040@gateway/web/> has quit IRC04:31
*** tardyp <tardyp!sid45259@gateway/web/> has joined #yocto04:31
*** ribalda <ribalda!sid306640@gateway/web/> has joined #yocto04:31
*** fury <fury!uid193779@gateway/web/> has quit IRC04:31
*** justinsg <justinsg!sid296040@gateway/web/> has joined #yocto04:31
*** paulbarker <paulbarker!sid269702@gateway/web/> has joined #yocto04:31
*** dagmcr <dagmcr!sid323878@gateway/web/> has quit IRC04:31
*** ric96 <ric96!sid234506@gateway/web/> has quit IRC04:32
*** behanw <behanw!uid110099@gateway/web/> has quit IRC04:32
*** fury <fury!uid193779@gateway/web/> has joined #yocto04:32
*** hugo___ <hugo___!sid6079@gateway/web/> has quit IRC04:33
*** ndec <ndec!sid219321@linaro/ndec> has quit IRC04:33
*** mirzak <mirzak!sid303002@gateway/web/> has quit IRC04:33
*** Tartarus <Tartarus!sid72705@gateway/web/> has quit IRC04:33
*** learningc <learningc!> has joined #yocto04:34
*** behanw <behanw!uid110099@gateway/web/> has joined #yocto04:34
*** ric96 <ric96!sid234506@gateway/web/> has joined #yocto04:34
*** catch22 <catch22!~catch22__@> has quit IRC04:35
*** ndec <ndec!sid219321@linaro/ndec> has joined #yocto04:36
*** hugo___ <hugo___!sid6079@gateway/web/> has joined #yocto04:36
*** Tartarus <Tartarus!sid72705@gateway/web/> has joined #yocto04:37
*** User__ <User__!> has quit IRC04:37
*** dagmcr <dagmcr!sid323878@gateway/web/> has joined #yocto04:39
*** AndersD <AndersD!> has joined #yocto04:53
*** AndersD_ <AndersD_!~AndersD@> has joined #yocto04:56
*** AndersD <AndersD!> has quit IRC04:59
*** dreyna <dreyna!> has quit IRC05:03
*** AndersD_ <AndersD_!~AndersD@> has quit IRC05:21
*** csanchezdll <csanchezdll!> has quit IRC05:50
*** learningc <learningc!> has quit IRC05:59
*** stdint <stdint!~stdint@> has joined #yocto06:25
stdintI have a problem with using meson in yocto06:26
stdintif the build option is a arrays, I need to use the comma to separate its value06:26
stdintbut I can't do that at PACKAGECONFIG, it would be regarded as next option for PACKAGECONFIG06:27
stdintany idea to avoid this problem?06:27
*** agust <agust!> has joined #yocto07:02
*** mirzak <mirzak!sid303002@gateway/web/> has joined #yocto07:05
*** cvasilak <cvasilak!> has joined #yocto07:07
*** fbre <fbre!91fdde45@gateway/web/freenode/ip.> has joined #yocto07:29
*** rob_w <rob_w!~bob@unaffiliated/rob-w/x-1112029> has joined #yocto07:41
*** jij <jij!jonashg@nat/axis/x-ywkjsjwiwlwvjuzr> has quit IRC07:48
*** fl0v0 <fl0v0!~fvo@> has joined #yocto07:58
*** mckoan|away is now known as mckoan08:20
*** jeanba <jeanba!~jbl@> has joined #yocto08:28
*** jeanba <jeanba!~jbl@> has left #yocto08:28
ndecstdint: check and how --with-dri-drivers or --with-vulkan-drivers are managed.08:32
ndecbasically you use an additional variable for the 'arrays'.08:33
stdintndec, I see I will thank you08:33
ndecPACKAGECONFIG is first parsed before variables it contains are expanded08:34
stdintndec, thank you08:35
*** florian <florian!~florian_k@Maemo/community/contributor/florian> has joined #yocto08:38
*** jij <jij!jonashg@nat/axis/x-rkcqywhmqqmnddqp> has joined #yocto08:39
*** Bunio_FH <Bunio_FH!> has joined #yocto08:48
*** yacar_ <yacar_!~yacar@> has joined #yocto08:54
fbreHi, poky/meta/recipes-core/packagegroups/ contains e.g. the line ${@bb.utils.contains("MACHINE_FEATURES", "phone", "packagegroup-base-phone", "", d)}... and later RDEPENDS_packagegroup-base-phone = " ofono".09:02
fbreMy question is: "phone", "packagegroup-base-phone" and "ofono" - what is each one called in yocto-language? Is it a package or packet or something? I'm not familiar with the proper category naming.09:02
*** fatalhalt <fatalhalt!> has quit IRC09:11
*** gtristan <gtristan!~tristanva@> has joined #yocto09:22
*** sk_tandt <sk_tandt!> has joined #yocto09:26
*** bluelightning <bluelightning!~paul@pdpc/supporter/professional/bluelightning> has quit IRC09:33
*** T_UNIX <T_UNIX!uid218288@gateway/web/> has joined #yocto09:39
*** cquast <cquast!~cquast@> has joined #yocto10:00
*** behanw <behanw!uid110099@gateway/web/> has quit IRC10:00
*** cvasilak <cvasilak!> has quit IRC10:05
fbreWhat is the difference between "phone" and "ofono"?10:11
fbreIs "phone" called "a group of packages" and "ofono" is called "a package"? How is it called?10:12 yocto nomenclature10:13
*** kuzulis <kuzulis!~kuzulis@> has joined #yocto10:13
LetoThe2ndfbre: ofono is a packages, unrelated to the yocto project. its just the name of the software.10:15
LetoThe2ndfbre: see:
LetoThe2ndfbre: the line that you pasted basically says: "if MACHINE_FEATURES contains the string 'phone', then add packagegroup-base-phone to packagegroup-base"10:16
LetoThe2ndfbre: and the latter just says that packagegroup-base-phone includes ofono10:16
kuzulisHi guys. How I can add an 'xrender' dependency to ny qt5 recipe? Because Qt5 requires 'xrender', but I don't know where I need to add this xrender..10:16
*** awe00 <awe00!~awe00@unaffiliated/awe00> has joined #yocto10:17
LetoThe2ndkuzulis: do you actually mean "libxrender"? or rather "xrandr"?10:18
LetoThe2nd(yet, in both cases, you should be fine to add it to your DEPENDS respective RDEPENDS)10:19
kuzulisLetoThe2nd: Seems it is libxrender..10:19
fbreLetoThe2nd: Thanx!! How do you call in yocto something like "packagegroup-base"? Is it a "package family"?10:20
LetoThe2ndfbre: we call it a "packagegroup" (SURPRISE!)10:20
LetoThe2ndkuzulis: well then, add libxrender.10:20
fbreuhmm OK :)10:20
fbreLetoThe2nd: And how do we call something like that "phone"? That is what I actually configure in my .conf file10:21
LetoThe2ndfbre: its a MACHINE_FEATURE10:22
*** jeanba <jeanba!~jbl@> has joined #yocto10:22
fbreah OK, and the other ones are DISTRO_FEATURE then. OK, thanx. I was not sure about the proper classifier naming10:22
fbreThe difficult thing is,  lists just the names of packages but not how they are called as MACHINE_FEATUREs or DISTRO_FEATUREs. So it's hard to track down what string I have to remove with DISTRO_FEATURE_remove to prevent certain packages in my own distro.10:25
fbreFirst, I tried this "DISTRO_FEATURE_remove = "ofono" and failed. :-(10:26
fbreThe documentation does not clearly describe how I get from "ofono" to "phone"10:28
LetoThe2ndfbre: usually i just do bitbake -g $MYIMAGE, and then inspect package-depends.dot10:29
*** jeanba <jeanba!~jbl@> has quit IRC10:29
*** blueness <blueness!~blueness@gentoo/developer/blueness> has quit IRC10:29
LetoThe2ndthere is some oe-scriptwhatevermagicthingy to aid in this too, but i just can't get its name into my head.10:29
fbreLetoThe2nd: Oh I haven't known that! Thanx. Having a look now...10:30
kuzulisLetoThe2nd: Hmm.. I have added: RDEPENDS_${PN}_append = "libxrandr" to my packagegroup-core-x11-base.bbappend, but now, the image recipe fails with: ERROR: Nothing RPROVIDES 'liberation-fontslibxrandr' (but /mnt/data/Yocto-miatech/yocto-miatech/sources/poky/meta/recipes-graphics/packagegroups/ RDEPENDS on or otherwise requires it)10:31
LetoThe2ndfbre: and a package usually does not directly correlate to an expression in DISTRO_FEATURES or MACHINE_FEATURES10:31
LetoThe2ndkuzulis: why do you tinker the packagegroup if your recipe depends on it?10:32
LetoThe2ndkuzulis: and adding a library to RDEPENDS is strange, as you usually already want to have it ready during compile stage.10:32
*** bluelightning <bluelightning!~paul@pdpc/supporter/professional/bluelightning> has joined #yocto10:33
LetoThe2ndkuzulis: and, liberation font are something totally different again.10:33
kuzulisLetoThe2nd: Because I is a full zero in yocto :)10:33
*** jeanba <jeanba!~jbl@> has joined #yocto10:33
LetoThe2ndkuzulis: and you even mixed up things. there is "libxrender" or "xrandr"10:33
kuzulisLetoThe2nd: I just don't understand, where I need to put the RDEPENDS and in what of a content? Should it looks like: RDEPENDS_${PN}_append = "libxrandr" ?10:35
LetoThe2ndkuzulis: no.10:35
LetoThe2ndyou have some qt application, probably. and if that application needs something, then add DEPENDS = "whatever" to the recipe of that application.10:36
LetoThe2ndkuzulis: seriously, do not tinker a packagegroup if you don't know what you're doing.10:37
*** jeanba <jeanba!~jbl@> has quit IRC10:39
kuzulisLetoThe2nd: Just DEPENDS += "libxrandr" ? :)10:40
*** la_croix <la_croix!> has quit IRC10:40
LetoThe2ndkuzulis: if your application needs libxrandr, then yes.10:42
kuzulisLetoThe2nd: A Qt5 itself needs libxrandr for building10:42
kuzulisLetoThe2nd: So, do I need just to add DEPENDS += 'libxrandr' to my qt5-base.bbappend recipe?10:43
LetoThe2ndkuzulis: i seriously doubt that you need to add it manually.10:45
kuzulisLetoThe2nd: Wow, seems it does work! :) Many many thanks :)10:45
*** jeanba <jeanba!~jbl@> has joined #yocto10:45
LetoThe2ndkuzulis: you are probably doing something wrong or unexpected somewhere else.10:46
kuzulisLetoThe2nd: This is probably true, I am doing something wrong .. But I do not have time to figure it out ... I need to do it quickly and to forgot it .. :) Many thanks, again.10:47
*** blueness <blueness!~blueness@gentoo/developer/blueness> has joined #yocto10:49
LetoThe2nd*sigh* in that case, save some time for the day it blow up into your face because you just did it quick and dirty now.10:50
LetoThe2ndbut, have fun.10:50
*** jeanba <jeanba!~jbl@> has quit IRC10:51
kuzulisLetoThe2nd: What if I need to add this option only if a X11 is enabled? Do I need to use it like: DEPENDS[x11] += "libxrandr" ?10:52
kuzulisLetoThe2nd: DEPENDS_append_x11 += "libxrandr" ?10:52
kuzulisLetoThe2nd: Brute force :)10:53
LetoThe2ndpackagegroup-base contains a lot of examples on how to evaluate DISTRO_FEATURES or MACHINE_FEATURES and modify depends accordingly.10:53
LetoThe2ndgo read there.10:53
*** yacar_ <yacar_!~yacar@> has quit IRC10:53
kuzulisLetoThe2nd: Ohh, ok, many thanks.10:53
yoctiNew news from stackoverflow: using require in layer.conf in yocto <>10:58
*** andycooper <andycooper!uid246432@gateway/web/> has quit IRC10:59
*** jeanba <jeanba!~jbl@> has joined #yocto10:59
*** paulbarker <paulbarker!sid269702@gateway/web/> has quit IRC10:59
*** justinsg <justinsg!sid296040@gateway/web/> has quit IRC10:59
*** paulbarker <paulbarker!sid269702@gateway/web/> has joined #yocto10:59
*** dagmcr <dagmcr!sid323878@gateway/web/> has quit IRC11:00
*** justinsg <justinsg!sid296040@gateway/web/> has joined #yocto11:00
*** dagmcr <dagmcr!sid323878@gateway/web/> has joined #yocto11:01
*** andycooper <andycooper!uid246432@gateway/web/> has joined #yocto11:02
*** jeanba <jeanba!~jbl@> has quit IRC11:20
*** armpit <armpit!~armpit@2601:202:4180:c33:45a8:35fd:6c1a:9e02> has quit IRC11:21
*** jeanba <jeanba!~jbl@> has joined #yocto11:21
*** armpit <armpit!~armpit@2601:202:4180:c33:45a8:35fd:6c1a:9e02> has joined #yocto11:23
*** jeanba <jeanba!~jbl@> has quit IRC11:30
*** m4nu3l <m4nu3l!> has joined #yocto11:32
*** berton <berton!~berton@> has joined #yocto11:32
m4nu3lstupid noob question, i want to have /tmp have its own tmpfs (syslog tends to fill up our /var/log). I wonder how I can best override initscripts/volatile (that creates the symlink from /tmp to /var/tmp) and potentially fs-perms.txt (which links from /var/tmp to /var/volatile/tmp)11:32
m4nu3li couldn't really find too much with my google fu except FILESYSTEM_PERMS_TABLE variable, which only does the second part of what i need11:33
*** la_croix <la_croix!> has joined #yocto11:34
LetoThe2ndm4nu3l: by default /var/log should be volatile anyways (see
LetoThe2ndm4nu3l: so if it is not in your case, try to find out what sets VOLATILE_LOG_DIR to "no"11:36
m4nu3lit is volatile11:39
m4nu3lwe limited it to 20 MB because a misbehaving program would quickly fill it up11:39
m4nu3l(and kill the system since it would take half the RAM)11:39
m4nu3lbut now we run into issues where shell scripts fail because /var/volatile/tmp is full and can't create here files11:39
m4nu3ldo i want both /var/log and /tmp to be volatile, but separate partitions11:40
LetoThe2ndm4nu3l: hum ok, so want two separate tmpfs. guess you have to dig deeper in then11:40
LetoThe2ndno specific idea, sorry.11:40
m4nu3lok, so no "configurable" way11:40
m4nu3lat least that you know of11:40
LetoThe2ndnone that i would be aware of.11:40
m4nu3lwon't be too bad i'm sure11:40
*** mort <mort!~mort96@snow/mort96> has joined #yocto11:44
morthey, the pulseaudio recipe inherits from the useradd class, which tries to run the command `groupmems --root /mnt/build/tmp/work/aarch64-poky-linux/pulseaudio/11.1-r0/recipe-sysroot  -g audio -a root -g audio -a root`. That's apparently an invalid command, because it prints the usage information, and pulseaudio's do_prepare_recipe_sysroot fails. Wh11:44
mortat's going on?11:44
*** mihai <mihai!~mihai@unaffiliated/mihai> has joined #yocto11:48
*** BCMM <BCMM!~BCMM@unaffiliated/bcmm> has joined #yocto11:51
*** yacar_ <yacar_!~yacar@> has joined #yocto11:55
*** rhadye_ <rhadye_!sid217449@gateway/web/> has joined #yocto12:06
fbremort: My yocto does not contain a file "recipe-sysroot". Where do you find it in your yocto directory tree "source"?12:11
fbresorry "sources"12:11
mortfbre: bitbake creates the recipe-sysroot folder when building a recipe12:12
fbreand what is pulseaudio recipe then? What is the path and filename in the sources tree?12:14
*** rhadye <rhadye!sid217449@gateway/web/> has quit IRC12:14
*** rhadye_ is now known as rhadye12:14
mortfbre: it's in poky/meta/recipes-multimedia/pulseaudio12:14
*** jwessel <jwessel!~jwessel@> has quit IRC12:16
mortgroupmem is executed by useradd_base.bbclass from useradd.bbclass, which inherits from12:17
fbremort: I've got the yocto version "sumo" and there is not useradd* file12:18
*** berton_ <berton_!~berton@> has joined #yocto12:18
*** JaMa <JaMa!> has joined #yocto12:19
mortI'm on poky commit 29f65bda6d2c9fea4adb125c4857ee64f9312b9f, which is sumo12:19
*** berton <berton!~berton@> has quit IRC12:20
*** vmeson <vmeson!> has quit IRC12:20
fbremort: you said poky/meta/recipes-multimedia/pulseaudio12:23
fbremort: but this is poky/tree/meta/classes/useradd.bbclass12:24
mortyeah, that's where the failing recipe is, but as I said, that recipe inherits from useradd.bbclass12:24
*** bluelightning <bluelightning!~paul@pdpc/supporter/professional/bluelightning> has quit IRC12:25
mortwhen doing the `perform_groupadd "$SYSROOT" "$OPT $opts"` on line 57 of useradd.bbclass, it seems like the opts variable contains " -g audio -a root -g audio -a root", which seems wrong12:25
*** radsquirrel <radsquirrel!> has quit IRC12:27
*** radsquirrel <radsquirrel!> has joined #yocto12:28
fbremort: I would hack it and comment out that line12:29
mortI don't understand where GROUPADD_PARAM is getting set to anything, the pulseaudio recipe certainly isn't setting it to "-g audio -a root -g audio -a root", I can't find anything related in my config, and this build system doesn't exactly have the most debuggable language12:30
fbremort: it seems to me they want to add the user "root" to the group "audio".12:31
mortyeah, but why is that in there twice?12:31
mortit's presumably somewhat important for the user root to be in the group audio, so just commenting it out seems like it would just subtly break our system12:32
mortsorry, it seems to be the perform_groupmems on line 89 which does "-g audio -a root -g audio -a root", not the identical perform_groupmems in the identically looking block on line 5712:33
fbremort: you could do:      EXTRA_USERS_PARAM = "useradd -P mom123 -G audio root"12:34
fbrethis should add root to audio12:35
fbrein recipes-core/images/core-image-minimal.bbappend12:35
mortbut this shouldn't be broken in the first place, and I'm very not comfortable just going around and patching core yocto files just because something seems to be broken12:35
fbrehmm yes. I'm not a yocto guru, and that would just be a hack/workaround12:36
Chaserhello, am a novice yocto user. I am trying to use SSTATE_MIRRORS pointing to - is there a way to tell bitbake to fail it doesn't find a package from yocto. i.e., I want to make sure everything is fetched and not built12:36
fbremort: I don't use audio of sumo, so I don't run into your problem :)12:37
fbresorry, can't help further12:37
LetoThe2ndChaser: and what would be the rationale behind that?12:37
mortI'm not too comfortable with how often yocto seems to just randomly not work, last week I tried upgrading to thud and abandoned that project because thud just literally broke the git submodule fetcher, and now the useradd class and/or pulseaudio recipe just doesn't work12:37
ChaserLetoThe2nd: to  see what packages are missing from shared state. I want to setup private sstate mirror and I want my local builds to fail if it doesn't find pre built package.12:38
*** georgem_ is now known as georgem12:39
LetoThe2ndChaser: hm. not sure if that approach really works. but might indeed be.12:40
ChaserLetoThe2nd: cool12:41
fbremort: You mentioned the git commit of your poky. Do you use the repo tool for sync'ing to git?12:44
*** dv_ <dv_!> has quit IRC12:45
mortfbre: no, wouldn't the regular git commands be enough?12:45
*** gtristan <gtristan!~tristanva@> has quit IRC12:45
fbremort: I just ask because I use the repo tool and I can tell you which git commit I use. I don't have build problems, even if I build everything.12:46
*** kaspter <kaspter!~Instantbi@> has quit IRC12:49
*** kaspter <kaspter!~Instantbi@> has joined #yocto12:50
*** vmeson <vmeson!~rmacleod@> has joined #yocto12:52
fbremort: my poky git revision is 64a257fa22126c4a40ff7e03424a404e360ebe1e12:53
fbremort: everything works there...  (hopefully ;-) )12:53
fbremort: at least compare that with your git revision12:55
fbremaybe the diff is small12:55
*** dv_ <dv_!~dv@> has joined #yocto12:57
mortit seems like the diff is pretty big, my version is from april and yours is from september and a lot happened in those 6 months it seems13:16
*** gtristan <gtristan!~tristanva@> has joined #yocto13:17
*** justanotherboy <justanotherboy!> has joined #yocto13:26
fbremort: this git version I use is what my eval board vendor advises13:26
fbremort: so I think it's a stable one13:26
*** otavio <otavio!~otavio@debian/developer/otavio> has joined #yocto13:33
*** rob_w <rob_w!~bob@unaffiliated/rob-w/x-1112029> has quit IRC13:43
*** nabakov <nabakov!~nabakov@> has quit IRC13:43
*** kaspter <kaspter!~Instantbi@> has quit IRC13:45
*** rburton <rburton!> has joined #yocto14:00
*** kaspter <kaspter!~Instantbi@> has joined #yocto14:11
*** pouet_forever <pouet_forever!~nico@> has joined #yocto14:15
psrcodeRP: little update regarding the lttng-tools ptest stuff, we fidlled with musl a bit before sending patches and turn out musl is having some spuradic issue in getgrnam (will post the backtraces on the musl mailing list and CC you), this lead me to do a cleanup in the tests suite to ensure that we run only test that are not dependent on lttng-ust... slowly getting there.14:16
*** chandana73 <chandana73!~ckalluri@> has joined #yocto14:16
psrcodespuradic -> sporadic...14:17
*** behanw <behanw!uid110099@gateway/web/> has joined #yocto14:19
*** Striking7 <Striking7!> has quit IRC14:27
*** Striking7 <Striking7!> has joined #yocto14:29
RPpsrcode: ah, interesting. The one that is really bothering me is the core-image-sato-sdk hang :/14:29
psrcodeRP: this is fixed, it is caused by this:
RPpsrcode: ah, great. So we'll get that fix when we upgrade lttng-tools?14:31
psrcodeit is not upstream yet but I plan on sending you a recipe update including all the necesseary fixes, we are currently putting a lot of effort on the 2.11 release so all other activity (bugfix/backport) is kind of halted for now.14:33
RPpsrcode: sounds good, I'd be happy to get that fixed with a patch until its upstream. I know you'll work on that when appropriate (I know all about release challenges!)14:34
*** Striking7 <Striking7!> has quit IRC14:36
*** Striking7 <Striking7!> has joined #yocto14:37
*** fbre <fbre!91fdde45@gateway/web/freenode/ip.> has quit IRC14:44
*** awe00 <awe00!~awe00@unaffiliated/awe00> has quit IRC14:52
*** awe00 <awe00!~awe00@unaffiliated/awe00> has joined #yocto14:53
sveinseIsn't FILES_${PN} += "${libdir}/*.so" what you need to put a /usr/lib/ into mypkg? I'm getting QA error on said lib into -dev, but I'm not getting the FILES_${PN} statement to take hold. What could I be missing?14:59
*** stryx` <stryx`!~stryx@unaffiliated/stryx/x-3871776> has quit IRC15:06
*** stryx` <stryx`!~stryx@unaffiliated/stryx/x-3871776> has joined #yocto15:08
kergothsveinse: not enough, no, ${PN}-dev is grabbing it first, as we don't normally package the symlink .so in the mian package15:08
kergothFILES_${PN} += "${libdir}/*.so"15:09
kergothFILES_SOLIBSDEV = ""15:09
kergoththat's how i usually do it15:09
kergothpackage it in ${PN}, and don't package it in -dev, respectively15:09
kergothremember that the first package in PACKAGES to grab a given file gets it, so order in PACKAGES matters15:09
sveinseYeah, this is this discussion about .so not being the actual file, just the symlink and this source puts the binaries in .so. Thanks kergoth15:10
sveinseIt's so nice where there is multiple practices out there :D (or the one half dont care how it is implemented)15:11
kergothheh, welcome to oe/yocto :) where there are 47 ways to do any given task15:11
kergothnatural consequence of the flexibility, really15:12
sveinseTrue :D15:12
*** chandana73 <chandana73!~ckalluri@> has quit IRC15:12
sveinseYou're not a yocto developer unless you have mastered grep and find!15:12
kergothideally need better linting tools to help catching less than ideal approaches, but that's not always easy to do15:12
sveinseAre there any mechanisms for debugging the package selector? as a specific example15:14
*** chandana73 <chandana73!~ckalluri@> has joined #yocto15:15
kergothyou'd need to elaborate. if you mean what binary packages from deploy get installed into the image, after the recipes are built, that depends on the package manager you're using, as we just construct the feeds and let them install them15:15
*** jeanba <jeanba!~jbl@> has joined #yocto15:15
kergothif you mean bitbake's selection of the recipe to build for a given provider, that's a different thing, and -d can add info a bout that, though along iwth a ton of unrelated stuff15:15
sveinseWell, the cause for my FILES_ rule to not work is because the -dev package picked it first. I would love to have a verbose log that sais "found file x: Placed into y-dev because of rule 'y'" or something similar. With it I would have seen that -dev picked up my file and why15:17
sveinsepackage split is the more correct term15:18
kergothhmm, check package.bbclass, that's what does it15:18
kergothmight be something in there, message wise15:18
*** Striking7 <Striking7!> has quit IRC15:19
kergothsee also the log file for do_package, perhaps15:19
sveinsekergoth: thanks, I'll check package.bbclass15:19
*** Striking7 <Striking7!> has joined #yocto15:20
*** stryx` <stryx`!~stryx@unaffiliated/stryx/x-3871776> has quit IRC15:21
*** stryx` <stryx`!~stryx@unaffiliated/stryx/x-3871776> has joined #yocto15:21
*** georgem_home <georgem_home!uid210681@gateway/web/> has quit IRC15:27
*** pouet_forever <pouet_forever!~nico@> has left #yocto15:31
*** jeanba <jeanba!~jbl@> has left #yocto15:45
yoctiNew news from stackoverflow: Qt - Module "QtQuick.Controls" is not installed <>15:59
*** yacar_ <yacar_!~yacar@> has left #yocto16:06
*** khem <khem!~khem@unaffiliated/khem> has quit IRC16:08
*** georgem_home <georgem_home!uid210681@gateway/web/> has joined #yocto16:10
*** Bunio_FH <Bunio_FH!> has quit IRC16:18
rburtonsveinse: top hint is to always remember that FILES are used in the order of PACKAGES16:21
rburtonsveinse: of course, oe-pkgdata-util will show you what files are in what package. then it's usually fairly simple to figure out *why*. bitbake recipe -e, search for FILES_{the package with files you don't expect in}, see why and where the value is set to what it is16:22
*** behanw <behanw!uid110099@gateway/web/> has quit IRC16:29
*** khem <khem!~khem@unaffiliated/khem> has joined #yocto16:30
*** fatalhalt <fatalhalt!> has joined #yocto16:36
*** mihai <mihai!~mihai@unaffiliated/mihai> has quit IRC16:38
*** chandana73 <chandana73!~ckalluri@> has left #yocto16:41
*** nighty- <nighty-!> has quit IRC17:16
*** mckoan is now known as mckoan|away17:18
*** cquast <cquast!~cquast@> has quit IRC17:25
*** awe00 <awe00!~awe00@unaffiliated/awe00> has quit IRC17:26
halsteadRP, armpit rburton Maintenance reboots complete on everything except arm and perf machines. We should be set to queue up builds.17:28
*** BCMM <BCMM!~BCMM@unaffiliated/bcmm> has quit IRC17:33
*** fl0v0 <fl0v0!~fvo@> has quit IRC17:34
*** vmeson <vmeson!~rmacleod@> has quit IRC17:35
RPhalstead: thanks!17:35
*** JaMa <JaMa!> has quit IRC17:36
*** Bunio_FH <Bunio_FH!> has joined #yocto17:38
*** radsquirrel <radsquirrel!> has quit IRC17:41
*** florian <florian!~florian_k@Maemo/community/contributor/florian> has quit IRC17:44
*** radsquirrel <radsquirrel!> has joined #yocto17:44
*** jae1 <jae1!~jaewon@> has quit IRC18:05
*** JaMa <JaMa!> has joined #yocto18:06
*** T_UNIX <T_UNIX!uid218288@gateway/web/> has quit IRC18:17
*** baldgeek <baldgeek!~dan@> has quit IRC18:19
*** baldgeek <baldgeek!~dan@> has joined #yocto18:30
halsteadAll done with the rest.18:39
JPEWarmin: Are you around?18:39
JPEWerr, armpit :)18:39
*** sa2ajj <sa2ajj!> has quit IRC18:50
*** WillMiles <WillMiles!> has joined #yocto18:53
*** awe00 <awe00!~awe00@unaffiliated/awe00> has joined #yocto18:55
*** sk_tandt <sk_tandt!> has quit IRC19:03
sveinseAre there any documented standard of naming "lib*.so.X" and symlink "lib*.so" -> "lib*.so.X" anywhere? Or is this just defacto standard and policy in OE?19:13
kergothit's not an oe convention, it's a linux one19:14
kergothnot just linux, either19:14
*** rcw <rcw!~rcw@> has joined #yocto19:14
kergotheven my mac sets SONAME to include the version and then symlinks without the suffix so ld can find it19:15
sveinsekergoth: yeah, just wondering if its documented anywhere. Because some tools does the oposite. Qt is a notable one19:15
kergothsame thing19:15
kergoththe filename should always match the SONAME, and then any link, if necessary, is to ensure the linker can find it despite the name not ending in .so19:15
kergothbut conventions for how to set soname, that i'm not certain of19:15
kergothnot sure19:15
sveinseWell, Qt does not adhere to this practice or standard. So of cource our Qt devs require proof that this is wrong19:16
kergothnot sure where it comes from historically19:18
sveinseYeah, was reading the latter one too19:19
kergothi know libtool enforces the convention, but not sure where it originated. apparently freebsd differs, or differed?19:22
* sveinse logging into my freeNAS server now to check19:22
sveinse*.so symlink to versioned .so.X files19:23
kergothah, good to know19:24
*** adelcast <adelcast!~adelcast@> has quit IRC19:25
JaMasveinse: where does Qt differ from the rest?19:27
JaMaobjdump -x /usr/lib64/ | grep SONAME SONAME     
JaMaroot@qemux86:~# ls -lah /usr/lib/*19:29
JaMalrwxrwxrwx    1 root     root          19 Mar 21 19:53 /usr/lib/ ->
JaMalrwxrwxrwx    1 root     root          19 Mar 21 19:53 /usr/lib/ ->
JaMa-rwxr-xr-x    1 root     root        6.6M Mar 21 19:52 /usr/lib/
JaMaroot@qemux86:~# objdump -x /usr/lib/ | grep SONAME SONAME     
*** vmeson <vmeson!> has joined #yocto19:30
kergothShared Libraries in SunOS, published 1987, section 4.3.5. Version Management of .so’s mentions the version suffix to the .so, including major and minor versions, though not the symlinking directly, the section on -Bdynamic and -l mentions the -l name not including the suffix, so it implies something has to resolve the mismatch, whether that's a link or gcc or ld itself isn't clear19:31
kergothtried to find Shared Libraries on UNIX System V, but can't seem to locate the paper so far19:32
kergothso anyone writing a project with a .so with an soname lacking a version suffix isn't even up to date with 1987 conventions ;)19:32
sveinseJaMa: you know what? I'm sorry, I'm wrong. In the later Qt5 (5.11 at least) this is now correct. At some point in the past, they changed it. The .so.x.y.z (and the other "lesser") pointed to .so. But now, it is indeed correct where the .so points to the .so.x.y.z as you say and as it should.19:33
JaMasveinse: interesting, in all my builds since Qt 3 I think it was correct19:34
sveinseI remember it clearly, because getting our own Qt app in recipe, involved some .so trixing to get it past the QA tests.19:35
sveinseLet me check back on older released. We've got artifacts from the first product release in 201219:35
*** willie2 <willie2!50d98c5d@gateway/web/freenode/ip.> has joined #yocto19:36
JaMaI've checked image with 5.6.3 and the same19:36
JaMasveinse: were you using meta-qt5 for that product?19:36
JaMaI've started meta-qt5 at the end of 201219:37
willie2What branch would be best to use for rpi3 B with adafruit 3,5' display?19:37
JaMaI don't have any image qith old qt4 or qt3 handy to check19:37
sveinseno, we changed to yocto in 201719:38
JaMamaybe it was your build system messing with the filenames, not Qt itself :)19:38
willie2I'm asking becouse I'm getting different results with krogoth,morty and master and i want to fix the "right" one19:39
sveinseJaMa: perhaps. At least with a different qmake config.19:39
JaMawillie2: these are different releases, see
JaMaeither of them should work to some extend, but if you need to fix something then you should start with master19:40
JaMaand then it can be backported to stable branches if needed19:40
willie2Jama: Okay, Well i can build stuff with krogoth but display settings are not supported as far as iknow19:41
willie2What would be the ideal branch to put some time into?19:42
JaMakrogoth is the oldest from the 3, IIRC it doesn't even have proper support for 3-B19:42
JaMawillie2: master19:42
willie2alright :)19:42
JaMaand if you don't want other layers to change under your hands too much, then use thud19:43
JaMabut then you'll still need to test and submit on master, before it gets accepted to thud19:44
sveinseJaMa, I found one in 4.8.5: ./usr/lib/ ->,  and  ./usr/lib/ -> , and no (symlink). Yet, not quite what I claimed thou19:44
JaMasveinse: that's correct, should be only in -dev package, so it isn't normally installed19:44
sveinseExcept this was from a tgz containing everyting the Qt installer installed19:45
*** jeanba <jeanba!~jbl@> has joined #yocto19:45
*** jeanba <jeanba!~jbl@> has left #yocto19:45
sveinseI'm wondinger if this was I was mistaking this with. We had some problems with this, as you had to specify so version when loading the lib in lack of the .so file19:46
sveinseI have to yield to be honest, I am very uncertain about all this, so I'm withdrawing my claim of Qt not doing this right. At least it is correct in the novel images19:47
sveinsesorry guys19:47
*** georgem_home <georgem_home!uid210681@gateway/web/> has quit IRC19:57
*** lazyape <lazyape!> has joined #yocto20:24
*** jeanba <jeanba!~jbl@> has joined #yocto20:26
*** lazyape <lazyape!> has quit IRC20:29
*** stew-dw <stew-dw!~stew-dw@> has quit IRC20:41
*** JaMa <JaMa!> has quit IRC20:44
*** stew-dw <stew-dw!~stew-dw@> has joined #yocto20:44
armpitJPEW, I am now20:44
*** jeanba <jeanba!~jbl@> has quit IRC20:49
JPEWarmpit: Not trying to cause trouble, but is there an estimate on when thud will get updated from thud-next? Our SDK builds are broken without the target-sdk-provides-dummy changes :(21:06
JPEWTrying to decide if I should wait or cherry-pick them locally21:07
armpitJPEW, I don't. I need to separate my poky changes into there corresponding repos. I was hoping to sort all that out over the weekend21:14
JPEWarmpit: That's fair. Thanks!21:15
*** berton_ <berton_!~berton@> has quit IRC21:17
*** vmeson <vmeson!> has quit IRC21:31
*** rburton <rburton!> has quit IRC21:34
*** georgem_home <georgem_home!uid210681@gateway/web/> has joined #yocto21:38
*** willie2 <willie2!50d98c5d@gateway/web/freenode/ip.> has quit IRC21:38
*** rcw <rcw!~rcw@> has quit IRC21:39
*** justanotherboy <justanotherboy!> has quit IRC22:02
*** WillMiles <WillMiles!> has quit IRC22:07
*** m4nu3l <m4nu3l!> has quit IRC22:40
*** vmeson <vmeson!> has joined #yocto22:48
*** agust <agust!> has quit IRC23:04
MarexRP: hey uh, does wic have some option to pad a rawcopy image with 0xff instead of 0x00 ?23:17

Generated by 2.11.0 by Marius Gedminas - find it at!